A User’s Handbook
Second Edition

A User’s Handbook
Second Edition

Charles M. Hansen

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2007 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 0-8493-7248-8 (Hardcover) International Standard Book Number-13: 978-0-8493-7248-3 (Hardcover) This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Hansen solubility parameters : a user’s handbook. -- 2nd ed. / edited by Charles Hansen. p. cm. Rev. ed. of: Hansen solubility parameters / Charles M. Hansen. c2000. Includes bibliographical references and index. ISBN 0-8493-7248-8 (alk. paper) 1. Solution (Chemistry) 2. Polymers--Solubility. 3. Thin films. I. Hansen, Charles M. II. Hansen, Charles M. Hansen solubility parameters. QD543.H258 2007 547’.70454--dc22 Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com 2006051083

Dr. John Durkee Consultant in Critical and Metal Cleaning Hunt, Texas U.S.A. Dr. techn. Charles M. Hansen Consultant Hoersholm, Denmark Prof. Georgios M. Kontogeorgis Technical University of Denmark Department of Chemical Engineering Lyngby, Denmark Prof. Costas P anayiotou Department of Chemical Engineering University of Thessaloniki Thessaloniki, Greece Tim S. P oulsen Sr. Research Scientist Molecular Pathology Glostrup, Denmark Dr. rer. nat. Hanno Priebe Sr. Research Scientist Chemical Development – Process Research GE Healthcare Amersham Health AS Oslo, Norway Per Redelius Research Manager Nynas Bitumen Product Technology Nynashamn, Sweden Prof. Laurie L. Williams Department of Ph ysics & Engineering Fort Lewis College Durango, Colorado U.S.A.

Preface to the First Edition
My w ork with solv ents started in Denmark in 1962 when I w as a graduate student. The major results of this w ork were the realization that polymer film formation by sol ent evaporation took place in two distinct phases and the de velopment of what has come to be called Hansen solubility (or cohesion) parameters, abbre viated in the follo wing by HSP. The first phase of film formati by solvent evaporation is controlled by surf ace phenomena such as solv ent vapor pressure, wind velocity, heat transfer, etc., and the second phase is controlled by concentration-dependent dif fusion of solvent molecules from within the film to the air sur ace. It is not controlled by the binding of solvent molecules to polymer molecules by h ydrogen bonding as w as pre viously thought. My solubility parameter work was actually started to define a finities between sol ent and polymer to help predict the de gree of this binding which w as thought to control solv ent retention. This was clearly a futile endeavor as there was absolutely no correlation. The solvents with smaller and more linear molecular structure dif fused out of the films more quickly than those with la ger and more branched molecular structure. HSP were de veloped in the process, ho wever. HSP ha ve been used widely since 1967 to accomplish correlations and to mak e systematic comparisons which one would not have thought possible earlier. The effects of hydrogen bonding, for e xample, are accounted for quantitati vely. Man y of these correlations are discussed later , including polymer solubility , swelling, and permeation; surf ace wetting and de wetting; solubility of inorganic salts; and biological applications including w ood, cholesterol, etc. The experimental limits on this seemingly universal ability to predict molecular affinities are apparently g verned by the limits represented by energies of the liquid test solvents themselves. There had/has to be a more satisfactory explanation of this uni versality than just “semiempirical” correlations. I decided to try to collect my e xperience for the purpose of a reference book, both for myself and for others. At the same time, a search of the major theories of polymer solution thermodynamics was undertak en to e xplore how the approaches compared. A k ey element in this w as to e xplain why the correlations all seemed to fit with an apparently “un versal” 4 (or 0.25 depending on which reference is used). This is described in more detail in Chapter 2 (Equation 2.5 and Equation 2.6). My present view is that the “4” is the result of the v alidity of the geometric mean rule to describe not only dispersion interactions but also permanent dipole–permanent dipole and hydrogen bonding (electron interchange) interactions in mixtures of unlik e molecules. The Hildebrand approach uses this and w as the basis of my earliest approach. The Prigogine corresponding states theory yields the “4” in the appropriate manner when the geometric mean rule is adopted (Chapter 2, Equation 2.11). Any other kind of averaging gives the wrong result. Considering these f acts and the massive amount of data that has been correlated using the “4” in the follo wing, it appears pro ven beyond a reasonable doubt that the geometric mean assumption is v alid not only for dispersion-type interactions (or perhaps more correctly in the present conte xt those interactions typical of aliphatic hydrocarbons) but also for permanent dipole–permanent dipole and h ydrogen bonding as well. For those who wish to try to understand the Prigogine theory , I recommend starting with an article by Donald P atterson.1 This article e xplains the corresponding states/free v olume theory of Prigogine and co workers in a much simpler form than in the original source. P atterson2 has also reviewed in understandable language the progression of developments in polymer solution thermodynamics from the Flory–Huggins theory, through that of Prigogine and coworkers, to the so-called “New Flory Theory.”3 Patterson also has been so kind as to aid me in the representations of the earlier theories as the y are presented here (especially Chapter 2). All of the pre vious theories and their extensions also can be found in a more recent book. 4 For this reason, these more classical

theories are not treated extensively as such in this book.The striking aspect about all of this pre vious work is that no one has dared to enter into the topic of h ydrogen bonding. The present quantitative treatment of permanent dipole–permanent dipole interactions and h ydrogen bonding is central to the results reported in e very chapter in this book. An attempt to relate this back to the pre vious theories is gi ven briefly here and more xtensively in Chapter 2. This attempt has been directed through Patterson,1 which may be called the Prigogine–Patterson approach, rather than through the Flory theory, as the relations with the former are more ob vious. I strongly recommend that studies be undertak en to confirm the usefulness of the “structura parameters” in the Prigogine theory (or the Flory theory). It is recognized that the effects of solvent molecular size, segment size, and polymer molecular size (and shapes) are not fully accounted for at the present time. There is hope that this can be done with structural parameters. The material presented here corresponds to my kno wledge and experience at the time of writing, with all due respect to confidentiality agreements, etc I am greatly indebted to man y colleagues and supporters who ha ve understood that at times one can be so preoccupied and lost in deep thought that the present just seems not to e xist. Charles M. Hansen October 19, 1998

1. Patterson, D., Role of Free Volume Changes in Polymer Solution Thermodynamics, J. Polym. Sci. Part C, 16, 3379–3389, 1968. 2. Patterson, D., Free Volume and Polymer Solubility . A Qualitati ve View, Macromolecules, 2(6), 672–677, 1969. 3. Flory, P. J., Thermodynamics of Polymer Solutions , Discussions of the Faraday Society, 49, 7–29, 1970. 4. Lipatov, Y. S. and Nestero v, A. E., Polymer Thermodynamics Library, Vol. 1, Thermodynamics of Polymer Blends, Technomic Publishing Co., Inc., Lancaster , PA, 1997.

Preface to the Second Edition
When the question about a second edition of this handbook w as posed, I w as not in doubt that several additional authors were necessary to meet the demands it w ould require. The writings of the fi e contributors that were chosen speak for themselv es. There is theoretical impact in Chapter 3 (Costas Panayiotou) and in Chapter 4 (Geor gios M. Kontogeorgis). Chapter 3 introduces statistical thermodynamics to confirm the d vision of cohesi ve ener gy into three parts enabling separate calculation of each. Chapter 4 describes ho w the Hansen solubility parameters (HSP) fit into othe theories of polymer solutions. The practical applications and understanding pro vided in Chapter 9 (Per Redelius) related to asphalt, bitumen, and crude oil should accelerate new thinking in this area and emphasize that simple e xplanations of seemingly comple x phenomena are usually the right ones. The thermodynamic treatment of carbon dioxide gi ven in Chapter 10 (Laurie L. Williams) is a model for similar w ork with other g ases and emphatically confirms the usefulness of Hanse solubility parameters for predicting the solubility beha vior of g ases in liquids and therefore also in polymers. Chapter 11 (John Durkee) goes through the process of demonstrating ho w “designer” solvents can be used in cleaning operations to replace, or partly replace, ozone-depleting solv ents, in spite of the problem of their HSP not being suf ficiently close to the HSP of the soils that are to be rem ved. I have added two chapters because of apparent need. Chapter 14 discusses environmental stress cracking (ESC). ESC is a major cause of unexpected and sometimes catastrophic failure of plastics. The recent impro ved understanding pro vided by HSP seemed appropriate for inclusion in this context. Chapter 16 discusses absorption and dif fusion in polymers. Many of the HSP correlations presented in this handbook cannot stand on HSP alone but must include consideration of absorption and diffusion of chemicals in polymers. These effects are often disguised by use of a molecular volume, as molecular size/volume correlates reasonably well with diffusion coefficients, especiall at low concentrations. Polymer surface layers are often significantly diferent from the bulk polymer. Surface mobility of polymer chain se gments plays an important role in surf ace dewetting, ESC, and resistance and/or delays to the absorption of chemicals. This chapter tries to unify the ef fects of a v erifiable sur ace resistance and v erifiable concentration-dependent di fusion coef ficients Solutions to the diffusion equation simultaneously considering these two effects explain the “anomalies” of absorption and also correctly model desorption phenomena, including the drying of a lacquer film from start to finis Each of the chapters in the first edition has been r viewed and added to where this w as felt appropriate without increasing the number of pages unduly . There is still a lack of significan activity in the biological area, in controlled release applications, and in other areas discussed in Chapter 18, such as nanotechnology . The relative affinity of molecules or s gments of molecules for each other can be predicted and in man y cases controlled in self-assembly with the understanding provided by HSP. Chapter 15 treating biological materials has been e xpanded more than the others included in the first edition. This was done with the help of Tim Svenstrup Poulsen. Perhaps the most surprising of the additions in Chapter 15 is a HSP correlation for the (nonco valent) solvent interactions with DNA. The δD;δP;δH values of 19.0;20.0;11.0 for DN A, all in MP a1/2, clearly sho w that h ydrogen bonding interactions (H) contrib ute much less to the nonco valent interactions that determine the structure of the DN A than the dispersion (D) and dipolar interactions (P). Only about 14% of the cohesion energy involved in what is commonly called “h ydrogen bonding” derives from hydrogen bonding.

then rethink.2 is not greatly expanded. Let us hope others will take up the ef fort and relate their findings for the benefit of al Charles M. In view of the large expansion in the number of pages over the firs edition it is hoped that the principles. The latter is due to the generous help of Hanno Priebe. are well illuminated. The majority of my work as a consultant has usually in volved agreements that prohibit or se verely limit publication of results paid for by pri vate sources. It has been satisfying to see ho w much can be interpreted with v ery simple observ ations and calculations.Table Appendix A. I want to once more thank those who have contributed to this second edition. Table Appendix A. both theoretical and practical. I ha ve regularly found this more specific data of considerable interest Once more resources and timing ha ve not been conduci ve to do a complete literature search to provide additional explanations of phenomena that should have had Hansen solubility parameters included in their interpretation. There have been too man y restrictions on what may be published to allo w any major e xpansion of this table.1 is greatly e xpanded both in number and in information. F or those who still lack information in a gi ven situation I can suggest a search using the k ey words “Hansen solubility parameters” followed by additional k ey words as required. the e xtent of which is clearly evident for those familiar with the first edition. However. There are close to 1200 entries in this table vs. This is true both for Internet searches as well as for searches in the more traditional literature. the approximately 860 in the firs edition. If it cannot be done simply . I have also included Appendix A. please be advised that most of these are calculated and not e xperimental values as indicated in the comments to the table. Hansen .3 with the original solubility data on which the di vision of the ener gy was based.

He is currently vice president of the Danish Society for Polymer Technology. he held leading positions with PPG Industries in Pittsburgh. also confirms the viability of the d vision of the cohesion energy into separate parts. Dr. 22 kilometers north of Copenhagen. sur ace science. He frequently re views papers for leading journals. Broendby. These have been found mutually confirming with the I. for the 17 years prior to the start of the current state of semi-retirement. which is reported in Chapter 3 of this second edition. and coatings science. Dr . Denmark. Hansen has published widely in the fields of polymer solubilit . After being awarded the Dr. de gree from the Technical University of Denmark in 1967. Hansen consults on the topics co vered by this book. and as director of the Scandina vian P aint and Printing Ink Research Institute in Hoersholm. He works from his home in Hoersholm. Hansen is perhaps best kno wn for his extension of the Hildebrand solubility parameter to what are no w called Hansen solubility parameters. and allo ws their independent calculation. Denmark. Hansen dealt with polymers at FORCE Technology. Prigogine corresponding states theor of polymer solutions and can be used to directly calculate the Flory–Huggins interaction coef ficient The statistical thermodynamics approach de veloped by Costas P anayiotou and co workers. Denmark. having recently completed a 5-year period as president. techn. and is on the editorial board of Progress in Organic Coatings. as well as being a member of the DanishAcademy of Technical Sciences (ATV). .The Author Charles M. Dr. He received a BChE from the Uni versity of Louisville and an MS de gree from the Uni versity of Wisconsin. diffusion and permeation in polymers and films.


Equation 1. de grees K Critical temperature.5 Ostwald coefficient in Equation 10. Equation 2. Equation 2. Figures 13. respecti vely. Equation 1. see Chapter 2.5 Distance in Hansen space.10) Solubility coefficient in Chapter 1 Molar entropy of mixing Absolute temperature “Total” number of solv ents used in a correlation as gi ven in tables (Normal) boiling point. Chapter 1.7 Maximum distance in Hansen space allo wing solubility (or other “good” interaction) Radius of interaction sphere in Hansen space Relative energy difference (Chapter 1.987 cal/mol K) Distance in Hansen space. atm.Key to Symbols Note: The symbols used in Chapters 3 and 16 are so numerous and dif ferent that the y have been placed in these chapters. de grees K Reduced temperature. Permeation coefficient in Chapter 1 Polar cohesion (solubility) parameter — in tables and computer printouts Pressure in Chapter 10 Solvent quality number Total pressure.1 Diffusion coefficient in Chapter1 Dispersion cohesion (solubility) parameter — in tables and computer printouts Dipole moment — debyes Dispersion cohesion ener gy Polar cohesion ener gy Hydrogen bonding cohesion ener gy Energy of vaporization (=) cohesion ener gy Number of “good” solv ents in a correlation. Equation 1. cm 3/gram molecular weight . (Chapter 13.9 or Chapter 2. Equation 2. used in tables of correlations Gibbs Energy in Chapter 4 Molar free ener gy of mixing Noncombinatorial molar free ener gy of mixing Hydrogen bonding cohesion (solubility) parameter — in tables and computer printouts Molar heat of v aporization Molar heat of mixing Henry’s law constant in Equation 10. A12 D D DM ED EP EH ΔEv G G ΔGM ΔGMnoncomb H ΔHv ΔHM KH L P P P Q P* R Ra RA RM Ro RED S ΔSM T T Tb Tc Tr V Energy difference defined by Chapter 2.5) Gas constant (1.4 and 13.12 Molar volume. see Chapter 1.

2 van der Waals volume Volume of mixture Constant in van der Waals equation of state (Chapter 4) Activity coefficient of the “i”th component in Appendix 10. and Chapter 10) Mole fraction in vapor phase (Chapter 13.V Vf V* VW VM a ai bi b c c ci f fi fi0 i k k n n nD p pi pis r r ts x y H Ω ΩI ∞ Σ ΔT α α β β δD δH δP δt δ ε ε γ γ Total volume in Chapter 4 Free volume (Equation 4. Equation 8.1 Compressibility in Chapter 10 Dispersion cohesion (solubility) parameter Hydrogen bonding cohesion (solubility) parameter Polar cohesion (solubility) parameter Total (Hildebrand) cohesion (solubility) parameter Prigogine normalized interaction parameter .A.A.1 Coefficients in Equations 10.1 Number of se gments in a gi ven molecule.A. Equation 7.1 Constant in van der Waals equation of state (Chapter 4) Dispersion cohesion ener gy density from Chapter 1.1 Mole fraction in liquid phase (Chapter 13.2 Index of refraction in Equation 10. Figures 13.21–10. 10.1 Component “i” in a mixture Constant in Equation 6. defined by Chapter 5.4 Coefficients (state constants) in Equations 10.17 and 10.2) Hard core or close pack ed volume in Equation 4.1 Fugacity at standard state in Appendix 10. and Chapter 10) Ratio of cohesive energy densities.A.25 Partial pressure (of carbon dioxide) in Chapter 10 Partial pressure of the “i”th component in Appendix 10. Chapter 2.8 Cohesive energy for a polymer se gment or solv ent in Chapter 2 Dielectric constant in Equation 10.1 to 5.1 Coefficient in Equaitons 10.6 Infinite dilution act vity coefficien Summation Lydersen critical temperature group contrib ution Thermal expansion coefficien Constant in Equation 4.1 Constant in Equations 10.15 Constant in Chapter 2. Figures 13.1 Saturation pressure of the “i”th component in Appendix 10.22. Chapter 2.1 Fractional solubility parameters.A.4 and 13.23 Coefficient in Equation 10.2 or Figure 1.17 and 10.5.5. Fugacity of the “i”th component in Appendix 10. Figure 1. Chapter 2 Ratio of polymer v olume to solv ent volume (Chapter 4) Sedimentation time. Equation 2. Equations 5.6 Bunsen coefficient (Equation 10.21. see Chapter 7. Equation 2. and 10.25 Surface free ener gy of a liquid in air or its o wn vapor Activity coefficient in Chapter .4 and 13.3 Concentration in Chapter 8. Equation 2.

Chapter 2. Chapter 2 Representative χ value from general literature Entropy component of χ (Subscript) indicates a solv ent (Subscript) indicates a polymer (or second material in contact with a solv ent) (Subscript) dispersion component (Subscript) polar component (Subscript) hydrogen bonding component (Subscript) dispersion component (Subscript) polar component (Subscript) hydrogen bonding component .4 Normalized intrinsic viscosity Volume fraction of component “i” Dipole moment Interaction parameter. Equation 2. Chapter 7. Equation 7.10 Density in Chapter 7. Equation 8.11 Contact angle between liquid and surf ace Advancing contact angle Receding contact angle Prigogine parameter for dif ferences is size in polymer se gments and solv ent. Equation 7.1 Viscosity of solution Viscosity of solv ent Intrinsic viscosity. Chapter 2 Interaction parameter — “Ne w Flory Theory” Critical polymer–liquid interaction parameter . Equation 2. Equation 2. Equation 7. Chapter 2.10 Polymer–liquid interaction parameter (Flory–Huggins).1 Solvent density in Chapter 7. Equation 7.1 Prigogine segmental distance parameter. see Chapter 8. see Chapter 2.η ηs ηo [η] [η]N ϕi μ ν Θ Θa Θr ρ ρ ρ ρp ρs σ χ χ12 χc χlit χs 1 2 D P H d p h Viscosity of solv ent.1 Density in Chapter 10 Particle density in Chapter 7.


................................................................................6 Calculation of the Dispersion Solubility P arameter δD ...................................................................................................................39 General Discussion ................................................................................................................................................. Dipolar................45 ..............................................................................................................................................17 Supplementary Calculations and Procedures ..................................................... and Hansen Solubility P arameters ................................................................................................18 Some Special Effects Temperature Changes ..................................................................................................................................................30 The χ12 Parameter and Hansen Solubility P arameters..........................................................................................1 Hildebrand Parameters and Basic Polymer Solution Thermodynamics ...............................................................................................................Table of Contents Chapter 1 Solubility Parameters — An Introduction ...................................................................................................................................27 Hansen Solubility P arameters (HSP) .....................................................................................35 Polyisobutylene........................28 Resemblance between Predictions of Hansen Solubility P arameters and Corresponding States Theories.......................................................................................................................................................................................... and Dispersion Solubility P arameters............................................................................................................................................................................................................................................................................................................................................27 Abstract ................................................................................................................................39 Postscript ....................39 Polyacrylonitrile .........................36 Polystyrene ......................................................19 Computer Programs ...........................45 Abstract ................19 Effects of Solv ent Molecular Size ...............................................................................................................................................................................................1 Introduction .....................................................................2 Hansen Solubility P arameters .............................................................21 Conclusion............................................34 Polybutadiene .....................32 Comparison of Calculated and Experimental χ12 Parameters ...................................................................................................4 Methods and Problems in the Determination of P artial Solubility P arameters ..................................................................................................................45 Introduction ...............................................27 Introduction ... χ12 Interaction Parameter...41 References ........................................................................................................38 Polyvinylacetate..........................................................................................................................................................................................24 Chapter 2 Theory — The Prigogine Corresponding States Theory................................................................................................................................................42 Chapter 3 Statistical Thermodynamic Calculations of the Hydrogen Bonding....................................................22 References ....................................................13 Calculation of the Polar Solubility P arameter δP .......................................................................................................................................................16 Calculation of the Hydrogen Bonding Solubility P arameter δH ............................................................................................................................................................................................................................................................................20 Hansen Solubility P arameters for Water ...17 Temperature Dependence ...................................45 Key words .........................................................40 Conclusion..........1 Abstract .............................................................

.........46 The Equation-of-State Framework............................................50 Applications ................90 List of Abbreviations................................................................................................97 Solubility — Examples ...................................................................................95 Abstract ...................................................................................................................................110 ...........................................95 Calculation of Polymer HSP ............66 Chapter 4 The Hansen Solubility P arameters (HSP) in Thermodynamic Models for Polymer Solutions ......107 Other Measurement Techniques ...................77 The Entropic Model .....................................................................................85 Mixed Solvent–Polymer Phase Equilibria ......................................................................................................................................................................................75 Abstract ....................................................................................................................................................................................................................................................................................................82 Flory–Huggins Models Using Hildebrand and Hansen Solubility P arameters (HSP) ..........................................................................................................................64 Appendix 3..........................................................................................................................106 Melting Point Determinations — Ef fect of Temperature..........................................................................106 Environmental Stress Cracking ............................................................76 The Group-Contribution Principle and Some Applications (Density......82 The FH/Hansen Model vs.............109 References .................................................................................................................46 The Contribution from Dipolar F orces ........................................................66 Appendix 3................. Unifac-FV) ...................................................................77 The UNIFAC-FV Model ..............95 Introduction ..............................................................................................63 References ......................................................................................................................................................................................................................................................84 Applications.I: The Acid Dimerization ...........................52 Discussion and Conclusions .....................................................80 Rules of Thumb and Solv ent Selection Using the Flory–Huggins Model and Solubility Parameters ........................................................................................................................59 Acknowledgments ............................................................................................................................................................................................................................................................................................................................................................................................92 Appendix 4.....................................................................................................................III: A Group-Contribution Method for the Prediction of δ and δD......................................................................................................................93 Chapter 5 Methods of Characterization — Polymers .............................................................................77 The Free-Volume Concept ................................65 Appendix 3.............................................................................................................................................107 Intrinsic Viscosity Measurements .............................................................................................................................................................................92 References ............Theory ...............................................................................................91 Symbols in this Chapter ................................................81 Activity Coefficients Models Using the HS ...............88 Conclusions and Future Challenges ..........................................II: An Alternative Form of the Polar Term....................... Solubility Parameters) ............................109 Conclusion.............................................................................85 Solvent Selection for P aints (Activity Coefficients at Infinite Dilutio .............................................................................................................................................................98 Swelling — Examples .................................................62 List of Symbols Special to this Chapter .................................75 Group Contribution Methods for Estimating Properties of Polymers ..................................................................................................................................76 GC Free-Volume-Based Models for Polymers (Entropic-FV ....................I: An Expression of the Flory–Huggins Model for Multicomponent Mixtures ...................................78 The Flory–Huggins Model and the Re gular Solution Theory ..... the GC Methods ............................

.....................122 Chapter 7 Methods of Characterization for Pigments...............................................................................151 Models of Bitumen ................................................................................................................................................................................................... and Fiber Surf aces ...............................................125 Introduction ..............................................................................................................................................113 Method to Evaluate the Cohesion Ener gy Parameters for Surf aces.............................................................................................................................134 References ......... Bitumen..................156 Hansen Solubility P arameters (HSP) ........................................................................................................................................................................................................................................................134 Chapter 8 Applications — Coatings and Other Filled Polymer Systems ...................................................................................................................................................................................................................................................................................................122 References ...............................................114 A Critical View of the Critical Surf ace Tensions..........158 ...................................................................................................................................................................................156 Hildebrand Solubility P arameters ............................113 Introduction ....................145 Hansen Solubility P arameter Principles Applied to Understanding Other Filled Polymer Systems ..........................................................................................148 Chapter 9 Hansen Solubility P arameters of Asphalt...................................137 Introduction .......................................................................................................................................................................................Chapter 6 Methods of Characterization — Surf aces.........................................117 Additional Hansen Solubility P arameter Surface Characterizations and Comparisons ............. Fillers...................................................................................................................................................................................................... Filler ................................................................................................................................154 Polarity.............................................125 Abstract ..........................................................................................132 Conclusion.......................................................................154 Molecular Weight ..................................................................................................................................................................................... and Crude Oils .......................................................................................................151 Introduction .........................................................................................................................................................................144 Polymer Compatibility ................................................................................................................... and Fibers ..................................................................131 Controlled Adsorption (Self-Assembly) .................................................................................................................................................................116 A Critical View of the Wetting Tension ...........................................................................118 Self-Stratifying Coatings...............................................................................................................155 Testing of Bitumen Solubility ..................................................142 Solvents and Surf ace Phenomena in Coatings (Self-Assembly) ........................151 Symbols Special to Chapter 9 ......................................................113 Hansen Solubility P arameter Correlations with Surf ace Tension (Surface Free Ener gy)..... Fillers.................................................................................................................................... and Fibers ...........................................................152 Asphaltenes ..122 Conclusion..................................137 Solvents ..............................131 Carbon Fiber Surf ace Characterization ................................................................................................................155 Solubility Parameters of Bitumen ...............................................................................................................................................147 References ...........................137 Techniques for Data Treatment.................................................................120 Maximizing Physical Adhesion ...........................................147 Conclusion.................137 Abstract ............126 Discussion — Pigments...................................127 Hansen Solubility P arameter Correlation of Zeta Potential for Blanc Fix e................................................................................................................151 Abstract .....................................125 Methods to Characterize Pigment..............113 Abstract .

.............227 ........212 Analysis of Capability of Designer Solv ents .....................................213 Conclusions ...............................................................................................204 Method for Calculating HSP of Composites (Soils or Solv ents) .............................................204 Pathology of Soils ..........................................................................................................................................................................................................................................169 Turbidimetric Titrations .....................................................................................................206 Reference Soils for Comparison ..................204 HSP of Multiple-Component Soils ........................177 Introduction ........................................................203 Introduction .............................................................................................................................................................................................................................................................................................191 Conclusion..........................................199 References .........................................................................................................................199 Ideal Solubility of Gases in Liquids .......................................................................................................................203 A Variety of Solv ents...........................................................................................................................................208 Identification of Designer Sol ents ......................................................................................................................................................197 Appendix 10...............................................................................205 More Realistic View about Evaluating HSP of Composite Soils ...................................................................................................................................................................................................197 References ........................177 Abstract ..............201 Chapter 11 Use of Hansen Solubility P arameters to Identify Cleaning Applications for “Designer” Solvents .....................................203 Abstract ...............................................................................................................................................................177 Methodology .................................................................................................170 BISOM Test ..........................206 Method for Choice of Suitable Solv ents ........................................................................................................187 Three-Component (Hansen) Solubility P arameters — Pure CO 2 ........................................................................................................164 Bitumen and Polymers .................................................................................................................................................215 Notes .....................................................159 Computer Program for Calculation and Plotting of the Hansen 3D Pseudosphere ............................................................................................................................173 References .............................................................196 Chapter 10 Addendum ..................................................................196 Acknowledgments ..................................The Solubility Sphere ................................................................................208 An Open Question — Answered ......................................A........................................................................................................................................................................................................................................189 Temperature and Pressure Ef fects on HSPs: δd...........................................................................................................................................................................................166 Crude Oil ................................................................................................................................208 Limiting R A Value for Expected Good Cleaning Performance ...................................................................................................................1: Ideal Solubility of Gases in Liquids and Published CO 2 Solubility Data ............................................................................................................196 Symbols Special to this Chapter ........................................170 Conclusion....................................................178 One-Component Hildebrand P arameter as a Function of Temperature and Pressure ..................................................................................................................................................................................................................................................................................................................191 Temperature and Pressure Ef fects on HSPs: δh......190 Temperature and Pressure Ef fects on HSPs: δp..............161 Components of Bitumen ..................210 Application of HSP Methodology to Cleaning Operations ...................174 Chapter 10 Determination of Hansen Solubility P arameter Values for Carbon Dioxide ......................................................

......................................................................................................................................248 Solubility Parameter Correlation of Polymer Swelling ..........................................................................................................................................................................238 Conclusion................................................264 Conclusion........256 References ............................................................................234 Tensile Strength ..........................................244 Solubility Parameter Correlations Based on Permeation Phenomena ..........................................................................................................................................................................................................269 Introduction .....243 Introduction .......................................................................................................................................................................................................................................................................................................................270 Hydrophobic Bonding and Hydrophilic Bonding (Self-Association) .............................................................................................................................................................259 Abstract .............................243 Concentration-Dependent Diffusion .......................................250 Solubility Parameter Correlation of Permeation Coef ficients for Gase ...............................................................239 References ....................................259 ESC Interpreted Using HSP ................................................................................234 Acceptable or Not — Plastics ....................................................................................................................................................233 PET Film Coating .........................................................................260 ESC with Nonabsorbing Stress Cracking Initiators ..................................................................................................................................................................................................................................231 Introduction ........................................................................................................................................................................................................................................................................240 Chapter 13 Applications — Barrier Polymers ..........................................237 Special Effects with Water..........................................................................................................................................................................................................................................245 Solubility Parameter Correlations of Breakthrough Times ..............231 Abstract ..............................................................................................................................................................................................................................................267 References .......................232 Effects of Solv ent Molecular Size .........................................................................................................................................................................................................................................................................255 Conclusion..................................................................................................................................................................................................232 Chemical Resistance — Examples ........................................................................................................259 Introduction ...................................................................................................................................................................................253 General Considerations ...............................................................................................................251 Laminates....................................................................245 Solubility Parameter Correlation of Permeation Rates ..............................................................................................Chapter 12 Applications — Chemical Resistance ................273 Cholesterol ......................233 Tank Coatings ..............257 Chapter 14 Applications — Environmental Stress Cracking in Polymers ...................................................................231 Chemical Resistance — Acceptable-or-Not Data ...............................................................275 Lard ......................................................263 Discussion ......................................................................................................................................................267 Chapter 15 Hansen Solubility P arameters — Biological Materials ....271 DNA ...........................................277 ....269 Abstract ...........................................................................................................................................243 Abstract .......................

.....................................................................................................................................................................................................................................................................................................296 The Diffusion Equation.....................................................................................................................................................311 Introduction ..................................................................................................................289 Surface Mobility .....309 Chapter 17 Applications — Safety and En vironment ............................................279 Wood Chemicals and Polymers .....................290 Chiral Rotation....306 General Comments ....311 Alternative Systems ...............................................................................................................................................291 References ............................................................304 Film Formation by Solv ent Evaporation ..................................................................................................298 Surface Resistance in Absorption Experiments ........315 Skin Penetration ..............................................................................................................................................296 Concentration Dependent Diffusion Coefficient ....................................313 The Danish Mal System — The Fan.....................................................................................................................................................................................................................................................................................296 Constant Diffusion Coefficient .........................291 Chapter 16 Absorption and Diffusion in Polymers .........315 Uptake of Contents by a Plastic Container .........................................................................................................................................................................................................................................304 Measuring Diffusion Coefficients with Sur ace Resistance and Concentration Dependence.....................................................................................301 Surface Resistance — A Discussion .................................................................................................................................................................................................................................................................................................................308 References ..............................318 .........................................................293 Introduction .......................279 Chlorophyll and Lignin ...................................................................................................................................................................................................................................................................................................302 Side Effects ...290 Conclusion..............................................................................................................................................................................................................................................................................................................................................................................277 Proteins — Blood Serum and Zein ...........................................................................................................................................................................................................308 Conclusion....................................................................................................................................................................................................297 Surface Resistance ........................................................................................................................316 Conclusion..............................................312 Solvent Formulation and Personal Protection for Least Risk ......305 Anomalous Diffusion (Case II....................................................................................317 References ..........................311 Abstract .............316 Transport Phenomena..................................................................................293 Abstract ......................279 Urea ........................................................................................................................... Hydrogen Bonding.......293 List of Symbols Used in This Chapter ..............................................................................................................................................................................................................................311 Substitution................................ Super Case II) ........................................................................................................283 Water ........................................298 Mathematical Background..............Human Skin...300 Surface Resistance in Permeation Experiments ....................................313 Selection of Chemical Protecti ve Clothing ........................................294 Steady State Permeation ........................................................................................................................................................................................................................................................................................................................ and Nanoengineering ....................................................

..........................333 Water............................................................................................................................342 Appendix A: Comments to Table A...........................................................................................................................................................................................................................................................................................................................................................................328 Characterizing Surfaces ........................507 Table A................338 Absorption of Chemicals in Plastics .................................................321 Abstract ..................................................................................................................................................................340 Theoretical Problems Awaiting Future Resolution .....................................................................................................341 Polymer Solubility .....................490 List of Trade Names and Suppliers .............................................................................................................................332 Surface Active Agents ....................................................................................................3 ...................................................................................................................................................................................336 Organic Salts ................................................................................................................................................................................................................334 Gases............341 Surface Phenomena .......................................................................339 Controlled Release..................................493 Appendix A: Comments to Table A...491 Table A.....................342 Conclusion......................337 Organometallic Compounds .......................................................................485 References ..........................................................342 References ....................................................321 Hansen Solubility P arameter Data and Data Quality .........................................337 Inorganic Salts ....................................................................................................................................................346 Table A..............511 .....................................................................................................................................................345 References ....................................................................................2 .339 Chemical Resistance.......................................................................................................................................321 Introduction ...............2 .................................................................................................1 ................................................................................................................................................................................339 Nanotechnology....................................................................................................................................................................332 Surface Mobility (Self-Assembly) ..........................................................................................330 Materials and Processes Suggested for Further Attention ...........................................3 ...............................................1 ............338 Aromas and Fragrances ...............................................................................................................................................................................................347 Appendix A: Comments to Table A...............................................................Chapter 18 The Future ..........................................................................328 Polymers as Points — Solv ents as Spheres ...........................................................................................................................................................................508 Index.....................................................................................................................................................................................................................324 Group Contribution Methods ...................................................................................................................


chemical resistance. additives. solv ency. One can no w easily predict how to dissolve a given polymer in a mixture of two solvents. this has been modified to “li e seeks like. provide some background for a basic understanding. The key factor is to determine those af finities that the important components in a system h ve for each other. Unfortunately.Parameters — 1 Solubility An Introduction Charles M. of k ey components.” More recently. this book cannot include discussion of all the significant e forts leading to our present knowledge of the solubility parameters. leading to compatibility . if necessary . fibers and fillers. and surf aces do not (usually) dissolv e. and permeation rates. external environment. Only those materials tha interact differently with dif ferent solvents can be characterized in this manner . For many products this means e valuating or estimating the relati ve af finities of sol ents. pigment surf aces. These include liquids used as energy probes and computer programs to process data. workplace environment. by e xperiment and preferably with agreement between the tw o. and even to characterize the surfaces of pigments. The skill with which solvents can be optimally selected with respect to cost. and substrates. INTRODUCTION The solubility parameter has been used for man y years to select solv ents for coatings materials. They are used in other industries. but for predicting affinities betwee different polymers. This chapter describes the tools commonly used in Hansen solubility parameter (HSP) studies. and af finities to sur aces to impro ve dispersion and adhesion. has impr ved over the years as a result of a series of impro vements in the solubility parameter concept and widespread use of computer techniques. and gi ve examples of uses in practice.” as many surface characterizations ha ve also been made. to measure the similarity . filler sur aces. The basic principle has been “like dissolves like. using well-define liquids as energy probes. to predict compatibility of polymers. Solubility parameters help put numbers into this simple qualitati ve idea. A lack of total success has stimulated further research. neither of which can dissolve the polymer by itself. Materials with widely different chemical structures may be v ery close in affinities. or lack of the same. and polymers will dissol e in solvents whose solubility parameters are not too dif ferent from their o wn. flash point. It can be e xpected that many inorganic materials. will not interact di ferently with these energy probes 1 . polymers. The goal is to arri ve at the HSP for interesting materials either by calculation or . It is note worthy that the concepts presented here ha ve de veloped to ward not just predicting solubility that requires high affinity between sol ent and solute. fiber sur aces. evaporation rate. Most commercial suppliers of solv ents ha ve computer programs to help with solv ent selection. such as fillers. etc. ho wever.. In these applications the solubility parameter has become a tool. Liquids with similar solubility parameters will be miscible. Hansen ABSTRACT Solubility parameters have found their greatest use in the coatings industry to aid in the selection of solv ents.An attempt is made to outline de velopments.

Equation 1. The energy of vaporization is a direct measure of the total (cohesi ve) energy holding the liquid’ s molecules together. This is discussed more in Chapter 2.15). as well as fillers li e barium sulf ate. The correct relation is 3–8: ΔGMnoncomb = ϕ1ϕ2VM(δ1 – δ2)2 (1. etc. ΔHM is the heat of mixing. When the same solvents that dissolve a polymeric binder are those which interact most strongly with a surf ace. Equation 1. ΔGMnoncomb. both or ganic and inor ganic. and co workers that ΔGMnoncomb is given by the right-hand side of Equation 1. The term cohesion energy parameter is more appropriately used when referring to surface phenomena. The solubility parameter is an important quantity for predicting solubility relations. Solubility parameters are sometimes called cohesion energy parameters as the y are deri ved from the energy required to convert a liquid to a gas.3 gives the heat of mixing as proposed by Hildebrand and Scott: ΔHM = ϕ1ϕ2VM(δ1 – δ2)2 (1.2 Hansen Solubility Parameters: A User’s Handbook as their energies are very much higher. Changing the sur ace energies by various treatments can lead to a surf ace that can be characterized more readily and often interacts more strongly with gi ven organic solvents. as can be seen from the follo wing brief introduction.1) Where V is the molar v olume of the pure solv ent. it can be e xpected that the binder and the surf ace have high affinit for each other. and also inorganic fibers (see Chapter 7). The free ener gy change for the solution process is gi ven by the relation: ΔGM = ΔHM – ΔTSM (1.3 and not ΔGM.0455 times larger than that in (cal/cm 3)1/2. Re gardless of these concerns. Thermodynamics requires that the free ener gy of mixing must be zero or ne gative for the solution process to occur spontaneously . and it has often been cited as a shortcoming of this theory in that only positive heats of mixing are allo wed. and VM is the volume of the mixture.4) The noncombinatorial free energy of solution. Equation . It has been sho wn by Patterson.3 is not correct. Delmas. The numerical v alue of the solubility parameter in MP a1/2 is 2. it has been possible to characterize pigments. and this has led to the concepts described in more detail later . All types of bonds holding the liquid together are brok en by e vaporation. The Hildebrand solubility parameter is defined as the square root of the cohes ve energy density: δ = (E/V)1/2 (1..2) where ΔGM is the free energy of mixing. and E is its (measurable) ener gy of vaporization (see Equation 1. and ΔSM is the entrop y change in the mixing process. includes all free energy effects other than the combinatorial entropy of solution that results by simply mixing the components. HILDEBRAND PARAMETERS AND BASIC POLYMER SOLUTION THERMODYNAMICS The term solubility parameter w as first used by Hildebrand and Scott 1.2 The earlier w ork of Scatchard and others w as contributory to this de velopment. T is the absolute temperature. An adsorbed layer of w ater on the high-energy surface can also play an important role.3) The φ1 and φ2 are volume fractions of solvent and polymer. zinc oxide.

The combinatorial entrop y enters as a constant f actor in the plots of solubility in dif ferent solv ents. These use solubility parameters as ax es and have experimentally determined boundaries of solubility defined by the act that the free ener gy of mixing is zero. in particular. can be found in Barton’ s extensive handbook. will al ways be problematic. “nonpolar” and “polar. An average solvent molecular volume is usually taken as about 100 cc/mol. even though solubility parameter considerations might predict the y would. particularly for coatings and adhesion phenomena. Therefore.4 predicting both positi ve and ne gative heats of mixing. in that the first has found wide use an the second greatly influenced the author s earlier activities. or rather the dif ference in solubility parameters for the solv ent–solute combination. found on simple mixing to result in a disordered mixture compared to the pure components. as the concentrations are usually constant for a gi ven study. has found numerous practical applications.9 10 The slightly older . can dissolve a polymer at larger solubility parameter differences than might be expected from comparisons with other solv ents with larger molecular volumes. and the succession of research reports that have attempted to improve on it. particularly in Chapter 2. ho wever. and the positi ve entropy change (the combinatorial entrop y change). both positi ve and ne gative heats of mixing can be e xpected from theoretical considerations and ha ve been measured accordingly . 11 who divided solv ents into h ydrogen bonding classes. The approach of Burrell. Chapter 12. It is clear that a match in solubility parameters leads to a zero change in noncombinatorial free energy. The Prausnitz article. for e xample. as defined by Hildebrand and Scott2 and does not account for association between molecules. It has been clearly sho wn that solubility parameters can be used to predict both positi ve and ne gative heats of mixing. will ensure that a solution is possible from a thermodynamic point of vie w. such as methanol and acetone. and Chapter 16. Chapter 13. The latter problem seems to ha ve been largely solved with the use of multicomponent solubility parameters. These effects are also discussed elsewhere in this book. e xcellent contrib ution of Gardon and Teas is also a good source of related information. such as those that polar and h ydrogen-bonding interactions w ould require. A practical aspect of this effect is that solv ents with relati vely low molecular v olumes. This is more in agreement with the use of the solubility parameter plots to follow. is important in determining the solubility of the system. for e xample. The maximum dif ference in solubility parameters that can be tolerated where the solution still occurs is found by setting the noncombinatorial free ener gy change equal to the combinatorial entrop y change: ΔGMnoncomb = TΔSMcomb (1. This discussion clearly demonstrates wh y the solubility parameter should be considered as a free energy parameter. the approach of Blanks and Prausnitz 12 divided the solubility parameter into tw o components. A more detailed description of the theory presented by Hildebrand. Lar ger molecular species may not dissolve. Pre vious objections to the ef fect that only positi ve values are allowed in this theory are incorrect. was .5) This equation clearly sho ws that an alternate vie w of the solubility situation at the limit of solubility is that it is the entrop y change that dictates ho w closely the solubility parameters must match each other for the solution to occur . Using the dif ference between two large numbers to calculate a relati vely small heat of mixing.4 is consistent with the Prigogine corresponding states theory (CST) of polymer solutions (see Chapter 2) and can be dif ferentiated to gi ve e xpressions3. It is important to note that the solubility parameter . The converse is also true.Solubility Parameters — An Introduction 3 1. however.” Both are w orthy of mention. A shortcoming of the earlier solubility parameter w ork is that the approach w as limited to regular solutions. This can be a difficulty in predicting the beh vior of plasticizers solely based on data for lower molecular weight solvents. the lack of accurac y with which the solubility parameters can be assigned will al ways remain a problem. It will be seen in Chapter 2 that solv ents with smaller molecular v olumes will be thermodynamically better than lar ger ones ha ving identical solubility parameters.

1 vs. dif ferently from the w ay the y do in situations where they make up their own environment.. a nonsolvent again with still further increase in temperature. see Chapter 5. This is one reason polymer–polymer miscibility is dif ficult to achi ve. although there are e xceptions. for example. Materials with similar HSP ha ve high af finity for each othe . ho wever. subsequently.e. The extent of the similarity in a given situation determines the e xtent of the interaction. This situation allows a nonsolvent. and it is surprising that so many different applications ha ve been possible since 1967 when the idea w as first published. The total cohesive energy. i. E. for e xample. whereas nitromethane is not. Thus the total cohesive energy is considered as being identical to the ener gy of v aporization. The free ener gy criterion dictates that polymer solubility parameters match e xtremely well for mutual compatibility . with or within themselv es.2 that the entrop y change is beneficial to mixing. Further discussion of these phenomena is b yond the scope of the present discussion. this w as the basis of the pioneering work of Patterson and coworkers. The basis of these so-called HSPs is that the total ener gy of v aporization of a liquid consists of se veral individual parts.e.2 Ethanol and nitromethane. It can be seen from Equation 1. but their affinities are quite di ferent. Indeed. Ethanol is water soluble. polymer–polymer miscibility can be promoted by the introduction of suitable copolymers or comonomers that interact specifically within the system. A rather large number of applications are discussed in this book. The basis of the approach is. However. It should also be noted that these cohesive energies arise from interactions of a given solvent molecule with another of its o wn kind. The entropy changes associated with polymer solutions will be smaller than those associated with liquid–liquid miscibility. Several cases are discussed where appropriate in the follo wing chapters. as pure liquids. The same cannot be said of the total or Hildebrand solubility parameter. but those of a liquid frequently decrease rapidly with temperature. Indeed. This is the usual case. breaking all the cohesi ve bonds. Increasing temper ature can also lead to a nonsolv ent becoming a solvent and. Others are found in the w orks of Barton. as there is little to be g ained from the entrop y contribution when progressi vely larger molecules are in volved. 25. These are usually “boundary” solvents on solubility parameter plots. Increasing temperature does not always lead to impro ved solubility relations. with a solubility parameter that is initially too high. Their work was done in essentially nonpolar systems. (molecular) permanent dipole–permanent dipole forces.1 MPa1/2. can be measured by e vaporating the liquid. respectively). v ery simple. therefore. have similar total solubility parameters (26. this postulate could ne ver have been made. this will w ork in the direction of promoting a more ne gative free energy of mixing. and (molecular) h ydrogen bonding (electron exchange). as the “monomers” are already bound into the config uration dictated by the polymer the y mak e up. 13–17 These arise from (atomic) dispersion forces. These appear most often where the en vironment causes the solv ent molecules to interact. Polymer solubility parameters do not change much with temperature.. to pass through a soluble condition to once more become a nonsolv ent as the temperature increases. such as Scatchard. HANSEN SOLUBILITY PARAMETERS A solubility parameter approach proposed by the author for predicting polymer solubility has been in wide use. mixtures of nitroparaf fins and alcohols wer demonstrated in man y cases to pro vide syner gistic mixtures of tw o nonsolv ents that dissolv ed . They are no longer free in the sense of a liquid solvent and cannot mix freely to contrib ute to a lar ger entropy change. i. without the work of Hildebrand and Scott 1. Needless to say. This is discussed in more detail in Chapter 2.1.2 and others not specifically reference here. When multiplied by the temperature.3–8 to show that subsequent increases in temperature can predictably lead to insolubility.4 Hansen Solubility Parameters: A User’s Handbook farsighted in that a corresponding states procedure w as introduced to calculate the dispersion ener gy contribution to the cohesi ve energy.9 A lucid discussion by Barton 18 enumerates typical situations where problems occur when using solubility parameters.

For the saturated aliphatic hydrocarbons. induced dipoles. These approaches will not be dealt with here b ut are described in Barton’s handbook 9 and else where. This would presumably parallel similar characterizations of surf ace-active materials. Induced dipoles ha ve not been treated specificall in this approach b ut are recognized as a potentially important f actor. As discussed in more detail later . In this perhaps o versimplified approach. They make possible estimates of the three parameters for additional liquids. the polar cohesive energy. orientation. F or e xample. Group contrib ution methods and suitable equations based on molecular properties were then derived from these. whereas the HSP concept readily confirms the reason for this e fect. The dipole moment is the primary parameter used to calculate these interactions. and proton acceptor. proton donor. The Hansen parameters have mainly been used in connection with solubility relations.13 This could ne ver ha ve been predicted by Hildebrand parameters. This can be called more generally an electron exchange parameter. Alcohols. mostly . A molecule can be mainly polar in character without being w ater soluble. such as nitroparaf fins propylene carbonate. The polar solubility parameters referred to here are well-defined. particularly for solv ents with zero dipole moments (see the Calculation of the Polar Solubility P arameter section). The author stopped with the three major types found in or ganic molecules. metallic bonds. electrostatic interactions. in the coatings and related industries. the description of or ganometallic compounds could be an intriguing study . . for e xample. induction. all molecules contain those types of attractive forces. As a single parameter. The most general are the nonpolar interactions. hence there is a misuse of the term polar in the general literature. The permanent dipole–permanent dipole interactions cause a second type of cohesion ener gy. E P. the most polar of the solv ents include those with relatively high total solubility parameters that are not particularly w ater soluble. but not e xclusively. xperimentally verified. These are inherently molecular interactions and are found in most molecules to one e xtent or another.19–21 The most e xtensive di vision of the cohesi ve ener gy has been done by Karger et al. these are essentially the only cohesi ve interactions. carboxylic acids. and other hydrophilic materials have high-hydrogen-bonding parameters. or h ydrocarbon counterpart.22 who developed a system with fi e parameters: dispersion. The strength of a particular type of hydrogen bond or any other bond is important only to the e xtent that it influences the cohes ve energy density.. glycols. The goal of a prediction is to determine the similarity or dif ference of the cohesion energy parameters. the Hansen hydrogen bonding parameter has serv ed remarkably well in the e xperience of the author and k eeps the number of parameters to a le vel that allows ready practical usage. and the ener gy of v aporization is assumed to be the same as the dispersion cohesive energy. These are derived from atomic forces and ha ve also been called dispersion interactions in the literature.Solubility Parameters — An Introduction 5 polymers. acid and base cohesion parameters — to allo w both positi ve and negative heats of mixing. Finding the dispersion cohesi ve energy as the cohesion ener gy of the homomorph. Solubility and swelling have been used to confirm the solubility parameter assignments of ma y of the liquids. E H. Other researchers ha ve divided this parameter into separate parts — for e xample. and can be estimated from molecular parameter as described later. There are three major types of interactions in common or ganic materials. E D. and tri-n-b utyl phosphate. Hydrogen bonding is a molecular interaction and resembles the polar interactions in this respect. this is based on a corresponding states calculation. The basis of this type of cohesi ve energy is attraction among molecules because of the h ydrogen bonds. where each part of the molecule requires separate characterization for completeness. As molecules are b uilt up from atoms. As noted previously. for example. It has been recognized that additional parameters could be assigned to separate ener gy types. or whate ver type of separate energy can be defined. The third major cohesi ve ener gy source is h ydrogen bonding. is the starting point for calculating the three Hansen parameters for a gi ven liquid. It is clear that there are other sources of cohesion ener gy arising in various types of molecules from. th hydrogen bonding parameter has been used to more or less collect the ener gies from interactions not included in the other two parameters.

the wetting of surfaces. Indeed.23–26 and have even led to systematic surface treatment of inorganic fiber so that the y could be readily incorporated into polymers of lo w-solubility parameters such as polypropylene27 (see also Chapter 7).29 so the interactions of molecules residing in surf aces should clearly be included in an y general approach to interactions among molecules.10.14. Man y widely dif ferent applications ha ve been discussed by Barton9 and Gardon. and in other important phenomena relating to nanotechnology (Chapter 18). certain biological phenomena (Chapter 15). In other w ords. it is emphasized that HSPs quantitati vely account for the cohesion energy (density). 13 This procedure in volved calculation of the nonpolar parameter according to the procedure outlined by Blanks and Prausnitz. such as sur ace science. whether surface or cohesive energies.7) (1. an e xperimental latent heat of v aporization has been considered a more reliable method to arrive at a cohesion energy rather than using molecular orbital calculations or other calculations based on potential functions. METHODS AND PROBLEMS IN THE DETERMINATION OF PARTIAL SOLUBILITY PARAMETERS The best method to calculate individual HSPs depends to a great extent on what data are available. can lead to rapid decisions and plans of action in critical situations that lack data. The basic equation go verning the assignment of Hansen parameters is that the total cohesion energy. An alternative method of calculating the three parameters has been presented. and the Prigogine corresponding states theory of polymer solutions discussed in Chapter 2. P . Up to this point of time. Surf ace mobility and surf ace rotation are important f actors in en vironmental stress cracking (Chapter 14). Flory also emphasized that it is the surf ace of molecules that interact to produce solutions. The author originally adopted an essentially e xperimental procedure and established v alues for 90 liquids based on solubility data for 32 polymers. A key to parameter assignments in this initial trial-and-error approach was .16. E. E = ED + EP + EH (1.6) Dividing this by the molar v olume gi ves the square of the total (or Hildebrand) solubility parameter as the sum of the squares of the Hansen D. where they have been used to characterize the wettability of various surfaces and the adsorption properties of pigment surfaces. must be the sum of the indi vidual energies that mak e it up. the goal of such e xtensive calculations for polar and hydrogen bonding molecules should be to accurately arrive at the energy of vaporization. and H components.8) To sum up this section. The statistical thermodynamics approach of Panayiotou and coworkers reported in Chapter 3 may ha ve changed this. E/V = E D/V + E P/V + E H/V δ2 = δ2D + δ2P + δ2H (1.28 Surface characterizations have not been given the attention deserved in terms of a unified similarity-of-ene gy approach. The author can certify that thinking in terms of similarity of energy. It is outlined in the follo wing section. b ut full evaluation of this ne w information has not been possible as yet. the e veryday industrial crisis situation often can be reduced in scope by appropriate systematic approaches based on similarity of ener gy. The division of the remaining cohesi ve energy between the polar and h ydrogen bonding interactions w as initially done by trial and error to fit xperimental polymer solubility data.12 This calculation procedure is still in use and is considered the most reliable and consistent one for this parameter .6 Hansen Solubility Parameters: A User’s Handbook HSPs do have direct applications in other scientific disciplines. The success of the HSPs for surf ace applications are not surprising in vie w of the similarity of predictions offered by these.

P atterson and co workers ha ve been especially instrumental in relating the Prigogine theory to solubility parameters and to the Flory–Huggins theory of polymer solutions. methanol. in that they dissolve a polymer even though they have solubility parameters placing them at a distance greater than the x e perimentally-determined radius of the solubility sphere. of course. separate. and h ydrogen bonding interactions all follo w the geometric mean rule.25) could be used to calculate the polar parameter quite well. Using the experimentally established. and Scatchard discussed pre viously. Scott. This dependence on molar v olume is inherent in the theory de veloped by Hildebrand. It should be recognized used. smaller molecular species such as acetone. nitromethane. that the constant 4 is theoretically predicted by the Prigogine corresponding states theory of polymer solutions when the geometric mean is used to estimate the interaction in mixtures of dissimilar molecules. ΔGM. as discussed in Chapter 2. permanent dipole–permanent dipole. There are. The approach of optimizing solubility data to spheres is still very much in use. Such smaller -molecular-volume species that dissolv e “better” than predicted by comparisons. well-defined partial-solubility parameters with a fourth. This is exceptionally strong evidence that dispersion. is equal to zero on the boundary . These values were also consistent with the e xperimental solubility data for 32 polymers a vailable at that time and with Equation 1. δP and δH parameters. When the scale for the dispersion parameter is doubled.Solubility Parameters — An Introduction 7 that mixtures of two nonsolvents could be systematically and synergistically (but predictably) found to dissolve given polymers. It will be noted here. should not necessarily be considered outliers. and others often appear as outliers. rather than spheroidal. Hansen and Skaarup15 found that the Böttcher equation (Equation 10. Smaller molar v olume f avors lo wer ΔGM. and Chapter 16).9) This equation was developed from plots of experimental data where the constant “4” was found convenient and correctly represented the solubility data as a sphere encompassing the good solv ents (see Chapter 5). re gions of solubility are found. F or e xample. in comparison with the other two parameters essentially spherical. The use of the geometric mean is basic to this agreement between the HSP approach and that of Prigogine (see Chapter 2). boundary regions where deviations can occur. Likewise. These are most frequently found to involve the larger molecular species as being less ef fective solv ents compared to their smaller counterparts that define the solubilit sphere. The total free energy of mixing. 3–8 The HSP approach of dividing the cohesive energy into parts derived from different types of cohesive forces has been confirmed both by xperimental studies. This in turn promotes solubility . ho wever. between tw o materials based on their respecti ve partial solubility parameter components: (Ra)2 = 4( δD2 – δD1)2 + ( δP2 – δP1)2 + ( δH2 – δH1)2 (1. reasonably accurate di visions into the three ener gy types were possible. these are especially important in correlating dif fusional phenomena with HSP (see Chapter 12. and this led to a revision of the earlier values to those now accepted for the same liquids. 30 The constant 4 differentiates between atomic and molecular intereactions. a spheroid. based on solubility parameters alone. Using a lar ge number of such predictably syner gistic systems as a basis. rather than multiplying the solubility parameters by the molar v olume raised to some power to redefine them The reason for the e xperimentally determined constant 4 in Equation 1. as well as the Prigogine theory . Furthermore. The molar v olume is frequently and successfully used as a fourth parameter to describe the effects of molecular size. The author has preferred to retain the three. or computer-optimizing boundaries of solubility by locating the maximum dif ference in solubility parameters allo wed by Equation 1.6. approximate. molar olume parameter. This meant that these had parameters placing them on opposite sides of the solubility re gion.9 will be discussed in more detail in Chapter 2.9 are both . Skaarup de veloped the equation for the solubility parameter “distance. Plotting regions of solubility based on experimental solubility data.” Ra. Chapter 13. This greatly aids two-dimensional plotting and visualization. Ro.

0 is a boundary condition. giving results generally consistent with Equation 1.1. as discussed in Chapter 2. 9 although Beerbo wer’s latest v alues ha ve only appeared in the National Aeronautics and Space Administration (NASA) document. Many of these have multiple groups. The revised set of parameters for the 90 original solv ents was the basis for group contrib ution procedures developed (most notably) by van Krevelen. and it becomes more and more difficul to make decisions as to ho w to treat them best. The trend has been to calculate HSP for lar ger and still larger molecules. Scanning a computer output for RED numbers les than 1. The latter quantity is important for relating the HSP approach to that of Huggins and Flory. The ratio (Ra/Ro) 2 = (RED) 2 is a ratio of cohesion energies. but it clearly has adv antages in some cases.6 to Equation 1.33 These v arious de velopments ha ve been summarized by Barton. Another promising approach to arri ve at the HSP for materials based on e xperimental data is to use multivariable analysis of one type or another . This may present problems for smaller sets of data. RED equal to or close to 1. as discussed in Chapter 5. rapidly allows location of the most interesting liquids for a given application. The majority of the data in this table. RED = Ra/Ro (1. reflecting th relative energy difference. The ratio Ra/Ro has been called the RED number. At times the HSP for the lar ger molecules can be estimated from the HSP of lar ger segments that mak e them up.13) and has been found quite reliable by K oenhen and Smolders. This is discussed in more detail in Chapter 5 and the definition of the limited s gment of the boundary of the HSP sphere deri vable from such correlations is based on Equation 1.0 indicate high affinity. If a boiling point is a vailable.2.8. and progress vely higher RED numbers indicate progressively lower affinities. the procedures for calculating δD have been used. 32 Table 1. e xcept as noted briefly late . Equation 1. RED numbers less than 1. as well as Table 1.1 is an e xpanded table of Beerbower group contributions. the poor solv ents or nonsolvents located near the boundary of the sphere that fix th boundary (and center) rather than the best solv ents in the middle.4 dif fers from this by the combinatorial entropy of mixing.35 This equation has been found reliable by the author as well. is the basis of the whole approach. Rather than e xpanding Table 1. again. the usual practice has been to locate chemicals with similar groups and to use their HSP v alues in a group contrib ution-type calculation.1 with additional data. which.9 is readily used on a computer (or on a hand calculator). requires that Ra be less than Ro. ha ve also appeared in Reference 34.17 which in volved only the dipole moment and the square root of the molar v olume.1 ha ve been used e xtensively to arri ve at the collection of HSP data in Appendix Table A.9. it should be noted that the ratio of Ra to Ro is really a ratio of quantities having the same units as the solubility parameter . It is ob vious that solubility. Beerbower also de veloped a simple equation for the polar parameter . b ut it is an adv antage when e xtrapolating into re gions of HSP higher than those of an y liquid that can be used in testing. If a dipole .32 and Hansen and Beerbower. If a boiling point is not a vailable. for example. or high af finit . and supplementary relations allow easier scanning of lar ge sets for data. Most of the chemicals of primary interest for which full data are available are presumably already in this table. This is also gi ven later (Equation 1.0.10) A RED number of 0 is found for no ener gy difference. The author’s preferred approach of locating the polymer HSP as the center of a sphere has a problem in that it is. in reality. the similarity with related molecules has been used.This type of approach has not been attempted by the author . Parenthetically. The group contrib utions in Table 1. which w as distributed among those who were in contact with Beerbo wer in the late 1970s.31 Beerbower.8 Hansen Solubility Parameters: A User’s Handbook that using the solubility parameters relating to ΔGMnoncomb in Equation 1. K oenhen and Smolders also give correlation coefficient for other calculation procedures to arri ve at the indi vidual Hansen parameters. The procedure has de veloped to the point where its principle features can be identified in th following table.17 who also used Fedors’ w ork.

This is necessary to pre vent δP for materials like plasticizers from being clearly too lo w. for example. Another reason is that some liquids are chameleonic. polyvinyl chloride.1). when dipole moments are a vailable for these. or wide variations in reported dipole moments. it has been found that the group contrib ution for the polar component of aliphatic esters should not be less than 300 cal/mol as gi ven in Table 1. and points out that Hildebrand or Hansen parameters must be used with particular caution where the e xtent of donor–acceptor interactions. Ho y38 cites the formation of c yclic structures for glycol ethers with (nominally) linear structure. The dipole moment for some liquids depends on their environment. Group contrib utions for sulfur . There is often a change in the group contrib ution as a function of molecular size.1. and other groups not found in Table 1. and 500 pesticides. The Hoy dispersion parameter. for a polymer for which there are limited data. especially .1. This is the main reason for the lack of e xpansion of Table 1. Sulfur containing compounds have also been somewhat difficult in this respect with major changes in estimated group contrib utions depending on molecular weight of the chemical in question. It is thought best that the uncertainty be clear to the user . 450 resins and polymers.37 has indicated a computer program with data for o ver 500 solv ents and plasticizers. In some cases. . Hoy has provided an excellent source of total solubility parameters.40 are not compatible with the customary Hansen parameters reported here. The extensive list of group contributions at the end of Chapter 3 pro vides what may be a partial replacement and/or a supplement for Table 1. as an e xample. In particula . F or example. A sizable number of materials ha ve been assigned HSPs using the procedures described here. The scope of this situation has been be yond the resources a vailable for its fully satisf actory resolution. Barton has discusse some of the situations where cohesion parameters need a more careful use. He independently arrived at the same type of di vision of cohesion ener gies as Hansen. and not for the molecule in question. this is due to questionable p ysical data.1. Large data sources greatly enhance a search for similar materials and the locating of ne w solvents. If necessary. The formation of h ydrogen-bonded tetramers of alcohols in a fluoropolymer has als been pointed out. P artial solubility parameters not tak en directly from earlier well-documented sources should be used with caution. as discussed later. different authors have used different group contribution techniques. This has been recognized and dealt with to the extent possible. The problem with this procedure is ob vious: any error or distortion of value for a class of compounds is perpetuated. but there are limits to what can be done with limited data. Still another reason for dif ficulties is the la ge variation of dipole moments reported for the same liquid.1 can be easily deri ved from the data on similar compounds reported in Appendix Table A. Many of these have not been published. Pure component data cannot be e xpected to predict the beha vior in such cases. the procedures gi ven here were used in preference to group contrib utions. Exxon Chemical Corporation 36. although the methods of calculation were quite dif ferent. A given solvent can be listed with dif ferent values in files to eep these phenomena in mind. for e xample.Solubility Parameters — An Introduction 9 moment is available. h ydrogen bonding within a compound. 38 as defined by H y. amides. for e xample.1. This would seem to be an unfortunate situation. Unfortunately . partial solubility parameter v alues found in Ho y’s e xtensive tables 9. 18 Amines. particularly where mixtures with w ater were in volved and structured species were postulated to e xplain phenomena based on specific interactions among the components of the mixtures. Many solvent suppliers ha ve also presented tables of solv ent properties and/or use computer techniques with these tables in their technical service. group contributions can be derived from similar molecules to the one in question. are kno wn to associate with each other . and there is a proliferation of dif ferent “Hansen” parameters for the same chemicals in the literature. 39 The term compound formation can be found in the older literature. The same is true of the δH component. although several of them appear with two sets of possible values awaiting experimental confirmation. in that the y adopt configurations depending on their e vironment. for latent heats of vaporization. The author’s files contain the three parameters for about 1200 chemicals (See Appendix Table A. based on their compatibility with. This requires some e xperience with the techniques in volved. is v ery dif ferent from that between compounds. b ut may ultimately pro vide benefits.

5 66.000 1.700 5.0 Same Same (31.360b 1.650 ± 500 1. 2 (cal/mol) ΔVδD Cyclo Aromatic Alkane Cyclo Aromatic Aliphatic Aromatic Aliphatic Polar Parameter.500 ± 300 c 1.500 ± 300 c ? 0 —e 950 ± 300 —f 3.0 97. 2 (cal/mol) ΔVδH Total Parametera ΔVδ2 (cal/mol) Aromatic Molar Volume.750 ± 200 (338.5 10.600b 6.4) Same Same Same 1.300 ± 100 3.870 ± 600 0 0 0 0 25 ± 10 18 ± 5 60 ± 10 — 0 0 1.650 ± 160 4.030 1.100b 3.770 ± 450 Same Same Same Same ? ? ? — 250 250 0 0 0 ? ? ? 1.500 ± 300 c ? 2.650 ± 150) Same Same Same Same Same .350 ± 400 ? ? 3.0 ϵF3 triplet f 66.300b 2.670b 4.350 ± 250 c 5.0 28.000 ± 200 c 1.650 ± 140 3.800 ± 250 c ? 1.0 –19.0 28.5 –5.0 –Cl ෇Cl2 twin f ϵCl3 triplet f –Br ෇Br2 twin f ϵBr3 triplet f –I I2 twin e ϵI3 triplet e 111.750 ± 250 4.030 — 250 250 1.100 ± 300 0 0 0 0 ? ? ? 50 ± 25 — 0 700 ± 100 500 ± 250 c ? 800 ± 100 400 ± 150 c ? 800 ± 100 400 ± 150 c ? 575 ± 100 400 ± 150 c ? 450 ± 150 950 ± 300 2.8 (23.000 ± 10%)/V 750 ± 350 800 ± 150 0 0 0 0 180 ± 75 180 ± 75 180 ± 75 — 0 0 0 0 0 100 ± 20 c 165 ± 10 c 350 ± 250 c 500 ± 100 825 ± 200 c 1.300 6.10 TABLE 1.250 ± 250 c 4.030 1.5 CH2< –CH< >C< CH2 = olefi –CH = olefi >C = olefi PhenylC-5 ring (saturated) C-6 ring –F ෇F2 twin f 40.760 4.180 820 350 1.1 Group Contributions to Partial Solubility Parameters London Parameter.370 ± 500 Same Same Same Same ? ? ? 7.0 81.0 52.050) 4.250 ± 100 800 ± 150 300 ± 100 1.000 11.100 ± 175 c ? 1.650 4.700 800 4.6 123.180 820 350 850 ± 100 875 ± 100 800 ± 100 — — — 0 0 0 1.2 28.0 62.0 60.9 30.120 Same Same Same Same Same Same Same 7630 — 250 800 b 1.600 5.100 ± 200 (56.2 31.200b 3.6 Hansen Solubility Parameters: A User’s Handbook 3.a ΔV (cm3/mol) Functional Group Aliphatic Aromatic Alkane 16.700b 6.150 (4.250 ± 150 2.000 ± 150 700 ± 250 c ? 1.700 1.600 7.950 ± 300 c 4.550 8.2) 18.450 ± 100 ? ? 1.4 — Same 22.300 ± 300 c 5. 000 ± 7%)/V 2.0 48.600 ± 400 1.5 — 16 16 18.350 ± 300 1.9 34.0 78.650 2.5 13.000 ± 12%)/V 500 ± 150 700 ± 200 0 0 0 0 ? ? ? — 0 0 ? ? ? 1.650 ± 250 c ? 450 ± 25 800 ± 250 d 1.350 ± 100 ? ? 600 ± 150 1.530 — 250 0 0 0 1.650 ± 400 0 0 0 0 ? ? ? 50 ± 50 c — 0 0 0 0 Same 180 ± 10 c ? 500 ± 100 800 ± 250 c ? 1.250 ± 100 800 ± 250 c 350 ± 150 c 1.000 ± 300 3.0 Same Same Same Same Same Same Same 71.200 ± 300 c ? 0 2.250 ± 100 800 ± 250 c ? 500 ± 150 (15.200 ± 100 400 ± 125 c 750 ± 150 475 ± 100 c 2.400 3.125 1.800 ± 400 c 2.350b (1.0 –O– ether >CO ketone –CHO –COO-ester –COOH –OH 24.8 10.000 ± 200 c 1.5 74.800 ± 325 550 ± 275 —f 3.000 ± 500 ? 300 ± 50 1.000 ± 250 c 4.400b 9.125 1.0 70.200 ± 250 c ? ? 0 2.0 CH3 33.960b 4.000 ± 200 1.500 ± 175 ? ? 2.750 ± 300 c 1.0 109.550 ± 250 1.2 35.700 ± 150 ? ? 1.1 –1.315b 2.400 ± 100 3. 2 (cal/mol ΔVδP Electronic Transfer Parameter.0 73.900 7.

0 19. M.050 ± 450 c ? ? ? 0 2.0 ? ? 1. c Based on v ery limited data.050 ± 300 1. Cop yright ASTM. Paint Testing Manual. R..000 ± 600 9. Sci..000 ± 10%)/V ? ? 600 ± 350 c ? ? ? 3.F .600 ± 600 600 ± 200 100 ± 50 ? (81. d Includes unpublished infrared data.300 ± 600 10. P f Twin and triplet v alues apply to halogens on the same C atom.000 ± 600 1.2 4.7) 28. Manual 17.000 2. American Society for Testing and Materials.440 Same ෇(OH)2 twin or adjacent –CN –NO2 –NH2 amine >NH2 amine –NH2 amide ϵPO4 ester 24. 388. with V for total compound.150 7.. These values apply to halogens attached directly to the ring and also to halogens attached to aliphatic double-bonded C atoms. 11 . C. P Solubility Parameters — An Introduction Source: From Hansen.500 ± 100 ? ? 9..700 ± 550 c 3.850) (7.0 24. e xcept that ΔVδ 2 also includes those on adjacent C atoms.000 (5. 14(2). A method for estimating both the solubility parameters and molar v olumes of liquids. values are for information only and not to be used for computation. With permission.350 ± 200 750 ± 200 2. in ma y cases. 472.600 ± 850 c 3. Ed.150 ± 225 ? —e a b Data from Fedors.550 ± 125 150 ± 150 c ? ? ? 4. K oleske. e Use formula in ΔVδ 2 column to calculate. Philadelphia. Polym. Limits sho wn are roughly 95% confidence.26.000 ± 800 c 3.5 (6. 1974.000 3.000 ± 500 400 ± 125 c 350 ± 50 c 2.. Reprinted with permission. J.400) Same Same Same Same Same 0 ? ? 1.000) Same (4.750 ± 300c 1.250 ± 200 d ? ? ? 4.0 Same Same Same Same 1. Eng. 147–154. V.0 Same 32.750 ± 100 800 ± 200 ? ? ? 500 ± 200 d 400 ± 50 d 1. 1995.

012 Cyclic.015 0.47 0.192 0.017 — — — 0.320 — — — 0.0493 0.0245 0.012 0.0521 0.0149 0.198 0.060 — — — — — — 0.0317 0.011 — — 0.09 0.0863 0.227 0. ΔP — 0.0131 0.17 0.0318 — 0.008 — — — — — — — — — ΔP T 0.0040 0.154 — 0.36 — — — — — — 0.2 Lydersen Group Constants Group CH3 CH2 >CH– >C< CH2 CH– C< CH aromatic CH aromatic –O– >O epoxide –COO– >C෇O –CHO –CO2O –OH→ –H→ –OH primary –OH secondary –OH tertiary –OH phenolic –NH2 –NH– >N– –CϵN –NCO HCON< –CONH– –CON< –CONH2 –OCONH– –S– –SH –Cl 1° –Cl 2° Cl1 twin Cl aromatic –Br –Br aromatic –F –I Aliphatic.13 — — — — — — — 0.0938 0.0897 0.014 — — 0.095 0. ΔP 0.011 0.021 — 0.0226 0.007 — 0.198 — — 0.50 — 0.031 0.0184 0.154 — — 0.12 — — 0.0345 0.018 0.018 0.0539 0.000 0.227 0.224 0.0546 0.0729 0.033 — — — — — — — — — 0.0093 0.0843 0.0392 0.0313 0.210 0. ΔT — 0.0440 0.06 — — — — –0.014 0.0343 –0.0274 0.184 0.020 0.0497 0.0178 0.006 — Aliphatic.0175 0.007 — — — — — — — 0.020 0.198 0.83 Cyclic.27 — 0.0400 0.048 — — — 0.0200 0.040 0.210 0. ΔT 0.0539 0.031 0.16 — 0.0060 0.018 0.0077 0.0311 0.0192 0.0129 0.135 0.02 0.000 — — 0.12 Hansen Solubility Parameters: A User’s Handbook TABLE 1.154 0.24 — — — — — — — — — .012 –0.024 0.0593 0.29 0.047 0.082 — — 0.015 0.02 — — — — — — — — 0.013 0.0267 0.33 — 0.010 — 0.035 0.0445 0.

as the Hoy procedure does not appear to fully separate the polar and h ydrogen-bonding ener gies. whereas the other two figures directly report the dispersion cohes ve energy density. A comparison with related compounds or the similarity principle gi ves better results than those found from the van Krevelen dispersion group contrib utions. the solubility parameters in SI units (MP a1/2) are 2.0118 0. 1995. Ed. J. All three of these figures h ve been straight-line e xtrapolated into a higher range of molar v olumes than that reported by Barton. ΔP — — — — — — — — — — — — Cyclic. Ho y subtracts estimated v alues of the polar and h ydrogen-bonding energies from the total ener gy to find the dispersion ene gy. These figures h ve been inspired by Barton. American Society for Testing and Materials. depending on whether the molecule of interest is aliphatic.0020 0. C.003 –0. ΔT — — — — — — — — — — — — Cyclic.2 (CONTINUED) Lydersen Group Constants Group Conjugation cis double bond trans double bond 4 Member ring 5 Member ring 6 Member ring 7 Member ring Ortho Meta Para Bicycloheptyl Tricyclodecane Aliphatic. In the follo wing. CALCULATION OF THE DISPERSION SOLUBILITY PARAMETER δD The δD parameter is calculated according to the procedures outlined by Blanks and Prausnitz.S. The latter is much simpler to use.3 can be used to find this paramete . Solubility parameters. Barton also presented a similar figure for the aliphatic sol ents. calculation procedures and e xperience are presented according to the procedures most reliable for the e xperimental and/or ph ysical data a vailable for a gi ven liquid. Energies found with these extrapolations have also provided consistent results. but values estimated independently using the v an Krevelen dispersion parameters are clearly lo w.0034 0. ΔP — — — — — — — — — — — — Source: Hansen..9 who converted earlier data to Standard International (SI) units. This allows for more calculational error and underestimates the dispersion energy.. The v an Kre velen dispersion parameters appear to be too lo w. in Paint Testing Manual. being completely dedicated to the full procedure based on corresponding states described here. The figure for the aliphatic liquids g ves the dispersion cohesive energy.0060 0. for example. As noted earlier. c.0015 0.0069 0. which still finds xtensive use in the U. P A. 12 Figure 1.0035 –0. Manual 17.Solubility Parameters — An Introduction 13 TABLE 1. cycloaliphatic. The author has not attempted these calculations. 383–404.0455 times lar ger than (ca1/cc) 1/2 in the older cgs centimeter gram second (cgs) system.0010 0. Koleske.V. ΔT — — — — — — — — — — — — ΔP T 0. Philadelphia..1 to Figure 1. ED. Reprinted with permission.M. is consistently lo wer than that found by Hansen. but it . in particular..0035 0. or aromatic. as one need only tak e the square root of the v alue found from the figure to find the respec ve partial solubility parameter.0095 Aliphatic.0010 –0.

.5 0. Paint Testing Manual. 389.70 30 20 10 0 50 100 150 200 250 V.70 50 60 70 80 90 3 300 250 100 110 120 130 V.1 Energy of v aporization for straight chain h ydrocarbons as a function of molar v olume and reduced temperature. cm /mol FIGURE 1. K oleske..60 0.60 0.. C. Philadelphia.) 400 350 c..45 0. J.5 5 0. Cop yright ASTM.2 Cohesive energy density for c ycloalkanes as a function of molar v olume and reduced temper ature. Ed.40 0.4 5 0.50 0.4 0 0. Cop yright ASTM. American Society for Testing and Materials. MPa Tr = 0. Ed. V.. 1995. Reprinted with permission.) .. (From Hansen.65 40 0. cm3/mol FIGURE 1.14 Hansen Solubility Parameters: A User’s Handbook Δ ED kJ/mol 70 60 = Tr 50 0 0. M. Paint Testing Manual. J. Reprinted with permission. V. K oleske. (From Hansen.65 0. M.55 0. 389. 1995. American Society for Testing and Materials. C. Philadelphia. Manual 17. Manual 17.

A table of the L ydersen group contributions. Tc. American Society for Testing and Materials.15 K. These corrections are applied by first finding the dispersion cohe ve energy from the appropriate figure. M. The critical temperature. it must be estimated. 1995. group contribution. The dispersion cohesive energy is then increased by adding on the correction factor. the desired groups may not be in the table. Its use is not recommended. Liquids such as chlorobenzene. is also required in this calculation. Paint Testing Manual. or some other technique. Tb.55 250 0.2. C.. This requires multiplication by the molar v olume for the c yclic compounds using data from the figures here. and a weighted a verage is tak en for the molecule of interest based on the number of significant atoms.11) where T has been tak en as 298. V. such as chlorine. etc. Such weighting has been found necessary to satisfy Equation 1. Cop yright ASTM.65 0. The normal boiling temperature. b ut not for oxygen or nitrogen that ha ve a similar size. It has been found that corrections are required for atoms significantly lager than carbon. Ed. sulfur. The size of the atom is important. When substituted cycloaliphatics or substituted aromatics are considered. Reprinted with permission. J.45 300 0. and ring compounds with alk yl substitutions that have only tw o or three carbon atoms ha ve been considered only as c yclic compounds. The dispersion parameter is based on atomic forces.40 c. hexyl benzene w ould be the arithmetic a verage of the dispersion energies for an aliphatic and an aromatic liquid.41 ΔT .3 Cohesive energy density for aromatic hydrocarbons as a function of molar v olume and reduced temperature. This correction factor for chlorine. as given by Hoy40 for calculation of the critical temperature is included as Table 1. and sulfur has been .. 389. Philadelphia. The dispersion ener gies are e valuated for each of the molecules in volved..11 and Equation 1. is required to use the dispersion ener gy figures.50 0. bromine. The end result does not appear too sensiti ve to these situations.Solubility Parameters — An Introduction 15 400 Tr = 0. The Lydersen group contrib ution method in volves the use of Equation 1. simultaneous consideration of the tw o separate parts of the molecule is required.12) (1. K oleske.6.60 0. (From Hansen.12 as follo ws: Tb/Tc = 0.567 + ΣΔT – ( ΣΔT)2 and Tr = T/Tc (1. toluene.70 80 90 100 110 120 130 V. each with the gi ven molar v olume of hexyl benzene.. a these figures g ve the cohesive energy densities. cm3/mol 140 150 160 170 FIGURE 1. Manual 17.) is inconsistent with the ener gy figure and in erro . which requires some educated guessing. In some cases. The carbon atom in hydrocarbons is the basis of the dispersion parameter in its present form. or example. MPa 350 0. bromine. This is not always available and must be estimated by similarity. If the critica temperature cannot be found.

13) The constant 37.4 gi ves this parameter in SI units. an entity that is lar ger than the corresponding compound containing only carbon (i. the combined polar and h ydrogen bonding parameters as used by Hansen. An outstanding e xample of v ariations of this kind can be found with carbon disulfide.13 has been consistently used by the author over the past years. of course. CALCULATION OF THE POLAR SOLUBILITY PARAMETER δP The earliest assignments of a “polar” solubility parameter were gi ven by Blanks and Prausnitz.6 to Equation 1.16 Hansen Solubility Parameters: A User’s Handbook taken as 1650 J/mol for each of these atoms in the molecule. The first Hansen pola parameters13 were reassigned new values by Hansen and Skaarup according to the Böttcher equation (Equation 10.08 in hexane. and the dielectric constant.4 in carbon tetrachloride.21 in nitrobenzene. both for interpretation of experimental data and to allo w Equation 1. its homomorph).6 (or Equation 1. The need for these corrections has been confirmed ma y times. as well as other relevant parameters. and it has a higher disper sion solubility parameter without an y special need for corrections. because square roots of rather large numbers are taken. in f act. There is a clear increase with increasing solubility parameter of themedia. similarity with other compounds. Larger problems arise because of Equation 1. It can be seen from the dispersion parameters of the c yclic compounds that the ring has an effect similar to increasing the ef fective size of the interacting species. The molar volume must be known or estimated in one w ay or another. particularly in view of its reported reliability.42 Other data sources also have the same. The reported dipole moments are mostly 0 for g as phase measurements.25). This leaves only the dipole moment to be found or estimated.4(DM)/V1/2 (1. and 1.6.8 to balance. with the most extensive listing still being McClellan. It must be noted that the f act of zero dipole moment in symmetrical molecules is not basis enough to assign a zero polar solubility parameter . Equation 1. This group appears to act as if it were. as the corrections are relati vely small. Di viding by the molar v olume and then taking the square root gi ves the (lar ge atom corrected) dispersion solubility parameter . and the y are higher for aromatic compounds than their corresponding c ycloaliphatics. Taking square roots of the lar ger numbers in volved with the lar ger molecular species reduces the errors in volved in these cases. as discussed later . but many interesting compounds are not included in the DIPPR. group contrib utions. When no dipole moment is a vailable.. 0. These are not available for many compounds. Similar effects also appear with the ester group.35 This reported reliability appears to be correct. and the calculation used is more dif ficult than the much simpler equation d veloped by Hansen and Beerbo wer17: δP = 37. the refracti ve index.e. lar ger for the smaller molecular species. The impact of these corrections is. 44 Many fluoropolymer . 0. This is one reason group contrib utions are recommended in some cases. in ef fect.8). Ener gy assigned to the dispersion portion cannot be reused when finding the other partial parameters using Equation 1. The Design Institute for Ph ysical Property Research (DIPPR) 43 database has been found useful for man y compounds of reasonably common usage.49 in chlorobenzene. or e xperimental data can be used to estimate the polar solubility parameter . Research is definitel needed in this area.15 This equation requires the molar v olume. supplemented by 0. and the y cannot be considered polar in the current conte xt. The careful evaluation of the dispersion cohesi ve energy may not ha ve a major impact on the value of the dispersion solubility parameter. and data such as latent heats and critical temperatures. The dispersion energies for cycloaliphatic compounds are lar ger than their aliphatic counterparts. Standard reference works have tables of dipole moments. The latter and the highest v alue has been found e xperimentally most fitting for correlating perme ation through a fluoropolymer film used for chemical protec ve clothing. 12 These parameters were. the dipole moment (DM).

47 . Here. At this stage. Latent heat data at 25°C ha ve been found consistently from those at another temperature. but also to arri ve at all three parameters by statistical mechanics. group contributions are judged to be the best alternative. This is still widely used where the required data are a vailable and reliable. ho wever. the polar solubility parameter has been found using the Beerbower table of group contrib utions. even moderate v ariations in reported heats of v aporization can cause severe problems in calculating the polar (or h ydrogen bonding) parameter when Equation 1. Diet yl ether has also presented problems as an outlier in terms of dissolving or not and permeating rapidly or not. and similar results are found by the approach of the first paragraph as well as with statistical thermodynamics SUPPLEMENTARY CALCULATIONS AND PROCEDURES The procedures listed pre viously are those most frequently used by the author in calculating the three partial solubility parameters for liquids when some data are a vailable. in f act. The question in each case is.Solubility Parameters — An Introduction 17 have considerable polarity. with a preferred v alue of 1.79 in chloroform. The above paragraph is not changed from the first edition of this handbook. The chameleonic c yclic forms of the linear glycol ethers would also seem to provide a basis for altered dipole moments in various media. the reported dipole moments 42 vary from 0. CALCULATION OF THE HYDROGEN BONDING SOLUBILITY PARAMETER δH In the earliest w ork. the similarity and group contrib ution methods are thought best. For large molecules.0. Therefore. These new polar group contrib utions then become supplementary to the Beerbo wer table. the hydrogen bonding parameter w as almost always found by subtracting the polar and dispersion ener gies of v aporization from the total ener gy of v aporization. In such cases.13 cannot be used. Similarity to related compounds can also be used. The latent heats are not generally a vailable with suf ficient accura y to allo w subtraction of tw o large numbers from each other to find a ery small one.38 When Equation 1.8 are strictly adhered to. Again. more reliable than estimating se veral other parameters to ultimately arri ve at the subtraction step just mentioned. of course. by similarity to related compounds and/or by subtraction of the dispersion and h ydrogen bonding cohesi ve ener gies from the total cohesi ve ener gy. latent heats found in a widely used handbook 45 are not clearly reported as to the reference temperature. and the result of such a procedure should be essentially the same as for using group contrib utions.6 or Equation 1.74 to 2. There is an indication that these are 25°C data. the accurate calculation of the relatively small polar (and h ydrogen bonding) contributions present special difficulties. Choosing a gi ven value seems rather arbitrary . in the absence of reliable latent heat and dipole moment data. but checking indicated man y of the data to be identical with boiling point data reported else where in the literature. Unfortunately .17 and 1. Subsequent editions of this handbook 46 have a completely different section for the latent heat of e vaporation. It no w appears possible not only to calculate the h ydrogen bonding parameter independently. Whether or not one understands all of the equations and methodology of Chapter 3. the procedure in itself confirm the need for (at least) three cohesion ener gy parameters. This is to emphasize the importance of the w ork of P anayiotou and co workers reported in Chapter 3. “Which data are a vailable and judged most reliable?” Ne w group contributions can also be developed from related compounds whose dipole moments are available. especially those with long h ydrocarbon chains. The lower dipole moments seem to fit in other instances. using the relation gi ven by Fishtine. This is clearly a major step forw ard. the group contribution techniques are considered reasonably reliable for most of the required calculations and. There are a number of other calculations and procedures that are also helpful.

Plotting the parameters for homologous serie among the esters. and glycol ethers has aided in finding th parameters for related compounds. When the latent heat at the boiling point is given in cal/mol.5 α) (1. δP . it is assumed that the parameters all demonstrate the same temperature dependence. in particular. TEMPERATURE DEPENDENCE Only very limited attempts have been made to calculate solubility parameters at a higher temperature prior to the second edition of this handbook. An additional aid in estimating HSP for many compounds is that these parameters can be found by interpolation or e xtrapolation. where α is the coef ficient of thermal xpansion17: dδD/dT = –1. δP . As the temperature increases. especially for homologous series. nitroparaf fins. E.22 × 10 –3 + 0. It appears as if the predictions are not significantly in erro when e xperimental data are a vailable for checking.15) where R is the g as constant and T is the absolute temperature. can be estimated by the following equations. In this direct but approximate approach. The g as-phase dipole moment is not temperature dependent. T. which will also change its cohesi ve energy density. The first member may no necessarily be a straight-line extrapolation. The inclusion of Chapter 3 and Chapter 10 in this handbook helps by pro viding a more accurate treatment of temperature dependence when the situation warrants it. of course. although the v olume of a flui does change with the temperature. and to date no problems with particularly faulty predictions ha ve been noted. Solubility parameter correlations of phenomena at higher temperatures ha ve generally been found satisfactory when the established 25°C parameters ha ve been used. and δH parameters for liquids with temperature. good or bad. The program then locates that set of δD. Equation 1. It might be noted in this connection that the ydrogenh bonding parameter.18) .5 αδP dδH/dT = – δH(1. in cgs units at this temperature: E = ΔEv = ΔHV – RT (1. The result is consistent with all the other parameters. that ha ve good solv ent quality. The change of the δD. etones.16) (1. A computer program has been de veloped by the author to assign HSP to solv ents. and this parameter will decrease more rapidly than the others. but comparisons with related compounds should always be made where possible to confirm assignments.18 Hansen Solubility Parameters: A User’s Handbook ΔHv(T1)/ΔH(T2) = [(1 – Tr1)/(1 – Tr2)]0. to find the total cohesion ene gy.14) This is done e ven if the melting point of the compound being considered is higher than 25°C. and outside the appropriate spheres where the solv ent quality is bad. into the program for a reasonably lar ge number of polymers where the solubility parameters. The procedure is to enter solvent quality. more and more h ydrogen bonds are progressi vely brok en or weak ened.15. Recalculation to higher temperatures is possible b ut has not generally been found necessary . and δH parameters for the solv ent that best satisfies the requirements of a locatio within the spheres of the appropriate polymers. RT equal to 592 cal/mol is then subtracted from this according to Equation 1.25 αδD dδP/dT = –0. is not the case. which. based on experimental data alone. is the most sensitive to temperature.17) (1. and appropriate radius of interaction for the polymers are kno wn. This has been used in se veral cases where the parameters for the gi ven liquids were desired with a high degree of accuracy.38 (1. alcohols.14 is used to estimate the latent heat at 25°C.

Water is also well kno wn as an exceptionally good plasticizer because of its small molecular size. SOME SPECIAL EFFECTS TEMPERATURE CHANGES Water (and methanol) uptak e in most polymers increases with increasing temperature. glycols. A related problem has been encountered with methanol. and glycol ethers become better solv ents for polymers of lower solubility parameters as the temperature increases. and. So it is not surprising . permeation. even though they may have identical solubility parameters. Figure 12.Solubility Parameters — An Introduction 19 Higher temperature means a general increase in rate of solubility/dif fusion/permeation.4 They can be explained either by the change in solubility parameter with temper ature or more completely by the Prigogine CST . The effects of temperature changes on solubility relations are most ob vious with systems ha ving a high h ydrogen-bonding character. whereas that of most polymers remains reasonably constant. This leads to premature f ailure of polymeric products used in such environments. The methanol that had been absorbed into the coating near the surface became insoluble as the temperature of the coating near the surf ace was lowered by the evaporation of e xcess methanol during the handling and weighing of the panels. This is because the solubility parameters of the w ater and the polymer are closer at higher temperatures. and this ef fect would be most significant for those liquids with la ger δH values. dif fusion. as mentioned earlier . and chemical resistance phenomena.3. Figure 8. as can be seen by a comparison of Equation 1. when highly crystalline polymers or solids are being tested for solubility . 1.29 Solvents with smaller molecular size ha ve also been repeatedly noted as being superior to those with lar ger molecular size. These phenomena are discussed in great detail by P atterson and coworkers. it is possible that a boundary solv ent can be a good solv ent at a gi ven temperature.3 and Figure 12.4 sho w how this effect can be measured e xperimentally with an increase in w ater content abo ve the equilibrium v alue when temperature cycling is encountered. Thus. δD. the Hildebrand solubility parameter theory also points to smaller molar volume solvents as being better than those with lar ger molar volumes. through Equation 1. and δH decrease with increased temperature. The δH parameter of w ater (and methanol) f alls with increasing temperature. This means that alcohols. Smaller molecules tend to be more readily soluble than lar ger ones. a f act that is often noted in practice. EFFECTS OF SOLVENT MOLECULAR SIZE The size of both solv ent and solute molecules is important for solubility . The presence of water in a film can also influence the upt e of other materials. One should be aware that the changes in the δH parameter would be larger than those in the other parameters. Blistering was encountered in the coating near the air surface shortly after the experiment began. with h ydrophilic materials being more prone to enter the film This can cause blistering on rapid cooling as discussed in Chapter 12 and in Reference 48 (see Chapter 8 and Chapter 12).3 sho ws how rapid cooling from a w ater-saturated state at higher temperature can lead to blistering. phenols.2 This fact of e xpected improved solvency for smaller molecules is also known from the Flory–Huggins theory of polymer solutions. increasing the temperature can cause a nonsolv ent to become a good solv ent. It w as intended to follo w the rate of uptake of methanol in an epoxy coating at room temperature by weighing coated-metal panels periodically on an analytical balance. This is a rather extreme case. such as in solubility parameter studies or resistance testing. The presence of dissolv ed water not only softens (reduces the glass transition temperature) a polymer as such. as well as larger solubility parameter spheres. Examples are given in the ne xt section for some special situations in volving water and methanol. As mentioned earlier . b ut it also means dif fusion rates of other species will be increased.18. As stated previously.16. use of the HSP (determined at 25°C) at ele vated temperatures can most often be done without too much trouble from a practical point of vie w. δP . b ut turn bad with either an increase or a decrease in temperature.

21 This may be because the size ef fect is most often not caused due to the thermodynamic considerations on which the solubility parameters are based. whereas SPHERE1 considers data for only those solv ents considered as “good. The similarities in the HSP approach and the Prigogine theory. relation between the Prigogine ρ (segment size parameter) and the δD parameter. but this has not been e xplored in relation to the present discussion. Smaller and more linear molecules diffuse more rapidly than lar ger and more b ulky ones. Sorting output data according to the molecular v olume of the test solv ents in a computer analysis helps to discover whether solvent molecular size is indeed an additional significant actor in a given correlation or testing program. if not identical.A n)1/n (1. This can be solution or not. The increase of δD with increasing molecular size among the aliphatic hydrocarbons. The DATA FIT approaches 1. and attainment of equilibrium.” including size of polymer molecules. The size and shape of the solvent molecule are also very important for kinetic phenomena such as diffusion. 16. They are very similar.0 as the fit impr ves during an optimization and reaches 1. and this may affect the evaluations. of course) parameters based on all the data. in some cases. fillers. permeation coefficients higher than a gven value.50 Polymer samples used for solubility parameter or other testing may well retain solvent or monomer for many years. breakthrough time being less than a given interval. b ut rather through a kinetic effect of diffusion rates or other free volume considerations. discussed in Chapter 2. 20. 49 Likewise. an fibers. The dif fusion coef ficient may be so l w that equilibrium is not attained for hundreds of years at room temperature. indicate a remarkably close. These are called SPHERE and SPHERE1. What is considered good or bad v aries according to the level of interaction being studied.0 when all the good solvents are included within the sphere and all the bad ones are outside the sphere. suggesting that molecular size dif ferences are at least partially accounted for in the δD parameter. COMPUTER PROGRAMS The author has used tw o computer programs extensively in his own studies and in collecting material for this book. Attempts to include the molecular v olume in ne w composite solubility and size parameters have not been particularly successful. A “1” indicates a “good” solv ent. whereas a “0” is used for a “bad” solv ent. a gi ven percentage of swelling or uptak e. the higher δD values for the larger units represented by cycloaliphatic and aromatic rings.. Specific xamples are given in Chapter 5 and Chapter 12.” It neglects the nonsolv ent data. SPHERE1 has been most useful in correlations with pigments.. This was demonstrated in common solvent exposures of rigid polymers like polyphenylene sulfide (PPS) with thicknesses o several millimeters. in chemical resistance. The program systematically e valuates the input data using a quality-of-fit function called th desirability function. etc. The Prigogine theory also has a parameter to describe “structural ef fects. permeation. and the need for corrections for lar ger atoms discussed earlier tend to support the molecular size dif ferences.19) where n is the number of solv ents for which there is e xperimental data in the correlation. as described in Chapter 7 The data input is by solv ent number follo wed by an indication of the quality of interaction with that solv ent. The function has the form: DATA FIT = (A 1 * A2 *.20 Hansen Solubility Parameters: A User’s Handbook that solvent molecular size can be an important fourth parameter in solubility and. the only difference being that SPHERE optimizes the polymer (or other material. the second stage in the tw o-stage drying process in polymer fil formation by solvent evaporation can last for man y years. long-time suspension of a pigment.51 This suggestion w as made by a reputed statistician man y years ago as the most appropriate statistical treatment for this type of problem. . It has been in use since the late 1960s.

For a bad solv ent inside the sphere. an error enters the D ATA FIT according to: Ai = e +(Ro – Ra) Such errors are often found for solv ents having low molecular v olumes. or a “0” for a bad one. Solvents with large and/or small molecules that gi ve the “errors” can sometimes be (temporarily) disregarded by generating a ne w correlation. and eight points around it are e valuated in a similar manner. given as D (δD).4.21) Such errors can sometimes be found for lar ger molecular species such as plasticizers. Different radii are evaluated at each of these points in the optimization process. and the total solv ents (T) in the correlation. with the current best values as center. based on an a verage of each of the HSP for the good solvents only. and finally to 0. This may suggest a new analysis of the data. especially the older ones. Experimental data for the solv ents are entered with solv ent number (comma) and a “1” for a good solvent.22) (1. such as for reactions or some e xtraneous effect not predictable by the solubility parameter . Errors in the correlations are indicated with an “*” in the SOLUB column where the e xperimental input data are indicated. As stated abo ve. which mak es general predictions v ery difficult. The initial length of the cube is 1 unit. The output of this program is for the least radius allo wing the maximum DATA FIT.The computer printouts all include a column for the RED number . and Ra is the distance from a gi ven solvent point to the center of the sphere. the number of good solv ents (G). It could either denote being good and outside the sphere or being bad and inside the sphere. This latter information has not always been recorded and may be lacking for some correlations.20) The Ai quotient for a given good solvent within the sphere and a bad solvent outside the sphere will be 1. The length of the edge of the cube is then reduced in size to fine tune the fit. F or a good solv ent outside the sphere. They may also be bad data. and rechecking data indicated with an “*” in the output has become a routine practice. unit in the last optimization step. the point with the best fit is t en as a ne w center. Nonsystematic errors may be real. P (δP). respectively. There is a special printout with the solv ents arranged in order of molecular volume that helps to analyze such situations. This continues until the DATA FIT cannot be impro ved upon. Water is still so unpredictable that its use as a test solv ent in solubility parameter studies . and H (δH). When better fits are found among the eight points. Ro is the radius of the sphere.3 unit. This is not unexpected for the reasons mentioned earlier .Solubility Parameters — An Introduction 21 Ai = e –(ERROR DISTANCE) (1. The error distance is the distance of the solv ent in error to the sphere boundary . the contrib ution to the D ATA FIT is Ai = e +(Ra – Ro) (1. Results from the SPHERE program reported in this book generally include the HSP . as well as the DATA FIT.0. The behavior of w ater often depends on its local en vironment. and Ro for the correlation in question. which is reduced to 0. The program then e valuates eight points at the corners of a cube. An example is found in Table 5. this gi ves an excellent DATA FIT for an abbre viated range of molecular v olumes. HANSEN SOLUBILITY PARAMETERS FOR WATER Water is such an important material that a special section is dedicated to its HSP at this point. systematic errors can sometimes be seen in the molar volume printout. The program assumes a starting point.

covers such high ener gies that no liquid can be found.1 13. This is true of w ater as a pure liquid or in mixtures. The HSP reported here are the center points of HSP spheres where the good solv ents are either those that are completely miscible or those that are miscible to only 1% or more. The averaged values are very questionable as water can associate. it has been found in a study involving water.9) is still used for these correlations. This is practical and almost necessary from an experimental point of vie w as most solv ents are so miscible as to not allo w an y e xperimental characterization in terms of a solubility sphere.3 16. It should also be mentioned that amines were a major source of outliers in these correlations. and 1. The last entry in Table 1. 52 as well as for the total w ater miscibility. The HSP for w ater as a single molecule. More recently. Based on SPHERE1 Program. Traditionally.1 12. No solids were included. ethanol. based on the latent heat at 25°C is sometimes used in connection with mixtures with water to estimate average HSP.3 is for an analysis using the SPHERE1 program.0 13. Their use to predict solubility relations for amines and for solids must therefore be done with caution.22 Hansen Solubility Parameters: A User’s Handbook TABLE 1. Table 1. suggest that about six w ater molecules associate into units. discussed at some length in Chapter 2. “Bad” solvents dissolve to a lesser e xtent. The cohesion energy is so high as to require solids. The second set of data is based on a correlation of the solubility of various solvents in w ater. The second and third entries in Table 1.1 18. A frequent problem is that a considerable portion of the HSP spheres. as discussed previously. The third set of data is for a correlation of total miscibility of the given solvents with water.5 16.3 includes data from various HSP analyses of the behavior of water.4 17. and w ater has a v ery small molar v olume as a single molecule.3. The first set of data is der ved from the energy of vaporization of water at 25°C. The data for the 1% correlation.9 15. where “good” solv ents are soluble to more than 1% in w ater. solvents are considered as points. An exception to this is the data for w ater reported in Table 1. primarily based on successes at lower levels of cohesion ener gies. such as in the case for w ater.1 δP 16.1 18.856 0.5 15. The HSP data are for the minimum sphere that encompasses only the good solv ents. The justification fo . CONCLUSION This chapter has been dedicated to describing the tools with which dif ferent HSP characterizations can be made and some of the pitf alls that may be encountered in the process. The constant 4 in the correlations (Equation 1.000 G/T — 88/167 47/166 47/47 Based on SPHERE program. This type of comparison usually results in some of the parameters being lo wer than when all the data are included.3 are based on the SPHERE program where both good and bad solv ents af fect the D ATA FIT and hence the result of the optimization. It almost appears to ha ve a dual character .9 δH 42.3 HSP Correlations Related to Water Correlation Water — Single molecule Water — >1% soluble in a Water — Total miscibility 1 a Water — Total miscibility “1 b” a b δD 15.5 Ro — 18. The bad solv ents are simply not considered in the data processing.9 FIT — 0.0 20.2-propanediol that the HSP for water indicated by the total water solubility correlation could be used to explain the behavior of the mixtures involved. is not recommended. but this is also supported by the comparison with the Prigogine cst of polymer solutions.880 1.

1-b utanol (B). The amines are ethyl amine (E) and propyl amine (P).4 δp vs.4 is included to sho w where many common solvents are located on a δp vs. The glycols are ethylene glycol and propylene glycol. the tools is further confirmed in Chapter 2 and Chapter 3. Hydrogen Bonding Parameter FIGURE 1. δH plot showing the location of various common solvents. ethanol (E). The alcohols include methanol (M). The amides include dimeth yl formamide (F) and dimeth yl acetamide (A). The ethers are symbolized by dieth yl ether. The phenols are phenol ( P) and m-cresol ( C). and 1-octanol (O). Figure 1. Polar Parameter Amides F B A Ketones 8 Alcohols M E E P Glycols 12 M C Chlorinated E T Phenols P C O E Amines 8 12 16 20 B 4 B Esters Ethers Tol P 4 0 0 I 24 28 δH. .Solubility Parameters — An Introduction 23 Alcohols Glycols Amides Phenols Chlorinated Chloroform Dichloromethane Nitriles Dimethylsulfoxide Esters Ethers Amines Methyl ethyl ketone Cyclohexanone THF Toluene Isooctane Nitromethane Bold type indicates relatively high δD 24 20 N Nitriles A S 16 δP. The nitriles are acetonitrile (A) and butyronitrile (B). and their use is demonstrated in all th subsequent chapters. δH plot. The esters are ethyl acetate (E) and n-butyl acetate (B).

Hildebrand. LA. K. 23. 42(550). 24. J. Hansen. Eng. Solubility parameters for film formers.. 8. Vol. and Long. J. 489–493. Patterson. Characterization of surfaces by spreading liquids. Anonymous [Note: This was. C.M. Barton.. 17. and Scott. 1967. ..M.. 1971.. Soc. A. Eds.F. 1972. H. Ne w York. 39(511). Fed. 29(394). Applications of solubility parameters and other cohesion ener gy parameters. Macromolecules. D. Fed. 9. 511–514. Hildebrand. J. C. Application of the acid/base concept describing the interaction between pigments. Hansen.. Thermodynamics of polymer solubility in polar and nonpolar systems. Congress Book. 8.. 505–510. and Somcynsky. Chem. A. 10. Introduction to thermodynamics of polymer solubility . Role of free v olume changes in polymer solution thermodynamics. Burrell. Paint Technol. and Patterson. Paint Technol. 104–117. A. Solubility parameters.M. 44(570). Copenhagen. 1968. P . Doc toral dissertation.R. J.... Fed. pp. C.M. 1967. J.L. Paint Technol.... Paint Technol. Prentice-Hall. Sci. S. August 25–30. 3. L. Ne w York. Polym. Paint Technol. 1967. and Delmas.. G. Using Dimethyl Sulfoxide (DMSO) in Industrial F ormulations.. in fact. 4. 2nd ed. and solv ents. 1970. J.. Ind. 660–664. 19.. Chem. 905–912. 14. Vernis. binders. and Prausnitz.S. 6. C. 102. 1969. Barton. J. Hansen.. Biros. C. E. 39(511).. Characterization of Coatings: Physical Techniques.L. The three dimensional solubility parameter — k ey to paint component af finities II. 1971. 27(369).M.. Chem. and Beerbo wer. 1991. 2126–2136. 12. Pure Appl. 1964. Dig.. 1985. 1972. 2. On the use of cohesion parameters to characterize surf aces. Ne w aspects of polymer solution thermodynamics. 1962. B. 1962.M. D.. 30. 1957. Slidell. CRC Press. J.W .24 Hansen Solubility Parameters: A User’s Handbook REFERENCES 1.M. Myers.M. 1962.. Dig. D.N. C. T. 13.... 34(450). Hansen. 57.. Patterson.. D. and Wallström. Marcel Dekk er.. Anal. J. Van Dyk. H. Part II. Hansen. 1964. 31–39. Paint Technol. 21–30. 20. The three dimensional solubility parameter — key to paint component affinities III. J. 3(1).R.. Ed. 275–286.. 16.. Reinhold. Ne w York.. 15. 47(602).M. 1978. D. Off. Karger. Surface dewetting and coatings performance. J. 2. 1992.. Soc. H. and Teas. J. Van Dyk. T. Regular Solutions. Technol. D. 50(14). and Skaarup. 3379–3389. 41(536). 18. 25.. 1159–1173.W. 39(505). Dig.. 57–60. in Treatise on Coatings. Fundam.. Paint Technol. 3rd ed.M. 16. Zeman. Danish Technical Press.. 21. Suppl.. C. Burrell. Soc. 726–758. L. but this does not appear on the b ulletin]. Ne w York. 5. in Kirk-Othmer Encyclopedia of Chemical Technology... Interscience. Prediction of the C parameter by the solubility parameter and corresponding states theories.. R.. FL. Sci. Boca Raton..L. R. and Scott. J. J. chap. Solubility parameters.. 2nd ed. Off. Gardon.M. Patterson. Bhattacharyya.. 11.. Sn yder. 889–910... Handbook of Solubility Parameters and Other Cohesion Parameters. Hansen. 15.. Delmas. 1983. 677–692..F. Paint Technol. 7. Patterson. J.. 1976. The Solubility of Nonelectrolytes. Thermodynamics of polyisobutylene-n-alkane systems. Englewood Cliffs. 30–35. Adhes. A solvent formulating chart. Polym.. J. Bulletin No. 1983.. Patterson. The Three Dimensional Solubility P arameter and Solvent Diffusion Coefficient. Blanks. 1991. R. J. Hansen. 1967.P. 1–8. 1950. NJ. Sørensen. Vol. Off.D. The principle of corresponding states and the excess functions of n-alkane mixtures. Part C.. Expanded solubility parameter treatment for classificatio and use of chromatographic solv ents and adsorbents. Hansen. Gaylord Chemical Corp. The three dimensional solubility parameter — k ey to paint component af finities I. Paint Technol. and Eon. VI Federation d’Associations de Techniciens des Industries des Peintures.. Standen. Polym. D. 79–98. C.. Sci. Paint Technol.F . C. 57(7). 1962. and Somcynsky. Emaux et Encres d’Imprimerie de l’Europe Continentale. 1276–1292. 4(1). P aper presented at the F ourth Chemical Congress of America. Physica.L. The use of the solubility parameter concept in the United States.. J.. Burrell. 22. 1975. A. J. R.

Properties of Gases and Liquids. Flory. W. K. North-Holland.. inPaint Testing Manual. K oleske.. Appl. 47. and Smolders. Amsterdam. The determination of solubility parameters of solv ents and polymers by means of correlation with other physical quantities.. 113–120. under contract DTCG50-89-P-0333 for the U. Hansen. Oil Color Chem. Lecture for the Danish Society for PolymerTechnology. 1963. J. Pennsylvania State University. J. S. 31. Prog. C. Ed.M. Anonymous.Solubility Parameters — An Introduction 25 26... 1957. Elastomer solv ent interactions III — ef fects of methanol mixtures on fluorocarbon elastomers. Hoy. Danish Isotope Centre.. 59(3). Centre for Polymer Composites (Denmark). CRC Press. 1st ed.M. Hansen. Polym.. R. Ind. Mod. (Editor-in-Chief).L. J. Reliable latent heats of v aporization. interf acial tension and b utt joint strength . 55–60. 1970. 1988–1989. and Pierce. New York.W. NASA Publication SP-318.J. 46–51.P. and Mathot. pp. 55(5). T. 1975. 27–43.J.D. 383–404. 41. Fishtine.A. Org.. Environmental Capability of Liquids. Copenhagen. Department of Chemical Engineering. Beerbower. 42(541). CRC Handbook of Chemistry and Physics. 46. J.. chap.F. Coat. Hansen. Eng. 50.H. FL. San Francisco.H. 1974. 1976.. 45. 582–596. Weast. Project Supervisors.. 1993.. DIPPR Data Compilation Project. 42. heat of mixing. WV. A.. Hennissen. Assoc.). J. van Krevelen.. 1973. 1996. and Sherw ood. K. P A. 39. Amsterdam.R. surface tension. 32.N. ISBN 87-7756-286-0. Selection of Chemicals for Permeation Testing Based on Ne w Solubility P arameter Models for Challenge 5100 and Challenge 5200.. Freeman. Fedors.L... 30.. T.L. University Park. Chem. Colloid Interface Sci. Myers.M.C. 1995. D... A.. 1958 (Lydersen Method — see also Reference 31). Cornell University Press. 29. Tables of Solubility Parameters. 1989. C. Appl.. 65th ed. Eng. Ind. 38.M. 48. Solvent Resistance of Polymer Composites — Glass Fiber Reinforced Polyphenylene Sulfide. Enthalpy of v aporization of or ganic compounds. 72nd ed. . Ind. American Society for Testing and Materials.S.A. 43.. 32.. Ne w York. 16. 55(6). Boca Raton. 1977.. Tables of Experimental Dipole Moments. Prod. 1968. 36. Project 801. 1953. Research and De velopment Dept.V.. 1986. 76–118... R.F. in Interdisciplinary Approach to Liquid Lubricant Technology. McClellan. Properties of Polymers: Their Estimation and Correlation with Chemical Structure. Polym.. 20–28. A. Koenhen. New values of the solubility parameters from v apor pressure data. 3515–3539. 1163–1179. 1995. Union Carbide Corp. D. J. pp. A mathematical description of film drying by sol ent evaporation. 49.. D. Critical review of concepts common to cohesive energy density. McGra w-Hill.. Paint Coat. 1974. 218–225. A method for estimating both the solubility parameters and molar v olumes of liquids. Chem. 51(1). Majer. Danish Technological Institute. Principles of Polymer Chemistry. and Caillault. 26. P. American Institute of Chemical Engineers Design Institute for Ph ysical Property Research. 19. 28.E. M.C. Eng. Else vier. The Molecular Theory of Solutions.. 1989. Taastrup. P. C. V. CRC Press. Hoy. 147–154... 1963. Brochure: Co-Act — A Dynamic Program for Solv ent Selection. Tables of Physical and Thermodynamic Properties of Pure Compounds. 365–431. (with the collaboration of Bellemans. 17. I. 35. Eng. 1–62. R. 40. A. C.M. Prigogine. Reid. C-672–C-683. Ne w developments in corrosion and blister formation in coatings. FL. Paint Technol.. Chem. 27. 55(4). and Hoftyzer. Dev. (Editor -in-Chief). Data Compilation. 34. South Charleston.K. 472. R. Polym. June 1989. Chem. Philadelphia. Manual 17. C. pp. A. Coast Guard. 33. Surface effects in coatings processes.. P. L.. C. Dante.M. M. Program calculates solv ent properties and solubility parameters. Lide.. and Daubert. Res. Sci. J. Systematic Modification of Filler/Fiber Sur aces to Achieve Maximum Compatibility with Matrix Polymers. 44. Boca Raton. Ind.. Sci. Danner . in Handbook of Chemistry and Physics. Hansen. 1985..E. 2nd ed. 6-100–6-107. 13(4). Sci.. Hansen. Copenhagen. C... I. and Abu-Isa. Eng. Solubility parameters. Bittar . February 10. 47–56. 37. J. 79(9). Gardon.M. Hansen. 1991–1992. 1969. Exxon Chemical International Inc.J.. 14(2)..L.. tensile strength.

1988. C.26 Hansen Solubility Parameters: A User’s Handbook 51.. 301–308.C. Jr. 494–498. and Andersen. Assoc. The affinities of o ganic solvents in biological systems. April 1965. Ind. 52. Qual. The desirability function.. Harrington. . Control. B. J. Hansen. Ind.M. 49(6).. E.. 21(10). Am.H. Hyg.

in the “New Flory” theory leading to the coef ficient χ12” is thought to be lar It does appear . χ 12 Interaction Parameter. INTRODUCTION The Flory–Huggins “chi” parameter. There is agreement in man y cases. especially for essentially nonpolar systems. Neither the Flory nor the Prigogine approaches can lead to the type of predictions possible with the HSP approach. b ut χ12 is not. the HSP approach can be e xpected to be useful to predict the Flory interaction coef ficient. It w ould be desirable to relate the widely used HSP3–10 more directly to χ12.2 but now the χ12 parameter derived from the New Flory theory is being currently accepted for general use instead of the older χ.1. Therefore. χ12.25) in the HSP correlations and the 0. Correlations using Hansen solubility parameters (HSP) confir the usage of the term cohesive energy difference proposed in the Prigogine CST . that the constant “4” (or 0. The Hildebrand and HSP approaches inherently use the geometric mean. b ut full understanding of the interrelationship has not yet been achie ved. The lack of accounting for permanent dipole–permanent dipole and h ydrogen bonding (electron interchange) gely responsible for this. The many correlations and other predictions contained in this book w ould not be possible with these theories. The re verse is not possible as a single χ12 parameter cannot be used to divide the cohesion ener gy into contrib utions from dispersion (nonpolar) forces. This implies that the geometric mean is capable of describing not only dispersion interactions but also those due to permanent dipoles and h ydrogen bonding. however. and Hansen Solubility Parameters Charles M. χ. that is. They modify the specifi interactions described by the Prigogine δ and also the polar and h ydrogen bonding HSP ( δP and δH).— The Prigogine 2 Theory Corresponding States Theory. permanent 27 . Hansen ABSTRACT Patterson has sho wn that the χ12 interaction parameter can be estimated from the corresponding states theory (CST) of Prigogine. This could imply that the Prigogine ρ attempts to describe what the δD parameter describes. This w ould allo w estimates of χ12 for systems where the HSP are known. The Prigogine theory must be used with the geometric mean to estimate the inter action between different species. the nondirectional (nonpolar) atomic interactions. has been used for man y years in connection with polymer solution behavior. Equations for this purpose are presented and discussed based on comparisons of calculated and e xperimental values for fi e polymers.25 in the leading term of the Prigogine theory ha ve identical functions. as the y do not separate the polar and h ydrogen bonding effects independently.

Both positi ve and ne gative heats of mixing are allowed. an increase in temperature leads to poorer solv ency. and found.13 The w ork of Delmas. 18. (See also the discussion in Chapter 1. β has a generally accepted average value near 0. In principle. which is the basis of the HSP . Biros et al. When the solv ent has lo wer solubility parameters than the polymer . 22 as a correction to the Flory combinatorial entrop y.13 (see also Chapter 1. it has been sho wn recently that HSP and the Prigogine/P atterson CST are mutually confirming and g ve similar predictions. This temperature is called the lower critical solution temperature. where VM is the v olume of the mixture and the ϕs are volume fractions of the solvent and polymer. These studies also sho w that subsequent increases in temperature lead to improved solvency when a solvent has higher solubility parameters than the polymer. Equation 1. in the current conte xt.1) where V is the molar volume of the solvent. implies strictly nonpolar systems. respectively.28 Hansen Solubility Parameters: A User’s Handbook dipole–permanent dipole forces.12. in the same manner as solv ency or swelling data being used for a similar purpose. and co workers has sho wn that predictions with the nonpolar Hildebrand solubility parameter and the Prigogine CST are in e xcellent agreement with each other with re gard to heats of mixing in essentially nonpolar systems.3 units lar ger than that found when using the Hildebrand solubility parameter.3 and Equation 1.2. just as they are used with weight g ain or intrinsic viscosity . The Hildebrand parameters are applicable to re gular solutions.6 These parts come from the nonpolar/dispersion (atomic) forces.19 Also. as sho wn later. R is the g as constant. Hildebrand solubility parameters ha ve been shown to reflect th 12. The customary equation to calculate χ12 from the Hildebrand solubility parameters for a nonpolar solvent and a nonpolar polymer is: χ12 = [V( δ1 – δ2)2]/RT + β (2. The Hildebrand solubility parameter is the square root of the cohesi ve ener gy density (ced). predicted. and h ydrogen bonding (electron interchange). E D. β. Patterson15 and coworkers17 have shown how to predict the χ12 parameter using corresponding states theories (CST). This is not true. The argument that solubility parameters are inadequate. noncombinatorial free ener gy directly via the first term in Equation 2. which. They use the symbol ν2 instead of χ12 for the same quantity . Patterson.1. Reliable χ12 values for numerous solvents and the same polymer can be used to determine the HSP for the polymer. The heat of mixing must be found by dif ferentiating this relation as sho wn by Delmas and co workers. Precipitation can e ven occur with increasing temperature.34. as the y do not allow for negative heats of mixing. 17 state that this v alue of β presents difficulties as a explanation of an error in the Flory combinatorial entrop y approximation. although combinatorial in origin. It w as previously thought that the heat of mixing w as given by the Hildebrand theory as φ1φ2VM(δ1 – δ2)2. These authors state that β should be interpreted as aligning the χ12 values from the solubility parameters with those found from CST. is not v alid. and T is the absolute temperature. the weighting schemes (described in Chapter 5) to a verage the solv ent parameters for obtaining the polymer HSP can also be used with the χ12 parameter. The CST predict χ to be about 0.11–17 as well as using the Hildebrand solubility parameter in (strictly) nonpolar systems. β is not required for essentially nonpolar systems when HSP are used in a relation similar to Equation 2.20. the permanent dipole–permanent dipole .21 This is discussed in more detail later . w as attached to χ12 in order to preserv e the Flory form of the chemical potential e xpression. The empirical constant β has been discussed as being necessary for polymer systems.4). δ is the Hildebrand solubility parameter for the solvent (1) and polymer (2).) HANSEN SOLUBILITY PARAMETERS (HSP) It has been shown that the total energy of vaporization can be divided into at least three parts.

4) δD. in solv ent selection. δP . is a modified di ference between the HSP for a solv ent (1) and polymer (2). as the y are based on cohesion energy density. and generalized corresponding states correlations for hydrocarbons. and ha ve been described in greater detail elsewhere7. (E D). barrier properties of polymers.0455 times lar ger than in the often used (cal/cm 3)1/2 units.2 to Equation 2. (E/V) 1/2. This experimental procedure is still thought to be the best way to arri ve at the HSP for polymers. V. It might be noted that the v alue of a solubility parameter in MP a1/2 is 2. surf aces. As described in Chapter 1. Ra must not e xceed Ro. in this equation.4. As described in Chapter 3. thus giving support to the approach used in Equation 2. polar . and h ydrogen bonding (molecular) forces.20.3) Ra. Subtracting this nonpolar contrib ution from the total cohesion ener gy then gi ves the sum of the permanent dipole–permanent dipole and h ydrogen bonding (electron interchange) contrib utions to the total cohesion ener gy. HSP also include v olume ef fects.4. . The latter is more generally called the electron exchange energy. HSPs have been applied to the study of polymer solubility and swelling. The quantities required are E. where all the good solv ents are within the sphere and the bad ones are outside. E = ED + E P + E H E/V = E D/V + E P/V + E H/V δ2 = δD2 + δP2 + δH2 (2.5) (2. E P. respectively.3 They only need to be located on opposite sides of the sphere of solubility for the gi ven polymer. a corresponding states calculation using h ydrocarbons as reference is used to find the part of the cohes ve energy of a liquid that is attributable to dispersion (nonpolar) forces. The constant “4” has been found empirically useful to convert spheroidal plots of solubility into spherical ones using δD and either of the other parameters (see Chapter 5). The three parameters described in Equation 2. These can then be separated by calculation and/or e xperiment into the polar and h ydrogen bonding parameters. the dipole moment (and perhaps the refractive index and the dielectric constant). and hydrogen bonding interactions.23–26 etc.4 are fundamental energy parameters that can be calculated from the mutual interactions of identical molecules in a pure liquid. the polymer HSPs being gi ven by the coordinates for the center of the sphere. The calculated values for a large number of the liquids ha ve been confirmed xperimentally by solubility tests.10 These procedures were described in Chapter 1. Group contrib ution methods and simpler calculational procedures ha ve also been established. Volume effects are also basic to the Prigogine CST. These correlations have been v ery convenient for practical use.Theory 29 (molecular) forces. the radius of interaction of a HSP solubility sphere. for e xample. It has been used with success in well o ver 1000 HSP correlations with a computer program that optimizes a solubility sphere according to Equation 2. E H. δ is the Hildebrand solubility parameter . The reliability of the spherical characterizations and the need to di vide the total cohesion ener gy (E) into at least three parts has been confirmed by systematically locatin nondissolving solvents that are syner gistic and dissolv e a given polymer when mix ed.8. and δH are the HSP for the dispersion.10 (see also the follo wing chapters). biological materials.5. This program was described in Chapter 1. if solubility is to be maintained. P anayiotou and co workers ha ve used a statistical thermodynamics approach to calculate all three parameters.2) (2. Both Ra and Ro ha ve the same units as solubility parameters. The usual equation used in HSP correlations is: (Ra)2 = 4( δD2 – δD1)2 + ( δP2 – δP1)2 + ( δH2 – δH1)2 (2.

RA. just as it w as used to arrive at the equations ha ving differences in solubility parameters.7.0 on the boundary surf ace of a sphere describing polymer solubility .6. The Prigogine ρ accounts for differences in the size of the solv ent and polymer segments. whereas H is a ratio of cohesion ener gy densities. it is adv antageous to define the Prigogine δ using cohesive energy densities as follows: δ = [(ced 2)1/2 – (ced 1)1/2]2/(ced1) = ( δ1 – δ2)2/δ12 (2. H is zero when the solubility parameters for the solv ent and polymer match and subsequently increases to lar ger values as the differences between solvent and polymer increase.30 Hansen Solubility Parameters: A User’s Handbook For present purposes of comparison. ρ is defined a ρ = ( σ2 – σ1)/σ1 2 (2. R M is the radius of a HSP sphere (spheroid) based on Equation 2.8) where ε is the cohesi ve energy for a polymer se gment (2) or for a solv ent (1). must be less than R M for solution to occur .17 have explained the Prigogine theory in concise form and simplifie some of the most important aspects. The geometric mean rule [ ε12 = ( ε1ε2)1/2] was used to arri ve at this result. P atterson states that the . The HSP difference between solv ent and polymer . cohesion ener gies (HSP) for solvents can be calculated whereas those for polymers currently require e xperimental data on solubility or other rele vant testing procedures. It can be seen that the quantities RA/R M and H = (RA/R M)2 will be 1. For the present discussion. which is in f act equal to χ12. A key parameter is the Prigogine δ.11) “ν2” includes effects from differences in segmental energy (the δ effect) and in segmental size (the ρ effect).5 must be normalized by 4R M2 to mak e its predictions consistent with the quantities commonly used in the literature. RESEMBLANCE BETWEEN PREDICTIONS OF HANSEN SOLUBILITY PARAMETERS AND CORRESPONDING STATES THEORIES Patterson and co workers 15. This describes normalized cohesive energy differences between polymer se gments and solv ents. R M = Ro/2 (2. As indicated abo ve. The Prigogine δ parameter is defined a δ = ( ε2 – ε1)/ε1 2.15 “ν ” is approximated by ν2 app = ( δ2/4 + 9 ρ2) app = ( δ/2 – 3 ρ/2)2 (2.6) (2.7) RM is the maximum solubility parameter dif ference that still allo ws the polymer to dissolv e based on Equation 2. The segmental distance parameter is s.10) Another k ey parameter in the Prigogine/P atterson CST is ν2. Equation 2. RA/R M is a ratio of solubility parameters. in connection with the CSTs and χ12: Η = (RA) 2/RM2 = [( δD2 – δD1)2 + ( δP2 – δP1)2/4 + ( δH2 – δH1)2/4]/RM2 RA = Ra/2.9) The numerator is the difference in cohesion energy densities between solvent and polymer. and this is normalized by the ced of the solv ent.

.6 relates directly to the second term (the ρ ef fect) in Equation 2.. i.Theory 31 coefficient in front of ρ is uncertain15 and. and the interaction is described by the dif ference in δD. P articularly.These terms describe normalized dif ferences in the respective types of cohesive energy in corresponding states terminology. I think. which must inevitably exist between any polymer and an y solvent and which gi ve a contrib ution to the chi parameter …. using different symbols. the Flory theory w as a v ery usable and particularly successful case of the Prigogine theory (which w as in f act difficult to use). In the future.” Patterson27 has recently helped the author to clarify some points about the relations among the theories. is that surf ace/volume fraction of the polymers are v ery different from those of solvents. i.e. These are interesting differences between the Flory theory and that of Prigogine. Flory always presented his theory as something absolutely different from that of Prigogine’s. I have used it with respect to mixtures of small molecules and not polymer solutions. those with polar and ydrogen bonding. Differences in ced arising from these sources in such systems are modified by a actor of 0.25 according to Equation 2. In reaction against this I ha ve al ways lik ed to call the whole thing the Prigogine–Flory theory. I think that the w ork by a lot of people has established the utility of the Prigogine–Flory theory . it can be seen that each of the three terms in this equation is in the form of a Prigogineδ as given by Equation 2. The Prigogine δ parameter becomes important in systems with specific interactions. and dif ferences in size of solvent and polymer se gments for breakage of solv ent–solvent (1–1) bonds and polymer–polymer (2–2) bonds to allo w formation of solv ent–polymer (1–2) bonds. different names for concepts. be ascribed to free-volume differences. a solv ent located on the boundary of a Hansen solubility sphere rather than the ced of a gi ven solvent under consideration. I feel a fe w quotes from these communications. ν2 accounts for se gmental ener gy dif ferences. Also: Very specificall . i.11. hence. One is led to the conclusion that the firs term in Equation 2. An additional point that Flory made much of b ut was only touched by Prigogine. . b ut the implications appear v ery clear to the author . If we no w consider Equation 2. the Prigogine δ is small (perhaps zero in this conte xt). and the y have been lumped together in the χ12 parameter. are in order at this point: In my opinion. More specificall . since about 1970. this relationship could be explored in more detail with the hope of experimental verificatio of the coef ficient in front of ρ.6. and Flory completely discarded it. the ced of the worst possible good solvent.9. too.6 are zero.. the rho parameter does not ha ve nearly as much importance as Prigogine thought. for small-molecule mixtures. In strictly nonpolar systems. and the quantity ν2 depends on se gmental-size differences only. But. furthermore.27 that “ρ was a (misguided) attempt to take into account se gment-size differences. etc.6 are normalized by R M2. again in my opinion. in addition to the one just cited. when using the term Prigogine–Flory theory. in the Prigogine–Flory approach. the Prigogine parameters delta and rho are no w out of f ashion. The discussion concentrates on the ν2 parameter.11. The cohesive energy differences in Equation 2. I ha ve done v ery little with polymer solutions and. the origin of the parameter beta in the solubility parameter approach w ould. the polymer is very bulky. In nonpolar systems. Ho wever. at least. The ideas have not been fully tested as of yet.e. the Flory theory . This demonstrates that there is not complete agreement among those who ha ve concerned themselves with these theories. The following is an attempt to unify all of these thoughts. the polar and h ydrogen bonding terms in Equation 2. being lo yal to the P atterson article 15 (sometimes referred to as the Prigogine–Patterson theory) where this part of the book got its start.e. or if you like.

11 (CST) where the geometric mean w as used. This same assumption w as made by Zellers and coworkers in their approach to correlate the swelling and permeation of elastomers used in chemical protective clothing. In HSP terminology.1 .An explanation for this is found by comparing Equation 2. in calculating χ12. which are not good solvents in spite of H being less than 1.34) in Equation 2. The geometric mean appears to be applicable to all types of ener gies discussed here. i..13 with Equation 2. using ν2 as a symbol for the same quantity. even when H is greater than 1. is to replace the Hildebrand solubility parameter dif ference of Equation 2. This is an a verage correction to these calculations because of the ne glect of some relati vely small b ut significant alues of ( δP1 – δP2) and/or (δH1 – δH2).15 can be deri ved from this comparison.0 cc/mol). The dependency of χc on polymer molecular size is also included in Equation 2.17 Therefore.2 = [( δD2 – δD1)2 + 0. it is e xpected that Equation 2.1 w as found from studies on almost nonpolar systems using the Hildebrand solubility parameters.34 for β. 1 This boundary v alue is called the critical chi parameter. β is not required in Equation 2. the polymer–solv ent interaction will be either polar or hydrogen bonding (or both) in nature. there will only be specific interactions Such differences in δP and δH will be modified by 0.14) This last equation assumes an a verage V. It is not worthy that the same f actor. THE χ12 PARAMETER AND HANSEN SOLUBILITY PARAMETERS Patterson and coworkers have shown that χ12 can be calculated. 28–31 An estimate for χ12 can also be found by noting that the total χ12 parameter in common solutions of polymers ha ving high molecular weight is required to be close to 0.7 cc/mol) and acetone (V = 74.e. modifies the Prigogine δ term in Equation 2.6 can be used in a similar w ay to predict χ12. when there are specifi interactions between solv ent and polymer. The same 0. This allows a simple estimate for χ12 for higher-molecular-weight polymers by the relation: χ12 = χc(RED)2 = Η/2 (2. this is a boundary solv ent with a placement directly on the sphere of solubility .e.25( δH2 – δH1)2] and χ12 is estimated from: χ12 = VA1. as this is a partial explanation for some of the results discussed later .25 in Equation (2.25.25( δP2 – δP1)2 + 0.6 being equal to 1.5 at the point of mar ginal solution or precipitation. just lik e the HSP correlations ha ve done up to the point.. and the quality is indicated by H in Equation 2.6 (HSP) and in Equation 2.. There remains a general belief that χ12 can be calculated by Equation 2.1 by a corresponding HSP term.15. This is seen with liquids lik e methanol (V = 40.6. It has been noted man y times that liquids with lo wer V are often better solvents than liquids having essentially identical HSP b ut with lar ger V.0 and for liquids lik e the phthalate and other plasticizers (V > 150 cc/mol). which are sometimes good solv ents.14.e.32 Hansen Solubility Parameters: A User’s Handbook If we no w consider a system where δD1 is equal to δD2.11. The change required to progress from the nonpolar Hildebrand solubility parameter to include polar and h ydrogen bonding ef fects with the HSP . according to the pre vious discussion.2 A1.13) (2. i.0.25 is present for a similar purpose both in Equation 2. Equation 2.1 using Hildebrand solubility parameters and a value of 0. 0. i. A1. χc.12) The empirical f actor β (0.

this added term leads to a correction of 1. when one of the molecules is a high polymer .5 MPa1/2 (Ra greater than 7.7.2).0 MPa1/2). where r 1 is approximately equal to r 2 and both are approximately equal to 1.14 has been retained for present purposes of comparison. alues based on Equation 2.15 indicates. r1 and r2 are the number of statistical segments in the molecules in question. For a solvent with V = 100 cc/mol.0 5. is placed adjacent to that of the solv ent molar volume to allow an easy mental multiplication by V/100 if desired. The correction term is e xpressed more generally by the relation in Equation 2. Some improvement in the estimates of χ12 using Equation 2. Values for R M greater than 5 MP a1/2 are common.0 2.0 . assuming the size of the solv ent molecule as being close to that of the polymer segment size. estimated from Equation 2. the correction amounts to 0. Equation 2. It is usually assumed that V for the average solvent is near 100 cc/mol. ho wever.5(1 + 1/r 1/2)RT/V} (2.15) “r” is the ratio of the polymer size to that of the solv ent and is usually considered as the approximate degree of polymerization.0 3.14 is possible by including V in a correction term.15 for polymers of molecular Table 2. Table A. This immediately points to potential problems in directly calculating χ12 from HSP data when R M is significantly la ger than about 3.03 for a polymer molecular weight of 100.000. A study of the kind discussed in the follo wing for smaller molecules w ould seem appropriate to help clarify some of the questions raised. the correction amounts to a f actor of about 2.1 gives the expected RM and Ro v weight high enough to ne glect the effect of r.1 Expected Solubility Parameter Differences for Marginal Solubility as a Function of the Molecular Volume of the Solvent V (cc/mol) 50 100 200 RM (MPa1/2) 5. the correction approaches 0.16) Here. The column for χ12. This is not generally the case as man y values of R M have been reported that are much higher than this 10 (see also Appendix. meaning the y are more easily dissolv ed than Equation 2.14 in Table 2.1 indicates that all polymers of reasonably high molecular weight will be insoluble in solv ents with V greater than about 100 cc/mol for a corresponding RA that is greater than an R M of 3.5 (as discussed earlier). TABLE 2.5 7.000 and to a correction of 1. Table 2.3 to Table 2. For mixtures of low molecular weight. when both the molecules ha ve v ery high molecular weights.5 Ro (MPa1/2) 10. although these are generally for lo wer-molecular-weight materials.16. with some polymer R M values being considerably larger.5 MP a1/2.Theory 33 RM2 = (Ro/2) 2 = {0.1 for a polymer molecular weight of 10. meaning compatible mixtures of polymers are v ery difficult to achi ve.5(r1–1/2 + r2–1/2)2} (2. because of its simplicity . Correction = {0.

62 0.34 Hansen Solubility Parameters: A User’s Handbook TABLE 2.7 δP 2.5 0.0 4.2 to Table 2.6 0.2 12. Chemische Werke Hüls Solvent Benzene Toluene Xylene Pentane n-Hexane n-Heptane n-Octane Chloroform Carbon tetrachloride Methanol Water V 89.14 ha ve been compared with χ12 parameter data in standard references.4 106.1 χlit 0.6 Polymers Polybutadiene Polyisobutylene Polystyrene Polyvinylacetate Polyacrylonitrile δD 17.2 Hansen Solubility Parameter Data for Polymers Selected for the Comparisons Given in Table 2. Paint Technol.8 11. with Calculations Based on HSP for Polybutadiene.5 16. Indicative Chi Parameter Data. χlit.9 21.7.3 14.. 39(505).7 18.2 with their HSP data.54 0.6 0.97 3. The polymers were chosen because . J.7 97. C. Their Importance in Surface Coating Formulation.13 1.58 0.1 δH 3.9 21.15 0.0 χ12 (Equation 2..13 and Equation 2.3 3.1 Ro 6.7 0.05 0. Hansen.1 40.04 0.8 123. The Three Dimensional Solubility P arameter and Solv ent Diffusion Coefficient. Doctoral dissertation.07 0.M.5 5.2 131. COMPARISON OF CALCULATED AND EXPERIMENTAL χ12 PARAMETERS The predictions of χ12 using Equation 2.14) 0.39 0.4 4.6 147.3 to Table 2.9 Note: Units are MPa1/2.5 7.54 0. TABLE 2. C.33 The polymers used for the comparisons are listed inTable 2.3 3.7 10.3 116. Danish Technical Press.62 0.08 0.3 Comparison of Experimental.3 0.3 χ12 (Equation 2.4 0. 1967.68 20.4 163.4 0.3 2.5 80.0455 times lar ger than in (cal/cm 3)1/2.52 0.7 13.32.02 0.04 0.5 Source: Solubility data from Hansen.43 0.M.13) 0. 104–117..12 0.10 0. Buna Hüls CB 10 cis-Polybutadiene Raw Elastomer. Values in these units are 2. 1967.3 9. The calculated and indicati ve e xperimental values for χ12 for the gi ven solvent–polymer systems are reported in Table 2.3 20.16 5. Copenhagen.7 9.

9 χ12 (Equation 2. but the size ef fect reduces solv ent quality .17 0.1 63. The solubility parameter predictions indicate that these are exceptionally good solvents.4 0. Ed.Theory 35 TABLE 2. Agreement for the aliphatic solv ents is considered excellent. Koleske. especially confirmed by the χlit for chloroform.3 to bring an agreement.4 Comparison of Experimental.3.35 0.19 0. POLYBUTADIENE The calculated and experimental χ12 values for polybutadiene are given in Table 2.43 0.14.8 156.20 0. Manual 17.8 provides these solvents along with their HSP. in Paint Testing Manual. A casual inspection of the measured and calculated χ12 values in Table 2. This is not borne out by the χlit values for the aromatics.37 0.7 would gi ve the impression that there are significant discrepancies between these alues which require further e xplanation.0 0.5 0. χlit.21 0. Lutonal® I60.7 116.2 131.3+ 1.23 0. Before one adds on a constant v alue of about 0. Some possible reasons for this are discussed in the foll wing sections. it should be noted that the calculated and e xperimental χ12 values for the aliphatic solv ents are in good agreement. with Calculations Based on HSP for Polyisobutylene.. HSP considerations indicate that chloroform and the aromatic solvents are near neighbors with similar HSP and might ha ve similar qualities.20 0.5 179. of similarities/differences in the HSP data. 383–404. Table 2.5 0. BASF Solvent Benzene Toluene Decalin Cyclohexane Pentane n-Hexane n-Octane n-Nonane Chloroform Carbon tetrachloride Methylene dichloride V 89.9 108.25 χ12 (Equation 2. Chloroform and carbon tetrachloride are predicted by HSP to be very good solvents.44 0.5 0. Philadelphia. The calculated and literature v alues for methanol and w ater are dif ferent enough to w arrant a comment.5 0.3 through Table 2.4 106.14) 0. Solubility parameters.10 0.5 0.7 97. b ut it points out clearly that there are factors which are not completely understood.13) 0. J.53 0.17 χlit 0. C.1. A problem of some significance in a y . pp. It should be noted that Ro is v ery near the ideal v alue for such calculations according to Table 2. which the χlit v alues do not indicate. whereas the χ12 values indicate that they are moderately good. The calculated and literature v alues for χ12 agree in some cases and differ significantly in others...05 0. American Society for Testing and Materials.6 163.45 0.13 and Equation 2.M.29 0. The solubility parameters for the higher-molecular-weight homologs are closer to those of the polymer.51 0.7 80.50 0.27 0.06 0.V. 1995.5 0.11 0.49 0. The first thre entries are for aromatic solv ents.6 Source: Solubility parameter data from Hansen. Indicative Chi Parameter Data. as well as the a vailability of suf ficient data on the χ12 parameter. HSP considerations indicate that the dif ference in beha vior between these tw o liquids should be sizable. This is not a complete e valuation of Equation 2. which are suspected as being too high for presently unkno wn reasons.

An extension of this type of situation can be cited in the tendencies of w ater to associate with itself. These local re gions may not necessarily reflect the same a finities as the polymer as a whole.4 Source: Solubility data from Hansen.87 2. Hansen.22 1.5 74. C.25 0.9 0.40–0.4.61 0. as well as with local re gions within polymers.92 3.M.7 128.38 0.0 0. with Calculations Based on HSP for Polystyrene.45 0.5 133.4 0.8 123. C.44 2..08 1. HSP finds chloroform .3 131.. for e xample.14) 0. however.66 0.35 0.5 0.4 0.0 156.8 0.12 1.5 0.23 0.4 106.1 125.0 98. 39(505). 1967. Indicative Chi Parameter Data. These local association ef fects can influenc results on swelling studies in both good and bad solv ents.5 Comparison of Experimental.5 0. Danish Technical Press. POLYISOBUTYLENE The calculated and e xperimental χ12 values for polyisob utylene are gi ven in Table 2.51 0. This has made simple predictions of its behavior impossible.49 0. Copenhagen.67 0.40 0.44 0.1 163.15 0.40 0. Polystyrene LG.8 0. that the potential dif ferences observed here between HSP predictions and observ ed χlit may be derived from such phenomena.5 0.6 χ12 (Equation 2.21 0.73 0.40–0..49 0.72 2.35 χlit 0.5 132.83 0. A detailed discussion of this is be yond the scope of this chapter .45 0.6 136. Doctoral dissertation. Their Importance in Surface Coating Formulation. whereas the estimates for the aliphatic solvents are in e xcellent agreement with Equation 2. J.2 MPa1/2 being near the ideal for a polymer of v ery high molecular weight.13) 0.1 115.6 147.5 0.99 1. It is suggested.52 1. in particular .41 0.54 χ12 (Equation 2. 104–117.5 0.9 108.6 0.38 1.8 104.3 123. The cyclic and aromatic solv ents are again better as judged by HSP than the χlit values indicate.24 0. such as are reflected by the solubl or-not approach commonly used in HSP e valuations.16 0. Other types of studies carried out at lo w-solvent concentrations can also be influenced by these s gregation/association phenomena. Again.22 1.50–1. 1967. BASF Solvent Benzene Toluene Xylene Ethyl benzene Styrene Tetralin Decalin (cis) Cyclohexane Methyl cyclohexane n-Hexane n-Heptane n-Octane Acetone Methyl ethyl ketone Methyl isobutyl ketone Cyclohexanone Ethyl acetate n-Butyl acetate sec-Butyl acetate V 89.09 0.29 1. χlit.0 90. study of solvents at low concentrations in polymers is that the smaller amounts of solv ent relative to the polymer can lead to preferential association of the solv ent with those local re gions/segments/groups in the polymer which ha ve similar ener gies (HSP). There are some similarities with polybutadiene both chemically and in the Ro value of 7.36 Hansen Solubility Parameters: A User’s Handbook TABLE 2.44 0.M.03 2.61 0. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient.26 0. Paint Technol.05 0.13.58 0.

and Immer gut.88 0.5 3.4 0.2 χ12 (Equation 2. 3rd ed.4 106.. F .1 75.61 2.0 χ12 (Equation 2.33 0.30 1.Theory 37 TABLE 2. Wiley-Interscience.3–0.44 0.13 2.7 80.0 2.33 0.0 90.8 98.56 0.91 2.7 128.36 0. J. Polymer Handbook. 1989.83 1. Mowilith® 50.58 0. New York.4 0.0 85.1 125.3 3.5 163.5 18.7 1. 1967. Doctoral dissertation.3–0.13) 0.40 0.16 1. pp.39 0.. Their Importance in Surf ace Coating F ormulation.80 1.5 1. Solvents and non-solvents for polymers.3–0.44 0..A.11 1.27 1.0 76.61 0.3 2.47 χ12 (Equation 2. C. 39(505).5 0.4 0.H.5 0. The Three Dimensional Solubility P arameter and Solv ent Dif fusion Coefficient.7 102.M. Indicative Chi Parameter Data.14) 0. 1967.2–0.N-Dimethyl acetamide Dimethyl sulfoxide Tetramethylene sulfoxide V 66. (a) Gundert..0 77.33 0.4 0. Copenhagen.7 Comparison of Experimental.3 Source: Solubility data from Brandrup.4 0.0 0.12 0.4 0.40 4. Danish Technical Press.. Farbwerke Hoechst Solvent Benzene Toluene Decalin (cis) Tetralin Cyclohexane Methyl cyclohexane n-Nonane n-Decane Acetone Methyl ethyl k etone Methyl isobutyl ketone Ethyl acetate Dimethyl phthalate Dioxane Chloroform Chlorobenzene n-Propanol V 89.99 1. .32 0.14) 0.0 92.33 0. J.88 0..46 0.51 0.7 195.5 2.3 0.64 0.5 71.25 χ12 (Equation 2.9 136.4–0. VII/379–407.54 0.05 3.95 0. Polymer solvent interaction parameters.00 6. and Wolf. Eds.8 58.9 74.2–1. χlit.92 3.53 χlit 0.4 0..21 0.92 1.4 0. Indicative Chi Parameter Data. E.97 0. VII/173–182. Paint Technol.3 179.49 0. B.29 0.3 90.M.8 156. (b) Fuchs. χlit. 104–117. Hansen.15 5.36–0. with Calculations Based on HSP for Polyvinylacetate.4 0.0 108.6 Source: Solubility data from Hansen..3 0.13) 1.46 0. with Calculations Based on HSP for Polyacrylonitrile Solvent Ethylene carbonate γ-Butyrolactone Ethanol Water N.03 6.45 0. O. C.6 Comparison of Experimental. TABLE 2.34 χlit 0.N-Dimethyl formamide N.76 0. pp.63 0.3 0.37 0.

using Equation 2.4 18.3 15.8 1.1 7.6 1. b ut are qualitati vely in agreement.5. HSP predicts that the aromatic and cyclic solvents are somewhat better than that e xpected from χlit.5 15.0 0.8 18.0 0.9 15. b ut is unkno wn.0 0. so the results are more in agreement with χlit .0 0. in agreement with solubilityor-not experiments.1 15.0 0.0 δH 6. An exception of some note is the well-kno wn good solv ent cyclohexanone that is predicted as a much better solvent by HSP than the χlit value would indicate. whereas the χlit values indicate these are not good or at best.0 16.7 MPa1/2 is now much higher than the ideal v alue indicated in Table 2. The agreement would be better if the χ12 values obtained from Equation 2.6 18.1 6.3 The value 0.3 7.0 0.8 18.8 15.14 for the aliphatic h ydrocarbons are also lo wer than χlit. The results of Equation 2.7 16.3 to Table 2.0 0.8 3.14.N-dimethyl acetamide Dimethyl sulfoxide Tetramethylene sulfoxide Water δD 15.4 5.1 1.7 15.7 Solvent Benzene Toluene Xylene Ethyl benzene Styrene Decalin (cis) Tetralin Cyclohexane Methyl cyclohexane n-Pentane n-Hexane n-Heptane n-Octane n-Nonane Acetone Methyl ethyl k etone Methyl isobutyl ketone Cyclohexanone Ethyl acetate Note: Units are MP a1/2.0 17.2 15.6 6.3 19.4 5.0 10.0 15.0 19.8 19.6 16.2 10.9 7.14 includes the boundary v alue of χc equal to 0.5.6 19.6 14.8 6. higher value in increasingly polar media is required because of induced dipoles methylene dichloride.0 15. b ut there is no consistent pattern.13 for the aliphatic h ydrocarbons are particularly in good agreement with χlit.0 3.5 δP 3.1 5.3 0.0 is v alid in nonpolar media and deri ves from a zero dipole moment.5 16.2 1. practically all the χ12 values calculated by Equation 2. Equation 2. a progressi vely 10 (see also Chapter 1).6 13.4 17.0 0. POLYSTYRENE The calculated and e xperimental χ12 values for polystyrene are gi ven in Table 2.9 0. are in generally good agreement with the literature v alues.8 16.1 4.3 12.1 0.0 17.14 than values indicated by χlit. As a consequence of the Ro v alue.7 11.8 15.7 2. and carbon tetrachloride as being v ery good.1 4.2 15.0 2.1 7.0 0.0 0.8 δP 0.0 7.4 19.0 5.6 4.4 1.3 8.4 4.3 10. There is considerably more differentiation in predictions of solvent quality found by Equation 2.1 42.0 17. The values found by Equation 2. The polymer molecular weight is thought to be reasonably high. One is tempted to di vide by a factor of 2 or 3.N-dimethyl formamide N.2 Solvent n-Butyl acetate sec-Butyl acetate Dimethyl phthalate 1.1.14 for these were increased by a factor 2.4 11.8 Hansen Solubility Parameters for the Solvent Included in Table 2.0 0.3 δH 2.7 3. The Ro v alue of 12.7 10. mar ginal in quality.0a 6.0 0.1 22.0 17.4-Dioxane Chloroform Chlorobenzene Carbon tetrachloride Methylene dichloride Methanol Ethanol n-Propanol Ethylene carbonate γ-Butyrolactone N.6 17.8 18.0 2.13 are too high.4 16.0 19.0 1.0 0.0 0.3 17.38 Hansen Solubility Parameters: A User’s Handbook TABLE 2.8 16.4 11. . The χ12 values for k etones and esters.5 16.0 2.8 21.0 6.3 5.0 18. a δD 18.2 9.4 18.0 0.4 9.

14 gi ves reasonably good approximations to χlit as long as the solv ents are good enough to dissolv e the polymer. A point of some concern is that ne gative values for χ12 are found in the literature. The correlations given in Table 2. Therefore. while considering a reasonably high molecular weight. in general. whereas it is a nonsolv ent. However. Normal here also includes the specific interactions attributable to permanent dipole–permanent dipole and h ydrogen bonding interactions. as discussed earlier . A negative χ12 implies a solvent of a quality superior to anything that normal polymer–liquid interactions could pro vide.7. R M is e xpected to decrease only slightly for mar ginally higher polymer molecular weight. Tetralin dissolves the polymer.6. γ-butyrolactone and dimeth yl sulfoxide. The molecular weight for this polymer is reported as being 260. The agreement with Equation 2. Equation 2. and R M is e xpected to decrease some what for higher polymer concentration. with the exception of the data for polyacrylonitrile where no polymer concentration is indicated in the original solubility data. which ag ain means χ12 values found from Equation 2. and the azeotropic mixture of ethanol and water actually dissolv es it. the HSP predictions of χ12 are generally lower than the data found in the literature.13 is certainly not generally acceptable as an instrument to predict χlit. Equation 2.2 can also be expected to change for higher polymer concentration and molecular weight. 33 The HSP data for correlations of the type gi ven in Table 2. No systems with ne gative χ12 are included in Table 2.000.13 will be higher than those found in the literature. Once ag ain. The nonsolvents are not in good agreement. however.9 MP a1/2 which. The present evaluations are at the same polymer concentration unless otherwise noted.7 MPa1/2. There is no obvious general explanation for this situation. one cannot help but wonder why. there is considerable scatter in the χ12 parameter data from different sources. once more. corrections of this type are not responsible for the dif ferences in the calculated and observ ed χ12 parameters. GENERAL DISCUSSION It should be noted in general that χ12 can either increase or decrease with concentration of the polymer. 3. An interesting fact to keep in mind is that a polymer with molecular weights (in millions) will only swell in e ven the best solvents. This difference varies considerably. ho wever. although this can vary. A closer review of this situation is desired. POLYACRYLONITRILE The calculated and experimental χ12 values for polyacrylonitrile are given in Table 2.2 to Table 2. but a factor of 2 to 4 is generally required to give reasonable agreement. Alcohols of higher and lo wer molecular weight do have a significant e fect on this polymer.7. No significant corrections of the type included in Equation 2. n-Propanol is an error from the HSP prediction of it being a mar ginal solv ent.16 are required. but these are not allowed in either the CST or HSP approaches. there are some major disagreements.3. but has χlit equal to 1. a f actor of 3 to 4 is required to bring the values into agreement. It can initially be noted that Ro is 13. especially for lower-molecular-weight species. as the polyme molecular weight is v ery high in all cases. b ut the agreement is not uniform when all the solv ents are considered.6 When dealing with nonsolv ents. An additional problem of some concern is that. This polymer has high polar and h ydrogen-bonding parameters and Ro equal to 10. is somewhat above the ideal range. Barton 32 presents data to e xamine the potential magnitude of this ef fect. and the seeming discrepancies do not contrib ute . Clearing up this situation is f ar beyond the scope of this book.14 is reasonably good for the good solvents.2 are based on whether or not the polymer dissolv es at a concentration of 10%.Theory 39 POLYVINYLACETATE The calculated and experimental χ12 values for polyvinylacetate are given in Table 2.13 agrees surprisingly well with the best solv ents. Equation 2.

and other size effects is accomplished. surface wetting. Barton cites a case where a coefficient of 0. Zellers and co workers used this coefficien as an adjustable parameter for individual solvents in their studies. but they certainly represent a good and satisf actory means for practical applications. g ves 0.5.25) in Equation 2. The “nearest neighbors” to a good solv ent are clearly e xpected to be of nearly comparable quality unless they are in a boundary re gion of the HSP solubility sphere. There are also structural parameters in the New Flory theory. of course. and there should al ways be a ree valuation of experimental data based on an initial correlation. if not all.12) should be a di ferent number. The experimentally-determined radius for the solubility spheres automatically tak es these f actors into account. The HSP for the solv ents relati ve to each other are correct for the majority of the common solvents.14.) The solvent δD.25 in Equation 2. of course.5. δP.2 in Equation 2. The solv ent parameters ha ve been used with success for many years in industrial practice to predict solv ent quality using computer techniques by most.12.32 so there are also problems in determining which v alues are best in this approach. b ut reliable calculation of the radius of interaction has not been possible as yet. .2 (rather than 0. The solvent quality indicated by the ratio Ra/Ro (the RED number) has been particularly satisfactory. This ratio was defined years ago as a ratio of solubility parameters.5 (correspondin to a coefficient of 0. One can also find ariations in HSP values for polymers from different sources.40 Hansen Solubility Parameters: A User’s Handbook to blind confidence in a y of the reported χ12 values. 35 The author has also e xplored situations in volving water. Care is required to generate the necessary data. as H is really nothing other than (RED) 2. 34 Skaarup has mentioned a case of 5 (rather than 4) as a v alue for the coef ficient in Equation 2. as this is indeed closer to an energy difference number and would agree more with the Flory approach as seen in Equation 2. There has been some discussion as to whether the coef ficient 4 in Equation 2. etc. POSTSCRIPT The author has al ways e xperienced consistenc y in the quality of the predictions using the HSP .5. enhance the possibility of impro ved calculations. These are usually based on summing up simple volume fraction times the solubility parameter values. Use of the ratio of cohesion ener gy densities is also possible. and reduce the need for e xperimental studies. The use of the coef ficient 4 is also confirmed in Chapter The Prigogine theory contains structural parameters that have not been explored in this context.6 and Equation 2.12 as determined. (An e valuation of the quality of the χ12 values in the literature could be made with precipitation e xperiments for mixtures to see whether a mixing rule gi ves consistent results for these as well. Mixing rules have been established for even complicated solvent blends. where the D ATA FIT was equal for either a constant 4 or 5 in Equation 2. Indicative χ12 values are used here. This would lead to a number that is slightly higher than the 4 in Equation 2. permeation. It is possible that the use of structural parameters will allo w better understanding. and it is the author’ s feeling that this could be the case whenever a complete understanding of the ef fect of solvent molecular volume. 28–31 One significant actor in this discussion is that solv ents with higher solubility parameters generally ha ve lower molecular v olumes. The χ12 parameter does not specificall account for permanent dipole–permanent dipole or h ydrogen bonding interactions. 10 It has become quite clear that the HSP for the solv ents are not precise enough for sophisticated calculations. as plotting and interpretation of data use solubility parameters. which. This means they will be better than e xpected by average comparisons of behavior. major solv ent suppliers. This fact tends to lead one to stretch the spheres a little more in the δP and δH directions to encompass these good solvents that would otherwise lie outside of the spheroids. and δH values that were established with e xtensive calculations ha ve been supported by tens of thousands of experimental data points based on solubility. which must be considered a major source of potential dif ferences.

it has not been its purpose to recite the de velopments of polymer solution thermodynamics in a historical manner with full e xplanations of each theory or modifications thereof. and the solv ent may associate with given segments or groups in the polymer and not reflect the beh vior of the polymer as a whole (see also the discussion in Chapter 5). An empirical f actor.Theory 41 As a result of ha ving reviewed all of this here. The Prigogine treatment is acknowledged as difficult to use in practice. This is not true of the HSP approach. The comparisons presented previously confirm some relation. CONCLUSION The Prigogine/Patterson CST and the HSP approach (which also in volves a corresponding states calculation) are shown to have very close resemblance. Reasonably good predictions are possible under f avorable circumstances. It is suggested that the structural f actors discussed by Prigogine be tried in this respect. Both can be used to estimate the Flory χ12. Two equations involving HSP are given for this purpose. 11 Use of the geometric mean in conjunction with the Prigogine theory brings the HSP and Prigogine approaches into agreement. b ut the resulting equation has not been studied enough to allo w general use of HSP to calculate χ12 parameters. and some problems still e xist and need to be solv ed before these calculations can be used with confidenc to estimate χ12 values for any solvent–polymer system. ut the single χ12 parameter may ha ve been o versimplified. Chapter 3 and Chapter 4 have been added to this edition of this handbook to gi ve broader co verage in this respect. β disappears when HSP are used for this purpose. This chapter has attempted to sho w relations between the classical theories of polymer solution ther modynamics and the HSP approach. χ12values for the better solvents are calculated by HSP at lo wer le vels than those found in the literature. at least as f ar as the major f actors discussed earlier are concerned. ” As stated in the Pref ace to this book. se gmental size.1 and Figure 5. The relation between the Prigogine–P atterson theory and HSP w as the most ob vious. The massive amount of experimental data presented in this book strongly supports the use of the geometric mean. Figure 5. as less-than-dissolving amounts of solv ent may be used. As a curiosity. β. The most f avorable circumstances are. These may not completely correspond to the type of e valuation often used to find χ12. The references cited in the Pref ace do this already . . χ12 values for nonsolvents are also generally calculated by HSP as being significantly l wer than the reported literature values. such that the more complete HSP approach cannot be immediately recognized. of course. equal to about 0. it might be noted that the use of the geometric mean (Lorenz–Berthelot mixtures) generated an ellipsoidal miscibility plot essentially identical to those given in Chapter 5. and polymer size are still required. The HSP values for the polymers used for the present comparisons are based on solubility-or-not type experiments which reflect the propertie of the polymer as a whole. the author senses that the HSP approach is a practical e xtension in complete agreement with the Prigogine–Flory theory when the geometric mean is used. F avorable circumstances in volve a system with an essentially nonpolar polymer whose Ro v alue is not too dif ferent from 7. which includes a quantitati ve accounting of both permanent dipole–permanent dipole and h ydrogen bonding interactions as an inte gral part. not al ways present. The ability of HSP to describe molecular affinities among so ma y different materials listed in this book speaks for the general application of both the Prigogine and the HSP treatments.34 appears in man y sources in the literature in connection with calculation of χ12 using Hildebrand solubility parameters.0 MP a1/2. Studies on the ef fect of molecular size.2. 36 This approach w as not continued because it w as stated that “the boundary of this ellipse is of little practical importance as there are no known cases of immiscibility in mixtures kno wn to conform to the Lorenz–Berthelot equations.

Bhattacharyya. C. Patterson. Doctoral dissertation. C. and resins. D. Role of free v olume changes in polymer solution thermodynamics . C.M. The Solubility of Nonelectrolytes. emulsifiers. 3379–3389. 17.. 16. Faraday Soc. J. 21. 1974. Ne w York. C.L. Sci. 218–225.. Interscience.N. 1276–1292.J. Reinhold. Regular Solutions. 9.. Trans.42 Hansen Solubility Parameters: A User’s Handbook REFERENCES 1.V. B. Hildebrand. pp. in Danish). Macromolecules. 14. New York. A. 43(1). 2nd ed... Chem. Patterson. J.. Prigogine. 1962. 1967. and P atterson. Faraday Soc. J. Eng. Dev. and Skaarup. Reinhold. D. 1974. Ind.M... Dev. 2035–2052. Verlag Chemie. Res. Färg och Lack Scand. 1969.. Chem.. D.. 1962. 1969. in Danish). 1971. Faraday Soc.. G. 2053–2060.M. 2066–2072. Independent calculation of the parameter components J. and Flory . Paint Technol. Copenhagen. Surface dewetting and coatings performance.. 1964. The Molecular Theory of Solutions. and Scott. Ne w York.M.. The universality of the solubility parameter . Polym. Ind... J. 6. 3rd ed. Ne w York. Principles of Polymer Chemistry.M. 13. 19. Dig. 4(1). 44(570).M.. 18–22. 10. L. Res. A.. 1972. and Mathot. K. Paint Technol. 1950. and Beerbo wer. Hansen. 39(511). Prod.. D.. 2. 104–117. in Danish).J. chap. Hansen. 1953. 17. C. 1970. 25. C. 78(9). pp. and pigments.. 12. and Somcynsky.E. Danish Technical Press. The Solubility of Nonelectrolytes. Suppl. R.. Introduction to thermodynamics of polymer solubility . J. Biros. 79–98. Hildebrand. 1950. 30. The principle of corresponding states and the excess functions of n-alkane mixtures .. Patterson. Soc. C. I. The three dimensional solubility parameter — k ey to paint component af finities I Solvents.. Thermodynamics of polymer solutions. J. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient. Characterization of surfaces by spreading liquids. 3rd ed. XII Federation d’Associations de Techniciens des Industries des Peintures. Polymer solubility — prigogine theory and Hansen Solubility parameter theory mutually confirmed (Polymeropløselighed — Prigogine Teori og Hansen Opløselighedsparameterteori Gensidigt Bekræftet. Hansen. Pigmenter.M. 11. mutual solubility and compatibilit . Thermodynamics of polyisobutylene-n-alkane systems. American Society for Testing and Materials. Paint Technol. inPaint Testing Manual. Surface effects in coatings processes. J. Hildebrand.M. Physica.E.. C. 677–692. Hansen. Flory.M. P. plasticizers. Ed. Standen. Polym. Sci. 15. Eichinger. 1971. Hansen. Patterson. 1968. C. 8. P .M. Trans. 30–35. 16. K oleske. D. D. Hansen. Hansen. and Scott. 34(450). Prediction of the χ parameter by the solubility parameter and corresponding states theories. Philadelphia.. Off. 1967. Congress Book. 1967. Paint Technol.. 4–6.M. P. J.. 23. 5–10. (with the collaboration of Bellemans.L. 13(4). pigments. 25 Years with solubility parameters (25 År med Opløselighedsparametrene.. 16. chap. Dan.. 73(8). Fed. J. 57–60. C.. Emaux et Encres d’Imprimerie de l’Europe Continentale. J. 889–910. and Pierce. Kemi. 24. Trans.. 1968. Vol. and Somcynsky. North-Holland.. Prod. 4. og Fyldstoffer. Delmas. Vernis. 57. The three dimensional solubility parameter — key to paint component affinities III. 1957.. (2).. Hansen. Hansen. Weinheim/Bergstrasse. Solubility parameters. Zeman. 505–510. Their Importance in Surf ace Coating F ormulation. Ed. Patterson. 64(1). 1967. Eng. Kemi. Hansen.. (3). 511–514. 2061–2065.. Ne w aspects of polymer solution thermodynamics..). Prentice-Hall. The three dimensional solubility parameter — k ey to paint component af finities II Dyes. and Delmas. Paint Technol. 39(511). 20. 91–99. 22. 1995. Amsterdam.L. NJ. 383–404. S.M.. 20. C.. D. R.. 1997. C. 18. 41(536).D... Englewood Cliffs.. J. Trans. 8(1). T. Paint Technol. 1992. 3.. Paint Technol. C.M. Part C. Faraday Soc. 5. (4). in Kirk-Othmer Encyclopedia of Chemical Technology. R.. and Scott.. 1997. Hansen. polymers. 1962. . J. Cornell University Press. A. Dan. Hansen. A.. 2–11. and fillers ( ohæsionsparametre for Overflade . 489–493. Cohesion parameters for surf aces.. Manual 17. 660–664. 42(550). T. 7. Hansen.. Solubility parameters. 39(505)..

Anna. Sulewski. 2081–2096. 32. Critical analysis of the graphical determination of Hansen’s solubility parameters for lightly crosslink ed polymers.H. Appl. CRC Press.. D. Liquids and Liquid Mixtures. 62.. pp.. Appl.. D.. Rowlinson. 31. 1997.S. J. pri vate communication.. Sci. and steady-state perme ation rates of or ganic solvents in Viton® gloves. 36.. and Wei. 50. 1997.. Sci. 513–530. J.M.. Pure Appl. Three-dimensional solubility parameters and chemical protecti ve clothing permeation. 1959 chap. Modeling diffusion coefficients. Chem.. Handbook of Polymer-Liquid Interaction Parameters and Solubility Parameters.. B. Sci. VII/173–182.-Z. 531–540. Technol. Polym. and Wei. 15(3/4). I. Eds.. J. Anna. 30. Brandrup. (b) Fuchs. Appl. Zellers. Zellers. E. R. Polym. 62.T. 2069–2080. 1996. E. Sci. E. Solv ents and non-solv ents for polymers.. 1990. D.A. Barton. E. VII/379–407. F . and Zhang G. 35.H. 9. 1993.. London. E. J. Patterson. Three-dimensional solubility parameters and chemical protecti ve clothing permeation.. 905–912. Hansen. 3rd ed. J. Zellers. Barton.F.. 275–286. Skaarup.T . 1996. 1989. Sule wski. 29. 1993. Zellers.M. Applications of solubility parameters and other cohesion ener gy parameters. 28.T. pp.. Modeling the solubility of or ganic solvents in Viton® gloves.T. Adhes. and Wolf. C. 34.F. . and Immer gut. Wiley-Interscience. Boca Raton. (a) Gundert. R. X.Theory 43 26.H. Sci. 27.. On the use of cohesion parameters to characterize surf aces. personal communication. II. Butterw orths Scientific Publications. New York. 50. 1983. Polym. E. X.... J. A. Polymer -solvent interaction parameters. A. FL... 57(7). J.M. O. Polymer Handbook... K. breakthrough times.. Impro ved methods for the determination of Hansen’s solubility parameters and the estimation of solv ent uptake for lightly crosslinked polymers. Polym. Polym.. 1985. Appl. and Wallström. 33.


available information on these separate components is exploited for extracting information on the thermodynamic behavior of the fluids o ver an e xtended range of e xternal conditions. and high polymers.The central principle behind the use of δ is the historic alchemist maxim. supercritical fluids. chemical resistance. dispersion. K ey to this approach is the development of explicit expressions for the contribution of hydrogen bonding. On the other hand. and Dispersion Solubility Parameters Costas Panayiotou KEY WORDS Statistical thermodynamics. and the drug deli very processes. ABSTRACT The main objecti ve of this chapter is the presentation of an equation-of-state frame work for the calculation of the h ydrogen-bonding component of the solubility parameter as well as the other partial solubility parameters. cohesive energy density. 45 .Thermodynamic 3 Statistical Calculations of the Hydrogen Bonding. the proposed method is essentially a predicti ve method for the hydrogen bonding component of the solubility parameter. solvent screening. characterizing surf aces. predicting solubility and degree of rubber swelling. Information on v arious thermodynamic properties of fluids is used in order to estimate the three solubility parameter components. Dipolar. ey parameters for selecting solv ents in industry . such as the supercritical fluid. δ. Originally introduced by Hildebrand. the coating. Extensi ve tables with the k ey parameters are presented. and for numerous other applications. the use of solubility parameter. and dipolar interactions to the potential ener gy of the fluid. or the closely related cohesive density is not always successful and this lack of total success stimulates continuing research. The approach is applicable to ordinary solvents. INTRODUCTION The conceptual simplicity of the solubility parameter . Hansen solubility parameters. similia similibus solvuntur (“lik e dissolv es lik e”). makes it most attracti ve in industry and 1 it remains today one of the k in academia as well. There is also much interest in utilizing solubility parameters for rationally designing new processes. When information on h ydrogen bonding interaction is a vailable from other sources. A ne w statistical thermodynamic approach has been de veloped for the estimation of these partial components o ver a broad range of temperature and pressure. polymer compatibility .2–8 Of course. permeation rates.

we have already proposed a systematic approach for estimating the solubility parameter o ver an extended range of temperature and pressure. δb. and δhb are very likely to be miscible. and δhb. ho wever. could not separate the dispersion and the polar components of the solubility parameters.P. is further subdivided into an acidic component.16 which w as modified in order to e xplicitly account for dipole–dipole interactions and. be a good guide in the study of solubility . however. This w as done by adopting the more recent and more accurate NRHB (nonrandom hydrogen-bonding) equation-of-state frame work. T . it became clear that the h ydrogen bonding contribution could be calculated rather accurately from the h ydrogen bonding part of the potential energy E and the v olume V of the system as obtained. the partial solubility parameters were determined for an enormous number of substances and led to critical compilations a vailable in the open literature. The bulk of the developments in solubility parameter reside on the principle of “similarity matching” of properties.14 we have proposed a group contrib ution method for the estimation of the total solubility parameter of a large variety of substances. in order to account for the Lewis acid and Le wis-base character of the substance. respectively. as long as it enables definition of the de gree of likeness in the given system with sufficient precision. for the dispersion.46 Hansen Solubility Parameters: A User’s Handbook probably the oldest rule of solubility . δp. heavily based on a more recent publication 15 in which the pre vious approach11–12 w as e xtended in an ef fort to account for all three components of the solubility parameter. the polar . from the lattice-fluid hydrogen bonding (LFHB) equation-of-state model. This rule can. a geometric characteristic of the molecule.1) Following previous practice. Thus. THEORY THE EQUATION-OF-STATE FRAMEWORK Let us consider a system of N molecules of a fluid at temperature T. for e xample. which are assumed to be arranged on a quasi-lattice that has a coordination number z. Each molecule is assumed to be di vided in r segments of se gmental volumes v*. δp. that a more appropriate principle w ould be the “complementary matching” of properties. The total number ( Nr) of lattice sites is given by: Nr = rN + N0 (3. indeed. 8–10 Over the years. and to ha ve zq = zrs external contacts. and a basic component. As it is recognized. In a recent publication.T ensemble and in its maximum term approximation: Q ( N . The very same method could be used for the estimation of δd. δp. as sho wn later. δhb. This need for precision in the definition of lik eness led to the di vision of δ into its partial components or Hansen solubility parameters 5 δd. and of volume V. number of sites Nr . δa. 11–12 In this w ork. e xplicitly calculate the polar component. This chapter is. ho wever. liquids with similar δd. 13 The model. The proposition w as made to calculate δd from the solubility parameter of the corresponding homomorph hydrocarbon. external pressure P. Although this proposal could be v alid for some classes of fluids. Starting from the original definition of cohesi ve energy density and solubility parameter .16 one may write for the configurational partition function of the fluid in the N. 3–5 These compilations are most v aluable sources of information for the nature of the substances and their interactions with other substances. N0 and that denotes empty sites. s being its surf ace-tovolume ratio.15. it could not be generalized. and the h ydrogen bonding contributions. P ) = QR QNR Qhb = Ω R Ω NR Ω hb exp −E p − Ed − E hb − PV exp exp exp kT kT kT kT (3.2) . thus. 9 the h ydrogen bonding component. Knowledge of the h ydrogen bonding and the dispersion components of the solubility parameter could lead to an estimation of the polar component as well.

4) whereas the total number of intermolecular contacts in the system is gi ven by: zNq = zqN + zN0 (3. ΩR. namely: N r ! N rl N ⎛ N q ! ⎞ ΩR = ω ⎟ N0 ! N ! ⎜ ⎝ Nr ! ⎠ N z 2 (3.16 Here. respectively. ΩR.2 are the dispersion. The superscript 0 refers to the case of randomly distrib uted empty sites. Ep. The most classical is Guggenheim’ s quasi-chemical e xpression17: ⎡⎛ N 0 ⎞ ⎤ N ! N ! ⎢ ⎜ r 0 ⎟ !⎥ ⎢ ⎣⎝ 2 ⎠ ⎥ ⎦ = 2 ⎡⎛ N r 0 ⎞ ⎤ N rr ! N 00 ! ⎢⎜ ⎟ !⎥ ⎢ ⎣⎝ 2 ⎠ ⎥ ⎦ 0 rr 0 00 2 Ω NR (3. and Ehb in Equation 3. the final equations are simply reproduced.5) In Equation 3.Statistical Thermodynamic Calculations 47 Ed. and for the h ydrogen bonding. The relation is gi ven by: f0 = N 0 N r − rN = = 1− f Nr Nr (3.6) 16 For the second f actor. respectively.7) where N rr is the number of e xternal contacts between the se gments belonging to molecules. In this random case. The detailed rationale behind the form of the combinatorial term. and h ydrogen bonding components. and N r0 is the number of contacts between a molecular segment and an empty site. In the following. of the potential ener gy of the system. we ha ve: 0 N rr = 1 qN z zqN = qN θr 2 N 0 + qN 2 (3. polar .8) . we will need the site fractions f0 and f for the empty sites and the molecular segments. N 00 is the number of contacts between the empty sites. ω is a characteristic quantity for each fluid that tak es into account the flexibility and the symmetry of the molecule. ΩNR. Ωhb. can be found in the pre vious work.3. we may use v arious expressions available in the open literature.3) where l= z r −q − r −1 2 ( ) ( ) (3. and this quantity cancels out in all applications of interest here. its correction factors for nonrandom distribution of free volume.

9) and N r00 = zqN where θr = 1 − θ0 = and the reduced v olume is defined as: v= V rNvv ∗ 1 = = ∗ ρ V rNv ∗ (3.13) N r 0 = N r00 Γ r 0 The nonrandom f actors. Γ. numbers must satisfy the follo wing material balance equations: 16–18 2 N 00 + N 0 r = zN 0 2 N rr + N 0 r = zqN By combining these equations. ρ The corresponding number of interse gmental contacts (N ij) in the nonrandom case are gi ven by the following equations: 0 N rr = N rr Γ rr = 0 N 00 = N 00 Γ 00 z qN θr Γ rr 2 (3. in these equations are equal to unity in the random case.15) .48 Hansen Solubility Parameters: A User’s Handbook 0 N 00 = 1 N z N 0 z 0 = N 0 θ0 2 Nq 2 (3.14) (3. we obtain: θ0 Γ 00 + θr Γ r 0 = 1 θr Γ rr + θ0 Γ r 0 = 1 These two equations along with the quasi-chemical condition: 16–18 These (3.12) q r q r + v −1 (3.10) ˜ being the reduced density .11) N0 qN = zN 0 = zqN θ0 = zN 0 θr Nq Nq (3.

17) Combining Equation 3. the equation of state of the fluid is gi ven by: ⎡ ⎤ ⎛l ⎞ z ⎛ q ⎞ z P + T ⎢ ln 1 − ρ − ρ ⎜ − ν H ⎟ − ln ⎜ 1 − ρ + ρ⎟ + ln Γ 00 ⎥ = 0 r ⎠ 2 ⎝r ⎠ 2 ⎝ ⎣ ⎦ ( ) (3. The reduced density needed in Equations 3.16) form a system of three equations from which one may obtain the f actor Γ. is obtained from the equation of state (cf.16.23) .22 sho wn later).15 and Equation 3.21) With these definitions. with the physically meaningful solution: Γr0 = 2 ⎤ 1+ ⎡ ⎣1 − 4θ0 θr 1 − A ⎦ (3. T*. ε* = z ε/2.18) (3. one has: ⎡ ⎣ N !⎤ ⎦ 2 Ω hb = NH ! ⎡ N − NH ⎣ ( ) ⎛ ρ −S ⎞ exp H ⎟ 2 ⎜ R ⎠ ⎝ rN !⎤ ⎦ NH (3. Equation 3.22) and the chemical potential by: μ dp μ H μ = + RT RT RT (3. of the fluid by: ε∗ = RT ∗ = P ∗ v ∗ whereas the reduced temperature and pressure are defined as: T= T P .15. The intersegmental interaction energy.Statistical Thermodynamic Calculations 49 ⎛ 2 * ε* / z ⎞ 4 N rr N 00 Γ rr Γ 00 = = =A exp 2 ⎜ N r20 Γr ⎝ RT ⎟ ⎠ 0 (3.13 In the case of a fluid with d proton donors and a proton acceptors forming NH hydrogen bonds.16 is related to the scaling temper ature. in Equation 3. and scaling pressure.19) 2 ( ) 1 Most general e xpressions for Ωhb may be found in the original w ork. P= * * T P (3. P*.20) where SH is the entrop y change upon h ydrogen bond formation and the NH is given by: B+d +a− N νH = H = rN (B + d + a) 2r 2 − 4ad (3. one can obtain a quadratic equation for Γr0.

22–24 In an o verwhelming majority of cases. a drastically dif ferent approach w as adopted that preserv es the simplicity of the formalism of the pre vious paragraph. in the absence of dipolar interactions we should ha ve εdp = ε and εdp* = ε*.27 we count se gmental interactions. we may write for the non-h ydrogen-bonding part (dispersion and polar): − E dp = N rr ε dp = Γ rr qN θr ε∗ dp (3.. however. this procedure led to underestimations of δp that often f all in the range of one to tw o orders of magnitude belo w the e xpected value. 20–21 as well as by using the perturbation model of Nezbeda and Pavlicek.28. m. the function f might be approximated by: .. As. of the fluid is zero. In the above formalism. the contributions from dispersion and polar forces are lumped into one contribution.22 are coupled equations and must be solv ed simultaneously for the reduced density and the nonrandom f actors. This function should be zero in the absence of dipole–dipole interactions or when the dipole moment. Equation 3. We may then write quite generally: ∗⎡ 1 + f m. thus.26) Equation 3. T . it was assumed that only first neighbor segment–segment interaction contacts contrib ute to the potential ener gy ( E) of the system and. only the contrib ution from dispersion forces. The heat of v aporization is given by: ⎡⎛ ⎛ ⎞ ⎤ ⎞ q q HV = rN ε∗ ⎢⎜ Pv − θr Γ rr ⎟ − ⎜ Pv − θr Γ rr ⎟ ⎥ + E H ⎡ N ⎢ ⎣ H r r ⎠ liq ⎥ ⎝ ⎠ vap ⎝ ⎢ ⎣ ⎦ (3. as in NRHB 16.15. In the previous paragraph. in Equation 3.50 Hansen Solubility Parameters: A User’s Handbook where μ dp RT = ln ⎡ q P 1 z q ⎤ zq ⎡ − ρl + ln ρ − q ln ⎢1 − ρ + ρ⎥ + ln Γ rr ⎤ − +r ⎣ ⎦ ωr 2 r ⎦ 2 T T ⎣ (3. ⎤ ε∗ dp = ε ⎣ ⎦ ( ) (3. An attempt will be made in the ne xt section to separate them.27) Obviously.24) is the chemical potential for the dispersion and polar interactions.25) ( ) liq − NH ( ) vap ⎤ ⎥ ⎦ (3. and μH d a = r ν H − d ln − a ln RT d − νH a − νH is the chemical potential for the h ydrogen bonding interactions. r.19.28) The crux of the problem is the e xplicit form of the function f in Equation 3. ρ. THE CONTRIBUTION FROM DIPOLAR FORCES In an initial attempt.. the contrib ution of dipole–dipole interactions w as approximated by the multipolar u-expansion of Twu and Gubbins19 by keeping the leading term of the point dipole–point dipole interaction and the P ade approximations. and Equation 3. that is. Thus.

Ho wever. ⎝r⎠ ( ) b ( ) (3..33) On the basis of the abo ve equation. His additional observ ation that δp is in versely proportional to V1/2 could be reconciled with Equation 3. has observ ed that δp is directly proportional to m. summarizing years of e xperience. This approach with the constant c replaced by the e xpression: c = πs 4 (3. = ⎜ ⎟ ⎝ r ⎠ Γ rr θr ( ) 2 (3.27 and Equation 3.30 along with Equation 3. it turned out that the follo wing simpler form of Equation 3.29) Hansen5.27 and Equation 3.34) ... One additional reason for adopting Equation 3.29 could be set equal to 2. r. lead to a simultaneously satisf actory estimation of δp and a satisf actory description of the thermodynamic beha vior of the fluids.. = ⎜ ⎟ g m.32) led to better results and pro vided both a satisfactory description of the thermodynamic behavior of fluids over a broad range of e xternal conditions and a satisf actory estimation of δp and the other partial solubility parameters for the o verwhelming majority of fluids. T . r. ρ. which implies that b in Equation 3. T . ρ.30: ⎛m⎞ f = ⎜ ⎟ πs 2 ⎝r⎠ 2 (3.Statistical Thermodynamic Calculations 51 ⎛m⎞ f m.. T ...29 by writing: ⎛m⎞ c f m. the partial solubility parameters are gi ven by: Γ rr qN θr ε∗ V δd = (3. the final form of the potential ener gy that w as adopted is: 2 ⎡ ⎛m⎞ 2⎤ ⎢ − E = Γ rr qN θr ε 1 + π ⎜ ⎟ s ⎥ − N H E H ⎝r⎠ ⎢ ⎥ ⎣ ⎦ ∗ (3.28. ⎝r⎠ ⎢ ⎥ ⎣ ⎦ Thus.30) c being a constant. The second attempt for the estimation of δp w as made by using Equation 3. indeed. r.32 is that the only change that should be made in the formalism of the pre vious 2 ⎡ ⎤ ⎛m⎞ paragraph is to replace ε* by ε∗ ⎢1 + π ⎜ ⎟ s 2 ⎥ .. ρ..31) could.

the entrop y change.2b compare the e xperimental5 solubility parameters with those estimated/predicted by our approach for a number of common fluids.16 This is a rather crude approximation b ut it permits a more direct comparison of the strength of the v arious types of h ydrogen bonds through the mere comparison of the energy change.35) δ hb = (3. and volume change upon hydrogen bond formation are also needed.52 Hansen Solubility Parameters: A User’s Handbook δp = ⎡ ⎛ m ⎞2 ⎤ Γ rr qN θr ε∗ ⎢ π ⎜ ⎟ s 2 ⎥ ⎢ ⎝r⎠ ⎥ ⎣ ⎦ V − N H EH V (3. as for alkanols. EH. Equation 3. the scaling constants for high polymers were obtained by correlating the available extensive experimental pressure-volume-temperature (PVT) data 29 with the equation of state. Table 3. ho wever. V. The parameter EH is also reported in Table 3. it does not suf fice to reliably determine SH as well. This equation. As in NRHB. was set equal to zero for all fluids. the energy. the energy change. cannot be used in the case of carboxylic acids where the main mode of h ydrogen bonding is dimerization. 16 the geometric constant s of each fluid w as obtained through the widely used group contrib ution calculation scheme of UNIFAC. APPLICATIONS In this section.I. we will apply the model presented in the pre vious two sections in a multitude of cases. The case of carboxylic acids is treated separately. is given in Appendix 3. 16. 13. Essentially. the orthobaric densities.21. VH. w as adjusted through the e xperimental value5 of δhb. EH. as shown in Appendix 3. In addition.36) where. which is the only adjustable parameter for the description of h ydrogen bonding in our approach. which tak es explicitly into consideration its dependence on temperature.25–26 These constants are reported in Table 3. In the overwhelming majority of cases. As a first step. In a similar manner .2a and Table 3. As observed and in vie w of the . In the case of hydrogen-bonded fluids. for lack of reliable information pertinent to h ydrogen bonding o ver an extended range of temperature and/or pressure. As this is only one datum.32.1a for a number of common fluids. The critical compilation (Design Institute for Ph ysical Property Research [DIPPR]) 25 was used as a source for the thermodynamic data and the dipole moments of the studied fluids.27–28 Having determined the scaling constants through the pre viously mentioned procedure. An alternati ve e xpression for the factor f in Equation 3. w as set equal to –26. of the system is gi ven by: V = rNvv ∗ + N H VH (3. and the heat of vaporization of fluids by determining their three scaling constants through a least-squares fit. SH. the total v olume.1b. and are reported in Table 3.37) VH being the v olume change upon h ydrogen bond formation. the number of hydrogen bonds in the system is obtained through Equation 3. entropy. we will describe the v apor pressure.22. this approach can estimate (essentially predict) the dispersion and the polar components of the solubility parameter o ver a broad range of temperature and pressure.1a. F or simplicity.II.5 JK1mol1. the v olume change.

746 0.857 0.769 0.786 0.941 11.757 0.833 0.079 10.2.833 0.mol–1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2720 3585 6421 6444 5100 5080 5120 8250 6274 11133 8705 6390 7384 6780 7547 6570 5860 8565 5116 4412 4275 25100 25100 24771b 25100 25100 25100 25100 25100 23000 22940 21130 s = q/r s 0.1A Characteristic Constants of Pure Fluidsa Fluid Propane n-Butane n-Pentane n-Hexane n-Heptane n-Octane n-Nonane n-Decane n-Undecane n-Dodecane 3-Methyl pentane 2.737 1.418 0.941 0.mol1 3319 3800 4295 4557 4734 4870 4999 5111 5212 5265 4514.888 1.451 2.131 14.141 14.010 1.903 0.759 0.867 0.650 1.278 1.130 0.313 1.790 1.mol–1 Nonpolar Fluids 9.978 14.961 12.700 1.928 11.690 1.630 0.823 0.836 0.894 0.330c 0.922 1.757 0.050 1.151 0.881 0.619 9.867 0.881 0.881 0.873 10.00 14.650 1.841 1.15 13.259 1.850 0.320c 1.551 1.266 1.660 1.028 1.004 10.760 0.880 0.960 10.016 1.818 0.317 1.264 1.290 1.113 1.857 0.182 10.548 0.637 13. 4821 4794 5171 4986 5247 5656 5639 5403 5845 5664 3197 5295 3888 4406 3992 4383 4226 3840 4460 1811 4827 4650 3859 5304 2632 3267 3787 3740 4130 4376 4683 4977 5294 6199 4824 ν* = ε*P*–1/cm3 .870 7.925 15.971 1.844 0.916 17.018 13.635 0.760 0.4-Trimethyl pentane Cyclohexane Benzene Toluene Styrene o-Xylene Tetralin Acetone Acetophenone Ethyl acetate n-Butyl acetate Vinyl acetate Methyl methacrylate -Caprolactone Diethyl ether 1.856 0.439 1.437 1.857 0.992 9.903 0.619 1.904 .122 12.575 1.25 1.739 10.700 1.040 ν*sp = ρ*–1/ cm3 .720 0.050 0.57 14.373 1.887 3.131 12.45 14.075 11.087 11.134 10.881 0.098 11.960 1.308 –EH/ J.205 m/ Debye 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.736 9.140 1.830 0.121 10.843 0.4-Dioxane Carbon dioxide Chloroform Dichloromethane Vinyl chloride Chlorobenzene Methanol Ethanol 1-Propanol 1-Butanol 1-Pentanol 1-Hexanol 1-Octanol 1-Decanol Phenol Ethylene glycol ε* = RT*/J.050c 2.680 1.220 0.820 1.317 15.757 0.748 13.903 0.753 0.451 1.156 14.680 1.283 1.140 10.g–1 1.400 2.392 1.423 0.800 0.4-Dimethyl hexane 2.360 0.896 0.426 1.124 11.909 0.252 1.350 0.839 0.780 1.720 0.869 0.126 11.431 0.690 1.830 0.698 1.Statistical Thermodynamic Calculations 53 TABLE 3.301 1.297 1.577 15.652 1.850 0.370 0.445 1.030 0.560 1.142 13.750 15.049 10.075 11.839 0.806 Polar/Hydrogen-Bonded Fluids 9.400 1.840 0.650 1.146 12.670 4.219 1.952 13.908 0.800 0.

902 0. P.818 0.527 ν*sp = ρ*–1/ cm3 .474 11.750 0.874 0.908 0.899 0.2) 3700(6.147–0.121 9.360c 3.887 0.855 0.000087P 0.915 10. per repeat unit.039 0.2) 3600(8. respecti vely.430 12.146 8.821–0.mol1 2325 2751 4377 4417 4025 4193 4610 3634 3065 4175 4703 4840 4676 4787 5028 5525 5166 6400 5610 1234 2676 4222 ν* = ε*P*–1/cm3 . Numbers in parentheses are proton-donors and proton-acceptors.870–0.876 0.920 1.957 1.000039P 1.1B Characteristic Constants of Pure Fluids/Polymersa Fluid (Polymer) Polyethylene-lin.692 8.5 JK–1mol–1 in all cases.739 1.710–0.569 11.413 12.810–0.650c 1.000049P 0.1) 4800(8.g–1 1. in column 4 is in MPa.614 11.824 1.160 6.991 m/ Debye 3.mol1 5401 5993 6335 4876 5862 5294 5840 4817 5607 5029 ν* = ε*P*–1/cm3 mol1 13.000164P 0.2) 68000(2.95 7.052 10.000017P 0.2) s = q/r s 0.627 4.910 0.000016P 0.2-Propylene glycol Glycerol Diethylamine n-Butylamine Tetrahydrofuran Formamide Dimethylformamide Acrylonitrile Dimethylsulfoxide Acetic acid Acrylic acid Propionic acid Butyric acid Methacrylic acid Octanoic acid Oleic acid Stearic acid Ammonia Water a b c ε* = RT*/J.472 9.670 2.391 1.322 6.036 1.935 0.041 1.341 0.800 0.025 0.898–0.700 1.855 0.709 10.108 1.1) 8700(3.197 0.767 0.896 0.000020P 0.000037P m/ Debye 0.976 13.229 1.963 0.1A (CONTINUED) Characteristic Constants of Pure Fluidsa Fluid 1.110 9.2)b 400(6.798 1.000083P 0.925 0.g–1 0.939 0.54 Hansen Solubility Parameters: A User’s Handbook TABLE 3.3) 7800(10.088 9. Polypropylene Polystyrene Poly(vinyl chloride) Polyacrylonitrile Poly(methyl methacrylate) Polycarbonate (bisphenol A) Poly(ε-caprolactone) Poly(vinyl acetate) Nylon 66 a b ε* = RT*/J.806–0.2) 9550(6.739 1.411 12.799 0.298 1.909 0.825 0.000 8.750 1. TABLE 3.807 3.348 11.036 1.823 0.959 12.783 Pressure.861 SH was set equal to – 26.861 0.676–0.080 0 1.938 1.925 0.887 0.169 13.970c –EH/ J.631 1. .mol–1 1350(4. Adjusted to fit δhb.762 1.014 19.780 0.mol–1 22450 22520 12268 12325 8700 8770 16238 16120 9152 11038 20882 31600 24700 28880 21638 21553 26369 21090 19200 12277 17493 18100 s = q/r s 0.169 ν*sp = ρ*–1/cm3 .600 4.432 5.992 0.717 3.751 1.752 10.75 8.1) 4400(3.460 1.650 16.450 10.843 0.265 1.000070P 0.888 0.922 0.130–0. Adjusted to fit δp.728 0.649 1.649 0.963 9.839–0.300 1.61 12.867 3.451 1.175 0.866 0.850 0.903 0.869 0.850 0.mol–1 13.810 2.922 0.667 0.00 –EH/ J.861 0.

40 5.90 2.00 2.80 15.11 5.10 4.59 5.60 13.08 16.36 7.32 19.10 14.80 18.80 6.10 19.94 20.00 2.30 4.58 0 15.48 3.07 18.24 6.07 18.67 14.98 24.12 20.10 14.48 18.43 2.79 5.15 5.00 1.33 17.40 1.32 17.85 3.87 20.75 3.68 18.2-Propylene glycol Exp 5 δ HB Exp 5 δP Exp 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.40 15.95 13.90 5.14 7.90 19.41 24.25 0 15.68 3.41 5 Calc Calc 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.43 19.77 5.65 0 14.07 4.58 18.73 5.33 1.27 11.05 9.48 1.48 0.05 29.30 6.52 0 14.61 29.70 33.86 3.20 9.05 2.90 17.20 15.40 7.52 29.00 2.22 3.00 21.47 20.90 5.32 23.73 20.08 26.04 6.2.40 19.09 18.43 6.50 15.90 0 16.05 2.07 2.39 0 13.27 2.80 6.64 4.80 15.70 15.10 0 14.79 5.18 5.92 0.33 5.60 15.90 14.63 24.94 19.50 4.04 6.91 14.16 12.73 5.40 17.28 7.36 6.33 11.41 18.64 25.05 18.09 17.91 2.72 4.95 7.Statistical Thermodynamic Calculations 55 TABLE 3.30 0 15.70 7.56 13.07 7.31 9.84 6.19 23.4-Trimethyl pentane Cyclohexane Benzene Toluene Styrene o-Xylene Tetralin Acetone Acetophenone Ethyl acetate n-Butyl acetate Vinyl acetate Methyl methacrylate ε-Caprolactone Diethyl ether 1.03 24.10 4.58 23.85 0 14.80 2.30 0 Polar/Hydrogen-Bonded Fluids 18.52 20.10 4.80 5.86 11.01 0.79 6.69 14.90 15.76 Nonpolar Fluids 12.43 8.95 20.08 19.00 10.90 2.23 12.65 14.21 0 15.19 17.99 3.90 15.90 19.35 22.89 22.73 .08 7.20 6.45 24.40 21.09 4.77 25.61 18.40 15.10 3.50 3.63 5.30 12.00 10.90 11.65 21.60 5.07 4.40 0 15.0 2.45 24.95 33.27 8.74 29.40 17.00 1.00 19.00 2.70 5.02 1.40 14.41 21.90 3.79 19.17 17.34 0 14.05 18.74 20.36 14.80 21.2A Total and Partial Solubility Parameters (in MPa1/2) of Pure Fluids δ Total Fluid Propane n-Butane n-Pentane n-Hexane n-Heptane n-Octane n-Nonane n-Decane n-Undecane n-Dodecane 3-Methyl pentane 2.00 1.27 2.56 11.12 0 15.92 17.4-Dioxane Carbon dioxide Chloroform Dichloromethane Vinyl chloride Chlorobenzene Methanol Ethanol 1-Propanol 1-Butanol 1-Pentanol 1-Octanol 1-Decanol Phenol Ethylene glycol 1.37 10.90 14.34 8.80 19.77 16.29 4.66 15.92 4.4-Dimethyl hexane 2.20 18.59 17.30 18.91 6.00 10.40 2.01 3.04 19.13 26.30 24.70 4.86 1.78 2.

1a and Table 3.93 28.10 5.50 7.28 19.24 17. Polypropylene Polystyrene Poly(vinyl chloride) Polyacrylonitrile Poly(methyl methacrylate) Polycarbonate (bisphenol A) Poly(ε-caprolactone) Poly(vinyl acetate) Nylon 66 Exp 5 HB Exp 5 P Calc 2.80 3.26 38.70 26.86 15.40 4.90 5.40 9.99 13.75 21.81 18.07 2.80 18.40 10.2B Total and Partial Solubility Parameters (in MPa1/2) of Common Polymers Total Fluid Polyethylene-lin.82 29.31 6.46 19.25 6.90 3.40 7.00 8.80 14.37 18.17 5 Calc 34.18 6.20 8. There are a number of comments that should be made re garding Table 3.90 8.81 5.46 36.80 10.56 Hansen Solubility Parameters: A User’s Handbook TABLE 3.66 4.29 26.26 19.18 30.50 8.65 23.75 8. the o verall picture is rather satisf actory.00 24.00 approximations made.70 12.00 11.80 42.30 15.50 17. are essentially predictions of .18 23.43 20.41 13.1b.30 3.20 11.49 1.20 10.16 Thus.30 4.25 10.96 23.62 22. reported in Table 3.26 21.60 10.09 14.90 12.00 21.15 19.48 18.95 4.68 Calc 14.33 18.34 16.80 1.00 19.20 11.87 26.85 23. their calculated solubility parameters.80 0.30 4.20 18.28 12.10 3.12 16.00 8.31 2.70 16.89 4.37 22.20 5.04 24.0 4. We are not a ware of an y similar predictive approach in the literature in order to mak e the analogous comparison.70 21.32 12.00 16.44 5.00 TABLE 3.42 9.70 16.00 0.2 21.80 16.50 5.22 6.80 42.35 20.52 48.21 10.08 14. First.16 6.2A (CONTINUED) Total and Partial Solubility Parameters (in MPa1/2) of Pure Fluids Total Fluid Glycerol Diethylamine n-Butylamine Tetrahydrofuran Formamide Dimethylformamide Acrylonitrile Dimethylsulfoxide Acetic acid Acrylic acid Propionic acid Butyric acid Methacrylic acid Octanoic acid Oleic acid Stearic acid Ammonia Water Exp 5 HB Exp 5 P Exp 29.13 8.80 1.09 Calc 0.95 20.2a.75 6.93 19.10 6.98 6.35 24.90 25.43 21.00 8.52 20.58 25.00 34.00 2. the scaling constants for the nonpolar substances are identical to those reported pre viously.20 27.0 4.10 5.00 2.08 10.80 11.90 3.10 2.42 33.87 2.00 2.70 5.01 19.59 26.52 14.10 10.90 8.25 20.25 6.55 27.80 2.98 5.20 13.90 5.40 4.20 13.50 5.00 2.80 0.00 23.90 Exp 5 Calc 17.50 5.38 19.10 8.87 2.61 16.63 47.35 2.51 6.90 12.67 21.00 18.08 8.44 23.30 17.22 8.00 22.80 10.26 18.10 6.10 19.95 21.40 5.16 14.80 3.19 12.61 18.

This type of diagram is most useful for designing applications involving subcritical or supercritical w ater.1 Fractional solubility parameters for w ater. the series of 1-alkanols is another case where the calculated solubility parameters are. In these cases. In this case. where neither protons nor dipoles are present. in the second entry the v alue of m was adjusted on the basis of the e xperimental5 δp.1b.III. In a similar manner . and the reported numbers in Table 3. m. Third. the picture is somewhat more complex. Thus. As was expected. Figure 3. in all other cases of polar substances there is always an e xperimental hydrogen bonding contrib ution5 even in cases where there is no ob vious proton-donor and proton-acceptor pair. in the case of polymers.III. for some fluids there are tw o entries in Tables 3. An analogous diagram for ammonia is sho wn in Figure 3. the total number of proton donors and acceptors (for the polymer) are the reported numbers in Table 3. Fourth. Fifth. the scaling constants ha ve been changed as the interaction energy is no w split into its dispersion and polar components. 50 δ 40 δhb δ / MPa½ 30 20 δd δp 10 0 250 300 350 400 450 T/K 500 550 600 650 FIGURE 3. b ut the ne w parameter. one may compare the temperature dependence of these components as estimated by the alternative approach of Appendix 3. . the contribution of the polar component is as important as that of the h ydrogen bonding component. the estimated δp on the basis of the literature value25 for m was largely deviating from the experimental one.1b multiplied by the de gree of polymerization of each polymer . Second. all hydrogens were considered as equi valent proton donors and (the π-electrons of) the aromatic ring as the proton acceptor. apparently.1 sho ws the calculated components of the solubility parameter of w ater o ver an extended range of saturation temperatures. A typical example is the case of carbon dioxide.Statistical Thermodynamic Calculations 57 the model. in aromatic hydrocarbons. 16 Of course. and a fictitious v alue of m w as adjusted on the basis of the corresponding e xperimental5 value for δp. especially at lo w temperatures. as the h ydrogen bonding parameters are the same as in the original NRHB model.1 of Appendix 3. the donor–acceptor interaction w as replaced by the acid–base or the electrophilic–nucleophilic (carbon–oxygen) interaction. the main contrib ution to δ of w ater.1a and 3.1b of proton donors and acceptors refer to this basis. In this case.III. The repeating unit was considered the basis for the calculations. is not an adjustable parameter . predictions of the model. There are cases where e ven the dipoles themselv es are not obvious and. and thus. and it appears to override the hydrogen bonding component at the supercritical region. In the same figure. the dipole moment of the polymer w as ag ain adjusted on the basis of the e xperimental5 δp. In such cases. the polar component δp refers to quadrupole or higher -order interactions. essentially. the dipole moment. is h ydrogen bonding.

2B Comparison of the calculated dispersion component of solubility parameters of 1-alkanols with the solubility parameters of the corresponding n-alkane homomorphs.0 δd. ho wever. as in our previous approach.5 δalkane 15.5 14.8 16.58 Hansen Solubility Parameters: A User’s Handbook 17.0 δalkane 14. alkanol 16.4 16.0 15. δd.alkanol = 6.5 16.5 FIGURE 3. they do not e ven f all on a straight line.6 13. in polar fluids from the total solubility parameter of the corresponding homomorph h ydrocarbon.8 14. In .2 δd.2A Comparison of the experimental dispersion component of solubility parameters of 1-alkanols 5 with the solubility parameters of the corresponding n-alkane homomorphs.2 14. δd.6 15. As shown. the route for the estimation of the solubility parameter components through the homomorph concept is not al ways a safe w ay.0 15.6 17.0 13.0 FIGURE 3. of alkanols is compared with the total solubility parameter of the corresponding homomorph hydrocarbon. This approach. 16. It is often proposed.2A.86 + 0.5 14.4 12.0 14. The correlation line is: δd.5 15.60 δalkane .11–12 to estimate the dispersion component.0 13. where the dispersion component. is not al ways successful as sho wn in Figure 3.5 13.4 16. As a consequence. alkanol 15. the experimental5 data not only f all away from the diagonal.5 15.

According to the calculations. as calculated using the scaling constants in Table 3. essentially . 5 Barton. if not kno wn. but also for the selection of the e xternal conditions (especially. The obtained scaling constants are: ε* = 5267 J/mol. then one may use this information and the abo ve formalism to estimate the scaling constants and the h ydrogen bonding ener gy of the fluid through a least squares fit. The e xperimental25 and the calculated (essentially. when we compare these tw o solubility parameters as calculated by the present approach. v * = 10.3 van Krevelen.33. and the normal boiling point (or the vapor pressure at some other temperature) are either kno wn or can be estimated with reasonable accuracy.9495 cm3/g.6.34 through Equation 3.4.3B.36. temperature) for the dissolution of the polymer or an y other solute. and tetralin are similarly compared in Figure 3. compilations by Hansen. the better is the solv ent for the polymer. this procedure leads to a reasonably accurate estimation of the thermodynamic properties of fluids. Ho wever. chloroform is the best of the three solv ents for this polymer . care must be exercised once ag ain.. the slope of the straight line is much lo wer than unity. the dipole moment is also needed and.31 The partial solubility parameters for (Bisphenol A) Polycarbonate as functions of temperature. and all these findings agree with e xperiment. especially the similarity of a polymer . the best solvent for polypropylene appears to be tetralin. The distances ( Ra) for polyprop ylene with three solv ents: tetrah ydrofuran. 5. is the solubility parameter distance. The distances ( Ra) of this polymer with three common solv ents are compared in Figure 3. As observed. as in this figure. It can be seen that all three components are nonnegligible and there is a cross-over in the polar and hydrogen bonding components for this polymer .2B. A refined discussion on Ra and the related quantities Ro and RED = Ra/Ro is pro vided by Hansen. Ra. the data seem to f all at least on a straight line. which is ag ain corroborated by the e xperiment. is most useful for an additional reason: If the partial solubility parameters and the molar volume are known (e.38) The idea is: the smaller the Ra. for it. it may be estimated by v arious rather widely available ab initio or semiempirical quantum mechanics calculations. and a potential solv ent.1B.g. polar .1A. These scaling constants can no w be used for the estimation of the basic thermodynamic properties of the fluid at an y temperature and pressure. Of course. rather close to the corresponding parameters reported in Table 3. This type of figure is most useful not only for the mere selection of the solv ent. a nonsolvent for the polymer . A sphere with radius Ro encompasses the good solv ents for this polymer .). v sp* = 0. respecti vely. predicted) vapor pressures and saturated liquid densities for acetophenone are compared in Figure 3.5. 2. We have applied the above procedure to acetophenone by using Hansen’ s compilation5 for the partial solubility parameters and molar v olume. chloroform. as sho wn in Figure 3. DISCUSSION AND CONCLUSIONS A new approach has been presented for the estimation of the partial solubility parameters of pure substances. A most useful concept that quantifies the similarity of tw o substances 1 and 2. As shown.3A and Figure 3.177 cm 3/mol.30 etc.Statistical Thermodynamic Calculations 59 contrast. follo wed by tetrah ydrofuran (THF). defined by: 5 Ra = ⎡ 4 δ d 2 − δ d 1 ⎢ ⎣ ( ) + (δ 2 p2 − δ p1 ) + (δ 2 hb 2 2 − δ hb1 ⎤ ⎥ ⎦ ) (3. 1. Heptane is the w orst and. which led to the e xplicit forms of Equation 3. The capacity of the approach appears satisf actory for both the estimation of the partial solubility parameters and the description of the thermodynamic beha vior of fluids o ver a broad . and hydrogen bonding components in Equation 3. and DIPPR 25 for the normal boiling point and the dipole moment. are shown in Figure 3. and EH = 6293 J/mol. The splitting of the potential ener gy into its dispersion.

The author and his co workers are not a ware of an y similar integral approach in the literature in order to mak e comparison. has a number of features.1 1.cm3 0. Second. range of temperature and pressure.5 2.3A Experimental (symbols) 25 and predicted (line) v apor pressures for acetophenone. it may utilize a vailable information on . applicable to an y fluid regardless of its size and shape.6 250 300 350 400 450 500 T/K 550 600 650 700 FIGURE 3.0 0.8 0.5 0. which is presented in this w ork.5 P / MPa 1.0 1. The equation-of-state approach for the estimation of the partial solubility parameters. Third.7 0.0 250 300 350 400 450 500 T/K 550 600 650 700 FIGURE 3. it permits the estimation of the partial solubility parameters over an extended range of temperature and pressure.3B Experimental (symbols) 25 and predicted (line) liquid densities for acetophenone.60 Hansen Solubility Parameters: A User’s Handbook 2.0 ρ / g. 1.9 0. First. it is in principle.

Statistical Thermodynamic Calculations 61 20 16 d 12 δ / MPa½ 8 p 4 280 320 360 400 T/K hb 440 480 FIGURE 3. 10 n-Heptane 8 Ra / MPa½ 6 THF¹ 4 THF² 2 Chloroform 0 290 300 310 T/K 320 330 340 FIGURE 3.5 The estimated solubility parameter distance. The lines mark ed with THF1 and THF2 were obtained by using the first and second set of scaling constants in Table 3. respectively.1A. of (bisphenolA) polycarbonate (MW = 100000) with three common solv ents.4 Partial solubility parameters for (bisphenol A) polycarbonate. as a function of temperature. . Ra.

Stef anis and I. Fourth.40. One further possibility is the use of a group contrib ution method for the estimation of the dispersion component in much the same way as suggested in. on which the abo ve approach resides. orthobaric densities.III. δd can be estimated with an average absolute error of 0. M. for the hydrogen bonding component. ACKNOWLEDGMENTS The contribution of E. . it may act as useful guide for the selection of appropriate solv ents and/or dissolution conditions. of (linear) Polyprop ylene with three common solvents. A no vel element in our approach is the w ay the potential ener gy is split into its dispersion. The calculation of the polar component. the polar component is obtained by a simple subtraction.1A. and h ydrogen bonding components. Tsivintzelis in the preparation of tables and figures of this chapter is gratefully ackno wledged. The lines for tetralin and tetrahydrofuran (THF) were obtained by using the second set of scaling constants in Table 3. the statistical thermodynamic model.30. ho wever. 14 Such a group contrib ution method f ails dramatically. The contribution of C. and δd are available. As shown. polar. heats of v aporization. as a function of temperature. δhb. The alternative approach in Appendix 3. However. etc. Once the estimations of total δ. less than half the corresponding error for the estimation of total with the same method. is rather oversimplified and there is much room for improvement if one wishes to use more in volved expressions for the function f in Equation 3. the partial solubility parameters for the estimation of the scaling constants of substances for which there are no a vailable e xtensive e xperimental data on v apor pressures.62 Hansen Solubility Parameters: A User’s Handbook 10 THF 8 Ra / MPa½ Chloroform 6 4 Tetralin 2 290 300 310 T/K 320 330 340 FIGURE 3. Ra. which enhances further the usefulness of the approach reported in the present work. Of course. the approach of this w ork is sufficient and most v aluable. can be used for a detailed description of the phase diagrams of pairs (mixtures) of fluids when the scaling constants and the binary interaction parameters are a vailable.II is one example. in particular . 14 as sho wn in Appendix 3. when a f ast screening of potential solvents is needed. Hansen through his v aluable comments is also gratefully ackno wledged.6 The estimated solubility parameter distance. Significant progress could be made if e xperimental information on the partial solubility parameters as functions of temperature and pressure were a vailable.

Statistical Thermodynamic Calculations 63 LIST OF SYMBOL SPECIAL TO THIS CHAPTER E G H K L N N Nr No Nij P R R m S s T Q(N. T ensemble.20) .2. p. T) V* V X Y z zq Potential energy Gibbs free ener gy Enthalpy Boltzmann’s constant Staverman’s parameter Number of e xperimental points Number of molecules Total number of lattice sites Number of empty lattice sites Number of contacts of type i-j Pressure Gas constant Number of se gments per molecule Dipole moment Entropy Surface to v olume fraction Temperature Configurational partition function of fluid in the N. see Equation 3. 46 Average segmental volume Volume ole fraction in liquid phase ole fraction in v apor phase Lattice coordination number Average number of e xternal contacts per molecule GREEK LETTERS ν* Γ Δ E Θ Θ M P φ Ω Ω Segmental volume Non random f actor Solubility parameter Interaction energy Surface (contact) fraction Hole-free surface (contact) fraction Chemical potential Density Segment fraction Geometric–flexibility parameter Combinatorial term (In Equation 3.7 and Equation 3. P. P .

Peña. Eng.M. T. 95. J. 4. 1982. Panayiotou. 27.. Regular Solutions. Applications of solubility parameters and other cohesion parameters. J. Panayiotou. and Weingerl.. Jones. Sci. Chem. 23. and Eon. Nezbeda. K. Karger. 25. I. I. Barton. Hansen. 2. 6253–6360. 27. Perry... Eng.. p Dm H O R Dispersion. and P anayiotou... Boca Raton. Eng. and Green. CRC Press. 125. 203.E. hydrogen bonding. T. 22.M. 43. E. hb... 267. J. Prentice-Hall. 1974. Ind. 21.. E. Rasaiah. surf aces. Fluid Phase Equilibria... and Gubbins.C. No. New York. 20.. Res... U. Fluid Phase Equilibria.. Sp yriouni. NJ. 1996. and Vera. C. J. Guggenheim. Jensen. P antoula... Polym.. B. E. 33. 1978. and polar component.. 43. Nezbeda. Constantinou. 51(1). Eds.L. 8. 51(2005). J. Tehrani. R. 24. J. Panayiotou. J. and P avlíek. Boca Raton.. I. K. 11. J. 1997.. 1393–1414.. Karakatsani. and Narang. 1987.L. 144–154. and Scott.A.. Ne w York. I. 10. Englewood Cliffs. E. C. E...F.H. The three dimensional solubility parameter — k ey to paint component af finities I.. Stell. 236. Twu. 2001. and P anayiotou. 1999. 131. Chem. 1985. 174.. Hansen Solubility Parameters: A User’s Handbook. respecti vely Quantity pertinent to dimer Hydrogen bonding quantity Property pertinent to holes Property pertinent to molecular se gments REFERENCES 1. 1976. 1086–1099. pp. Phys. Res.F.. 99.E. Clarendon Press. 2004..L. Chem. Phys.R. 2004.B. 35. Hansen.. Org. 19. Ed. 1962.. Tsivintzelis. Oxford.. C.64 Hansen Solubility Parameters: A User’s Handbook SUPERSCRIPT ~ * L V Reduced quantity Scaling constant Liquid phase Vapor phase SUBSCRIPT d. D. I.. and Barra. Hildebrand.H. J. 9.. L. Technol. 1996.. 14. 13. Sn yder.. C. J. A. 1952. and Economou... 57(7). Mixtures. in Surface and Colloid Science in Computer Technology. Mittal..M. Stefanis. M. 1998. 21–35. 116. FL. Fredenslund. 39(505). A. P. Paint Technol. Ed. AIChE Symp. 12.. New York. K.. 40hh-40mm.. 141–150. A. 55–66. Panayiotou. Chemical Engineers’ Handbook. American Institute of Chemical Engineers. I. 16. 18. February 1993. 71–88. 10090–10097. L. 17... 104–117. 1985.. J. Am.A.. Kraska. 1595–1606. Eng. 1983. and Gubbins. . Sci. Stef anis. 26.. 1975. C. T. 5. R.M.E. Panayiotou. C.. 4727–4737. 240. and Prausnitz. Ser. 6592–6606.. R.. M.. Phys. C. 2005. 2004. Fluid Phase Equilibria. Mol. 1991. Daubert. 14. CRC Press.. Lab. 863–878. Ind.. and dif fusion in polymers. C. Prog. G.. Chem. 530–536.. 2006. Aspects of solubility ... 1999.. and Sanchez. Chem. and Economou.P. 7. Int. 681–694. C. Tsivintzelis. McGra w Hill.. J.L.M.. 27–59... Res. 21. 6. Stefanis.. Barton. C. Coat. 905–912. Pharm. Data Compilation Tables of Properties of Pure Compounds. R.M. and Danner . CD. AIChE J. AIChE J. Handbook of Solubility Parameters and Other Cohesion Parameters. Ind. Chem. Fluid Phase Equilibria. Bustamante. 2328–2342.. Chromatogr. Plenum Press. Polym. 3.. FL. W. C. Mol. 15. H. C. I.C. 1967. Hansen. Pure Appl.

Hansen.. Coat.I.2) ways. 77–84. in Models for Thermodynamic and Phase Equilibria Calculations. 30.Statistical Thermodynamic Calculations 65 28.I. A. 29.M. P.I. Technomic Publ.3) is the free energy change upon formation of one dimer. Prog. van Krevelen. Marcel Dekker. 13 one may derive the formalism for the acid dimerization in a rather straightforward manner . (2 N ) ( dm N! ! N − 2Ndm ! ) (3. 1994. 31. 1976. Zoller. One can select these dimerized molecules out of the N acid molecules in. New York. Fifty years with solubility parameters — past and future. the hydrogen bonding factor in the partition function becomes: QH = ⎛ ρ ⎞ N! N dm ⎜ ⎝ rN ⎟ ⎠ N − 2 N dm ! N dm ! 2 N dm ( ) ⎛ N G ⎞ exp ⎜ − dm dm ⎟ RT ⎠ ⎝ (3. Basel. Lancaster . 2004. we will consider dimerization only .I.. PVT Data for Polymers. Org. If Gdm = E dm + PVdm − TS dm (3..4) The equilibrium number of dimers per mol of segments of acid. Properties of Polymers. Let Ndm be the number of dimers in the system. 51(1). S. F or simplicity .... C.6) 1 − K dm 1 4 + 2 K dm K dm 4r (3.I.I: THE ACID DIMERIZATION Following the LFHB practice. and Walsh.. and form the Ndm dimers in (2 N ) ( dm N! N! 2 N dm − 1 2 N dm − 3 .1 = ! N − 2 N dm ! N − 2 N dm ! N dm ! 2 N dm ) ( )( ) ( ) (3.I.W. D.5) In this case of dimerization. Elsevier.I. Fredenslund.. as dimers are the overwhelming majority of the association species in h ydrogen-bonded acids.7) . APPENDIX 3. 2nd ed. νdm..1) ways. Sandler. the h ydrogen bonding contrib ution to the chemical potential is: μH 1 = r ν dm − ln RT 1 − 2 r ν dm (3. K dm = ⎛ −Gdm ⎞ ρ exp ⎜ r ⎝ RT ⎟ ⎠ (3. D. 1995. is obtained from the above equation through the usual free ener gy minimization condition. Co. Amsterdam. or 2+ ν dm = where.. Ed..

which is no w given by: ⎡ 4 ⎛ m ⎞ 2 s2 ⎤ ⎥ Γ rr qN θr ε∗ ⎢ ⎜ ⎟ ⎢π ⎝ r ⎠ T ⎥ ⎣ ⎦ V δp = (3. the tw o alternati ve approaches for the estimation of the polar component lead to differences not only in δp but also in δd and to the total . This is a rather gross simplification as one w ould expect the dipolar forces to be in versely proportional to temperature or to a function of temperature.II. ho wever.III. the scaling constants of the fluids.III. intact. almost identical manner to the one obtained by the corresponding scaling constants of the main text.66 Hansen Solubility Parameters: A User’s Handbook APPENDIX 3.II: AN ALTERNATIVE FORM OF THE POLAR TERM The factor f in Equation 3.2.14 Two kinds of functional groups are used: First-order (UNIF AC groups) and second-order groups that are based on the conjug ation theory. which now becomes: 2 ⎡ 4 ⎛ m ⎞ s2 ⎤ ⎥ − N H EH − E = Γ rr qN θr ε∗ ⎢1 + ⎜ ⎟ ⎢ π⎝ r ⎠ T ⎥ ⎣ ⎦ (3. In addition.II. 19–24 This has been e xplored by adopting the follo wing alternative form for the f actor f: 2 ⎛ m ⎞ s2 π⎜ ⎝r⎟ ⎠ T 2 f = (3. This is important.II.III. the predicted partial solubility parameters by the tw o sets of the scaling constants are compared in Table 3.1 and Equation 3.III: A GROUP-CONTRIBUTION METHOD FOR THE PREDICTION OF δ AND δD The details of the group contrib ution method may be found in the original w ork.32 for the dipolar interactions implies that the dipolar forces depend on temperature and/or v olume in the same manner as do the dispersion forces.2 do not lead to any clear improvement in this respect either. As observ ed. The hydrogen bonding component appears. The basic equation that gives the value of each property according to the molecular structure is: f(p) = ∑n F i i i + ∑m j j Sj (3.II.2) This change will change. These ne w scaling constants are reported in Table 3.1 for some representati ve fluids. the formalism of the main te xt remains the same e xcept for the equation for the potential ener gy. as δhb may be used in approaches lik e the one reported in Appendix 3.3) As sho wn in Figure 3.II. APPENDIX 3. These constants describe the key thermodynamic properties of the fluids in a similar .1) . Equation 3.II.1) With this expression.1. The essential difference is the dependence of δp on temperature. of course.II.

026 10.060 12.216 10.798 1.mol–1 5041 5236 5858 4207 5875 4444 4741 5046 4040 4435 4859 5062 5517 3798 4042 4656 5280 6868 5681 4386 2595 4471 4927 4491 5110 4981 2506 2070 ν* = ε*P*–1/cm3 .178 1. δ.000 12. δ.500 23.704 10. at 25°C are presented.2 sho ws the second-order group contrib utions for the same properties. A least-square analysis has been carried out to estimate the first-order and second-order group contrib utions for the solubility parameters.685 11.865 0.048 12. and is selected after a thorough study of the ph ysicochemical and thermodynamic beha vior of the property .262 1.145 0.709 10. Sj.078 1.800 30.997 –EH/J.Statistical Thermodynamic Calculations 67 TABLE 3.III. The determination of the contrib utions is done by a two-step regression analysis for the F is and the S js.747 0.mol–1 3724 3590 5055 7970 6192 10915 8674 5383 6630 6501 8610 4992 4269 25100 25100 25100 25100 23300 21775 22400 24600 12240 11980 8810 23735 23822 11940 18198 SH was set equal to 26. In Table III.257 1.171 0.902 0. respectively.399 0. the first-order group contrib utions for total solubility parameter . p.g–1 1.5 JK –1mol–1 in all cases. Adjusted to fit δhb.170 13.250 14.040 14.029 0.2) .873 1.293 11. and the dispersion partial solubility parameter (Hansen). f(p) is a single equation of the property .526 10.940 0. at 25°C ((kJ/m 3)(1/2)): δ=( ∑n F i i i + ∑m j j S j + 75954. Table III.054 1.958 1. where Fi is the contribution of the first-order group of type i that appears n i times in the compound and.II.962 1.034 10.776 7.14 (3.031 15.805 20.232 14.4-Dioxane Chloroform Dichloromethane Chlorobenzene Methanol Ethanol 1-Butanol 1-Octanol Phenol Ethylene glycol 1.000 1.617 0.899 14.942 0.019 7.484 17.033 1.383837 − 56.684 11.961 0.1 Characteristic Constants of Pure Fluidsa Fluid Benzene Toluene Tetralin Acetone Acetophenone Ethyl acetate n-Butyl acetate Methyl methacrylate Diethyl ether 1.1.862 ν*sp = ρ*1/cm3.019 1.mol–1 9.155 1. The selected equations for the estimation of each property are the follo wing: Total solubility parameter.192 0.145 1.261 7. δd. is the contribution of the second-order group of typej that appears mj times in the compound.2-Propylene glycol Glycerol Diethylamine n-Butylamine Tetrahydrofuran Acetic acid Butyric acid Ammonia Water a b ε* = RT*/J.1) 0.420 13.093 1.710 14.

20 6.86 14.41(18.74(23.50 23.05 2.80 43.70 3.52 10.00 2.16(0.93 23. − X exp .20(20.22) 4.68 Hansen Solubility Parameters: A User’s Handbook TABLE 3.86 1.86(4.25 6.41 18.10(14.93(1.11 7.43 15.70) 18.70 29.90 25.52 34.31 19.71(15.63 33.14(34. As observed.95 3.36 12.90 6.0 10.35 20.72) 3.27 5.46 21.04(6.18) 20.15 Expa 1.48) 20.94 20. δd.93) 27.30 19.90) 20.75) 6.96) 28.32 19.61 18.40 2.64) 31.79) 5.43 8.2 Total and Partial Solubility Parameters (in MPa1/2) of Pure Fluids Polar/Hydrogen-Bonded Fluids Total Fluid Benzene Toluene Tetralin Acetone Acetophenone Ethyl acetate n-Butyl acetate Diethyl ether 1.70) 2.84 3.4-Dioxane Chloroform Dichloromethane Methanol Ethanol 1-Butanol 1-Octanol Phenol Ethylene glycol 1.70(27.02 1.01(12.86(29.31) 17.67(20.84(0.12 16.08) 19.10 15.98) 15.00 13.04) 21.41 12.79 29.09 22.30 4.27)a 17.84(15.39(4.97(16.36 5.51) 16.95 3.07) 4.58) 24.72(17.98(8.14) 7.16) 16. at 25°C ((kJ/m 3)(1/2)): δd = ∑n F i i i + ∑m j j S j + 17.08(19.69) 33.20 6.92) 1.70 16.01) 0.73 4.10(15.73) 19.20 27.32 29.90(29.98) 7.46(7.84(1.31(5.15(24. The quantity mjSj is considered to be zero for compounds that do not ha Table 3.70(5.11(11.73 18.II.32(18.III.34) 8.40 47. the method is rather quite satisf actory.2) 13.05 2.29(20.80(48.41) 10.59(24.00 8.78) 18.48) 1.27 8.95 21.95(17.97(11.21) 17.10 8.10 8.09 24.27(3.91 6.75) 6.27) 24.63) 12.87 24.00 2.07 7.31) 2. and Average Absolute Error = AAE = ∑X est .70(16.91) 3.68) Values in parenthesis from Table 3.93(12.04) 5.3) ve second-order groups.68 9.61 26.82 Calc 2. Dispersion partial solubility parameter .34) 16.48) 18.60(12.89) 4.80 11.59 5.70 2.49(8.00 13. Standard Deviation = ∑ (X est − X exp )2 1 N N .36 5.68 9.66(18.82 18.11 7.97(33.35 20.00 8.47 18.30 5.60 17.48 17.26(26.08) 22.80 5.19) 34.73 4.3231 (3.90 25.94) 14.52 10.80) 11.56(5.77 23.00) Expa 18.44(10.3 illustrates the statistics concerning the o verall improvement in the estimation of solubility parameters that has been achie ved after the introduction of second-order groups in the regression.85(18.32 3.69(5.34(3.07 2.2-Propylene glycol Glycerol Diethylamine n-Butylamine Tetrahydrofuran Acetic acid Butyric acid Ammonia Water a HB Calc Expa 2.02(19.30 5.80 19.50 5.46(19.80 42.70 7.30(20.59 15.08) 20.66 20.89) 26.98 4.80) 18.2A.90 11.III.56) 12.44 30.52) 47.05(18.73(26.92(22.43) 15.70(6.80) 21.90 6.92) 30.98(3.49(2.05 9.24) 5.00 P Calc 0.

Fi –0. δd = ∑n F i i i + 17.II.Statistical Thermodynamic Calculations 69 30 (A) 25 δT 30 (B) 25 δT 20 δ / MPa½ δ / MPa½ δhb δp δd 20 δhb 15 δp 10 δd 15 10 5 290 300 310 320 330 340 350 360 370 380 T/K 5 290 300 310 320 330 340 350 360 370 380 T/K FIGURE 3. C Occurrences.39814 17. where N is the number of data points.1A.02686 –0.3231 = 15. Example of Prediction of the Hansen Partial Solubility Parameter.08058 –0. such as δhb.92496 (kJ/m 3 )(1/2) . Average Absolute Percent Error = AAPE = 1 N ∑ Xest − X exp p X exp 100% . ni 1 3 1 — — Contributions. B: with the constants from Table 3.1.34621 — — niFi –0. Syntactic isomers can be distinguished.1 Fractional solubility parameters for ammonia. and X exp the experimental value. A: with the scaling constants from Table 3. The estimation of one of the other Hansen solubility parameters. X est is the estimated v alue of the property .3231 No second-order groups are in volved.34621 –1. for 1-Butanol First-Order Groups –CH3 –CH2 –OH ΣniFi Universal constant. δd.97135 –0. It is w orth pointing out that the solubility parameters of comple x structures that occur in aromatic or multiring compounds of biochemical interest can no w be predicted by only using their molecular structure and without an y other data kno wn. whereas stereoisomers cannot. could lead to the estimation of δp as well.97135 –0. as described in the main te xt.II.

III.3-Dimethyl-2-butene (1) 1.26857 –1.2 –2766.4-Dioxane (2) 1-Amino-2-propanol (1) Isopropylamine (1) n-Methylaniline (1) di-n-Propylamine (1) ylamine (1) Trimethylamine (1) Triethylamine (1) Aniline (1) 2-Methacrylamide (1) n-Methylacetamide (1) N.23203 –0.8 3381.8 –206.3 Contributions to δd –0.8 4686.6-Dimeth 1.35506 0.84464 0.58280 0.8 12228.1 First-Order Group Contributions to (Total) δ and δd at 25°C First-Order Groups -CH3 -CH2 -CH< >C< CH2 = CH-CH = CHCH2 = C< -CH = C< >C = C< CH2 = C = CHCHECCEC ACH AC ACCH3 ACCH2CH3CO CH2CO CHO COOH CH3COO CH2COO HCOO COO OH ACOH CH3O CH2O CHO CH2O (CYCLIC) CH2NH2 CHNH2 CH3NH CH2NH CHNH CH3N CH2N ACNH2 CONH2 CONHCH3 CON(CH3) 2 C 5 H 4N C 5 H 3N CH2SH CH2S I BR CH2CL CHCL Contributions to δ –2308.9 8456.8 2166.9 3650.4 14930.7 1229.27531 –0.2 9477.53723 0.3 6574.1 5194.65178 0.1 1601.52883 –0.69868 –0.57169 0.11050 0.03098 0.3 –6.02686 0.70 Hansen Solubility Parameters: A User’s Handbook TABLE 3.4 8616.29130 na 0.20386 –0.1 –480.9 12770.40303 –0.64501 1.26226 0.6 –277.9 –980.34621 0.2 3183.4 3269.1 7274.29100 –0.1 –355.8 1865.06889 na 0.2 1716.0 3671.1 3733.87693 1.8 –3745.2 –381.88334 0.5 2169.3 426.6 9228.65267 –0.2-Butadiene (1) Propyne (1) 2-Butyne (1) Benzene (6) Naphthalene (2) Toluene (1) m-Ethyltoluene (1) Methyl ethyl ketone (1) Cyclopentanone (1) 1-Butanal (1) Vinyl acid (1) Ethyl acetate (1) Methyl propionate (1) n-Propyl formate (1) Ethyl acrylate (1) Isopropanol (1) Phenol (1) Methyl ethyl ether (1) Ethyl vinyl ether (1) Diisopropyl ether (1) 1.1 27386.8 2163.58277 0.48289 0.35922 –1.2 1887.27971 1.77971 0.97135 –0.2 3585.81162 na di-Isoprop 0.2 5398.4 684.0 –975.44622 Sample Group Assignment (occurrences) Propane (2) Butane (2) Isobutane (1) Neopentane (1) Propylene (1) cis-2-Butene (1) Isobutene (1) 2-Methyl-2-butene (1) 2.05949 0.54006 0.21737 0.5 –176.46810 1.3 –221.44822 na 2-Meth na 2.00476 –0.N-Dimethylacetamide (1) ylpyridine (1) ylpyridine (1) n-Butyl mercaptan (1) Diethyl sulfide (1) Isopropyl iodide (1) 2-Bromopropane (1) n-Butyl chloride (1) Isopropyl chloride (1) .01116 na 0.20284 0.7 2191.7 560.8 1923.5 –2662.05853 0.69325 –0.2 4183.8 1023.

00-15.6 7805.00 MP a1/2 Percentage error = (16. experimental5 δd =16.7 –2965.33919 n-Butyronitrile (1) –0.92 MP a1/2 and δhb = 15.4 na 0.2 na 4770.2-dichloropropane (1) na Benzotrichloride (1) 0.11704 Fluorobenzene (1) 0.6 1164.1 –795.3 –2326.00 = 0.70693 2-Fluoropropane (1) –0.55021 Sulfolene (1) 0.III.48988 Thiophene (1) 1.72 MP a1/2.6 1234.1-Dichloropropane (1) 2.7 5026.6-Trimethylpyridine (1) 0.0 na 0.7 12706.8 –1643.7 6303.28043 3-Methyl-1.17971 Sample Group Assignment (occurrences) t-Butyl chloride (1) 1.6 636.925 MP a1/2.70 MPa1/2.1 2694.84750 m-Dichlorobenzene (2) 0.4 –7339.22931 Perfluorohexane (2) na 1-Nitropropane (1) na 2-Nitropropane (1) 1.80 MPa1/2.41953 Nitrobenzene (1) –0. .77470 N-Methylthiopyrrolidone (1) na T riphenylphosphine (1) –0. Contributions to δ –1415.22893 2. this equation-of-state approach estimates = 22.3-Pentadiene (1) na Diisopropyl ketone (1) 0.17069 4722.4 na na Contributions to δd 2.54378 Triphenylamine (1) Thus. The experimental value5 is 5.04716 Divinyl ether (1) 0.1 (CONTINUED) First-Order Group Contributions to (Total) δ and δd at 25°C First-Order Groups CCL CHCL2 CCL2 CCL3 ACCL ACF CL-(C=C) CF3 CH2NO2 CHNO2 ACNO2 CH2CN CF2 CF C 4 H 3S F (except as abo ve) CH2 = C = C< CH = C = CHCHCO O (except as abo ve) Cl (except as abo ve) NH2 (except as abo ve) >C = N-CH = NNH (except as abo ve) N = NCN (except as abo ve) NO2 (except as abo ve) O = C = NCHSH CSH SH (except as abo ve) S (except as abo ve) SO2 >C = S >P>C = 0 (e xcept as abo ve) N (except as abo ve) Note: na = not a vailable.II.Statistical Thermodynamic Calculations 71 TABLE 3.36532 –1208.2 14215.925)/16.5 3380.6 10253.8 2467.3-Dichloropropene (1) –0.96719 Isoquinoline (1) na Dibenzopyrrole (1) na p-Aminoazobenzene (1) 0.5 –5199.4 3459.04271 2-Mercaptobenzothiazole (1) 1.75755 1.8 2230.0 1655.97290 Perfluoromethylcyclohexane (5) Perfluoromethylcyclohexane (1) na 2-Meth ylthiophene (1) –0.8 –3464.2 –701.4.7 1332.30737 2. estimated δd =15.47% According to Table 3.5 9359.5 –473.5 10030.2-butadiene (1) na 2.3 –841.0 26271.2.08615 cis-Crotonitrile (1) na Nitroglycerine (3) –0.22562 Hexachlorocyclopentadiene (2) na Melamine (3) –0. These data combined with the group contribution result for δd give: δp = 4.43429 Anthraquinone (2) 1.13065 n-Butyl isocyanate (1) na Cyclohexyl mercaptan (1) na tert-Butyl mercaptan (1) 1.

4 1267.0 –3745.04509 –0.1 11489.06801 na –0.24676 –0.2.4 142.6 3306.5 511.27979 –0.00000 –0.29806 –0.56461 na na 0.9 –762.1 134.0 –359.1 –1634.3 1581.7 –228.3.7 –1257.4 –2875.48916 0.22183 0.5 –23.9 Contributions.25679 0.1 592.18466 –0.5 –3673.8 9215.22930 na –0.08088 –0.4 5020.38742 –0.29563 na 0.2-Propanediol (1) 1-Amino-2-propanol (1) Cyclohexanol (1) Ethyl vinyl ether (1) Methyl phenyl ether (1) Cyclopentimine (1) Tetrahydrothiophene (1) Bromobenzene (1) Iodobenzene (1) 2-bromo-propene (1) Methyl isopropyl ketone (1) Cyclopropane (1) Cyclobutane (1) Cycloheptane (1) Methyl benzoate (1) Naphthalene (1) Diketene (1) Diphenyl ether (1) Ethylbenzene hydroperoxide (1) di-t-Butyl peroxide (1) 2-Pyrrolidone (1) Oxazole (1) Methylal (1) Dibenzopyrrole (1) 2.02003 na na –0.08764 0.08394 na –0. Contributions.6 –4646.66808 0.C = CCH3-C = -CH2-C = >C{H or C}-C = string in c yclic >CHCHO CH3(CO)CH2C(cyclic) = O ACCOOH >C{H or C}-COOH CH3(CO)OC{H or C}< (CO)O(CO) ACHO >CHOH >C < OH -C(OH)C(OH)-C(OH)C(N) C(in cyclic)-OH C-O-C = C AC-O-C >N{H or C}(in c yclic) -S-(in cyclic) ACBr ACI (C = C)-Br CH3(CO)CH< ring of 3 carbons ring of 4 carbons ring of 7 carbons ACCOO AC(ACHm)2AC(ACHn)2 Ocyclic-Ccyclic = O AC-O-AC CHn-O-OH CHm-O-O-CHn NcycHm-Ccyc = O Ocyc-CcycHm = Nc yc -O-CHm-O-CHnAC-NH-AC C(= O)-C-C(= O) Note: na = not a vailable.8 11.4-Pentanedione (1) . δ 142.0 –492.6 –2637.20629 0.40589 na 0.III.07853 –0.1 –9764.11231 –0. δd 0.07377 na na na –0.5 2493.2 –426.7 2389.13554 –0.0 –3076.5 –2029.2.2 Second-Order Group Contributions to (Total) δ and δd at 25°C Second-Order Groups (CH3)2-CH(CH3)3-C-CH(CH3)-CH(CH3)-CH(CH3)-C(CH3)2-C(CH3)2-C(CH3)2ring of 5 carbons ring of 6 carbons -C = C .3-Dimethylbutane (1) 2.4 4022.72 Hansen Solubility Parameters: A User’s Handbook TABLE 3.5 2002.27069 0.32357 –0.9 3315.2 2678.1 –8721.3 2.12341 0.37514 0.3-Trimethylbutane (1) 2.4 337.3-Tetramethylpentane (1) Cyclopentane (1) Cyclohexane (1) 1.1 na –437.7 –524.5 –69.4 5677.52196 –0.19450 na –0.6 –620.37724 0.4 –83.48615 Sample Group Assignment (occurrences) Isobutane (1) Neopentane (1) 2.8 –3668.04604 –0.4 –1486.2 626.3-Butadiene (1) Isobutene (2) 1-Butene (1) 3-Methyl-1-butene (1) Ethylcyclohexane (1) 2-Methylpropanal (1) Methyl ethyl ketone (1) Cyclopentanone (1) Benzoic acid (1) Isobutyric acid (1) Isopropyl acetate (1) Acetic anhydride (1) Benzaldehyde (1) 2-Propanol (1) Tert-Butanol (1) 1.

61 1.47 0.44 0.41 5.67 2.Statistical Thermodynamic Calculations 73 TABLE 3.3 Comparison of the First.and Second-Order Approximations Standard Standard Data Deviation Deviation AAE AAE AAPE (%) AAPE (%) Property Points First-Order Second-Order First-Order Second-Order First-Order Second-Order δ 1017 δd 344 1.90 0.15 2.III.62 4.00 0.31 0.58 1.42 .


colorants. Compatibility of plasticizers in PVC 7. etc. polyolefin industr . Examples include: 1. The control of emissions from paints as well as the swelling of the film in the presenc of water 4. The recycling of polymers based on ph ysicochemical methods like selective dissolution 5. In several of these cases it is not suf ficient to empl y only the Hansen solubility parameters (HSP). As several of the literature approaches used for 75 . polymer molecular weight.. for e xample. Free-volume effects. LCST . dendrimers). such LLE may tak e v arious forms. The compatibility of polymer blends including those with no vel structures (star -like. including solv ent activities. This is the purpose of this chapter .Hansen Solubility 4 The Parameters (HSP) in Thermodynamic Models for Polymer Solutions Georgios M. aqueous tw o-phase systems based on polymers for separating proteins This is only a short list. In the biotechnology . and temperature. by-products. Polymers form highly nonideal liquid solutions with lo w-molecular weight chemicals and liquid–liquid phase separation (LLE) is the rule rather than the e xception in polymer -solvent mixtures. and other additives after polymerization 2. for solvent emission assessment or even full phase diagrams and at both lo w (e. It is rather tempting to combine the e xtensive use/tables available for the HSP with a rigorous thermodynamic approach and compare the performance of this method to more adv anced approaches. closed loop.g. as much more detailed calculations may be needed. fewer VOC) 3. The selection of mix ed solv ents in the paints and coatings industry to ward designing environmentally-friendly paints (w ater-based. gas permeabilities in polymers). permeabilities of g ases in fl xible polymeric pipes used in the North Sea and other major oil and g as producing areas for transporting of h ydrocarbons on the seabed and from the seabed to the surf ace 6. biotechnology) and high pressures (e. special structures and crystallinity cause additional comple xities. The removal of unreacted monomers.g. and polydispersity have great effects. Kontogeorgis ABSTRACT Polymer thermodynamics plays an important role in a lar ge number of processes and in the design of many different polymer-based products.. toxic compounds. UCST . and man y more applications of polymer thermodynamics e xist.. Moreo ver.

76 Hansen Solubility Parameters: A User’s Handbook comparison purposes are based on the group-contrib ution principle.6 GCVOL can predict satisfactorily the density of solvents. and one CH3CO group. methanol) because further di vision is not possible if good results are to be obtained. Similarly . including copolymers. e. instead of kno wing the parameter v alues of a specific property for thousands of molecules..g. including also mechanical and environmentally-related properties.1. Van Krevelen provides extensive GC tables for the Hildebrand and Hansen solubility parameters as well.. one CH2. However.. the solubility parameter (Hildebrand and HSP). it is an approximation.7 The great advantage of the group contribution method is its simplicity: Even though there may be thousands of different molecules (and mixtures). the melting and glass transition temperatures. Problems can also be e xpected for multifunctional groups and where more than one polar groups are close to each other (e. 1 Van Krevelen2 gives an overview of the application of group contrib ution methods to se veral properties of pure polymers. The GC methodology has been applied to man y properties and for both lo w molecular weight compounds and polymers. melting point. ni is the number of groups of type i and Fi is the corresponding group v alue. the corresponding number of groups is significantly smaller (no more than 100 or so). Thus. it is su ficient to kn w the group parameters for a much smaller number of groups. GC methods perform much better than for others.g. Many properties of pure polymers and polymer solutions can be estimated with group contrib utions (GC). butanone has one CH3.g.4. oligomers. and polymers. density. then the total v alue of the property for the whole molecule is often e xpressed by a general additive rule of the form: F= ∑n F i i i (4. For example. e.. the general GC equation has a more complicated form than that indicated by Equation 4. the UNIFAC model by Fredenslund et al. F or some properties.g. often within 2%. The GC methodology is a v ery useful technique leading to good results in man y cases. e.g. in alcohols and acids with more .5 A list of GCVOL parameters is provided elsewhere. Fi values are also functions of temperature for temperature-dependent properties such as the v olume and the v apor pressure. n-hexane (CH 3-(CH2)4-CH3) can be considered to ha ve two CH 3 and four CH 2 groups. In Equation 4.1) or similar additive equations. An alternative GC method for the polymer (and solvent) density has been de veloped by Elbro et al. Specific molecules are assigned a separate groups (e. activity coefficients. F or several properties. Activity coefficients h ve also been predicted with group contrib utions. based often on a some what unjustified d vision of the molecule into groups. 3 (GCVOL) and was recently extended to cover several group families. In some cases. and the sur ace tension. If the group values are known for a specifi property F. such as for density . Two limitations of the approach should be k ept in mind: 1. a short introduction to this principle is pro vided first GROUP CONTRIBUTION METHODS FOR ESTIMATING PROPERTIES OF POLYMERS THE GROUP-CONTRIBUTION PRINCIPLE SOLUBILITY PARAMETERS) AND SOME APPLICATIONS (DENSITY. The GC method is based on the assumption that the properties of molecules can be estimated using additive rules from the v alues of the corresponding groups the y are composed of..1.

28 and 1. The concept of free-volume (FV) is rather loose b ut still very important. In Equation 4. i. for selecting suitable compounds having a required set of properties. which is obtained from the group increments of Bondi and is tabulated for almost all existing groups in the UNIFAC tables. For example.35V w. Elbro 8 showed. the GC principle is e xtensively used for property calculations for specific molecules and also. even two different methods for the same property may emplo y different definitions fo the groups. NRTL.2. The exact definition of groups may be di ferent from method to method. The UNIFAC-FV Model The original UNIF AC model does not account for the free-v olume differences between solv ents and polymers.e. with the e xception of w ater and polydimeth yl siloxane (PDMS). For example. The latter technique is called computer aided product design. The latter problem will be further discussed later . In some cases. and Bondi mentions that for many organic molecules it can be expected that the ratio V*/Vw should be between 1. the division of butadiene rubber according to the van Krevelen and GCVOL methods for density is dif ferent.2).The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 77 than one OH and COOH groups.10 suggested using V* = Vw . The most well-known is the UNIF AC-FV model by Oishi and Prausnitz 12: . and for the same GC method.2) originally proposed by Bondi 9 and later adopted by Elbro et al. 10 and others 11 in the Entropic-FV model. Empirical modified UNI AC v ersions (de veloped in L yngby and Dortmund) with e xponential segment fractions are also inadequate for polymer solutions. But what is actually the hard-core volume? This also is rather difficult to determine. and most of them emplo y the energetic (residual) term of UNIFAC. consequently . it highly underestimates the solv ent activities in polymer solutions. All of these approaches include the FV effects. and UNIFAC models. Ho wever.9 later) pro vides a first approximation fo polymer solutions. 2. UNIFAC-FV) The classical Flory–Huggins model (See Equation 4. Both the combinatorial and the energetic terms need improvement via inclusion of free-volume effects and nonrandom local-composition terms such as those of the UNIQ UAC.. GC FREE-VOLUME-BASED MODELS The Free-Volume Concept FOR POLYMERS (ENTROPIC-FV. Elbr 8. that the FV percentages of sol ents (40–50%) are greater than those of polymers (30–40%). One of the simplest and most successful equations is: V f = V − V * = V − Vw (4. Various modifications — xtensions of the classical UNIF AC approach to polymers — ha ve been proposed. Many mathematical expressions have been proposed for the FV. using a simple definition for the FV (Equation 4. the y systematically o verestimate the solvent activities. In other cases. free-volume is simply the “empty” volume available to the molecule when the molecules’ own (hard-core or closed-packed V*) volume is subtracted.43. for h ydroxyl-acids or for alkanolamines). equal to the v an der Waals volume (V w). most agree today that V* > Vw due to the closed packed structure of molecules. Other investigators interpreted the hard-core volume somewhat differently. which are neglected in the UNIFAC combinatorial term. despite these problems. the closed-packed cubic structure suggests that V* = 1. a particular molecule can be di vided into groups in tw o different ways that may yield different results. in the r verse w ay.

In the UNIFAC-FV model as suggested by Oishi and Prausnitz 12 the parameters ci (3ci is the number of e xternal de grees of freedom) and b are set to constant v alues for all polymers and solvents ( ci = 1. fv ∑ x (V − V j j j xi (Vi − Vwi ) wj (4.w (4. which can be readily e xtended to multicomponent systems and liquid-liquid equilibria.5) ( ) In Equation 4.W are all expressed in cm3/mol.5. and it is based on the Flory equation of state: /3 ⎡ v1 i −1 ⎢ = 3ci ln ⎢ 1/ 3 ⎢ vm − 1 ⎣ ln γ fv i ( ( ) ⎤⎥ − c ⎡⎢⎛ v ) ⎥⎥⎦ ⎢⎣⎜⎝ v i i m −1 ⎞⎛ 1 ⎞ ⎤ − 1⎟ ⎜ 1 − 1/ 3 ⎟ ⎥ ⎠⎝ v1 ⎠ ⎥ ⎦ (4.13 The UNIFAC-FV model was originally developed for solvent activities in polymers and does not gi ve satisfactory results for polymer activities. the v olumes Vi and the van der Waals volumes Vi. 11: ln γ i = ln γ icomb− fv + ln γ ires ln γ icomb− fv = ln ϕ ifv = ϕ ifv ϕ fv +1− i xi xi = ∑x V j j xiVi . i.3) The combinatorial (comb) and residual (res) terms are tak en from original UNIF AC.4) where the reduced v olumes are defined as vi = Vi bVi.1 and b = 1.6) ) ln γ ires → UNIFAC . thus.e. 10 and Kontogeorgis et al.w + x2V2. Some researchers ha ve suggested that.. fv j . The performance of UNIF AC-FV is rather satisf actory. estimated based on vapor-liquid equilibria data for low molecular weight compounds. for a lar ge v ariety of polymer -solvent systems.w x1V1 + x2V2 vm = b x1V1. as sho wn by man y in vestigators. it has not been applied to polymer-solvent LLE. better agreement is obtained when these parameters ( ci and b) are fitted to xperimental data. is the so-called Entropic-FV model proposed by Elbro et al. in some cases. The Entropic Model A similar to UNIF AC-FV b ut some what simpler approach.28). An additional term is added for the free-volume effects.78 Hansen Solubility Parameters: A User’s Handbook fv ln γ i = ln γ comb + ln γ res i i + ln γ i (4.1 All the ener getic parameters in the residual term are the same as in original UNIF AC. The FV term used in UNIFAC-FV has a theoretical origin. No parameter reestimation is performed.

as mentioned. which yields better results for athermal polymer solutions.6. focusing especially on its combinatorial-FV term.10 The first term in Equation 4. using a suitable form of the generalized v an der Waals partition function.g.e. whereas the latter term is a re gular solution theory-type term.7) This parameter table has been developed using the combinatorial term of the original UNIFAC model (which. these can be estimated with the GC methods mentioned previously. Kouskoumvekaki et al. GCV OL. The residual term of Entropic-FV is taken by the so-called new or linear UNIFAC model. The same group parameters are used in linear -UNIFAC and Entropic-FV.2 is employed. especially the former. copolymers. i. free-v olume fractions are used.8 is the same as in Entropic-FV with Vw = b. If not a vailable. combinatorial and free-volume effects are combined into a single e xpression.8) ∑ x (V − b ) j j j j xi (Vi − bi ) δi = ai Vi ϕi is the v olume fraction as defined later in Equation 4.6 with V* = 1. Both UNIFAC-FV and Entropic-FV require as input the v olumes of solvents and polymers (at the temperatures of interest). are rather sensiti ve to the density v alues employed. The combinatorial term of Equation 4.. LLE. mixed solvents. the Entropic-FV model has been deri ved from the van der Waals partition RT . As already mentioned. compared to the assumption V* = Vw adopted in the original Entropic-FV model. the free-v olume definition g ven by Equation 4.15 For example. which uses a linear temperature dependent parameter table: 14 amn = amn.The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 79 As can been seen from Equation 4. It is considered one of the state-of-the-art models in the field . The combinatorial–FV expression of the Entropic-FV model is deri ved from statistical mechanics..2V w.6 is v ery similar to that of Flory–Huggins. In this w ay. Various efforts in impro ving Entropic-FV ha ve been published. However. e.6. Entropic-FV has been extensively applied to various types of phase equilibria (VLE.2 T − To ( ) (4. no parameter reestimation has been performed. and SLE) of polymer–solv ent and polymer–polymer (blends) systems as well as solutions including dendrimers. they are reviewed elsewhere. As with UNIFAC-FV. does not account for free-v olume effects). The similarity of the model with the v an der Waals equation of state P = -----------V – b V2 becomes apparent if the latter is written (when the classical Van der Waals one fluid mixing an classical combining rules are used) as an acti vity coefficient model ln γ i = ln γ icomb− fv + ln γ ires ⎛ ϕ fv 2 ϕ fv ⎞ ⎛ V ⎞ = ⎜ ln i + 1 − i ⎟ + ⎜ i ( δ i − δ j ) ϕ 2 j⎟ ⎠ xi ⎠ ⎝ RT ⎝ xi ϕ ifv = (4. This is in agreement with what is stated pre viously.1 + amn. instead of v olume or se gment fractions.16 suggested using Equation 4. a co volume being higher than the v an der Waals volume. and paint-related polymers. Activity coef ficient calculations with UNI AC-FV and Entropic-FV .– ----afunction.

the FH parameter is typically not a constant and should be estimated from e xperimental data.35 + V1 δ1 − δ 2 RT ( ) 2 (4. F or practical applications. the FH model cannot be used for predictions unless a predicti ve scheme for the FH parameter is a vailable. Such a predicti ve scheme can be based on a solubility parameter . we can relate the FH parameter with the solubility parameters. Usually it v aries with both temperature and concentration. the so-called FH interaction parameter χ12.A. whereas the second rather empirical (v an Laar -type) ener getic term includes the only adjustable parameter of the model. The parameter r is the ratio of the polymer volume to the solvent volume V2/V1 (approximately equal to the de gree of polymerization).c.b.80 Hansen Solubility Parameters: A User’s Handbook THE FLORY–HUGGINS MODEL AND THE REGULAR SOLUTION THEORY The Flory–Huggins (FH) model is the most well-kno wn approach for polymer solutions and can. but in some cases a reasonable v alue of the FH parameter can be obtained. using the follo wing equation: χ12 = χs + χ h = 0. The FH theory can be extended to multicomponent systems (see Appendix 4. it should be mentioned that the adjustable parameters of such equations (a. Accurate representation of miscibility curv es with the FH model is possible using rather comple x equations for the temperature and the concentration-dependence of the FH-parameters: ΔG = RT ∑ x ln ϕ + g i i i 12 ϕ1ϕ 2 ⎞ ⎛ c g12 = B(ϕ)C (T ) = a + bϕ 2 ⎜ 1 + + dT + e ln T ⎟ T ⎝ ⎠ ( ) (4.e) ha ve no apparent ph ysical significance.1).11) Although we could use equations like Equation 4. Due to the similarity of the v an Laar term with the re gular solution theory (see Equation 4.d. th y cannot be generalized and are specific for each polyme -solvent system. which renders the FH model useful basically for correlating e xperimental data.A. This is an approximate approach.8).11. unfortunately.9 is due to combinatorial e fects and is deri ved from the lattice theory.9 with a composition-independent Flory–Huggins interaction paramete . Even in this w ay. it often suffices to use Equation 4. Moreover. either the Hildebrand or the Hansen.9 is that the Flory–Huggins interaction parameter ( χ12) is independent of composition (see Appendix 4. and the v olume fraction is defined as ϕi = xiVi xiVi + x jV j (4.10) The first term of Equation 4. for binary systems.12) .9) The most important assumption in deri ving Equation 4. be e xpressed under some assumptions as follo ws for the solv ent acti vity coefficient ln γ 1 = ln ϕ1 ϕ + 1 − 1 + χ12 ϕ 2 2 x1 x1 ϕ ⎛ 1⎞ 2 = ln 1 + ⎜ 1 − ⎟ ϕ 2 + χ12 ϕ 2 x1 ⎝ r⎠ (4.1) but (at least) one χ12-value is required per binary .

9).17 In the intermediate re gion (between 6 and 8). the empirical f actor 0. Using Hansen solubility parameters (HSP). χ12 ≤ 0. A chemical (1) will be good solvent for a specific polymer (2). If the polymer and the solvent have very different polar and hydrogen bonding degrees: 4 δd1 − δd 2 ( ) + (δ 2 p1 − δ p2 ) + (δ 2 h1 − δh2 ) 2 ≤R (4. These deficiencies become vident when comparing experimental data for athermal polymer and other asymmetric solutions to the results obtained with the FH model. Moreover.35 is added for correcting for the deficiencies of the FH combinatorial and residual terms. and these rules of thumb are summarized here.The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 81 Equation 4.Values of the infinite dilution actvity coefficient ab ve 8 indicate nonsolvency.8 and Equation 4.12 for estimating the FH parameter. in other words.5 (the lo wer the Flory–Huggins parameter v alue.14) where R is the Hansen solubility parameter sphere radius. the tw o compounds will be miscible if one (or more) of the follo wing “rules of thumb” are v alid: 1. or in other ords. and although satisf actory results are obtained in some cases. Rules of Thumb and Solvent Selection Using the Flory–Huggins Model and Solubility Parameters The FH model and the solubility parameters of fer v arious alternati ve approaches for solv ent selection for polymers. This latter rule of thumb requires some further e xplanations. without the empirical 0.12. the greater the miscibility or . 3.5 indicate nonsolvency.8 ⎜ 3 ⎟ ⎝ cm ⎠ 1/ 2 (4. it is dif ficult to conclude i the specific chemical is a sol ent or a nonsolv ent. A consistent underestimation of the acti vity coef ficient data i observed. the better a solvent is a specific chemical). is deri ved from the re gular solution theory (compare Equation 4. corrects in an empirical w ay for these free-v olume ef fects.Values much above 0.35 f actor. Ω1 ≤ 6 (the lower the infinite dilution act vity coefficient of the sol ent.13) 2. If the polymer and the solvent have “similar polar and hydrogen bonding degrees:” δ1 − δ 2 ⎛ cal ⎞ ≤ 1. we cannot generally recommend using Equation 4. The term. ∞ 4. for many nonpolar systems with compounds having similar solubility parameters. the greater the solvency of a chemical). Using Hildebrand solubility parameters. The constant 0. which contains the 0. The weight-based acti vity coefficient at infinite dilution is defined γ i∞ = lim xi →0 γ i ⎛xγ ⎞ ∞ M2 Ωi∞ = lim wi→0 ⎜ i i ⎟ = γ 1 M1 ⎝ wi ⎠ .35 should be dropped.35 term. which is often attributed to the inability of the FH model to account for the free-v olume differences between polymers and solv ents or between compounds dif fering significantly in siz such as n-alkanes with v ery dif ferent chain lengths. Ho wever.

dif ferent w ays of expressing the composition fraction ϕi: 1.15) This model is hereafter abbre viated as FH/Ha(nsen) or FH/HSP .1. Table 4. ACTIVITY COEFFICIENTS MODELS USING THE HSP FLORY–HUGGINS MODELS USING HILDEBRAND PARAMETERS (HSP) AND HANSEN SOLUBILITY Several of the problems of the Flory–Huggins model are associated with the dif ficulties in predictin the FH interaction parameter and the f act that this parameter depends on both temperature and concentration. Segment fractions.10) together with α = 0.. the latter being defined via Equation 4. PEMA. A thermodynamically more correct method is to calculate the acti vity–concentration diagram with a thermodynamic model lik e Entropic-FV or UNIFAC-FV. In all cases.2 pro vides results for se veral polymer solutions with the FH/Hansen model (FHHa. 358 data points ha ve been considered for solutions containing acrylates and acetates (PBMA. i. Lindvig et al. 19 ha ve tested three dif ferent combinatorial e xpressions. The best results are obtained when the v olume-based combinatorial term is used (Equation 4.e.25 ( δ p1 − δ p 2 ) + 0. . these can be estimated by thermodynamic models lik e the ones mentioned abo ve (UNIF AC-FV. as can be seen in Figure 4. Based on v olume fractions (Equation 4. In total.18 otherwise. and Flory–Huggins).82 Hansen Solubility Parameters: A User’s Handbook (The latter part of the equation is v alid for a binary solv ent(1)-polymer(2) mixture.10) 2. FV fractions (Equation 4. A minimum does exist for the different types of solutions.6 (see also Figure 4. PMMA. the maximum indicates phase split. PV Ac). whereas a monotonic increase of acti vity with concentration indicates a single liquid phase (homogeneous solution). Recently. However. the minima for the nonpolar and h ydrogen bonding solvents are v ery close.15) using the volume-based combinatorial with both the optimum parameter and α = 1. the results depend not only on the accuracy of the model b ut also on the reliability of the rule of thumb. The expression for the solv ent activity coefficient in a binary sol ent-polymer solution is: ln γ 1 = ln ϕ1 ϕ + 1 − 1 + χ12 ϕ 2 2 x1 x1 2 V 2 2 χ12 = α 1 ⎡( δ d1 − δ d 2 ) + 0.25 ( δ h1 − δ h 2 ) ⎤ ⎦ RT ⎣ (4.1). the results are better when the optimum v alue is used than when α is set equal to one or when the term with the Hansen solubility parameters is ignored ( α = 0). 19 proposed an e xtension of the Flory–Huggins equation using the Hansen solubility parameters for estimating activity coefficients of compl x polymer solutions. Lindvig et al. Thermodynamic models often perform better for this type of calculation rather than for predicting full LLE phase diagrams. Entropic-FV . whereas there is little sensiti vity to the parameter for polar solv ents. These are shown for the v arious combinatorial terms in Table 4. ut with v an der Waals The universal parameter α has been fitted in each case to a la ge number of polymer -solvent VLE data. which in turns depends on the assumptions of the Flory–Huggins approach. In particular . This means that a uni versal α value can be established. Equation 4.) ∞ Extensive collections of experimental Ω1 data are available.10 volumes used instead of v olumes.1.6) 3.

. FH/Hansen performs in all cases better than FV/Hildebrand which is the best possible model implementing the Hildebrand parameters. 24% (polar). This may be e xpected as entropic ef fects are not accounted for and compensation is required by a higher parameter value. the pre viously described Entropic-FV (EFV) and UNIF AC-FV (UFV) acti vity coef ficients and an equation of state.3 pro vides an o verall comparison of the FH/Hansen model using all three choices for the combinatorial term and the three GC models mentioned abo ve. 20 The presentation of the results is organized into three categories according to the nature of the solv ents (nonpolar. It seems that the HSP account not only for ener getic effects but also for some residual-free v olume contributions. The FV/Hildebrand model is gi ven by Equation 4.L. 2002. polar. Th. Michelsen. Finally. th GC–Flory (GCFl) model by Bogdanic and Fredenslund.1 Influence of the α-parameter on the performance of Equation 4.3 0.8 with solubility parameters being the total Hildebrand parameters. .8 0.. and hydrogen bonding).5 0. Moreover. For the FH/Hansen model: in all cases better results are obtained when the optimum αparameter is used compared to α = 1.2 0. (From Lindvig. Reprinted with permission. and K ontogeorgis. It can be concluded that: 1. 247. 203.) Results are also sho wn with three dif ferent group-contribution models.The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 83 100 90 80 Average absolute percentage deviation 70 60 50 40 30 20 10 0 Combinatorial part based on volume fractions PBMA PEMA PMA PMMA PVAc Average 0 0.M. G. 2. M. Results are sho wn for all systems considered in the database for the estimation of the α-parameter as well as tw o commercial epoxy resins for which acti vity coefficient data are vailable. and 48% (h ydrogen bonding).4 0.21 The a verage de viations with FV/Hildebrand are: 36% (nonpolar).1 0.6 0. Table 4.7 Value of the correction constant 0.15 for all polymer solutions in the database when the Flory–Huggins part of the model is based on v olume fractions.. The α-parameter is higher when v olume and especially se gment fractions are used in the combinatorial term. Fluid Phase Equilibria.9 1 FIGURE 4.

55 20 37 40 0. Equation 4. A difficulty with the FH/Hansen model is that arious values of HSP are sometimes reported for the same polymers in the literature. for the tw o epoxy resins.. and Kontogeorgis.) Fraction Volume αopt % AAD (αopt) % AAD (α = 0) % AAD (α = 1) αopt % AAD (αopt) % AAD (α = 0) % AAD (α = 1) αopt % AAD (αopt) % AAD (α = 0) % AAD (α = 1) Non-Polar 0. An advantage of the FH/Hansen model compared to the GC methods is that the xact e knowledge of the structure of the polymers is not needed. 22 Use of different main groups in models lik e EFV or UFV will ha ve different results.L. The de viations are within the reported e xperimental uncertainty for infinite dilutio activity coefficients.60 22 41 40 0.25 28 31 76 Polar 1. 21 COOCH 2. 3. 4.84 Hansen Solubility Parameters: A User’s Handbook TABLE 4.60 25 54 53 0. Reprinted with permission.B.00 34 48 34 0. which is typically between 10–20% The FH/Hansen Model vs..1 (From Lindvig.05 20 20 33 H. H. Results are sho wn for all systems a vailable in the database (denoted as “total”) and for the dif ferent types of solv ents. G.85 19 47 22 0. 2002. and other chemicals. The FH/Hansen model is as accurate as the other group-contrib ution models for the systems used in the database for estimating the α-parameter. the FH/Hansen method does not suffer from the often problematic assignment of groups in the GC methods.B. Moreover. 5.80 25 51 29 0. indicates h ydrogen bonding solvents.M. for e xample.12 and COO. The most important similarities are that the y both need as input the density of polymer and solv ents (though not GC–Flory. Michelsen. 203.40 22 40 87 Total 0. PBMA: CCOO. 0. The only information required is the HSP and the densities.1 Optimum Values of the α-Parameter with the FH/HSP Model. Th.. the GC Methods There are similarities and dif ferences between the FH/Hansen and the GC methods. we can mention that three dif ferent definitions h ve been proposed for the “acetate” group of.75 28 63 40 0. As an example. Fluid Phase Equilibria. especially the well-kno wn UNIFAC-FV. It is particularly better than the GC models for h ydrogen bonding solv ents. which are a vailable for many polymers. 19 . thus their application is limited to lo w pressures. which is an equation of state) and that the y are formulated as acti vity coefficien models. or can be readily estimated. M.00 23 31 23 1. 247. and it is not always apparent beforehand which group should be chosen. The FH/Hansen model is more accurate than the other models. solvents.30 26 31 71 Segment Free volume Note: The a verage absolute de viations (AAD) are pro vided using these optimum values as well as when the α-parameter is equal to zero or one.

21 have performed an e valuation of various models for solv ent selection for paints.6 FHHa 18 44 24 38 14 24 α = 0.6 FHHa 35 27 22 35 17 25 α = 0. The results of these e valuations are summarized in Table 4. These included Entropic-FV. whereas selected results are sho wn in Table 4. APPLICATIONS Solvent Selection for Paints (Activity Coefficients at Infinite Dilution) Lindvig et al. GC-Flory as well as the the FH/Hansen model presented previously and the classical Hansen method (Equation 4. For the three group-contrib ution models.2 Average Absolute Percentage Deviations (% AAD) between Calculated and Experimental Activity Coefficients for Various Groups of Solvents at Infinite Dilution in Polymers often Used in Paints and Coatings Applications Non-Polar Solvents Polymer PBMA PEMA PMA PMMA PVAc Total Nsys 11 4 5 5 8 33 NDP 60 5 17 17 103 202 EFV 22 47 45 40 20 31 UFV 22 49 46 25 19 29 GCFl 11 36 28 26 17 20 α=1 FHHa 65 86 18 19 53 51 α=1 FHHa 28 22 37 25 17 25 α=1 FHHa 65 86 18 19 53 53 α = 0. respectively. the number of systems and datapoints. the solv ent selection is based on the rule of thumb: ∞ Ω1 ≤ 6 : good solv ent ∞ Ω1 ≥ 8 : poor solv ent (nonsolvent) .14).6 FHHa 18 44 24 38 14 25 Polar Solvents Polymer PBMA PEMA PMA PMMA PVAc Total Nsys 4 2 2 3 4 15 NDP 12 6 6 10 30 64 EFV 9 10 35 25 36 23 UFV 13 11 36 25 25 21 GCFl 11 11 25 23 24 19 Hydrogen Bonding Solvents Polymer PBMA PEMA PMA PMMA PVAc Total Nsys 6 3 2 3 2 16 NDP 40 10 4 21 17 92 EFV 65 46 16 108 15 57 UFV 87 47 34 117 63 26 GCFl 24 21 9 12 15 18 Note: Nsys and NDP are.4a . UNIFAC-FV.The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 85 TABLE 4.4b. 19.

Thermodynamics of paint-related systems with engineering models.e. Th. TABLE 4.14) FV/Hildebrand (Equation 4. et al. 2001. 247. AIChE J.17) Entropic-FV UNIFAC-FV GC-Flory Correct Answers 102 99 86 91 78 72 Incorrect Answers 20 23 23 19 21 26 No Answer — — 13 19 17 14 No Calculation 7 7 7 0 13 17 Source: Adapted from Lindvig. The FH/Hansen and FV/Hildebrand models are summarized as follo ws: . 47(11). For the FH/Hansen and FV/Hildebrand models. Th. All the FH/Hansen results shown in this section are with the “best” combination. et al. 2002. The last tw o columns sho w predictions for tw o epoxy resins.6.86 Hansen Solubility Parameters: A User’s Handbook TABLE 4.. Fluid Phase Equilibria. 247. The density of the epoxies is estimated using the GCVOL method.3 Average Absolute Percentage Deviations (% AAD) between Experimental and Calculated Activity Coefficients for PaintRelated Polymer Solutions Using the Flory-Huggins/Hansen Method and Three Group Contribution Models Model FH/Hansen volume fractions FH/Hansen segment fractions FH/Hansen FV fractions Entropic-FV UNIFAC-FV GC-Flory %AAD (systems in database) 22 25 26 35 39 18 %AAD Araldit 488 31 — — 34 119 29 %AAD Eponol-55 28 — — 30 62 37 Note: The second column represents the systems used for optimization of the uni versal parameter (solutions containing acrylates and acetates). using the v olume fraction-based combinatorial and α = 0. the solvent selection is based on whether the value of the FH parameter is belo w or abo ve 0. Lindvig.4A Validity of the Solubility Answers Obtained from Five Methods for Solvent Screening in Various Polymer-Solvent Systems Model FH/HSP (Equation 4.5 as e xplained pre viously. i. Source: Adapted from Lindvig. 203..16) Original Hansen (Equation 4. 2573. Th. et al. 203... No answer is obtained when the infinite dilution act vity coefficient alue is between 6 and 8. Fluid Phase Equilibria.. 2002.

7 ( – ) 29.5 (NS) 8.18 (S) 0.9 (S) 8.27 (S) 1.16.8 (NS) 8.17) χ12 = V1 ( δ1 − δ 2 ) 2 RT . fv ∑ j xi (Vi − Vwi ) x j (V j − Vwj ) (4.4B Prediction of the Solubility for Characteristic Polymer-Solvent Systems Using Various Rules of Thumb and Models for Solvent Selection System PBMA/nC10 PBMA/xylene PBMA/CHCl3 PBMA/acetone PBMA/ethyl acetate PBMA/ethanol PMMA/acetone PMMA/ethyl acetate PMMA/butanol PEMA/MEK PEMA/diethyl ether PEMA/nitropropane PVAc/hexane PVAc/methanol PVAc/ethanol PVAc/nitromethane PVAc/THF Experiment NS S S S S NS S S NS S S NS NS S NS S S EFV 6.16) xiVi xiVi + x jV j ∑ j xiVi .4 (NS) 11. S = good solv ent.6 (S) FH/HSP 1. FH–Hansen (FH/HSP) ln γ 1 = ln χ12 = 0.2 (NS) 10. et al. .36 (S) 0.14 (S) 0. NS = non solv ent. 2002.2 (NS) 6.20 (NS) 0.7 (NS) 7.25 (δ h1 − δ h 2 )2 ⎤ ⎦ RT ⎣ (4.4 (S) 38.5 ( – ) 2.67 (NS) 0.4 (NS) 14.0 (NS) 6.63 (NS) 0.57 (NS) 0. fv x jV j .8 (S) 5.9 (NS) 3.8 (S) 4.9 (S) 0.01 (NS) 0.4 (NS) UFV 6.43 (S) 0.2 (NS) 3.1 (NS) 6.2 (S) 0.41 (S) 0..The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 87 TABLE 4.11 (S) 1.25 (δ p1 − δ p 2 ) + 0.71 (NS) 0.6 ϕi = FV–Hildebrand ln γ 1 = ln ϕ ifv = ϕ ifv ϕ fv + 1 − i + χ12 ϕ 2 2 xi xi = ϕ1 ϕ + 1 − 1 + χ12 ϕ 2 2 x1 x1 2 V1 ⎡ (δ d1 − δ d 2 )2 + 0. Th.7 (NS) 18. 247.3 (NS) 16.1 (NS) 5.= no answer according to the rule of thumb .6 (S) 9. 203.4 (NS) 38.6 (NS) 19. Source: Adapted from Lindvig.09 (S) 0.6 (– ) 26.09 (NS) 0.3 (S) 1.9 (NS) 15.05 (S) Note: The values in the table for EFV and UFV indicate infinite dilution act vity coefficient while those of FH/HSP are the Flory-Huggins parameters estimated based on Equation 4.5 (S) 38.1 ( – ) 3.7 ( – ) 31.1 (NS) 14. Fluid Phase Equilibria.6 ( – ) 1.

the FV/Hildebrand method 3. which has tw o. In addition to this distinction (purely predicti ve and molecular models). For most polymers (PBMA. TABLE 4. All interaction parameters ha ve been fitted to binary xperimental data as e xplained by Lindvig et al. Most models ha ve problems with nitrocompounds and PEMA 5. Purely predicti ve GC models for which calculations can be made without re gressing parameters from binary data for the systems considered (EFV . GC-Flory. and the accurac y of the data may in some cases be doubtful.5 Presentation of the Characteristics of the Various Models Tested for Mixed Solvent–Polymer Phase Equilibria Model SAFT EFV/UQ FH Pa-Ve EFV/UN UFV GCFl GCLF FH/Ha Correlative X X X X X X X X X Fully Predictive Activity Coefficient Model X X X X X X X X X X X X X Partially Equation of State X Partially Group-Contribution . on a verage. Mixed Solvent–Polymer Phase Equilibria Lindvig et al. and GCLF). The FH/Hansen method is. FH/Ha. i. The behavior of the v arious models for PVAC is rather mix ed and peculiar . All calculations are based on e xisting parameters (typically group-based ones). 2. The FH/Hansen method is better than the “best” combination based on the Hildebrand parameters.6 divided according to the experimental source. where only the FH/Hansen method performs well 4. each model has dif ferent strengths and weaknesses.. and their v arious characteristics are summarized in Table 4. as good as the original method of Hansen and the three GC methods 2. the models tested for multicomponent systems ha ve similarities and dif ferences. UFV. PMMA. Two fundamentally different modelling approaches ha ve been tested: 1. Only a few experimental data are a vailable for such multicomponent systems. 23 extended the applicability of v arious models to mix ed solvent-polymer VLE and typical results are presented in Table 4. and unlik e the other polymers.5.23 All correlative models contain a single binary parameter e xcept EFV/UNIQUAC. This is a more time-consuming approach than the former one b ut is still a predicti ve one for ternary mixtures in the sense that no multicomponent data are used for the re gression of the model parameters. PEMA) the GC methods have problems for ketonecontaining solutions.88 Hansen Solubility Parameters: A User’s Handbook The basic conclusions are: 1.e. These models are the f astest and simplest tools for thermodynamic calculations. Molecular models using binary molecular interaction parameters estimated from e xperimental data for the corresponding binary systems.

PS-benzene-toluene at 308 K. 38: 388. PV Ac-acetone-methanol at 303 K.24 SAFT by Chapman and co workers. PMMA-butanone-toluene at 308 K. and FH/Hansen) and the GC-Flory equation of state. Sys. 1 2 3 4 Variable P y P y P y P y SAFT 11 4 4 5 – – – – EFV/U 6 3 2 2 – – – – FH 6 3 14 5 – – – – Pa-Ve 6 3 8 1 – – – – EFV/U 2 3 2 2 3 3 4 18 UFV 1 3 2 2 3 3 4 18 GCFI 21 3 – – 12 2 13 19 GCLF 8 3 2 4 5 4 9 19 FHHa 2 3 11 5 18 4 5 15 Note: 1.25 and P anayiotou-Vera. Source of experimental data: Katayama et al. 35: 1012]. PMMA-benzene-toluene at 308 K. 23 Besides the acti vity coefficient models mentioned pr viously (Entropic-FV. PS-toluene-c yclohexane at 303 K. three adv anced equations of state are considered: the GCLF by Lee and Danner . 4. No. No.15 K. 194–197: 1067–1075]. 1 Variable P y SAFT 14 17 EFV/U 16 17 FH 11 16 Pa-Ve 4 11 EFV/U 17 17 UFV 19 13 GCFI 93 2 GCLF 5 18 FHHa 52 14 Source of experimental data: Tanbonliong and Prausnitz (1997).26 The basic conclusions are: . PS-chloroform-carbon tetrachloride at 323. UNIFAC-FV. No. (2002). Source of experimental data: Liu et al. 1974. Kagaku Kogaku. 1971. PS-toluene-ethylbenzene at 303 K.The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 89 TABLE 4. 2002.6 Average Absolute Logarithmic Percentage Deviations between Experimental and Predicted Equilibrium Pressures and Average Absolute Deviation (X100) between Calculated and Experimental Vapor-Phase Compositions (Mole Fractions) for Various Ternary Polymer-Mixed Solvent Systems Sys. PV Ac-acetone-ethyl acetate at 303 K. 4. 1 2 3 4 Variable P y P y P y P y SAFT 13 28 7 14 – – – – EFV/U 16 30 4 13 – – – – FH 15 29 12 14 – – – – Pa-Ve 16 29 7 14 – – – – EFV/U 16 32 8 8 30 17 72 4 UFV 36 32 14 8 37 18 57 5 GCFI 16 28 149 6 16 8 52 4 GCLF 15 27 14 23 22 5 32 4 FHHa 17 21 7 13 50 5 20 4 Note: 1. Adapted from Lindvig et al. (1971) [Kag aku K ogaku. Polymer. 2. Sys. Fluid Phase Equilibria. 38: 5775. 3. 3. PMMA-b utanone-acetone at 308 K. Matsumara and Katayama (1974). 2.

theoretically-based equations of state lik e SAFT and the Group Contrib ution Lattice Fluid. The predictive group contribution methods are less successful for the prediction of liquid–liquid equilibria as their parameters are based on VLE. 4. have been applied mainly to or ganic polymer solutions of rather simple structures andVLE/solvent selection studies.6. Such methods require accurate v alues of the densities and. 5. for example. The methods can be. For 9 out of the 13 systems considered. for Entropic-FV. However. This is a surprising result as it is e xpected that the molecular information in a model should lead to a better representation of phase equilibria for multicomponent mixtures. In many cases the results with the various models are closer to each other than to the e xperimental data. b ut the o verall differences are minor . Thorlaksen et al. It may be that some experimental data for such multicomponent mixtures are of lo w accuracy.27 As such models include the liquid volume as input parameter. Moreover. solv ent selection for paints. GC-Flory. UNIFAC-FV. and we belie ve that the y fall within the e xperimental uncertainties of the data considered. and solv ent selection. and this has been indeed done. and VLE for mix ed solvent–polymer systems. and SAFT. However. Future developments can include complex structures such as dendrimers (where EFV and UFV already ha ve been applied 29). This investigation does not point to a “clear winner” among the models. solvent selection can be based on the use of infinite dilution act vity coefficient for which these models are quite successful.16) includes a single universal parameter that has been re gressed to e xperimental data for man y polymer solv ent solutions. star -like. although some comple x paint-related polymers ha ve been considered as well. . and GC-Flory) in predicting infinite dilution act vity coefficients including com plex epoxy polymers. and more data are required for further in vestigations. both the correlative and the predicti ve. 3. successful results have been reported even for high pressure LLE. An alternative equally successful approach is offered by the combination of the Flory–Huggins model with the Hansen solubility parameters (HSP). So far most acti vity coefficient models.90 Hansen Solubility Parameters: A User’s Handbook 1. This chapter has been limited to v apor–liquid equlibria (both at finite concentrations and infini dilution).15 Better results are obtained when a molecular local composition model is used as. e xtended to polymer–solvent LLE. this preliminary study pro vides confi dence for use of the FH/Hansen approach. good predictions of both the equilibrium pressures and the vapor phase compositions are obtained by all models. 28 have recently combined the Entropic-FV term with Hildebrand's re gular solution theory and de veloped a model for estimating g as solubilities in elastomers. A similar approach can be adopted for the FH/HSP model presented here. The models perform sometimes dif ferently for isolated systems. The correlative models do not of fer any improvements over the group-contribution models. are based on the a vailability of group parameters in the UNIF AC tables. which contain corrections for the FV effects. Models like Entropic-FV and UNIF AC-FV can be used for such calculations and are sho wn to satisf actorily predict the solvent activities and vapor–liquid equilibria for binary and ternary polymer solutions. mixed solvents. moreo ver. h yperbranched polymers as well as v arious copolymers. in the Entropic-FV/UNIQUAC model as shown by Pappa et al. GCLF. CONCLUSIONS AND FUTURE CHALLENGES Many successful calculations in polymer thermodynamics can be carried out using simple groupcontribution methods based on UNIFAC. FH/HSP is as successful for mix ed solvent–polymer phase equilibria as comple x. for e xample. Some results are summarized in the recent literature. in principle. The FH/HSP (Equation 4. The FH/HSP model is sho wn to be as successful as the state of the art GC models (EntropicFV. The combinatorial term that gi ves the best results is based on v olume fractions. including FH/HS . The FH/Hansen is as successful as the GC methods and other more comple x models. 2.

and cross-linking./exp. inorganic polymers and polyelectrolytes. PVAC Res S SAFT SLE UCST U-FV UNIFAC UQ vdW vdW1f VLE VOC Average percentage absolute de viation Combinatorial Combinatorial-free volume Critical solution temperature Experimental Entropic-FV (Same as EFV) Entropic-FV (using UNIF AC for the residual term) Entropic-FV using UNIQ UAC as the residual term Flory–Huggins (model/equation/interaction parameter) The FH model using the Hansen solubility parameters.The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 91 liquid–liquid and solid–liquid equilibria. Equation 4. LIST OF ABBREVIATIONS % AAD Comb Comb-FV CST Exper. including the ef fect of crystallinity. EFV EFV/UN EFV/UQ FH FH/HSP FV GC GC-F(lory) GCLF GCVOL HB HSP LCST LLE MW NS Pa-Ve PBMA PDMS PEMA PMMA Pred.16 Free-volume Group contribution (method/principle) Group contribution Flory equation of state Group contribution lattice flui Group contribution volume (method for estimating the density) Hydrogen bonding Hansen solubility parameters Lower critical solution temperature Liquid–liquid equilibria Molecular weight Nonsolvent/nonsoluble Panayiotou–Vera equation of state Polybutyl methacrylate Polydimethylsiloxane Polyethyl methacrylate Polymethyl methacrylate Predicted Polyvinyl acetate Residual Solvent/soluble Statistical associating fluid theor Solid–liquid equilibria Upper critical solution temperature UNIFAC-FV Universal functional activity coefficient UNIQUAC Van der Waals equation of state Van der Waals one fluid (mixing rules Vapor–liquid equilibria Volatile organic content .

5 Degree of freedom in Equation 4. 23): G E = G E.2) Hard core or close pack ed volume in Equation 4.10) Activity coefficient in Chapter Flory–Huggins interaction parameter Infinite dilution act vity coefficien APPENDIX 4.1 Total volume Free volume (Equation 4.5 Constant in Equation 4. b ut it w as soon modified also to account fo energetic interactions.1 Group value for given property in Equation 4.comb + G E.res G E.comb = RT G E.res = RT ∑ n ln φ x i i =1 N N i i =1 j =1 N i i ∑∑φ φ a j ij χij = 2 aij νi .4 and Equation 4.4 Property in Equation 4.15 Constant in Equation 4.2 van der Waals volume Solubility parameter (as in rest of handbook) Volume fraction (see also Equation 4.I: AN EXPRESSION OF THE FLORY-HUGGINS MODEL FOR MULTICOMPONENT MIXTURES The Flory–Huggins model w as originally de veloped as a model for the entrop y of mixing for mixtures containing molecules of dif ferent size.92 Hansen Solubility Parameters: A User’s Handbook SYMBOLS IN THIS CHAPTER G R R T a amn b ni r x v a b c F Fi V Vf V* VW δ φ γ χ12 ΩI∞ Gibbs energy Radius of Hansen solubility sphere Gas constant (in connection with T) Temperature Constant in van der Waals equation of state Coefficient defined in Equation 4 Constant in van der Waals equation of state Number of given groups of type “i” in molecule Ratio of polymer v olume to solv ent volume Mole fraction Reduced volumes in Equation 4. The model can be formulated in terms of the excess Gibbs energy as follows (Lindvig et al.

and Prausnitz. Ind. FL. Boca Raton. and Tassios. Process Des. and Rasmussen. P. New York. Elbro.W. Fredenslund.. Extension of the GCV OL method and application to some comple x compounds.. Chem. and Rasmussen.. Group contrib ution method for the prediction of liquid densities as a function of temperature for solv ents. Kalospiros. Bogdanic. Ind. in Handbook of Surface and Colloid Chemistry.. Fredenslund. J. Technical University of Denmark. Res. G.. Extension and re vision of the group contribution method GCVOL for the prediction of pure compound liquid densities . Eng.. Res. Elsevier.. Estimation of solv ent acti vities in polymer solutions using a groupcontribution method... Fried. 333. John Wiley & Sons. Phase Equilibria of Polymer Solutions — with Special Emphasis on Free Volumes. 2nd ed.. 5. A. 1641. 2003. N. Simple acti vity coefficient model for th prediction of solv ent activities in polymer solutions.C.M. Aa.. and Prausnitz. Fredenslund. Dev. H.L.S. 1994..R. J. Ind. Sci.N. Chem..M. Kontogeorgis. Eng. 1991.. Prediction of VLE for mixtures with co-polymers. Elbro.16. Van Krevelen.. Elbro.. Their correlation with chemical structure. Aa. and Fredenslund. I. Eng. 3. 324. D... 4707.S. P . Ed. Oishi. 4. Ind. 2. 7. 10. their numerical estimation and prediction from additive group contributions.. 1086–1098. 362.. 25(1). 408–412.S. 1968. 13.M. Macromolecules. Fredenslund. 1975. E. Ph.. Aa. .S.. Jiang. CRC Press. Jones. Eng... 1992... 95. 17(3). G. Res. 33. 1978. and polymers . Thermodynamics of polymer solutions.S. 32. Aa. Comput. although there is an interrelationship based on the simple equation sho wn above. oligomers. Res.P. 34. G. E. 8. H. 6.D thesis. K. Properties of polymers. Chem. the multicomponent equation reduces to the traditional FH residual term: res 2 ln γ 1 = χ12 ϕ 2 REFERENCES 1. and Yeh. D. Chem. 12. D.. and Gmehling. Kontogeorgis. the follo wing expression for the acti vity coefficient is obtained ln γ i = ln γ comb + ln γ res i i where the combinatorial term is gi ven by: ln γ icomb = ln and the residual term is: ϕi ϕ +1− i xi xi = 2 vi ln γ res i ∑ j =1 NC ϕ j aij − vi ∑∑ϕ ϕ a j k j =1 k =1 NC NC jk The abo ve formulation of the FH model is slightly dif ferent from the con ventionally used formulation using the Flory–Huggins interaction parameter (χ12)... M. Res. 2576. 1990. Liquids and Glasses. Chem. 30. 1990.. 9. A new simple equation for the prediction of solvent activities in polymer solutions. and Tassios. Ind.. 1992.P. Polym. 11.. Group contrib ution estimation of acti vity coeffi cients in nonideal liquid mixtures. 42(2).The Hansen Solubility Parameters (HSP) in Thermodynamic Models for Polymer Solutions 93 Using basic thermodynamics. Ihmels. Department of Chemical Engineering. Physical Properties of Molecular Crystals. Chem. J. 23.. H.. R. 1965. 1993. AIChE J. Eng. Bondi... Ind. 2003. Tsibanogiannis. chap. 2. Eng. J. For a binary mixture.S. Birdi. Aa. T.

7..L. 28. 2004.. 2573. Ind. 2001. Aa. 4848. 247. An improved lattice-fluid equation of state for pure component poly meric fluids. 20. Th.. Chem. and Gaani. . Michelsen... 2002.M.. et al. 1992. G. M. 1990.. Polymer Solution Handbook.M. G. G. and Eng. 1709–1721. 2001.. and Danner .P. 17. DIPPR 881 Project. Abildskov. G. Ind. 1988.H.G. B.. Liquid-liquid phase equilibrium in polymer -solvent systems: correlation and prediction of the polymer molecular weight and the pressure effect. B. Thermodynamics of paint-related systems with engineering models. Modeling of multicomponent v apor-liquid equilibria for polymer -solvent systems. Progress in Organic Coatings. and K ontogeorgis. 47(11). Thorlaksen. Kouskoumvekaki. and Tassios. 25. I. 837. 29. Sci.. Design Institute for Physical Property Data. and Vera. Fluid Phase Equilibria. 2002. in Computer Aided Property Estimation for Process and Product Design. Res.C. M. Eds. Polym.. and Kontogeorgis. Voutsas. 1992. 2003. I... I. Eng. Chem. 128. Kouskoumvekaki... Holten-Andersen. and K ontogeorgis. K.M.M. 345. UNIF AC with lineary temperature-dependent groupinteraction parameters. Danner. R. Economou. Chem.. R. J. Coto. Kontogeorgis.. Th. Lindvig. Res.. J.M. Kontogeorgis. AIChE J. R. 26.G. G.L. 27. Chapman. chap. 19. M. Hansen.. Prediction of infinite dilution act vity coefficients in polymer solutions comparison of prediction models. Res. 211.94 Hansen Solubility Parameters: A User’s Handbook 14. Pappa. 220. Ind.. 22. and K ontogeorgis. 17. M. M. 325. Fluid Phase Equilibria. Prediction of polymer -solvent phase equilibria by a modified group contribution EoS. 16. 324. 34.P. 4654. Fluid Phase Equilibria. R. Prediction of g as solubilities in elastomeric polymers for the design of thermopane windo ws. W. Lee. 1995.. Eng.L.D.. G. Michelsen. P. D. 15... Th. Activity coef ficients in polymer solutions. Michelsen. Giesen.S. 29. 16.P. and K ontogeorgis.. Else vier.C. 2002. IVC-SEP Internal Report 9212. Lindvig. 1997. 41. Eng.P.. Michelsen. 2004.. 18..... G. Fluid Phase Equilibria. Fluid Phase Equilibria. E. 1996. High. G. Panayiotou.. AIChE. G. B. 202(2). M. Ind.L. Lindvig.C.K.. 42.L. 23. 1982.. and Danner . Models for polymer solutions. 77. 22. 97.. and K uhlmann.M. 203. G... Eng. H. Eng.. Lee. Chem. and Fredenslund.. 24.M... R.M. 11–20. Free-v olume acti vity coefficient models for dendrimer solutions. 21. Ne w reference equation of state for associating fluids. R. C. Bogdanic.. Res. J.P. B. 40(21). and K ontogeorgis.. Michelsen. and Danner .

These include swelling. three describing the nonpolar . Hansen ABSTRACT The simplest e xperimental method to determine the Hansen solubility parameters (HSP) for a polymer is to e valuate whether or not it dissolv es in selected solv ents. must be used with care.9) to produce spheres in all three plots. polar. In earlier w ork simple plots were used. One may then observ e or measure a number of different phenomena including full solution at a gi ven concentration. The author’s usual approach to generate data in solubility parameter studies is to contact a polymer of interest with 40 to 45 well-chosen liquids.1 The spheroids in the figure including the δD parameter gave problems. Adding more good solvents well within the sphere or more bad solvents well outside of it will not change an ything but the data fit The goal of the e xperimental work is to arrive at a set of data sho wing differences in behavior among the test solvents. and an y other measurement reflecting di ferences in polymer af finities among the sol ents. one which is considered “good” and the other which is considered “bad. These differences are then traditionally used to di vide the solv ents into tw o groups. surface attack. Experimental data can be generated and treated in v arious ways to arrive at the v alues of interest. v olume change. surface attack. Those solvents dissolving the polymer will ha ve HSP closer to those of the polymer than those solv ents that do not. etc. barrier properties. a radius of interaction for the type of interaction described. Polymer HSP can be higher than the HSP of an y of the test solv ents. The approximate determination of polymer HSP can be done with three plots of experimental data using the HSP parameters pairwise. INTRODUCTION Experience has sho wn that if it is at all possible. Equation 1. ” Such data can be entered into the SPHERE program as discussed in Chapter 1.. the author uses a set of solv ents as described belo w. melting point reduction. Other types of evaluations can also lead to polymer HSP. clarity. viscosity measurements. Ro. Hansen and Skaarup 2 simply used a scaling f actor of 2 (the coefficient “4” in Chapter 1. A simple plot of δP vs. Whenever possible. Supplementary test solv ents are usually in the boundary re gions as it is these that determine the parameters of the sphere. Figure 5. Examples are included in the follo wing.of Characterization 5 Methods — Polymers Charles M. those which use a verages of solv ent HSP to arrive at the polymer HSP . degree of swelling by visual observation or by measurement of weight change. chemical resistance. A computer program or graphical method can then be used to find the HSP for the polyme . The object of the studies is to determine differences in affinity of the polymer for the di ferent solvents. an e xperimental evaluation of the beha vior of a polymer in contact with a series of selected liquids is the best w ay to arrive at its HSP. i. and h ydrogen-bonding interactions for the liquids and the fourth. The author has most often used computer techniques to e valuate the data to find the polyme HSP.e.3 demonstrate how this was attempted initially. As Ro must 95 . This means that some of the methods suggested in the literature to interpret data. often supplemented by selected solv ents depending on the purpose of the investigation. These data are then processed to arrive at the four parameters characteristic of HSP correlations.1 to Figure 5. δH is also helpful in man y practical situations to get guidance as discussed in Chapter 8.

This gives three fractional parameters defined by Equation 5. δH for the solubility of polymeth yl methacrylate (Polymer B in Table 5.2 contains a listing of the polymers included in Table 5. The individual Hansen parameters are normalized by the sum of the three parameters. (From Hansen.96 Hansen Solubility Parameters: A User’s Handbook 12 10 8 δp 6 (P.3 fd = 100 δD/(δD + δP + δH) fP = 100 δP/(δD + δP + δH) fh = 100 δH/(δD + δP + δH) (5.1 Two-dimensional plot of δP vs. Calculations for points in doubt can be made using Chapter 1.4 to Figure 5.2) (5.) be the same in all of these plots. Examples are found in Reference 5 to Reference 7 and in Chapter 8.6. T able 5. Units are (cal/cm 3)1/2. Some accuracy is lost.1.1) (5.1 to Equation 5. With permission. Plots with the modified δD axis are gi ven for the solubility of polystyrene3 shown in Figure 5.1.2). This allows the use of the special triangular technique.3) The sum of these three fractional parameters is 1.M.9.. where comparisons are made between the “hand” method and results of the SPHERE program. The circle is the projection of a sphere on the gi ven coordinates. are modified HS parameters. An idea of the accurac y of the graphical approach can be found in T able 5. 71. C. Specific solubility data are g ven for these polymers in 88 solv ents in Appendix A.0. The triangular plotting technique uses parameters for the solv ents.H) 4 R 2 2 4 6 8 δh 10 12 14 FIGURE 5. and δH to see how well the separation into good and bad solv ents is accomplished. and there is no theoretical justification for this plotting technique . which. δP . a single compass setting is tried for a set of δD. Färg och Lack. These are the original figures fro this thesis. 17(4). and the numbers refer to a table of solv ents found there. 1971.3. Teas4 has developed a triangular plotting technique which helps visualization of three parameters on a plain sheet of paper. Equation 1. in f act.

M. The results are not yet fully satisf actory.2). epoxy. and other polymers potentially useful in self-stratifying coatings ha ve been reported by Benjamin et al. Group . Units are (cal/cm3)1/2.9 is v ery often if not most often significant.2 Two-dimensional plot of δH vs. Rasmussen and Wahlström10 pro vide additional HSP data in relation to the use of replenishable natural products (oils) in connection with solv ents. There are other sources of HSP for polymers in the literature.3. but a full review of these and their uses is beyond the scope of this book. One of the more significant e forts in this has been made by Utracki and coworkers.15. HSP for the polymers and film formers discussed in the foll wing examples are given in Table 5. because it w as described in a standard reference book very shortly after it w as developed. and should not be replaced with a “1 ” either. but there is hope for the future. Values for a number of acrylic.4 sho ws how such a plot can be used in finding suitable solvent when dealing with such an older oil painting. C. With permission. also point out man y of the problems encountered with these characterizations. 1971. This plotting technique is often used by those who conserv e old paintings.16 They assumed the δD parameter for polymers did not dif fer too much between polymers and interpreted e valuations of polymer–polymer compatibility using calculated v alues for δP and δH. 71.. Zellers et al. Barton 6. CALCULATION OF POLYMER HSP Calculation of the HSP for polymers is also possible.8 has also provided solubility parameters for many polymers. These data are based on solubility determinations unless otherwise noted. (From Hansen. 9 (see Chapter 8).) but one does get all three parameters onto a tw o-dimensional plot.Methods of Characterization — Polymers 97 12 10 8 δp 6 4 2 2 4 6 δh 8 10 12 14 FIGURE 5. Expansion of the δD scale by a f actor of 2 w ould yield a circle (a sphere in projection). The data processing techniques and data accumulated by Zellers and co workers11–14 on elastomers used in chemical protecti ve clothing are also useful. Färg och Lack. 17(4). δD for the solubility of polymeth yl methacrylate (Polymer B in Table 5. A word of caution is advisable here and that is that the constant “4” in Equation 1. Such problems are also discussed belo w.7 Figure 8.

The author has never been particularly successful in calculating the same values as were found experimentally. 1971. This is probably the best calculation approach currently a vailable.9. although the estimated dispersion parameters are thought to be too lo w. These are essentially those included in Table 5. The group contributions given in Chapter 1 can be used for this purpose. as discussed in the following. 17(4). Units are (cal/cm3)1/2.M. See Chapter 3. although good laboratory practice should allo w use of the ones . It can be presumed that th errors involved in either process will cancel internally . The solvents should ha ve different HSP chosen for systematic e xploration of the three parameters at all levels. a starting point could be the series of liquids used by the author for many years.10) near 1. Equation 1. It is suggested that HSP for polymers determined by these calculations not be mix ed with e xperimentally determined HSP until confirmation of agreement is found. As indicated earlier. C. It is actually the boundary which is used to define the center point of the spher using Chapter 1. Equation 1. b ut improvements are thought possible.0 are located near the sphere boundary. although a serious ef fort to use weighting and similar f actors. δD for the solubility of polymeth yl methacrylate (Polymer B in Table 5. as solv ents with RED numbers (Chapter 1.98 Hansen Solubility Parameters: A User’s Handbook 12 10 8 δp 6 4 2 2 4 6 8 δd 10 12 14 FIGURE 5. SOLUBILITY — EXAMPLES The most direct method to determine the three HSP for polymers or other soluble materials is to evaluate their solubility or de gree of swelling/uptak e in a series of well-defined sol ents. With permission.3 Two-dimensional plot of δP vs..4. Färg och Lack. (From Hansen. has never been tried. Some changes are also possible to remove or replace solvents which are now considered too hazardous.2). 72.) calculations were used. Sometimes boundaries are define better by inclusion of additional test solv ents. A computer analysis quickly gives a choice of many of these. but these may not necessarily be the same for the calculated results as for the e xperimental ones. Expansion of the δD scale by a f actor of 2 w ould yield a circle (a sphere in projection).

indicated. Each of these error situations reduces the data fit.3. The input data to the SPHERE program described in Chapter 1 are included in Table 5.4 Two-dimensional plot δP vs. as noted in the follo wing.2). and the boundarie are rather poorly defined. The entry for polyethersulfone (PES) found in Table 5. where it should not. . Perfect fits are rather easily obtained with small sets of data. The solubility of PES w as evaluated in 41 dif ferent solvents. there is an indication of the data fit.000. H.3 w as determined from data included in the computer output reported in Table 5.0 is actually preferred. as the computer program has the optimized the data to a single set of v alues that are so close to being correct as the y can be within experimental error. and R for the solubility of PES are given at the top. In addition. which is 0. Table 5.999 here. which means it is an outlier. These can be found easily by listing the solv ents in order of their RED numbers and choosing .4. The HSP generally in use for liquids ha ve all been e valuated/calculated at 25°C.Methods of Characterization — Polymers 99 16 14 SOLUBLE NOT SOLUBLE STRONG INTERACTION 12 10 δp 28 45 8 43 6 38A 29 RA 69 21 19 19A 15A 18A 18 42 15 67A 89 25 4 2 32 23 0 δd 0 2 4 6 8 δh 10 12 14 16 FIGURE 5. It w as found that fi e of them actually dissolv ed the polymer . One can continue testi with additional solv ents located in the boundary re gions of the established sphere as stated pre viously. A data fit slightly less than 1. These same values can also be used to correlate ph ysical phenomena related to solubility at other test temperatures with some care.0 when ver this result is found. A perfect fit is 1. Several examples of HSP correlations based on solubility are found in Table 5.4 in the SOLUB column. δH of solubility data for polystyrene (Polymer G in Units are (cal/cm 3)1/2. which means the center is also poorly defined. D. and a 0* means a bad solv ent lies inside the sphere. An unknown number of sets of the parameters can gi ve a data fit of 1. A 1* means that a good solv ent lies outside the sphere. A “1” means a good solvent and a “0” means a bad solv ent.

but such an e xtensive experimental study w as not undertak en to confirm the predictions A special problem that can be encountered is when only a fe w solvents with very high solubility parameters dissolve a polymer.5 Two-dimensional plot δH vs.3. Finally.1 of the first edition of this handbook. of the solvents in cc/mol. Units are (cal/cm 3)1/2. Data on good and bad solv ents17 for polyacrylonitrile (PAN) have been used as input to the computer program.0 can be e xpected to dissolv e this polymer .6 The entry in Table 5. H ving found the HSP for a polymer in this manner. The RED number is gi ven for each solv ent in the RED column. A quality number.2).3 places a big question mark o ver the solubility parameters. V.3. one can then search a database for additional solvents for the polymer in question.100 Hansen Solubility Parameters: A User’s Handbook 16 14 SOLUBLE NOT SOLUBLE STRONG INTERACTION 12 10 δh 8 9 67A21 19A 69 15B 35 25 29 23 45 42 47A 28 RA 32 45 56 15 6 4 2 0 δp 5 6 7 26A 8 63 48 9 δd 10 11 12 FIGURE 5.0 and the perfect fit of the data. The data fit of 0. Q = 1 – RED.4 indicating the molar volume. An example is polyvinyl alcohol with true solvents being 1-propanol and ethanol in a data set with 56 solv ents. This was done for the HSP database with o ver 800 solvent entries in Table A. but the data are still useful in finding the HSP for thi polymer. is also conceptually useful. Equation 1.The computer analysis quickly encompasses the tw o good solv ents in the data set within a small sphere as the y . as well as with the radius 4. These test solv ents do not dif fer as widely from each other as the test series suggested earlier .0. and 23 indicated as bad. there is a column in Table 5. There was no need to analyze the influence of this parameter in the present case A second example of this type of approach is gi ven in Table 5. A significantly la ge number of the 123 additional solvents found to ha ve RED numbers less than 1. There are 13 solvents indicated as good. those with v alues not too dif ferent from 1. Expansion of the δD scale by a factor of 2 has given a spherical representation according to Chapter 1. δD of solubility data for polystyrene (Polymer G in Table 5.9.931 is good for this kind of data. These are reported in Table 5.

Methods of Characterization — Polymers





10 δp


40 16 38A

44 26B 45 43




69 29 21 58 8 RA



67A 23 25


32 8


0 δh





8 δd





FIGURE 5.6 Two-dimensional plot δP vs. δD of solubility data for polystyrene (Polymer G in Table 5.2). Expansion of the δD scale by a factor of 2 has given a spherical representation according to Chapter 1, Equation 1.9. Units are (cal/cm 3)1/2.

have reasonably similar parameters. Based on reasonable similarity with other solubility correlations for w ater-soluble polymers, one anticipates spheres with a radius much lar ger than the distance between these solv ents. This result is not good and should not be used. Another e xample of determining HSP for a polymer with v ery high solubility parameters is Dextran C (British Drug Houses). Only 5 out of 50 solv ents were found to dissolv e Dextran C. 18 In this case, there w as enough spread in the solubility parameters of the test solv ents such that the spherical model correlation (Chapter 1, Equation 1.9) forced the program to find a radius of 17. MPa1/2. This appears to be a reasonable number for this situation. The problem can be made clearer by noting the dissolving solv ents with their RED numbers in parentheses. These were dimeth yl sulfoxide (1.000), ethanolamine (0.880), eth ylene glycol (0.880), formamide (0.915), and glycerol (0.991). Some dissolving liquids had RED equal to 1.0 or higher and included dieth ylene glycol (1.000), propylene glycol (1.053), and 1,3-butanediol (1.054). These helped to define the boundar of the Hansen solubility sphere. Note that the HSP for the polymer are in a re gion higher than that defined by the alues of test liquids. Any technique using an average of the HSP for the test solvents will inherently underestimate the solute HSP in such a situation. The solubility data for the polymer Dextran C led to the HSP data reported in Table 5.3 when the SPHERE program used a starting point based on a verages of the HSP v alues for the good solv ents. When the starting point w as 25 MPa1/2 for D, P, H, and Ro, respecti vely, a perfect data fit as found for D, P, H, and Ro equal to 26, 26, 26, and 24, all in MP a1/2. When the starting point w as for D, P, H, and Ro equal to 30, 30,


Hansen Solubility Parameters: A User’s Handbook

TABLE 5.1 Calculated vs. Trial-and-Error Solubility Parameter Data for Various Polymersa
Computed Handtrials A B C D E F G H I J K L M N O P Q (First Values) (Second Values) δD δP δH 8.60 9.2 9.11 9.2 9.95 8.5 9.98 8.5 9.79 9.4 9.09 8.5 10.40 8.6 10.23 9.3 10.17 9.5 7.53 7.0 9.90 9.3 9.08 9.5 11.37 9.0 9.65 9.4 10.62 8.9 8.58 8.5 9.87 4.72 5.3 5.14 5.0 5.88 5.5 1.68 2.5 2.84 3.2 2.13 4.3 2.81 3.0 5.51 5.0 4.05 4.0 7.20 7.0 3.09 3.7 6.22 6.0 3.20 4.0 5.68 5.3 0.46 3.0 4.58 4.7 6.43 1.94 2.1 3.67 4.2 5.61 5.5 2.23 3.0 5.34 5.1 6.37 5.5 2.10 2.0 4.72 4.0 7.31 6.4 4.32 4.3 2.64 2.1 5.38 6.0 4.08 5.5 7.13 7.4 4.17 3.8 7.00 6.5 6.39 Computed Handtrials Ro 5.20 5.3 4.20 4.0 6.20 4.7 6.70 5.3 5.70 5.0 5.20 4.8 6.20 3.5 6.70 4.9 6.20 4.7 5.60 5.5 5.20 4.2 3.70 4.5 9.70 6.4 6.20 5.5 7.70 4.5 5.20 5.0 5.70 FIT 0.960 0.923 0.945 0.923 0.853 0.829 0.974 0.957 0.930 0.929 0.948 0.871 0.955 0.915 0.891 0.855 0.924 0.909 0.933 0.918 0.949 0.933 0.921 0.896 0.978 0.923 0.897 0.867 1.000 0.952 0.942 0.940 0.942 R S T U V X Y Z A B C D E F G L (First Values) (Second Values) δD δP 9.04 9.2 10.53 8.8 8.58 8.7 9.10 9.3 8.10 8.5 7.10 7.8 8.57 8.8 8.52 8.2 9.60 8.7 9.95 9.5 8.05 8.5 10.34 9.2 8.58 8.5 8.91 9.4 9.49 8.8 9.86 10.8 9.3 4.50 4.5 7.30 7.0 1.64 1.8 4.29 4.5 0.69 1.5 1.23 1.0 1.10 2.5 –0.94 0.8 2.31 2.5 4.17 4.0 0.18 1.0 6.63 5.8 0.58 1.5 3.68 4.5 2.68 2.7 7.14 7.0 6.2

δH 2.40 2.6 6.00 5.5 1.32 1.8 2.04 2.0 –0.40 1.5 2.28 3.6 1.67 1.2 7.28 5.7 3.80 3.5 5.20 5.5 1.39 2.0 6.26 4.2 1.76 1.8 4.08 3.5 2.82 2.7 7.35 8.8 4.7

Ro 5.20 5.0 8.20 6.0 3.20 3.5 4.70 4.7 4.70 3.4 6.20 4.0 3.20 3.8 4.70 2.9 5.20 4.2 7.20 7.0 4.20 3.4 6.70 5.0 3.20 2.6 1.70 3.2 4.70 4.0 5.70 7.1 4.2

FIT 0.985 0.972 0.910 0.879 0.974 0.965 0.969 0.950 0.974 0.964 0.921 0.881 0.950 0.914 0.971 0.954 0.942 0.951 0.980 0.976 0.966 0.960 0.964 0.868 0.968 0.956 0.992 0.895 0.961 0.963 0.970 0.936 0.892

Note: Units are (cal/cm 3)1/2.

See Table 5.2 for polymer types.

Source: From Hansen, C.M., Färg och Lack, 17(4), 73, 1971. With permission.

30, and 30, all in MP a1/2, a perfect correlation w as found to D, P , H, and Ro equal to 30, 28, 28, and 32, all in MPa1/2. These data show that extrapolations into regions where there are no data can be problematic. It is thought that the data gi ven in Table 5.3 for Dectran C are the most representative, because of the data fit being slightly less than 1.0 g ving a better definition of a boundar . The properties of good solv ents alone cannot al ways lead to a good estimate of the solubility parameters for these polymers, and the radii of spheres using only a fe w solvents with high solubility parameters will be very uncertain. One can sometimes find better results by correlating d grees of swelling or uptak e, rather than correlate on solubility or not. The work of Zellers and co workers

Methods of Characterization — Polymers


TABLE 5.2 List of Polymers and Resins Studied
A B C D E F G H I J K L M N O P Q R S T U V X Y Z A B C D E F G L Lucite® 2042-poly (eth yl methacrylate), E. I. du Pont de Nemours & Co., Inc. Poly (methyl methacrylate), Rohm and Haas Co. Epikote® 1001-epoxy, Shell Chemical Co. Plexal P65-66% oil length alk yd, Polyplex. Pentalyn® 830-alcohol soluble rosin resin, Hercules Incorporated. Butvar® B76-poly (vin yl butyral), Shawinigan Resins Co. Polystyrene LG, Badische Anilin- und Soda F abrik. Mowilith® 50-poly (vin yl acetate), F arbwerke Hoechst. Plastopal H-urea formaldeh yde resin, Badische Anilin- und Soda F abrik. H Sec. Nitrocellulose-H 23, A. Hagedorn and Co. Parlon® P10-chlorinated poly (prop ylene), Hercules Incorporated. Cellulose acetate, Cellidora A-Bayer AG. Super Beckacite ® 1001-Pure Phenolic Resin, Reichhold Chemicals Co. Phenodur 373U-phenol-resol resin, Chemische Werke Albert. Cellolyn 102-modified pentaerythritol ester of rosin, Hercules Incorporated Pentalyn 255-alcohol soluble resin, Hercules Incorporated. Suprasec F5100-blocked isocyanate (phenol), Imperial Chemical Ind. Ltd. Plexal C34-34% coconut oil-phthalic anh ydride alkyd, Polyplex. Desmophen 850, Polyester -Farbenfabriken Bayer AG. Polysar 5630 — styrene-b utadiene (SBR) ra w elastomer, Polymer Corp. Hycar® 1052-acrylonitrile-butadiene raw elastomer, B. F. Goodrich Chemical Corp. Carifl x IR 305-isoprene ra w elastomer, Shell Chemical Co. Lutanol IC/123-poly (isob utylene), Badische Anilin- und Soda F abrik. Buna Huls CB 10-cis poly b utadiene raw elastomer, Chemische Werke Huels. Versamid® 930-polyamide, General Mills, Inc. Ester gum BL, Hercules Incorporated. Cymel® 300-hexamethoxy melamine, American Cyanamid Co. Piccolyte® S100-terpene resin, Pennsylv ania Industrial Chemical Corp. Durez® 14383-furfuryl alcohol resin, Hook er Chemical Co. Piccopale® 110-petroleum h ydrocarbon resin, Pennsylv ania Industrial Chemical Corp. Vipla KR-poly (vin yl chloride), K = 50, Montecatini. Piccoumarone 450L-cumarone-indene resin, Pennsylv ania Industrial Chemical Corp. Milled wood lignin — special sample from Prof. A. Björkman.

TABLE 5.3 Hansen Solubility Parameter Correlations for Selected Materials
Material PES solubility PAN solubility PP swelling Polyvinyl alcohol ? (see te xt) Hexamethylphosphoramide PVDC melting temperature 110°C PVDC melting temperature 130°C Dextran C solubility Note: Units are (cal/cm 3)1/2. δD 19.6 21.7 18.0 17.0 18.5 17.6 20.4 24.3 δP 10.8 14.1 3.0 9.0 8.6 9.1 10.0 19.9 δH 9.2 9.1 3.0 18.0 11.3 7.8 10.2 22.5 Ro 6.2 10.9 8.0 4.0 — 3.9 7.6 17.4 FIT 0.999 0.931 1.00 1.00 — 0.992 0.826 0.999 G/T 5/41 13/36 13/21 2/56 — 6/24 13/24 5/50


Hansen Solubility Parameters: A User’s Handbook

TABLE 5.4A Calculated Solubility SPHERE for PES Solubility
D = 19.6 P = 10.8 H = 9.2 RAD = 6.2 FIT = 0.999 NO = 41 Solvent Acetone Acetophenone Benzene 1-Butanol Butyl acetate γ-Butyrolactone Carbon tetrachloride Chlorobenzene Chloroform Cyclohexanol Diacetone alcohol o-Dichlorobenzene Diethylene glycol Diethyl ether Dimethyl formamide Dimethyl sulfoxide 1,4-Dioxane Ethanol Ethanolamine Ethyl acetate Ethylene dichloride Ethylene glycol Ethylene glycol monob utyl ether Ethylene glycol monoethyl ether Ethylene glycol monomethyl ether Formamide Hexane Isophorone Methanol Methylene dichloride Methyl ethyl k etone Methyl isobutyl ketone Methyl-2-pyrrolidone Nitroethane Nitromethane 2-Nitropropane Propylene carbonate Propylene glycol Tetrahydrofuran Toluene Trichloroethylene Note: Units are MP a1/2.

D 15.5 19.6 18.4 16.0 15.8 19.0 17.8 19.0 17.8 17.4 15.8 19.2 16.6 14.5 17.4 18.4 19.0 15.8 17.0 15.8 19.0 17.0 16.0 16.2 16.2 17.2 14.9 16.6 15.1 18.2 16.0 15.3 18.0 16.0 15.8 16.2 20.0 16.8 16.8 18.0 18.0

P 10.4 8.6 0.0 5.7 3.7 16.6 0.0 4.3 3.1 4.1 8.2 6.3 12.0 2.9 13.7 16.4 1.8 8.8 15.5 5.3 7.4 11.0 5.1 9.2 9.2 26.2 0.0 8.2 12.3 6.3 9.0 6.1 12.3 15.5 18.8 12.1 18.0 9.4 5.7 1.4 3.1

H 7.0 3.7 2.0 15.8 6.3 7.4 0.6 2.0 5.7 13.5 10.8 3.3 20.7 5.1 11.3 10.2 7.4 19.4 21.2 7.2 4.1 26.0 12.3 14.3 16.4 19.0 0.0 7.4 22.3 6.1 5.1 4.1 7.2 4.5 5.1 4.1 4.1 23.3 8.0 2.0 5.3

SOLUB 0 1 0 0 0 1 0 0 0 0 0 0 0 0 1 0*a 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0

RED 1.371 0.955 2.129 1.777 1.741 0.998 2.301 1.576 1.483 1.467 1.321 1.204 2.101 2.183 0.915 0.996 1.493 2.077 2.241 1.547 1.007 2.837 1.563 1.395 1.618 3.044 2.745 1.094 2.575 0.990 1.368 1.782 0.655 1.580 1.899 1.387 1.429 2.457 1.237 1.978 1.485

V 74.0 117.4 89.4 91.5 132.5 76.8 97.1 102.1 80.7 106.0 124.2 112.8 94.9 104.8 77.0 71.3 85.7 58.5 59.8 98.5 79.4 55.8 131.6 97.8 79.1 39.8 131.6 150.5 40.7 63.9 90.1 125.8 96.5 71.5 54.3 86.9 85.0 73.6 81.7 106.8 90.2

Outlier (a bad solv ent lying inside sphere).

Methods of Characterization — Polymers


ALTERNATE TABLE 5.4B Calculated Solubility SPHERE for PES Solubility (Listed in RED Order)
D = 19.6 P = 10.8 H = 9.2 RAD = 6.2 FIT = 0.999 NO = 41 Solvent Methyl-2-pyrrolidone Dimethyl formamide Acetophenone Methylene dichloride Dimethyl sulfoxide γ-Butyrolactone Ethylene dichloride Isophorone o-Dichlorobenzene Tetrahydrofuran Diacetone alcohol Methyl ethyl k etone Acetone 2-Nitropropane Ethylene glycol monoethyl ether Propylene carbonate Cyclohexanol Chloroform Trichloroethylene 1,4-Dioxane Ethyl acetate Ethylene glycol monob utyl ether Chlorobenzene Nitroethane Ethylene glycol monomethyl ether Butyl acetate 1-Butanol Methyl isobutyl ketone Nitromethane Toluene Ethanol Diethylene glycol Benzene Diethyl ether Ethanolamine Carbon tetrachloride Propylene glycol Methanol Hexane Ethylene glycol Formamide Note: Units are MP a1/2.

D 18.0 17.4 19.6 18.2 18.4 19.0 19.0 16.6 19.2 16.8 15.8 16.0 15.5 16.2 16.2 20.0 17.4 17.8 18.0 19.0 15.8 16.0 19.0 16.0 16.2 15.8 16.0 15.3 15.8 18.0 15.8 16.6 18.4 14.5 17.0 17.8 16.8 15.1 14.9 17.0 17.2

P 12.3 13.7 8.6 6.3 16.4 16.6 7.4 8.2 6.3 5.7 8.2 9.0 10.4 12.1 9.2 18.0 4.1 3.1 3.1 1.8 5.3 5.1 4.3 15.5 9.2 3.7 5.7 6.1 18.8 1.4 8.8 12.0 0.0 2.9 15.5 0.0 9.4 12.3 0.0 11.0 26.2

H 7.2 11.3 3.7 6.1 10.2 7.4 4.1 7.4 3.3 8.0 10.8 5.1 7.0 4.1 14.3 4.1 13.5 5.7 5.3 7.4 7.2 12.3 2.0 4.5 16.4 6.3 15.8 4.1 5.1 2.0 19.4 20.7 2.0 5.1 21.2 0.6 23.3 22.3 0.0 26.0 19.0

SOLUB 1 1 1 1 0*a 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

RED 0.655 0.915 0.955 0.990 0.996 0.998 1.007 1.094 1.204 1.237 1.321 1.368 1.371 1.387 1.395 1.429 1.467 1.483 1.485 1.493 1.547 1.563 1.576 1.580 1.618 1.741 1.777 1.782 1.899 1.978 2.077 2.101 2.129 2.183 2.241 2.301 2.457 2.575 2.745 2.837 3.044

V 96.5 77.0 117.4 63.9 71.3 76.8 79.4 150.5 112.8 81.7 124.2 90.1 74.0 86.9 97.8 85.0 106.0 80.7 90.2 85.7 98.5 131.6 102.1 71.5 79.1 132.5 91.5 125.8 54.3 106.8 58.5 94.9 89.4 104.8 59.8 97.1 73.6 40.7 131.6 55.8 39.8

Outlier (a bad solv ent lying inside sphere).


Hansen Solubility Parameters: A User’s Handbook

reports extensive studies of this type. 11–14 It should be noted, ho wever, that the HSP-sphere parameters usually v ary some from correlation to correlation based on the same data when dif ferent criteria are used for good and bad solv ents. This is because the absorbed solv ent tends to locate in regions with similar solubility parameters, and there are local variations in HSP within most, if not all, polymers. This is particularly true of polymers which are not homopolymers. This situation relates to self-assembly . Solvents or se gments of molecules with similar HSP will tend to reside near each other . An e xample of this is w ater residing at local h ydrophilic sites, such as alcohol groups, in polymers. Utilization of the HSP af finity between molecules or s gments of molecules is a viable w ay to control self-assembly . See also Chapter 18.

The correlation for swelling of polypropylene reported in Table 5.3 is based on solvent uptake data reported by Lieberman and Barbe at 22°C. 19 The limit of 0.5% w as arbitrarily set to dif ferentiate good solv ents from bad ones. As mentioned earlier , e xperience has sho wn that a dif ferent limit usually gives different parameters. It should be noted that swelling data reflect the properties o the regions in the polymer where the solv ent has chosen to reside because of ener getic similarity (self-assembly). The principle is not necessarily “lik e dissolves like,” but rather “lik e seeks lik e.” If the solv ent is homogeneously distrib uted in the polymer , the solubility parameters found will reflect the properties of the whole polyme . Crystalline re gions will not contain solv ent. If the solvent collects locally in re gions with chemical groups dif ferent from the b ulk of the polymer , then the HSP so deri ved will reflect at least partially the p ysical nature of these chemical groups. The parameters reported in Table 5.3 seem appropriate for what is expected in terms of low polarity and low hydrogen-bonding properties for a polyprop ylene-type polymer. An example of a characterization using swelling data that did not result in a good correlation is that for Viton® (The Du Pont Compan y, Wilmington, DE). This problem has been discussed by Zellers and Zhang 11,12 and is also discussed in Chapter 13. If one tries to force-fit data where ther are several different comonomers into a single HSP sphere, the result is usually reflected in a poo correlation coef ficient. Figure 13.3 sh ws that impro vements can be made by using a separate sphere for each comonomer. One reason for the poor correlation of swelling beha vior is that Viton is not a homopolymer , and also contains a cross-linking chemical. The dif ferent se gments ha ve different af finities. Indeed, there are s veral qualities of Viton, each of which has significantl differing chemical resistance. Swelling of Viton has also been treated by Ev ans and Hardy 20 in connection with predictions related to chemical protecti ve clothing, and by Nielsen and Hansen, 21 who presented curv es of swelling as a function of the RED number .

Partly crystalline polymers that are placed in dif ferent liquids will ha ve melting points which are lowered to a de gree depending on the solv ent quality of the indi vidual liquids. The melting points of polyvin ylidine chloride (PVDC) ha ve been measured in dif ferent solv ents.22 These data ha ve been analyzed by e valuating solubility parameter re gions based on those solv ents which dissolv e the polymer at 110°C and above and also at 130°C and above. As expected, there are more solvents which dissolve the semicrystalline polymer at the higher temperature. The results for these correlations are included in Table 5.3. The main reasons for the somewhat lower data fit at 130°C includ two nondissolving solv ents within the solubility parameter sphere. These are dimeth yl phthalate, where the lar ge molecular size is a f actor, and benzyl alcohol, where temperature ef fects can be larger than e xpected compared with the other solv ents as discussed later and in Chapter 1. The solubility parameters for PVDC at this temperature, based on tab ulated solvent values at 25°C, are not affected significantly by this type of situation. A single room temperature solv ent for PVDC

Methods of Characterization — Polymers


is reported by Wessling.22 This is hexamethylphosphoramide and its solubility parameters are also reported in Table 5.3 for comparison.The change in the v alues of the indi vidual solubility parameters with temperature is discussed in Chapter 1 (Equation 1.16 to Equation 1.18). Chapter 3 also treats the temperature dependence of the HSP. See also Chapter 10 where the HSP of specifically carbo dioxide are treated in depth as a function of temperature and pressure.

Environmental stress cracking (ESC) is unfortunately a v ery frequent mode of f ailure for plastics. For this reason a whole chapter is de voted to the topic (see Chapter 14). It has been possible to correlate HSP with ESC phenomena, and this can also provide an estimate of the HSP for the given polymers. Care is advised since the stress/strain le vel is important, as is the molecular size and shape of the chemicals in volved. Several collections of ESC data in the older literature 23–25 should not be forgotten in these days of “Internet and only Internet.” Such collections have particular value as it is considered impossible to get a commercially a vailable polymer without some additi ves. These can also affect ESC behavior. These older data were the basis of the ESC correlations gi ven in Chapter 14.

One of the more promising methods to e valuate polymer HSP for limited data is that using the intrinsic viscosity. Van Dyk et al. found a correlation with the intrinsic viscosity of an acrylic polymer (polyethyl methacrylate) in v arious solvents and the polymer HSP 26 (see the discussion on polymer compatibility in Chapter 8). Segarceanu and Leca 27 have devised a method to calculate the polymer HSP from data on its intrinsic viscosity in different solvents. The intrinsic viscosities will be higher in the better solvents because of greater interaction and greater polymer chain e xtension. The intrinsic viscosity gi ves an indication of the solv ent quality. It has been used earlier to calculate the Flory–Huggins chi parameter, for e xample.28 In the ne w technique, the intrinsic viscosities are normalized by the intrinsic viscosity of that solvent giving the highest v alue. These normalized data (numbers are 1.0 or less) are then used in a weighted a veraging technique to arri ve at the center of the Hansen sphere. δD2 = Σ(δDi × [η]i)/Σ[η]i δP2 = Σ(δPi × [η]i)/Σ[η]i δH2 = Σ(δHi × [η]i]/Σ[η]i (5.4) (5.5) (5.6)

The subscript 2 is for the polymer , and the respecti ve solv ents are indicated by an “i. ” The intrinsic viscosity in the i-th solv ent is given by [ η]i. Those solvents with the greatest weighting f actor have higher intrinsic viscosities and are closest to the geometric center of the sphere. Those solv ents which do not dissolv e the polymer were assumed to have a zero weighting factor. The HSP for a polyesterimide were reported as an e xample. HSP values were assigned both by the “classical” e valuation and with this ne wer approach. These data are included as the first entries in Table 5.5. This is a v ery promising method of arri ving at the polymer HSP with limited data. There are se veral aspects of this w ork which deserv e comment. It w as demonstrated earlier that many polymers ha ve higher solubility parameters than an y of the solv ents which are or can be used to test them. The present method only allows for polymer HSP within the range attainable


Hansen Solubility Parameters: A User’s Handbook

TABLE 5.5 HSP Data for the Same Polyesterimide Polymer Based on Data Given in Reference 27
Correlation Classicala1 Newa1 HSP SPHERE a HSP SPHERE b Classicala Classicalb Newa Newb δD 17.4 18.0 20.0 19.0 δP 12.3 11.1 11.0 11.0 δD 8.6 8.8 10.0 9.0 Ro 4.1 8.6 8.3 7.0 FIT — — 1.000 1.000 0.426 0.447 0.506 0.364

Note: Units are MP a1/2. Indicates use of the solubility parameters for the solvents given in Reference 27. b Indicates use of the solv ent HSP data in the author’ s files

by the test solv ents. The method will lead to v alues that are too lo w in some cases, including the example with the polyesterimide used as an e xample in Segarceanu and Leca. 27 It is not surprising that the polymer HSP are often higher than solv ent HSP, as the y are in a ph ysical state between that of a liquid and a solid. When the cohesion energy becomes too high, a material is a solid rather than a liquid. Lo w molecular weight solids frequently ha ve HSP some what higher than the HSP of liquids. Man y examples can be gi ven, including urea, eth ylene carbonate, etc. When the data (as soluble or not) for the 11 solvents were processed by the SPHERE computer program, the parameters found were those gi ven by the third set of HSP inTable 5.5. The agreement with the “new” method is acceptable, even though none of the test solv ents have δd as high as that of the polymer . Further inspection sho wed that the solubility parameters used in the study were not in agreement with those published in the latest reference to Hansen listed by Se garceanu and Leca.27 It also appears that the radius of the HSP sphere for the classical determination is in error , being far too low. To further clarify the situation, se veral runs with the SPHERE program were done with the parameters listed in this book, as well as with those listed in the article being discussed. In both cases the data fit is not good for the HSP reported by S garceanu and Leca. 27 In the classical case, the data fit is only 0.426 (1.0 is perfect), and four of the f e good solvents are located outside of the sphere. Only N-methyl-2-pyrrolidone is inside. In the ne w case, the data fit is not much bette , being 0.506. Here, four of the fi e bad solvents are inside the sphere with only one being outside. It has been possible to estimate the polymer parameters within acceptable v ariation, but the radius of the sphere has not been accounted for in a satisf actory manner. Further inspection of the data suggests that morpholine, the solv ent with the highest [ η] that was used to normalize the data, is not as good as might ha ve been e xpected from the intrinsic viscosity data. This can be seen in Table 5.6. The reason for this is unkno wn, but experience has shown that amines often are seen to react with v arious materials in a manner which does not allo w their inclusion in correlations of the type discussed here. To conclude this section, it is noted that a similar weighting technique w as used by Zellers et al.13,14 where the weighted measurements were solv ent uptake by elastomers customarily used to make chemical protecti ve clothing. The same precautions must be tak en in analyzing this type of

Methods of Characterization — Polymers


TABLE 5.6 Calculated Solubility SPHERE for Polyesterimide (Listed in RED Order)
D = 19.0 P = 11.0 H = 9.0 RAD = 7.0 FIT = 1.000 NO = 11 Solvent Methyl-2-pyrrolidone Dimethyl formamide Dimethyl sulfoxide γ-Butyrolactone Morpholine Cyclohexanone Diacetone alcohol Acetone Diethylene glycol monomethyl ether Ethylene glycol monoethyl ether Ethylene glycol monoethyl ether acetate Note: Units are MP a1/2.

[η]a N 0.970 0.947 0.182 0.689 1.000 0.718 — — — — —

D 18.0 17.4 18.4 19.0 18.8 17.8 15.8 15.5 16.2 16.2 15.9

P 12.3 13.7 16.4 16.6 4.9 6.3 8.2 10.4 7.8 9.2 4.7

H 7.2 11.3 10.2 7.4 9.2 5.1 10.8 7.0 12.6 14.3 10.6

SOLUB 1 1 1 1 1 1 0 0 0 0 0

RED 0.427 0.682 0.809 0.832 0.874 0.937 1.031 1.044 1.055 1.131 1.283

V 96.5 77.0 71.3 76.8 87.1 104.0 124.2 74.0 118.0 97.8 136.1

Normalized intrinsic viscosity data from Reference 27.

measurement, but as the polymers studied were reasonably nonpolar, some of the solvents had HSP which were higher than those of the polymers studied. Zellers et al. 14 and Athey29 also describe multiple variable statistical analysis techniques to find the HSP of a g ven polymer. Barton’s work6 contains man y literature sources of intrinsic viscosity studies using the solubility parameter for interpretation.

There are man y other techniques to dif ferentiate between the beha vior of dif ferent solv ents in contact with a polymer . Man y of these are discussed in the follo wing chapters and will not be treated here. These include permeation measurements, chemical resistance determinations of a vrious kinds including ESC, and surface attack, etc. Some of the techniques can be very useful, depending on the polymer in volved. Others will present problems because of the probable influence of othe factors such as solvent molar volume and length of time before attainment of equilibrium. Se veral of these phenomena can be correlated with HSP, but the techniques used in the measurements will present problems in using the data for direct HSP characterization of polymers because other ef fects are also important.

HSP for polymers can be e valuated experimentally by correlations of data where a suitably lar ge number of well-chosen solvents are brought into contact with the polymer . The observed behavior which can be correlated includes true solubility, swelling, weight gain, dimensional change, degree of surf ace attack, reduction of melting point, permeation rate, breakthrough time, and tensile strength reduction. Correlations for simple e valuations of chemical resistance of the suitable-or not type and ESC are also possible. In each case, the molecular size of the liquids used can af fect the result and should be considered in some w ay. The use of w ater as a test liquid is not recommended for these purposes.


Hansen Solubility Parameters: A User’s Handbook

1. Hansen, C.M., Solubility in the coatings industry , Färg och Lack, 17(4), 69–77, 1971. 2. Hansen, C.M. and Skaarup, K., The three dimensional solubility parameter — key to paint component affinities III. Independent calculation of the parameter components, J. Paint Technol., 39(511), 511–514, 1967. 3. Hansen, C.M., The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient, Their Importance in Surf ace Coating F ormulation, Doctoral dissertation, Danish Technical Press, Copenhagen, 1967. 4. Teas, J.P., Graphic analysis of resin solubilities, J. Paint Technol., 40(516), 19–25, 1968. 5. Gardon, J.L. and Teas, J.P., Solubility parameters, in Treatise on Coatings, Vol. 2, Characterization of Coatings: Physical Techniques, P art II, Myers, R.R. and Long, J.S., Eds., Marcel Dekk er, Ne w York, 1976, chap. 8. 6. Barton, A.F.M., Handbook of Solubility Parameters and Other Cohesion Parameters, CRC Press, Boca Raton, FL, 1983; 2nd ed., 1991. 7. Torraca, G., Solubility and Solvents for Conservation Problems, 2nd ed., International Centre for the Study of the Preserv ation and the Restoration of Cultural Property (ICCR OM), Rome, 1978. 8. Barton, A.F.M., Handbook of Polymer-Liquid Interaction Parameters and Solubility Parameters, CRC Press, Boca Raton, FL, 1990. 9. Benjamin, S., Carr , C., and Wallbridge, D.J., Self-stratifying coatings for metallic substrates, Prog. Org. Coat., 28(3), 197–207, 1996. 10. Rasmussen, D. and Wahlström, E., HSP — solubility parameters: a tool for de velopment of ne w products — modelling of the solubility of binders in pure and used solv ents, Surf. Coat. Int., 77(8), 323–333, 1994. 11. Zellers, E.T., Three-dimensional solubility parameters and chemical protecti ve clothing permeation. I. Modeling the solubility of or ganic solvents in Viton® gloves, J. Appl. Polym. Sci., 50, 513–530, 1993. 12. Zellers, E.T . and Zhang, G.-Z., Three-dimensional solubility parameters and chemical protecti ve clothing permeation. II. Modeling diffusion coefficients, breakthrough times, and steady-state perme ation rates of or ganic solvents in Viton® gloves, J. Appl. Polym. Sci., 50, 531–540, 1993. 13. Zellers, E.T., Anna, D.H., Sulewski, R., and Wei, X., Critical analysis of the graphical determination of Hansen’s solubility parameters for lightly crosslink ed polymers, J. Appl. Polym. Sci., 62, 2069–2080, 1996. 14. Zellers, E.T., Anna, D.H., Sule wski, R., and Wei, X., Impro ved methods for the determination of Hansen’s solubility parameters and the estimation of solv ent uptake for lightly crosslinked polymers, J. Appl. Polym. Sci., 62, 2081–2096, 1996. 15. Luciani, A., Champagne, M.F., and Utracki, L.A., Interfacial tension in polymer blends. P art 1: Theory, Polym. Networks Blends, 6(1), 41–50, 1996. 16. Luciani, A., Champagne, M.F ., and Utracki, L.A., Interf acial tension in polymer blends. P art 2: Measurements, Polym. Networks Blends, 6(2), 51–62, 1996. 17. Fuchs, O., Solv ents and non-solv ents for polymers, in Polymer Handbook, 3rd ed., Brandrup, J. and Immergut, E.H., Eds., Wiley-Interscience, New York, 1989, p. VII/385. 18. Hansen, C.M., The universality of the solubility parameter , Ind. Eng. Chem. Prod. Res. Dev., 8(1), 2–11, 1969. 19. Lieberman, R.B. and Barbe, P .C., Polypropylene polymers, in Encyclopedia of Polymer Science and Engineering, 2nd ed., Vol. 13, Mark, H.F ., Bikales, N.M., Ov erberger, C.G., Menges, G., and Kroschwitz, J.I., Eds., Wiley-Interscience, New York, 1988, pp. 482–483. 20. Evans, K.M. and Hardy , J.K., Predicting solubility and permeation properties of or ganic solvents in Viton glove material using Hansen’s solubility parameters, J. Appl. Polym. Sci., 93, 2688–2698, 2004. 21. Nielsen, T.B. and Hansen, C.M., Elastomer swelling and Hansen solubility parameters, Polym. Testing, 24, 1054–1061, 2005. 22. Wessling, R.A., The solubility of poly(vinylidine chloride), J. Appl. Polym. Sci., 14, 1531–1545, 1970. 23. Wyzgoski, M.G., The role of solubility in stress cracking of n ylon 6,6, in Macromolecular Solutions, Seymour, R.B. and Stahl, G.A., Eds., Per gamon Press, New York, 1982, pp. 41–60. 24. Mai, Y.-W., Environmental stress cracking of glassy polymers and solubility parameters, J. Mater. Sci., 21, 904–916, 1986.

S.. J. Impro ved method to calculate Hansen solubility parameters of a polymer .Methods of Characterization — Polymers 111 25. Testing coatings: 6. Org. Appl.A. Prediction of Flory-Huggins interaction parameters from intrinsic viscosities. 367–372.L. Segarceanu. Polym. and Leca. Ind.. Weissman.D. Appl. Athey. 353–362. 473–478. 29. C. Dev. H... and Rudin. 26. 1982. 307–310.. M. 1982.. Polym. Res. R. 27. Van Dyk. J. 2675–2682. 1993. Prod. White. O.W. D.. Sci. and Kambour .. Solubility.M. 31(4). R. Kok. 28.. Solubility parameter determination. 24(3). Resistance of a polyetherimide to en vironmental stress crazing and cracking.. Frisch. 1985. and Wu. 5. and solubility parameters. 1997. Eng. Sci.T. 27... Prog. S. Coat. . Coat. J... solvency. 27. J. A.R. Chem. Eur.P.


it is not surprising that v arious correlations of cohesion parameters and surface phenomena can be found.. plastics. The dewetting behavior (wetting tension test) of the liquids must also be the same. Hansen solubility parameters in a total characterization of surf ace energy. The orientation of adsorbed molecules is a significant added e fect that must also be considered in man y cases.2 The various treatments and correlations in the literature will not be explicitly dealt with here.of Characterization 6 Methods — Surfaces Charles M. i. A complete match of surface energies of two surfaces requires that exactly the same liquids (in a lar ger number of well-chosen test liquids) spontaneously spread on both surfaces. Cohesion parameters (solubility parameters) can be used with full theoretical justifi cation to characterize man y surf aces. coatings. In this chapter .1) 113 . Solubility as such does not necessarily enter into the ener getics of interf acial phenomena. Hansen ABSTRACT Relations between cohesion parameters and surface energy parameters and their practical significanc are discussed. 1 As molecular surf ace-to-surface contacts control both solution phenomena and surface phenomena. including substrates. in that the same liquids should not retract when applied to the surf aces as films INTRODUCTION Interfacial free ener gy and adhesion properties result from intermolecular forces. This chapter will emphasize methods of surface characterization using HSP. Important molecular relations between a binder in a coating or adhesi ve and its surroundings then become ob vious. solubility parameters are called cohesion (energy) parameters and refer more specifically to HS . This correlation with surface tension had been long lost in an internal report to members of the Danish Paint and Printing Ink Research Laboratory in 1967. Use of cohesion parameters. pigment and fille surfaces. The “like dissolves like” concept is e xtended and applied as “lik e seeks lik e” (self-assembly). b ut the energy characteristics of surf aces can still be correlated with HSP .. clearly shows how the single point concepts of the (Zisman) critical surf ace tension and the wetting tension fit into a la ger energy concept. other than those directly related to Hansen solubility parameters (HSP).0688V 1/3[δD2 + k( δP2 + δH2)] (6. HANSEN SOLUBILITY PARAMETER CORRELATIONS WITH SURFACE TENSION (SURFACE FREE ENERGY) Skaarup w as the first to establish a correlation between liquid sur ace tension and HSP . 3. as well as in an abstract for a presentation to the Nordic Chemical Congress in 1968. etc. This idea has been well e xplored and dealt with elsewhere. It has been recognized for man y years that molecules interact by (molecular) surf ace to (molecular) surf ace contacts to enable solutions to be formed. in addition to the binder or polymer used in a gi ven product.e.4 γ = 0.

It is this diference between atomic and molecular interactions that is basic to the entire discussion of similarity between HSP and the Prigogine corresponding states theory in Chapter 2. assuming that the molecules tend to occup y the corners of re gular octahedra.6 With the exception of aliphatic alcohols and alkali halides. for xample. and h ydrogen bonding (electron interchange) interactions. They also ha ve the same coef ficient when solubility is correlated (see Chapter 1. h wever. Equation 2. related both to b ulk and surf ace phenomena.4 or Chapter 7. Table 7.6). The reason for this is the molecular orientation in the specific interaction derived from permanent dipole–permanent dipole and h ydrogen bonding (electron interchange) interactions. Beerbower independently published essentially the same type of correlation in 1971. This causes local dipoles and attraction among atoms. and 10.7147 in the empirical correlation. This includes dispersion. It is interesting to note thatδP and δH have the same coef ficient in the surace tension correlations. and k is a constant depending on the liquids in volved. however. v alues in dyn/cm are numerically equal to those in mN/m. This k was reported as 0.This is a completely different type of interaction and requires a dif ferent coefficient in the correlations.2. permanent dipole–permanent dipole. which sho ws that the geometric mean is valid for estimating interactions between dissimilar liquids. Ho wever.9 o Chapter 2. The dispersion or London forces arise because of electrons rotating around a positi ve atomic nucleus. The units for the cohesion parameters are (cal/cm 3)1/2. METHOD TO EVALUATE THE COHESION ENERGY PARAMETERS FOR SURFACES One can determine the cohesion parameters for surf aces by observing whether or not spontaneous spreading is found for a series of widely dif ferent liquids.3 for n-alk yl benzenes. Table 5. The contact angle need not necessarily be measured in this simplified procedure.2. Beerbo wer found γ = 0.0715V 1/3[δD2 + 0. The correlations presented by K oenhen and Smolders 7 are also rele vant to estimating surf ace tension from HSP. It is strongly suggested that none of the liquids be a mixture. In fact. Droplets of each of the liquids are applied to the surf ace and one simply observes what happens. there is an adv ancing contact angle and the cohesion ener gy/surface energy of the liquid is (significantly) higher than that of the sur ace. there is presumed . as this introduces an additional f actor into the evaluations.114 Hansen Solubility Parameters: A User’s Handbook γ is the surface tension. Equation 1. 5. The author has never explored them in detail. If a droplet remains as a droplet.2) where γ is the surf ace tension.7152 by a mathematical analysis in which the number of nearest neighbors lost in surf ace formation w as considered. If spontaneous spreading is found. Contact angles have generally been found to increase for greater dif ferences in cohesion parameters between the surface and liquid 9 (see also Figure 6. The constant w as separately deri ved as being equal to 0.632( δP2 + δH2)] (6. and those of the surf ace tension are dyn/cm in both Equation 6.5). The previous discussion and the contents of Chapter 1 and Chapter 2 clearly indicate that the author is not in full agreement with this viewpoint. The liquids in the series often used by the author are indicated in Chapter 5. 0.265 for normal alcohols. The finding that the polar and ydrogen bonding (electron interchange) ef fects require the same coef ficient for both ulk and surf ace correlations suggests that the net effects of the (often mentioned) unsymmetrical nature of h ydrogen bonding are no dif ferent from the net ef fects occurring with permanent dipole–permanent dipole interactions.1 and Equation 6.8 for se veral homologous series. so they are not discussed here. The lack of specific consideration that ydrogen bonding is an unsymmetrical inter action led Erbil 8 to state that HSP has limited theoretical justification. The liquids used in standard solubility parameter determinations are suggested for this type of surf ace characterization. it appears that this book presents massi ve experimental evidence. The constant w as actually found to be 0.

1 sho ws a complete ener gy description for an epoxy polymer surf ace 10. Receding contact angles will generally increase as one progresses to liquids with still higher HSP. The Hansen polar and hydrogen bonding parameters δP and δH are used to report the data. Further explanation of these parameters themselves can be found in Chapter 1. E. Spontaneous spreading is more related to adhesion since such liquids w ant to cover the surface spontaneously. 1983. (From Hansen. Units are MPa1/2. δD.Methods of Characterization — Surfaces 115 15 18 * Wetting Tension – ADV = O 10 δP 24 * 25 20 * 19 * C 17 * 1 * B A 11 * 12 * 13 * 6 5 * 4 3 * 2 * 26 21 * 22 14 15 * 5 27 * 23 42 7 8 16 9 5 10 δH 15 20 FIGURE 6. Liquids that are found in the re gion above the upper curve will retract when applied as films.1 HSP surface characterization of an epoxy surf ace showing regions of spontaneous spreading of applied droplets (A). If a given liquid does not spontaneously spread. This test determines whether or not there is a receding contact angle under the gi ven conditions. the f act of spontaneous spreading for a given liquid does not mean that its HSP are identical with those of the surf ace being tested..1 shows two curves that are concave toward the origin. As discussed in greater detail later . The wetting tension test uses . whereas liquids applied as films will remain as films This region deserves more attention in future research. it can be spread mechanically as a film and be obser ed to see whether it retracts.1 is a re gion where liquids applied as droplets will remain as droplets. Figure 6. With permission. and Wallström. This can be done according to ASTM D 2578-84 or ISO 8296:1987 (E). has not been specifically accounted for in the t o-dimensional figure Figure 6. C. The apparent similarity may be misleading. for e xample. Note that thi characterization may not be valid for all epoxy surfaces. and d wetting of applied films (C). J. A test method to determine this is found in the ASTM and ISO standards gi ven previously. Belo w the line one finds tha liquids applied as droplets will spontaneously spread. 281. The lower of these divides the test liquids into tw o groups based on spontaneous spreading or not. The circular lines can be considered as portraying portions of HSP spheres.11 based on the testing procedure described previously. lack of dewetting of applied films (B). 15[3/4]. Intermediate between the two curves in Figure 6. The energy properties of these liquids are not as close to those of the surf ace as are the ener gy properties of the liquids that spontaneously spread. b ut the third Hansen parameter. except that one uses a lar ge number of pure liquids instead of the liquid mixtures suggested in the standards..) to be some “similarity” in the ener gy properties of the liquid and the surf ace. Adhes.M.

whereas liquids with lo wer surface tensions than the critical v alue will spontaneously spread. .11 This is both impractical and impossible. for example. there is still a problem in simplifying these results for easier use and improved understanding. liquid surf ace tension and e xtrapolate to the limit where the contact angle is no longer present (see Figure 6. At the same time. Hydrophilic se gments can (perhaps) rotate toward a water droplet at some rate.2 for a plasticized polyvin yl chloride (PVC) and in Figure 6. for e xample. This is similar to the condition for a RED number equal to 1.1 it is almost in the middle of the region describing spontaneous spreading of the liquids. and until it is found. The polymer dissolves in some of them. The result of this is that the center of the normal HSP sphere for describing spontaneous spreading can be assigned sizable ne gative values. Liquids with higher surface tensions than this critical v alue allow measurement of a contact angle.116 Hansen Solubility Parameters: A User’s Handbook an e xternal force to spread the liquids.0 discussed in Chapter 1 and Chapter 2.13 The Zisman critical surf ace tension is determined by measuring the e xtent that af finity is lackin (contact angles) for a surf ace using pure liquids or liquid mixtures in a series. This is discussed more in Chapter 18. but in Figure 6. Nature reduces the free ener gy level of the surface by requiring he xane to cover the epoxy coating. The following discussion relates these to the HSP-type characterizations discussed earlier . and increase the hydrophilic nature of the surf ace accordingly. consequently . Measuring the critical surf ace tension has been and still will be a useful technique to better understand surf aces. and the de gree of swelling of the polymer in question is measured in those liquids which do not dissolve it fully. but it should be done with the following in mind. assigns the polymer solubility parameters corresponding to those of this boundary solvent. The surface tension of each of the liquids is kno wn. it sheds some light onto the true meaning of the critical surf ace tension. The limiting critical surf ace tension 12. for example. One can then plot cosine of the contact angle vs. One subsequently determines the solubility parameter of the polymer by e xtrapolating the de gree of swelling to infinit .4). The fact that the liquid with a surf ace tension just under the critical v alue spontaneously spreads is often tak en as an indication of high af finit . He xane is within the re gion of spontaneous spreading because it has a lo wer surface energy (surf ace tension) than the epoxy surf ace. This is dif ficult to understand and appears to be misunderstanding.13 has v ery little to do with the “best” solvent for the surface. In the mean time. This method should clearly ne ver be used to determine solubility parameters for polymers.3 for a polyethylene (PE). This extrapolation can be done by plotting 1/(de gree of swelling) vs. which corresponds to total solution. interest will still be focused onto the usual test method(s) for determining surf ace tensions based on the Zisman critical surf ace tension plots (lack of adv ancing contact angle) or by using the ASTM procedure for wetting tension (lack of receding contact angle). A CRITICAL VIEW OF THE CRITICAL SURFACE TENSIONS 12. known solubility parameters. A better method of handling this situation is still desired. Hexane. Hexane will not contrib ute to a “bite” into an epoxy coating for impro ving intercoat adhesion with a subsequent coating. solv ent composition (solubility parameter). One assumes that there is no better solv ent than this one and. This is exactly what one does when the critical surf ace tension is measured. after which the y may continue to remain as a film. One now focuses attention upon that liquid which (by e xtrapolation) just dissolves the polymer. Who would determine the solubility parameter for a polymer by the follo wing method? One makes up a series of liquids with dif ferent. Additional surface characterization plots for spontaneous spreading and wetting tension using HSP are included in Figure 6. The mobility of the surf ace layer(s) will play a role in the wetting tension test. As mentioned earlier . which are more desirable in most cases. does not dissolv e an epoxy polymer. It is more appropriately compared with a v ery poor solvent which can only marginally dissolv e a polymer . one must presumably refer to simple plots or other simple comparisons rather than to refined computer techniques.

C. Hansen and Wallström11 compared the critical surf ace tension plotting technique with one where a dif ference in HSP w as used instead of liquid surf ace tension. Adhes.4 and Figure 6. 280.M. E. A CRITICAL VIEW OF THE WETTING TENSION A region larger than that for spontaneous spreading will be found on a δP vs..1 to Figure 6. In practice. 15[3/4].3.2 HSP surface characterization of spontaneous spreading of applied droplets and wetting tension for applied films for plasticized polyvi yl chloride (PVC). Units are MPa1/2 (From Hansen.5..1 to Figure 6. Mixtures of formamide and eth ylene glycol monoethyl ether are usually used in practice for these measurements according to ASTM . With permission. 1983. This is explained by the fact that the cohesion parameter regions of the type shown in Figure 6.3 are not symmetrical around the zero axis. The corresponding total surf ace tension will v ary from intersection to intersection as mentioned earlier .Methods of Characterization — Surfaces 117 15 18 * 10 24 * δP 25 * 26 * 5 22 * 27 * 23 8 16 9 10 5 7 * 21 * 14 15 * * 13 * 6 5 12 * 20 * 19 * Wetting Tension 11 * 2 * C B A – ADV = O 17 * 1 * 4 * 3 * 10 δH 15 20 FIGURE 6. This comparison is made in Figure 6.) If we now consider the re gion for spontaneous spreading in Figure 6. δH plot when one plots data for those liquids that remain as films (do not break up or contract) when th y are applied as films. The individual liquid series used to determine the critical surf ace tension will intersect the cohesion parameter spontaneous spreading boundary at dif ferent points. it can be seen that the critical surf ace tension is a point on its boundary . and Wallström. Note that these characterizations may not be v alid for all PVC surfaces. one finds di ferent critical surface tensions for the same surf ace depending on which liquids (or liquid mixtures) are used. This type of experiment measures the wetting tension. One arrives at the same general conclusions from both types of plotting techniques. J.

. their plots must be the same. “A Critical View of the Critical Surface Tensions.1. 279. If tw o different surfaces are to have the same wetting tension beha vior.3 HSP surface characterization of spontaneous spreading of applied droplets and wetting tension for applied films for a polyet ylene (PE) surf ace. Examples include the work of adhesion on mercury. The results of the ASTM test are usually stated in terms of the surf ace tension of the liquid or liquid mixture which just stays intact as a film for 2 sec. J. 15[3/4]. frictional properties of untreated and treated polyeth ylene . A single point determination may not al ways be sufficient information an certainly neglects the complete picture possible from HSP considerations. Adhes. Note that these characterizations may not be v alid for all PE surfaces. One can also use the same liquids suggested earlier for cohesion parameter determinations and mak e a plot lik e that in Figure 6.M. E. It is hoped the reader no w better understands the total ener gy conte xt of the simple ASTM wetting tension measurements.118 Hansen Solubility Parameters: A User’s Handbook 15 18 * 10 δP 24 * 25 * 26 * 5 22 * 27 23 A 8 16 9 10 5 7 * B C 21 * 14 15 * * 13 * 6 5 12 * 20 * 19 * 11 * Wetting Tension 4 3 * 2 * 1 * – ADV = O 10 δH 15 20 FIGURE 6. This simple single point determination corresponds to determining a single point on the boundary of the HSP plot describing wetting tension for all liquids. Comments identical in principle to those included in the earlier section.” on measurement of the critical surf ace tension are also v alid here. With permission.. (From Hansen.) D 2578-84 or ISO 8296:1987 (E). and Wallström. ADDITIONAL HANSEN SOLUBILITY PARAMETER SURFACE CHARACTERIZATIONS AND COMPARISONS Beerbower14 has reported man y other correlations of surf ace phenomena with HSP . C. Units are MP a1/2. 1983.

.4 0. E.6 1 0. J. Adhes. 0. C. 282. and Wallström. (From Hansen. 15[3/4].M.. J.4.5 Critical HSP plot of cosine of the static adv ancing and receding contact angles vs..8 COS θ 3 2 0. Equation 1. and Wallström.) 1. liquid surface tension for low density polyethylene.M. (From Hansen. 1983. Adhes. 1983.9 for l w density polyeth ylene. With permission. E.4 0.8 COS θ 3 2 0. C.0 7 ADV.Methods of Characterization — Surfaces 119 1.m-1 40 50 60 FIGURE 6.6 1 0. The same data are used in Figure 6. With permission. 282. REC. The same data are used in Figure 6. the HSP difference as defined by Chapter 1..4 Zisman critical surface tension plot of cosine of the static advancing and receding contact angles vs.5.2 16 18 20 RA 22 24 26 28 FIGURE 6.2 20 30 γ mN. 15[3/4].) .0 7 ADV REC 0.

The test surface may be a polymer. e ven though recent progress and understanding in the area are much more adv anced. or other material. Results may be a fected when molecules in a surface can orient differently from their original state upon contact with a liquid. but the critical strains were somewhat higher on an a verage. and the Rehbinder ef fect — crushing strength of Al2O3 granules under various liquids. respectively. B. is not able to retard sedimentation where it may spontaneously spread.3. category A liquids gave ESC. etc. A significant ad antage in this testing method is that he xane. as discussed abo ve. silicones. spontaneous spreading and de wetting. a simple test of applying a droplet of liquid and possibly spreading it.).120 Hansen Solubility Parameters: A User’s Handbook for 2 and 5 min. NT POL Y 176.” This method is based on visual observation of droplets of known surface tension after they are applied to a test surface. it is appropriate to mention that thinking of the type described above has led to a Nordtest Method. The spreading surface tension is found to within ±1 mN/m by locating two liquids in a series where one of them spreads spontaneously and the other with a slightly higher surf ace tension does not. pressure sensitive adhesive. metal. The reason for this may be the e xtent (or depth) of the adsorption layer . which is important in these characterizations and not just spontaneous spreading. has been established through measurements of en vironmental stress cracking (ESC). SELF-STRATIFYING COATINGS A newer development in the coatings industry is to apply a single coat of paint which separates by itself into a primer and topcoat. The preferred set of liquids is made with ethanol and w ater with a dif ference of 2 mN/m between them. Adsorption of given liquids onto the particle or fiber sur ace slows the sedimentation rate. for e xample. with H2S2O7. The HSP differences outweighed the expected spreading based on surface tension differences. if it does not do so itself. The characterizations using standard HSP procedures indicate it is truly high affinity for the sur ace. A large number of test liquids in each category were used to e valuate the critical strain required for ESC. The test method used is to determine sedimentation rates for the materials of interest in the same lar ge number of solv ents traditionally used in HSP studies. An indirect correlation between HSP and the phenomena discussed above. and indeed some (fine particles with rather high densities suspend for years in or ganic liquids with rather modest densities. c yclic olefinic copolyme (COC). These include characterizations of the surf aces of pigments. and orientation (ceiling in a tunnel. against this to determine the presence of contamination. the Joffé effect — effect of liquid immersion on fracture strength of soda-lime glass. is a rapid w ay to assess a potential problem. Surf aces of man y different geometries (from 4 μm diameter wire to ships being painted). for e xample.17 Misev has also discussed formulation of this type of product using . and C as sho wn in Figure 6. wherever they may be.15 Additional surface characterizations using HSP are reported in Chapter 7. Before lea ving this section. Category C liquids could also gi ve ESC in some cases. Both ganic and inor ganic materials have been characterized.) ha ve been tested with remarkable success using this simple test. with water (see the discussion in Chapter 18).16 As discussed in Chapter 14. and fibers. Although these observ ations should not replace testing. or both. as well as whether the adsorption occurs at specific sites. and acrylonitrile/b utadiene/styrene (ABS) terpolymer could be described in terms of the regions A. etc. All category B liquids also g ave ESC. Hexane is not an isolated example of this beha vior. A special issue of the journal Progress in Organic Coatings was devoted to this type of coating. The polymers polycarbonate (PC). inside pipes.1 to Figure 6. The usual procedure is to assign a v alue to a clean(ed) surface and then compare test surfaces. ESC correlates with the strain and the HSP and molecular size and shape of the cracking agent. “Spreading Surf ace Tension by the Applied Droplet Method. 14 It appears that these f actors should still ha ve some consideration. states of contamination (from clean for internal medical use to contamination with oil. fillers. Beerbo wer has also brought cohesion parameters into the discussion of wear and boundary lubrication. It w as found that in e very case tested. nhexane was a category C liquid for some of these polymers in spite of its lo w surface tension.

The binder with the lowest energy (surface tension/cohesion parameters) will naturally migrate toward the low energy air interf ace and. including polymer molecular weight. CRC Press. There are a number of other f actors which are important for the process. The interface between the topcoat and primer is formed from an otherwise homogeneous system. This discussion is included because it once more demonstrates ho w cohesion parameters are coupled with surface energy and also to interf acial energy. FL. it is widely known among those who work with partially compatible polymers that this is indeed the case. therefore. (From Birdi. Without going into greater detail. Ed. so it will be located in the common region to the spheres portrayed. A solvent is required which dissolves both polymers.) . b ut these will not be discussed here. The polymer with HSP nearest the origin will be the topcoat.6 shows the principles involved for selecting the solv ent which can make these work.Methods of Characterization — Surfaces 121 HSP concepts. 1997.6 Sketch of HSP principles used to formulate a self-stratifying coating from an initially homogeneous solution (see discussion in te xt). Handbook of Surface and Colloid Chemistry. rate of solv ent e vaporation. The solv ent must just dissolve the binders such that the y become incompatible when it be gins to e vaporate. The previous considerations lead to the e xpectation that the magnitude of the interf acial surface tension between tw o incompatible polymers is closely related to the dif ference in their cohesion parameters. It is clear that selection of the optimum solv ent for this process of designed generation of an interf ace is aided by systematic use of HSP . Figure 6. Boca Raton. This is a prime e xample of selfassembly where proper formulation can be aided by the concepts discussed abo ve. etc. PRIMER POLAR PARAMETER PARAMETERS REQUIRED FOR COMMON SOLVENT TOP COAT (LOWEST ENERGY) HYDROGEN BONDING SOLUBILITY PARAMETER FIGURE 6. 324.. this determines which of the binders mak es up the topcoat.S.20 See also Chapter 9 where partial compatibility in bitumen (asphalt) is discussed.. Mutual solubility of tw o polymers is promoted when the solvent favors the polymer which is most dif ficult to dissol e.19.21 This is usually the one with the higher molecular weight.. p. K. as it has the lo wer (surface or cohesion) energy of the two.18 The separation of the binders into primer and topcoat must occur while the coating is still liquid enough to allo w the necessary transport processes to occur .

This allo ws mutual interactions to be inferred by comparisons of which materials are similar and which materials are different in terms of their HSP . Technol. Chem.. and alcohol. Surf ace Tension and 3-D Solubility P arameters (Ov erfladespænding og 3-D Opløse lighedsparametre). Skaarup. Any dif ferences in this spontaneous spreading or wetting tension behavior can be interpreted as being a dif ference in ph ysical similarity . allo w insight as to ho w the single point measurements fit into th overall ener gy picture for the product. The discussion of forming good anchors on pigments and other surf aces found in Chapter 8 is also rele vant to the present discussion. The dif ferences in the behavior of liquid droplets or films that are obser ed may suggest which steps can be tak en to minimize differences. The previous discussion suggests that physical similarity can be obtained when tw o criteria are met.e. Nordiske Kemikermøde. 1985. København. 2. Beerbower. Ed. Guidelines for systematically changing the af finities o surfaces can also be obtained from HSP concepts. Sci. Colloid Interface Sci. A. in Kirk-Othmer Encyclopedia of Chemical Technology. HSP. the ph ysical similarity (same HSP) of the tw o interfaces being joined must be as close as possible. i. 1968 (in Danish). Flory. Applications of solubility parameters and other cohesion ener gy parameters. 1953. pp. if this is required. but also the HSP for the v arious surfaces which these encounter . 1971.1) for more precise comparisons.. and primary amine groups characteristically increase δH. J. The critical surf ace tension and wetting tension are single point determinations...M. Pure Appl. 57(7). for polymers and solvents it contains. Reference can be made to the table of group contrib utions in Chapter 1 (T able 1. C. Barton. Polym. 54 (TM 2-67). Should one add aliphatic segments to reduce the polar and hydrogen bonding contrib utions? Should alcohol and/or acid groups be incorporated to increase the hydrogen bonding in the system? This type of approach can be used to establish guidelines for action relative to each of the HSP parameters. 1971. Similar materials in this conte xt have similar HSP re gardless of differences in composition. 126–132. 5. Suppl. A. The surf aces of substrates.122 Hansen Solubility Parameters: A User’s Handbook MAXIMIZING PHYSICAL ADHESION If one wishes to maximize ph ysical adhesion. CONCLUSION Greater insight into the mak eup of a product is possible when one not only kno ws the cohesion parameters. 4. Hansen... K.. 889–910. fibers. P. Ne w York. Solubility parameters. Lakog Farveindustriens Forskningslaboratorium. 2nd ed. Cornell University Press. C. København. 35. plastics. fillers..M. Vol. K... The second criterion is that exactly the same liquids maintain films when spread (ASTM method for wetting tension) on eac of the surf aces to be joined. and oth materials can also be characterized by HSP (see Chapter 5 and Chapter 7). Cohesion parameters allow a more complete characterization of surfaces than do these single point measurements and. and Beerbo wer. 6. Skaarup. The first criterion is that xactly the same liquids spontaneously spread on each of the surf aces to be joined. and Hansen. Interscience. New York. Standen. A. Aromatic character and halogens other than fluorin characteristically increase δD. acid. A..M..J. Rapport No. REFERENCES 1.F. as such anchors can also be used to enhance adhesion. nitro and phosphate groups characteristically increaseδP. pigments. 905–912. Both the spontaneous spreading re gion and the wetting tension re gion on HSP plots for tw o different surfaces must be identical if the y are to ha ve identical o verall surface characteristics. The Three-Dimensional Solubility Parameter and Its Use (Det Tredimensionale Opløselighedsparametersystem og dets Anvendelse). 1967 (in Danish).. Surface free energy: a new relationship to bulk energies. at the same time. . 3. Principles of Polymer Chemistry.

1996... 1163–1179. L. A. P art 1: Theory. pp.. Birdi. Washington. 41. 1972. Advances in Chemistry Series No. Technol. Verlag Chemie. 42(550). Surf ace wetting and the prediction of en vironmental stress cracking (ESC) in polymers. Boundary Lubrication — Scientific and Technical Applications Forecast. The determination of solubility parameters of solv ents and polymers by means of correlation with other physical quantities. J. 21. 1996. Appl. Coat. 89. 51–62. and Hansen. Adhes. Congress Book.. 48–56. 1974. 23–28... July 1996.. CRC Press. 8. 15..M. 15(3/4). Thermodynamic analysis of phase separation in self-stratifying coatings — solubility parameters approach. Department of the Army.. Dev.. T. 19. D.. Weinheim/Bergstrasse. AD747336. Coat.. 218–225. Emaux et Encres d’Imprimerie de l’Europe Continentale. Res.. Interf acial tension in polymer blends.A.. Koenhen.. 17. 44(564). J. and Wallström. 13(4). Paint Technol. Eng. Org. A. Friction at the Atomic Scale. A. Luciani. 43.F .. M. On application of the three dimensional solubility parameter to the prediction of mutual solubility and compatibility. FL. 20. Hansen. 1974. Champagne. 513–516. Boca Raton. P art 2: Measurements. 28(3). Prod. and Utracki. J. Misev.F. C. 132–138.S. Luciani.C. C. 1970. Washington. 1997. Surface energetics of wetting. 9. D. 18. Degradation Stability. 275(4). XII Federation d’Associations de Techniciens des Industries des Peintures. Krim.E. Sci.A. Networks Blends. 6(1).M. Surface tension of polymers..A. 2005. Relation of the equilibrium contact angle to liquid and solid constitution.A.Y. 660–664. Zisman.A. 275–286. C. Characterization of liquids by spreading liquids. Polym. Ed. Beerbower. Office of the Chief of Research and D velopment. M. Ind. chap. Gould. 1. in Handbook of Surface and Colloid Chemistry.M. 13. E. H. 6(2).. C. R. Paint Technol.. C.M. 10. 1991. Hansen. 19.B. Interfacial tension in polymer blends. Special issue de voted to self-stratifying coatings. Zisman.A. Surface effects in coatings processes. On the use of cohesion parameters to characterize surf aces. L. Nielsen. W. 16. J. Vernis. 63(795). and Pierce. 1975. spreading. and Utracki. Champagne. 1964. T. D. 265–312. P .N. Prog. W. Polym. Scientific American. 11. Färg och Lack.F . 14. 1972.C. Erbil. 91–99. . and Smolders. K. 1967. October 1996.Methods of Characterization — Surfaces 123 7. in Contact Angle....... Hansen.. 12. Networks Blends. pp. J. C.. Polym. Polym. and adhesion. 1983. Hansen. Wettability and Adhesion. American Chemical Society. 13(6).. 41–50...M.. Ed. Chem. J.


Cohesion parameter data are gi ven for some ne wer pigments. Hansen ABSTRACT Cohesion parameters for pigments. Shareef et al.2B. Data from such evaluations can be computer-processed to assign Hansen cohesion parameters (HSP) to the material in question. and Fibers Charles M. The absolute sedimentation rates are generally not of primary interest. 3 have also characterized pigment surfaces. In the future. such that a relative sedimentation rate can be used for comparisons. fillers. Inorganic fibers h ve also been characterized.6 Each se gment of such molecules requires its o wn HSP. more systematic selection of dispersion aids should be possible. It has been sho wn by calculation that hemicelluloses act lik e surface-active agents. They reflect molecule–molecule interactions whethe at surfaces or in b ulk. and fibers can often be valuated by observation of the suspension and/or sedimentation beha vior of particulate matter in dif ferent liquids. 5 All of these characterizations ag ain confirm th universality possible with these parameters. judging from the relati vely small number of publications appearing on the topic.of Characterization 7 Methods for Pigments. Table 7. fillers and a carbon fiber to demonstrate the principles INTRODUCTION The possibilities offered by cohesion parameter characterization of pigments. and δH parameters (HSP) and a characteristic radius of interaction e xactly analogous to the polymer characterizations discussed in Chapter 2 and Chapter 5. Data should be interpreted by accounting for differences in the densities and viscosities of the test liquids. Hansen and Beerbower have touched on this topic. Those liquids with stronger interaction can suspend finer fractions of solids indefinitely or reta sedimentation.1. 125 . and fibers ve not been generally recognized.1. including metal oxides. These data together with some more recent pigment characterizations are included in Table 7. Fillers. The discussion in Chapter 15 for the interactions within cell w alls in wood demonstrates how this could be done. with some se gments seeking lo werenergy lignin regions and some segments (those with alcohol groups) orienting to ward the higherenergy cellulose. Gardon and Teas4 clearly sho wed the dif ferences between zinc oxides treated and untreated with or ganic phosphate using a cohesion parameter characterization. fillers. Pigments and a few fillers were characterized in some of the author s first publication dealing with the solubility parameter . compared with the other liquids. as these can also be described with the same energy parameters. These characterizations are based on relatively stronger adsorption by some of the liquids compared with others. and Table 7. δP.2A.2 These were gi ven δD.

it will float.7 Ro 17.1 HSP Correlations for Older Inorganic Pigments1.2 14. This can vary from sample to sample.3 12. Danish Technical Press.126 Hansen Solubility Parameters: A User’s Handbook TABLE 7. C. t s .0 8. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient. et al. FILLER.1. compared with the other test solvents. K.0 14. February 1986.3 16.8 METHODS TO CHARACTERIZE PIGMENT. With permission. after compensating for differences in density and viscosity.M. J. A characterization is less certain when there are only 4 or 5 good liquids out of the perhaps 40 to 45 tested. Some pigments ha ve portions that suspend for years in spite of lar ge-density dif ferences and relatively low-solvent viscosity. From Hansen. in a direction perpendicular to the incidence of a laser light. Coat. A corrected relative sedimentation time (RST) can be found by modifying the sedimentation time. and some initial experimentation is usually advisable. as some samples seem to coat out rapidly on glass surf aces. When the solid has a lo wer density than the test liquid.1 19.. The sedimentation rate is most easily e xpressed as the time at which the amount of particles.8 δH 19. and then sedimentation or lack of the same is observed.6 7. Observ ations can be made visually .8 10.9 6.2 4. 58(733). 35–44.3 13. Copenhagen. A visual observation is required in an y event. If the pigment or filler particle siz is large — say o ver 5 µm — the surf ace effects become less significant compared with a sampl where the particle size is only 0.0 20. Satisfactory results from this type of measurement require some experience regarding what to look for .5 11. b From Shareef. Technol.5 11. With permission.M.1 17.4 7..A. Problems arise when the pigments are soluble enough to color the liquid such that sedimentation cannot be e valuated. a δD 24. although this depends some what on which liquids are in volved.6 6.7 16.1 16.9 16.2 and Metal Oxides3 Material Kronos® RN57 TiO2a Aluminum pulver lack 80 a Red iron oxide a Synthetic red iron oxide b Synthetic yellow iron oxide b Zinc oxide Note: Units are MP a1/2.2 9. b ut the term sedimentation will be retained here for both sedimentation and floating. or perhaps instrumentally .9 11. Doctoral disse tation. 1967. Rates of floati have also been noted.3 8.3 15. at a gi ven point in the test tubes. has f allen to some small amount.01 µm.2 δP 14.6 12.5 15. “Good” in this context means suspension of particulates is prolonged significantl . The amounts of solid sample added to the liquids can v ary depending on the sample in question.. perhaps zero. Sedimentation rates have still been used successfully in some of these cases.2 Small amounts of the pigments are shak en in test tubes with a gi ven volume of liquid (10 ml) of each of the test solv ents.1 12. AND FIBER SURFACES The cohesion parameter (HSP) approach to characterizing surf aces gained impetus by experiments where the suspension of fine particles in pigment p wders was used to characterize 25 organic and inorganic pigment surfaces. The larger particle size pigments and fillers may sediment ery rapidly.0 10.

RST = t s(ρp – ρs)/η (7. (C. 5.I.M.I..I. Frankfurt (M). Køge Chemical Works. 9500. Køge. Frankfurt (M). With permission. inde x). C. Heliogenblau B Pulv er. Pigment Blue 3 (C. 1967.I. Køge.I. (1949). 19. W. 69801).I. Fanalrosa G Supra Pulv er. (C. Inde x-pigment mixture).Methods of Characterization for Pigments. Køge Chemical Works. C. Source: From Hansen. Pigment Violet 23. Pigment Red 81 (C. 15850) (Ca Salt). B ASF. C. and Fibers 127 TABLE 7. PigmentYellow 81.I. 12370).I. Pigment Green 7. Phthalocyanine Blue. HSP Results are Given in Table 7. 6. Danish Technical Press. Køge. B6. C. 3. Pigment Blue 15 (C. FILLERS. 25. (C. 13. Smith Ltd. . Description TiO2. F arbwerke Hoechst.I. Frankfurt (M). 12075). 45160). Køge Chemical Works. DISCUSSION — PIGMENTS.I. Køge Chemical Works. There are apparently no publications indicating a systematic modification of pigment sur aces to achieve a given set of cohesion parameters.1) ρp and ρs are densities of particle and test liquid. Germany. Heliogenblau LG.I. Refl x Blau TBK Ext. Characterizations based on these techniques tend to place emphasis on the nature of the surf aces for the smaller -particle-size fractions. German y. C. Isol Benzidene Yellow GA-PR. C. Pigment Orange 13 (C.4. 74260). 14. 21. F arbwerke Hoechst. I. 24. 10. 11. du Pont de Nemours and Co. Denmark. 7522. Pigment Red 48 (C. German y.I. De gussa. F arbwerke Hoechst. F arbwerke Hoechst. German y.. Isol Ruby BKS 7520. 74100). or fiber is most beneficial when the pigment su ace and the binder in question ha ve the same cohesion parameters. Copenhagen. Pigment Red 112. 8. Frankfurt (M). Permanent Bordeaux FRR. Farbwerke Hoechst.I. Pigment Red 12 (C. B ASF. German y. Isolbonared Nr. respecti vely.I. AND FIBERS It can be reasoned that a pigment. I. Denmark. German y. Pigment Blue 16 (C. Eckart-W erke. Denmark. and T.I. Doctora dissertation. C. A. Denmark. Denmark. Carbon Black. Pigment Yellow 3. C. London. Germany. Titan Co. (No C. 4. Kronos RN 57. German y. Ludwigshafen.I.I.I. Permanent Red. C. Vat Blue 4 (C. Hansagelb 10 G.I. Isol Benzidine Yellow G 2537. 21110). Pigment Orange 5. Fillers. Køge. German y.I.3.I.I.I. The Three Dimensional Solubility P arameter and Solv ent Diffusion Coefficient. 74160).I. Frankfurt (M). F arbwerke Hoechst. Peerless Carbon Black Isol Fast Yellow IO GX 2505. (C. Pigment Yellow 3 (C. C. 2. Frankfurt (M). Imperial Chemical Industries. (No C. Heliogengrün GN. 16. B ASF.I.I. Permanent Orange G. German y. Pulv er. Køge Chemical Works.I.2B Pigment 1. A prolonged RST implies greater adsorption of the gi ven solvent onto the surf ace in question. German y. 12505). Permanent Violet RL Supra. Ludwigshafen.. Aluminum Pulver Lack 80. An e xception is the first one in the tabl where the analysis is based on only three good solv ents that were able to e xtend sedimentation significantly relat ve to the other solv ents tested. Germany. and η is the liquid viscosity . E.I. Permanentgelb H 10 G. Køge Chemical Works. 20. Køge. 17. 21090). 851 Fürth/Bayern.I. Køge.I. FGR Extra Pulver. C. Ludwigshafen. 42140). Po wder. C. 11710). Red Iron Oxide. 22. fille . B ASF. C. C. Pigment Red 57 (C. 9..2A List of Pigments Studied.I. J. Monolite Fast Blue 3 RS. Denmark. 7. 18.I. 23. A/S. These data indicate that their respecti ve surf aces are essentially identical. 12. Frederikstad. C. Farbwerke Hoechst. Ludwigshafen. Frankfurt (M). An example of a data sheet used for such studies is included in Table 7. Printe x V (5519-1). Norw ay.I. C. 15865) (MnSalt). Isol Fast Red 2G 2516. C. 12385). Pigment Yellow 12. Frankfurt (M). Pigment Yellow 12 (C. German y. The characterizations for some or ganic pigments are gi ven in Table 7. C. 15. Brillfast Sky Blue 3862.

3 10. 19. 16. 24. 2.5 7.3 9.5 11.1 10.3 9. 15. 4.0 7.1 10.5 2.2 10.0 4.5 3.0 3.0 6.2 11. As these solvents dissolve the most commonly used binders.5 13.8 6.6 11.0 6.5 Comments Suspension Few suspending solv ents Few suspending solv ents Suspension Color only Mixed color and suspension Suspension Color only Color only Suspension Primarily suspension Suspension Color only Mixed color and suspension Mixed color and suspension Suspension Color only Color only Primarily color Suspension Suspension Suspension Suspension Suspension Suspension Note: Units are (cal/cm3)1/2.5 10.0 7. 25.7 10.2 14. one can conclude that the common binders will adsorb readily onto these pigment surf aces.5 5. provided the solvent is not so good for the binder that it can remo ve the binder from the pigment surf ace. 10. 21.2 15.3 3.2 13.0 12.5 4.0 8.0 6.5 10.0 4. Schröder7 (BASF) confirms that the optimum polymer adsorption will be found when the binde and pigment surface have the same HSP. Source: From Hansen.4 2.8 9.9 3.2 3. 17.9 13.2 10.0 7.8 13.3 6.7 3.1 3.1 3.7 10.0 5. 12.6 9.6 6. He indicates that the solv ent should be v ery poor for the pigment and located on the boundary re gion for the binder . as well as for conditions where the solv ent has lower HSP than the pigment.0 δP 7.3 9.7 10. 9.5 4.5 5. Doctoral dissertation.0 3.4 7.9 5.5 3.2A Pigment 1. 6.1 4.3 2. With permission.1 δD 11.0 5. with the solv ent and binder on opposite sides of the pigment.6 2.4 9.0 4. The solv ents most frequently appearing as good for adsorption onto these surf aces include several chlorinated solvents.2 4.8 10.5 2. 11.8 10.5 7.8 1. These results suggest that pigment manufacturers have essentially arrived at the same result — a surf ace ener gy compatible with a wide v ariety of currently used binders.9 6.0 4.7 9.3 5. 20.8 5.0 2. and tetrahydrofuran.6 3.7 δH 9.2 3.5 13.0 8.8 3..0 4.0 7.8 10.8 3.8 9. 7.9 10. 18.7 3. 8. He prefers the pigment to ha ve HSP values placing it intermediate between the solv ent and binder . This will give a good result. C. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient.1 8. 13. .5 9.128 Hansen Solubility Parameters: A User’s Handbook TABLE 7.0 3.6 5. 23.0 12. 14.9 3. 22.4 2. This is suggested for conditions where the solv ent has higher HSP than the pigment.5 6.8 2.6 9.1 9. 5.0 4.7 13. 1967.3 5.M. toluene.3 8. means the composite v ehicle has parameters v ery closely matching those of the pigment.5 10.2 3.2B Characteristic Parameters for Various Pigments Given in Table 7.5 6. This situation. 3. δt 16.2 5.0 13. Danish Technical Press.0 10.4 10.5 2.5 5. Copenhagen.4 3.6 3.2 13.8 10.3 11.2 5.2 4.5 6.3 Ro 8.5 6.

72 0.30 0.59 1.47 Viscosity 20°C 0.87 1.98 1.05 1.31 107.66 0.55 0.04 1.84 20.20 1.65 4.55 0.72 3.05 1.29 1.00 3.92 0.99 1.82 0.62 0.00 68.13 0.23 0.0 1.59 1.27 35.80 0.95 0.81 0.Methods of Characterization for Pigments.10 1.D s From To Relative Sedimentation Time (RST) From To .25 1.91 0.89 0.99 0.92 0.03 0.3 Sedimentation Study Date Sample: Density of sample (Dp): Reference No.60 0.74 1.37 1.59 0. Fillers.12 0.44 0.80 2.4-Dioxane Dipropylene glycol Ethanol Ethanolamine Ethyl acetate Ethylene dichloride Ethylene glycol Ethyelene glycol monob utyl ether Ethyelene glycol monoeth yl ether Ethyelene glycol monometh yl ether Formamide Hexane Isophorone Methanol Methylene dichloride Methyl isobutyl ketone Methyl-2-pyrrolidone Nitrobenzene Nitroethane Nitromethane 2-Nitropropane Propylene carbonate Propylene glycol Tetrahydrofurane Toluene Trichloroethylene D(Ds) 20°C 0.90 2.80 56.96 1.82 1.87 1.95 1.75 2.89 1. and Fibers 129 TABLE 7.31 1.93 0.88 0.94 1.00 0.79 1.90 2.10 0.13 0.35 1.03 1.48 0. 4 6 13 28 30 37 40 41 44 47 48 56 61 75 82 90 94 96 98 104 105 106 120 121 123 124 126 131 140 148 153 162 167 172 177 178 179 181 204 205 222 225 229 D p.33 2.90 0.78 0.96 1.20 1.90 0.21 1.00 0.03 0.03 0.22 24.10 1.: Solvent Acetone Acetophenone Benzene 1-Butanol Butyl acetate Butyrolactone Carbon tetrachloride Chlorobenzene Chloroform Cyclohexane Cyclohexanol Diacetone alcohol o-Dichlorobenzene Diethylene glycol Diethyl ether Dimethyl formamide DMSO 1.04 0.59 0.12 0.33 0.79 1.70 0.58 Sedimentation Time (min) No.43 0.

the zeta potential.0 9. the cohesion ener gy parameters.5 4.0 means that there most probably are other sets o the same parameters that will ha ve a data fit of 1.1 δH 10.00 — 0.00 0. Data analysis which only considers the good solv ents to define the least spher possible. It is conceptually and.e.3 14.95 G/T 3/35 5/34 4/34 6/33 5/37 7/28 23/23 23/31 5/19 Inorganic Pigments. a system can still function satisfactorily. Eur.3 9. 38–42. Units are MP a1/2. It should not be too dif ficult t modify an organic surface to an alternative organic surface with satisfactory properties. b Data from Winkler. Cabot Hochdisperse a 16. although man y ha ve clearly found them beneficial.4 using data from Winkler.99 1.99 0..0 δP 3. It does not clearly indicate specific a finity relations of a g ven binder for the pigment surf ace. Low zeta potential means sedimentation is rapid in all solv ents.7 3.5 10.0 11. the current problem is the lack of data. A data fit of 0.9–11 The author has not found it necessary to resort to this type of approach in an y acti vity.7 16. 1997. It would seem that the HSP approach allows predictive ability that is not possible with an acid–base approach.95 1.. as a result of a given surface treatment.4 4. With permission. i. G/T is the number of good (G) liquids and the total (T) number of liquids in a correlation.88 0.9 18.0 and also define a sphere th surrounds all the good solv ents.1 4. An HSP correlation for the zeta potential of blanc fi e is gi ven in Table 7. also requires a greater de gree of coverage to mask the base inor ganic surface. The latter are generally lacking. a There is also a relation between ho w clearly a pigment can be characterized by sedimentation measurements and its zeta potential.4 HSP Correlations for Selected Materials Material Organic Pigments Paliotol® Gelb L1820 B ASF Heliogen ® Blau 6930L B ASF Socco Rosso L3855 B ASF Perm Rubin F6B Hoechst Perm Gelb GRL02 Hoechst Perm Lackrot LC Hoechst δD 18.7 2. and this type of characterization becomes difficult or perhaps impossible.5 4.8 4.7 20..7 Cabot Hochdisperse 19.79 0.0 2.7 3. The producers of inor ganic pigments and fillers must either g ve their products suitable .8 This is discussed further in the follo wing section. To obtain a complete picture of the ener getics of the surf ace. for example. One can suspect that some pigments ha ve such high-intensity zeta potential — at some cost — that e ven though the cohesion parameters match poorly with a gi ven binder .130 Hansen Solubility Parameters: A User’s Handbook TABLE 7.7 12.4 Fit 0.7 5. if desired.5 Note: A perfect data fit of 1. i. J. See discussion of the SPHERE1 program in Chapter 1. Mor research is needed to fully understand the successes of the acid–base as well as the HSP approaches. one needs an intensity f actor.7 19. depending on ho w the test liquids interact with these.e. as well as a qualitative factor.0 5.3 16. The zeta potential reflect the intensity (percentage co verage and number of layers) of the surf ace energy characteristics. The characterizations may reflect both a sur ace treatment and the surface of the base particles.1 3. However. J. Coat. more dif ficult to modify an ino ganic surf ace to mak e it compatible with or ganic systems.0 17.0 4. etc.. Organic pigments normally have a good organic substrate on which to base an or ganic surface modification.5 5. This requires a significant change in sur ace energy from high to much lower and. 1–2/97. This is given by HSP. in practice. Acid–base theories ha ve been popular .5 4. Fillers. presumably.3 Zeta Potential Blanc Fix eb 26.99+ is preferred to define t optimum sphere for this reason.5 19.0 11.5 Ro 5.

probably after much ef fort. This correlation supports the contention that cohesion parameters are significant for characterization of pigment. It is felt that those who understand the use of cohesion parameters are able to more systematically modify surfaces of inorganic materials to optimize or impro ve their compatibility with or ganic polymers and binders. fill . amine.Methods of Characterization for Pigments. One of these sets has δD. According to Winkler. This lack of e xpected performance may be also partly due to solvent size. The adjustments for optimum local association must be made by the solv ent molecules alone in the sedimentation testing. with a RED of 0. with success. The relati vely high local cohesion parameters in the binder that are associated with these groups w ould indicate a high af finity fo the high-cohesion-energy surface of the inor ganic material. th polymeric macromolecule is no longer a solid in the good solv ents. This has been done for inor ganic Rockwool® fibers that are to be incorporate into polypropylene. such as acid. or else one needs one or more superef fective additives to be able to achieve a good and stable dispersion. There were 19 liquids included in this careful study. There were 5 systems with zeta potentials greater than about 10 mV and 14 systems with zeta potentials less than about 5 mV . These liquids could easily be divided into two groups. This is a good correlation and supports the views presented earlier. and fiber sur aces. This is discussed in more detail in the follo wing section.δH equal to . or combinations of these. These phenomena appear to be a significant area for future research HANSEN SOLUBILITY PARAMETER CORRELATION OF ZETA POTENTIAL FOR BLANC FIXE Winkler8 has reported zeta potentials measured for 1% v/v blanc fi e with 0. in printing inks for the same reason. A rigid solid surf ace does not ha ve this potential for adopting ener getically desirable positions for its acti ve groups.5 It must be presumed that this type of systematic procedure can guide surf ace treatment of other inor ganic materials in a more directed w ay toward a desired goal. Fillers. After considerable refinement of the xperimental technique. it w as determined that tw o separate sets of HSP are required to describe the fiber sur ace. This provides a good. alcohol. giving a large number of possible (dynamic) structures that can be formed with the solv ent. When testing is finished. some solvents may not be able to retard sedimentation because the y cannot adopt the geometrical positioning required to do this without the help of a mobile substrate. It helps to incorporate gi ven (high-cohesion energy) groups in a grinding resin. A contribution to the formation of the ener getically desirable geometrical structures is not possible from the mo vement of rigid solid surf aces. The only major “error” was for hexamethylphosphoramide. Panex 33 from Zoltek. Binders with high-acid numbers are frequently used. whereas the predictions based on the behavior of other solv ents that do significantly retard sedimentation indicate that this shoul be the case. The solvent will not dissolve that polymer or polymer segment away from the surface. A macromolecule has v arious possibilities for contortions and the positioning of significant act ve groups in solution (or when swollen). Data fits h ve been generally lo wer for characterizations of particulate solid surf aces. There are many solvents that do not retard sedimentation significantl .3 includes the results of the correlation of these data with cohesion parameters. those that ha ve much lower cohesion parameters.9 mV . stable anchor on the pigment surf ace. these local re gions of adsorbed polymer se gments are not particularly soluble or are insoluble in the cheaper h ydrocarbon solvents — for e xample.34% moisture content. Therefore. CARBON FIBER SURFACE CHARACTERIZATION Hansen solubility parameters ha ve been assigned to a carbon fiber sur ace. and Fibers 131 surfaces.951 and a zeta potential of 1. At the same time. whereas the particulate fille remains a solid. the location of the active groups. etc. there was no correlation with the acceptor or donor numbers (acid–base). such as fillers. than for characterizations of polymers based on solubilit .δP. Table 7.

3 8.3 12.1 5. Ministry of Science.7 11. De gussa HT indicates a special heat treatment w as performed prior to testing. ether.5 8. This work was done as a part of the Frame work program Interface Design of Composite Materials with the support of the Danish Research Agency.8 3.3 3.9 13. ag ain in MP a1/2.0 7.1 4. With permission. The h ydrocarbon-like surface can be the backbone of the polyacrylonitrile (PAN) precursor for the fibe.6 10.9 2. whereas the analyses were done at the Risø National Laboratory.9 Comments Sedimentation rate 3mm fiber Sedimentation rate 3mm fiber Suspension Sedimentation (slow) Sedimentation (rapid) Sedimentation (rapid) Sedimentation (rapid) Sedimentation (rapid) Sedimentation (slow) Solubility Solubilityc Printex V (5519-1). Carbon. Two separate sets of HSP assignments are confirme by x-ray photoelectron spectroscop y (XPS) analysis.1. carboxyl. 13. 1591–1597. and Smith.7 Ro 10. Denmark.L.2. Some guides are discussed in the follo wing. The second set is characteristic of a h ydrocarbon material with δD.7 δP 17. It is a well-known fact that a small percentage of acid groups (or alcohol groups) on a polymer chain will promote adhesion and adsorption to man y surfaces.2.4.. all in MPa1/2.4. Broendby. This corresponds to a highly polar surface with a significant ydrogen bonding component as well. One can consider the cohesion ener gy properties of formic acid ( δD.4 18.2.2 2.4 contains the HSP data used to construct Figure 7. Table 7.9 16. The cohesion ener gy parameter of an isolated acid group is high.3.0 8.17.4 17. These data confirm that “carbon can be man y things with widely dif ferent surface energies. CONTROLLED ADSORPTION (SELF-ASSEMBLY) Significant tasks for formulators are to control the sur ace and interf acial ener gies of products. The origins of the material.8 3. carbonyl. and to pro vide/ensure adequate and durable adhesion to gi ven substrates.4 9.132 Hansen Solubility Parameters: A User’s Handbook TABLE 7.5 2. Data from Hansen. to maintain stable dispersions.2 21. especially if they are water reducible. Guidelines for courses of action are frequently a vailable when cohesion ener gy parameters are referred to. Technology and Innovation (STVF). Log mole fraction solubility greater than –3.9 16.7 19.9 δH 14.1 17. as well as the method in which it has been handled or treated. a b c d δD 13.0 11.9 21.3 8.5 20.5 18. The hydrogen bonding and polar contrib utions arise from the both bound and unbound (sizing/finish agent) chemical functionalities mainly in the form of ydroxyl. The HSP characterizations were done at FORCE Technology. if this is possible.14.7 12.0. 2004.2 11. and nitrile groups.5 10.δP. A. Two separate regions are found to coe xist on the carbon fiber sur ace.3 8. . 42(8–9).5 20. This is required to allow substrate wetting.δH equal to 17.2 9.7 8. C. The carbonaceous backbone of the carbon fibe primarily accounts for the nonpolar region.M. can completely dominate the nature of the surface of the gi ven materials and their solubility . Roskilde.7 5.6 6. amide.4 12.5 HSP Correlations for Various “Carbon” Materials Material Carbon fiber (high r gion)b Carbon fiber (l w region)b Carbon black a Carbon black 1 Carbon black 1 HT b Carbon black 2 HT b Carbon black 3 HT b Carbon black 4 HT b Petroleum coke Coal tar pitch C60 fullerene d Note: Units are MP a1/2. Denmark.8.1 7.

26-03-0160). δH = 14.9. 16. Unit are MPa1/2. If. low Fullerene Petroleum Coke 2 0 0 2 4 6 8 δH 10 12 14 16 FIGURE 7.Methods of Characterization for Pigments. and Fibers 133 20 18 Carbon fiber. This would provide an anchor that the product will not be able to remo ve. The work on which this figure is based as supported by the Framework program Interface Design of Composite Materials (STVF fund No. the solvent were too good for the anchor. Reproduced with permission.3. This type of adsorption may be called hydrophilic bonding. Fillers. The polar cohesion energy parameter of an acid group is not so high. . high 16 14 12 δp 10 “Carbon Black” 8 Carbon Black 1 Coal tar pitch 6 4 Carbon fiber. δP. It w ould seem logical to systematically use acid groups for adsorption to high-energy surfaces and to mak e certain that the cohesion ener gy parameters for the solv ent and bulk of the product are much lo wer. on the other hand. such that isolated acid groups w ould not be dissolv ed. 11.1 Characterization of a carbon fiber and comparisons of this with other carbon materials.6) as an isolated part of a polymer chain.

1 MP a1/2). 39(511). Hansen.134 Hansen Solubility Parameters: A User’s Handbook it could be presumed that e ven an acid group may be too readily dissolv ed off the surf ace or at least tak e part in a dynamic equilibrium of adsorption and desorption. Mahmood Ali. This method has generally allowed useful characterizations. Shareef..A. 2. For example. b ut the y were included in the lesser interacting group by the arbitrary di vision into good and bad groups. ch acterized with the same ener gy parameters should pro vide incentive for impro ving on the present state. The positive effects of associative thickeners can be counteracted by the presence of solv ents preferentially locating where the hydrophobic bonding is to occur. The total cohesion ener gy parameters for ethanol and nitromethane are v ery close: 26. The Three Dimensional Solubility P arameter and Solvent Diffusion Coefficient. Paint Technol. A challenge to the creati ve mind is to deri ve ne w uses for high-ener gy groups that are not particularly water soluble or sensitive.M.1 MP a1/2. P.. Doc toral dissertation. J. plasticizers.M. respectively.4 MP a1/2) compared with nitromethane (5. such as the h ydrophobic bonding in proteins. the data are often scattered and not nearly of the quality usually found when observing polymer solution beha vior. Absorbed/adsorbed w ater can sometimes interfere with such anchors at high-ener gy surfaces. CONCLUSION Many pigments and fillers h ve now been characterized by Hansen cohesion parameters (HSP). Danish Technical Press. The three dimensional solubility parameter — k ey to paint component af finities II.... M. C.1 and 25. Certain se gments of gi ven molecules have such low cohesive energy parameters that they are no longer soluble in the media. Local adsorption by so-called active groups (alcohol. Ethanol is soluble in water. 1967. acid. Copenhagen.J. Ethanol has a relati vely high-hydrogen bonding parameter (19. which is usually aqueous. One assumes that maximum physical adsorption is accompanied by closely matching HSP . A method based on relati ve sedimentation time and/or suspension is described for doing this. and Reddy. J. February 1986. This scatter of data for untreated surf aces in particular may cause some to disre gard the method. Coat. 58(733). Hansen. although some experience is helpful. . they can develop a better one. Would not the nitro group be a suitable anchor analogous to the pre vious discussion concerning acid groups? Also. hopefully. C. The obvious advantages of ha ving solvents. 3.. 505–510. can gi ve anchors on a surface that may no longer be soluble in the continuous media.. The hydrophobic bonds lose strength or may even dissolve away. Yaseen. Several of the pigments reported in Table 7.. The division of the cohesion ener gy into at least three parts allows these considerations to be made in a reasonably quantitati ve manner. REFERENCES 1. This mak es all the difference. The reverse of this thinking is systematically used by those designing associati ve thickeners and also by nature. and the y either seek out their o wn kind (associate) or perhaps adsorb on or penetrate into a low-energy surface where cohesion energy parameters more suitably match. fillers. it w ould not be h ydrophilic with the inherent problems of w ater sensitivity associated with high-hydrogen bonding parameters.M. both in terms of numbers of characterizations as well as impro ved methodology.. etc.4 did indeed ha ve moderate af finity for the nitropara fins. Many e xamples are gi ven. and therefore will remain in place as required. Technol. One can choose nitro groups or perhaps groups containing phosphorus as e xamples of species characterized by highpolar-cohesion-energy parameters and lo w or moderate h ydrogen bonding parameters. polymers. nitromethane is not. K. fibers. 35–44. amine) ha ving the required match. pigments. 1967. M. for xample.

. Ed. A. 6. Interscience. L. 34(10).. Using Hansen solubility parameters to correlate solubility of C 60 fullerene in or ganic solvents and in polymers.. R. in Treatise on Coatings.. 10. 5/98. 334–340. Gardon.L. Soerensen. J. pp. Soerensen.M.R. 1988 (in Danish). 7. Solubility parameters. 12. 363–372. Suppl. Colloid chemistry aids to formulating inks and paints. Ne w York. Färg och Lack Scand.. 8. P art II. J. J.. Application of the concepts solubility parameter and pigment char ge. J. 1998.. and Teas. 8. 1976. 81–93. Cohesion parameters used to formulate coatings (K ohaesionsparametre an vendt til formulering af f arver og lak). Solubility parameters.. Application of the acid/base concept describing the interaction between pigments... 9. 1996. Vinther.. Chim. J. Schröder. and Smith. 1591–1597. Marcel Dekk er. Vol. J. J. Ne w York. and Beerbo wer. and solv ents. Lecture for the Danish Society for PolymerTechnology.. . 1–2/97.. 889–910. 47(602). and Fibers 135 4.. J. A. A. 1971. A.P. Myers. 31–39.. Eds. Copenhagen. Carbon. in Kirk-Othmer Encyclopedia of Chemical Technology. Zeta potential of pigments and fillers. Fillers.S. C. Winkler. 2nd ed. Paint Technol. C. binders. Eur. Hansen. 42(8–9).. Coat. Hennissen. P. 11.. 5. Peint. 1975. 38–42.L. Standen. 2.. Systematic Modification of Filler/Fibre Sur aces to Achieve Maximum Compatibility with Matrix Polymers. Hansen. February 10. chap. (England).Methods of Characterization for Pigments. 1997. and Long.M. Coat. 1971. 2004. Vol. Characterization of Coatings: Physical Techniques. P. 34(4)..Eur.


printing inks. optimization increasingly includes considerations of worker safety and the e xternal environment. etc. plasticizers. b ut not necessary. both liquid and solid. one must ha ve or estimate its HSP . understanding and controlling processes. SOLVENTS In order to find the optimum sol ent for a polymer . Computer techniques are helpful. and. etc. to mention a fe w. They also aid in substitution to less hazardous formulations in various other types of products such as cleaners. INTRODUCTION There are man y applications documented in the literature where HSP ha ve aided in the selection of solvents. drugs. aromas. The same principles useful for under standing the behavior of coatings are useful in understanding beha vior in a lar ger number of related products. b ut this procedure does not necessarily optimize the ne w situation. Hansen ABSTRACT Hansen solubility parameters (HSP) are widely used in the coatings industry to help find optimu solvents and solvent combinations. Volatile organic compounds (VOC) are to be reduced to the greatest extent possible. etc. This allows their relative affinities to be established. The behavior of solvents in connection with surf aces of various kinds and the use of HSP to understand and control surf ace phenomena is especially emphasized. fillers. surf actants. which can be characterized by HSP . and chewing gum. Matching the HSP of an already e xisting solvent or combination of solv ents can be done. 137 . A solvent with the highest possible affinity for the polymer is both xpensive and probably not necessary and will rarely be optimum. The discussion in this chapter includes the ph ysical chemical reasons wh y solv ents function as the y do in man y practical cases. polymers. There are man y examples of controlled self-assembly. Chapter 11 is devoted to replacing ozone-depleting chemicals in cleaning operations. in general. pigments. as well as to the surf aces of substrates. Products where HSP concepts can be used in a manner similar to coatings include other (filled) polymer system of various types such as adhesi ves. including adhesives. che wing gum. can also be characterized by HSP to infer ho w they behave in seemingly comple x systems. These contain widely different materials. printing inks.. This chapter emphasizes coatings applications and discusses the practical application of HSP to solv ent selection. In more recent years. and fiber Various additives such as resins. The optimum depends on what is desired of the system. printing inks. Pr vious chapters ha ve discussed ho w to assign HSP to solv ents. adhesi ves. of fering guidance where affinitie among materials are of prime importance 1–5 (see also the follo wing chapters and e xamples below).— Coatings and 8 Applications Other Filled Polymer Systems Charles M. scents. fl vors. and resins.

cannot be neglected in every case. it has been found that their dispersion parameters will be rather close regardless of structure. when comparing noncyclic solvents in practical situations. etc. It is usually good practice to include a small or moderate amount of slo wly evaporating solvent of good quality and lo w water sensitivity to take care of this situation.. This same trend is expected for oligomers of a polymer as molecular weight increases. The mixture of equal parts xylene and n-butanol has been widely used in conjunction with man y polymers such as epoxies. such as a k etone. The total solubility parameter for aliphatic hydrocarbon solvents is identical with their dispersion parameter and increases only slightly with increased chain length. This relation has generally pointed to the use of . such as chlorine. The real benefit of this concept is in interpretin more complicated behavior. and a solubility parameter approach is particularly valuable in quickly limiting the number of candidates. will ha ve higher dispersion parameters. These are located in boundary regions of the solubility region for epoxies. These can lead to a situation where solvent quality changes in an adv erse manner. Here again. or glycol ether. and the polar and h ydrogen-bonding parameters are higher than if n-butanol had been added to the same concentration. Balance of solv ent quality on e vaporation of mixed solvents is also necessary. and simple two-dimensional plots using δP and δH can often be used by those with limited x eperience with these techniques to solve practical problems. bromine. In spite of these pressures to let the computer do the thinking. 1 Glycol ethers also can be added to hydrocarbon solvents with adv antage. δD. Some safety mar gin in terms of e xtra solvency is advised because of temperature changes. These have frequently been slo wly evaporating ketones and esters. as some soent retention must be anticipated. As indicated previously. b ut a third solv ent. but. There are man y possibilities. Much more w ork needs to be done in these areas. An oxygenated solv ent which is frequently added to h ydrocarbon solvents and has been cost effective in increasing the v ery important h ydrogen-bonding parameter has been n-butanol (or sometimes 2-butanol). such as affinities of polymers with polymers and polymers with suraces as described in the follo wing. ester. Most coatings applications involve solvents reasonably well within the solubility limit which is indicated by the boundary of a solubility plot such as that sho wn in Figure 8. By putting this together with other necessary considerations such as flash point. it is becoming almost mandatory that computers be used. Neither xylene nor n-butanol satisfactorily dissolves an epoxy of higher molecular weight by itself. and calculations of solvent quality can be made at all stages of evaporation. A major f actor in this almost immediate o verview is the decrease in the number of solvents useful in coatings. availability. the follo wing e xamples are intended to illustrate principles on which to base a systematic course of action. metals. for example. cost. most solvent suppliers and many large users of solvents have computer programs to predict solution behavior as well as e vaporation phenomena.. the experienced formulator who knows the HSP for the relati vely few solv ents possible in a gi ven situation will be able to select a near -optimum combination by a process of e xclusion and simple mental arithmetic.138 Hansen Solubility Parameters: A User’s Handbook Whereas hand calculations and plotting of data are still quite useful and at times more rapid than computer processing. Relative costs for impro ving solvency from a h ydrocarbon base solv ent can be estimated by n-butanol. and the least v olatile solvent is enriched and left behind. Re gardless of the means of processing data. etc. but on opposite sides of the characteristic Hansen spheres (see Figure 8. proper vaporation rate. etc. The nonpolar cohesion parameter. is often included in small amounts to increase the polar parameter/solv ency of the mixture. 1 A maximum of cheaper hydrocarbon solvent is also desired and can frequently be used to arri ve at such a situation for common polymers used in coatings. for the relation ( δP2 + δH2)1/2/cost. To this end.2). such as increased water sensitivity and poorer corrosion resistance of the final film. and those containing atoms significantly la ger than carbon. computer approaches are possible. Cyclic solvents. an experienced formulator can often arri ve at a near -optimum result without recourse to paper or to computers. odor. computer techniques can be v ery useful but are not always necessary. b ut the following gives an indication of what might be e xpected. potential variations in production.1. The addition of glycol ethers or other w ater-soluble solv ents can ha ve adv erse ef fects. This does not mean the use of HSP is on the w ay out.

Solvents can be ranked in this manner to arrive at the least cost solutions to given solvent selection problems. Of course. which also get filtered out and th y are not therefore a vailable to do their job in the film when the ater evaporates. penetrating the substrate f aster than the polymer particles. (From Hansen. Ketones (MEK. The distribution between the w ater phase and the dispersed polymer phase depends on the relative affinities for ater and the polymer. making certain they are where they are needed when the y are needed.M. . Such coalescing solvents may be preferred for applications to porous substrates. methyl ethyl ketone). or other solv ents such as glycol ethers and their acetates (here ethylene glycol monoeth yl ether and eth ylene glycol monoeth yl ether acetate) can be used to do this. n-butanol). Otherwise. alcohols (B. Solvents which are not particularly w ater soluble will preferentially be found in the polymer phase. 17(4). 1971.) 8 10 12 • E GMEEA example. 69–77. Aliphatic hydrocarbons (ALH) and aromatic h ydrocarbons (ARH) do not al ways dissolve well enough so other solv ents must be added to bring the mix ed solvent composition into the re gion of solubility for the binder .Applications — Coatings and Other Filled Polymer Systems 139 12 10 8 6 δp 4 MEK • • E GMEE •B 2 ARH 0 ALH 0 2 4 6 δH FIGURE 8.. Färg och Lack. Coalescing solvents in water-reducible coatings are often (but not always) those with somewhat higher hydrogen-bonding parameters than the polymer . as a cost-ef ficient sol ent to increase the h ydrogen-bonding parameter in particular . w ater-soluble coalescing solv ents w ould tend to follow the aqueous phase. Units in the figure are in (cal/c 3)1/2. With permission. The expected solvent improvement at least cost is discussed in the te xt as the quantity ( δP2 + δH2)1/2/cost.1 Sketch showing location of typical solvents relative to the HSP of a binder. this af finity of the coalescence sol ent for the polymer is a function of its HSP relati ve to those of the polymer . C. which also means the y are w ater soluble or have considerable w ater solubility. The choice of solv ent today w ould involve glycol ethers based on prop ylene glycol as discussed in Chapter 18. the solv ent must dissolv e to some e xtent in the polymer to promote coalescence. When w ater e vaporates.

(From Hansen.M. in particular. whether nonionic or ionic. meth ylene chloride (MC). Eth ylene glycol monob utyl ether. Units are (cal/cm3)1/2.) Amines are frequently added to w ater-reducible coatings to neutralize acid groups b uilt into polymers. Increases in temperature especially lead to lo wer hydrogen-bonding parameters (see Chapter 1. the hydrophobic portion may be forced to remain in the interf acial region. These three very volatile solvents have often been used in analytical work. thus providing a water-solubilizing amine salt.2 Sketch sho wing formulation principles using tw o relati vely poor solv ents in combination to arrive at a good solv ent. which means the y are to be found in a stabilizing role in the interf ace in the aqueous phase while still being attached to the polymer . Färg och Lack. Chapter 3.7 The hydrophobic end of such molecules may reside within the polymer if HSP relations dictate this. as well as contrib uting to lowered surface tension. if the HSP differences are too great. Surface-active agents. because the y dissolv e all of the typical coatings binders sho wn in the figure. has been shown to be a good coupling agent. and Chapter 10). F or this reason. etc. paint removers. and the h ydrophobic end will seek out an environment where energy differences are lowest (self-assembly). not being accepted by the aqueous phase either . solv ents with high h ydrogen-bonding parameters. Such amine salts ha ve been characterized separately to demonstrate that the y have higher solubility parameters than either (acetic) acid or organic bases.140 Hansen Solubility Parameters: A User’s Handbook 12 10 8 δp 6 MEK • MC • 2 THF • •B 4 0 2 4 6 8 δH 10 12 14 16 FIGURE 8. glycol ethers. or meth yl eth yl ketone (MEK) among others. such as glycols. 6 These salts are h ydrophilic and ha ve very little af finity for the polymers used i coatings. The hydrophilic end with a high h ydrogen-bonding parameter will seek the aqueous phase. 17(4). and the dispersed solubilized polymer can be visualized in terms of a porcupine with raised quills. are also to be found where the af finities o the respective parts of their molecules dictate their placement. become better solv ents for most polymers at higher . and alcohols. Xylene (X) can be mix ed with n-butanol (B) to arri ve at a mixture which can be improved by additions of tetrah ydrofuran (THF). 1971. Labeling requirements h ve dictated other choices in more recent years. Electrostatic repulsion contrib utes to stability as well. C. Amine in excess of that required for total neutralization of the acid groups acts lik e a solv ent... Otherwise. With permission. It might be noted here that some solv ents have surfactant-like properties as well. lik e seeks like. 69–77.

If the SWEA T w ater cannot rapidly dif fuse out of the coating.3 Sketch of water uptake in a polymer as a function of temperature. these temperature ef fects are an advantage in water-reducible.Applications — Coatings and Other Filled Polymer Systems 141 % H2O T1 T2 T3 T1 > T2 > T3 Time FIGURE 8. and epoxy). The amount and rate of uptak e is higher for the higher temperatures. Excess w ater freed by this mechanism has been called SWEA T (soluble water e xuded at lo wered temperatures).3. The use of supercritical g ases as solv ents has become possible in recent years.3 where w ater uptake curves are sho wn for three temperatures. These may dissolv e too readily in the swollen.7. Prog. Ne w developments in corrosion and blister formation in coatings. Rapid cooling to belo w the solubility limit at a lo wer temperature means the system is supersaturated.. This can be seen in Figure 8. becomes more fluid. With permission from Else vier Science. polyester.7. and 5. Quenching to a lower temperature (arrow) leads to excess water in the film and possibly to ater blisters and delamination (see te xt for further discussion).) temperatures.. for example. it will appear as a separate phase. and ven in EPDM rubber. and δH equal to 15. These are now revised as discussed in Chapter 10 to δD. 26.3 and Figure 12. This can markedly affect hot-room stability in water-reducible coatings. as poor solv ents at room temperature become good solv ents in the o ven after the w ater has evaporated. It has been documented in man y cases that a rapid quench from hot water to cold water can cause blisters in coatings. This effect is not restricted to w ater. dispersed polymer.M. ut ultimately at h ydrophilic sites or at a substrate. an has altered af finities for stabilizing sur ace-active agents. perhaps first as clusters. A very special destructi ve effect of w ater is caused by the reduction of its h ydrogen-bonding parameter with increases in temperature. The solubility of w ater in most polymers is higher at a higher temperature than it is at a lo wered temperature because the HSP for the polymer and w ater match better at the higher temperature. Examples of measurements of this type are sho wn in Chapter 12. Higher temperature leads to more rapid uptake and to higher equilibrium levels. C. oven-cured coatings.4 for an EPDM rubber g asket and for a poly(ether sulfone) tensile bar . which swells. 113–120. Figure 12. The cooling due to the methanol evaporation was sufficient to produce methanol blisters nea the air surface of the coating because of e xcess amounts of methanol over that soluble at the lower temperature resulting from the methanol e vaporation. leading to higher film int grity. 8 Previously dissolved water within the film n w becomes in excess of that soluble in the film. Org. The solubility parameters for carbon dioxide have been reported 9 earlier and in the first edition of this handbook based on the room temperature solubility of the g as in different liquids. HSP v alues are reported as a . as more of the solv ent will partition to the polymer phase. Coat. it has also been seen for an epoxy coating that w as repeatedly remo ved from room temperature methanol to measure weight gain. This special effect has been noted by the author in coatings (alk yd. When carefully controlled. The coating fails by blistering or delamination. δP. 1995. 6. in rigid plastics such as poly(phen ylene sulfide) and poly(ether sulfone). (Reprinted from Hansen.

1. TECHNIQUES FOR DATA TREATMENT As mentioned earlier. of course. this can be adjusted if necessary. Solv ent quality can be adjusted by the RED number concept. This same type of analysis can be used to e valuate the temperature and pressure ef fects for other g ases. The varnish remover should be in this region. a copolymer of eth yl methacrylate and methyl methacrylate from Rohm and Haas. ut one does get all three parameters onto a tw o-dimensional plot with enough accuracy that its use has survi ved for this type of application (at least). Some accurac y is lost.1 to Equation 8. These parameters are found using the solvents that dissolve more than the theoretical amounts of carbon dioxide that are reported in Table 10. The individual Hansen parameters are normalized by the sum of the three parameters. which. The Teas plot in Figure 8. and are. Plasticizers should be included in the calculations. A helpful simplifying relation to use in solv ent selection calculations using solubility parameters is that the resultant v alues for mixtures can be estimated from v olume fraction a verages for each solubility parameter component. There is a re gion in the lo wer. which is discussed in Chapter 1 (Equation 1. A varnish which could be considered for use is P araloid® B72. h ydrogen-bonding parameters with a circle (or estimated circle) for the polymer in question. They will be very slow to dissolve rigid polymers. and low labeling requirements. HSP correlations for materials of interest in restoration of older paintings are included in Table 8. a simple approach to man y practical problems is to mak e a two-dimensional plot of polar vs. in particular. or graphically as described abo ve. Mixtures of h ydrocarbon solvent and ethanol are located in this re gion and could be considered. This gives three fractional parameters defined by Equation 8. low VOC. are modified HS parameters.2. but their use has been limited to relati vely smaller items because of the size of pressure equipment.2) (8.3) The sum of these three fractional parameters is 100 in the form the equations are written. The art in volved in this stage of the conserv ation process is to remove the old varnish without attacking the underlying original masterpiece. nonvolatile for all practical purposes.1) (8. 2 One can mention the formulation of solv ent cleaners based on v egetable oils as an additional example. HSP principles ha ve been used since the late 1960s for selecting solv ents and solv ent blends for this purpose.10).4 includes an estimate of the solubility/strong attack of older . in f act. Subsequently. See also Chapter 3. One can then plot points for potential solvents and quickly arri ve at a starting composition for an e xperiment. A special plotting technique for solv ent selection de veloped by Teas11 is used frequently by those who restore old paintings. A linear mixing rule based on the volume (or weight) fractions of the solvent components is usually satisf actory. This allows use of the special triangular technique. right-hand part of this plot where the v arnish is soluble and the dried oil is not. . dried oil paint.142 Hansen Solubility Parameters: A User’s Handbook function of temperature and pressure. The circle should encompass the good solv ents. The use of such g ases is considered an adv antage for the environment.10 Such “green” products ha ve found increasing use.3 fd = 100 δD/(δD + δP + δH) fP = 100 δP/(δD + δP + δH) fh = 100 δH/(δD + δP + δH) (8. Solvent technology has also been used in a wide v ariety of other products and processes as listed by Barton. and there is no theoretical justification for this plotting technique. as ha ve those solvents with low volatility. 12 The triangular plotting technique uses parameters for the solv ents.

Solv ents with RED less than 1.915 1.8 7. Equation 1. TABLE 8. cleaners.0 Ro 9. evaporation rate.0 δH 5.000 0. etc.0 FIT 1.6 7. A nearest neighbor search involves calculation of the quantity Ra (Chapter 1. leads to the most promising candidates for the substitution. . A similar application is to predict which other solv ents will probably be aggressi ve to a chemically resistant coating where very limited data have indicated a single solvent or two are somewhat aggressive. for example. cost. heptane (H).4 16.2 6.9) for a whole database. The potentially most aggressi ve liquids are at the top of the list.Applications — Coatings and Other Filled Polymer Systems 143 0 100 fh fp DRIED OIL + PAR B72 E PAR B72 0 fd 100 CH 0 100 FIGURE 8.0 δP 7. etc.6 18.000 G/T 17/26 30/56 9/22 A computer search with the SPHERE computer program (Chapter 1) for “nearest neighbors” for a gi ven single solv ent has been used man y times to locate alternates for a wide v ariety of product types including coatings of v arious descriptions.3 5. and ethanol (E) (see te xt for further discussion).4 4.0 are “good” and therefore easily recognized.4 8.1 HSP Correlations for Materials of Interest in the Conservation of Older Paintings MATERIAL Paraloid® 22 solubility Dammar gum de waxed Dried oil (estimate) δD 17. Equation 1. and then arranging the printout in RED number order (Chapter 1.4 Teas plot for a typical painting conserv ation situation where a v arnish is to be remo ved or applied without attacking the underlying original oil painting. Solvents indicated are cyclohexane (C).10). Sorting out these possibilities considering toxicity .

as well as for conditions where the solv ent has lower HSP than the pigment. 400. will probably dissolv e the polymer too well. the desired solv ent quality in man y coatings is just slightly better than that of a mar ginal solvent. Solvent 2 would probably be too e xpensive and. Schröder13 (B ASF) confirms that the optimum polymer adsorption will be found when th binder and pigment surf ace have the same HSP. Cop yright ASTM. One reason for the desired mar ginal solv ent quality is that this will ensure that the polymer adsorbed onto pigment surf aces during pigment dispersion has little reason to dissolv e away from that surf ace. These depend on their HSP relati ve to each other . If this polymer is dissolved away. and surfaces can be inferred by their relati ve affinities. the polymer. Pigment δp 2 1 3 Polymer δh FIGURE 8. and the li e. (Reprinted from Hansen.0 relati ve to the polymer will be sought. means the composite v ehicle has parameters v ery closely matching those of the pigment. He prefers the pigment to ha ve HSP values placing it intermediate between the solvent and binder. The solvent in this case should have a RED number for the pigment surf ace greater than 1.1 for discussion). C. and perhaps e ven rheological dif ficulties.M. 14th ed. and the pigment surface should have high affinity for each othe . in addition. or at least reasonably high. The dispersion stabilizing polymer should remain on the pigment surf ace where it is desired. Of course. or some portion of it. p. A v ery similar type of result w as found by Skaarup. 1995. fillers. This situation.5 where the mar ginal solvent is number 1.0. undesired settling. with the solv ent and binder on opposite sides of the pigment. This is suggested for conditions where the solv ent has higher HSP than the pigment. With permission. This means RED numbers just under 1. 14 who especially emphasized that optimum color strength w as found for solv ents marginal in quality for the binder and poor for the pigment in question.) . pigments. of the Gardner -Sward Handbook.144 Hansen Solubility Parameters: A User’s Handbook SOLVENTS AND SURFACE PHENOMENA IN COATINGS (SELF-ASSEMBLY) Chapter 6 and Chapter 7 ha ve been de voted to the characterization of surf aces for substrates. As stated pre viously. and the RED number concept can be quite useful. polymer.. Paint and Coating Testing Manual.5 Sketch showing influence of sol ent quality on e xpected pigment dispersion stability (see te xt and Figure 1. to aid in the planned af finity approach to pigment dispersion stabilit . This means the interplay between solvent. A sketch of the optimum relations in coatings is gi ven in Figure 8. He indicates that the solv ent should be v ery poor for the pigment and located in the boundary re gion for the binder. the result is most likely pigment flocculation which leads to color change.

Applications — Coatings and Other Filled Polymer Systems 145 In special applications. The advantages of a copolymer containing the monomers of A or B and C should also be e vident. Lower RED numbers with respect to this polymer are desired. where it can be seen that dif ferent percentage blends of solvents 1 and 2 will have different relative affinities for the polymers. whereas polymer C would not. 400. Cop yright ASTM. Miscible blends of tw o polymers ha ve been systematically found using a solv ent mixture composed exclusively of nonsolvents. Polymers A and B should be compatible.5) w ould adsorb onto the pigment surf ace preferentially.6. No other alternat ve theory of polymer solution thermodynamics can duplicate this predictive ability.) .15 This is demonstrated schematically in Figure 8. POLYMER COMPATIBILITY In some cases. In such cases the solvent may ha ve high af finity for the pigment sur ace as well as for dispersing polymer . (Reprinted from Hansen. This means a better -than-marginal solvent for the polymer is desired. 1 1+2 δp A 2 B δh FIGURE 8.6 Sketch showing how two otherwise immiscible polymers can be brought into a homogeneous solution by the use of mix ed nonsolvents. and pigment dispersion stability would be poor. Such a copolymer will essentially couple the system together. This is true when tw o polymers are mix ed and one of them precipitates. C. Polymer miscibility is enhanced by lar ger overlapping solubility regions for the polymers as sk etched in Figure 8. Such a systematic analysis allows modification of a g ven polymer to provide more o verlap and enhanced compatibility . of the Gardner -Sward Handbook.. A good anchor has high af finity for the pigment sur ace and mar ginal affinity for th solvent. This is most likely the polymer with the lar ger molecular weight. p. an extended polymer chain configuration is desirable. An extension of this thinking may be required for pigment pastes and other very highly filled products. Solvent 3 (Figure 8. while still maintaining af finity for th other polymer. 1995. closer -than-usual matches between solv ent and polymer solubility parameters are required. An ideal situation here is where all the ingredients ha ve the same HSP. With permission. 14th ed. ut a solid anchor to the pigment surf ace is also desired. and it must be dissolv ed even better. there is little dispersing ehicle relative to the pigment.7.M. Paint and Coating Testing Manual. and the solv ent must be considered as being part of the dispersing v ehicle. In these cases.

ormulation principles have been discussed in detail elsewhere. and a controlled porous membrane results. assuming that the δD parameters would not be too dif ferent.) Van Dyk et al. Cop yright ASTM.. [ η]. This is an expression reflecting polymer chain xtension in solution.146 Hansen Solubility Parameters: A User’s Handbook C δp A B δh FIGURE 8. The (8. further differentiation between the polymers and impro ved group contribution methods may offer even more impro vement.19. Another example of controlled phase relations during a dynamic process is found in the formulation of self-stratifying coatings. This is interesting in that the solubility parameter is a thermodynamic consideration. the solvent quality becomes bad. As stated above additional uses of HSP (and the total solubility parameter) in solv ent technology can be found in Barton. . 1995.2 but these are too numerous to include here. with higher v alues reflecting greater chain entanglements because of greater polyme extension. p.8. [η] = ( ηs/η0)/c . a couple of examples relating to guided polymer compatibility are w orthy of special mention. 16 ha ve correlated the inherent viscosity of polymer solutions with HSP inherent (intrinsic) viscosity used in this study .M. With permission. (Reprinted from Hansen. Although this is a good starting point to pro ve the procedure has possibilities. 14th ed. 401. 21. C. some of the recent work on miscible polymer blends should also be noted. 17. These are the formulation of asymmetric membranes for separations. The HSP principles involved in this type of coating can be seen in Figure 8. of the Gardner-Sward Handbook.20 Before concluding this section.7 Sketch describing e xpected polymer miscibility relations (see te xt for discussion). This is discussed in Chapter 6 in terms of the creation of interf aces and therefore interfacial surface tension. whereas the viscosity is a kinetic phenomenon.4. is given by Equation 8. η0 is the solv ent viscosity. The solution is immersed in w ater. and c is the solution concentration. However.18 where polymer solutions — ha ving gi ven HSP relations — and at least one solv ent soluble in w ater are used.22 This work used group contribution estimates of the δP and δH parameters only in an effort to correlate interfacial tension between polymers. Higher [ η] were found for solv ents with HSP nearest those of the polymer . Paint and Coating Testing Manual. The solvent must dissolv e both the topcoat and primer and allo w the lo wer surface tension topcoat to migrate to the surf ace during film formation.5 g/dl. The concentration used was about 0.4) ηs is the solution viscosity .

Ed.Applications — Coatings and Other Filled Polymer Systems 147 PRIMER POLAR PARAMETER PARAMETERS REQUIRED FOR COMMON SOLVENT TOP COAT (LOWEST ENERGY) HYDROGEN BONDING SOLUBILITY PARAMETER FIGURE 8. If the y do not. CONCLUSION Many practical uses of the solubility parameter concept ha ve been described in detail. 324. a desired product characteristic is that the release of the taste components should be controlled. Boca Raton.8 Sketch illustrating the principles of solv ent selection for self-stratifying coatings.25 He concludes that all the (solid) ingredients in a paint formulation should ha ve the same ener gy characteristics. p. With permission.. One can just as well make the formulation stable from the start. and then e verything will remain stable just where it is because there are no driving forces for anything to move around. When all of the materials in volved in a gi ven product and application can be characterized with the same af finity (solubility/cohesion) parameters. and other performance considerations. In addition to rheological beha vior. there will be a driving force for this to occur. the truth of the matter is well documented in v ery practical terms. Greater differences in solubility parameters between fl voring agents and w ax-free gum bases lead to enhanced fl vor release.S. FL. This can lead to problems. including optimizing solvent selection. appearance. 1997. K. improving polymer compatibility. Although this sounds expensive. obvious. the possibility xists to predict interactions . (From Birdi. CRC Press. Handbook of Surface and Colloid Chemistry. Similarity of HSP can lead to stopping the desired release too soon. Perhaps the most important practical w ork dealing with solubility parameters and the stability of pigment dispersions is that attributable to Stephen.) HANSEN SOLUBILITY PARAMETER PRINCIPLES APPLIED TO UNDERSTANDING OTHER FILLED POLYMER SYSTEMS Recent characterizations of inor ganic fillers and fibe 23 have confirmed that HSP concepts can b applied to engineered fibe -filled systems such as those based on polypro ylene. and perhaps too simple. and enhancing pigment dispersion.. 24 The behavior of chewing gum can also be analyzed in terms of solubility parameter principles.

14(2). July 1996. J. Polym. Chem. L. 16(3). Chem..S. 1985. Chem.. 2. Farbe und Lack.. 1996. Solubility parameters. 2nd ed.. II. 6. 21. 6(2).. T.. 1975.. Solubility parameters. 18.. solvency. R. Polym. 26.A. Boundary Lubrication — Scientific and Technical Applications Forecast. A. Polym. 28(3).. C. K. Standen. J. 1967. Lecture for the Danish Society for PolymerTechnology.. and Utracki. J. Ind. 8. Prod. 22. 23–28. This is true e ven in complicated situations.. Coat. Schröder. Hennissen. C. E. 1971. 1978. 3. On application of the three dimensional solubility parameter to the prediction of mutual solubility and compatibility. 83(7). A. Paint Technol. Solubility in the coatings industry . Systematic Modification of Filler/Fibre Sur aces to Achieve Maximum Compatibility with Matrix Polymers.M.. in Kirk-Othmer Encyclopedia of Chemical Technology. 13(6)... 8. Hansen. in Danish). Thermodynamic analysis of phase separation in self-stratifying coatings — solubility parameters approach. 19. Appl. Coat. . and Wu..M. AD747336.. Färg och Lack. C. Rome. 1968. Sci. Hansen. 20. 77(8). 1723–1730. Eng. 14. Kemi. J. 1995. HSP-solubility parameters: a tool for development of new products — modelling of the solubility of binders in pure and used solv ents. Ind. Prod.. J. G. 889–910.M. Prog.A. 323–333. and Utracki. Department of the Army. 1977.Eur. Ind. Frisch... 19. 1976. Coat. 473–478. 40(516). Copenhagen. Ne w developments in corrosion and blister formation in coatings. Boca Raton. and Teas. C. Dev. Solubility and Solvents for Conservation Problems. Hansen. 266–268. and Beerbo wer. pp.. J.. Ed. Dev. 5/98. J.. Hansen. 14(3). 113–120.R. Vol. Hansen. Eng.A. Beerbower. International Centre for the Study of the Preservation and the Restoration of Cultural Property (ICCR OM).. M. 1972. Some aspects of acid/base interactions (Einige Aspekte der Säure/Base-W echselwirkung.L. Solubility. Interscience. 1991. and Long. 17. 16. Special issue de voted to self-stratifying coatings. Characterization of Coatings: Physical Techniques. P art 2: Measurements. Office of the Chief of Research and D velopment.M. Pigmented systems. solubility parameters. 18–22.P.L... Surf. skandinavisk tidskrift for Fårg och Lack.. 9. 12..F. 334–340. 4.. 51–62. C. 2nd ed. 2. Int. 1992. P art 1: Theory.S. C. 595–598. Graphic analysis of resin solubilities. A.M. Färg och Lack. in German).. Klein. printing inks.M. 15..C.. Teas. D. Res. A.. 00153 Rome) 13. J. C. Coat.. Res. REFERENCES 1..M.148 Hansen Solubility Parameters: A User’s Handbook among them. Rasmussen.. Handbook of Solubility Parameters and Other Cohesion Parameters. Org. M. 24(3). A. FL. Champagne. 73(8). D. Networks Blends. The three dimensional solubility parameter and its use.. Vol.W. 7. Via Di San Michelle. 17(4).F . (13. 45–56. Eng.M. 132–138. Res. A. Interfacial tension in polymer blends.S. 63(795). Dev. 11(2). in Treatise on Coatings.M. Interf acial tension in polymer blends. J. Chawla. 1971. 1998. and other filled polymer system including chewing gum. Use of solubility parameters for the preparation of hemodialysis membranes. Marcel Dekk er.. Misev. chap. 2nd ed. 23. 207–210. Ne w York. Org. H. Colloid chemistry aids to formulating inks and paints. 1983. and Wahlström.. Luciani.. 1996. and Smith. 1977. 10. L. such as the formulation of v arious types of filled systems including coatings. J. Assymetric membrane formation. 11. 25 years with solubility parameters (25 År med Opløselighedsparametrene. 5. L. 6(1).K. Coat. and Chang.. Van Dyk. Washington. 1994.. Hansen. 1968. J. Eds. 41–50. Prod. 1991. Ne w York.P. Myers. 1996. February 10.. adhes ves. CRC Press. Networks Blends. 69–77... 1968. Dan.. Hansen.. 19–25. E. D. 1972. Torraca. A. T. P art II.F.T. Technol. Barton. Prog. Gardon. Suppl. 28–42. Champagne. Luciani. Skaarup. Solv ents in w ater-borne coatings..

S. 1994. J. Assoc. M.501. IL. 1986.286. Oil Colour Chem. U. Chicago.H. 25.. H. February 15. Company. Parameters controlling colour acceptance in late x paints. 69(3).G..A. 53–61. assigned to Wm. 5.. Petroleum Wax-Free Chewing Gums Ha ving Improved Flavor Release. J. Patent No. Song. . Wrigley Jr. and Reed. Stephen.Applications — Coatings and Other Filled Polymer Systems 149 24.


Road surf aces can also be grey to white in color. ” It consists of a mixture of hydrocarbons of different molecular sizes containing small amounts of heteroatoms such as sulfur . each representing a direction in the HSP space. and Crude Oils Per Redelius ABSTRACT Hansen solubility parameters (HSP) are sho wn to be a useful ne w tool for understanding compatibility relations among bitumens and crude oils. nitrogen. Bitumen behaves as a viscoelastic thermoplastic solid at ambient temperature and turns into a viscous liquid at high temperature. which is mix ed with crushed rock aggre gate. Bitumen and crude oils are comple x mixtures of hydrocarbons which are k ept in solution mainly by their mutual solubility . thus eliminating expensive trial and error testing. Bitumen. we all know the “black” roads on which we drive every day. A more precise determination of the extension of the ellipsoids can be found by turbidimetric titrations with three dif ferent titrants. Bitumen is a semisolid material that can be produced from certain crude oils by distillation. respecti vely. Although the solubility of the hydrocarbons is mainly determined by the dispersive interactions. as well as traces of metals like vanadium and nickel. it is not possible to mak e correct estimates of their stability without also taking polar interactions and h ydrogen-bonding interactions into consideration. SYMBOLS SPECIAL TO CHAPTER 9 C P FR pa po Amount of bitumen/total amount of solv ent and titrant Stability index given by Equation 9. Defined by Equation 9.Solubility Parameters 9 Hansen of Asphalt. INTRODUCTION Even if most of us are not familiar with bitumen. which ma e it ideal in the manuf acture of asphalt for road 151 .3 Volume of solv ent/total volume of solv ent plus titrant Defined by Equation 9. They are not colloidal dispersions as previously thought. in which case an alternative binder has been used: Portland cement concrete. It can also be found in nature as “natural asphalt. Relations between the HSP of dif ferent materials is visualized using 3D plots showing the HSP as ellipsoids. The majority of road surf aces are black because the binding agent used to manuf acture the surfacing is bitumen. HSP ha ve proven their ability to gi ve a good estimate of the stability of bitumen and/or crude oil ha ving different origins in relation to solv ents and polymers. and oxygen. It is no w possible with the help of simple laboratory e xperiments to predict the consequences of dif ferent courses of action. It presents unique adhesive and waterproofing properties.

. so the de velopment of new PMB has to a lar ge extent been done on a “trial and error” basis. Although the reserves of such crude oils are v ery large worldwide. most scientists agree that the y have been formed by de gradation and transformation of ancient or ganisms. whereas others are semisolid materials that ha ve a viscosity making them impossible to handle at room temperature. 50/70. A large number of dif ferent polymers ha ve been tested as modifiers for bitumen.152 Hansen Solubility Parameters: A User’s Handbook construction and to use in a wide range of industrial application. whereas the Ring&Ball gi ves the stif fness close to the highest expected temperature in practice. Sometimes bitumen is confused with tar. The low viscosity crude oils contain lar ge amounts of fuel b ut very little bitumen. Another reason for modification with polymers is to increase durabilit . Tar is produced by dry distillation of coal or w ood. there have been very few tools for prediction of compatibility between the polymer and bitumen. It is also common. In the U. the dynamic shear rheometer (DSR) — has made it possible to describe the full rheology of bitumen. In Europe bitumens are graded according to their penetration at 25 °C — for e xample. and bitumen has been used for about 100 years.S. As the road construction area is very conservative. This will be improved if a proper polymer is selected. The main restriction for the choice of polymer is the e xpected improvement of the rheological properties in comparison with the price of the polymer .S. most tests are empirical and have been used for a long time. which is a product of completely dif ferent origin. The better understanding of the true nature and the solubility properties of bitumen provided by Hansen solubility parameters (HSP) has given a new tool for understanding of polymer compatibility with bitumen as discussed in the follo wing. the term asphalt is used instead of bitumen. Although the true origin of crude oils is still under discussion. The most common process for production of bitumen is by distillation under acuum v of properly selected crude oils. Until now. There are ho wever just a limited number of crude oils which permit direct distillation to proper bitumen grades suitable for production of road asphalt. It is desired to ha ve a reasonable stif fness of the binder e ven at the highest surf ace temperatures a road can reach on a hot summer day . ho wever. The main reason for modification of bitumen is to impr ve the rheological properties. The penetration gives a measure of the stiffness of the bitumen at most common service temperatures of a road.. for e xample. a viscoelastic material with a comple x rheology and can thus not be completely described by simple penetration testing and softening point. The properties of crude oil v ary depending on age and conditions during formation. where the tw o numbers gi ve the highest and lo west limit for the particular grade. Some crude oils are liquids with lo w viscosity. as well as being reasonably fl xible at the lo west temperatures on a cold winter day. to use viscosity gradation based on viscosity at 60 °C. and the high viscosity crude oils contain v ery little fuel b ut large amounts of bitumen. particularly in the U. In the end just a fe w of them ha ve reached lar ger commercial use. it is used for hea vy crude oils. from w aterproofing in constructio to sound dampening in the automoti ve industry. whereas in Canada. Bitumen is. But e ven more important is the compatibility or the solubility of the polymer in the bitumen. MODELS OF BITUMEN Crude oils ha ve been found in man y places around the w orld. they are not primarily produced as the y contain too small amounts of fuel. Two of the most common tests are penetration at 25 °C and softening point Ring&Ball. if an y. Thus. During the last 20 years we ha ve seen an increase in the use of polymer modified bitume (PMB) with impro ved properties. The development of modern and reliable rheometers — for e xample. which is the most important and profitable product for refiner The functional properties of bitumen are usually related to its use as binder in asphalt for roads. particularly to mak e the binder less sensiti ve to temperatures. In Europe the term is defined as ab ve. The term bitumen is not completely unambiguous as it has been gi ven different meanings in different parts of our world. the most common properties are related to the rheology of bitumen.

aromatics. and particularly the residue after distillation. 5 and the micro structural model as a result from the SHRP development program in the U. Bitumen. e xcept in the case of polymers. There is no major population of identical molecules. The constituents in these fractions have been characterized by modern analytical techniques until almost every single component has been identified. In spite of the solubility model . and sulfur . The material is black and viscous. From these data tentati ve structures of the molecules ha ve been suggested. These fractions have been given names like “resins. The most common approach for chemical characterization of bitumen in volves a separation into generic fractions based on chromatographic principles.S. ethane. and Crude Oils 153 From a chemical point of vie w crude oil is an e xtremely complex mixture of h ydrocarbons. The dispersion is assumed to be stabilized by the resins. The smallest molecules are the gaseous methane. Usually small amounts of heteroatom lik e nitrogen.6 Recent research has shown that the asphaltenes do not form micelles b ut are soluble in the maltenes. The separation of crude oils into different fractions is done in refineries by distillation. ha ve escaped such detailed characterization. There are se veral reason wh y this has been a superior challenge: • • • • • The number of dif ferent molecules is v ery large.8 These models describe bitumen as a solution of organic material of different polarity and dif ferent molecular weight ha ving a kind of mutual solubility in each other. The heaviest molecules ha ve molecular weights higher than 1000 and are thus h ydrocarbons with 70 carbon atoms or more. which is found in most books and papers on bitumen chemistry . The precipitation beha vior of asphaltenes is what could be predicted from regular solution theory and could be described as flocculation. 1 In fact hardly any one single molecule from the complex mixture has been chemically analyzed. The precipitated fraction is called asphaltenes and is defined as the fraction of bitumen that is insoluble inn-heptane. as well as trace amounts of metals like vanadium and nickel. and thus no micelles can exist in the bitumen. The lo w-boiling fractions consist of g asoline and g as oils. what is the mechanism for its stabilization?” Other models that question the e xistence of micelles ha ve also been proposed. The heavier fractions (heavy gas oil). 2 The model was later refined by Pfei fer and Saal. 3 Although the model might be attracti ve for mechanical engineering. The boiling point is higher than approximately 450°C. it is more dif ficult to accep for an organic chemist. oxygen. most of the understanding is in terms of a verages of dif ferent chemical functional groups or structures. teaches that bitumen is a colloidal dispersion of asphaltenes in maltenes.Hansen Solubility Parameters of Asphalt. Bitumen may be produced only after proper distillation process of a selected crude oil using v acuum. particularly since colloidal dispersions of h ydrocarbons in other h ydrocarbons are rare. Examples of models are the continuous thermodynamic model by P ark and Mansoori 4 and Buduszynski et al. and propane.The most common separation procedure is called SARA analysis (saturates.” “aromatics.. It consists of two principally different steps: first. although the content v aries depending on type and origin of the crude oil. When a solvent such as n-heptane is added to the system. The first one t introduce this concept w as Nelensteyn (1924). making the molecules a firly large. The range of molecules of dif ferent polarities and sizes is continuous. resins. creation and precipitation of a solid fraction by dilution of the bitumen wit n-heptane. and asphaltenes). These are dissolved in the liquid h ydrocarbons. and then a separation of the soluble fraction with respect to polarity . P art of the system precipitates.” and “saturates. The n-heptane soluble fraction is named maltenes and is further separated by polarity into three more fractions.” The most common and widespread h ypothesis about the structure of bitumen. are found. the balance is disturbed. Although the residual oil and bitumen have been e xtensively analyzed with modern equipment.7. with the di ferent fractions being collected based on their boiling points. A number of questions are immediately raised: “Do the asphaltenes have enough different chemistry to permit dispersion rather than dissolution?” and “If it is a colloidal dispersion. Most residual oils are further upgraded by dif ferent refining processes to fuels.

se veral h ydrocarbons of lo wer molecular weight that are not soluble inn-heptane (for example. depending on the source of asphaltenes. transport. coronene or dibenz(a. Thus. Most of the characterization w ork has been conducted on precipitated asphaltenes. This indicates that the asphaltenes associate in solution. and FIMS lik ely gives a more correct value. This assumption might have been correct if the asphaltenes were a colloidal fraction in bitumen. which might range from as lo w as 300 up to more than 1000. and very little attention has been gi ven to asphaltenes in their natural en vironment in the bitumen. ho wever. Recent studies with fluorescence depolarization techniques h ve confirmed the FIMS results 13 It may be speculated that lar ge size molecules are less soluble in n-heptane. ved surprisingly lo w ASPHALTENES During production. Much confusion has come from the misuse of the term asphaltenes to mean all kinds of precipitates from bitumen. As will be pro ven later in this chapter the cause of formation of precipitates is more related to general solubility rather than just solubility in n-heptane. There is a clear definition of the term asphaltenes11.These results also vary depending on crude oil source. it has recei acceptance in the research on bitumen and crude oils. It is orth discussing some of the more common statements about the chemistry of asphaltenes and to compare them with e xperimental facts. This varies with different crude oils b ut may be estimated as being 500°C. and thus asphaltenes should consist mainly of high molecular weight material. suggesting that the insolubles inn-heptane could represent precipitates in general.h)anthracene. less for polyc yclic aromatics and more for nalkanes. The true molecular weight of asphaltenes has been under discussion for man y years. 1 Other attempts to determine molecular weight using field ionization mass spectrometry (FIMS) r veal an apparent molecular weight of 700–1000. This roughly corresponds to h ydrocarbons with 35 carbon atom. but this it is not the case. A high dependency of molecular weights on solubility is well kno wn from polymers. if the formation of precipitates could be controlled. There are.10 as well as mechanisms of formation of the precipitates. . It is thus reasonable to assume that the lowest molecular weight in the asphaltenes is equal to the smallest molecule with a boiling point at the cut-point of the bitumen. The apparent molecular weight is strongly dependent on the solv ent.12 as the material that precipitates on dilution of bitumen or crude oil with n-heptane. e xtensive research has been performed to in vestigate the chemistry of asphaltenes9. The mechanism of precipitate formation is certainly not only an academic matter b ut is of major importance for the whole oil industry as precipitates may cause blocking and fouling of equipment used in crude oil production as well as in transport and refining. The deposits have been claimed to be asphaltenes. and refining of certain crude oils there are sometimes problems wit the formation of precipitates and deposits. MOLECULAR WEIGHT A general statement about the molecular weight of asphaltenes w ould be that they are high molecular weight material. although there might be a risk that some de gradation has tak en place in the ion source. and similar molecules may be part of the asphaltenes fraction. 5 It is ob vious that the VPO overestimates the true molecular weight due to interactions between the molecules. Investigations using vapor phase osmometry (VPO) on precipitated asphaltenes dissolv ed in different solvents have shown molecular weights from 1000 up to 10000.154 Hansen Solubility Parameters: A User’s Handbook being a more precise description of the true nature of bitumen. where the very high aromatic content leads to v ery high dispersion parameters compared with the relatively low dispersion parameter for n-heptane in the HSP concept). and therefore there is considerable interest in them and mone y spent to sa ve. The conclusion is that the asphaltenes fraction lik ely consists of a range of molecules of different molecular weight.

These are polarizable and thus may act as polar molecules in contact with other polar molecules. It is thus more correct to state that the asphaltenes are not polar in a chemical sense. and the claims that they are highly polar have without any doubts been misleading in the attempts to understand the role of the asphaltenes in bitumen and crude oils. when applied to bitumen. The first complication comes from the act that bitumen is very black.Hansen Solubility Parameters of Asphalt. Another complication is that se veral solvents . and methanol. ho wever. the asphaltenes remain mainly nonpolar. cannot be applied to such complicated mixtures as bitumen. The best approach is probably to mak e solubility tests of the material in a lar ge number of solvents with known solubility parameters and then try to find the best verage of the good solvents. 15–20 The focus in these investigations was to study the onset of precipitation of asphaltenes and their solubility properties. for e xample. Use of common methods based on ph ysical and chemical parameters. In spite of this. 14 This is not more than about one to three nitrogen and oxygen atoms per asphaltene molecule assuming a molecular weight of about 1000. and it is rather difficul to see with the e ye whether the solution is clear or not. The choice of good solv ent was usually toluene and the poor solv ent was usually n-heptane. Ev en this seemingly simple approach is rather complicated.24. but not all solv ents in this range were good solv ents. This is not enough to mak e them particularly polar . and Crude Oils POLARITY 155 Asphaltenes are claimed to be a “highly polar” fraction in bitumen. toluene. This indicates that the prediction could be slightly impro ved if more than three types of interactions are used. in contrast to the more nonpolar oils (maltenes). ho wever. As bitumen and crude oil mainly consist of h ydrocarbons. ho wever. but they might be considered as more polar than the other hydrocarbons in bitumen and crude oil. As nitrogen and oxygen are the only atoms in asphaltenes that could contrib ute significantly to a permanent polarit . probably because there are other interactions between the molecules that are not tak en into consideration. SOLUBILITY PARAMETERS OF BITUMEN The first attempts to determine the solubility parameters of bitumens were made using the Hilde brand solubility parameter concept. b ut the best estimation w as given by the Hansen three. as long as it is in accordance with the definitio of asphaltenes. In these in vestigations traditional systems using ratios between a good solv ent and a poor solv ent are used. This statement is based on the f act that asphaltenes are insoluble in n-heptane. a nonpolar solvent. easily soluble in relati vely nonpolar solvents like benzene. When the solubility properties of bitumen are e xtended to more v aried types of solv ents than aromatic and aliphatic hydrocarbons. and particularly not if the assumed colloidal model would be correct. The asphaltenes are. but sometimes other n-alkanes were used. the simple Hildebrand solubility parameters were belie ved to gi ve a good prediction of solubility properties. This sho ws that the Hildebrand solubility parameters are not appropriate for bitumen. The authors of this paper and others23 found that using two-dimensional solubility parameters gives a better description of the solubility properties. In fact. the good agreement with the Hildebrand solubility parameter is to some extent lost. 21 The authors of Reference 22.25 There are still some de viations. This approach gives reasonably good results. whereas the y are insoluble in polar solv ents lik e w ater. and dichloromethane. b ut this will make the model unnecessarily complicated. found that all good solv ents for bitumen fall between = 15 MP a1/2 and = 23 MP a1/2. Bitumen. glycerine. an estimation of the relati ve polarity can be made by considering the amounts of nitrogen and oxygen atoms compared to the amount of carbon atoms. be increased by the content of polyaromatic compounds in some asphaltenes. it is not completely evident that solubility parameters should be applicable for such mixtures. which easily can estimate the solubility parameter of pure compounds.dimensional solubility parameter. Elemental analyses have revealed that the total amount of oxygen and nitrogen in the asphaltenes is usually lo wer than 4%. The apparent polarity might. Determination of solubility parameters of bitumen and crude oil is rather complicated as these consist of a very complex mixture of hydrocarbons.

In most cases the samples were left to dissolv e for at least 24 h and sometimes for up to 48 h. The range of solubility parameters is slightly higher than claimed in Reference 21. ranging from completely soluble to completely insoluble. The bitumens were selected to co ver a wide v ariation of dif ferent properties. Some samples were taken from the mark et. 3. The effect of the mutual solubility is that a higher concentration of bitumen results in better solubility . The third complication is that one has to tak e the mutual solubility between the bitumen molecules into consideration. for e xample.8 MP a1/2 and = 25. In all experiments. 2. In the testing of the solubility we find that most sol ents give a kind of partial solubility with more or less residue. HILDEBRAND SOLUBILITY PARAMETERS Solubility data for 15 different bitumens are given in Table 9. b ut if the staining of the filter paper is a uniform darkish br wn. Totally dissolved: no residue by filter paper test Almost totally dissolv ed: light residue w as noticed by filter paper test Partly dissolved: large residue w as noticed in dark bro wn liquid. polar interactions. a fi ed concentration was used in all experiments to get comparable data. In uncertain cases a drop of the solution can be placed on a filte paper. h ydrogen bonding. although the final calculation requires only “soluble” or “not soluble ” In an e xperiment using 15 different bitumens. All solvents with no visible residue (grade 1) were considered as “good solv ents.8 MP a1/2. If a black dot appears at the spot of the drop.5 g bitumen was dissolved in 5 ml solv ent. 26 Each le vel of solubility w as designated as a solubility grade according to the follo wing rules: 1. it does not contain an y precipitate. As it is so difficult to estimate true solubilit . One may speculate that the reason could be the presence of other kinds of interactions in bitumen such as. and some were made e xperimentally for this purpose. Very slightly dissolv ed: mainly residue in bro wnish liquid. As bitumen is v ery black. but it is also ob vious that several “poor solv ents” are found in the same range. 0. the solution contains precipitate. Not dissolved: colorless or almost colorless liquid. . which is contradictory to normal solubility theory that teaches that a saturation le vel for the solute is reached. In this case solubility becomes better for higher concentrations.156 Hansen Solubility Parameters: A User’s Handbook may partly dissolve the bitumen. the solubility w as determined in 6 dif ferent grading levels. TESTING OF BITUMEN SOLUBILITY Solubility testing may be used for calculation of the solubility parameters of bitumen by the method given in the follo wing. it is sometimes better to gi ve a grading of the solubility in several steps.” A bar diagram of the solv ents for bitumen No. leaving a small precipitate or residue.” and all others were considered “poor solv ents. and also indicates that a better correlation may be achie ved when more interactions are tak en into consideration. it is sometimes dif ficult t notice small traces of precipitate. which is probably due to a slightly different selection of solvents and bitumen types. If these interactions are of significant importance. it xplains the poor correlation with the Hildebrand solubility parameter .Thus. It is e vident that the majority of the “good solv ents” can be found in a range between = 17. 6. It is known that the solubility of bitumen is concentration dependent. Slightly dissolved: large residue w as noticed in red-bro wn liquid. 1 in relation to the Hildebrand solubility parameter is given in Figure 9. 4. The results confirm the earlier findings that the Hildebrand solubility parameter is of little or value to predict solubility properties or compatibility between solv ents or other materials with bitumen. or π-interactions between the molecules.1. 5.1.

. and Crude Oils 157 TABLE 9. and Name 7 – Acetone 11 – Acetophenone 46 – Aniline 52 – Benzene 92 – 1-Butanol 102 – n-Butyrolactone 115 – y-Butyrolactone 122 – Carbon tetrachioride 148 – Chloro benzene 182 – Cyciohe xanol 209 – Diacetone alcohol 234 – Dichlorobenzene 255 – Dieth ylether 263 – Dieth ylene glycol 297 – Dimeth ylformamide 303 – Dimeth ylsulfoxide 306 – 1.Hansen Solubility Parameters of Asphalt. Bitumen.2-Dichloroethane 368 – Eth ylene glycol 375 – Eth ylene glycol b utyl ether 376 – Eth ylene glycol eth yl ether 380 – Eth ylene glycol meth yl ether 397 – F ormamide 417 – n-He xane 438 – Isophorone 456 – Methanol 481 – Meth ylethyl ketone 491 – Meth ylisobutyl ketone 521 – N-Meth yl-2-pyrrolidone 524 – Meth ylene chloride 532 – Nitroethane 534 – Nitromethane 536 – 2-Nitropropane 584 – Prop ylene carbonate 585 – Prop ylene glycol 617 – Tetrahydrofuran 637 – Toluene 649 – Trichloroethylene 698 – Xylene 1 4 1 4 1 5 2 5 1 1 3 5 1 3 6 5 5 3 5 6 4 1 6 3 5 5 6 3 1 6 3 2 3 1 5 5 4 6 6 1 1 1 1 2 4 2 4 1 5 2 5 1 1 4 5 1 3 6 4 5 3 6 6 4 1 6 3 5 5 6 3 1 6 3 1 3 1 4 5 4 5 6 1 1 1 1 3 4 2 4 1 5 3 5 1 1 5 5 1 3 5 4 5 3 5 5 4 1 6 4 5 5 6 3 1 6 4 3 3 1 4 5 4 6 6 1 1 1 1 4 4 3 4 1 5 2 5 1 1 3 5 1 3 6 4 5 3 5 5 4 1 6 3 5 5 6 3 1 6 3 2 3 1 4 5 4 5 6 1 1 1 1 5 4 3 4 1 5 2 5 1 1 3 5 1 3 6 5 5 3 6 6 4 1 6 3 5 5 6 3 1 6 3 2 3 1 5 5 4 6 6 1 1 1 1 Bitumen Sample – Code No.1 Solubility Test of 15 Different Bitumens in 42 Different Solvents Solvent HSP No. 6 7 8 9 10 11 4 1 4 1 5 2 6 1 1 3 5 1 2 6 4 5 3 6 6 3 1 6 3 5 5 6 3 1 6 3 2 3 1 5 5 5 6 6 1 1 1 1 Solubility Grade 4 4 4 4 1 1 2 2 4 5 4 4 1 1 1 1 5 5 5 5 2 3 3 2 5 5 5 5 1 1 1 1 1 1 1 1 3 4 3 3 5 5 5 5 1 1 1 1 2 3 3 2 6 6 6 6 4 5 4 4 5 5 5 5 2 3 3 3 6 6 6 6 6 6 6 6 3 3 4 4 1 1 1 1 6 6 6 6 3 3 3 3 5 5 5 5 5 5 5 5 6 6 6 6 3 3 3 3 1 1 1 1 6 6 6 6 3 3 3 3 2 3 2 2 3 4 3 3 1 1 1 1 4 5 4 5 5 5 5 5 4 4 4 4 6 6 6 6 6 6 6 6 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 2 4 1 5 2 5 1 1 3 5 1 2 6 4 5 3 6 6 4 1 6 3 5 5 6 3 1 6 3 2 3 1 4 5 4 6 6 1 1 1 1 12 4 2 4 1 5 2 5 1 1 3 5 1 2 6 4 5 3 5 5 4 1 6 3 4 5 6 3 1 6 3 2 3 1 4 4 3 5 6 1 1 1 1 13 4 2 4 1 5 2 5 1 1 3 5 1 2 6 4 5 3 6 6 3 1 6 3 5 5 6 3 1 6 3 2 3 1 4 5 4 5 6 1 1 1 1 14 4 1 5 1 5 2 5 1 1 5 5 1 3 6 4 5 3 6 6 4 1 6 3 5 6 6 3 1 6 3 2 3 1 4 5 4 6 6 1 1 1 1 15 4 2 4 1 5 2 5 1 1 5 5 1 3 6 4 5 3 6 6 4 1 6 3 5 5 6 3 1 6 3 3 3 1 5 5 4 6 6 1 1 1 1 Note: The solubility is graded from 1 (completely soluble) to 6 (completely insoluble).4-Dioxan 325 – Ethanol 326 – Ethanolamine 328 – Eth yl acetate 367 – 1.

2-DICHLOROETHANE 2-NITROPROPANE 1.158 Hansen Solubility Parameters: A User’s Handbook FORMAMIDE ETHYLENGLYCOL ETHANOLAMINE PROPYLENGLYCOL METHANOL DIETHYLENGLYCOL PROPYLENCARBONAT DIMETHYLSULFOXIDE ETHANOL γ-BUTYROLACTONE DICHLORBENZENE NITROMETHANE DIMETHYLFORMAMIDE GLYCOLMETHYLETHER GLYCOLETHYLETHER METHYLPYRROLIDONE NITROETHANE ANILINE CYCLOHEXANOL 1-BUTANOL CHLORBENZEN ACETOPHENONE GLYCOLBUTYLETHER 1.0 20.5 FIGURE 9.4-DIOXAN METHYLENECHLORIDE ISOPHORONE ACETONE TETRAHYDROFURAN METHYLETHYLKETONE TRICHLORETHYLENE BENZENE TOLUENE ETHYL ACETATE o-XYLENE C-TETRACHLORIDE n-BUTYLACETATE DIACETONALCOHOL METHYLBUTYLKETONE DIETHYLETHER n-HEXANE 0.1) in dif ferent solv ents of kno wn Hildebrand solubility parameter.1 in Table 9.1 Solubility of bitumen No 1 (T able 9.0 MPa0. each gi ving a quantitati ve v alue for the dispersion (D).0 Hildebrand solubility parameter 35. HANSEN SOLUBILITY PARAMETERS (HSP) The data set for bitumen No.0 10.0 15. HSP consists of three components.1 was used for testing whether HSP gi ves a better model for bitumen solubility than Hildebrand solubility parameters. polar (P).0 5. gray bars = good solv ents.0 30. and h ydrogen bonding (H) .0 25. White bars = poor solv ents.

” All other solv ents were considered as “poor solvents. whereas the solubility in Table 9.” Although the dif ferent bitumens are selected to represent a range of products produced di e siv er sp 25 ter in ac tio ns . Ev en if we cannot completely rule out the possibility that there e xist other types of interactions. The program permits only tw o le vels of solubility . conclude that the HSP estimate is good enough.2.” This confirms that the solubility properties of bitume can be reasonably well predicted by HSP . however.2 where all “good solv ents” are f alling within a certain region separated from the “poor solv ents. transport.1 w as determined in 6 grades.” In the first calculation only the bes solvents (grade 1) were selected as “good solv ents” and then in a second calculation the tw o best grades (1 and 2) were tak en as “good solv ents. in a 3D. It can be assumed that the same situation is valid also for crude oils. It is ob vious that the calculated HSP for the dif ferent bitumens become slightly dif ferent. bitumen No. the HSP were calculated using tw o dif ferent criteria for “good solv ents. THE SOLUBILITY SPHERE Chapter 1 includes a discussion of a computer program called SPHERE for calculation of the best estimated HSP as well as the radius of the best fitted pseudo sphere.” The result is illustrated in Figure 9. respectively. where each axis is one of the Hansen solubility parameters. The suitability of HSP may be illustrated by using a three-dimensional (3D) diagram where each axis constitutes one of the interactions.1. F or comparison.” based on a set of solubility data. and processing. and Crude Oils 159 Bitumen 25 ctions polar intera insoluble ▼ soluble 20 15 10 5 ▼▼ ▼ 0 ▼ ▼ ▼ ▼ 15 ▼ 20 25 0 20 15 ng 10 nd i o b 5 en rog hyd FIGURE 9.” The results are listed in Table 9.Hansen Solubility Parameters of Asphalt. which indicates that the use of HSP w ould be a valuable tool.1 for calculation of the best estimate for HSP for 15 bitumens. we may. it seems like the latter two types of interactions are still of fundamental importance for understanding the properties of bitumen. depending on the choice of solubility grade for the “good solvents.2 Plot of the solv ents in Table 9. and particularly for understanding the true nature of bitumen. for crude oil production. Bitumen. The program w as applied on the data in Table 9. x-y-z plot. also. interactions. “good solv ents” and “poor solv ents.” ho wever. All solv ents with a solubility grade 1 were considered as “good solv ents” and all other grades as “poor solv ents. which includes the “goo solvents” and e xcludes the “poor solv ents.1. Although the “good solv ents” are found in a re gion of relatively high dispersion interaction and relati vely low polar and hydrogen bonding interaction.

0 2.0 2.7 2. sho w that there is a lar ge v ariation in performance of the bitumens. If solv ents giving a small residue (solubility grade 2) are accepted as “good solvents.0 6.” the v ariation between the binders becomes more e vident.56 H 3.5 4.2 show.8 5.8 4.9 18.3 5.0 3.0 4. It is e xpected that some of these dif ferences should be reflected i the different chemical compositions of the bitumens and that these same dif ferences should also be reflected in the HS .5 5.7 6.5 in the same units.5 5.0 4.9 17.5 4. The “sphere” radius (RAD) is 5.3 5.2 MP a0.1 4.9 17.9 17.3 3.9 6.4 17.5 4.1 5.8 5. the dif ference in HSP is surprisingly small.0 18.3 3.1 3.1 5.1.8 4.5 6. that there are only very small differences.5.5 4.9 18.5 5.9 17.5 4.5.5 5.0 2. The small variation in HSP between the dif ferent bitumens is a result of the small v ariation in solubility as seen in the data from Table 9.5 3.0 5.8 5.8 4.3 3.4 17. Although the 15 bitumens listed in Table 9.0 17. .6 7.3 3.13 from different crude oils as well as dif ferent process conditions.64 RAD 5.5 5.0 4.5 4. however.4 17.5 5.4 17. The results given in Table 9. such as styrene/b utadiene/styrene (SBS). P = 4.5 4.5 17.” If “grade 2” is also accepted as “good solv ent. but the variation between the different binders becomes more e vident.0 2.5.3 5.2 2. The main general trend is a small shift to ward lower h ydrogenbonding interactions and a lar ger radius of the solubility sphere.1 are primarily intended for use in the water-proofing industr . Laboratory e xperiments.2 Calculated HSP for 15 Bitumens Using Two Levels of Solubility as the “Good Solvents” Grade 1 = “Good Solvents” Bitumen 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Average D 18.5 17.6 MP a0. particularly when calculated with only the best solv ents as “good solv ents.9 17.5 5.9 MP a0.6 6. which varies to a lar ge extent.5 4.23 RAD 5.1 3.1 17. with only a fe w solvents giving different solubility for different types of bitumen. One example is the compatibility with polymers. The larger radius is an e xpected consequence when more solv ents are accepted as “good solv ents.7 4.9 18.2 3.6 6.8 4.8 5.0 18.1 2.5 5. and H = 3. In applications where bitumen is used — for e xample road building and w ater proofing — i is well kno wn that bitumen produced by dif ferent methods and from dif ferent crude oils ha ve different performance.” The decrease in h ydrogen bonding is more dif ficult to xplain.9 17.9 18.6 5.0 17.0 17.9 17.2 2.2 3.0 3.5 5.7 5.160 Hansen Solubility Parameters: A User’s Handbook TABLE 9. but a comparison with kno wn composition and performance still does not allo w a simple correlation.6 17.46 Grades 1 and 2 = “Good Solvents” D 17.0 17.2 2.3 3.0 4.63 H 3.91 P 4.6 6.8 5.3 3.7 2.5 5.9 2.5 5.3 3.8 3.3 3. The average HSP for bitumen based on calculations using “grade 1” as “good solv ents” are D = 17.72 P 5.8 4.5 4.1 3.0 18.2 3.3 4.3 3.9 18.3 5. and fiel experience for some of the samples.2 3.9 17.5 5.5 5.3 4.4 17.” one still gets a very similar average HSP.5 4. they are selected and manufactured to cover a wide variety of crude sources as well as dif ferent types of manuf acturing processes.7 4. b ut it might indicate that the “sphere” is not completely symmetrical.1 17.

A disadvantage with the SPHERE program is the lack of a tool for plotting the ellipsoid in a diagram that w ould be beneficial for illustration purposes Thus. an improved program which permits 3D plotting of the ellipsoid w as developed. And finall . If the HSP of other types of materials than bitumen are going to be measured.5. Another approach is to perform turbidimetric titrations to estimate the e xact HSP at the precipitation point calculated from the ratio of a “good solv ent” and a “poor solv ent” at precipitation. base oils. polar.5 MP a0. P = 3. which is similar to the binder used to obtain the data reported in Table 9. the HSP is D = 18. also in the petroleum area. and a radius of 5. is in f act an ellipsoid (spheroid). ” with the three dif ferent types of interactions as coordinates.” There should be an option to mak e other types of fitting than the SPHERE program t the available data.4 MP a0.2. This requires. and h ydrogen bonding interactions is desired. The better precision may be achieved in two ways. These solvents have RED between 0 and 2 related to the estimated HSP of bitumen as presented above.” “good solv ents. During the development it was discussed that although the factor 4 has been proven to be a good approximation for most materials there might be complex mixtures which could give a better fit with xperimental data if other f actors were used.3. as these are close to the boundary .” and “borderline solvents. The new program has the follo wing features: • • • • • Permits plotting of the HSP solubility ellipsoid in a 3D diagram. and the radius of the sphere is 6.3 in the same units. petroleum w axes. Bitumen. of course. As much v ariation as possible with respect to the dispersion. The input data should be based on “poor solv ents. The first impr vement is to use a better selection of solv ents for the solubility testing. whereas most bitumens are mixtures of h ydrocarbons where the differences in chemical properties are relatively small. nontoxic and inexpensive solvents are preferred.1 are selected to co ver a large area in the 3D solubility space.5.5.9 MP a0.0 MP a0. Negative values of HSP interaction coef ficients are not ta en into consideration. and Crude Oils 161 This lack of correlation is without an y doubt disappointing. This approach is further discussed as BISOM titrations below.5. A suggested set of solv ents.Hansen Solubility Parameters of Asphalt. optimized for determination of HSP of bitumens and similar materials is presented in Table 9. 9 (D = 17. Solv ents that have HSP close to the border of solubility are preferred to better define the borde .5.3 MP a0. distillates. it is suggested that other sets of solv ents may be needed to get the best precision.2.3.5. Obviously it is necessary to have much better precision than the solubility testing as shown in Table 9. We may however speculate that it is mainly due to a lack of precision. COMPUTER PROGRAM FOR CALCULATION AND PLOTTING OF THE HANSEN 3D PSEUDOSPHERE The SPHERE program described in Chapter 1 has gi ven very good approximations of the HSP as well as the diameter of the (solubility) sphere for a lar ge number of materials. H = 3. When using this set of solvents for a Venezuelan binder. A comparison can be made with binder No. In the SPHERE program. a f actor 4 is used as a multiplier for the dif ference in the dispersion interactions of the species concerned.5 MP a0. The RED (relative energy difference) concept is discussed in Chapter 1.6 MP a0. etc. that an approximate HSP of the material is already a vailable. This means that the “sphere. This set of numbers is dif ferent from the pre viously estimated v alues in Table 9.5). P = 4. An improved selection of solvents should focus on solvents with RED v alues around 1. H = 3. Permits plotting of up to three ellipsoids representing dif ferent materials in the same 3D diagram. The solvents in Table 9. . Examples are light crude oils.

3 15.7 5.3 6.7 13.8 19.9 17.4 P 8.2-Dimethoxybenzene Ethyl acetate Ethyl benzene Ethyl lactate 2-Ethyl-hexanol Ethylene glycol dib utyl ether Hexadecane Hexyl acetate Isopropyl acetate Laurylalcohol Mesityl oxide Methyl acetate Methyl benzoate Methyl ethyl k etone 1-Methyl naphthalene Methyl oleate 3-Methyl-2-butanol Methylene dichloride Nitrobenzene Oleyl alcohol Pyrrolidine Salicylaldehyde Tetrahydrofuran Tetrahydronaphthalene 1.9 4.2 17.6 3.1 15.3 8.5-Tetramethylbenzene Toluene 2-Toluidine D 19. 56 93 717 1060 118 955 156 182 183 184 188 194 1019 791 269 1084 889 328 333 353 345 758 412 419 440 1063 450 464 472 481 500 502 1029 524 531 546 1051 704 617 618 885 637 953 Solvent Benzophenone 2-Butanol 2-Butyl octanol Butyraldehyde Caprolactone (epsilon) 1-Chloro pentane Chloroform Cyclohexanol Cyclohexanone Cyclohexylamine Cyclopentanone cis-Decahydronaphthalene 1.1 7.7 5.2 14.4 16.2 9.5 10.1 7.1 6.3 4.7 2.0 7.7 16.8 17.0 6.4 6.6 19.5 15.0 15.8 3.0 9.7 8.5 17.2 7.5 1.6 14.2 3.4 17.9 3.9 5.4 1.4 5.4 7.2 15.8 17.0 19.8 14.0 17.8 5.4 Solubility 1 0 0 0 0 1 1 0 1 1 0 1 1 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 0 1 0 0 1 0 0 1 0 1 1 1 1 0 .5 HSP No.5 3.2 0 5.3 17.7 3.8 16.3 3.4 15.9 0 7.9 0.0 14.2 16.7 4.8 H 5.1-Diethoxy ethanol (acetal) Ethylene glycol monoethyl ether acetate Diisopropylamine 1.5 5.8 18.162 Hansen Solubility Parameters: A User’s Handbook TABLE 9.8 19.6 18.2 0 2.3 6.2 6.7 3.0 20.9 15.1 4.3 15.5 9.6 7.3.4 5.4 5.7 16.1 6.3 0.2 9.6 15.0 0.2 1.9 8.9 19.6 5.7 5.1 8.5 5.5 11.2.5 0 2.7 14.5 9.4 12.4-Dichlorobutane 1.1 4.1 15.2 20.6 4.0 2.8 16.8 6.1 1.0 16.2 16.6 18.5 2.1 4.6 10.0 0.9 18.6 2.4 14.8 4.7 13.1 11.8 17.2 8.9 5.6 18.3 9.6 6.3 Solvents Used for Determination of the Solubility of Bitumen with Their HSP in MPa0.

2. 6.27 The program permits 6 different kinds of fit to create a 3D bod . The computer program hsp3D was developed on a MATLAB platform.3. based on a lar ge set of solubility data.0 H 6.5 HSP No. which probably is the best figure to describe the solubility properties.4-Trimethylpentane o-Xylene D 18.Hansen Solubility Parameters of Asphalt.8 4. 648 653 667 670 698 Solvent 1. b ut with v ariable coefficients for the three a es (the three types of interactions).8 P 5. In each case all good solvents are included and all poor solv ents are e xcluded. From Figure 9. and Crude Oils 163 TABLE 9.2-Trichloroethane Tricresyl phosphate 1. which is similar to the Axis-aligned ellipsoid above but allows the body to rotate and tilt to obtain a better fit. Minimum enclosing ellipsoid is the body with the smallest v olume that encloses all the “good solvents. In the normal Hansen fi a factor 4 is used for transformation of the dispersion interactions. The Hansen fi is the same type of fit as in the SPHERE program using Equation 1. .” The features of the impro ved computer program hsp3D were further e xamined using the solubility data from Table 9. Rotated ellipsoid with convex hull center and volume. which is similar to the Hansen fit a ve. In all of the fits a ve it is assumed that the axis of the ellipsoid is aligned along the three ax es. so the results compared to the SPHERE program might be slightly dif ferent. number of outliers. and fitting coe ficient.1 17. Rotated ellipsoid fit.2 19.0 0 3. 5.3 12. and at the same time also optimize the transfo mation factors for the ax es. 1. Bitumen.3 1. Convex hull fit which could be described as the points for the good sol ents being wrapped with a fl xible membrane.0 0 1. The results from the dif ferent available fits were compared in 3 diagrams with three different fits in each (Figure 9.1. The body for this fit has the sam center coordinates and v olume as the con vex hull b ut attempts to align with the “good solvents” to minimize distance to its surf ace. Axis-aligned ellipsoid fit.3 and Figure 9.3 (CONTINUED) Solvents Used for Determination of the Solubility of Bitumen with Their HSP in MPa0. in the axis-aligned fit this actor as well as the factors for the other two axes are adjusted to optimize the fit 4.9 seems also to be v alid for such complicated mixtures as bitumen. 3. as it is the truest body constructed withou approximations. This might be the first choice if di ferent materials are going to be compared.0 14. ” This set of solv ents better defines the boundary r gion as discussed in the te xt. In the rotated ellipsoid the program can tilt the ax es to impro ve the fitting.4 we see a comparison between the con vex hulls.9 The search algorithm is ho wever slightly dif ferent. This indicates that the f actor 4 in Equation 1.3 it is vident that there is a v ery small difference between the resulting ellipsoids using dif ferent fitting algorithms Transformation or tilting of the axis did not give any major improvement compared to the SPHERE data. Another way of comparing the quality of the fit using the di ferent algorithms is to compare some indicators like volume.5 1.4). In Figure 9. 2.4-Trimethylbenzene 18.0 2. This fit ma es use of only the good solv ents.1 Solubility 1 0 1 0 1 Note: Good solv ents are indicated with a “1” and poor solv ents are indicated with a “0.2.

The best solubility body is the one ha ving the smallest volume. most of the corners mathematically will f all outside the ellipsoid. This sho ws that the approximation with an ellipsoid is rather rob ust. and the highest fitting coe ficient. most outliers.4 that the HSP for the particular Venezuelan bitumen.11. and rotated ellipsoid are compared It can be seen in Table 9. which gi ves the smallest volume. If a body with multiple corners is transferred to an ellipsoid with the same v olume. Hansen fit. and less good fitting. The Hansen sphere and the axis aligned ellipsoid gi ve almost the same results. ” In almost all of the literature about bitumen and asphalt it is . The rotated ellipsoid gi ves a smaller volume b ut at the e xpense of more outliers and less good fitting. bitumen is usually di vided into fractions that are defined by the selection of the separation method. and most likely also for other bitumens. The most e xtreme case is the ellipsoid with the same center point (HSP) and the same v olume as the con vex hull.164 Hansen Solubility Parameters: A User’s Handbook Venezuelan bitumen soluble Venezuelan bitumen insoluble 14 12 δH H-bonding 10 8 6 4 2 0 15 10 5 16 δP Polar 0 δD Dispersive The ellipsoids according to 18 20 FIGURE 9. the least number of outliers. It is almost impossible to isolate chemically uniform fractions. a result of the algorithm. axis-aligned ellipsoid. the definition of asphaltenes is the material that precipitates upon dilution of bitumen (or oil) with nxtraction of n-heptane soluble heptane.12 The fractionation could also be considered as an e molecules from the bitumen. ” The asphaltene-free fraction from bitumen is called “maltenes. also be due to an uneven selection of solvents rather than properties in the material. lea ving a residue named “asphaltenes. This might. COMPONENTS OF BITUMEN Bitumen is a very complex mixture of different hydrocarbons but yet with very similar properties. instead. Perhaps the most commo separation of bitumen is the precipitation of asphaltenes from the maltenes.This is. of course. is more or less independent of the fitting method. The f act that the coordinates are dif ferent indicates that the con vex hull is sk ewing for this material compared to the Hansen Sphere. however. As stated abo ve.3 3D solubility body of bitumen using computer program hsp3D.

5 3.4 18.0 18. The isolated maltene fraction is also tested for solubility in the same set of solv ents. That this is not correct can easily be pro ven by solubility testing and plotting of the solubility ellipsoids using the hsp3D program.987 0.6 4. erroneously claimed that the asphaltenes are dispersed in the maltenes as a colloidal dispersion.4 P 3.1 4.798 0.5 confirms that there is no verlap of the HSP for n-heptane and the ellipsoid for asphaltene. and also the minimum enclosing ellipsoid.4 Plots of fit of bitumen solubility data using hsp3D sh wing the Convex hull model.9 3. and Crude Oils 165 Venezuelan bitumen soluble Venezuelan bitumen insoluble 14 12 10 δH H-bonding 8 6 4 2 0 15 10 20 5 0 16 δD Dispersive 18 δP Polar FIGURE 9.1 3.3 18.980 0. The solubility ellipsoids for the two materials are plotted using the hsp3D program (Figure 9. TABLE 9.4 18.7 Volume 399 399 242 99 371 Outliers 3 3 5 10 6 Fit Coeficient 0.4 4. the ellipsoid with the same center and v olume.3 to Ellipsoids Type of Fitting Hansen Sphere Axis aligned ellipsoid Rotated ellipsoid Ellipsoid: convex hull c and v Minimum enclosing ellipsoid D 18. and it can be considered that the y are so f ar apart that the asphaltenes are not soluble in n-heptane.3.983 Note: Outliers = number of “good solv ents” with RED > 1 + number of “poor solv ents” with RED < 1. It is also vident that the HSP of the .939 0. Asphaltenes isolated by the standard method ASTM D6560 12 have been tested for solubility in the set of solv ents listed in Table 9. Bitumen.9 4.Hansen Solubility Parameters of Asphalt.5).4 Precision Indicators for Fitting the Data in Table 9.1 H 3.6 3. Figure 9. This agrees with the definition of asphaltenes.

less likely as long as the continuum in the asphaltenes and the maltenes is k ept intact. maltenes is dif ferent from the HSP of n-heptane. and thus could still be dispersed rather than dissolved. The temperatures are representative of the highest and lowest temperatures on the surface of an asphalt road during winter and summer . Probably all possible types .166 Hansen Solubility Parameters: A User’s Handbook 14 12 10 δH H-bonding 8 6 Asfaltenes Maltenes 2 0 15 10 20 18 16 0 δD Dispersive n-heptane δP Polar FIGURE 9. The reason is that the continuum has been brok en and w ould probably not reflect the conditions in the original sample In fact. In f act. there is such a lar ge overlap between the solubility ellipsoids of the maltenes and the asphaltenes that the y are quite likely to be soluble in each other . The most common and well kno wn modification is the addition of di ferent types of polymers to bitumen. there is no reason to belie ve that the asphaltenes will appear in the same state in maltenes as inn-heptane. although in reality the maximum and minimum temperatures v ary considerably with geographical location. The reason is to impro ve the lo w temperature properties by making the bitumen softer at v ery lo w temperatures (<20°C) and at the same time mak e the bitumen more stif f at higher temperatures (+60°C). Thus. 28 If a fraction of the “insoluble” asphaltenes is mixed with the maltenes the y might be insoluble. removal of fractions from either the maltenes or the asphaltenes will create a risk for phase separation. This is also the reason wh y one should be v ery careful in making an y claims or predictions of bitumen properties based on the properties of fractions. This strongly suggests that the asphaltenes are not dispersed in the maltenes as a colloidal dispersion b ut are more lik ely dissolved. In some e xperiments the asphaltenes ha ve been further fractionated into “soluble” asphaltenes and “insoluble” asphaltenes.5 Solubility ellipsoids for asphaltenes and maltenes compared with n-heptane. This is. BITUMEN AND POLYMERS It is a v ery common practice to impro ve bitumen properties by adding dif ferent additi ves. The fact that they are insoluble in n-heptane is no e vidence that the y are insoluble in the maltenes. It might be ar gued that some of the asphaltenes molecules with e xtreme HSP might not be soluble in the maltenes. however.

however. Another important factor is the cost ef ficien y. To mak e it simple we selected tw o polymers. which means that a good impro vement of the bitumen properties is achieved with fairly small levels of polymeric additives. and Crude Oils 167 of polymers ha ve been tested in bitumen — for e xample. In case PES is mix ed with bitumen it will be dispersed rather than dissolv ed. based on predictions of compatibility between dif ferent polymers and bitumen. One of the most important requirements is the “compatibility” between the bitumen and the polymer. If the HSP of a particular polymer is not kno wn.6. Neither of these polymers is a common modifier for bitumen. The requirements on such products are. v ery strict. The solubility ellipsoids of the tw o polymers compared to the HSP sphere of Venezuelan bitumen are illustrated in Figure 9. the meaning of “compatibility” is the stability ag ainst phase separation. the maximum le vel of modificatio is typically below 5%. for which solubility data are presented in Chapter 5 and Chapter 18. so in practice v ery fe w polymers ha ve found a wider use as modifiers fo bitumen. as the solubility ellipsoid is almost completely outside the ellipsoid of bitumen. respect vely. In the roofing industry higher l vels are generally accepted as product quality is more important than price. 20 15 δH H-bonding Polyether sulphone 10 5 Bitumen 0 25 20 15 10 5 0 δP Polar δD Dispersive 16 18 20 Polyethylene sulphide FIGURE 9.6 Solubility ellipsoid for bitumen compared with solubility ellipsoid for polyether sulfone and polyethylene sulfide . To illustrate the usefulness we may compare tw o types of polymers with kno wn HSP with the HSP of bitumen.Hansen Solubility Parameters of Asphalt. It is evident that PES is not soluble in bitumen. elastomers. In the road building industry where the volumes are large and the price constraints are strong. Bitumen. The Hansen solubility parameter concept pro vides a good tool for selection of suitable polymers. In this case. polyethersulfone (PES) and polyethylensulfide. and e ven rec ycled rubber and plastics. it can easily be determined with a simple solubility test as described above. tw o component curing systems. plastomers.

sunn y summer days. The effect is the same physical crosslinking as in pure SBS. In reality the situation is e ven more complicated as SBS consists of tw o dif ferent kinds of polymer se gments based on b utadiene and styrene. It is e vident that there is a considerable o verlap between the SBS and the bitumen. respectively. In case of polyeth ylenesulphide the solubility ellipsoid is inside the ellipsoid of bitumen. As the polymer is completely soluble we do expect the effect to be related to the concentration of the modifie. No problem with storage stability is foreseen. This polymer gi ves a good modification e fect at f airly lo w concentration (3–5%). It has been proposed that the same mechanism is also giving a good effect in bitumen. These become plastic-lik e and can be processed at higher temperatures.168 Hansen Solubility Parameters: A User’s Handbook 10 H-bonding SBS 5 Bitumen 0 15 10 Polar 5 18 16 0 14 Dispersion 20 FIGURE 9. at the same time having rubber-like properties at room temperature because of physical crosslinking caused by polystyrene and polyb utadiene being mutually insoluble.7. and therefore SBS is airly f workable at temperatures above 100°C b ut is still rubber -like at lo wer temperatures. SBS belongs to the group of thermoplastic elastomers.This implies partial solubility . Each of these segments has different HSP. Figure 9. The polystyrene blocks ha ve a glass transition temperature of approximately 100°C. The HSP of SBS was determined with a solubility test as abo ve and the solubility ellipsoid was plotted together with bitumen in Figure 9. The effect will be an increase in stif fness at temperatures where PES will act as a solid fille . This risk is highest on hot.7 Solubility ellipsoid of bitumen compared to a solubility ellipsoid of SBS. and thus polyethylenesulphide is e xpected to be completely soluble (compatible). The higher stiffness will decrease the risk for rutting (permanent deformation). The major adv antages are increased stif fness at f airly high temperatures (60°C) and improved fl xibility at low temperature. None of the polymers discussed abo ve have been frequently used for modification of bitumen The polymer most commonly used for bitumen is styrene b utadiene block copolymer (SBS) or similar polymers based on styrene and b utadiene. with the polystyrene being soluble/compatible in bitumen at the mixing temperature (180°C) b ut becoming less soluble or insoluble at lo wer temperatures.7 supports this .

Under certain conditions of storage and transport of crude oils there are sometimes problems with the formation of precipitates and/or deposits. and Crude Oils 169 10 δH Hydrogen bonding 8 6 4 2 0 15 10 δP Polar 5 0 12 14 16 18 20 22 24 δD Dispersive FIGURE 9. picture as the part of the SBS ellipsoid located outside bitumen presumably represent the polystyrene. and the North Sea.8 Comparison between a hea vy Venezuelan crude oil (Laguna) and a light crude oil (Leadon) from the North Sea. These deposits are sometimes blamed on asphaltenes and sometimes on w axes. China. It is sho wn that Leadon is co vering a lar ger space than Laguna and is thus expected to have better solubility properties. Some crude oils are v ery light and contain a lar ge percentage of the most desired products. although this has not been v erified by xperiments. The use of HSP to study the precipitates in comparison with the solubility parameters of the crude oils is a good tool for better understanding of the precipitation mechanism. The heaviest of the crude oils have such a high viscosity that they can not be pumped at normal ambient temperature b ut always have to be handled at higher temperature. Crude oil is pumped up to the surf ace where it is transported by pipeline or ships to refineries for further processing into desired products. The crude oils are v ery dif ferent. North America. depending on origin. whereas other crude oils are heavy and bitumen-rich. Bitumen. CRUDE OIL Crude oil is found almost all o ver the world with large reserves in the Middle East. g asoline and diesel fuel. may be illustrated by 3D plots of the solubility ellipsoids of each crude oil calculated with the hsp3D program (Figure 9. hea vy Venezuelan Laguna and medium Leadon from the North Sea. It is produced by drilling wells in the ground or under the sea.8). just to mention a fe w examples. To have the complete picture it is also necessary to understand ho w temperature and pressure influence th HSP of dif ferent molecules in the crude oil. Venezuela. The difference between tw o crude oils.Hansen Solubility Parameters of Asphalt. This is probably an effect of Leadon being a lighter crude which contains more lo w viscosity oils. These are better solv ents than the . There are lar ge economic benefits to be ained if the problem with deposits could be decreased. Russia. These might decrease the capacity of pipelines by formation of solid contaminants in the crude oil. The exact nature of these precipitates and the mechanism of their formation are not fully understood and is thus the subject for intense research. Only a fe w selected crude oils can be used for production of high quality bitumen suitable for making asphalt for roads.

2). In the case of BISOM titration the selected “poor solv ents” are 2. The selection of components is mainly done on a “trial and error” basis. going from the HSP of the good solv ent toluene to an y of the three poor solv ents. This means that the Laguna crude should be completely soluble in the Leadon crude.4-trimethyl pentane (isooctane). The separation tendency can be o vercome by a proper selection of bitumen. but sometimes. particularly at concentrations of SBS between 10 and 15%. Three nonsolv ents ha ve been selected.170 Hansen Solubility Parameters: A User’s Handbook higher molecular weight components of the Laguna. BISOM TEST The procedure discussed in the follo wing has been developed at Nynas Bitumen based on turbidimetric titrations to precisely determine the boundary of the surf ace of solubility. These ha ve HSP placing them just outside the surface of the solubility ellipsoid. by a better selection of test solv ents and by better methods to optimize the ellipsoid. The ef fect is seen as a higher concentration of bitumen giving better solubility . The HSP of the mixture is proportional to the concentration of each solv ent. and 2-eth yl hexanol. and particularly with modification with SBS polymers. Most mixtures of bitumen and SBS sho w a tendency for separation if the y are stored at high temperature for a long time. This has partly been met. This is seen as a hard precipitate at the bottom of the bitumen tank. and no problems with the formation of precipitates are to be expected by dilution of Laguna crude oil with Leadon crude oil. among bitumens h ving v ery similar solubility ellipsoids (Table 9. The principle of the test method is visualized in Figure 9. This situation is contradictory to what is usually kno wn for solubility of pure substances. The procedure is called BISOM. as discussed abo ve. A typical separation test is made at 180°C for 3 d. but better precision is required than can be obtained with simple solubility testing with pure liquids. an acron ym for BItumen Solubility Model. The separation is usually seen as the polymer floating to the surace. For this purpose. a solvent with a solubility parameter as close to the center of the ellipsoid as possible has to be selected. there is still a desire for better precision. it has to be diluted to decrease the viscosity to permit proper stirring during the titrations. This comes from the f act that bitumen is k ept homogeneous by the mutual solubility of all its dif ferent molecules. The titration can be considered as a dilution of the bitumen with a mixture of a good solv ent and a poor solv ent.9. One of the most important properties is the separation stability. TURBIDIMETRIC TITRATIONS Although the determination and visualization of solubility parameters for bitumen and other mineraloil-derived materials have proven to be very illustrative. also in the bitumen. It is also seen that the solubility ellipsoid for the Laguna is located completely inside the solubility ellipsoid for the Leadon. a phase separation can tak e place. although there are some empirical rules.2. The HSP ellipsoid of the bitumen is constructed using the “poor solvents” illustrated as solid triangles in the figure and the “good sol ents” being illustrated with open triangles. To understand this phenomenon we must consider that the first sign of turbidit . Bitumens that are seemingly very similar with respect to solubility give still very different properties after mixing with SBS. The titration is illustrated as arro ws. A precipitate will appear when the HSP of the mixture has a v alue placing it on the surf ace of the solubility ellipsoid. HSP may be an e xcellent tool for selection of suitable combinations of bitumen and polymer . They may be seen as the black triangles in the center of the circles in Figure 9. An important ef fect to tak e into consideration in the turbidimetric titration of bitumen is the concentration effect. F or the BISOM titration we have selected toluene (or in some cases xylene) as the good solv ent.9. 2-b utanone (methyl ethyl ketone). alternatively by selection of a suitable polymer . As bitumen is a high viscosity liquid or a semisolid material. There are very obvious differences when it comes to practical applications. Impro ved estimation of the best solubility ellipsoid is required for optimal use of the HSP concept.

” In this titration only one good solvent.4 0. were used. The concentration effect was first described by Hithaus 29 He developed a kind of turbidimetric test for what he called “peptization of asphaltenes. toluene. Heithaus performed the titration at several different concentrations of bitumen.10 gi ves an illustration of the principle.10 Illustration of the Heithaus titration of a Venezuelan bitumen. is interpreted as “insolubility . The details of the calculations can be found in Reference 29.9 Summary of the BISOM titration with the “good solvent” and the three “poor solvents” titrants.2 0 VT = volume titrant VS = volume solvent -0. n-heptane. Bitumen.” but more precisely.4 0.6 0. and one poor solv ent. Heithaus Titration 1 FR = Vs/(VS + VT) 0.1 0.Hansen Solubility Parameters of Asphalt.5 -0. .3 0.1 0 0.2 C(g/mL) = Wa/(VS + VT) FIGURE 9. b ut Figure 9. it is the insolubility of the molecule(s) that are least soluble in the particular titrant/solv ent mixture that has been used. and Crude Oils 171 Laguna B180 soluble Laguna B180 insoluble 14 12 δH H-bonding 10 8 6 4 2 0 15 10 2-ethyl-1-hexanol MEK Toluene 20 5 0 18 16 δD Dispersive Iso-octane δP Polar FIGURE 9.2 0. To account for the concentration ef fect.8 0.

During the addition of titrant both FR and C become smaller and smaller. illustrated by the four titration arro ws showing the decrease of the FR and the C v alue during the titration. At infinite dilution FR = 0 and C = 0. the BISOM data may be combined with solubility data for better precision.A straight line is fitted throug the four points using the least R-squared method. . and four black dots indicating the first sign of precipitation.10 the dilution ratio FR = volume solvent/total volume solvent and titrant is plotted against C = amount of bitumen/total amount of solvent and titrant. A higher number indicates a higher stability . F or each experiment the titration is repeated several times using different concentrations of bitumen.2) (9.10 the titration has been repeated four times. The fact that a higher concentration of bitumen requires more titrant to reach the precipitation point confirms that the solubility of bitumen increases as the concentratio increases. The solvent to titrant ratio can be used to calculate the HSP at the precipitation point for infinit dilution of the bitumen. The computer program hsp3D also has the possibility to handle HSP of solv ents or solv ent mixtures considered as being e xactly on the surf ace of the ellipsoid. The results from a BISOM titration may be reported in dif ferent ways. There is a modified ersion of the ATR instrument 31 which is suitable for BISOM titrations.1) (9. In Figure 9. P is a balanced stability index describing the internal stability of the bitumen. At the start of the titration FR = 1 as no titrant (V T = 0) has been added. The most simple is to give the FRmax for the intercept C = 0 and Cmin for the intercept FR = 0. This is the case with HSP calculated from the ratio of good and poor solv ents at the precipitation point. whereas at the same time C is equal to the concentration of bitumen in the solvent.172 Hansen Solubility Parameters: A User’s Handbook In Figure 9. One of the instruments measures the transmission of light through a cuvette with a short beam length30 and the other instrument uses variation in the intensity of a reflected beam of light (attenuated total reflectance principle TR]). Thus. and finall . This effect is also visible with v ery dilute solutions. The meaning of the intercept for FR = 0 is the lo west concentration of bitumen that is needed to gi ve full stability in pure titrant. ut before this a precipitate has been formed. Heithaus29 proposed further calculations of f actors: pa = 1 – FR max ⎡⎛ 1 ⎞ ⎤ p 0 = FR max ⎢⎜ ⎟ + 1⎥ ⎣⎝ C min ⎠ ⎦ P= p0 1 − pa (9. The equation for the line is used for extrapolation to find the intercepts for FR when C = 0 and for C when FR = 0. The point where the first sign of precipitate is noticed i marked with a black dot in the diagram. p0 is related to the solubility po wer of the bitumen. There are at least tw o commercial instruments which can be used for BISOM titrations although they are both originally developed for automatic Hethaus titration. pro vided that the titrant has been properly selected. The meaning of the FR v alue at C = 0 is the ratio of solv ent to titrant where there is solubility for all concentrations of bitumen.31 Both instruments can be equipped with more than one titrant for BISOM titration. The testing procedures used by the instruments are described in tw o ASTM standards. It could also be expressed as the maximum titrant which can be added to bitumen without causing precipitation.10.3) where pa is considered to be related to the solubility of the molecules in bitumen with the lo west solubility. In practice the precipitation point at the BISOM titration can be determined by dif ferent methods. The BISOM data may easily be recalculated from FR max to the FR v alue for the same concentration used in the solubility testing by using the equation for the line in Figure 9.

To have the best precision. The program can also estimate the best coef ficients for the ellipsoid model to illustrat the extension of solubility re gions in a 3D diagram.5. particularly since the e xact center point is not kno wn. however.Hansen Solubility Parameters of Asphalt. A computer program hsp3D can estimate the best fit to the solubility data and from thi calculate the HSP. each of which could be considered to be on the surface of the solubility ellipsoid.5.10. Comparison of HSP for bitumen.9 as a multiplier for the dif ference in the dispersion parameters for the materials being considered seems to be v alid.0 MPa0. P = 3. and bitumen should thus not be considered as a colloidal dispersion as is frequently claimed. and ho w close to insolubility the y are. also for comple x mixtures lik e bitumen. the test solv ents should be selected with respect to the material to be tested. Bitumen. To ha ve a complete picture it is necessary to ha ve several titrants with dif ferent HSP as the traditional determinations of internal stability using only n-heptane as precipitant will usually gi ve an incomplete picture. BISOM titration is well suited for measurement of the internal stability of comple x hydrocarbon mixtures lik e bitumen or crude oils. The BISOM could also be seen as an identification of those molecules that are the leas soluble in the bitumen or crude oil. The use of three titrants permits calculation of the HSP at three precipitation points. The precipitation point at a certain concentration can easily be estimated from data such as are reported in Figure 9. BISOM titration is also a determination of the least soluble molecules in bitumen or crude oils. The best estimated HSP v alues for Venezuelan bitumen are D = 18. H = 3. rather than determination of HSP . The interpretation of the BISOM titration is a determination of the internal stability of the bitumen or crude oil. . although it is based on the principles of HSP. The factor 4 in Equation 1.4 MP a0.6 MP a0. Up to three dif ferent materials can be compared in the same 3D plot. the concentration at the solubility testing has to be tak en into consideration. The program is not limited to bitumen and crude oils. If the HSP calculated from turbidimetric titrations are going to be compared to the HSP ellipsoid of the same material.5. not possible to estimate the solubility ellipsoid from only three points on the surf ace. and maltenes confirms that asphaltene are soluble in maltenes. It is preferred to ha ve solvents near or at the borderline of solubility . asphaltenes. b ut could equally well be used for other types of materials. This procedure is called a BISOM titration. or other types of bitumen or crude oil. and Crude Oils 173 A high internal stability can be seen as a high allo wance for blending with additi ves. It is. A procedure for turbidimetric titrations has been de veloped to further impro ve the precision to determine the surf ace of the HSP solubility ellipsoid. polymers. CONCLUSION • • • • • • • • • • • • The Hansen solubility parameters of a comple x material such as bitumen or crude oil can be estimated by determination of its solubility in a lar ge number of solv ents with known HSP. It is not e xpected that the material precipitated by addition of n-heptane would be the same as that precipitated by the addition of 2-eth yl-1-hexanol or 2-b utanone. solvents.

. Washington. and Scott. J. Soc. Am. Relation of the micelle to the medium in asphalt. Pet. 2–8. J.A. 4. 20.N..Y. Speight. 47(11). 1990. L. 7. Johnson. Neumann.. Rahimian.174 Hansen Solubility Parameters: A User’s Handbook REFERENCES 1.B.. 12. 82.. R.A. 511–514.J.M... November.. F. O. Redelius. June 25–27. P . Chem.. New York. 14. 18.. 10.nynas. D. 265 North 9th Street.H. New York. Hirschberg.M. G. Elsevier Science BV.. 5. 2004... and Zenk e.H..N.. Mitchell. 25. Zenk e. 1087–1092. and Redelius.E. Technol. 1986 (in German)... 1984. A. Turner.. Annual book of ASTM Standards Section 05. H. Nellensteyn. www . 14. September 9–14. 1964. Div... Buch. . N. 1440–1446. E. M... Sirota. Chem. Assoc. 1: Physical Characterization. The solubility of asphaltenes in h ydrocarbon solv ents. 1989 (in German). Robertson. Einfluss der Löslich eitseigenschaften von Asphaltene auf die Rückstandsverarbeitung.G. 2nd ed.. Hildebrand. L. 21. Boduszynski. Wyoming 82070–3380..G. I.F.. 19. J.G. Zum Verhalten organicher Lösemittel ge genüber Bitumen. 27–35. 1984. Kim... Warnez.G. Asphaltenes in petroleum asphalts: composition and formation. Jones. Test Method for Determination ofAsphaltenes (Heptane insolubles) in Crude Petroleum and Petroleum Products. Measurement and modelling of asphaltene precipitating. 278–286. Meet. 134–138. Asphaltene phase beha viour: a molecular thermodynamic model ...J. A. unpublished results from Icopal – Nynas joint research project: Correlation between Bitumen Characteristics and Fire Properties of Roofing Membranes.M. Paper 097. Fuel. H. R. Phase Behaviour and Stability. Proc.J. Soc. 1973...E. Paint Technol. The program hsp3D is a vailable as share w are in Europe from: Nynas Petroleum. C. 42(7/8). Lira-Galena. 13. Molecular size of asphaltene fractions obtained from residuum h ydrotreatment. Characterisation of asphalt by solubility profiles. S. O. Fundamentals and Applications. J. 16. Fuel. Bitumen solubility model using Hansen solubility parameter . 8. Bitumen 1. 49. 6. 1994.M. 1995. Solubility parameters and bitumen .C. 23. Phys. 109–125.F . Pfeiffer. C. Schipper . T. J. Rahimian. Hansen. 9. J... 52.. Energy Sources. G.G. pp..G. 53. Lyon. J.. G. G.com or in USA from Western Research Institute. and Saal..P . www. 1980. Kashou.westernresearch. Test Method for n-heptane Insolubles. 26.A . L.. 2001. Sweden. and Mullins.L.. S-149 82 Nynäshamn.. 2003. D. P. Eng.. 17.E. Aggregation and deposition of hea vy organics in petroleum crudes. J. ASTM D6560-00.S. Branthaver.org. A. 430–435. Marcel Dekker. Asphalt Paving Technol. Binder Characterization and Ev aluation. H. Influence of temperature and pressur on asphaltene flocculation. 3rd Euroasphalt and Eurobitume Congress Vienna. 11... and Bahia.R. J. 1979. Randolf.. 283–293.. and Lathham.C. Harnsberger. R. Buenrostro-Gonzales. Christiansen. Pet.F..03. Asphaltenes. K. New York. Erdöl und Kohle –Erdgas. deJong....U. Characterisation of Heavy Crude Oils and Petroleum Residues.J. Chem. 930–939. J. Burke.V.C.. Plenum Press. 1988. Laux. and Rahimian. 79.J. Laramie.. H. 24.C. S. Do ver. R. Meijer . Pet. 124–143.. Asphaltenes and Asphalts. J. Asphaltenes where are you? Proc. D. and Speight.. Duvall.M. Redelius. 1967.. and Chilingarian. Amsterdam 1994. M. D. R. The three dimensional parameter — k ey to paint component af finities III. 10.. I. Technol. E. 22. Skaarup.. R. Understanding the Physical Structure of Asphaltenes to Optimize Bitumen Manufacture. 44 139–149. E. B. Park. 2000.. D. Yen. Groenzin. Sheu.04. Boduszynski. Anderson.. Strategic Highway Research Program SHRP-A-367. P . M. Pet... Erdöl und Kohle — Erdgas — Petrochemie vereinigt mit Brennstoff-Chemie. Petersen. Asphaltic bitumen as colloidal system. 18. 1994. Inst. J.. I. Mullins. Hagen. Annual book of ASTM Standards Section 04. 149–152.B. 3. and Mansoori. Colloid-disperse Crude Oil Systems. Vol. P. June 1984. The Solubility of Nonelectrolytes..L. 1928. 2.J. and Redelius. F. Hofner . B. Hirschberg. 39(511). P . Assoc of Asphalt Paving Technologists. Independent calculation of the parameter components. 1075–1084.F. 15.I. 2001–2003 27. Hobbs. S. M. 3rd ed. Energy Fuels. ASTM D327997..W. S. 2004. The Chemistry and Technology of Petroleum. 119–137.P. 1991. Andersen. Fuel. J. and Hermans. 1 in Developments in Petroleum Science 40 A. 1940. McKay .

29. H. Colloid Interface Sci. ASTM D6703-01.. Heithaus J.Hansen Solubility Parameters of Asphalt. 271(2). Standard Test Method for Determining Stability and Compatibility of Hea vy Fuel Oils and Crude Oils by Hea vy Fuel Oil Stability Analyzer (Optical Detection). 2004. P. Inst.. 31.. 45–53. and Fogler. Nydén. and Crude Oils 175 28. Measurement and significance of asphatene pepetization.03.. Annual book of ASTM Standards Section 04.. Characterisation of fractionated asphaltenes by UV-VIS and NMR self-dif fusion spectroscopy. 48(458).. Annual book of ASTM Standards Section 05. 30.. M. 1962. Östlund. Bitumen.S.J. J. Pet. 372–380. Wattana.04. ASTM D7112-05a . J. J.-A. Standard Test Method for Automated Heithaus Titrimetry.


pigments. A methodology for e xtending this reference set of HSP v alues to an y temperature and pressure has been de veloped utilizing an equation of state (EOS) of the form P = f(ρ. The methodology adopted for this is based on room temperature solubility of the g as in different liquids of known HSP values.of Hansen 10 Determination Solubility Parameter Values for Carbon Dioxide Laurie L. polar ( δp). so that 2 2 δ t2 = δ 2 d + δ p + δh (10. polar (dipole–dipole and dipole–induced dipole) interactions ( Ep). in 101 liquid solv ents have been g athered and e valuated.2 MP a1/2. HSPs ha ve been applied to biological materials. at 25°C and a CO 2 partial pressure of 1 atmosphere. yielding the v alues: δd = 15. T). 1 The basis of the HSP approach is the assumption that the total cohesi ve ener gy ( E) of a pure compound is made up of the additi ve contributions from nonpolar (dispersion) interactions (Ed).8 MP a1/2. and hydrogen bonding or other specific association interactions including Lewis acid–base interactions ( Eh): E = Ed + EP + Eh Dividing each contrib ution by the molar v olume.6 MP a1/2.2) (10. In addition.1) gives the square of the total solubility parameter as the sum of the squares of the Hansen dispersion (δd). E Ed E p Eh = + + V V V V (10. and h ydrogen bonding ( δh) solubility parameters. and fibers.3) where 177 . as well as the characterization of surf aces. barrier properties of polymers. and δh = 5. δp = 5. fillers. CO2 gas solubility data. Williams ABSTRACT Reference values of the Hansen solubility parameters (HSP) for carbon dioxide (CO 2) have been determined. INTRODUCTION The HSP concept is widely used for selecting suitable solvents for organic compounds.

178 Hansen Solubility Parameters: A User’s Handbook δ2 d = Ed .5 and the total solubility parameter value is calculated from Equation 10. polar . a set of HSP v alues at a single reference temperature and pressure is determined. and δh. δp. The Ostwald coefficient. T is the absolute temperature. Where multiple CO 2 gas solubility data were a vailable for an indi vidual solvent.7) where R is the g as constant. and h ydrogen bonding parameters in Table 10. pCO2 is the partial pressure of CO 2. reduced to 0°C and 0.6) The mole fraction of dissolv ed CO 2 can then be calculated using ⎡⎛ RT ⎞ ⎤ = ⎢⎜ + 1⎥ 0 ⎟ Lp V ⎢ ⎥ ⎝ ⎠ CO 1 2 ⎣ ⎦ −1 xCO2 (10. an average value for that solv ent was used.1 MP a. and xCO2 is the mole fraction of dissolv ed CO 2. it has been necessary to correct the reported v alues of CO 2 gas solubility to a common set of units. In the follo wing section.1 MP a. All the solubility data sho wn in Table 10. Ω. is defined as the v olume of g as.5) where KH is the Henry’ s law coefficient. all measured at the same temperature. hereafter called data set #1. Using the collected v alues of xCO2 at 25 °C and pCO2 = . L. published data of CO 2 gas solubility at 25°C and a CO 2 partial pressure of 1 atmosphere in a lar ge number of liquid solv ents are e valuated. under a gas pressure of 1 atmosphere.3 If the solubility is small and the gas phase is ideal. V δ2 p = Ep V . Those liquids that show the highest solubility for the gas are assumed to have HSPs closer to those of the g as than those liquids that ha ve lo wer solubilities for the g as. the Ostwald coefficient is independent of total pressure and these tw o coefficients are simply related by 4: L= T Ω 273 2 (10.1 MPa.3. and V10 the molar v olume of the pure solvent. or has been corrected to pCO2 MPa using Henry’s law.1 MP a. T. xCO2. and δ 2 h = Eh V (10.4) The determination of HSPs for compounds that are gases at ambient conditions is usually based on room temperature solubility of the g as in a range of dif ferent liquids of kno wn δd. From this data.1. The Bunsen coefficient.1 are from Hansen’s solublity parameter handbook.1 w as either e xperimentally determined at a CO 2 partial pressure of 0. METHODOLOGY Literature values of CO 2 solubility in v arious liquid solv ents at 25°C ha ve been collected and are summarized in Table 10.2 xCO2 = pCO2 KH (10. The Hansen dispersion. the Bunsen coefficient or the Ostwald coefficient. . noted earlier . is defined as the ratio of the v olume of g as absorbed to the v olume of the absorbing liquid. In addition to correcting CO 2 partial pressure to 0. or as one of the tw o dimensionless quantities. Generally . dissolv ed per unit v olume of solvent at a system temperature. HSP v alues were calculated based on a simple weighted a verage using the data set in Table 10. solubilities ha ve been reported as mole fraction of dissolv ed CO 2.1.

4 18.5 14.1 7.1 3.0 7.0 0.0 8.4 10.00734 0.00938 0.0 0.00912 0.20 23.00774 0.8 17.50 22.50 21.20 22.0 6.6 4.0 17.2 10.1 2.26 21.7 5.4803 0.00962 0.0 7.5 19.3 15.77 16.7 5.4 10.34 22.02800 0.1 3.3 10.8 17.4979 0.4401 0.1 3.0 5.0 0.01903 0.9136 1.74 20.02092 0.0 11.00487 Aniline 0.95 18.1407 0.8309 0.5114 0.20 20.0 0.0 7.51 17.05 16.3 4.4 10.2129 0.4 19.01896 0.2151 0.3444 0.1 6.8 16.4897 0.4 18.1 δh (MPa)1/2 13.00880 Benzyl chloride Bromobenzene Butanol Butanol Butanol 2-Butanol t-Butanol 0.8279 0.01089 0.00887 0.43 21.4095 0.80 16.5576 0.2 17.00760 0.8 17.3 5.5 13.94 19.8635 0.4 10.0 10.3 3.4757 0.19 21.5 20.01126 Methyl ethyl benzene Cycloheptane Cycloheptanone Cyclohexane Cyclohexane Cyclohexane Cyclohexanol δd (MPa)1/2 14.02452 0.5 16.4 10.0 7.3948 0.95 18.47 18.26 17.9865 0.6 0.1 7.2 10.4283 0.94 19.02612 0. 29 30 29 29 30 31 32 33 34 35 29 30 29 29 29 30 33 3 30 29 36 29 29 29 32 37 38 39 40 38 29 29 41 35 36 29 29 34 29 34 30 29 42 43 44 2 42 45 0.2183 0.95 17.0 0.43 18.0 0.0708 1.94 19.1 5.8 15.01122 0.4 19.0 0.7 10.8 17.5 10.00725 Butyl oleate Butyric acid Carbon disulfide Carbon disulfide Carbon tetrachloride Carbon tetrachloride Carbon tetrachloride Chlorobenzene Chlorobenzene Chloroform 0.81 19.4 10.8193 0.3380 0.01277 0.01896 0.4 18.1 0.00721 0.5 15.8279 0.00981 0.0 4.7 3.0 7.00771 0.6 0.8 19.58 18.9 20.3872 0.3319 0.8 15.80 22.4 19.51 20.4201 0.0 0.3166 1.00789 0.00660 0.1 CO2 Solubility in Various Solvents at T = 25ºC and a CO2 Partial Pressure of 1 Atmosphere Mole Fraction CO2 0.2121 0.94 17.8 5.1 5.00904 0.51 20.40 Ref.4 7.94 19.01140 0.0 16.1 5.01235 0.56 17.3148 0.62 23.00718 0.7 14.2882 0.7 5.3983 0.4 δp (MPa)1/2 8.6 0.0 7.4918 0.0 4.0940 0.4899 0.0 2.5 δt (MPa)1/2 21.8 15.6 0.26 21.0 2.2129 0.8 20.6220 0.5 16.5 15.1 15.3207 0.51 18.6 0.2 0.50 22.6934 0.2057 Formula C2H4O2 C2H4O2 C4H6O3 C3H6O C3H6O C3H6O C3H6O C3H6O C3H6O C7H14O2 C7H14O2 C7H14O2 C5H11Br C5H11Cl C6H7N C6H7N C6H7N C6H7N C7H6O C7H6O C 6H6 C 6H6 C7H7Cl C6H5Br C4H10O C4H10O C4H10O C4H10O C4H10O C22H42O2 C4H8O2 CS2 CS2 CCl4 CCl4 CCl4 C6H5Cl C6H5Cl CHCl3 CHCl3 CHCl3 C9H12 C7H14 C7H12O C6H12 C6H12 C6H12 C6H12O Solvent Acetic acid Acetic acid Acetic anhydride Acetone Acetone Acetone Acetone Acetone Acetone Amyl acetate Amyl acetate Amyl acetate Amyl bromide Amyl chloride Aniline 0.37 22.3 2.01876 0.8 16.5 15.00486 Aniline 0.8 20.3 4.01100 0.4 19.01008 0.8 14.0 19.8 1.6 0.01059 0.50 22.3843 0.94 19.8 0.81 17.0 15.7 0.00215 0.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 179 TABLE 10.5 15.01297 0.81 17.0 16.4 19.00471 .4624 0.2 0.26 17.3367 0.01977 0.5394 0.01424 Actual/ Ideal 0.02790 0.5 5.2 7.1432 0.7 5.58 19.01121 Chloroform Chloroform 0.5665 0.7 5.8 15.8 15.37 21.4 4.1 5.80 16.7 5.75 15.00328 0.2 10.2 5.0 15.3 3.20 23.5 17.2 14.0 5.00487 Benzaldehyde Benzaldehyde Benzene Benzene 0.2 13.3 3.5 15.00493 Aniline 0.0 2.4 3.29 19.2 16.8 17.1 17.0 0.6 2.5 15.0 4.01171 0.1 6.01588 0.5 14.8 16.2227 1.02259 0.3135 0.4 0.

2-Dibromoethane 0.70 15.01250 0.01258 0.87 20.5197 0.50 17.47 16.0 0.29 21.64 36.01318 0.01428 0.8 17.9 6.00491 0.04 22.00642 0.90 14.0 13.8 17.2 9.02300 0.68 26.2-Dibromoethane 0.6 6.9 15.5 16.0 1.3-dichloro-2-propanol Dichloromethane Dimethyl cyclohexanone Dimethyl sulfoxide Dimethyl sulfoxide Dimethyl sulfoxide Dimethyl formamide 1.4 17.6201 0.4 4.0 0.30 15.4393 0.00809 0.5202 0.5459 0.6441 0.01202 0.17 20.3 7.01600 0.7 12.00838 0.7 δh (MPa)1/2 5.0 29.8 19.5801 0.8 10.7 15.00725 0.3 0.0 16.2 10.17 20.01290 0.01475 0.2804 0.4 7.1 16.17 20.83 21.0 16.4125 0.3962 0.16 15.5212 0.9921 0.0 12.0 2.0 7.7166 0.4 18.01174 0.1 16.4 18.52 26.30 16.2 11.4 6.66 22.7 6.11 20.9 14.3509 0.01420 0.4469 0.0 7.6 6.5069 0.00009 0.0 0.5 15.4 13.8 15.3 0.87 26.40 6.0 8.68 26.6235 0.7336 0.5633 0.30 14.4 11.7031 0.0 0.8 5.5 17.20 17.00220 0.0 15.8 11.5256 0.85 20.1 15.01006 0.8 8.2996 0.00688 0.8 0.0 0.4 19.0 0.1 3.5 17.0 0.9 15.4 15.56 17.9 0.0 7.01023 0.3014 Formula C6H10O C8H16 C8H16 C5H10 C5H8O C10H22 C10H22 C10H22O C2H4Br2 C2H4Br2 C2H4Br2 C2H4Br2 C3H6Cl2O CH2Cl2 C8H14O C2H6OS C2H6OS C2H6OS C3H7NO C4H8O2 C12H26 C12H26 C12H26 C12H26O C2H6O C2H6O C2H6O C4H8O2 C8H10 C2H4Cl2 C2H6O2 C10H12O2 C5H9NO2 C3H8O3 C7H16 C7H16 C7H16 C7H16 C7H16O C16H34 C16H34 C6H14 C6H14 C6H14O C5H5I C5H12O C4H10O C4H10O Solvent Cyclohexanone Cyclooctane Cyclooctane Cyclopentane Cyclopentanone Decane Decane Decanol 1.0 5.01133 0.4 17.0 0.50 16.1 0.00690 .0 0.3 15.5495 0.00642 0.70 20.3546 0.4946 0.8 17. 46 42 47 42 48 32 49 50 34 29 30 45 29 38 51 44 52 33 32 38 53 49 32 32 30 31 38 38 38 29 38 29 54 29 53 49 36 52 32 32 49 32 49 32 29 30 2 29 0.0039 0.2 5.8 19.3 0.68 24.0 8.01328 0.6 6.5754 0.00686 0.30 21.86 20.02272 0.00804 1.4-Dioxane Dodecane Dodecane Dodecane Dodecanol 0.01194 0.0 0.6 7.32 20.0 0.0 11.00973 Actual/ Ideal 0.0 0.4 6.6987 0.3481 0.0 0.3004 0.7 5.0 0.3 15.8 15.2 18.1 19.15 17.98 26.90 20.9 14.47 16.00746 0.7908 0.01641 0.3258 0.5 18.4249 0.0 8.01811 Ethanol Ethanol Ethanol Ethyl acetate Ethyl benzene Ethylene chloride Ethylene glycol Eugenol N-formyl morpholine Glycerol (glycerin) Heptane Heptane Heptane Heptane Heptanol Hexadecane Hexadecane Hexane Hexane Hexanol Iodobenzene Isoamyl alcohol Isobutanol Isobutanol δd (MPa)1/2 17.0044 0.3 15.00812 1.0 0.8 5.0 17.11 15.0 14.7 1.0 10.9 δt (MPa)1/2 19.2-Dibromoethane 1.66 Ref.2805 0.0 0.5459 0.1 0.3 15.1 13.4 19.52 18.5197 0.3 15.01250 0.8 8.3166 1.01680 0.01610 0.00945 0.0 10.00908 0.0 0.2585 0.1 δp (MPa)1/2 6.180 Hansen Solubility Parameters: A User’s Handbook TABLE 10.5 8.8 15.3530 0.00907 0.4 19.50 22.3965 0.8 17.1 (CONTINUED) CO2 Solubility in Various Solvents at T = 25ºC and a CO2 Partial Pressure of 1 Atmosphere Mole Fraction CO2 0.6 17.2 1.3 16.3 14.01161 0.0 15.00696 0.0 0.00592 0.5 15.30 15.2146 0.5248 0.00 19.4 16.8 17.01191 0.2 15.2-Dibromoethane 0.30 22.2 0.3 8.8 17.4 0.0 0.4 16.01204 0.8 15.80 33.4 9.0 16.00 16.00797 1.01190 0.0961 0.0 0.00 16.0 0.2 10.0 7.0 6.0 0.3041 0.8 16.7 17.0 0.01190 0.30 15.17 24.3 10.5124 0.3659 0.52 26.1 26.

0 14.01210 0.17 27.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 181 TABLE 10.15 22.8 8.1 15.0 7.4266 0.9 4.1 δt (MPa)1/2 16.0 12.7 3.8 6.65 17.00654 0.01570 0.5 15.0 0.5231 0.5094 0.7 5.46 22.1 4.00680 0.01074 0.01481 0.62 19.3285 0.8 17.61 29.01660 0.0 18.6 7.1747 0.6943 0.00927 0.6 16.4062 0.3 12.22 27.01020 0.4 18.2 0.0 16.0 20.8 3.3 15.2 7.00635 Actual/ Ideal 1.5474 0.50 21.22 21.1 4.0 18.4 17.60 24.9 5.61 29.5284 0.3 22.7 4.0 16.6048 0.00752 0.5104 0.6 12.0 0.14 15.2 16.01360 0.46 19.2774 0.1 15.9123 0.5390 0.00957 0.60 19.6465 0.2463 0.02429 0.6 2.2460 0.4455 0.22 27.0 16.1 12.2969 0.8 6.28 19.5 18.4688 0.7 15.3 6.8 15.9118 0.0672 0.01171 0.3489 0.6 1.0 0.3414 0.1 7.6 8.8 19.4046 0.3 7.9 7.98 25.0 19.4 12.5 0.4 22.4180 0.2797 0.00997 0.5939 1.75 21.4 20.11 Ref.00930 0.9 5.3 6.02444 0.60 24.70 16.80 14.20 19.00563 0.5 16.4 16.3 0.3 7.96 15.9 8.01385 0.4 17.2771 Formula Solvent δd (MPa)1/2 15.1 4.2 16.01410 0.3 0.0 19.3316 0.2 16.1 14.6856 0.37 24.01167 0.0 13.95 21.9 5.8 8. 29 29 38 32 30 31 29 29 34 55 56 33 38 38 29 34 38 32 32 32 49 50 38 32 32 29 41 34 57 58 29 29 34 34 29 32 33 38 29 30 34 38 32 32 49 38 38 29 C6H12O2 Isobutyl acetate C4H9Cl Isobutyl chloride C3H8O Isopropanol CH4O Methanol CH4O Methanol CH4O Methanol CH4O Methanol C3H6O2 Methyl acetate C3H6O2 Methyl acetate C7H14 Methyl cyclo hexane C7H12O 2-Methylcyclohexanone C4H8O Methyl ethyl ketone C11H10 1-Methyl naphthalene C19H36O2 Methyl oleate C6H5NO2 Nitrobenzene C6H5NO2 Nitrobenzene C5H9NO N-methyl-2-pyrrolidone C9H20 Nonane C9H20O Nonanol C8H18 Octane C8H18 Octane Octanol C8H18O C18H34O2 Oleic acid C15H32 Pentadecane C5H12 Pentane C5H12O Pentanol 0.01254 0.0 8.0607 0.02088 0.1 4.01169 0.01193 0.75 21.9 12.1 15.5 15.3 9.0 0.59 12.3 3.9837 0.75 21.01234 0.0 3.01590 0.0 0.0 20.01198 0.3983 0.5211 0.39 16.61 18.0 2.02700 0.77 15.0 11.70 20.1 0.58 29.2856 0.03 19.00759 0.7 15.0 15.1 15.9 10.9 16.0 0.8 7.8 δh (MPa)1/2 6.8 14.70 18.00674 0.6 14.8 5.5073 0.3 15.7 2.5 7.4355 0.4 5.6 19.3 7.7249 1.86 16.20 16.1 (CONTINUED) CO2 Solubility in Various Solvents at T = 25ºC and a CO2 Partial Pressure of 1 Atmosphere Mole Fraction CO2 0.5115 0.3 17.6 0.3 22.3 22.3519 0.00806 C7F16 Perfluroheptane C8H7N Phenyl acetonitrile C3H8O Propanol C3H8O Propanol C3H8O Propanol C3H6O2 Propionic acid Propionitrile C3H5N C5H10O2 Propyl acetate C3H6Br2 Propylene bromide C4H6O3 Propylene carbonate C4H6O3 Propylene carbonate C4H6O3 Propylene carbonate C5H5N Pyridine C5H5N Pyridine C5H5N Pyridine C9H7N Quinoline C4H8O2S Sulfolane 0.01231 0.02500 0.01162 0.0 20.00782 0.5376 0.05 21.3 12.0 3.5284 0.0 17.6158 0.00564 0.00977 0.00 23.91 22.0 18.50 21.5 16.0 3.50 15.02089 0.0919 0.5 15.15 22.01677 0.3 14.0 16.65 23.2 0.4 0.7323 1.01210 0.77 15.0 0.02690 0.0 8.00641 0.0 4.3 δp (MPa)1/2 3.2943 1.0 5.0 0.0 19.9 0.1 5.0 6.1790 0.00912 0.3 4.00799 C14H30 Tetradecane C14H30 Tetradecane C4H8O Tetrahydrofuran C10H12 Tetrahydronaphthalene C7H9N m-Toluidine .6 7.3 12.7 5.0 20.00635 0.0 12.01 17.5 17.3 0.0 4.5 20.7 15.61 29.02253 0.0 19.0 19.5 0.

5 15.00 20.0 0.3 42.1 3.0 0.8 14.1 3.0 0.2641 0.40 17.01823 0.00 14.7 2.01387 0.4642 0.3 1.6 0.182 Hansen Solubility Parameters: A User’s Handbook TABLE 10.01042 m-Xylene 0.8) i δ CO 2 p = ∑xδ i=1 n n pi ∑x i=1 n ’ (10.0 2.3 0.0 1. 29 29 55 52 34 38 59 32 33 59 32 32 35 29 30 31 60 29 60 60 0.00994 0.4 6.0 16.0 16.81 47.0 1.5 15.16 18.4-Trimethylpentane Undecane Undecanol 0.3 14.7 16.0 18.4340 0.00994 Toluene 0.5014 0.01010 Toluene 0.00060 0.0 0.0 16.83 14.2.1 (CONTINUED) CO2 Solubility in Various Solvents at T = 25ºC and a CO2 Partial Pressure of 1 Atmosphere Mole Fraction CO2 0.0 11.4411 0.4744 Formula C7H9N C 7H 8 C 7H 8 C 7H 8 C 7H 8 C12H27O4P C2Cl3F3 C13H28 C6H15N C8H18 C11H24 C11H24O H2O H2O H2O H2O C8H10 C8H10 C8H10 C8H10 Note: Solvent o-Toluidine Toluene 0. (10.70 17.8 17.4550 0.01708 Water Water Water Water o-Xylene m-Xylene 0.4 1.2 1.9) i δ CO 2 h = ∑xδ i=1 n i hi ∑x i=1 .16 18.0 4.79 16.10) i .01039 Tributyl phosphate Trichlorotrifluoroethane Tridecane Triethylamine 2.0261 0.0 δh (MPa)1/2 10.01042 Toluene 0.1 3.4552 0.0 18.4 18.3 42.4 15.5 15.0 2.7 16.01175 0.7961 0.6057 0.1 16.4 17.00070 0.4 1.10 17.01087 x ideal CO 2 δ CO 2 d = ∑xδ i=1 n n i di ∑x i=1 i ’ (10.81 47.16 18.3 3.0 1.4 17.0 1.10 16.1 δt (MPa)1/2 22.19 47.5502 0.0306 0.7457 0.4 0.5132 1.0 15. δd (MPa)1/2 19.2 42.0 1.00605 Actual/ Ideal 0.70 Ref.81 47.03550 0.55 18.5 17.0 16.01148 0.0134 0.00059 0.0 18.4 δp (MPa)1/2 4.0229.00061 0.0 0.4 1.81 18.4 17.0 6.70 17.01063 p-Xylene = 0.16 18.3 42.0258 0.0 2.4536 1.02321 0.4339 0.0267 0.

hereafter called data set #2.3 7.02500 0. xCO 2 = 0.0 15.3 6.02429 0.7 9.5 15.1 MPa. A second HSP evaluation was undertaken using a subset of data set # 1.2 7.02253 = 0.7 16.3 15.8 MP a1/2 Data set #2: δd = 15.5 δp (MPa)1/2 6.02300 0.0229.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 183 TABLE 10.5 MP a1/2 δh = 5.3 5.A.02690 0.2 MP a1/2 δh = 5.8 MP a1/2 .6 MP a1/2 δp = 5.02800 0.2 δh (MPa)1/2 4.02444 0.1).8 14.3 3.0 3.4 8.02790 0.03550 0.0 4.3 15.8 14.02700 0.3 5.8 15.3 3. is comprised of the 10 solvents shown in Table 10.4 MP a1/2 δp = 5.1 16.7 6.4 5.2 Solvents Showing Greater Than Ideal Solubility for CO2 at 25°C Solvent Tributyl phosphate (C12H27O4P) Amyl acetate (C7H14O2) Butyl oleate (C22H42O2) Tetrahydrofuran (C4H8O) Methyl oleate (C19H36O2) Isobutyl acetate (C6H12O2) Methyl ethyl ketone (C4H8O) Propyl acetate (C5H10O2) Ethyl acetate (C4H8O2) Methyl acetate (C3H6O2) Note: PCO2 = 1 .7 3. These two evaluations resulted in the follo wing HSP values for CO 2 at 25°C: Data set #1: δd = 16. The subset was chosen to consist of solv ents where the measured CO 2 solubility w as greater than the ideal solubility at ideal 25°C and pCO2 = .1 3.6 7.0229.1 7. (The calculation of this ideal solubility v alue is given in the Appendix 10.6 0.9 3. This data subset. ideal xCO 2 Exptl xCO 2 δd (MPa)1/2 16.2.

whereas an unfavorable interaction is one where the CO 2 solubility is less than ideal at these same conditions. δp. and similar v alues for δp and δd. and is considered a fourth parameter in HSP v alue determinations.The solubility sphere approach is essentially a trial and error method. where R o incorporates all good solv ents and e xcludes all bad solv ents. δh. in a plot of δd vs. kno wn as the solubility sphere.5 ( Ra ) 2 = 4 δd 2 − δd1 ( ) + (δ 2 p2 − δ p1 ) + (δ 2 h2 − δ h1 ) 2 (10. The HSP values derived from the two data sets g ave identical results in terms of the h ydrogen bonding parameter. This solubility sphere approach is essentially a trial and error method whereby all the good solvents are included within a sphere inδd. δh. suspension time of a pigment. δh. For the present e valuation. The radius of this sphere is known as the interaction radius.1 Interaction radius. is then used to plot a sphere in a three-dimensional plot of δd vs. a second approach. Based upon the selected criteria. δh. therefore. In this e valuation. This solubility sphere evaluation is aided by an equation developed by Skaarup for determining the straight-line distance. δp vs. An extended discussion of the validity of this coef ficient is found in Chapter 2. δp. δh between tw o materials based on their respecti ve HSP values. δp. two-dimensional plots are produced for δd vs.11) where δd2. whereby all the good solv ents are included within a sphere in δd. Figure 10. dissolution. δd vs. and δh space. and δp vs. and δh2 are associated with a gi ven solvent and δd1. and δh space while simultaneously excluding the bad ones. The resulting radius for the three plots of δd vs. The criterion of good v ersus bad is arbitrary and is defined based on the particular interaction being e valuated. a f avorable interaction is defined as CO 2 solubility greater than ideal at 25°C and pCO2 = 0. and δp vs. etc. δp2. barrier breakthrough time. Ra. permeation coef ficients higher than a gi ven value. whereas simultaneously excluding all the bad solv ents.1 is a schematic representation of the solubility sphere approach. δp. Ro.1 MPa. should be considered for further study .184 Hansen Solubility Parameters: A User’s Handbook 25 δh (M Pa)½ 20 15 10 5 5 δd (M Pa)½ 10 15 20 25 0 5 10 15 20 0 δp (M Pa)½ FIGURE 10. solvents that ha ve not yet been e xperimentally tested for the desired interaction can be quickly screened and. δh. δp vs. and therefore where the (attracti ve) CO 2–solvent interactions are greater than . we are concerned with optimizing solv ents for their CO 2 solubility. and δh1 with the center of the optimized solubility sphere. It is clear that for cases where CO 2 solubility is greater than ideal. δd vs. This equation w as developed from plots of e xperimental data where the leading constant of 4 in the leading right-hand term w as found to correctly represent the solubility data as a sphere encompassing the good solv ents. δp1. such as de gree of polymer swelling. The adv antage of the solubility sphere approach is that once an interaction radius has been determined. and the three circle radii are adjusted until a single optimized radius for all three plots is found.5–7 was also used to evaluate the published solubility data and resulting HSP v alues for data set #1 and data set #2. Further confirmation is found in Chapter 9. δh. To assist in the determination of a final set of HSP v alues.

δp. T=25°C 1 Data Set #1 2 Data Set #2 20 18 16 14 ½ 12 10 8 6 4 2 0 12 13 14 15 16 17 18 ½ δd (MPa) (a) 19 20 21 22 δP (MPa) Alcohols Alkanes/Nonpolar compounds Halogenated compounds Simple acids/Carboxylic acids Esters Ketones Amides Cyclic Alkenes Nitro compounds Aldehyde Amines Nitriles Anhydrides Sulfoxide compounds FIGURE 10. (b) δh vs. As can be seen from these plots. such as the 101 solvents listed in data set #1. the solubility sphere analysis for data set #1 results in the inclusion of 7 bad solv ents (2-meth ylcyclohexanone. whereas for data set #1. propylene bromide. dichloromethane. (a) δp vs. PCO2 = 1 atm.2-dibromoethane). occurrences of good solvents falling outside of the sphere radius. whereas a RED ≈ 1 represents a boundary condition between good and bad. and 1. In addition. and (c) δh vs.0 for data set #2 incorporates the good solvents. δd. Progressi vely higher values of RED indicate progressively more unfavorable interactions. which has been called the relative energy difference (RED) number5 RED = Ra Ro (10. and Figure 10..12 allo ws for easy scanning of lar ge data sets. and bad solv ents falling inside the sphere radius can be vie wed as an indication of the goodness of the fit. In terms of the solubility sphere technique. Ra should be less than Ro. an interaction radius Ro = 4.2c.2b. c yclohexanone.12) For an indi vidual solv ent. A convenient index for relati ve goodness of a solvent is the ratio Ra/Ro. oleic acid. . δd. trichloromethane. The solubility spheres optimized for data set #1 and data set #2 with the two HSP center points are sho wn in Figure 10. a v alue of RED < 1 indicates f avorable CO 2–solvent interactions.7 is necessary to incorporate the good solvents. whereas the sphere analysis generated for data set #2 results in the inclusion of only 1 bad solv ent (oleic acid). Figure 10. 5 CO2 Solubility in Liquid Solvents.2a. an interaction radius Ro = 4.2 Two-dimensional plots of CO 2 in organic solvents.11 and the RED from Equation 10. Computing Ra from Equation 10.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 185 solvent–solvent interactions.

.3 MPa1/2. δp = 5. 1. 5 who observ ed that the approach of using all solv ents to establish the center of a solubility sphere can result in this sphere boundary (and center) being determined by the poor solv ents or nonsolv ents..2 MPa1/2. the HSP values from data set #2 were selected as the optimum reference values for CO2 at T=25°C.3dimethoxybutane: δd = 15.5 Similar v alues are reported for liquid solvents such as diprop yl k etone: δd = 15. δp = 5.8 MPa1/ 2 This determination is further supported based on problems noted by Hansen.7 MP a1/2.6 MPa1/ 2 δ p = 5. It should be noted. and eth yl acetate: δd = 15. δp = 5.9 MP a1/2. however. and δh = 4. that the CO2 optimum HSP reference v alues correspond to a reference temperature of 25°C and a reference pressure of 90.5 MPa8. a higher operating pressure than found in common industrial applications. T=25°C 1 Data Set #1 22 20 18 16 14 δh (MPa)½ 12 10 8 6 4 2 0 2 4 6 8 10 δp (MPa) (b) FIGURE 10.2 (Continued) 12 ½ 2 Data Set #2 Alcohols Alkanes/Nonpolar compounds Halogenated compounds Simple acids/Carboxylic acids Esters Ketones Amides Cyclic Alkenes Nitro compounds Aldehyde Amines Nitriles Anhydrides Sulfoxide compounds 14 16 18 20 From the refinement of the tw o sets of CO2 HSP values.8 MPa1/2.2 MP a1/2. and δh = 5. δ d = 15.5 MP a1/2.8 MP a1/2. A comparison of these CO 2 values can be made with the lar ge database of HSP v alues found in Hansen Solubility Parameters: A User’s Handbook.6 MP a1/2.186 Hansen Solubility Parameters: A User’s Handbook CO2 Solubility in Liquid Solvents. rather than the best solv ents in the middle. PCO2 = 1 atm.2 MPa1/ 2 δ h = 5. using the solubility sphere methodology . A methodology for e xtending this reference set of HSP v alues to an y temperature and pressure has been developed and is discussed in the follo wing sections. and δh = 7. that is.

n.0 × 10-18 (ESO) or 3. the internal pressure and ced are approximately equal. and where specific interactions (particularly hydrogen bonding) are largely absent. PCO2 = 1 atm.13: ⎛ ∂E ⎞ n ΔE = ⎜ ⎟ V ⎝ ∂V ⎠ T (10.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 187 CO2 Solubility in Liquid Solvents. This same work also demonstrates that n is near unity for nonpolar liquids and also for polar liquids where the dipole moment is less than 2 D . . Hildebrand and coworkers found that for nonpolar/nonassociating liquids. 10 1932. 11 and 1950 12 where the tw o terms.13) where (ΔE/V) is defined by Hildebrand as ced.2 (Continued) ONE-COMPONENT HILDEBRAND PARAMETER AS A FUNCTION OF TEMPERATURE AND PRESSURE Hildebrand’s solubility parameter theory was derived from an approximation of the internal pressure of a fluid. Whereas no direct e valuation of the v alue of n has been found in the literature for carbon dioxide (CO 2). based on w ork conducted in 1928. as shown in Equation 10. a comparison of the v alues found by Hildebrand and others 13–16 strongly suggests that the value of n for CO 2 is e xpected to be near unity . 1 D is equal to 1. This was later termed the cohesi ve energy density (ced). internal pressure and cohesive energy density were found to be related by the quantity .9 1929. and (∂E/∂V)T is the internal pressure .336 × 10–30 Cm.. As a result. where intermolecular interactions are weak. n is not f ar from unity 9–11 and the equality of ced and internal pressure is a good approximation. T=25°C 1 Data Set #1 20 18 16 14 δh (MPa)½ 12 10 8 6 4 2 0 13 14 15 16 17 18 ½ 2 Data Set #2 Alcohols Alkanes/Nonpolar compounds Halogenated compounds Simple acids/Carboxylic acids Esters Ketones Amides Cyclic Alkenes Nitro compounds Aldehyde Amines Nitriles Anhydrides Sulfoxide compounds 19 20 21 22 δd (MPa) (c) FIGURE 10.

using the empirical EOS of Huang et al. '2 '3 '4 ⎝ ⎠ T T T T T '5 ⎟ ⎛ c c ⎞ .188 Hansen Solubility Parameters: A User’s Handbook Accordingly. ⎛ ΔE ⎞ ⎛ ∂E ⎞ ≈⎜ δ=⎜ ⎝ V ⎟ ⎠T ⎝ ∂V ⎟ ⎠T 1/ 2 1/ 2 (10.15) Total (one-component) solubility parameters can therefore be calculated using an EOS of the form. P = f (ρ. This approach is used in this w ork to calculate the total solubility parameter for pure CO 2.18) (10.12 can also be approximated by the square root of the internal pressure for nonpolar/nonassociating fluids.17) ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ( ) ( ) ( ) ( ) ( ) where T ′ = T Tc .16) (10. ⎛ ∂E ⎞ ⎛ ∂P ⎞ =T⎜ −P ⎜ ⎝ ∂V ⎟ ⎠T ⎝ ∂T ⎟ ⎠V so that ⎛ ∂E ⎞ ⎛ ∂P ⎞ δ2 ≈ ⎜ =T⎜ −P ⎟ ⎝ ∂V ⎠ T ⎝ ∂T ⎟ ⎠V (10. b4 = ⎜ c10 + 11 ⎝ ⎠ T' ⎟ ⎛ c ⎞ b5 = ⎜ c12 + 13 ⎝ ⎠ T' ⎟ The state constants (c i) are as defined in the Huang reference.T).19) . ⎛ c c c c c ⎞ b2 = ⎜ c1 + 2' + 3 + 4 + 5 + 6 . Equation 10.14) And the internal pressure can be calculated fromthe thermodynamic equation of state. ρ′ = ρ ρc . Δρ = 1 − 1 / ρ′ (10. Hildebrand's solubility parameter.15. and. defined as the square root of the ced. 17 ⎡ ⎢1 + b ρ' + b ρ' 2 + b ρ' 3 + b ρ' 4 + b ρ' 5 + b ρ' 2 exp p −c21ρ' 2 + b8 ρ' 4 exp −c21ρ' 2 2 3 4 5 6 7 ⎢ ⎢ '2 ⎡ −c ΔT 2 ⎤ + c Δρ exp ⎡ −c Δρ 2 − c ΔT 2 ⎤ P=ρRT ⎢ 25 27 ⎢ +c22 ρ exp ⎣ ⎢ ⎥ ⎢ 27 ⎥ 23 ρ' ⎦ ⎦ ⎣ ⎢ ⎢ 2 2 ⎢ +c24 Δρ exp ⎡ −c26 Δρ − c27 ΔT ⎤ ' ⎢ ⎥ ⎣ ⎦ ⎢ ρ ⎣ ( ) ( )⎥ ⎥ ⎤ ⎥ ⎥ (10. ⎛c ⎞ b6 = ⎜ 14 ⎝ T' ⎟ ⎠ ⎛c c c ⎞ b7 = ⎜ 15 + 16 + 17 ⎝ T '3 T ' 4 T '5 ⎟ ⎠ ⎛c c c ⎞ b8 = ⎜ 18 + 19 + 20 ⎝ T '3 T ' 4 T '5 ⎟ ⎠ ΔT = 1 − T ′. b3 = ⎜ c 7 + 8' + 9 ⎝ ⎠ T T '2 ⎟ ⎛ c ⎞ .

the modified solubility parameter proposed by Ikushima et al. and the EOS model proposed by P anayiotou. It should be noted that there are a wide range of a vailable EOSs for carbon dioxide and a comparison of Huang’s and others can be found in a re view by Span and Wagner. The pressure volume temperature (PVT) EOS that calculates the total (Hildebrand) CO 2 solubility parameter value (Equation 10.16) w as used to determine the combination of pressure and molar v olume ⎛ ⎛ ∂P ⎞ ⎞ corresponding to T = 25°C and δt = ⎜ T ⎜ − P⎟ ⎟ ⎝ ⎝ ∂T ⎠ V ⎠ P = 91.18 These equations and the appropriate deri vatives are then used to calculate CO 2 solubility parameters over the temperature and pressure range for which the EOS is stated to be v alid (220 K ≤ T ≤ 420 K.16 and Equation 10..g. h ydrogen bonding).1 MP a ≤ P ≤ 60 MPa). e.13 cm 3 mole 1/ 2 = 17. no such studies appear to ha ve been done.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 189 25 25 20 δ (MPa)½ 15 10 5 0 25 0 30 20 LIQUID SUPERCRITICAL FLUID δ (MPa)½ 15 10 5 0 GAS 0 0 35 40 0 45 0 0 T (K 10 20 30 40 50 M P( Pa ) ) FIGURE 10. such as Lee-K elser or modified Redlich-Kwong for evaluation).21 This later work utilizes the lattice fluid theory and a lattice fluid h ydrogen bonding model to e valuate solubility parameters and tw o separate components: ph ysical (or van der Waals) and chemical (or specific. and 0. 20 (where the solubility parameter is expressed in terms of reduced parameters).3 Total (one-component) solubility parameter of pure CO 2 calculated using Equation 10. Figure 10.20) .3 is a plot of the resulting one-component solubility parameters.7 MPa VCO2 = 39. ho wever.17. THREE-COMPONENT (HANSEN) SOLUBILITY PARAMETERS — PURE CO2 Extending the HSP methodology to supercritical fluids would significantly enhance the understanding of their solv ent properties. Other notable w orks include Allada’s19 proposed generalized solubility parameter (that uses analytical equations of state (EOSs).4 MP a1/2 that g ave: (10.

originally derived by Hansen and Beerbo wer. A suggested approach for this calculation is outlined as follo ws where the temperature deri vatives.3. Both temperature and pressure will influence total solubility parameters. or supercritical fluid. not found in an y literature search. and n is about 1.23 Generally.21 into Hansen’s definition for the dispersion solubility parameter. are derived in a manner parallel to the temperature deri vatives.23) ⎛ n + 1⎞ ⎛ 1 ⎞ = −δ d ⎜ ⎝ 2 ⎟ ⎠⎜ ⎝V ⎟ ⎠ and.5 for normal (nonassociating or van der Waals) liquids. where the total CO 2 solubility parameter was calculated using Equation 10.P ⎝ 2 ⎟ ⎠ ⎡ − ( n + 1) ⎤ ⎢V 2 ⎥ ⎡V −1 ⎤ ⎦ ⎥⎣ ⎢ ⎣ ⎦ (10. Substituting Equation 10.17. Both of these trends can be seen in Figure 10. The temperature and pressure dependence of individual HSPs as a function of temperature and pressure has apparently not been e valuated for an y liquid. In addition. Similarly.24) . k1/2 δ d = − ( n+1) 2 V (10.5. g as. other than Giddings’ extension of the one-component (Hildebrand) solubility parameter model to supercritical fluids.24 are v erified.21) where k is a constant dependent upon the nature of the particular liquid.7. ∂δ d ∂V = −1. using pressure and temperature inte gral functions.16 and Equation 10.22) allows one to calculate the change in δd produced by a change in v olume by differentiating Equation 10. an increase in pressure at constant temperature will increase the total solubility parameter through an increase in the solv ent density. Ho wever. TEMPERATURE AND PRESSURE EFFECTS ON HSPS: δd Hildebrand12 in his 1950 w ork considered the ef fect of temperature on solubility parameters by recalling the e xpression for the dependence of E on the v olume: E=− k Vn (10. which will be deri ved subsequently. inte gral forms are developed.25 δd V (10.22 there appears to be no published reports on methods to calculate total solubility parameters as a function of pressure and only limited reports on the calculation of solubility parameters as a function of temperature.22. ⎛ ∂δ d ⎞ ⎛ n + 1⎞ = k 1/ 2 ⎜ ⎜ ⎟ ⎝ ∂V ⎠ T . an increase in temperature at constant pressure will decrease the total solubility parameter . Pressure deri vatives.190 Hansen Solubility Parameters: A User’s Handbook HSP values at other pressures and temperatures will be based on this set of HSP v alues.

This is still widely used where the required data are available and reliable. are tak en from Bondi.4 and Table 10.28) (10.4.4 μ ⎞ =− =− ⎜ 1/2 ⎟ ⎠ 2V ⎝ V 2V and ∂δ p δp =− ∂V . 25 From Equation 10. is defined as: δ2 h = Eh V (10.25) A simplified equation w as later developed by Hansen and Beerbo wer. Results of both deri vations are shown in Table 10. Hansen.P ⎝ ∂V ⎟ δp 1 ⎛ 37. noting that there is no rigorous w ay of arri ving at v alues of the temperature dependence of the h ydrogen bonding solubility parameter .4 μ ) ⎜ ⎝ 2⎠ ⎠ T . which. δ2 P = ⎡ cal ⎤ 12108 ε − 1 2 nD + 2 μ2 ⎢ 3 ⎥ 2 V 2 2 ε + nD ⎣ cm ⎦ ( ) (10. in turn. or integrated. the hydrogen bonding parameter w as almost always found by subtracting the polar and dispersion energies of vaporization from the total energy of vaporization. which in volves e xperimental heats of v aporization data for h ydrogen-bonded substances. δh.29) .Determination of Hansen Solubility Parameter Values for Carbon Dioxide 191 Equation 10. Results of both deri vations are shown in Table 10.24 δp = 37.26) where μ is the dipole moment (in Debyes). ⎛ ∂δ p ⎞ ⎛ 1⎞ = − ⎜ ⎟ V − 3 2 ( 37. the h ydrogen bonding solubility parameter .24 can no w be dif ferentiated for either a change in temperature or pressure. TEMPERATURE AND PRESSURE EFFECTS ON HSPs: δh In Hansen’s early work.5. de veloped an empirical approach for the determination of the temperature dependence of δh.4μ ⎡ MPa1/ 2 ⎤ ⎦ V 1/ 2 ⎣ (10.25.28 can no w be dif ferentiated for either a change in temperature or pressure. 5 however.4 and Table 10. sho wn here as Equation 10.5. 2V (10.27) Equation 10. This equation is utilized for determining the change in δp with respect to either temperature at constant pressure or with respect to pressure at constant temper ature. TEMPERATURE AND PRESSURE EFFECTS ON HSPS: δp The first v alues of δp were assigned by Hansen and Skaarup using the Böttcher equation. or integrated.

and di viding by the a verage e xcess heats of v aporization (heat of v aporization of the hydrogen bonding compound minus the heat of v aporization of its nonpolar homomorph) results in the following form of Equation 10.30 with respect to temperature at constant pressure.25 Averaging the rate of change of the h ydrogen bond heat of v aporization with temperature (dEh/dT).3 along with e xperimentally derived values of Eh. and substituting in the isobaric coef ficient of thermal expansion.25 through e xploratory calculations. has sho wn that the dif ference between the heat of vaporization of a h ydroxylic compound (a compound displaying strong h ydrogen bonding) and that of its h ydrocarbon (or other nonpolar) homomorph constitutes a good measure of h ydrogen bond strength.32) Bondi. ⎛ 2.33) . rearranging terms. Reference curves of (dEh/dT) were constructed23 for various functional groups and are sho wn in Table 10.192 Hansen Solubility Parameters: A User’s Handbook so that E h = V δ2 h (10. This work also discusses the decrease in the heat of formation of the hydrogen bond with increasing temperature. ⎛ ∂E h ⎞ ⎛ ∂δ ⎞ ⎛ ∂V ⎞ = V 2δ h ⎜ h ⎟ + δ2 h ⎜ ⎜ ⎟ ⎠P ⎝ ∂T ⎠ P ⎝ ∂T ⎠ P ⎝ ∂T ⎟ ( ) ⎛ ∂δ ⎞ ⎛ ∂E ⎞ ⎛ ∂V ⎞ 2V δ h ⎜ h ⎟ = ⎜ h ⎟ − δ 2 h ⎜ ⎝ ∂T ⎠ P ⎝ ∂T ⎠ P ⎝ ∂T ⎟ ⎠P (10.31) ⎛ ∂δ h ⎞ = ⎜ ⎝ ∂T ⎟ ⎠P α.30) where Eh is the hydrogen bonding contribution to the total cohesive energy.32. Differentiating Equation 10. ⎛ ⎛ ∂E h ⎞ ⎞ ⎜⎜ ⎟ ⎟ ∂ ⎛ ∂δ h ⎞ ⎜ ⎝ T ⎠P − α⎟ = δ h ⎜ ⎜ 2 Eh 2⎟ ⎝ ∂T ⎟ ⎠P ⎜ ⎟ ⎝ ⎠ (10. ⎛ ∂E h ⎞ ⎛ ∂V ⎞ − δ2 h ⎜ ⎜ ⎟ ⎝ ∂T ⎠ P ⎝ ∂T ⎟ ⎠P 2V δ h Simplifying.32 × 10 −3 + ⎟ 2⎠ ⎝ (10.64 × 10 −3 α ⎞ ⎛ ∂δ h ⎞ = − δ + ⎟ h ⎜ ⎜ 2 2⎠ ⎝ ∂T ⎟ ⎠P ⎝ ⎛ α⎞ = −δ h ⎜ 1.

Eh (cal/mole) 4650 1350 500 2750 ± 400 ± 200 ± 200 ± 250 dEh/dT (cal/mole·K) –10 –4. The derivative forms are summarized in Table 10.4.32 × 10 −3 β β ⎞ ⎛ ∂δ h ⎞ = δ + ⎟ h ⎜ ⎜ 2⎠ α ⎝ ∂P ⎟ ⎠T ⎝ δp δh .5.4 Equations (Derivative Form) for the Temperature and Pressure Effects on HSP Temperature Increment Pressure Increment δd ⎛ ∂δ d ⎞ ⎜ ∂T ⎟ = −1.25δ d β ⎝ ⎠T ⎛ ∂δ p ⎞ ⎛ β⎞ ⎜ ⎟ = δp ⎜ ⎟ ⎝ 2⎠ ⎝ ∂P ⎠ T ⎛ 1.5. The total solubility parameter. incremented for small changes in temperature and pressure.25δ d α ⎝ ⎠P ⎛ ∂δ p ⎞ ⎛ α⎞ ⎜ ⎟ = −δ p ⎜ ⎟ ⎝ 2⎠ ⎝ ∂T ⎠ P ⎛ ⎛ ∂δ h ⎞ α⎞ = −δ h ⎜ 1. TABLE 10.36) (10.36 can be rearranged to a form that can also be easily inte grated.5 –7.0 –2. Table 10.34) Equation 10. ∂T β =− ∂P α so that ⎛ 1.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 193 TABLE 10.32 × 10 −3 + ⎟ ⎜ ⎟ T 2⎠ ∂ ⎝ ⎝ ⎠P ⎛ ∂δ d ⎞ ⎜ ∂P ⎟ = 1. can be calculated from equations (deri vative form) in Table 10.4 and the inte grated forms in Table 10.35) (10.9 The change in the δh with respect to pressure at constant temperature is obtained by utilizing the relationship: ∂δ h ∂δ h ∂T = ⋅ ∂P ∂T ∂P where.32 × 10 −3 β β ⎞ ⎛ ∂δ h ⎞ = δ + ⎟ h ⎜ ⎜ 2⎠ α ⎠T ⎝ ∂P ⎟ ⎝ (10.3 Experimentally Determined Values of Eh and (dEh /dT) Functional Group –OH (aliphatic) –NH2 (aliphatic) –CN (aliphatic) –COOH (aliphatic) Hydrogen-Bond Parameter.

0 MP a1/2. when the two are compared in terms of HSP values.7 MPa1/2.4.6 MP a1/2.7 MP a1/2. a solv ent that CO 2 is often closely compared to 26–28 (δt = 14. are shown in the CO 2 HSP surf ace diagrams illustrated in Figure 10.5 ⎥ 0.38) where the reference v alues are as determined earlier.32 × 10 Tref − T − ln ⎜ ⎟ ⎢ ⎝ V ⎠ ⎥ ⎣ ⎦ ( ) 2 ⎡ ⎤ ⎡ ⎤ ⎛ ∂δ p ⎞ ⎛ ∂δ p ⎞ ⎛ ∂δ d ⎞ ⎛ ∂δ d ⎞ 2 ⎢δ p + ⎥ + δ = ⎢δ d + ⎜ + + Δ T Δ P Δ T Δ P ⎥ ⎜ ⎟ ⎜ ⎟ ⎟ ⎟ ⎜ ⎢ ⎥ ∂ ∂ T P T P ∂ ∂ ⎠ ⎝ ⎠ ⎝ ⎝ ⎠ ⎝ ⎠ ⎢ ⎥ P T ⎣ ⎦ P T ⎣ ⎦ 2 ⎡ ⎛ ∂δ ⎞ ⎛ ∂δ ⎞ + ⎢δ h + ⎜ h ⎟ ΔT + ⎜ h ⎟ ΔP T ∂ ⎝ ⎠P ⎝ ∂P ⎠ T ⎢ ⎣ or from the equations (inte grated form) in Table 10.0 MPa1/2.2 MP a1/2. and Tref =298.25 δp δ pref δp −0. It is also interesting to note that the total solubility parameter v alue of he xane. isopropyl ether: δd = 13. δp = 6.3 MPa1/2.32 × 10 −3 T − T − ln ⎛ Vref ⎞ ⎟ ⎥ ⎢⎜ ⎢⎜ ⎥ ⎥ ⎟ ⎟ ref ⎜ V ⎟ ⎢ ⎥ V V ⎠ ⎠ ⎜ ⎝ ⎠ ⎟ ⎢ ⎢ ⎥ ⎥ ⎣⎝ ⎣⎝ ⎦ ⎦ ⎠⎥ ⎝ ⎢ ⎣ ⎦ 2 ( ) (10.37) 2 2 ⎡ ⎤ ⎡ ⎡ ⎤ ⎤ ⎢ ⎥ ⎢ ⎢ ⎥ ⎥ ⎢ ⎥ ⎢ δ pref ⎥ ⎢ δ dref ⎥ δ href 2 ⎢ ⎥ δ =⎢ + +⎢ −1.3 MPa1/2. δp = 4.3 MP a1/2. δp = 4.5 ⎤ ⎥ ⎥ ⎦ 2 (10.5 ⎞ ⎥ ⎢ ⎛ ⎢ ⎛ Vref ⎞ ⎢ ⎛ Vref ⎞ ⎥ ⎥ ⎢ exp ⎜ −1. little similarity is noted. δhref = 5.15 K.2 MPa1/2 can also be compared to liquid solv ent HSP values. From this w ork though. Comparible liquid solvents include chlorodifluoromethane: δd = 12.2 MPa1/2. δp = 0.9 MP a1/2. δpref = 5.4.9 MP a1/2.9 MP a1/2.6 MP a1/2.0 MP a1/2.5 δh δ href δh .13 cm 3/mole. δdref = 15. δd = 12. and δh = 2. Yet. δp = 3. as a function of temperature and pressure. . δt = 14.3 MP a1/2. and δh = 0.5 ⎡ ⎛ Vref ⎞ ⎤ −3 ⎢ ⎥ = exp −1.194 Hansen Solubility Parameters: A User’s Handbook TABLE 10. and δh = 7.7 MP a1/2. and δh = 5. and δh = 5.5 Equations (Integrated Form) for the Temperature and Pressure Effects on HSP δd δ dref δd ⎛ Vref ⎞ =⎜ ⎟ ⎝ V ⎠ ⎛ Vref ⎞ =⎜ ⎟ ⎝ V ⎠ −1. CO 2 HSP v alues at a temperature of 25°C and a pressure of 200 bar .9 MP a1/2) is very near the total solubility parameter of CO 2 (at 25°C and 200 bar). CO2 HSP values calculated with the equations in Table 10. and vinyl trifluoroacetate: δd = 13. Vref = 39.25 ⎥ −0. he xane: δd = 14.8 MPa1/2.

5 5 2. 6 6 5 δp (MPa)½ 4 3 2 1 0 25 0 3 00 3 50 5 4 3 2 1 0 δp (MPa)½ T (K ) 40 0 0 45 0 10 20 30 40 50 Pa ) M P( (b) CO2 polar parameter as a function of temperature and pressure 45K) 7 6 5 4 3 2 1 0 2 50 30 0 3 50 7 6 5 4 3 2 1 0 δh (MPa)½ T (K ) 40 0 0 45 0 10 20 30 40 50 Pa ) M P( (c) CO2 hydrogen bonding parameter as a function of temperature and pressure FIGURE 10.5 5 2. (b) CO2 polar parameter as a function of temperature and pressure.5 15 12.4 HSP values for CO 2 as a function of T and P. (c) CO 2 hydrogen bonding parameter as a function of temperature and pressure.5 10 7. (a) CO 2 dispersion parameter as a function of temperature and pressure.5 0 ) δd (MPa)½ M P( (a) CO2 dispersion parameter as a function of temperature and pressure.5 15 12.5 10 7.5 0 2 50 30 0 0 35 T (K ) 40 0 45 00 10 20 30 40 50 Pa 20 17. δh (MPa)½ δd (MPa)½ .Determination of Hansen Solubility Parameter Values for Carbon Dioxide 195 20 17.

CHAPTER 10 ADDENDUM Research has been the k ey to man y progressi ve de velopments. this set of HSPs were refined using the RED af finity number and Hansen solubility plots to correlate the solubility of carbon dioxide in the solv ents identified in Table 10. . both one-component Hildebrand and three-component Hansen parameters.6 MP a1/2. δp. to re vise the contents of this chapter using these slightly dif ferent numbers.4). Further.3 MPa1/2. The HSP for carbon dioxide reported in the chapter ( δd. to correlate g as/liquid solubility data for a wide v ariety of chemically dif ferent solvents. Therefore. Table 13. for a given solvent–solute combination has been foremost in determining the mutual solubility of the system. The ability to e xpress molecular interactions.3 MPa1/2.2 MP a1/2. δp = 5. It has not been possible. and the criterion for a good solvent was solubility in excess of the theoretical. using CO 2 as an e xample. for a pure supercritical fluid. for a pure fluid solv ent in a w ay that unites the liquid.7 MP a1/2 were found to gi ve a perfect data fit of 1. The δd. and these result in still other impro vements in the same direction when other resources are applied. and supercritical phases. and 5. or rather the dif ference in solubility parameters. It is important to note that the solubility parameter . nor has it been deemed necessary . 5.5 An analogy to “lik e dissolv es lik e” is appropriate.8 MP a1/2. δp.2 MP a1/2.7 MP a1/2.000 (v ersus 0. based on the room temperature solubility in different liquid solvents of known HSP values: δd = 15. ACKNOWLEDGMENTS Special thanks to James Rubin of Los Alamos National Laboratory for his excellent assistance and collaboration in the de velopment of this w ork.6 MPa1/2. in terms of HSPs. This note is only to indicate wh y a slightly dif ferent set of HSP is reported else where in this handbook (T able A-1 and Chapter 13. whether by hand or by computer .8 MP a1/2) were found by a plotting technique with a radius for the solubility sphere of 4. and δh = 5. g as.1. the accurac y to which the solubility parameters for a binary pair can be kno wn will be v aluable in predicting the system’ s behavior. 6.981) with a radius of only 3. Significant steps to ward impro ved understanding have been made with limited resources. A computer analysis of the same data has no w shown that it is possible to describe the solubility of carbon dioxide with a slightly dif ferent solubility sphere. and δh values 15. The data used were collected from a wide variety of sources. This work introduces a theoretical methodology for generating solubility parameter v alues. and 5.0 MP a1/2. represents an advancement in the understanding of the role of solv ents in both e xisting and new applications. Both correlations emphatically sho w the ability of this procedure.196 Hansen Solubility Parameters: A User’s Handbook CONCLUSION A set of Hansen solubility parameters at T = 25°C ha ve been determined for CO 2. and δh equal to 15.

7 Pressure Activity coefficient of the “i”th component in Appendix 10. Chem.M. Battino. 33–38. 7. Rubin...J. 1973. Vol. Williams.1. 1950. 1–9.H. The Properties of Gases and Liquids. 12. Calculation of Hansen solubility parameter v alues for a range of pressure and temperature conditions. J. Copenhagen Danish Technical Press. H.. .. H.L. 6. 13.W. Reid.A. New York. p.. The Solubility of Nonelectrolytes.. B. 3rd ed. Internal presssures and cohesive energy densities of simple liquids. M.1. Barton. 5.. 3. J. 49. G. and Battino.A. R. 11. 16. The thermal pressure and energy-volume coefficients of the methyl alcohol-water and t-b utyl alcohol-water systems . 3592–3603.1. Can. J. and Cle ver. 1977.1 Fugacity of the “i”th component in Appendix 10. Intermolecular forces and chain fle xibilities in polymers: I. Hansen..R. Vol. 334.. E. Vol. 1975. J. L. C. Paint Technol.. Prausnitz...A. Rev. 2nd ed. Equation 10. L. J.. 324–327. 135–144.E. Vol. 15. J. 1928. 4.23 Partial pressure of the “i”th component in Equation 10.A. 54. F.6 Compressibility Density REFERENCES 1. 1. Equation 10.. and Starling.A. J. Chem.. pp. Vol. 17. and Poling.. 28. I. A study of v an der Waals forces between tetrahalide molecules. J. 1643–1648..H... Equation 10. 1965. 1987.. 8. Gee. M. The internal pressure of pure and mix ed liquids. and Scott.21–Equation 10. 34. 490–496. 1960. Denmark. Chem. Chem. Dack.J. Vol.M. The solubility of g ases in liquids. 1967. Solutions and Solubilities. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient. Rev. Phys. Eng. 7.L. J. and Carter. K. Polymer. 984–993.. Hildebrand.2 Saturation pressure of the “i”th component in Equation 10.5) Ostwald coefficient in Equation 10. J. Vol. 31. Chem. Reinhold. Wilhelm.H. Boca Raton. Li.J.. G. 379.F..... The three dimensional solubility parameter — k ey to paint component af finities. Westwater.C. Intermolecular forces in liquids. 6. Renuncio. 10. FL. Boca Raton.. and Hyne. 1932.A..M. J. and c yclohexane.1 Fugacity at standard state in Appendix 10. 505. John Wiley & Sons. Aust.. McGraw Hill. Lee. 14. Chem. 1999. R. 39. Solvent structure. Ed.D. Phys.B. chap. J. Hildebrand.M. No. 1971. 1985. Chem.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 197 SYMBOLS SPECIAL TO THIS CHAPTER KH L P ai fi fio k pi pis Ω β ρ Henry’s law constant (Equation 10. The use of internal pressure and cohesi ve energy density to examine contributions to solv ent-solvent interactions. No. including the supercritical fluid re gion.. No. and Wilson.M. polymers. Vol. Solvents. CRC Handbook of Solubility Parameters and Other Cohesion Parameters.. 4th ed. Am. A.M.A. H. Frantz. J. 81. G.A. 43. Allen. benzene. Hildebrand.W. Vol.. R.A. 9. CRC Press.L.R. 1.... J. Hansen. Huang.F.. plasticizers. p. CRC Press. Dack.. Ne w York. New York. Breedveld. M.. An accurate equation of state for carbon dioxide..J.. Jpn.1 Constant in Equation 10. Res. 4. Rev. MacDonald. 2004. No. and Edw ards. C.H. W.R.. and Hildebrand. and resins. and Prausnitz. Ind. 2636–2642.. 1929. Vol. J. D.. Vol. 4. 456–476. Soc. 1967. Hansen Solubility Parameters: A User’s Handbook.. J. No. 4967–4972. 1991. Eng. 18. 73. R. Thermodynamic functions of the solubilities of gases in liquids at 25°C. Internal pressures and solubility parameters for carbon disulfide. Hansen.2 Bunsen coefficient in Equation 10.M. 2..B. FL. Phys. G. 104–117. C.

R.J. Chem. benzene. and Urieta. Kunerth.C.P. and v an der Veen. A. Zaynalov.A...M. W. and K eller.K. 1993.. F ang. pp. 131.L. 159–172. The solubility of carbon dioxide. 71.. 10. 30.G..M.. 1509. J. J..P. Chem... 49. Santafe. Shestakova. T. 342. Fluid Phase Equilibria.. Hansen.. Y... Vol.. Scand. Acta Chem. CO2 and SF6 in various solvents. 21. 1993. Res. CH 4.S.. S.. 155. D. John Wiley & Sons. Ro yo. Cao. Pardo. 133–140. Vol. Process Des. 1968. J.Acta Chem. J. 3067–3079.H. Solubility parameters. M. Solubility of CO 2 and N 2O in certain solv ents. 21. E. F. Chem. Z. 42. 4145.N. Vol. 1996.C.N. 109. 25 No. 87.. McLaren. 28. oxygen. pp.N. 537–544. J. 37..G.. Polyuzhin. Chem. 1998. Johnston. Phys. Solubility parameters of supercritical fluids. 67–73. New York.D. M. C. Span. 11. 50 1992 International Union of Pure and Applied Chemistry (IUPAC). Bondi. 37. A simple method for determining the solubility of g ases in liquids: application to CO 2-cycloparaffin systems. 153–161.N. Podvigaylova. 66.. Wilkinson. M. I.V. Vol. Russ.A. Vol. 29–37. 23. J. 1968..K. Gjaldbaek. Solubility of CO2 in Organic Solvents. 134. Pardo.A. Z.. S... J. Y. 473. carbon monoxide.M. I. 1984. C. Solubility of g ases in b utanols. 5. J. J. Liquids. 37. Ro yo. 4. E. Vol.G. Panayiotou. F. J. Vol... Vol. 1997. and Simonov. Fluid Phase Equilibria. and Wagner. c yclohexane. 19. Vol.. 43.S. T. 5097–5101. Solubility of Gases and Liquids: A Graphic Approach.R. Lopez. 1995. R. Physical Properties of Molecular Crystals. 19.. 1957. A.. 1987. Kobatake. 127–137. M. Ref. Gjaldbaek. and aqueous solution in aerosol. 36.. N. D. A.. J. Research Triangle Park. Scand. Goto.L. carbon tetrachloride. J. and Mel’nik ova.. The Soviet Chemical Industry. Vol.. 27 Fedotov.G. Solubility and entrop y of solution of He. Sinevich. M. Chem. R. Fluid Phase Equilibria. Pardo. K erschner. 20. A. 1997..G. Hyatt. C. Phys. 41. Vol. G.C. Russ.M.L.... Kirk-Othmer Encyclopedia of Chemical Technology.. Ind.. Giddings. 1970. 60. Dev. 38. III. A. Phys.V . 40. and Glasses. Rev.. Lopez. Science... Allada. Vol. 29. Vol. and Urieta. Chem. IV.. 331–334.. and La vecchia. Ed. Bull. 1922.. Interscience. NC. J. Gironi. Vol..... A. J. 344. Lopez. Org. 1984. regularity of gas solubilities. Data. Y. A. Pirig. Chem.C. 1975. Ro yo. Ind. I. 39.. Eng. 1990.A. Carbon Dioxide in Nonaqueous Solvents. 4. 3 No. Radzhabov. Chem.. and Hildebrand. 954.J. 841–851. Q.. E. C 2H6. L. No.S. Vol.T. 26. 65. R. F... J. Suppl. C. 162. 691–695. Mainar .. Urieta. 512–524. M.. Solubility parameter revisited: an equation-of-state approach for its estimation. Vol.. Heats of v aporization of h ydrogen bonded substances. 34... Smith.. Just. P. Gerrard.. Fluid Phase Equilibria. Solubility of g ases in b utanols.. Ne w York. Fogg. No. On the applicability of Pierotti’ s theory for the solubiliy of gases in liquids. A new equation of state for carbon dioxide co vering the fluid re gion from the triple-point temperature to 1100 K at pressures up to 800 MP a. No. and Losa. and Urieta. and nitrogen in polar solv ents. Jpn. 1953. Ikushima. 8. Shagidanov. 6. O2. 19–21.A.. 1971. Solubility of 15 non-polar g ases in cycloheptanone. Lopez.I. K orbutova. 35. 32.. M. 23. 33. J. Plenum Press. High Pressure Gas Chromatograph y of Nonvolatile Species. and Arai. L. Eng. 73. Solubility of g ases in b utanols. 58.H. Gallardo. The solubility of carbon dioxide in perfluoro-n-heptane. 31. Carbon dioxide solubility . New York. Solution Chem. O’Neill. N 2. J.. 7. K. 22. J. normal heptane. 1900–1903. 25. Vol. W. Vol. No.. p. Liquid and supercritical carbon dioxide as or ganic solv ents.P.. and Makitra. P art 2. A.198 Hansen Solubility Parameters: A User’s Handbook 18. Myers. C.. .G. 1976.... and Simkin. 1997. Bondi. De Ligny.. Vol.. Vol. Vol. carbon disulphide. and Beerbo wer. Soc.. Fluid Phase Equilibria.. Kruglikov. J. 1999. Solubility of homopolymers and copolymers in carbon dioxide.C. W. p. M. Phys. Chem. 1398–1413. and Anderson. AIChE J. Vol. Intermolecular interactions of supercritical carbon dioxide with polymers of dif ferent types..T. Appl. 1961. S. 1901. 4.. Phys. M. F.. C.A..S... J. Modified solubility parameter as an inde x to correlate the solubility in supercritical fluids. Vol. Mayoral.. 24. and Jureller. Fluid Phase Equilibria. B. Solubility Data Series.C.

. 1977. and Losa. 71. Phys. Chem. 4–8. 10. 57. Gallardo. 223–233. J. CF 4. 10... H.. L. 1975. Melendo.1 IDEAL SOLUBILITY OF GASES IN LIQUIDS Solutions that come close to approximating ideal solutions are those which are very dilute. Walter. 50.. Eng.. 65. 515–522.. M.. and he xadecane. Vol. and Losa. 55. Jou. J. the intermolecular forces on a solute molecule may be quite dif ferent than in the pure solute phase.2dibromoethane.. Solubility of nonpolar g ases in 2. Field. F . J.. De Ligny.15 K at 101. Chem. 1975.E.. 47.. J. and Wilhelm. Chem. 51.B. 20. Eng. Data.H. N 2. 2675–2682. J. 59–61. E. 1977. Eng.-Y. 1972.H. Solubility of nonpolar g ases in 2-methylcyclohexanone between 273. D2. 1.G. H 2. 9. Deshmukh. M.. Vol.15 and 303.) The concept of an ideal solution is often an appropriate approximation .D. 56..J. Lopez. F.. Chem. Vol. 1989. .15 K at 101. triethylamine and 1. and Al-Najjar. Hayduk. Vol. B. 6. Can. 1.32 kP a partial pressure of g as.J. C. Chem.L. 1989. h ydrogen sulphide and propane in some normal alkane solv ents I. Vol.M. Kr. Phys. The solubilities of carbon dioxide.J. C 2H6.. J. Chem. Thermodynamics. R. 1974. Urieta.. (In very dilute solutions.R. C. J.. Dymond. J. Sci..15 and 303.. E.C. 1829–1831. The solubility of g ases in liquids 10..15 K at 101.. J. Urieta.R.. 1990. CH 4.. Battino. and Hildebrand. 52.S. Byrne.. Rusz. 32.. O 2. Chem. 27.. 7. 58. C. 111–115. Thermodynamics. but the solute molecules are far enough apart that solute–solute interactions do not manifest themselves. 68.15 and 303. 1967. 809–811. J. 241–1246. 8. CO 2 and CH 4 in NFormyl Morpholine. Faraday Trans. J.. CCl2F CClF2. J. R... 6.G. and 2. L. Melendo..S. Vol. Vol. Melendo. M. The solubility of gases in liquids.D. Chem.F.. and Mather . Otto.G. Chem. J. R. dodecane... and Losa. Xe.32 kP a partial pressure of g as. SF 6. Chem.. Data.. A test of Pierotti’s theory for the solubility of gases in liquids. No. 9. Eng. a mixture of 1-chloro-2-bromopropane and 2-chloro-1-bromopropane. No.B. J. 49. R.. 45. J. H. 1. Wilcock. Ind.1. Wilhelm. J.G. Solubility of H 2S. 67. Vol.. J. Urieta... and van der Veen. Vol.A.A. Fluid Phase Equilibria.G. Chem. P ... E. Hung. J. 1972. Chem. Soc. Tokunaga. 1.6dimethylcyclohexanone. Vol.. 391–401. M. Maget. Solubility of carbon dioxide in c yclohexanol. No.. Begley. J. 1965.. 435–439. Wilcock. 1978.. 41–46. Data. Chem. and Wilhelm. Thermodynamics.24-(CH3)3C5H9.. Solubility of non-polar g ases in c yclohexanone between 273..S. M. APPENDIX 10. Battino. The solubility of a series of g ases in c yclohexane and dimeth ylsulfoxide. Vol. 817–822. J. 213–214. Vol. and CO 2 in c yclopentanone from 273. Thermodynamics.1: IDEAL SOLUBILITY OF GASES IN LIQUIDS AND PUBLISHED CO2 SOLUBILITY DATA. 1964. by means of literature data on solubility and entrop y of solution. H. 41–46.A. M.. 7. and Losa. 60.M. Two of these solv ents. 237–243. A.. W. R.. J.. Gallardo. Published CO2 gas solubility data in 103 liquid solvents were gathered from the available literature and is presented in Table 10.E. R. Gallardo. Hiraoka.A. Chem. J. Vol.W. Makranczy. 50. E.. Can. C.M. No. Chem. The solubility and entrop y solution of certain g ases in (C4F9)3N. No.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 199 44. 53. 48. C 2H4. and Wilhelm. Ne. The solubility of g ases in liquids. W. and Battino. J. Vol. and Simpson. 2198–2202. 17. and mineral oil. J. Vol. Gallardo. 1989.S. 1. Can. and Williams. King. Danforth. N. 1979. 54.. Sci. Battino. J. Eng. Ar. or those where the molecular species are so nearly alik e that a gi ven molecule is subject to the same intermolecular forces (both attracti ve and repulsi ve) in the mixture as in its o wn pure phase. 46. 85. Solubility of propane and carbon dioxide in heptane. Chem. 68. Vol. 59. Urieta. K..32 kPa partial pressure of g as. R. Solubility of He. Vol. I. C. J.A.. 1987.4-dioxane were subsequently deleted from the data set based on their kno wn tendency to chemically react with CO 2. and Balog-Me gyery. J..

A.1 e valuates the ideal solubility of CO 2.0177 0. In this situation.50 71. at the solution temperature. In Equation 10. for g ases dissolv ed in liquids. calculated using Equation 10.1) The acti vity of a substance gi ves an indication of ho w acti ve a substance is relati ve to its standard state. Equation 10. most g ases are only sparingly soluble in typical liquids.60 . xi.1. a.2) Equation 10. and the fug acity terms in Equation 10.0199 0.19 56.12 0. for the temperature range 0°C to 30°C.2 usually gi ves correct order of magnitude results pro vided that Pis is not lar ge and the solution temperature is well belo w the critical temperature of the solv ent and not e xcessively abo ve the critical temperature of the g aseous solute. Pis.2. 4 Table 10.A. as it provides a measure of the difference in chemical potential at the state of interest and that at the standard state. and the mole fraction. Raoult’ s la w predicts an ideal CO 2 solubility of xCO 2 = 0. equals the mole 3 fraction.40 50. xi = pi Pi s (10. xi = ai = fi fi o (10.A. over the entire composition range and o ver a nonzero range of temperature and pressure.1 approach the partial pressure and saturation pressure of i. generally at lo w pressures. is referred to as the ideal g as solubility. ideal From Table 10.A. xi . the ratio of the partial pressure and saturation pressure can no w be used to express the mole fraction.1 is an empirical rule suggested by Le wis and Randall 5 that assumes imperfect g as mixtures to beha ve as ideal mixtures. 4 The term fugacity. The ideal solubility calculated from Equation 10.1 Ideal Carbon Dioxide Solubility Calculated Using Raoult’s Law T (°C) 0 15 20 25 30 s PCO2 (atm) ideal =1/P s xCO2 CO2 34. 2 Thermodynamically.1. 2. an ideal solution is defined as one in which the activity. as calculated from Raoult’ s law.A.60 63. fi is the fug acity of component i at partial pressure pi. w as introduced by Le wis5 as a measure of thermodynamic escaping tendency and is equal to the effective gas pressure corrected for deviations from ideality.0291 0.A.0141 Note: pCO2 = 1 atm. the ef fect of pressure on the fug acity of component i is ne gligible. therefore.A.0157 0.2 is kno wn as Raoult’s law. f. respectively.A. as at modest pressures. and this v alue has been used in se veral of the published CO 2 solubility studies.A. and fio is the fugacity at the saturation pressure of i. When de viations from the ideal g as la w are small.200 Hansen Solubility Parameters: A User’s Handbook TABLE 10.A.A.0157 at T = 25°C.4.

9944 . 6.9941 . New York.5 71.6 63. Vol. G..G. J. I. Molecular Thermodynamics of Fluid-Phase Equilibria. carbon monoxide and nitrogen in polar solv ents. Therefore.0157. R. This value was also used by Gjaldbaek et al. J. Prausnitz. E. Table 10. Reid.Determination of Hansen Solubility Parameter Values for Carbon Dioxide 201 TABLE 10. 8 1398–1413. . 3.. benzene.M.4 50.. communication. Pergamon Press Ltd.. Carbon dioxide solubility . Scand. Pirig. The Properties of Gases and Liquids..2 Ideal Carbon Dioxide Solubility Calculated Using Raoult’s Law and Fugacities T (°C) 0 15 20 25 30 s PCO 2 (atm) fCO2 PCO2 = 1 atm .H. R. The solubility of carbon dioxide. de Azevedo. Vol. 1954. 9. carbon disulphide and aqueous solution of aerosol. P art 2. The ideal CO 2 solubility at 25°C and 1 ideal atmosphere partial pressure. J.0229 0. Lichtenthaler .9947 .17 0. 537–544. 4. Gjaldbaek. 5. R.9928 .A..2 gi ves the saturation pressures of CO 2 for the temperature range 0°C to 30°C. 7. Acta Chem. The solubility of carbon dioxide in perfluoro-n-heptane.. Prausnitz. J. 1970. From Table 10. Prentice-Hall. U.. 28 39 42 46 49 34.E.A. 1953. REFERENCES 1. B.19 56. and Makitra. Prausnitz..0251 0. Chem.C.51 36. oxygen.04 39.. J. and Poling. New York..K.. 1986. McGrawHill..L. Hansen. It has been noted by Prausnitz et al.2..V.0157 0.0177 0. 1969.9 in their work comparing e xperimental and calculated CO 2 gas solubilities. Van Nostrand Reinhold Company. Polyuzhin. McGraw Hill. Acta Chem. it is necessary to consider the error in using pi /pis instead of fi /fi.N. 1993.B. No. Lewis. Appl.. 2nd ed. and Randall. Y. when the saturation pressure of the g as is high. In doing so. Oxford.N.12 26.M. M. 66. Regular and Related Solutions. Scand. 8.0199 0.. and Scott.C.A.0211 Note: pCO2 = 1 atmosphere.K. Engle wood Cliffs.G. Charles M. in cases where the saturation pressure of the g as is above 1 0 atmosphere. and errors in the use of this estimation technique for the solubility of g ases in liquids can be significant. E. New Jersey.9950 o fCO 2 at s PCO 2 ideal xCO 2 s = 1/PCO 2 ideal xCO 2 o = fCO2/fCO 2 % Diff.2 it appears that the assumption of Raoult’s law for the determination of ideal CO2 solubility in liquids results in significant error ..7 is xCO 2 = 0. New York..3 47.0141 0. R. they caution that se veral assumptions are made. especially . and Anderson.M. 1923. normal heptane. J.N. Hildebrand. Phase Equilibrium in Mixtures.0375 0. Russ.Thermodynamics and the Free Energy of Chemical Substances.. Equation 10.0276 0. King. however.4 that the simplest way to reduce Equation 10.0229 compared with ideal a Raoult’s law prediction of xCO 2 = 0. carbon tetrachloride. c yclohexane. J.A..6 43. 1987.. Gjaldbaek. as well as the fug acities and calculated ideal solubilities using both methodologies.1 to a more useful form is to rewrite it in the manner suggested by Raoult’ s law. 4th ed.C. 691–695. 8. M. Vol.A. 7.0291 0. 4. 2. as calculated from CO 2 fugacities...


In this chapter. such as vapor degreasing. 3 203 . These HSP characterizations of solv ents and soils can be compared as v alues in HSP space just as similar values of solvents and polymers are compared to determine if a solv ent can dissolve a polymer. As designer goods vs. The purpose of this chapter is to illustrate ho w to use HSPs to aid in identifying v aluable applications for these and other solv ents in cleaning applications. six representati ve soils and 13 base solv ents. it will be sho wn that the base azeotropic feedstock must display significant polar an hydrogen bonding intermolecular forces if the azeotrope is to clean a soil composite that also contains significant polar and ydrogen bonding intermolecular forces. Given the relative inertness of these designer solvents to both humans and the en vironment. Chief . these benefits are a relat vely nonhazardous e xposure to humans and compliance with nearly all en vironmental re gulations (including the Montreal Protocol) of most nations. The approach tak en is the traditional one used in cleaning work: matching the solvent to the soil.of Hansen Solubility 11 Use Parameters to Identify Cleaning Applications for “Designer” Solvents John Durkee ABSTRACT Choosing the most ef fective solv ents for cleaning requires characterization of both solv ents and soils. Hansen solubility parameters (HSP) can be used whether or not single or multiple components are present in either the soil composite or the solv ent blend. were selected as feedstocks for formulation of binary azeotropes. In this chapter . which limits applications. The selected solvent azeotropes were predicted to clean fi e or six representati ve soils. so-called “designer” solv ents. 2 HSP values of the solvents will be those matched to soil materials. INTRODUCTION Since the early 1990s. using a literature database. more than a dozen ne w solvents have been commercialized for cleaning 1. it is not surprising to find them to be inert to ma y soils. A previously developed value of R A for cleaning operations allo ws prediction of useful cleaning performance by solv ent on soil. the y are high priced. Another limitation is their character as solv ents. The sobriquet “designer” has been applied to them as their structure w as designed to pro vide certain benefits. commodities. and other applications.

the better the result. The better matched cleaning agents and processes are to a soil. the components must have similar intermolecular forces. Pre vention of a mismatch of intermolecular forces between a cleaning solvent and a soil (an incompatibility) is a role uniquely fulfilled by HS . often inadv ertently produced by users. while poor matches will still be apparent. the higher melting fractions of the coating wax probably will not be well-cleaned because they are not heated to a high enough temperature. A good e xample is h ydrocarbon-based w ax. If the soil material is a single phase (presumably a solution and not an emulsion). Microcrystalline wax. have excessive or limited v olatility.5 A soil material is probably soluble in a solv ent (or solv ent blend) if the Hansen parameters for the solv ent lie within the solubility sphere for the soil. the soil is the enemy to be conquered. It can be hard to discriminate among them. Cleaning of multiple soil components usually produces multiple cleaning results. They may be physically unsuited to the application because of mismatched ph ysical properties such as surf ace tension. no cleaning solv ent is perfect. Yet. which must normally be melted in order to be removed from surfaces.or multiple-phase mixture. All ha ve fl ws. ”) HSP OF MULTIPLE-COMPONENT SOILS How then is a user to match a single (or azeotropic) cleaning solv ent with a multicomponent soil? The answer combines observ ation along with conventional solubility theory and is seen in the following list: • First. carcinogenic chemicals such as benzene. In other words. or simply be too e xpensive. (See “More Realistic View about Evaluating HSP of Composite Soils. has a sharp melting point because the range of molecular weights of its component paraf fins is narr w. safety. used as a binder in polishing compounds. Sol ents with general fl ws4 may pose environmental. or viscosity. They are composed of occasionally unkno wn ingredients prepared with unkno wn and v ariable proportions. if their cleaning w ork is to meet their requirements. 6 PATHOLOGY OF SOILS Those managing cleaning w ork must know at least as much about soils as the y do about cleaning agents.204 Hansen Solubility Parameters: A User’s Handbook A VARIETY OF SOLVENTS There are almost too man y types of cleaning solv ents. the specific f w that is usually f atal to an application is the solv ent’s not being compatible with the soil materials. density . However. They may include “tramp” materials or contain une xpected components that are not al ways present or recognized at best. and lo w-cost common chemicals such as w ater ha ve all been used as cleaning solv ents. more comple x chemicals such as N-methyl-2-pyrrolidone. or health hazards. ozone-depletin chemicals such as carbon tetrachloride. After all. Wax used in box-coating operations has a mixture of molecular weights and a broad melting point. which are not homogenized. One of the strengths of solv ent-cleaning technology is the v ariety of solvents that can be used in v arious solvent cleaning processes. it is lik ely that a cleaning process must . Soil can also be the outcome of incomplete cleaning work. and a useful solvent can probably be found from their study. If the boiling point of the cleaning solv ent only matches the a verage melting point of both w ax mixtures. one observes7 if the soil is a single. both general and specific. flammable chemicals such as acetone. Soils are multicomponent mixtures. If the mixture contains multiple phases. the soil least matched to the cleaning solv ent is likely to be the soil least cleaned. Simple chemicals such as he xane.

in volves some compromise. The choice of a solv ent (or blend) similar to a soil component (or composite) is accomplished in a similar manner as solv ent for a polymer 10 is chosen (see section on “Method for Choice of Suitable Solv ents”). In other w ords. In this case. for e xample) or the chosen cleaning process must ha ve multiple steps. the required compromise solvent should be no further than the outer boundary of the lar gest circle (see discussion following).9 Third. then it is assumed that the observ ation of the soil form as a single phase w ould mean that the unrepresented components must be similar to those represented. HSP values for the two that represent the greatest volume fraction should be used. without specific data.1. Information about the individual soil components is used here. each of the three HSPs (disperse. That is. it is normally assumed that there is no olume change upon mixing of solv ents. if the soil contains multiple components.1. the composition v alue to be used in calculating solubility parameters for solv ent mixtures is the volume fraction (φ) for each component. 12 For a binary (tw o-solvent) mixture. Whereas it is perfectly possible (see next two sections) to compute the composite HSP using the entire composition.Use of Hansen Solubility Parameters 205 Polar HSP (soil or solvent) #2 Soil Component #1 HSP Parameters for Common Soil or Solvent H2 Bonding HSP (soil or solvent) FIGURE 11. Second.15. one estimates the HSP v alues for a composite of soil materials.1) Traditionally. polar . That is: . 14 This equation is correct for more than two components where the HSP 14 values are known. and hydrogen bonding) of a solv ent mixture is a linear function of composition. one chooses a solv ent whose HSP v alues are similar . too. 8 If that means other components are not represented in estimation of HSP v alues.16 δ blend ≡ ⎡ ⎣ϕ comp1 × σ com1 ⎤ ⎦+⎡ ⎣ϕ comp 2 × σ com 2 ⎤ ⎦ (11. METHOD FOR CALCULATING HSP OF COMPOSITES (SOILS OR SOLVENTS) Solubility parameters11 of mixtures are linear. the equation for all four13 solubility parameters is Equation 11. the recipe for that composition is only seldom kno wn. This usually involves some compromise. For example. This.1 • • be chosen that is not based on solvency but on application of mechanical force (aqueous cleaning. This is illustrated in Figure 11.

In this w ay. And other components of the composites will also ha ve dif ferent rates of solubility . In some cases. with component number 3. Equation 11. In other words. δHydrogen% % ϕ δDispersion δPolar Bonding 67.1 0.1. Fraction ⎞ ⎜ ⎟ ⎝ Density ⎠1 (Vol.0 133.0 6. one component of the soil composite will al ways be less soluble with the cleaning solvent in comparison with the composite of soil components as a whole. Fraction ⎞ ⎛ Wt . the solv ency character of the azeotropic blend is v ery different from that of the neat cyclohexane. or nonazeotropic blends.2. is: . That distance should be as small as possible (R A ≈ 0). 20. polar .5 2. The least compatible component may well be the one that should define the sol ency of the cleaning solvent. hydrogen bonding intermolecular forces to the blend.1 ⎛ Wt . the volumetric-proportional HSP v alues (disperse.2 Isopropanol 60. Also note that the value of adding isopropanol (2-propanol) to cyclohexane is to add polar and.789 16. cleaning (where all soil components must be equally well remo ved) is a situation where the v olumetric average may not represent the true situation.2 33. Fraction ⎞ ⎟ ⎜ ⎟ +⎜ ⎝ Density ⎠1 ⎝ Density ⎠2 (11.1 Calculation of HSP for Mixtures via Equation 11. and h ydrogen bonding) of the soil may not represent the actual nature of the cleaning problem.2 16.0 67. in Figure 11. 22 which defines A.1 is useful for both solv ents and soils. Fraction)1 = ⎛ Wt .206 Hansen Solubility Parameters: A User’s Handbook TABLE 11.8 15.8 0. This is because the indi vidual solvent density values are essentially the same. 19 METHOD FOR CHOICE OF SUITABLE SOLVENTS RA is the distance in HSP space between the soil and the solv ent.8 16.2) An azeotropic mixture of cyclohexane and isopropanol is used as an e xample of this calculation. Cyclohexane has essentially none. especially.5 Cyclohexane 84. It is also useful for azeotropes. that the v olume and weight 17 concentrations are essentially the same. This is illustrated. the well-mixed composite being the limit of cleaning performance.1) may not lead to the right choice of cleaning solv ents. 18 Unless the soil is a single component.1 Component 1 2 Solvent Mol Density δDispersion δPolar δHydrogen-Bonding Wt. Calculated v alues of HSP for a composite soil (as are shown in Figure 11.4 Wt.21 The equation to be used for solv ent selection. MORE REALISTIC VIEW ABOUT EVALUATING HSP OF COMPOSITE SOILS Judgment is also a factor required for evaluation of the HSP of composite soils. The data needed and results are gi ven in Table 11. Vol. for this e xample only. Note. This happens when a single component in the composite soil limits cleaning performance vs.0 5.779 0. near azeotropes. g/cc MPa1/2 MPa1/2 MPa1/2 0.

2 2 RA Solution = ( δ Polar for Solvent – δ Polar for Soil ) 2 + ( δ H 2 – Bonding for Solvent – δ H 2 – Bonding for Soil ) 2 + ⎡ 4 × ( δ Disperse for Solvent – δ Disperse for Soil ) ⎤ ⎣ ⎦ 2 (11. The independent variable in the equation is the choice of solv ent.3 is the three pre viously estimated HSP values for either a soil composite (via Equation 11.3 23).Use of Hansen Solubility Parameters 207 Polar HSP (soil or solvent) #2 Soil Component #1 #3 HSP Parameters for Common Soil or Solvent H2 Bonding HSP (soil or solvent) FIGURE 11.3 . the matching of solv ent (or blend of solv ents) to soil (of one of more components) is described in Reference 3 as Figure 4. that is. If the distance between the loci Polar HSP (soil or solvent) Soil or Solvent #1 HSP Parameters for Common Soil or Solvent Soil or Solvent #2 H2 Bonding HSP (soil or solvent) FIGURE 11.6 (reproduced here as Figure 11. The dependent variable in the equation shows how well the soil is dissolved. Graphically.1 or that for the single soil component judged to limit cleaning performance as sho wn in Figure 11.2. HSP data and a spreadsheet can aid in making this comple x choice because they allow evaluation of solution performance for man y solvents (or blends) ag ainst all components of the soil. The individual values for HSPs for the soil are either those computed from Equation 11. how closely R A approaches zero.3) The preceding equation comprises the follo wing data: • • • The basic data for Equation 11.1) or HSP v alues for a component e xpected to be limiting.

based on partially fluorinated propane. and solubility properties are gi ven in Table 11. the fully fluorinated sol ent based on butane (PFC-506025) is also included for comparison. and the proposed solv ent cleaning process w ould be e xpected to be an unsatisfactory choice. HFC-365mfc (based on partially fluorinated propane). The two-dimensional graphs26 of the latter are illustrated in Figure 11. Their structural formulas. Hydrocarbon ether solvents partially based on silicon (not on carbon): OS-10.” AN OPEN QUESTION — ANSWERED It is interesting to note the o verlap of Figure 11. HFE-7200.”). HCFC-225 ca/cb. No parts need be found. wet. which does find applications a nonazeotropic cleaning solvent.3. All three ha ve ozone depletion potentials (ODP) of significance. as well as three or f e hydrogen atoms. Four families of solvents are considered: Hydrofluoroethers (HFEs): HFE-7100. Where there is HSP data for soil components and solv ents.5. But in general. These reference soils serve both a specific and a general purpose. This situation is common to that found in mark eting. Each represents a soil type frequently encountered in industrial operations. Numerically.2. and their collecti ve HSP values cover a broad range of HSP space. b ut the y are based on a silicon to oxygen bond rather than a carbon to oxygen bond. Solvents more useful as refrigerants which form azeotropes: These are HFC-245f a (based on partially fluorinated ethane).4. In addition. IDENTIFICATION OF DESIGNER SOLVENTS Choice of the cleaning solvents of the designer type is somewhat arbitrary. and HFE-8200 24 are molecules uniquely containing both fluorine and oxygen. . Specific azeotropes of design solvents will be proposed to clean each soil. and OS-30 are molecules also containing ether linkages. their production and use in unde veloped countries is permitted for se veral years be yond the publication date for this book. they total a “baker’s dozen. There is observ ed to be little o verlap. They are identified and described in Table 11. and CFC-113.5. Ho wever. The h ydrocarbon meth yl group is plentiful.208 Hansen Solubility Parameters: A User’s Handbook of HSP values for solvent and soil is too great (RA is too large).4 and Figure 11. a (HFC-4310mee). weighed. The question is: Against which soils should these solv ents (and blends) be positioned (e valuated)? The approach taken here is to choose six different single-component organic soils for reference. cleaned. physical. one must be able to position them relati ve to common soils. This overlap would represent the similarity (see Endnote 8) of these designer solv ents for use in cleaning typical industrial soils. The HSP values of each soil are plotted as in Figure 11. OS-20. an optimum choice of solv ents (or solvent blends) can be made (see section on “Identification of ‘Designer Solv ents. there will be little graphical overlap between the tw o spheres. or e xamined. which has led to a ban on their production in industrialized countries pe the Montreal Protocol. the method of de veloping successful solvent cleaning application is illustrated by matching solv ents to soil components. Solvents with harmful environmental impacts but useable in de veloping countries: HCFC141b. REFERENCE SOILS FOR COMPARISON To e valuate the suitability and limitations of designer solv ents and their azeotropes in solv ent cleaning operations.

Use of Hansen Solubility Parameters 209 Table 11. Tricresyl Phosphate 6 15.6 6. lubricants. Ester. Questions asked during the last decade by those who practice cleaning science are:What should these designer solv ents be used for?. and “how can we g ain the value of the United States v olatile organic compound (USVOC) exemption and negligible ODP promised by their manuf acturers.5 3. or isobutyltrimethylpentane) is a component of gasoline (Octane Rating 100) and representative of the general class of hydrocarbon chemicals. solvents or cosolvents.2. It and its derivatives have applications in the manufacturing of soaps. these designer solv ents are no more functionally useful than a simple solv ent such as heptane (which costs perhaps one twentieth as much per kg).7 CH3. plastics.9 13. Ester Castor Oil 27 3 13. when these designer solvents appear useless to remove typical industrial soils? ” . and as flame retardant. waxes and polishes.8 7.5 Phosphate. It is also used as a plasticizer (PVC and alkyd resins).3 0 0 CH3 Butyl Stearate 2 12. Acid Double Bond Hydroxyl Aromatic.6 6. and oil carriers in agricultural operations. Double Bond Ethyl Cinnamate 4 16. Double Bonds (6) (Triglyceride Ester of mainly Linoleic Acid28). paint and ink additives. which are used as used as raw material of emulsifiers or oiling agents for foods.9 13.3 6. hydraulic and brake fluids. dyes. ASTM Fuel “A” 1 14. Other than cleaning h ydrocarbon-based soils that do not ha ve substantial polarity or h ydrogen bonding character.5 CH3. Esters (3). surfactants and base materials for perfumery. It is the ester primarily responsible for the smell of cinnamon. Representative of a class of soils as fatty acid esters. as a detergent. pharmaceuticals and perfumes. a natural vegetable oil.5 5 Linseed Oil 13. coatings. Castor oil is a vegetable oil.0 10. spin finishes and textiles. Aromatic This material is found in common metalworking fluids: anti-wear agents or lubricants.2 Reference Soil Materials Soil # δ Dispersion δ Polar δ Hydrogen-Bonding Molecular Model Image MPa½ MPa½ MPa½ Contained Functional Groups Industrial Use Isooctane (2. lubricants for plastics. inks.5 3.0 10. (Ethyl Phenyl Acrylate) is a natural chemical and a common fragrance and flavoring component.4-trimethylpentane. paints. wh y should manuf acturers char ge e xorbitant prices?. is the most important drying oil of the oil painting industry.1 CH3.

MPaˆ (½) 12 10 8 6 4 2 0 0 2 4 6 8 10 Polar HSP. including their blen HSP v alues calculated through literature component v alues from Reference 15 and Equation 11.4 FIGURE 11. Naturally . the type of cleaning29 solvent evaluated in this study is a binary azeotrope with one component being a designer solv ent and the other being a more common (and lo w-priced) chemical. MPaˆ (½) Hydrogen Bonding HSP. .5 sho ws that approach to be fruitless) b ut in combination with secondary components in binary azeotropes identified by the manuacturers of these solvents. To extend the power of designer solvents by using them not in neat form (as a comparison of Figure 11. and A library of azeotropic formulations identified (See Reference 16). MPaˆ (½) 12 14 HSP OF SINGLE COMPONENT SOLVENTS 13 ‘Designer’ Solvents 14 12 10 8 6 4 2 0 0 2 4 6 8 Polar HSP.210 Hansen Solubility Parameters: A User’s Handbook HSP OF REPRESENTATIVE SOILS Chosen to Cover a Broad Range of Value 14 Hydrogen Bonding HSP. To use e xisting solvent data of ph ysical and chemical properties as well as HSP v alues to match one of the six common industrial soils to a binary azeotrope one of whose components as a designer solv ent. MPaˆ (½) 10 12 14 FIGURE 11. It is that R A is related to the amount of soil not remo ved (uncleaned) from the parts being cleaned. In all cases.4 for HSP alues).5 The approach to answer these questions without commercial bias will be tw o fold: 1.4 and Figure 11.S. In other words. The questions remain as to what standard of cleaning performance speaks to how well the soil is dissolv ed and how closely does R A approach zero? The answer is in a study also reported in Reference 16. LIMITING RA VALUE FOR EXPECTED GOOD CLEANING PERFORMANCE A good cleaning performance that is attained by limiting R • • • A requires the follo wing: A method for calculation of HSP values of blends that are identified (See Equation 11. the cleaning process contemplated is the con ventional vapor degreasing. Cleaning trials30 have been and are being conducted by the Surface Cleaning Laboratory within the Toxic Use Reduction Institute (TURI). located at the Uni versity of Massachusetts–Lowell.1) Target soils to be identified (See Figure 11. The work is aimed at pro viding practical. this formulation of solv ent blends (reduces) purchase price and blends (dilutes) en vironmental benefits 2. useful. and en vironmentally sound solutions to b usinesses within the New England re gion of the U.1.

3 Properties of “Designer” Solvents Solvent Structure Specific Mol.0 12.7 HFC-365 mfc 148.Use of Hansen Solubility Parameters 211 Table 11.0 1.820 152.0 4.0 1.6 0.0 15.0 0.0 1.9 1.1 1.0 0.2 1.550 53.1 1.4 0.0 12.0 0.1 1.0 16.2 0.0 11.1 3.7 1.6 HFC-43-10 mee 252.9 HFE-7200 264.7 2.7 0.0 13.4 OS-20 (OS-20 Octamethyldisiloxane) 236.560 47.4 0.9 12.680 55 12.0 0.430 76.5 0.2 15.6 OS-10 (OS-10 Hexamethyldisiloxane) 162. g/cc Point.0 16.9 1. Boiling δ Dispersion δ Polar δ Hydrogen-Bonding δ Overall Wt.0 14.9 14.8 .1 13.2 1.0 12.0 13.1 Solvents with Harmful Environmental Impacts But Usable in Developing Countries HCFC-141B 116.4 PFC-5060 288.0 14.0 1.0 13. °C Values in Mp½ Cleaning Solvents Designed for Minimal Environmental Impact HFE-7100 250.2 HCFC-225 ca / cb 202.7 OS-30 (Decamethyltrisiloxane) 310.2 Solvents More Useful as Refrigerants.3 0.270 40.3 2.0 0.5 CFC-113 187.580 54.7 4.7 0.2 16.6 0.6 12.1 0.4 1.854 193.520 60.5 HFE-8200 250.8 11.3 15.1 1. But Which Form Azeotropes HFC-245 fa 134.320 15.764 100.43 76 13.0 13.0 11.1 1.230 32.8 14.0 0. Gravity.0 1.4 11.0 0.7 0.

6 and Figure 11.7 . in which the cleanliness performance data are e xhibited. the solv ent cleaning operation does not ha ve a good chance to be successful.7 are similar to that in Figure 11. The limiting value of 8 will be used for R A in this study . Quality of cleaning is related to the distance in HSP space (R A) between solvent and the soil least compatible with the solv ent.212 Hansen Solubility Parameters: A User’s Handbook A group of more than 50 tests with v arious solvents (and blends) in vapor degreasing operations were examined by this author . This means less soil is left on parts in actual cleaning tests when R A is smallest.6. or a process dif ferent from v apor degreasing should be considered.6 and Figure 11.7 mak es 100% 80% % Soiled 60% 40% 20% 0% 0 ▲ ▲ ▲▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲ ▲ 5 10 15 MAX RA 20 25 30 FIGURE 11. If the R A calculated exceeds 8. the solvent cleaning operation has a good chance to be successful. The results are enlisted as sho wn in the follo wing: 1.7. APPLICATION OF HSP METHODOLOGY TO CLEANING OPERATIONS Figure 11. The basis for Figure 11. Only the ranges ha ve been changed. These solution cleaning results are consistent with and similar to the teachings of Reference 3. should be considered as the tw o-dimensional spherical solubility plots in Reference 3.6 10% 9% 8% 7% 6% 5% 4% 3% 2% 1% 0% ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲ ▲▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ % Soiled 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 MAX RA FIGURE 11. This means that: If the R A calculated for soil components and a single solv ent or azeotrope is less than 8. which are about solution of polymers or other materials within solv ents (or the re verse). Another solvent or blend. Data in Figure 11.16 It w as found that the “bright line ” between 31 acceptable and not acceptable cleaning performance is represented by an R A value of about 8.

and economic consequences than are manifested by the 13 chosen designer solv ents. The information is sorted to meet the needs of users by gi ving the various designer solvents that can be blended to clean the stated soil.20. or separation) between one type of behavior (acceptable cleanliness) and the opposite (unacceptable cleanliness).5. as well as the appendices in this book. chosen to dif ferentiate acceptable and unacceptable cleanliness. Each azeotropic blend noted abo ve35 is a fluid with a single boiling point. and there is good compatibility .4 and Figure 11. can be used as sources of HSP v alues when the six soils chosen to be representati ve in this study are not representati ve of those in the current application.8 through Figure 11. The proper metric 36 for discrimination among these azeotropes is the ef fective completion of the proposed cleaning work.15 are gi ven and illustrated in Figure 11. This is because the amount of polar and h ydrogen bonding intermolecular forces needed to produce a compatible solution cannot be obtained by adding a second solv ent rich in those forces to designer solv ents that are nearly de void of them. Users can use HSP to dif ferentiate one solv ent or solv ent blend from another based on the values of the soil found in the specific application. The average value in Appendix Table A. As viewed within the HSP space. Quantization of this mismatch is produced for specific soils by the parameter A. and is essentially the approach described in the section on “An Open Question — Answered.21. The value of R A (8).2 of Reference 3 and Appendix Table A. using HSP methodology.3 and in Table 11. en vironmental impacts. health.16. There are v ery fe w or ne gligible anomalies (situations where HSP v alues suggest that there should be poor compatibility .” to find verlap between Figure 11. The capability to clean v arious soils with azeotropes of designer solv ents is gi ven in Table 11. is similar to the interaction radius between polymers and solv ents typical for selected correlations given in Appendix Table A. There are no results of poor cleaning when R A is 8 or belo w. . F our to fi e units (MP a1/2) of polar HSP v alue can be added through addition of secondary solv ents to form constant-boiling 37 azeotropes. as well as the re verse). “bright line. This is illustrated schematically in Figure 11. ut with dif ferent physical properties.2 of this Handbook.2 is about 10.” demarcation.4 33 through Table 11. HSP methodology further allo ws recognition of the limits of azeotropic blends based on designer solvents.5. the v alue of the components from which these azeotropes are formed is to add h ydrogen bonding and polar intermolecular forces to designer solv ents that generally do not display those forces. Results of calculation of the HSP values for each binary32 azeotrope blend based on the solvents in Table 11. Information in the preceding tables can be use to screen azeotropes formulated from designer solvents and reduce the needed burden of feasibility testing. As described in the section on limiting R A value for good cleaning performance (at least 95% remo val of all soils) is around 8. Reference 3 or Reference 15. the material not cleaned (percentage soiled) by a solv ent is related to the soil least soluble. This analysis.Use of Hansen Solubility Parameters 213 physical sense. 2. clearly shows the superiority of the fluoroether structur (HFE) as a feedstock for azeotropes. safety . 3. There is a limit (breakpoint. ANALYSIS OF CAPABILITY OF DESIGNER SOLVENTS The limited utility 34 of neat designer solv ents has been e xperienced in the decade since the y were commercialized and is characterized via HSP methodology in Figure 11. That metric is the magnitude of the mismatch between intermolecular forces of the solvent (single or azeotropic blend) and composite soil. No composition is e xpected to be capable of cleaning soil number 6 (tricresyl phosphate).

008.12.390 1.9 15.541 1.S. are not likely candidates from which azeotropes can be formulated to clean an y but single hydrocarbon soils.4 5. 6.689.5 4.453 1.179 WO 96/36689.15 and Figure 11.327 WO 96/36689.9 2.0 1.3 5.P.7 δP 2. 6. 3.4 Azeotropes with HFE-7100 Primary Azeo Component HFE-7100 Isobutyl alcohol Type of Solvent Azeotrope Wt (%) HFE-7100 100% 99. 2.P.950 13.93 1.0 55.4 1.S.S. 5 1. 3. 6.3 Soils Cleaned RA < 7. 2.0 53.4% 22.2 2.304 U. U.426.3 1.327 WO 96/36689.327 U. 5 2.5 2.7 13.0 46.6.496 1.2-Dichloroethylene (CIS) n-Propyl bromide Ethyl formate 1. that no azeotrope of HFE-7100 can clean soils based on castor oil or tricresyl phosphate. 6.327 δD 13. 5 1. 5 2-Butanol Azeotrope 98.0 15.327 U.P. 3. 6.8 through Figure 11. U.S.426.0 14.327.0 52.P.8 2. 3.020.7 8.0 1. 6.20 are cluttered. U. 6.7% 25. 5 2-Propanol (IPA) Azeotrope 1-Chlorobutane Azeotrope HCFC-225 ca/cb Ethanol Azeotrope Methanol Azeotrope 1.428 1. 6.9 3.P.9 15.0% 93.P. 4.18. U.426.0 5.214 Hansen Solubility Parameters: A User’s Handbook TABLE 11.4 1. At a single glance. that HCFC-141b is a lik ely candidate to be a feedstock from which one or more azeotropes can be formulated to clean soils based on castor oil.S.S.327 WO 96/36689.S.5 1.179 WO 96/36689. 5 1. U. 4.P. 6. 5 1.426.S.646.P.327. 4.2 14. 5 1.0% 94.438 1.304 U. U. 5 1.0 31.S.8 3.753. 5 1.179 WO 96/36689.P.0 14. 4. 2. U.3 δH 1.2-D ichloroethylene (TRANS) Methyl acetate Methyl formate 2-Chloropropane Near Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope 65.0% 56.0 52. 5 1.427 1.426. 4. U.0 58. 6.P.0% 44. that OS-10 is a v ery useful feedstock from which one or more azeotropes can be formulated to clean a v ariety of soils.2 35.0 57.1 1.0 54.507 HSP Reference WO 96/36689.2 40.0 13.327 WO 96/36689.0% BP (ºC) 60. 6.6 2.6 2.3 3.S. respecti vely.6 50.S. 4. 2. U.0 SpG (g/cc) 1.5 2.327 WO 96/36689.P.426.S.426. 6.499 13. 6. 2.P. 6. .9% 56.1% 28.1 2.327 WO 96/36689. 5 1-Propanol (nPA) Azeotrope 97.426.008.6 15.753.S.361 1. 2.S. U.1% 39.2 5.2-Dichloropropane Azeotrope Azeotrope 94. In Figure 11.3 2.P.8.S. rich in information.S.426.S. U.103 0.753. and capable of ef communication.S.426.327.327 U.6 1.P.9 13. 6.426.8% 64.4 1.7 15.0 1. 2.P. 6.0 3.0 3.8% 95.9 13.304 U. 5 1.1% 87. U. In Figure 11.0 59.492 13.426.P.5% 93. 6. 2.5 2.734 U.1% 31.P.7 1. 5 2.P.426.520 1.P. 6.0% Figure 11.430 1.S. that HFC-245f a and HFC-365mfc.S. 2. 2. 2.9 1. U.7 5.100 1. 5 t-Butyl alcohol 1. one can estimate: fective In Figure 11.4% 58. 2.S.5 4.2 1. 2. 5 1.376 1.16. 6. 5 2.0 7. 3.008.P.5 14.8 7.0 1.232 1.0 54. 6.8 2. In Figure 11.9 14. U. 2.P.

4 1. and good cleaning is observ ed when R A is about 8 or less.5 2. thirteen “designer” solvents were so characterized.0% 84. CONCLUSIONS The developments described in this chapter are original and support the follo wing conclusions: 1.8 3.3 13.1% 84. In this chapter.8 2. characterization in a quantitative manner.0 1. six soil materials representati ve of common industrial soils were so characterized.8 14. tw o-component azeotropes of these designer solvents were identified. the blend HSP v alues for solvent azeotropes were compared to those of soil composites with the standard for comparison being that the R A value between them be less than 8.018 WO 96/36688 & USP 6.8 2.018 WO 96/36688 & USP 6.288. safety.6 6.018 USP 6. 5 1.3 4.293 1.0 SpG. These comparisons were tab ulated and plotted.288.0 53. °C 76.018 WO 96/36688 & USP 6.0% 56.288. 3 1.0 63. This outcome allo ws soils to be characterized in a quantitati ve manner . 2.288.1 62.0 2. or health hazards. 2.5 2.288.018 WO 96/36688 & USP 6. 2. In this chapter .8 2. 3 1.6 Soils Cleaned RA < 7. The same approach can be follo wed for a solv ent and multiple component solv ent blends with the same outcome — that is. In the same pre vious publication (Reference 16).018 WO 96/36688 & USP 6. Similarity means the distance in HSP space between the soil and the solv ent materials.6 1. .5 Azeotropes with HFE-7200 Primary Azeo Component HFE-7200 Ethanol 1. their intermolecular forces can be characterized by HSPs.307 HSP Reference WO 96/36688 & USP 6. g/cc 1. 2.9 13.829 δD 13. 2. And as a mixture of chemical components.Use of Hansen Solubility Parameters 215 TABLE 11.2 δH 1.699. their o verall HSP v alue can be computed through a con ventional volumetric blend rule. it has been sho wn that the ef fectiveness of solvent cleaning systems (v apor degreasers) can be predicted by the similarity of inter molecular forces between soil composites and solv ent blends.5 6. As chemicals. They can be described by the component chemicals of which the y are comprised.0 69. 4.0% 87. 1. 2. 3.0 67.385 1. In a pre vious publication (Reference 16).7 4. The quantitative term is R A.288.0% 87. In this chapter.262 1.0 73.93 1. 3. 2. Soils are chemicals..0 48. 2.8 15.9 δP 2.0% BP.430 1.0% 68. These solvents are notable for being similar to simple h ydrocarbons with lo w-polar and h ydrogen bonding forces and being considered to foster limited concerns about en vironmental.8 13.8 13. 4 It w ould be dif ficult to imagine h w another technical approach could allo w ef ficient iden tification of sol ent capability for industrial cleaning from a table of azeotropes in a handbook.388 1.2-Dichloropropane 1-Chlorobutane 1-Butyl Alcohol n-Propyl Bromide Methanol 1. 5.417 1.2-Dichloroethylene (TRANS) Type Solvent Azeotrope Azeotrope Near Azeo Azeotrope Azeotrope Azeotrope Azeotrope wt % HFE-7200 100% 88.0 4.5 13.266 1.

3. 5 1.6 Azeotropes with HFE-8200 Primary Azeo Component HFE-8200 HCFC 225 ca/cb Isobutyl alcohol 2-Butanol Azeotrope t-Butyl alcohol 1-Propanol (nPA) Azeotrope Azeotrope Type of Solvent Azeotrope Azeotrope Wt (%) HFE-8200 100% 30.5 3. 2.338 1. 6. 6.2 2.2 4. 5 Azeotrope Azeotrope 65.8 1. 6. 6. 5 1.0 40.7 1. 6.304.327 U. 5 .104 15. 5 2.2% Azeotrope 79.327 U.0 1.327 U.0 56.327 U.0 3.426.327 U.4 1.4 5.3 2.0% 98.5% 93.S.426.P.753. 6.3 13. U.S. 4.P.0 4.0 13.426. 2.417 1. 5 2.4 4.P.P.3 1.3% 87. 5 1.122 15.P.7 1.1% 50.402 1. 3.511 1.426.3% 1. 6.426. 5 1.0 2.0 4. 2.P. 3.P.P.1 3. 4. 2.S.0 58.0 58. 6.304. 6.7 δP 4. U.0 52.5 4.S. 2. 2. 3.S.S.426.2-Dichloroethylene (CIS) Methyl acetate Azeotrope 25.327 U.S.9% BP (ºC) 76.8% 89.319 1.287 HSP Reference U. 6.419 1.426.1 2.P.1 13.2-Dichloropropane Azeotrope 2-Propanol (IPA) Azeotrope 1-Chlorobutane Azeotrope Ethanol Azeotrope Ethyl formate n-Propyl bromide Methanol Azeotrope 1.2 δH 1. 5 1.6 5. 5 1.5 6. 6.2 13.P.P. U. 4. 4.2 4.S.426.1 7.0 57.S.426.S. 5 1.646.430 1.0% 97.2 7.5 14.1 14.0 31.P.0 0.329 1.327 U.426.327 U.426.4% 94. 5 1. 6. 6.020. 6.4 5. 3.426.376 1.9 2.0 52.P.7 13.0 56.3 13.362 1.327 U.327 U.1 4.6 13.426.753.1 4. 4.942 15. 6.S.1 13.426.8% 95.0 55.7 1.S.S. 5 1.3 4.327 U. 2.426.376 1.327 U. 5 1.356 1.P.0% 35.P. 3.8 5.0 1.1 Soils Cleaned RA < 7.412 1.1 7.P. 6.216 Hansen Solubility Parameters: A User’s Handbook TABLE 11.2-Dichloroethylene (TRANS) Methyl formate Azeotrope Azeotrope 50. 6.3 1. 2. 2.0 53.S.8 1. 6.6 6.P. 5 1.3 1.8% 94.S.3 SpG (g/cc) 1.0 13.6% 54.7% 44. 2. 5 1. 6.327 δD 13.S.0 59.426.P. 5 2-Chloropropane Near Azeotrope 22.327 U.P.3 13.S. 6.S.327 U.356 1.S. 2.S.S.327 U.327 U.2 13.0 55.P.753.93 1.4 1. U.6% 99.0 46. 6.P.426. 2.304.4 14. 2.4 5.0% 40. 2.

5 1.Use of Hansen Solubility Parameters 217 TABLE 11. The comparison showed that expected cleaning performance of designer solvents can be enhanced by combining it with other solv ents into binary azeotropes.6 100. 6.0% 73.4 13.7% BP (ºC) 54.6 2. 5773.0% 82.S.1-Dichloroethane Azeotrope Acetone Azeotrope 1.734 U.0% 87. 2.S.0 0. 3.9 Soils Cleaned RA < 7.9 δH 0.770 0. 5773.0 δP 0.774 0.P.7 3.403 6.0% 61.9 13.9 12.7 Azeotropes with HFC-4310mee Primary Azeo Component HFC-4310mee 2-Propanol (IPA) Ethanol Methanol n-Propyl bromide Type of Solvent Azeotrope Azeotrope Azeotrope Azeotrope Wt (%) HFC-4310mee 100% 97.4 39.P. such as tricresyl phosphate.8 1.5 2.580 1.9 3.0 SpG (g/cc) 1.3 0.93 1. 5 Type of Solvent Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope Azeotrope Reference U.7 76.4 Soils Cleaned RA < 7.3% 23. 3.540 1.403 U.P.2 12. . 5773.5 12.0 57.6 2.5 δH 0.535 1. 2. 2.P.8 95.P. 5834.0 1.7 SpG (g/cc) 0.5 2.509 1.P. 5 1. 2.6 14.416 U.8 1. 5 2. are not e xpected to be remo ved in a solv ent cleaning process emplo ying azeotropes of these designer solv ents.S.403 U.416 U.0 0.7 96.4 2.S.4% 97.493 U.0 0.9 13.764 0.0 1.5 1.S. Ho wever. 2.P. 5824.5 δP 0.1 0. soils that display high levels of polar and hydrogen bonding forces.S.443 HSP Reference Vertrel XP Vertrel XE Vertrel XM U. 5 1.8 Azeotropes with OS-10 Primary Azeo Component OS-10 Hexamethyld isiloxane Sec-butyl acetate 2-Pentanol Propylene glycol methyl ether n-Propyl acetate 2-Propanol (IPA) Ethanol Methanol Wt (%) OS-10 Hexamethyld isiloxane 100% 96.8 11.1 4.P.S.6 42.788 0.3% 63. 2.7 7.1% HSP BP (ºC) 100.0 0.515 1.137 Vertrel MCA δD 11.8 11.S.S.1 1.0 1.93 1.2 1.1 1.P. 4. 5196.9% Azeotrope 67. 5834.4 71. 5 1. 3.493 U. That enhancement is a g ain of about four or fi e units (MP a1/2) of total (Hildebrand) solubility parameter . 2.768 0. 2.7 13. 5 1.2-Dichloroethylene (CIS) 1.0% 83.689.9 2.775 δD 12.P. which allows prediction of ef fective cleaning of se veral common industrial soils.1% 95. 5196.4 12. 2.5 43.9 1.4 6. 2.5 97. 5 1.808 0.471 1.445 1.3 1. 5 1. 5478. 5 1.1 2.9 8. 5 1.773 0.4 58.0 43.8 12.9 52. 5478.0% 54.3 39.361 1. 2. 5 1.4 45.S.9% Azeotrope 61. 2.8 13.6 13.P.8% 59. 5 1.6 11. 5 1.137 U.S. 5 1.2-Dichloroethylene (TRANS) TABLE 11.634 U.

8 3.4 0.632 1. It appears possible 16 to identify an azeotropic solv ent pair that is capable of being expected to clean an y proposed soil composite. 5516.866 U.S.P.93 1.0 1.8 0. 5824. DPGME) Propylene glycol monobutyl ether (PGBE.8 SpG (g/cc) 0.S.833 δD 11. 5454.1 δP 0.1 2.9 Azeotropes with OS-20 Primary Azeo Component OS-20 Octamethyltrisiloxane Propylene glycol n-butyl ether Propylene glycol n-propyl ether Ethyl lactate Wt (%) OS-20 Octamethyltrisiloxane 100% HSP BP (ºC) 152.10 Azeotropes with OS-30 Primary Azeo Component OS-30 Decamethyltrisiloxane Dipropylene glycol n-propyl ether Dipropylene glycol methyl ether acetate Dipropylene glycol methyl ether (DPM. 3.S.0% 170.S. 5454.P.0% Azeotrope 60.S.826 Soils Cleaned RA < 7. “5824. 5 0.3 Azeotrope 89.5 2. 3. U. 5 Azeotrope 15.4 HSP δD 12.P.1 1.888 13.844 13.1 1. 5628.874 U. This is a crucial point because additional solv ents other than the designer solv ents can be used as a feedstock to formulate azeotropes. 3.632 0.0 0.0% 139.0 1. 2.5 13.7 193. 5824. 5824.2 1.970.8 7.P. .0% Azeotrope 61. 2.0 Soils Cleaned RA < 7.859 U.0 1.93 Type of Solvent Reference Azeotrope 91.632 0.1 δH 0. 2. 5 Type of Solvent Reference U.218 Hansen Solubility Parameters: A User’s Handbook TABLE 11. 2.P. PnB) Wt (%) OS-30 DecamethylBP SpG trisiloxane (ºC) (g/cc) 100% 193. 5628.0% 141.8 170. 5 TABLE 11. 5 Azeotrope 63.0% Azeotrope 89.0 0. 5628.S.4 0.632 0.6 14.820 0.P.833 U.S.4 4.P. 5 The methodology used in this chapter can be easily emplo yed with other soil composites and solvent blends.8 151.889 U. 4.4 δH 0. 2. 2.8 1.833 U.8 2.854 186.0% 0. Such additional solv ents will not be as limited with re gards to polar and h ydrogen bonding intermolecular forces as are the designer solv ents chosen for this study .9 5. U.S. 5 4.2 12.4 12.4 δP 0.9 0.P. 3.7 12.P.S.

93 1.P.365566 U. 5 1.S.3 1.P.0% 96. 6.Use of Hansen Solubility Parameters 219 TABLE 11. 5 1.3% 96.2 14.P.2 0.6 3.020 δD 16.11 Azeotropes with HFC-245fa Primary Azeo Component HFC-245fa Trichloroethylene (TCE) HCFC-123 Near Methylene chloride 2-Propanol (IPA) 1-Propanol (nPA) Methanol Near Ethanol Near Type of Solvent Wt (%) HFC-245fa 100% BP (ºC) 15.7 15.3 2.8% 96.7 16.5% 1.291 1.241 1.320 1.0 0.P.283 1. 6.P. 5445757 U.765 U.977 U. 5 Near-Azeotrope 98. 5 1.4 0.2 14.100.743. 5 .3 0. 5 1.7 15. 6.8 15.7 3.2Near-Azeotrope Dichloroethylene (TRANS) Water Near -Azeotrope 83.322 1.258 1.974 U.0% 7. 6.153 U. 6.9 1.0 0.P.8 5.7 15.P. 1. 1. 5683.S.9 0.6 0.2 Soils Cleaned RA < 7.7 15.0 0.974 U. 5851.0 16.4 14.2% 60.974 U.S.S.009 Reference U.5 1.93 1.8% 37.263 1. 5 5 1.229 U.100. 5 1.4 0.0 1.4 16.1 δH 0.646.2 38.P. 2. 6. 5 1.P.3 Soils Cleaned RA < 7. (g/cc) 1. 5 1.1 0.323 1.294 1.6 33.255273 U.9 1.251 15. 6.S.295 HSP Reference U.2 δP 0.293 1.8% 97.450 1.4% 98.12 Azeotropes with HFC-365mfc Primary Azeo Component HFC-365mfc Ethanol Azeotrope Water Azeotrope Methanol Azeotrope HCF2OCF2OCF2H NearAzeotrope 2-Chloropropane Azeotrope HSP Type of Solvent Wt (%) HFC-365mfc 98. 5683.S.S.S.1 δP 0.0% BP.7 15.974 U.S.1% 46.P.S.514.5 0.1 0.1 0.7 15.320 1.0 0.3 δH 0.7 15. 5683.3 13.1 1.5 14.0 15.4 16.P.0% 95.4 0.S.P. 5 1.362.0% 98.S.0 SpG.0 0. 5 TABLE 11.S.3 14.1% -Azeotrope Near-Azeotrope Near-Azeotrope N ear Azeo -Azeotrope -Azeotrope 98.4 16.S.P.0 — 32.7 3.P. 5 1.1 0. 6. 5683.229 U. (ºC) 39.4 — SpG (g/cc) 1.1 0.928 δD 15.

0 0.1 20.14B Azeotropes with HCFC-225 ca/cb Primary Azeo Component HCFC-225 ca/cb HFE-7100 Azeotrope Wt (%) HCFC-225 ca/cb 100% 28.0 26.0 31.5% 98.861 USP 5085798 USP 5126.1 14.511 13.0% HSP BP.6 4. 5 1.1 4. °C 32.959 δD 15.0 Soils Cleaned RA < 7.327 U.7 5.0 0. 6.716 TABLE 11. 2.861 TABLE 11.1 14.7 3. 5 Reference USP 5494. 5 Type of Solvent Azeotrope Azeotrope Reference USP 5560. 2. 2.561 δD 12.8 δP 4.216 1.7 15. 5 1. 2.0 SpG (g/cc) 1. 5 1.067 USP 5085797 USP 4.2 15.222 1.0 1.6 0.0 31.9 4. 5 1.1 4.8 14.0 1.2 32.218 1.6 2.334 1.5 Soils Cleaned RA < 7.420 1.426. 2.2 32.3 δH 1.0% 70.9 δH 1.0 1.7 13. 5 1. 5 1. g/cc 1.0 δP 3. °C 56.420 1. 2.0 32.4 16. g/cc 1.2 2. 3.0% 50.0% 95.426. 4.3 3.5% HSP BP (ºC) 54.0 1.2 31.7 15.7 15.0 1. 5 . 2.861 USP 5560.541 δD 14. 4.0% 50.680 1. U.201 1.0 26.6 4. 5 1.230 1.7 δP 0. 5 Type of Solvent Reference WO 96/36689.5% 98.5 SpG.169 PROMOSOL 141b MC USP 5318.0% BP.2-Dichloroethylene (TRANS) 2-Chloropropane Ethanol Methanol HCFC-123 Methylene Chloride 2-Propanol (IPA) Type of Solvent Azeotrope Azeotrope Near Azeo Near Azeo Near Azeo Near Azeo Blend Near Azeo Blend Near Azeo wt % HCFC-141b 100% 50.230 1.7 15.5 1.0 2. 2.1 4.4 Soils Cleaned RA < 7.836.14A Azeotropes with HCFC-141b HSP Primary Azeo Component HCFC-141b Perfluoropentane (PFC 5050) Perfluoroh xane (PFC 5050) Cyclopentane 1.S.1% 98.5 10. 6.93 1.4 15.6% 50.0 26.0 2.257 0.S.9 30.P.13 Azeotropes with PFC-5060 Primary Azeo Component Perfluorohexane (PFC 5060) HCFC-141b HCFC-123 wt % Perfluorohexane (PFC 5060) 100% 50. 5 2.0 2.6% 53.327 HFE-8200 Azeotrope 30.947 PROMOSOL 141b MS USP 5194.2 13.7 15.6 0. 5 1.93 1.3 2.0 53.0% 50. 5 1.9 δH 0.414 1.5 SpG.601 USP 5560.6 1.4 1.5 31.93 1.0 1.3 2.220 Hansen Solubility Parameters: A User’s Handbook TABLE 11.550 1.P.0% 98.

3.0 RA < 7.9 3.815 δD 14.S.3 2.2 1. 5 1.9 .560 1.817 U. 2. 5 Type of Solvent Reference U.8 HSP OF BINARY AZEOTROPES 7 Azeotropes with HFE-7200 Hydrogen Bonding HSP.5 1.0 45. MPa^(1/2) 12 14 ASTM 0 FUEL “A” 0 2 FIGURE 11.Use of Hansen Solubility Parameters 221 TABLE 11. MPaˆ (½) 12 14 FIGURE 11.3 15.999.8% 87.S.5 HSP OF BINARY AZEOTROPES 17 Azeotropes with HFE-7100 14 Hydrogen Bonding HSP.P.1.4 14.0 HSP SpG (g/cc) 1.P. 5 1. 2.7 2.93 0.522 1.15 Azeotropes with CFC-113 Primary Azeo Component 1.391 Soils δH Cleaned 0.6 Nitromethane Azeotrope Methylene chloride Azeotrope Acetone Azeotrope 14.8 37.2-Trichlorotrifluoroethane BP (CFC 113) (ºC) 100% 47.999. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL Methyl Formate ETHYL CINNAMATE BUTYL STEARATE Methyl Acetate Ethyl Formate LINSEED OIL Ethanol TRICRESYL PHOS ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.0 1.S.P.1% 85.5% 46.2-Dichloropropane 4 6 8 10 Polar HSP.544 1.213 U.1.2-Trichlorotrifluoroethane (CFC-113) Wt (%) 1.573. 2. MPa^(1/2) 14 12 10 8 6 4 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE Methanol BUTYL STEARATE t-Butyl Alcohol 1-Chlorobutane LINSEED OIL ?????????????? ?????????????? 2 Ethanol 1.7 δP 1.6 97.

MPaˆ (½) 12 14 FIGURE 11. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE Methyl Formate BUTYL STEARATE Methyl Acetate Methanol LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP. MPaˆ (½) 12 14 FIGURE 11.10 FIGURE 11.222 Hansen Solubility Parameters: A User’s Handbook HSP OF BINARY AZEOTROPES 17 Azeotropes with HFE-8200 14 Hydrogen Bonding HSP.12 .11 HSP OF BINARY AZEOTROPES 7 Azeotropes with OS-10 Hexamethyldisiloxane 14 Hydrogen Bonding HSP. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL 2-Propyl (PA) Ethanol ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL TRICRESYL PHOS ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.

MPaˆ (½) 12 14 FIGURE 11. MPaˆ (½) 12 14 FIGURE 11.15 .14 HSP OF BINARY AZEOTROPES 9 Azeotropes with HFC-245fa 14 Hydrogen Bonding HSP. MPaˆ (½) 12 14 FIGURE 11. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE Propylene Glycol n-Propyl Ether LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.13 HSP OF BINARY AZEOTROPES 4 Azeotropes with OS-30 Decamethyltrisiloxane 14 Hydrogen Bonding HSP. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE Propylene Glycol n-Propyl Ether LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.Use of Hansen Solubility Parameters 223 HSP OF BINARY AZEOTROPES 3 Azeotropes with OS-20 Octamethyltrisiloxane 14 Hydrogen Bonding HSP. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.

MPaˆ (½) 12 14 FIGURE 11. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL Methanol 2-Chloropropane 4 6 8 10 Polar HSP.17 . MPaˆ (½) 12 14 ASTM 0 FUEL “A” 0 2 FIGURE 11.224 Hansen Solubility Parameters: A User’s Handbook HSP OF BINARY AZEOTROPES 5 Azeotropes with HFC-365mfc 14 Hydrogen Bonding HSP.16 HSP OF BINARY AZEOTROPES 2 Azeotropes with Perfluorohexane (PFC) 14 Hydrogen Bonding HSP. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.

19 .18 HSP OF BINARY AZEOTROPES 2 Azeotropes with HCFC-225 ca/cb 14 Hydrogen Bonding HSP. MPaˆ (½) 12 14 FIGURE 11. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP. MPaˆ (½) 12 14 FIGURE 11. MPaˆ (½) TRICRESYL PHOS CASTOR OIL 12 10 8 6 4 2 ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP.Use of Hansen Solubility Parameters 225 HSP OF BINARY AZEOTROPES 10 Azeotropes with HCFC-141b 14 Hydrogen Bonding HSP.

HCFC-225 ca/cb. HFC365mfc. HCFC-141b. OS-20. HFE-7200.16 Capability of Azeotropes of Designer Solvents Soil Number 1 2 3 4 5 6T Soil Name ASTM fuel “A” Butyl stearate Castor oil Ethyl cinnamate Linseed oil ricresyl phosphate Designer Solvents from which Azeotropes Can Be Blended to Clean the Stated Soil (RA < 8) HFE-7100. HCFC-141b. OS-20.20 Polar HSP (soil or solvent) Soils Soils Soils Azeotropes ne nt s m po Soils “Designer” Solvents H2 Bonding HSP (soil or solvent) FIGURE 11. OS-20. OS-10. HFE-8200. OS-30. CFC-113 HFE-7100. OS-30. HCFC-225 ca/cb. HFE-8200. PFC5060. OS-10.21 TABLE 11. OS-20. OS-10. OS-30. HFC-365mfc. HFE-8200.226 Hansen Solubility Parameters: A User’s Handbook HSP OF BINARY AZEOTROPES 3 Azeotropes with 1. CFC-113 None Ad d iti on of Co . MPaˆ (½) 12 14 FIGURE 11. OS-10. CFC-113 HFE-7100. HFE-7200. HFC-245f a. HCFC-141b HFE-7100. PFC-5060. OS-30. HCFC-141b. PFC-5060.2-Trichlorotrifluoroethame 14 Hydrogen Bonding HSP. HFE-8200. HFE-7200. HCFC-225 ca/cb. HFE-7200. HCFC-141b HFE-7100. MPaˆ (½) 12 10 8 6 4 2 CASTOR OIL TRICRESYL PHOS ETHYL CINNAMATE BUTYL STEARATE LINSEED OIL ASTM 0 FUEL “A” 0 2 4 6 8 10 Polar HSP. HFC-245f a.1. HFC365mfc. HFE-7200. OS-10. OS-30. HFE-8200.

It is understood that φcomp 1 + φcomp 2 = 1.B. Durkee. there is a combined ef fect. because it is that basis by which solv ents are normally sold. It is that solv ents are miscible.. In Chapter 11. 16. and δ is an y solubility parameter . 10. 13. 3.. In Equation 11. and soils are soluble in solvents. Refers to the total or Hildebrand solubility parameter . Hansen Solubility Parameters — A User’s Handbook. Management of Industrial Cleaning Technology and Processes.Use of Hansen Solubility Parameters 227 NOTES 1. 2006 by Elsevier Science. CRC Press. 4. The appendix to Reference 3 is an excellent source of HSP values for solvents (Appendix Table A. φ is the v olume fraction of component 1. C. An efficient approach to making this obser ation is to carefully apply some of the liquid soil composite to a glass slide..9 and Equation 1.M. See Chapter 1. although the y are sold by weight. J. 1991. polar. the e xtent of graphical area from the center point must be measured. and Equation 11. One must be cognizant of Heisenber g’s uncertainty principle in preparing to make this observ ation. 61. regions that are multiple phases can usually be delineated.1 as a circle. Chapter 1. The radius (that will be identified as R ) of each circle (sol ent) represents their sphere of similarity in two dimensions.9 in Chapter 1 in Reference 3. When the coated slide is illuminated from belo w. H wever. as the soil sample must be equi valent to the actual residue. This can be found in Durkee. The center point of the circle for each solvent is the locus of their polar and hydrogen bonding HSP values. November 2005. Volume fraction is easy to compute because solv ents are stored in pails or drums and used by v olume.1 is used to calculate that effect for any blend of solvents. CRC Handbook of Solubility Parameters and Other Cohesion Parameters. a sufficient condition for a successful coating application. 5. the region is visualized in Figure 11.. FL.1.3 in “Method for Calculating HSP of Composites (Soils or Solv ents)” in this chapter . Reference 3. Data found here are one feedstock for implementation of Equation 11.B. . Figure 11. The capacity of a v apor degreaser sump is gi ven in g allons or liters.M. Subboiling (cold cleaning) is not the process for which use of the solv ents mentioned in this chapter is envisioned.1 as circles. Because these two solvents are different but miscible. 6. e Composition is normally gi ven on a weight basis in this chapter . 7. This is a re gion or zone in HSP space within which HSP v alues are suf ficiently similar to those a the center point. Boca Raton. In two dimensions. J. cleaning solv ents are assumed to be used in a v apor degreasing process. 11. FL. Hansen. Oxford and Amsterdam. Boca Raton. Durkee. See Equation 1. thorough solutioning is usually a necessary condition for either type of application to be successful. its material safety data sheet (MSDS) will lik ely contain infor mation that either is the recipe or from which a suitable estimate of the recipe can be inferred. Note that thorough solutioning of a soil by a cleaning solv ent is no more a sufficient condition for a successful cleaning application than is thorough solutioning of a polymer in a coating carrier solv ent. 8. That is the center point for the combined circle. 2nd ed.There are two aspects of this idealized representation of solv ent character as graphical areas: (1) solvents or soils or polymers are compatible (lik ely miscible) when their areas overlap (2) whereas the center point HSP v alues can be computed from molecular characteristic (properties). CRC Press. The latter will be shown in the section on “An Open Question — Answered. 2006 by Else vier Science. Another useful reference is Barton.B. Equation 11.. and Chapter 8 of Reference 3. Disperse.1). Equation 1. or total. Similarity has a practical meaning.1. Oxford and Amsterdam. though that is not the case in this xample. Cleaning Times column in Metal Finishing Magazine. polymers. 2.1 is an idealized representation. J. Chapter 2. hydrogen bonding.” 9. What W ould the ‘Perfect’ Cleaning Solv ent Be?. A. 14.10. p. and some soils (Appendix Table A. 1999. Solubility behavior of the tw o solvents is represented in Figure 11. The reason for inclusion of molecular weight v alues in the basic data is because the mixture concentrations are often given in molar concentration. 15. 17. Management of Industrial Cleaning Technology and Processes. The other feedstock is composition information about azeotropic solv ent composition. 12.F.2).. If the soil is a commercial product.

21. The other major acidic component of the triglyceride is linolenic acid that has one fe wer double bond. 27. 23. Castor oil is also referred to as the triglyceride of this acid. 28. because of its inertness.1 is a generalized presentation. selection of a suitable cleaning solv ent based on similarity of intermolecular forces is but a necessary first step in d velopment of cleaning process applications. In cleaning.” 26. HFE-7100 and HFE-8200 are considered equivalent in terms of the boiling point and composition of azeotropes. Certain selection criteria are listed as follo ws: • A solv ent that has a strong af finity for a soil ut a lo w holding capacity for it (mass solubility) would be a poor choice. all components are generally compatible with the carrier solv ent. Ho wever. all components must be manageable. a poor choice is a solv ent that only gradually penetrates and swells the soil and allo ws it to be removed by rinse fluids • Finally. soil components can be partially soluble. United States Pharmacopeia (USP) 6. Finding the number 4 in Equation 11. insoluble. This difference has been retained in the calculations of solubility beha vior of the azeotropes. the HSP properties of HFE-7100 and HFE-8200 are slightly dif ferent. 511.228 Hansen Solubility Parameters: A User’s Handbook 18. where there can be destruction of coatings. “A chain is only as strong as its weak est link.. These values are relatively similar. without successful completion of a screening evaluation via use of HSP.” 20. Note that Figure 11. development of cleaning process applications becomes either resource intensi ve or based on hearsay e vidence. HFE-8200 is actually a component of the product sold as HFE-7100. Said in a colloquial manner . Hansen. 19. • A solvent selected to do cleaning w ork must be able to dissolv e the soil to the e xtent desired in the time allotted under the pre vailing conditions. and lo w surf ace tension. The three dimensional solubility parameter — key to paint component affinities. high specific gr vity. or soluble.. Vol. 511–514. 1967. There are two justifications for it: (1) it co verts three-dimensional spheroidal plots with dispersion solubility parameters to spherical ones. in this analysis. or soils can be matched to other soils for the purpose of e valuating if solution-based or force-based (aqueous technology) should be used. is the most common (52%) acidic component of linseed oil. solv ents can be matched to other solv ents for the purpose of producing a blend with certain ph ysical properties. Yet. Paint Technol. its use is limited “by design. management involves hydrodynamic (mechanical) force. In a coating. K. In cleaning. See Chapter 5 in this book. The structure shown represents the most common (89%) component (ricinolenic acid. and (2) theoretical considerations associated with the Prigogine theory of corresponding states. It has been found empirically useful o ver several decades of practical experience. No.M. Generally . linoleic acid. HFE-8200 is the pure perflu oroisobutyl methyl ether. Hence. . Consequently. Soils can be matched to solv ents for the purpose of solvent selection. depending upon whether the soils are e xpected to combine in a single or multiple phases. swollen. 12-h ydroxyoleic acid) of castor oil. The structure sho wn.327 teaches that azeotropic behavior of mixtures of the tw o isomers and another component is relati vely independent of distribution of isomers in the HFE material.426. PFC-5060 is a designer solv ent that on emission migrates without de gradation to the Earth ’s stratospheric layer and acts as a global warming agent. Unfortunately. 25. a cleaning process can be designed to manage remo val of soil components from surf aces without re gard to their relationship to the solv ent.3 may be surprising. • Also. 24. and solv ency. it finds use i cleaning work as a rinsing agent. Pressurized contacting equipment is expensive. a solvent that efficiently dissol es an adequate mass of soil in an ef ficient time ut only at a temperature above its boiling point is nearly useless. The disperse HSP v alue is momentarily ne glected in the interests of graphical clarity when a tw odimensional graph is used. and Skaarup. In solv ent cleaning. J. Cleaning is a transport process. 305. heat. There is considerable discussion in the literature about the need for it. 22. Owing to its inertness as a solv ent. cleaning is unlik e construction of coatings. C.The latter is mixture of Perfluoro n-butyl methyl ether (35%) and perfluoroiso utyl methyl ether (65%). In this w ay. Yet.

A secondary standard is to fulfill the co ventional requirements upon which solv ent cleaning w ork (vapor degreasing) is based.S.T . In this study . As all of these designer solv ents have commercial v alue as unique compounds. the composition (like that of a single component solv ent) remains constant. F or these laboratory trials. but not the ro ws that are not sorted by that parameter . Normally. Readers are reminded that the normal interpretation of U. health. Elimination of those concerns was roundly and justifiably applauded by users. In a sense. This phrase refers to the identifying characteristic of an azeotrope. 31.S. all azeotropes noted are minimum-boiling azeotropes. their concentration in th applied soil mix w as not. Boiling points are also gi ven. 35. There is no published table of these binary azeotropes e xcept that found in Reference 15. one can say with 95% conf dence that if the R A value is less than 8. S. if these binary azeotropes are not commercial products. the limit of acceptability re garding cleaning performance is set by the requirements of the next processing or use step. and test e xposure times were used. these azeotropes are claimed in U. Only blends defined by measured p ysical property and boiling point data are included. No denigration of designer solvents is intended in this work.org. patents whose grant date is during the 1990s. patent.14 are sorted by increasing v alue of polar HSP with that of the base designer solv ent given in the initial ro w. Azeotropes in Table 11. then the w orst level of cleanliness of the least compatible soil component w ould be 94%. that research w as too successful. In other w ords. 30. 33. But in vapor degreasing operations. and (3) whereas the soil components were identified. In other w ords.S. ASIN: B0006R4J04. They were identified in 2004 during an xtensive search of the world patent literature. Locating All Heterogeneous and Reactive Azeotropes. Both ternary and quaternary azeotropes containing designer solv ents have been identified and patented Some find applications in aerosol-p wered cleaning products. The patent assignee is most often the manuf acturer of the designer solv ent. the ternary mixture is still co vered by the U. Most w ork was done around 2004. (January 1. The method of assaying cleanliness is gra vimetric. (2) soil components were identified b chemical abstracts service (CAS) number from the manuf acturer’s MSDS that is not al ways a complete listing of ingredients. there w as considerable le vel of replicates a vailable through the use of multiple runs where the same solvent was used under a dif ferent product identification. 34. as the entire v olume applied evaporated quickly after application. patent law is that when a third component is added to a claimed binary composition. In the main. designer solv ents were those whose use raised fe wer concerns. users cannot formulate them without permission from the patent holder. There are se veral uncertainties about use of these 53 data points: (1) a v ariety of v apor degreasing processes. as the mixture boils. Theoretical or predicted azeotropes are not included because the y cannot be purchased as blends for cleaning work. The expected value of cleanliness w ould be around 97%. no such information w as available. The parts used are those from the b usiness operation.4 through Table 11. 37. as a class of product. these solvents represent the result of significant and successful chemical research into identification and man acture of chemical structures. 1998). 32. which is the venue of this chapter. The soils are those identified by the usiness as being of concern and are applied as pure components to pre viously cleaned parts (or coupons). which lists around 450 of them.cleanersolutions. the azeotropes the y form with other solv ents are claimed in patents. All information is publicly a vailable. Here. U. which alle viate man y of the safety . Rather. binary) azeotropes with use cannot be ne glected..S. . See http://www. distrib uted by the American Institute of Chemical Engineers. See Reference 2 for details. there is little concern about composition change in use. Yet. b ut whose value in use w as more limited. patent numbers are gi ven as references where available. Seemingly. Binary azeotropes based on designer solv ents are an attempt to span the g ap between absence of concern and absence of v alue in use. the concern about composition change of ternary and quaternary (vs. and en vironmental concerns about solv ents commonly used in cleaning (and other) w ork in the late 20 th century. that the composition of the v apor and the composition of the boiling liquid are identical. A good source of information about predicted azeotropes is Harding. test conditions.Use of Hansen Solubility Parameters 229 29. 36. Hence.


and simple e valuations of chemical resistance of the suitable-for -use or not type. de gree of surf ace attack. Some surf ace studies may in volve evaluation of a more direct form of chemical attack. HSP correlations are presented for chemical resistance studies of epoxy and zinc silicate tank coatings. as well as else where. tensile strength reduction.13–18 can also provide some insight into chemical resistance. Obtaining a good HSP correlation of chemical resistance that allo ws reliable predictions depends v ery much on careful treatment of the available data or generation of data with such a correlation in mind. Examples of HSP correlations included here are for solubility . there are certain additional pitfalls to be aware of when correlating chemical resistance. This has been discussed in Chapter 8 and also in References 1 through 12. etc. The HSP correlations of surf ace phenomena. POM. These include (lack of) attainment of equilibrium. then it is quite unlikely that it will attack a fully crosslink ed epoxy coating or glue. PUR. 19 This is discussed in more detail in Chapter 14. the general considerations of the HSP characterizations discussed in earlier chapters are the same for the HSP correlations of chemical resistance reported here.— Chemical 12 Applications Resistance Charles M. PPS. However. the effects of molecular size of the test chemical. This means that chemical resistance for some polymers can be partly inferred from HSP correlations of their solubility and/or swelling. En vironmental stress cracking is discussed in more detail in Chapter 14. Chapter 16 treats absorption and diffusion in polymers with this in mind. it can be used to calculate the behavior of other systems that have not been tested. The fact of solution is in itself clearly one simple form of chemical attack of the polymers the y dissolve. Liquids that spontaneously spread were found to induce en vironmental stress cracking at lower critical strains than those liquids that do not. which have been discussed in Chapters 6 and 7 and else where. PEI. Liquids not wetting a surf ace are not as likely to attack it as those that do wet it. An example is that if a chemical does not dissolve an epoxy component or the curing agent. Hansen ABSTRACT Hansen solubility parameters (HSP) can correlate dif ferences in ph ysical beha vior observ ed in chemical resistance testing of polymers and polymer -containing systems when a suf ficiently la ge number of dif ferent organic solvents are included in a study . This example is discussed in more detail later . In each case. PA6/66. A relation between spontaneous spreading and de wetting of liquids and en vironmental stress cracking has been found. A common problem is that tests with larger molecular weight liquids have not reached equilibrium absorption within the timeframe of the e xposure. although there are no guarantees. and acid/base reactions. Once a reliable HSP characterization of chemical resistance is a vailable. These correlations can then be used to predict the chemical attack e xpected in systems that have not yet been tested. PET. Whatever is being correlated. very few 231 . Unfortunately . HSP correlations of this type ha ve been discussed in Chapter 5. such as the attack/whitening of PET coated with “amorphous” PET to improve weldability. Neoprene ®. INTRODUCTION HSP are widely used in the coatings industry to select solv ents for dissolving polymers and binders. the molecular size of the liquids used can affect the result and should be considered in some w ay. dif ference in local se gments of polymers (e ven in homopolymers).

20 Some of the results of this study are discussed in Chapters 15 and 18. say . e ven though it may not ha ve been tested at that temperature (and in principle the HSP are only v alid at room temperature). Larger molecular species which are “poorer” than expected by comparison with the data for the other solvents are often seen at the bottom of the list. Likewise. particularly when the list of agents is long. it is often possible to find xcellent solubility parameter correlations using chemical resistance data of the acceptable-or -not type. it must be clear that the chemical attack discussed in this chapter does not include true chemical reactions leading to covalent bonding or destruction. with the use of due caution. Chemical reactions forming ne w compounds are often found with amines and or ganic acids. CHEMICAL RESISTANCE — ACCEPTABLE-OR-NOT DATA Additional sources of data for solubility parameter characterizations include chemical resistance tables reported by ra w material suppliers 21–26 or collected in books 27–29 and other sources such as those supplied on electronic media by the Plastics Design Library .30 Although these data are certainly valuable in themselv es. Additional precautions with re gard to data of the acceptable-or -not type include whether a molecular size ef fect is present as discussed in the follo wing. if the top of the list includes the type of “error” where the smaller molecular species are “better” thanxpected e by comparison with all the other solv ents. and 5 that the size of solv ent molecules is important for polymer solubility. and this ef fect is rarely confirmed or ven considered. true chemical attack with acids and bases must sometimes be sorted out. What appear to be errors in the correlation may become systematically arranged. 20°C. One can focus upon the molecular size range of greatest interest in such cases and repeat the correlation. then it will also attack it at. Nevertheless. all else being equal. that is. These reactions often lead to discoloration in systematic solubility parameter testing with one or more of the amines used as test solv ents. such as with acids and bases. 2. So it is not surprising that solv ent molecular size can be an important fourth parameter in correlations of chemical resistance. being more aggressive than expected. This is known from the theories of polymer solubility discussed in Chapters 1 through 4. An appropriate w ay to check this is sorting output data from a computer (or other) HSP optimization according to the molecular v olume of the test solv ents. it can be assumed that if a chemical attacks a polymer at. respecti vely. and also from the discussion of barrier polymers and dif fusion found in Chapter 13 and 16. whether they are or ganic or inor ganic. Also. ne glecting those species which are outside of this range of molecular v olumes. As discussed earlier . penetrating more rapidly. 70°C. Discolored systems should simply be neglected in HSP correlations of ph ysical (reversible) solubility. It can then be included as data in a 70°C correlation. Solvents with low dif fusion coef ficients will appear to be less aggress ve than the y might become at longer exposure times or at higher concentrations. more aggressive from a thermodynamic point of vie w than larger ones. HSP correlations ha ve confirmed that this e fect is even more important when chemical resistance is being considered. or reducing mechanical properties more severely. it has been found that gi ven data sets are not al ways as reliable/consistent/coherent as could be desired for solubility parameter correlations. EFFECTS OF SOLVENT MOLECULAR SIZE It has been emphasized in Chapters 1. say . The correlation is then strictly v alid only .232 Hansen Solubility Parameters: A User’s Handbook studies of chemical resistance have been designed with the purpose of generating HSP correlations. This may tak e the form of une xpectedly dissolving. and the chemical reagents have not been chosen for the purpose of solubility parameter correlations. the data are often limited in number and scope. The products of reactions of well-defined o ganic bases with well-defined o ganic acids ha ve actually been studied systematically from a solubility parameter point of vie w. It can easily be seen. Attainment of equilibrium may not have been achieved. Smaller molecules are e xpected to be better. Also.

b ut no HSP correlation has been made. Data in chemical resistance tables are often of the type +. and this may also af fect the e valuations. b ut rather through a kinetic ef fect of diffusion rate. the solv ents with V less than the lo wer limit can be expected to attack if their RED numbers are less than 1. Various examples of HSP correlations of chemical resistance are included in the follo wing. +/–. However. permeation. They can also be neglected. The liquids which attack are clearly good solv ents for the material in question and will be located within the HSP spheres with RED numbers being successi vely lower for more severe attack. Some indication of the beh vior of the solvents with V larger than the upper limit is possible if their RED numbers are greater than 1. CHEMICAL RESISTANCE — EXAMPLES Chemical resistance means dif ferent things in dif ferent contexts. it will be repeated here that smaller and more linear molecules dif fuse more rapidly than larger and more bulky ones.0. Polymer samples used for solubility parameter or chemical resistance testing may contain retained solvent or monomer for many years. the ef fects of w ater can be e xtremely rapid as discussed in the follo wing. The data and correlations are about 20 years old. –. the second stage in the tw o-stage drying process in polymer film formation by sol ent evaporation can last for man y years. Epon® 1001 (Shell Chemical Corp. Unless otherwise specified the results are for room temperature. These would not be e xpected to attack under an y circumstances. 31 Such ef fects lead to comparisons where some systems may ha ve reached an equilibrium uptak e.32 Attempts to include the molecular v olume into a ne w composite solubility parameter and size parameter have not been particularly successful. the size and shape of solv ent molecules are v ery important for kinetic phenomena such as dif fusion. The treatment used in the individual correlations presented here is indicated in the follo wing. There are also situations where the attack is mild. The HSP data for the correlations discussed are included in Table 12.0.1. and whether or not satisfactory results are found with a product depends on its use. Likewise. questionable) either in the attacking (NR) group or in the nonattacking (R) group. These are for a tw o-component epoxy type and for a zinc silicate type.Applications — Chemical Resistance 233 for the size range specified. respectively. Lik ewise. The data in Table 12. all other things being equal. 33. Experimental data are always preferred over predicted behavior based on a correlation.31 See Chapter 16. not knowing which group to put them into. A HSP correlation of the solubility of a lo wer molecular weight epoxy. or satisf actory/questionable/unsatisfactory. discuss correlations of HSP and elaborate on dif fusion phenomena in more detail.1 include two older HSP correlations for chemical resistance for two types of tank coatings supplied by Hempel’ s Marine Paints. The correlations can include the solvents with mild attack (+/. However. and attainment of equilibrium. a good HSP correlation can be used to find ma y chemicals that will clearly attack or that will clearly not attack.). Chapters 13 and 16. Lar ger numbers of solv ents are needed in the study if this is to be done with an y benefit As stated abo ve. The numbers are not too dif ferent from those of the HSP correlation for chemical . Ho wever. Impro vements in chemical resistance are kno wn to ha ve been implemented in a ne wer epoxy tank coating. is included for comparison. recommended/not recommended (R/NR). whereas others ha ve not. 4. TANK COATINGS Chemical resistance is important for tank coatings used in the transport of b ulk chemicals. The diffusion coefficient may be so l w that equilibrium is not attained for hundreds of years at room temperature in common solv ent e xposures of rigid polymers lik e polyphenylene sulfide (PPS) with thicknesses of s veral millimeters.34 This may be because the size ef fects are not necessarily caused through the thermodynamic considerations on which the solubility parameters are based. or something similar . They are included here for purposes of demonstration.

700 0.0 9. HSP correlations of this type are included for PET .7 5.9 18.6 7.7 3.1 15.3 6.5 17.0 6.3 5. The data for these are all found in Reference 21.9 16.000 0.9 8.3 17.0 9.1 17.937 0.0 16.0 4.993 0. Although the data fit numbers were not recorded at the time.2 6.9 8.000 0.2 18.6 8.4 11.8 The fact of a successful HSP correlation for a completely inor ganic type of coating lik e zinc silicate is surprising. resistance for the tw o component epoxy tank coating.4 19.4 7.1 9. POMH.950 0.4 9.2 FIT —a —a —a 1.7 7.0 9.1 18.710 0.1 17.234 Hansen Solubility Parameters: A User’s Handbook TABLE 12.3 6.0 6.6 9. ACCEPTABLE OR NOT — PLASTICS Several examples of HSP correlations of data reported in the form acceptable-or -not are included in Table 12.2 Ro 7.7 2.0 7. a b Data not a vailable. glycols.865 0.8 8.3 3.000 1. This correlation is based on only 11 well-chosen data points b ut clearly sho ws that attack for man y chemicals can be e xpected.0 6. See text: NR — not resistant.7 4.0 5.0 6.1 18.931 0.9 6.967 0. modified PET coating on PET films to imp ve their weldability.991 1.2 6. is the attack of the amorphous.1 23.8 4. with the understanding that G solv ents are within the HSP correlation spheres and are not recommended for use.609 0. POMC.999 G/T —a —a —a 7/11 7/34 16/26 2/28 11/28 9/31 2/102 18/92 49/91 48/74 30/48 9/16 3/20 4/20 9/20 8/45 6/25 5/41 Note: The symbols G for good solv ents and T for total solv ents are maintained.7 7. R — resistant.1 19.7 3.3 4.3 5.7 4.9 7.5 7.8 18.6 9.8 δH 10.0 9.800 0. or lack of the same. th two chemical resistance correlations reported here were clearly considered reliable.1. Other data sources for these are also a vailable.0 18.1 18.6 δP 9.3 9.7 17.7 8.4 18. These three correlations ha ve been reported earlier. and .5 10.9 10. Among those chemicals not attacking are h ydrocarbons.7 5.3 18.0 17.9 6.9 7.7 4. and glycol ethers and higher alcohols which ha ve a reasonably high h ydrogen bonding character. PUR.5 7.4 17.5 8. PET FILM COATING Another e xample of chemical resistance.981 0.3 11.955 0. This is still another demonstration of the uni versality of the applications possible with the HSP concept.6 4.952 0.6 17.1 5.1 Hansen Solubility Parameter Correlations for Selected Materials Material Epoxy tank coat (tw o component) Epoxy solubility (Epon 1001) Zinc silicate coating PET-amorphous coating PET-CR (+/– R b) PUR-CR (+/– R b) POMC/POMH (+/ Rb) (+/– R b) POMC (+/– NR b) PA6/PA66 (+/– NR b) Halar 300 ECTFE 23°C Halar 300 ECTFE 50°C Halar 300 ECTFE 100°C Halar 300 ECTFE 120/149°C Neoprene-CR (+/– R b) PPS tensile strength <60% 93°C PEI ULTEM 1000 600 psi PEI ULTEM 1000 1200 psi PEI ULTEM 1000 2500 psi PEI ULTEM 1000 solubility PES mechanical properties PES solubility δD 18.2 6.5 11.

Additional solvents may be required to make an improved correlation based on the improved coverage possible. which is readily attack ed by far more solvents. The first three correlations of this type consider reporte evaluations of minor attack which will require further e valuation (+/–) as if these systems were suitable for use (resistant). It is for this reason that more e xtensive tables of HSP correlations of chemical resistance are not reported here.” as there appears to be symmetry about a central point. etc.. such as mustard. The HSP data for these correlations are also included in Table 12. a correlation of chemical resistance data for Neoprene ® rubber (Du Pont)22 is included. such as a cross-linking segment. In spite of these pitf alls. The compositions are also different. This approach has been used to assign HSP to some liquids when calculations were uncertain. some reasons have been given earlier and others are in the follo wing. The most se vere attack or swelling may occur in one region or another of the polymer or maybe e ven on a third component. A systematic analysis of wh y this is found will not be attempted for reasons of space. Viton® is discussed in Chapter 5 and Chapter 13 in more detail. b ut their usefulness becomes clearer with some e xperience. the data fit sh ws that there are man y outliers. This is to demonstrate that dif ferences are found. as e xpected.Applications — Chemical Resistance 235 PA6/PA66 using data from Reference 21. When the solv ents demonstrating minor attack are considered as being in the attacking group. b ut w orking in the . and the probable effects of solv ent molecular size may become apparent. There is also a question as to which segments of the polymer may be subject to attack by which solv ents. The data on which these correlations are based are of the recommended-or -not type and were found in Reference 26. are considered as being in the nonattacking group for this correlation. This could perhaps be done by composition in some cases. The data fits are not particularly good at the higher temperatures To complete this section. certain detergents and oils. The last two correlations consider this condition as not suitable for use (not resistant). Ho wever. depending on ho w the data are considered and that outliers are often found when correlating this type of data. The effects of temperature on the chemical resistance of poly(eth ylene co-chlorotrifluoroeth ylene) (ECTFE) Halar ® 300 can be seen in Figure 12. A suitable goal for a future project is to determine the ef fective HSP for various media frequently encountered in resistance lists. Too much space is required to try to e xplain wh y gi ven results are outliers. Solv ents in the intermediate cate gory. Previously it w as indicated wh y HSP correlations of this type lead most often to guidance rather than to a firm recommendation. In other cases. e ven if all the outliers could be e xplained in one w ay or another. Viton® is an e xample of this. as more solv ents then become more severe in their attack. The HSP found for a polymer in this type of correlation may not be representati ve of that polymer in all aspects of its beha vior.1. The POM correlations ha ve typical problems in that the correlation considering all minor attack as ne gligible is based on only tw o severely attacking solvents among the 28 solv ents tested. This figure ha affectionately been dubbed a “b ullseye. There are man y pitfalls to be a ware of both in generating such correlations as well as in using them. one could see whether beha vior paralleled that of a kno wn chemical. although this is not strictly true as the HSP data confirm. whether or not equilibrium has been attained can be inferred. A computer program was de veloped similar to the SPHERE program as described in Chapter 1.e. The HSP v alues for PET based on chemical resistance are some what different from those of the amorphous PET coating. Outliers can be reconsidered.1. There is a question as to co verage over the whole range of solvent HSP possible by the test solvents. it is strongly suggested that those generating this type of resistance data should try HSP correlations to evaluate the consistency of the data before reporting it. that of a questionable-for -use recommendation. The radius of the chemical resistance spheres increases with increasing temperature. A third method is to determine these parameters by recognizing a similarity to all materials attack ed and a dif ference from those not attack ed. Block copolymers may demonstrate two separate (o verlapping) correlations that cannot be reasonably force fitted into a single HS correlation. i.

9 R 2.5 FIT 0.7 7.9 7. Liquids within the spheres (circles) are not recommended at the gi ven temperatures.3 P 8.1 18. . δH TEMP.9 8.1 Chemical resistance of Halar® 300 ECTFE at various temperatures.993 0.236 Hansen Solubility Parameters: A User’s Handbook SOLVENT RESISTANCE OF HALAR 300 AT 4 DIFFERENT TEMPERATURES 14 13 12 11 120° = 149° HANSEN POLAR PARAMETER.1.8 8. 23° 50° 100° 120/149° D 16.5 7. HSP data gi ven in Table 12.1 18.7 5.4 7.8 18. δH 10 100° 9 8 7 6 5 4 3 2 1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 120°✚ 23° ● 100°▲ 50° ✳ 23° 50° HANSEN HYDROGEN BONDING PARAMETER.7 H 7.5 7.71 0.2 6.70 0.80 NO(NR) 102 (2) 92 (18) 91 (49) 74 (48) FIGURE 12.

The solvent parameters were then systematically v aried by the program to reduce the collecti ve error. Glass Fibre Reinforced Polyether Sulphone (PES). (From FORCE Institute. not all the predictions based on the HSP thus assigned to the solv ent agreed with the e xperimental data.2. 1994. the data fits for this procedur were comparable to those found for polymer solubility using the SPHERE program. The solubility of a number of polymers was evaluated in the solvent.24 It is clear that the chemical resistance is dependent on the stress level. which lasted 336 h. Additional HSP tensile strength correlations have been generated for polyetherimide. that is. More extensive correlations for PPS are found in Reference 31.) . but the errors were small. The solvents considered as being those which attack led to cracking during the study . 23 A HSP correlation of these data using the “good” solv ents as those which reduce tensile strength under these conditions to less than 60% of the initial v alue is found in Table 12.Applications — Chemical Resistance 237 opposite manner. With permission. using data reported by General Electric. Solv ent Resistance of Polymer Composites. Center for Polymer Composites. PPS δD δP δH 5 FIGURE 12..2 HSP correlation of the tensile strength reduction of Ryton ® PPS. In other w ords. Within the sphere are liquids which reduce the tensile strength to less than 60% of the original v alue after e xposure for 1 year at 93°C. One of the more direct techniques to measure such effects is to determine the tensile strength. Higher stress levels lead to more severe attack by a larger number of chemicals. Some led to earlier cracking than others. 31. TENSILE STRENGTH The long-term exposure of polymers or polymer composites to solv ents normally leads to changes in mechanical properties.A more rapid attack is expected from the better solvents with the smallest size and shape. to locate the best possible set of HSP for the solv ent. 1st ed. The correlations all ha ve high data fits. The tensile strength reduction for glass fiber reinforced polyphe ylene sulfid (PPS) after e xposure to a number of solv ents at 93°C for 12 months has been reported.1 and Figure 12. which could be treated in a separate correlation. ULTEM® 1000. but this has not been done. In general.

238 Hansen Solubility Parameters: A User’s Handbook The next entry in Table 12.1. The HSP correlation for the recommendation from the supplier 25 for glass fiber reinforced PES (Ultra son® E. whereas the c ycled samples suddenly be gin to gain weight well beyond the equilibrium value. it is suggested that data for w ater used as a test solv ent not be included in HSP correlations.1. Figure 8. in particular . especially if there are hydrophilic sites in the polymer or at an interface to which water will rapidly diffuse. This special type of failure has been discussed in more detail elsewhere32 (see also Chapter 1 and Chapter 8).3 and for PPS in Figure 12. b ut it also means diffusion rates of other species will be increased. One follows weight g ain by rapidly weighing samples that are surf ace dry. with h ydrophilic materials. It may tak e several days of exposure to fresh air (to reduce the amount of methanol to acceptable le vels) before subsequent direct contact with w ater or sea water can be tolerated. If a coated tank has been used to store methanol. This correlation is not directly comparable with the pre vious ones for ULTEM 1000 as the number and range of solubility parameters included in the test solv ents are different. Control samples that are not c ycled reach equilibrium and stay there. The previous study24 did not include a sufficient numbe of solvents having widely different HSP to gi ve a true total picture of the UL TEM 1000.3). the presence of w ater in a fil can influence the upta e of other materials. the w ater diffusing into the coating will associate with the methanol. There is a dif ference. polyesters. BASF) can be compared with the HSP correlation for simple solubility of the polymer in standard test liquids. This is for polyethersulfone (PES) based on mechanical properties after e xposure to v arious liquids. but it is not lar ge. Water uptak e in most polymers increases with increased temperature. This is because the solubility parameters of the water and polymer are closer at higher temperatures. These data were generated in a standard solubility parameter study . The phase separated w ater has been called SWEA T (soluble w ater e xuded at lowered temperatures). alkyds. This has been discussed in detail in Reference 35 for PPS and PES. Water is an e xceptionally good plasticizer because of its small molecular size. The extra weight is phase-separated water within the samples. Therefore. This mechanism can be confirmed xperimentally by cycling samples continually e xposed to water between two relevant temperatures using a quench from the higher one to the lo wer one.4. polyethersulfone (PES). The increasing . These droplets can quickly collect into blisters. A final HSP correlation of the suitable-fo -use or not type is presented in Table 12. and perhaps hot methanol in particular. The author has observ ed this phenomenon as a mechanism of f ailure for epoxies. Typical results for the SWEAT phenomena for EPDM are seen in Figure 12. SPECIAL EFFECTS WITH WATER As stated else where in this book (Chapter 1 and Chapter 15. If there is too much methanol retained in the coating. The parameters for the polymer found in this correlation are those e xpected to reflec its (thermodynamic) affinities most closel . Water can be a v ery aggressi ve chemical. The presence of water not only softens (reduces the glass transition temperature) a polymer as such. This has been discussed in Chapter 1 and Chapter 8. Some of the water already in the polymer is now in e xcess and suddenly appears as small clusters or droplets of freed liquid w ater within the polymer itself (see Chapter 8. A related problem can be encountered in chemical resistant coatings for tanks that ha ve been in contact with methanol. but it is repeated here with e xamples for those interested in chemical resistance. F or this reason.1 is for true solubility of UL TEM 1000. in particular). and ven EPDM rubber. being more prone to enter the film The increase of w ater uptak e with increased temperature can cause special problems with blistering if the temperature of a w ater-saturated polymer falls rapidly to a lower temperature. the coating is more than lik ely saturated with methanol. The previously soluble water can no longer be truly dissolved. the seemingly unpredictable behavior of water has often led to its being an outlier in HSP correlations. polyphenylene sulfide (PPS). The very high δH parameter for w ater decreases more rapidly with increasing temperature than the δH parameter for most polymers. also gi ven in Table 12.

HSP correlations including systematic consideration of the solvent molar volume (or other suitable size parameter[s]) should be an inherent part of all future studies of chemical resistance. Blister form and total delamination can occur . Water in e xcess of the equilibrium v alue at longer cycling times is SWEAT. de gree of surf ace attack. and if it is included as an outlier .50 3.00 0.3 A rapid quench to a lo wer temperature can free w ater already dissolv ed in a polymer in the form of SWEAT.50 4. cracking. It is strongly suggested that data for water not be included in these correlations as well.00 2. Examples include HSP correlations of true solubility and swelling. and correlations for simple e valuations of chemical resistance of the suitable-or-not type. √DAYS 5 6 7 FIGURE 12. Data for true acidic or basic chemical attack must not be included in HSP correlations.00 WATER UPTAKE. water content in the mixture of methanol and w ater will ultimately cause the solubility parameters of the mixture to be suf ficiently high so that it becomes incompatible with the coating. These correlations aid in the determination of whether equilibrium has been attained. .50 1. W/W% 3.50 0. CYCLING 120°-15° EPDM-GASKET 5. the molecular size of the liquids v ealuated will affect the result. this f act will force a correlation with less predictive ability than had it been ne glected.00 4.00 1. tensile strength reduction. The data in the figure are weigh gain for EPDM with c ycling in w ater between 120 and 15°C. These blisters are often near the substrate. as HSP correlations reflect p ysical attack and not chemical attack. CONCLUSION HSP can correlate dif ferences in ph ysical beha vior observ ed in chemical resistance testing of polymers and polymer containing systems when a suf ficiently la ge number of dif ferent organic solvents are included in a study. SWEAT can lead to blistering. It is reemphasized that. as well as pro vide insight into the beha vior expected from untested solv ents whose HSP are stored in a solv ent database or can be calculated. and delamination.50 2. as this is where the retained methanol will be found at highest concentrations. and this should be considered in some w ay.Applications — Chemical Resistance 239 WATER ABSORPTION. in each case.00 0 1 2 3 4 EXPOSURE TIME. Its behavior is too unpredictable compared with other test liquids.

SWEAT can lead to blistering. plate thickness 1 and 2 mm 2 90 C. Ne w York... and resins. inPaint Testing Manual.4 A rapid quench to a lo wer temperature can free w ater already dissolv ed in a polymer in the form of SWEAT. 8(1). Hansen. 1 mm ✛ ▲ ▼ ▼ ▼ ▼ 140 C. J. Vol. ▲ ▲ -. Hansen..M.240 Hansen Solubility Parameters: A User’s Handbook WATER ABSORPTION VS. Hansen... emulsifiers. Independent calculation of the parameter component .. The data in the figure are weigh gain for PPS with cycling in water between 90 and 23°C. Manual 17. Solubility parameters. 18–22. 4.M. 1967.. C. Hansen. 1971. The Resistance of PPS. 1967. C.. 3.5 0 -1 0 10 20 30 40 50 60 70 SQUARE ROOT OF TIME / THICKNESS SQRT (HRS)/mm 80 FIGURE 12. 1992..M. Hansen. and Skaarup. American Society for Testing and Materials. C. C. 1 mm ■ WEIGHT GAIN (%) 1 90 C. 1 mm 1. cracking.. C... Res. Suppl. 1994. Copenhagen. Philadelphia.. 2nd ed. Paint Technol. The three dimensional solubility parameter — key to paint component affinities III. 383–404.5 ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▼ ▼ ✕ ▲ ▲ ▲ ▼ ▲ ▲ ▲ ✕ ✛ ▼ ■ ✛ ▼ ▼ ▼ ▼ ▼ ▼ ▼ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ■ ▲ ▼ ▲ ▲ ▲ ▲ ▲ ▲▲ ▲ ■ ▲ ▲ ▲ ▲ ▲ ▲ ■ ■ ■ ▲ ▲ 50 C. 505–510. J. 39(511). 1995. in Kirk-Othmer Encyclopedia of Chemical Technology..M. 2.M.M.V. C.) REFERENCES 1. polymers. K oleske.5 ■ 90 C. 39(505). 39(511). 5. PES and P A Polymer Composites to Temperature Cycling During Water Exposure. The three dimensional solubility parameter — k ey to paint component af finities II Dyes.M.. K. Paint Technol.. 1967.. A. 1969. 7. C. in Danish). and delamination. 6. 2–11. (From Hansen. Danish Technological Institute. and pigments . Ind. Taastrup. Ed. Danish Technical Press. The three dimensional solubility parameter — k ey to paint component af finities I Solvents. pp. plasticizers. 2 mm ▼▼ . Hansen. 1 mm ✕ 23 C.TEMPERATURE RYTON R4 XT-HV. J. and Beerbo wer. 11. J. Dan. P A. Doctoral dissertation. The universality of the solubility parameter . Standen.M. 889–910. Paint Technol. 8. With permission. pp. Dev. mutual solubility and compatibilit . Interscience.. Eng. 511–514.. 1967. Solubility parameters. 104–117. Their Importance in Surf ace Coating F ormulation. Ed. Hansen. Center for Polymer Composites (Denmark). Chem. C. Hansen. C. 73(8). A. Kemi. 2 mm ▲ ▲ ■ ▲ ▲ ■ ✕ ✕ ✛ ✕ ✛ ✕ ✛ ✛ ■ ■ . Water in excess of the equilibrium value is SWEAT. 25 years with solubility parameters (25 År med Opløselighedsparametrene. Prod. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient.M.

.. Hansen. C. 57–60.. On the use of cohesion parameters to characterize surf aces. Chemical Resistance of Plastics and Elastomers used in Pipeline Construction. Hansen.M . 7.. 1989. 15. Solvent Resistance of Polymer Composites — Glass Fibre Reinforced Polyphen ylene Sulfide. Mod. 24. Plastics Design Library.D. Res. 17. Surf ace wetting and the prediction of en vironmental stress cracking (ESC) in polymers.. Du Pont Compan y.M. P.W .M..M. 22.M. in German). 30... Philip Petroleum Co. Brochure: Co-Act — A Dynamic Program for Solv ent Selection. 1983. 1993. but this does not appear on the b ulletin]. The Resistance of PPS. Degradation Stability. 126–132. 33. and Reddy . Surface dewetting and coatings performance.. 1992. August 25–30. Sha wbury. 1992. 1971. A. 58(733). Hansen. FL. 42(550). Anonymous. 23. Plastguide.. Polymer Products Department. Speyer am Rhein. October 1991. and Pierce. 660–664. 16. Handbook of Solubility Parameters and Other Cohesion Parameters. J. Slidell. FL. 13. C. 1991. Gaylord Chemical Corp. Verhalten v on Ultrason ® ge gen Chemikalien — B ASF Technische Information TI-KTE/TH-01 d 82132. P aper presented at the F ourth Chemical Congress of America. Anonymous. 2005. Taastrup.S. Org. Barton. CRC Press. Hansen. Colloid Interface Sci. 113–120. Bulletin No. 102. Adhes. 1980. Taastrup. Characterization of surfaces by spreading liquids. 79(9). J. Prog. Thermoplastic Plastics (in German).. Inc. Anonymous. Inc. Fluid Resistance of Viton®. E. J. Using Dimethyl Sulfoxide (DMSO) in Industrial F ormulations. DE. M. 89. Anonymous. 1995. Surface effects in coatings processes. J. A. Characterization of pigments surfaces in terms of three-dimensional solubility parameters of solvents. pp. 1989. Thermoplastische Kunststoffe.. J. 1974.M. C. 32. 1990. Shareef.. Expanded List — Chemical Resistance of Halar ® Fluoropolymer. Anonymous. A.. Shrewsbury. Danish Technological Institute. T. Shropshire. C. and Hansen.M. Hansen.. 26. in fact. Dev. Paint Technol. 1983..J. U. Ind.. 1970.. Some aspects of acid/base interactions (Einige Aspekte der Säure/Base-W echselwirkung. Yaseen.Applications — Chemical Resistance 241 9. Anonymous. NY. B. Geor ge Fischer +GF+. Norwich. Denmark. Ausimont.. Danish Technological Institute. 21.. Van Dyk. RYTON® PPS Polyphen ylene Sulfide Resins — Corrosion Resistance Guide. Anonymous.. February 1986. 19. Surface free energy: a new relationship to bulk energies. Anonymous. Anonymous.E. Handbook of Polymer-Liquid Interaction Parameters and Solubility Parameters. J.. 11. Chemical resistance data. LA. Carlowitz. 218–225. 31. 13(4). Ne w York. 1989. Hansen. 29. C. 15(3/4). 1–62. Pittsfield. Farbe und Lack. C.A.F. Program calculates solv ent properties and solubility parameters. Coat. Ne w developments in corrosion and blister formation in coatings.. Bittar . PES and P A Polymer Composites to Temperature Cycling During Water Exposure. Coat. Prod. . 1994. 595–598. C. Van Dyk. J..B. Centre for Polymer Composites (Denmark). SCS Dukadan AS. J. 10.F.M. 35. Ultem ® Resin Design Guide.. Suspension interaction of pigments in solv ents.. Centre for Polymer Composites (Denmark). 1990. 20. 35–44. Technol. 18. 1977. Beerbower. Eng. 26... A. Zechner und Hüthig. Hansen.M.J. 14. GE Plastics. 275–286. Elastomers Division. Barton.. 1993. new ed.. Paint Technol. Dante. C. William Andrew. and Caillault. Boca Raton. and Wallström. 44(570). Chemical Resistance Data Sheets Volume 1 Plastics.. Polym.. 513–516. MA. 46–51. 1972. USA. Chem..W. Boca Raton. Wilmington. M. Anonymous [Note: This was. 34. K. Nielsen. C. 28. Randers. Rapra Technology Limited.. 1989 25. O. 27. 12.F.M. CRC Press.M. Volume 2 Rubbers. Hansen. Exxon Chemical International Inc. 35. Paint Coat.M.


e ven though they may have different solubility parameters (b ut not so dif ferent that both are able to dissolv e in the polymer at the le vel of comparison). S: P = DS. and.2–4 The polymers in these studies were a copolymer of 87% vin yl chloride and 13% vinyl acetate.1 This difference in diffusion coefficients is a result of di ferences in molecular size. it has been found that the rate of dif fusion at the same concentration is about the same for dif ferent solvents with approximately the same size and shape. The diffusion coefficient indicates h w fast the permeant molecules can mo ve through the polymer. S will be lo wer when the HSP of the barrier fil and a solv ent are v ery different. and barrier polymer swelling. As the amount of permeant being dissolv ed increases with closer matches of the HSP for permeant and barrier polymer . Likewise. The diffusion coefficient for oxygen in polyvi yl chloride (PVC) is well over a million times greater than that ofn-hexane (at low concentrations) in the same polymer. of a liquid or a g as through a polymer is gi ven by the product of the diffusion coefficient. This ef fect can be v ery significant. its plasticizing ef fect on the polymer becomes significant. Absorption and dif fusion in polymers is treated extensively in Chapter 16. At lo w permeant concentrations D is a constant. Larger molecular size and more comple x and bulky molecular shape are major f actors that lead to lower diffusion coefficients. the permeation coef ficient. and the solubility coef ficient.— Barrier 13 Applications Polymers Charles M. The solubility coefficient indicates h w much of the permeant can be dissolv ed in the polymer. are functions of the HSP match. Hansen ABSTRACT The permeation coef ficient. S correlates with the Hansen solubility parameters (HSP). permeation rates through barrier polymers. INTRODUCTION The permeation of a liquid or a g as through a polymer can be described by the relation P = DS (13. HSP correlations are gi ven for breakthrough times in chemical protecti ve clothing. assuming the same diffusion coefficient mass transport will be proportionately higher . the concentration gradient is correspondingly higher. and therefore P. D. an the dif fusion coef ficient increases mar edly. When solubility is higher . and the solubility coefficient. S. A significant actor af fecting D is the molecular size and shape of the permeant molecules. Both liquids and g ases are treated. the end result is that both S and D. D.1) P. as the permeant concentration increases. P. and polymeth yl methacrylate. is the product of the di fusion coefficient. The successful correlations of permeation phenomena with HSP are thought to be lar gely a result of this e xceptional dependence of D on the dissolved permeant. polyvin yl acetate. 243 . The amount dissolv ed in the polymer determines the concentration gradient o ver a film and the concentration gradient is the dri ving force for mass transport. Ho wever.

5 See Chapter 16.1 Diffusion coefficients in polyvi yl acetate at 25°C for methanol (A). S also Chapter 16. This difference corresponds to the speed of a snail in the w oods compared with a modern jet airliner . (From Hansen. 1998. Original data are in Reference 3. 3–7 This rule of thumb assumes that the polymer beha ves as a rigid polymer over the concentration range being considered. As this behavior is general for solvents in polymers. a rule of thumb indicates that the local dif fusion coef ficient for sol ents in rigid polymers can increase by a f actor of about one million when about 20 v ol% solvent is present compared with the solvent-free state. 7. Water. It can be seen in this figure that for moderate sol ent concentrations in this rigid polymer. chlorobenzene (C). The solvent diffusion coefficient data in Figure 13. The diffusion of the smaller species (and other species) becomes f aster as their local concentration and plasticizing ef fect become greater.M.) . 7 A CM2/SEC 8 B 9 10 11 12 13 14 15 C D -LOG DIFFUSION COEFFICIENT 0 8 16 24 VOLUME FRACTION PENETRANT 32 FIGURE 13. Pharmaceutical and Medical Packaging 98. eth ylene glycol monomethyl ether (B)..1 were first presented in Reference 3. The diffusion coefficient for ater shown in the figure as found by absorption and desorption e xperiments in thin films where a correction for the sur ace resistance w as also required.12 With permission. for e xample. and c yclohexanone (D). This e xtra free v olume allo ws easier polymer se gmental motion.244 Hansen Solubility Parameters: A User’s Handbook CONCENTRATION-DEPENDENT DIFFUSION Low molecular weight liquids are plasticizers for polymers if they can be dissolved in them. This figure sh ws diffusion coefficients for s veral solvents in polyvin yl acetate (PVAc) at 25°C. the local dif fusion coefficient increases by a actor of about 10 for an increase in solvent concentration of about 3 to 4 v ol%. Permeability of polymers. C. The low molecular weight materials can greatly reduce the glass transition temperature of their mixtures with a polymer as the y have considerably more free v olume associated with them than the polymers themselv es. can significantly soften ma y polymers e ven though it is dissolv ed to only a fe w percent. The data point for w ater (*) is included for comparison.

Do. The diffusion coefficients in P Ac for meth ylene chloride. In addition to demonstrating concentration dependence. which does not dissolv e this polymer . an alternate strate gy. is as follo ws. 8 It is also discussed here because it is a major factor in the success of HSP correlations of permeation phenomena. 13 wt% vinyl acetate. w as introduced by placing a completely dry polymer film in closed container o ver toluene vapors. leads to concentration-dependent dif fusion coefficients because the amount ta en up at the contact surf ace on liquid contact is v ery often more than 15%. b ut much higher than those for EGMME. such as a laminated system. Water has one “significant” atom. even though the number of significant atoms is the same. unless the rigidity is such as to allo w decided holes of suitable size to enable quite rapid dif fusion of much smaller molecules. for example.1 also sho w the well-established relations that those solv ents with lar ger and more complicated chemical structures are those with lo wer diffusion coefficients. in practice. such as meth ylene chloride. the significance of which will increase with increasin amounts of the dissolv ed materials. those with higher glass transition temperatures. Here. An example of how to estimate diffusion coefficients in P Ac for other liquids. the rule o thumb is that the dif fusion coefficient increases by a actor of about 10 for an increase in solv ent concentration of about 15 vol%. i. If. Another type of comparison which is possible is to state that the dif fusion coefficients for toluene are xpected to be close to those for chlorobenzene because of a similarity in molecular size and shape.11 The definition of a “good” sol ent which was used for these correlations w as that the breakthrough time w as less than some selected .7 This shows that liquid contact with chemical protecti ve clothing. A list of HSP for barrier polymers used in chemical protecti ve clothing has been published 12 based on data by Forsberg and Olsson. The Crank-Nicholsen finite di ference treatment for concentration-dependent dif fusion8 was extended by Hansen3 and used to describe film formation by sol ent evaporation. this leads to water sensitivity. particularly with respect to transport of w ater in b uilding materials.5–7 Klopfer 9 de veloped analytical solutions in volving concentration-dependent dif fusion for man y situations found in practical b uilding applications. Union Carbide).000 times lar ger than that for the latter. Planar chlorobenzene dif fuses more rapidly than nonplanar cyclohexanone.4 to explore what is termed anomalous diffusion. These considerations lead to the best combination of properties for a barrier polymer as being one with a high glass transition temperature and with HSP far removed from those of the permeant.. may be required.5 to develop an easy method to e valuate data leading to concentration-dependent diffusion coefficients 6 and to account for the effects of concentration-dependent diffusion and surface boundary resistance simultaneously .11 Such data can also be used to establish correlations with HSP . the diffusion coefficient data for P Ac in Figure 13. This was confirme by solvent retention studies where toluene and chlorobenzene were retained in identical amounts in a film of VYHH® (87 wt% vinyl chloride. Concentration-dependent diffusion can be handled properly without great dif ficulty for most situations of practical interest Neglect of this ef fect can lead to errors.10. with three significant atoms.10 Some of these correlations have been improved in most instances by correlating the more e xtensive data of F orsberg and Keith. Concentration-dependent diffusion has been discussed at length by Crank. and eth ylene glycol monometh yl ether (EGMME) has fi e. Toluene. is 10.e. The diffusion coefficient for ater in PVAc at lo w concentration. Diffusion can be e xpected to be slo wer in more rigid polymers. can be xpected to be some what lower than those for methanol. SOLUBILITY PARAMETER CORRELATIONS BASED ON PERMEATION PHENOMENA SOLUBILITY PARAMETER CORRELATIONS OF BREAKTHROUGH TIMES Extensive permeation studies and collections of permeation data are a vailable within the chemical protective clothing industry. methanol has tw o.Applications — Barrier Polymers 245 Concentration-dependent diffusion coefficients are also found for elastomers.

8 6.3 7. With regard to the protecti ve ability of the dif ferent garments.2 7.1 for the breakthrough times of less than 1 h for neoprene rubber used in chemical protecti ve clothing. This includes butyl rubber. 66 50 86 77 58 116 32 Neoprene® 16. This size effect is in agreement with what has been known about solvent retention in coatings 2–4.5 9. and al yl alcohol with shorter breakthrough times than predicted and the phthalate plasticizers which ha ve longer breakthrough times than predicted.3/177.1 13. NH).0 1.0 Viton® 15.1 give polymer HSP in the range of those e xpected from their composition.12.9 84.1 13.907 0. Units are MPa1/2 Type δD δP 8.0/175.15 Evans and Hardy.0/172 71. means that the dif fusion coefficients at the reasonably l w concentrations e xpected in better barrier polymers do not v ary too greatly from each other. Outliers were usually these larger molecular species and permeants with smaller or more linear structure.574 1.0 2.6 2.0 Neoprene 19.0 0.2 None None FIT 0. more complicated structures required much longer times for breakthrough for a gi ven protective membrane type than comparison with other solv ents would indicate.16 and by Nielsen and Hansen. satisfactory correlations could only be found when the differences in HSP between the permeant and the barrier polymer were combined with a size (and shape) parameter(s). the solv ents with lar ger.925 0. and polyeth ylene (PE). This.1 gives a very poor data fit (0. 14. An e xcellent e xample of this type of impro ved correlation is included in Table 13.2 3.0 Butyl 17. The first correlation for this material listed in Table 13. carbon disulfide. The molecular v olume. Merrimack. Challenge ® 5100 (Chemical F abrics Corporation.969 No. Viton® (The Du Pont Compan y).6 PE 16.8 8.1 HSP Correlations of Breakthrough Times for Barrier Polymers Typically Used in Chemical Protective Clothing. There were 46 out of 66 liquids which had breakthrough times shorter than 1 h.6 Nitrile 19. it w as found that.1 includes some of these impro ved HSP correlations based on a 1-h breakthrough time for commercial film thickness HSP alone cannot always correlate barrier properties unless comparisons are limited to solv ent molecules with approximately the same size (and shape).000 0.1 Ro 10.13 and what has been discussed pre viously. This eliminates these “outliers. Printing the correlation data arranged in increasing order of permeant V clearly sho wed whether the molecular size w as important.896 0. 17 . where dif fusion is much more rapid than expected in average comparisons. The HSP correlations for 1-h breakthrough times for other barrier polymers discussed in Table 13. 1 h. of course. in general.5 Note: “Good” solvents in these correlations have breakthrough times of less than 1 h.3 7. w as found to be a reasonably successful single parameter for this purpose. and as e xpected.7 δH 4.3 7.9 V Limits None 71.6/148.” This correlation is based on 39 liquids with breakthrough times of less than 1 h out a total of 50.246 Hansen Solubility Parameters: A User’s Handbook TABLE 13.0 0. V. It is clear from closer analysis of the details of the correlation that the outliers are methanol.8 2. HSP correlations of the swelling of Viton are discussed here as well as in Chapter 5 and by Zellers.0 7. The thicknesses of all of the films discusse here are those commonly used in chemical protecti ve clothing. either 20 min.574). A perfect fit is found when the molecular olume range of the permeants included in the correlation is abbreviated to between 71 and 172 cc/mol.8 72. In many cases.6 13. or 4 h.902 0. value. Table 13. nitrile rubber .8 Challenge 5100 ® 16.

The effect with the terminal double bonds clearly indicates the preferential direction of motion for this type of molecule. Molecules with still lo wer molar v olumes appear to come through with only a slight dependence on the RED number . Cop yright ASTM.0 FIGURE 13. (Reprinted from Hansen. the dif ference in HSP between a solv ent and a polymer . This is given in HSP correlations using Chapter 1. American Society for Testing and Materials. . The data are plotted using permeant molar volume vs. The number. 1992. Molecules with molar v olumes greater than about 75 cc/mol require a terminal double bond and lo wer RED numbers to breakthrough under these conditions. RED numbers increase as the solv ent quality decreases. “good” permeants ha ve breakthrough times of less than 3 h. the solv ent quality is considered “bad.M.. 5. 906.Applications — Barrier Polymers 247 150 D. Eds.6. is a fluoropolymer supported by fib abbreviations for the permeants in Figure 13. F or RED numbers greater than 1. This correlation shows that molecules with molar volumes greater than about 100 cc/mol will not have breakthrough times of less than 3 h. With permission.The Performance of Protective Clothing: Fourth Volume.0 1.2 are e xplained in Table 13.W.0. ASTM STP 1133. The radius of the sphere is actually determined as the dif ference in HSP of the “w orst” good solvent(s) and the HSP for the polymer (which is the center point of the sphere). N..) The RED number (Chapters 1 and 2) is a k ey parameter to judge solv ent quality. R = 16. Ro. In this correlation. Figure 13.0 RED NUMBER 4.0 5. 3.18 The barrier material. Challenge 5100.2 sho ws one w ay to graphically use RED numbers to present data from HSP correlations of permeation phenomena. The data are plotted using V vs. Philadelphia.e.10 as Ra (Chapter 1.2. divided by the radius of the correlating sphere. J.997 FOR 68 < MV < 98 ■ BTC ● BCN BREAKTHROUGH TIME “=” “NO=” < 3 HR > 3 HR Evaluation uncertain ■ ■ x ● ● NONE MOLAR VOLUME ● MIK BCL 100 MSO ■ ● STY ● CHA ● EAC ■ TCE EET MMA ■ ●● NTB ● TCR ●CHK DEN●● CRB ● ETA● TTE ■ ATC DOUB– EVE ■ ● ACA MAM BNZ ■ MEK ANL● LE CRPVAM ■TCL■ ● ● BTR ■ ● ● DOX MAN ■ THF● PRA BOND ■ ALC ● FFA EPC ● PYR ■ MVK ■ CRF●● VOC ■ EDC● ● ACI ALN ● DMF ACB● ● ■ ALM ● ACE POX●■ ACN ■ DCM● ARL MIC● CBB● ● AAC ● AAD ● MBR ALL ■ ALA ATN ● ● NME 50 0. Equation 1. P.P. HSP correlation for breakthough times of less than 3 h in Challenge 5100.. solv ent–polymer af finit .. The molecules in effect worm their way through the barrier polymer .2 Graphical method to present HSP correlations. RED number. C. H.8 FIT = 0. re gardless of RED number. and Henry . 3. Symbols used are explained in Table 13. et al.9). Equation 1. ” although swelling may still occur .0 2. the RED glass. i.2. RED numbers near zero indicate v ery good solv ents (rapid breakthrough). McBriarty.

et al.-Tetrachloroethylene Trichloroethylene 1.1. American Society for Testing and Materials. SOLUBILITY PARAMETER CORRELATION OF PERMEATION RATES Permeation rates for dif ferent permeants in a polymer can also be correlated to find HSP for th polymer. C.1-Trichloroethane Tetrahydrofuran Toluene Tetrachloroethylene Vinyl acetate 1.2 List of Symbols Used in Figure 13.W. 1992. Eds. With permission. Cop yright ASTM.1..2. 903. N.2 Symbol AAC AAD ACA ACB ACE ACI ACN ALA ALC ALM ALN ANL ARL ATC ATN BCM BCN BNZ BTC BTR BUT BZN CAC CBB CBT CCF CHA CHK CLA CLB CRB CRF CRP DCM DEN DMF DOX DSO EAC Compound Acetic acid Acetaldehyde Acetic anhydride Acetyl bromide Acetyl chloride Acetone Acrylonitrile Allyl alcohol Allyl chloride Allyl amine Allyl cyanide Aniline Acrolein Allyl isothiocyanate Acetonitrile Bromochloromethane Butyl acetate Benzene Butyl acrylate Butyraldehyde Butane Benzonitrile Chloroacetylchloride Carbon disulfid Carbon tetrachloride Dichloromonofluoromethane (Freon 21 Cyclohexylamine Cyclohexanone Chloroacetone 1-Chlorobutane Chlorobenzene Chloroform Chloroprene Dichloromethane Diethylamine Dimethyl formamide 1.2. McBriarty..M. The “good” solv ents will ha ve permeation rates greater than an arbitrarily selected v alue.. and the “bad” solv ents will ha ve . This is done by di viding a data set into tw o groups.P.248 Hansen Solubility Parameters: A User’s Handbook TABLE 13. and Henry .1-Dichloroethylene Source: Reprinted from Hansen.4-Dioxane Dimethyl sulfoxide Ethyl acrylate Symbol EBR EDC EET EIM EPC ESH ETA EVE EVK F12 FFA HXA MAL MAM MAN MAT MBR MEK MES MIC MIK MMA MSO MVK NEE NME NTB POX PRA PYR STY TCE TCL TCR THF TOL TTE VAM VDC Compound Ethyl bromide Ethylene dichloride Diethyl ether Ethyleneimine Epichlorohydrin Ethanethiol Ethyl acetate Ethyl vinyl ether Ethyl vinyl ketone Dichlorodifluoromethane (Freon 12 Furfural Hexane Methanol Methyl acrylate Methacrylonitrile Methyl acetate Methyl bromide Methyl ethyl ketone Methyl sulfid Methyl isocyanate Methyl isobutyl ketone Methyl methacrylate Mesityl oxide Methyl vinyl ketone Nitroethane Nitromethane Nitrobenzene Propylene oxide Propylamine Pyridine Styrene 1. J. Philadelphia. ASTM STP 1133. The Performance of Protective Clothing: Fourth Volume.

5 6. A problem of some concern is the permeation through b uried water pipes by chemicals or oil products which somehow reach them. With permission.7 18.6 10. 7. Permeability of polymers.9 13.5 3. The permeation coef ficient data.0 24. a b Source: From Hansen. either by general pollution or by g asoline or oil spills. 1998.1 20. Reasons for this are not vident but may include additives in the polymer .9 δH 4.7 7.7 14.5 9. 20 This correlation is based on data by Lieberman and Barbe 22 and is discussed in more detail in Chapter 5. These pipes are often made from polyolefins .0 5.0 18. g Polyamide 6/66 chemical resistance based on rather uncertain data 24 (see discussion in Chapter 12).9 3.4 7. are reported by P auly19 for lo w density polyeth ylene (LDPE).3 5. “Good” solv ents are arbitrarily considered as those which ha ve permeation coef ficients in these units which are greater than 1. The parameters reported correlate the data well b ut are some what different from those which might be expected for a polyolefin. Recommendation of uncertain-for -use is considered as acceptable for use.7 10.742 0. Swelling data is from Reference 23.3 7.925 1.3.1 8. (g x mm)/( 2 x d).2 5.M.865 0. Attacking solvents are within the correlating HSP sphere. Such a correlation based on permeation coef ficients fo various liquids in PE is included in Table 13.8 4. at 21.0 14.6.0 11.984 0.0 12.3 17.000 0.3 HSP Correlations Related to Barrier Polymers Material LDPE permeation coefficient 21.0 3.9 δP 5. four liquids dissolv ed and one (morpholine) swelled very strongly. 14) Viton® swell c >10 wt% 20°C EVOH sol/swell d Polyvinyl chloride swell e Cellophane — >25% swell PETP chemical resistance (+/– OK) f PA6/PA66 chemical resistance (+/– OK) g PA6/PA66 chemical resistance (+/– bad) g δD 16. One clearly expects more extensive permeation by chemicals that have HSP not too different from those of the polymer from which the pipe is made. all other things being equal.1 11. Recommendation of uncertain-for -use is used as indicated. or some other local v ariation in the composition of the polymer . e Visual observation of v ery strong swelling and/or solubility .0 19.000 0.2 17.927 0.3 6.0 6. Units are MP a1/2.0 8.950 G/T 26/47 4/13 13/21 6/26 4/21 35/50 36/57 5/24 13/47 4/22 7/34 2/31 9/31 Note: Data fit and the number of good liquids (G) and total number of liquids (T) in th correlations are also indicated.9 1.1° Permeation viable skin a PP swelling b ACLAR® 22 >5 wt% swelling ACLAR® 22 2%<swelling<5% Psoriasis scales swelling (Chap.4 3. c This correlation is discussed in Chapter 5.9 12.0 3.0 14.000 1. C.4 9. This correlation is discussed in more detail in Chapter 15 and is based on limited data.7 Ro 8.5 18.000 1. This is why high density PE is a better barrier polymer than lo w density PE.000 1..Applications — Barrier Polymers 249 TABLE 13. local oxidation.6 18. d Ethylene vin yl alcohol copolymer (EV OH).1 18.6 13. It should be remembered that permeation occurs in the amorphous re gions only. f Polyethylene terephthalate (PETP) chemical resistance based on rather uncertain data24 (see discussion in Chapter 12).7 18.5 6.7 2.0 6.1°C. the higher densities are attrib utable to a higher percentage of crystallinity . Pharmaceutical and Medical Packaging 98.0 FIT 1. Attacking solvents are within correlating HSP sphere.7 16.6 11.000 1.9 12.955 0. permeation rates lo wer than this v alue.

250 Hansen Solubility Parameters: A User’s Handbook A HSP correlation has been possible in a v ery special case of polymer permeability where the barrier polymer is viable human skin. How different are the HSP of the test liquids? What are their v alues compared with the predictions desired? The parameters reported in Table 13. An extension of this type of situation can be cited in the tendencies of w ater to associate with itself as well as with local hydrophilic regions within polymers. The HSP correlations for chemical resistance based on data of the acceptable-for -use or recommended/not recommended type are not as reliable as those usually found for solubility and swelling where a suitably lar ge number of liquids are used in the testing.3). Like seeks like.5% weight g ain was arbitrarily set to dif ferentiate “good” solv ents from “bad” ones.3 (see also the data reported in Chapter 5). 20 This is discussed in more detail in Chapter 15.3. has made simple predictions of the beha vior of water very difficult Correlations of polymer solubility and swelling ha ve led to se veral of the HSP data sets reported in Table 13. among other things. As little as 1% of hydrophilic additive can ef fectively destroy the w ater barrier properties of a polymer film but this small amount cannot be measured in swelling or solubility studies leading to HSP correlations. Copolymers. SOLUBILITY PARAMETER CORRELATION OF POLYMER SWELLING Solubility is a major f actor in the equation P = DS. Some chemicals readily permeate this boundary . A different limit might gi ve different parameters.15 also had difficulty correlating the swelling of Viton. A poor data fit can be anticipated when a single HSP sphere is used to describe what should b represented by (at least) tw o overlapping HSP spheres (see also Figure 13. such as Viton. a problem of some significance in a y study of solv ents at lo w concentrations in polymers is that the smaller amounts of solv ent relative to the polymer can lead to preferential association of solv ent with those local re gions/segments/groups in the polymer that have energies (HSP) most similar to their own. These local association effects can influence results on swelling studies at l w solvent uptakes in both good and bad solv ents. The reasons for this are discussed in depth in Chapter 12. This f act. This correlation 21 which is also also has a relation to the HSP correlation for the swelling of psoriasis scales. for e xample. Zellers 14. Swelling data for Viton23 are correlated by the HSP v alues included in Table 13. discussed in Chapter 15.3 for polyprop ylene seem to accurately reflect what is xpected in terms of lo w polarity and lo w h ydrogen bonding properties for this type of polymer . and this f act has been used to establish a tentative HSP correlation for the permeation rate of viable human skin. As stated pre viously. so correlations of solv ent uptake in polymers are important to understand their barrier properties. The amount of water taken up at equilibrium is not reflected by an verall HSP correlation of polymer solubility or swelling. 22 The limit of 0.3 were tak en from Reference 24. The predictive ability of the correlation will depend on the number of test liquids used in the study and their gi ven HSP v alues. These local regions may not necessarily reflect the same a finities as the polymer as a whole. are particularly susceptible to this problem. such as are indicate by the totally soluble-or -not approach most commonly used in HSP e valuations. . The data used in the chemical resistance correlations reported in Table 8. The correlation for swelling of polyprop ylene reported in Table 13.3 is based on solv ent uptak e data reported by Lieberman and Barbe. Other types of studies carried out at low solvent concentrations can also be influenced by this s gregation/association phenomena. Human skin is a polymeric barrier with se veral functions. The HSP found in a gi ven correlation of swelling depends on which polymer se gments the smaller amounts of permeant prefer to associate with. one of which is to help k eep undesirable chemicals out of the body .

7 6.8 (14. b ut other g ases. h ydrogen sulfide. etc.0. Table 13.000 1. As the data fit is 1.3 Bimodal HSP correlation(s) for uptak e of liquids in ACLAR® 22.9 6. H wever.0 2. It is not surprising that there are HSP correlations of permeation coef ficients fo gases in dif ferent polymers as a function of their solubility parameter dif ferences.999 for the >5% correlation and 0.0 4 6 8 10 12 14 δH.7) 2 (18. It has RED numbers of 0.0) 0 0 2 Ro = 4.4 gi ves the δD. δP.Applications — Barrier Polymers 251 ACLAR®22 Bimodal Uptake HSP Correlations Uptake 5 . the numbers are approximately correct.000 ● ■ ✕ 14. Hydrogen Bonding Parameter FIGURE 13.25 who showed .0 for both correlations.10 % 2-5% > 10 % δD δP δH R FIT 1. and δH parameters for a number of gases. Units are MP a1/2. One such correlation using the total solubility parameter has been given by König and Schuch. and δH parameters.8 18. will h ve significant alues for all three parameters. For strictly nonpolar g ases. other sets of parameters can also gve data fits of 1.0 1.0 Trichloroethylene within both 12 10 δP. such as carbon dioxide. δP.978 for the correlation of uptak e between 2 and 5%. the values of δP and δH will be zero. Trichloroethylene uptake is the lar gest among the test solv ents because it is the only solv ent found within both re gions.. SOLUBILITY PARAMETER CORRELATION OF PERMEATION COEFFICIENTS FOR GASES Gases can also be assigned δD.0 4.7 3. Polar Parameter 8 6 Ro = 4 6.

7 17. See Chapter 10 Addendum.0 14. Nitrogen has slo wer permeation than oxygen.0 10.5. those with lo w oxygen permeation coef ficients were those whose solubility parameters were most dif ferent from the solubility parameters of oxygen.252 Hansen Solubility Parameters: A User’s Handbook TABLE 13. An example of ho w HSP principles can be applied to interpreting the beha vior of g as barrier films can be found in the performance of poly(chlorotrifluoroe ylene).9 12.4 δP 16. .8 15.3 4.2 0 0 0 0 0 11. Values changed from 1st Edition.5 11. 1. The barrier properties of films made from this material are no nearly as good for carbon dioxide as the y are for nitrogen or oxygen. The amounts of gases dissolved at low pressures are usually low. and carbon dioxide based on this correlation are 5.0 15. respecti vely. but the enhanced solubility of carbon dioxide in the polymer o verrides this e xpectation. and various laminating possibilities e xist.7 0 0 2. One might ha ve expected the permeation rate of carbon dioxide to be lo wer than that of nitrogen and oxygen as it is a lar ger molecule.5 13.0 14.26 These films are xcellent barriers for w ater and oxygen.0 0 20. and constant diffusion coefficients are xpected.7 11. The RED numbers for w ater.e.0 1. whereas the poorer barriers include polyolefins and polytetrafluoroe ylene (PTFE).5 15. b ut close to the HSP of carbon dioxide. in general agreement with this ranking.1. i. and both are much slower than carbon dioxide.3 15.4 HSP for Common Gases of Interest in Permeation Phenomena Gas Water Ammonia Chlorine Sulfur dioxide Carbon dioxide a Carbon monoxide Ethane Ethylene Helium Hydrogen Hydrogen sulfid Methane Nitrogen oxide Nitrogen Nitrous oxide Oxygen Acetylene δD 15.3. e See Chapter 10. a that the better barrier polymers for oxygen. oxygen. The data on which the example is based are tak en from the commercial literature supplied by Allied Signal concerning their barrier films under the tradename of ACLAR®.0 8.7 14.0 0 4. including polyethylene.7 10.6 15.48. nitrogen. A contributing factor in this is that the HSP of the polymer is some what different from the HSP of oxygen and nitrogen.3 17.9 0 2.2 δH 42. and 0.0 5. This may not be true at higher pressures where solubility parameters for the g ases increase more rapidly than those of the polymers and polymers can absorb them to a greaterxtent.1 17. The better barrier polymers for oxygen include polyacrylonitrile and polyvin yl alcohol.0 11..0 0 17. A HSP correlation for the swelling of ACLAR 22 to greater than 5 wt% is included in Table 13.8 0. 1. and polyeth ylene terephthalate.4.4 6.7 0 0 10.0 5.7 0 0 6. polyvinyl chloride.9 Note: Units are MP a1/2.

29–31 HSP for gases can be calculated using the procedures outlined in Chapter 1 with the special figure for ases included in Chapter 18 (Figure 18. will be good barrier polymers for w ater because of low water solubility. whereas the secondary HSP region is what might be expected from a chlorinated species. indicated from the total ener gy of v aporization is 31. it will be plasticized. polyamides (P A). Correlations of g as–liquid solubility with the solubility parameter are included in Figure 13. Depending on the nature of the additi ve and the amounts present. with solubility parameters f ar different from those of w ater. and neither is there a hydrogen atom. The phenomena correlated in this figure confirm the xpectation that nonpolar polymers. or polyeth ylene terephthalate (PET).Applications — Barrier Polymers 253 Figure 13. even if there are supporting estimates from the techniques of Chapter 3. It might be noted that the scale in Figure 13. The abbre viations used in Figure 13.2 MP a1/2 to be distrib uted between permanent dipole and hydrogen bonding (electron transfer) ef fects.3. some of these can totally change the barrier performance of the base polymer . as well as to optimize cost with performance.4 may be found in standard references. as discussed earlie . δt. such as eth ylene vinyl alcohol copolymers (EVOH). y . and the mole fraction in the liquid phase. This clearly requires e xperimental data to resolv e the distribution of the ener gy of vaporization into components for all of these ef fects. The first of these. These comments and generalities are not necessarily v alid for polymers containing additives. Such secondary re gions can potentially allo w higher permeation rates and greater absorption of unpredictable materials based on a single HSP correlation. The primary uptak e region has HSP that might be expected from a fluoropolyme . As mentioned earlier. etc. This is generally true. There is one liquid in the data which is common to both of the HSP re gions pictured in Figure 13. The total (Hildebrand) solubility parameter . P*. If the inner barrier polymer takes up water. such polymers include the polyolefins as well a chlorinated and fluorinated polymers. x.3 MPa1/2. is for liquid uptak e to greater than 5% by weight. Sometimes an indirect approach allo ws prediction of the uptak e of a g as in a polymer . The solubility parameters for g ases not found in Table 13.. This involves determining the uptake of the gas in a liquid ha ving solubility parameters that are similar to those of the polymer . The second of these is for uptake between 2 and 5% by weight. the mole fraction in the g as phase. of course. This approach e xpands the usefulness of g as–liquid equilibrium data.4 for the equilibium values for water27 and in Figure 13. Searching a database of solv ents. This is trichloroeth ylene (which w as assumed to be a good solv ent in both of the perfect correlations in spite of being absorbed to o ver 10% by weight).3. The techniques of Chapter 3 ha ve not been e xplored in this conte xt ho wever.5. and its barrier properties will be reduced. There can be problems of dividing the total cohesive energy into three parts.4 for the uptak e of w ater in v arious liquids is exponential with data covering almost fi e decades in concentration. this solv ent is absorbed more than an y of the other solvents tested because of this property . plasticizers. 28 The quantity P*y/x is gi ven by the total pressure.5 for the equilibrium v alues for nitrogen.6 MP a1/2. δD found by the usual techniques is 15. There is no dipole moment. LAMINATES Laminated barrier polymer systems are designed to mak e the best use of the properties of each of the individual layers. w ould sho w which of these could behave in an une xpected manner.5 are e xplained in Table 13.A polyolefi .3 sho ws that there are tw o distinct spherical characterizations possible for ACLAR 22. This leaves a residual corresponding to a solubility parameter v alue of 27. These interior barrier polymers often ha ve relatively high solubility parameters. Ev en though trichloroethylene has high RED numbers in both correlations.2). This second correlation is also reported in Table 13. aromatic compounds. Sulfur trioxide is a good e xample of how one cannot come further in dividing the energy of vaporization into components without experimental data. The most common type has a polyolefin on the xterior surfaces to protect the inner barrier polymer from w ater.

5 H from Perry: “CHEMICAL ENGINEERS’ HANDBOOK” (1963). C.1 20. 106 105 ᭪H2S ᭪ NO ᭪ CO ᭪H6 ᭪N2 ᭪ 2 CH4᭪ ᭪ C2H6 O ᭪H2 104 ᭪ C2H4 H = P+ y/x 103 ᭪ C2H2 ᭪ N2O ᭪ CO2 ᭪ C12 102 ᭪ SO2 101 ᭪ NM3 100 0 10 20 30 40 DISTANCE TO “1% IN WATER” MPa½ FIGURE 13.21 and gas–water equilibrium data from Perry et al. Kemi. With permission. y is the mole fraction in the g as phase.254 Hansen Solubility Parameters: A User’s Handbook D P H 1 % IN WATER 15. 21. various combinations of laminates can be systematically designed using HSP considerations as one of the design parameters. . Depending on the performance desired.M.. 27 P* is the total pressure. Dan.4 HSP correlation of gas–water equilibrium data using water HSP values found with a correlation using a limit of >1% liquid soluble in water as a “good” solvent. 1992.4 16. 73(8). (From Hansen.) laminated to such a potentially w ater-sensitive barrier film can significantly delay this upt e and loss of barrier properties and maintain reasonable costs. and x is the mole fraction in the liquid phase.

20. There is a strong dependence of the dif fusion coefficient of permeants in polymers on thei size and shape.Applications — Barrier Polymers 255 5000 NITROGEN D = 11. y is the mole fraction in the g as phase.0.0 4000 MTA 3000 ETA P y⁄x 2000 BEN ETY t 1000 MTC HEP PEN DEC HEX PTF PFM PFH ISA 0 0 1 RED ⁄ 10 2 3 FIGURE 13. 73(8). The HSP correlations are preferably based on systems in thermodynamic equilibrium.5 HSP correlation of nitrogen–liquid equilibrium data at temperatures near 25°C and lo w pressure. This . as t o permeants with identical HSP will ha ve different D if their sizes and shapes are significantly di ferent. Kemi. C. 1992. Dan. With permission..M. P* is the total pressure.9.0 H = 0. (From Hansen. P = 0. RAD. This can clearly affect HSP correlations of permeation coef ficients. These correlations were possible because of the e xceptional dependence of the permeation phenomena on the amount of permeant being dissolv ed. although the correlations presented pre viously on breakthrough times are an e xception to this.) GENERAL CONSIDERATIONS HSP correlations have been possible for many phenomena where differences in behavior on contact with different solvents have been studied. and x is the mole fraction in the liquid phase. = 1.

for e xample. and swelling of v arious types of polymers. b ut this is explained by thermodynamic considerations rather than a relati vely faster diffusion process. This reduces the concentration gradient and pre vents significant self-plasticizatio of the polymer . for exposures of thick samples (3 to 4 mm) of rigid polymers used for tensile testing after solvent exposure for given times. i. HSP correlations have been presented for breakthrough times in chemical protecti ve clothing.5 Key to Symbols Used in Figure 13. a closer HSP match. This is clearly a problem in the determination of equilibrium de gree of swelling and lo w amounts of uptak e. In all of these cases.256 Hansen Solubility Parameters: A User’s Handbook TABLE 13.e. ut this is a w arning that the dif fusion process for these liquids has not yet achie ved equilibrium and that the ef fects of such liquids can be e xpected to be more se vere at still longer times than those used in the study . HSP correlations can be used in this way to find those xposure liquids which have not reached equilibrium at the exposure time chosen for evaluations. See Chapter 16 for more discussion of dif fusion in polymers. Ne glecting such data for the sak e of a correlation can be justified. all other parameters being equal. Selection of polymer–permeant combinations with widely different solubility parameters will ensure low solubility of the permeant in the polymer. Both g ases and liquids are treated. The self-plasticization leads to concentration-dependent dif fusion coefficients. permeation rates in barrier polymers. as well as to determine reliable HSP for the polymers studied. The time required for attainment of equilibrium with the lar ger diffusing molecules can be so long as to be prohibiti ve for their reasonable inclusion in HSP correlations. Crystalline polymers also have a tendency to be more readily soluble in solvents with lower V. where not enough e xposure time has been allowed for attainment of equilibrium.5 Symbol PFH PTF PFM PEN HEX HEP MYC ISA ETY BEN ETA MTA Compound Perfluoroheptan Perfluorotri utylamine Perfluoromet ylcyclohexane Pentane Hexane Heptane Methyl cyclohexane Isobutyl acetate Ethyl acetate Benzene Ethanol Methanol differential in diffusion rate based on solv ent size and shape can also gi ve apparent errors in HSP correlations of polymers for their chemical resistance. a effect which becomes more significant with increasing amounts of permeant being dissol ed. for example. CONCLUSION Successful HSP correlations for the permeation and solubility behavior of selected barrier polymers have been presented to demonstrate the use of simple principles to arri ve at optimum barrier systems. the majority of the outliers in the correlations are the test liquids with higher V. . This problem has also been found. It is suggested that dif fusion rates be carefully considered when liquids with v ery high V are outliers in HSP correlations..

.. J. Dif fusion coef ficient measurements by sol ent absorption in concentrated polymer solutions. 1967.. 651–660. Hansen. Randers. 3rd ed. Ed. 13. P.M. Dif fusion in polymers.. Selection and use of molecular parameters to predict permeation through fluoropolyme -based protective clothing materials. 1993.. C. 12.M. S. Polym. breakthrough times. Elastomer swelling and Hansen solubility parameters. pp. K. E. 93. Eds. 609–614. I. Van Dyk. 7. J... Eng. Nielsen. C. 22. Anonymous. Anonymous. 37(480). Ind. C. Vol.Z. pp.M. Klopfer.B. The measurement of concentration-dependent diffusion coefficients — the xponential case. Chem. Appl. Evans. Sager.. 20(4). 2688–2698.. SCS Dukadan AS. J. J. 2. Wiley-Interscience. Modelling diffusion coefficients.. K. Am. New York. 1956. 1980. C. in Encyclopedia of Polymer Science and Engineering. Färg och Lack. Assoc. Crank.J. and Henry. P . and Bentz. 169–186. 20.M. 1968. 50. Oil Colour Chem. 1993. 1054–1061. 1995. American Society for Testing and Materials. and Kieth. Appl. Am. 2nd ed.. Polym.. J. DE.. Eng... C. Lieberman.. Wilmington. pp. Philadelphia.M. C. Off.. C. Menges. C. C. PA. Zellers. 9. Guidelines for Selecting Chemical Protective Gloves. Hansen. 1990.. C.. and Barbe.M.. 6(4). Hansen... J. J. and Immergut. Du Pont Compan y. E..M. Solubility parameter prediction of the barrier properties of chemical protective clothing. Hansen. and Hardy ... Three-dimensional solubility parameters and chemical protective clothing. pp. and Hansen... and Nielsen.. J. G. 1964. Hansen. N. Eds. Modelling the solubility of organic solvents in Viton gloves. 15. I. Oxford University Press.O... 1985. Three-dimensional solubility parameters and chemical protecti ve clothing permeation. Austin. A mathematical description of film drying by sol ent evaporation. Philadelphia. E. Assoc.M.B. ASTM STP 1133. Comyn. H. Bauv erlag GMBH.B. A. 16. Hansen. Hansen..M. ASTM STP 989. Elastomers Division. K. 10(7). Bikales. Ind. 26.M. 3. 482–483. Wiley-Interscience. Forsberg. 24. 14. 1988.K. Zellers. R. 2005. and Olsson.E. London.C. 513–530.F .. Sci. Pauly. Sci. VI/445–446. in Swedish). C. Ov erberger..H. Branderup. Hansen.. J. American Society for Testing and Materials.. Billing. 2004. Polypropylene polymers. Polymer Products Department.. 1981. Danish Technical Press.. Jr. and steady-state permeation rates of o ganic solv ents in Viton gloves. 24. 27–43. 17. 11. N. TX. Sci.P. 3311–3315. Doc toral dissertation. R. Wiesbaden. ..G.P. Permeability of commercial solv ents through living human skin.. Christensen. Polym. Hansen. in Polymer Permeability. Chemical Protective Clothing Performance Index.. Permeability and diffusion data..A. ( Riktlinjer för val av kemiskyddshandskar. Oxford. Denmark. Ind. 1988. McBriarty. 21.M. Performance of Protective Clothing: Second Symposium. 56. J. 19... C. 301–308. pp. 1989. Eds.H. 894–907. 5. Jensen. New York. C.. Assoc. The affinities of o ganic solvents in biological systems. 50. J. 10. 8. The Three Dimensional Solubility P arameter and Solvent Diffusion Coefficient. Förening Teknisk Företagshälso vård (FTF). Elsevier Applied Science. C. Predicting solubility and permeation properties of or ganic solvents in Viton Glove material using Hansen’s solubility parameters.T. 18. S. and Ebbehoej. Sci. J..M. 1965. Plastguide.I..Applications — Barrier Polymers 257 REFERENCES 1. 11–73. Stockholm. Polym. Mansdorf. Polym. 49(6). 6. 252–258. Rogers. 1989. H.. J. in The Performance of Protective Clothing: Fourth Volume. Hansen. J. Hyg. Hansen... Appl. 4. Testing.. Fundam. Appl. in Polymer Handbook.. Some aspects of the retention of solvents in high polymer films. 51(1). 531–540. Instant Reference Sources.T. and Andersen. J. Fluid Resistance of Viton®.. 1974.M. L.... Water Transport by Diffusion in Solid Materials (Wassertransport durch Diffusion in Feststoffen. Hyg. B. Sci. K.M. 13. Permeation of gases and vapours in polymers. The free v olume interpretation of plasticizing ef fectiveness and dif fusion in high polymers. T. The Mathematics of Diffusion. Eds. Ursin. P A. 1991. Forsberg. 1992. 23.H. Copenhagen.W. K. 57–77. and Hansen.. in German). 4th ed. C. 1988. Polym. and Kroschwitz. 1985.G..M. 197–208.W.. Mark.P. 1967. C.. Dig. II.

1962.. Boca Raton.M. 73(8). Ne w York. 2–7. H.F. R. 28. and Scott. and Scott. 18–22. 1963. Regular Solutions. Hildebrand. in German). 1977.M... in Danish).D. AlliedSignal — Advanced Materials. Chemical Engineers’ Handbook. 26. Dan. Kemi.258 Hansen Solubility Parameters: A User’s Handbook 25.. 3rd ed. C. 25 years with solubility parameters (25 År med Opløselighedsparametrene. Hansen. Structure and permeability of polymers (K onstitution und Permeabilität von Kunststoffen. Kunststoffe. 30. ACLAR® Barrier Films.. S.H. 27–31. Reinhold.L. and Kirkpatrick. CRC Press... and Schuch.. 27. Section 14. Barton. Eds. NJ.. König. Engle wood Cliffs.H. McGraw-Hill. C.. 1950. Allied Signal Inc. Prentice-Hall Inc. FL. 4th ed.L. Anonymous. A. J. U. 29. Mor ristown. . 1992. J. New York. Perry... 1990. 31. NJ. 67(1). Hildebrand. Chilton.. Handbook of Polymer-Liquid Interaction Parameters and Solubility Parameters. The Solubility of Nonelectrolytes. pp. R. J.

In the latter case the similarity of the HSP of the challenge liquid may be to the HSP of an entity in the polymer chain that otherwise might be oriented away from the polymer surf ace. 1 A failure by ESC can appear almost immediately . V. Under given conditions of strain it may prefer the environment of the ESC initiator once there is contact with the polymer . can initiate the cracking process.1 The present chapter is in man y ways a supplement to this e xcellent work. such as rotation of the polymer chain segments at the contact surface. Hansen ABSTRACT Hansen solubility parameters (HSP) can be used to help predict which chemicals can cause environmental stress cracking (ESC) in polymers. There is a re gion at lo w RED including those challenge liquids that dissolve the polymer or are v ery aggressive. It has generally been assumed that some similarity in HSP is required such that some absorption occurs. The ESC re gion on these plots increases in size with increased tensile stress and/or increased critical strain. INTRODUCTION The pre vious chapter dealt with the chemical resistance of polymers and briefly touched o environmental stress cracking (ESC) as one aspect of chemical attack.— Environmental 14 Applications Stress Cracking in Polymers Charles M. An excellent general source of the ESC literature is Wright’s encompassing book on f ailures in polymers in general. but it is not complete as discussed belo w. As the structure weakens locally. It is theorized that some physical movement. This form of failure represents at least 25% of all f ailures in plastics and therefore deserv es special attention. the generally accepted mechanism for ESC has been that some absorption of the active chemical weakens the polymer structure locally . after hours. or e ven after years. the molar v olume. There is a re gion at high RED where the absorption is zero or not great enough to matter or else the absorption rate is slow enough to allo w relaxation of the polymer in preference to ESC. This may be suf ficient to cause a craze or crack. An understanding of this phenomenon seems to be e volving. The tensile stress must be suf ficient to locally pull polymer chains or s gments of chains from the bulk. there is added stress in adjacent re gions. such that the tensile stress increases in adjacent regions. although e xamples are given where ESC takes place without measurable absorption for good matches in HSP at relatively high stress/strain. More recently it has been recognized that measurable absorption is not al ways necessary for ESC to occur in some polymers. As stated above. There is a considerable literature on ESC. and ESC is not found as such. This chapter e xpands the previous discussion of this special type of ph ysical chemical attack on polymers. after minutes. ESC can occur in an intermediate re gion where there is some absorption of challenge liquid. It is no w clear that the polymer and the chemical that initiates the stress cracking must ha ve similar or reasonably similar HSP . There are three distinct regions on this plot. and often occurs without prior w arning. ESC requires tensile stress and correlates when the RED number (relati ve energy difference in the polymer -solvent interaction) found from HSP considerations is plotted vs. The ESC initiator need not dissolv e the polymer as such. a molecular size parameter . ut in man y cases further absorption or further chemical penetration seems necessary to repeat the same process until finally the stress become 259 .

and styrene-acrylonitrile copolymer (SAN). In some cases absorption of liquid permits stress relaxation. There are balances in properties and processing ability that are required for both amorphous thermoplastic polymers as well as partly crystalline polymers in order to achie ve maximum ESC resistance and still maintain other desired beha vior. It is such tie-molecules that gi ve the ultimate resistance to ESC.10) refers to the correlation in question. polymethyl methacrylate (PMMA). the polymer is plasticized and/or weak ened by the absorption of these chemicals. One must either break the tie molecules or pull them out of one of the lamella. HSP correlations for ESC in PEI using data from 8 were also reported.260 Hansen Solubility Parameters: A User’s Handbook too high and a craze or crack occurs. polyamides (P A). All of these HSP correlations are included in Table 14. which is not al ways well defined. and there is potentially a dif ferent type of problem to cope with. Emphasis is placed on the consequences of similarity of HSP for the polymer and challenge chemical. PMMA. whereas the outer sphere encompasses these as well as those gi ving ESC for the gi ven samples.Another point to be remembered is that the RED number (Equation 1. and tie-molecules or chains that e xtend from one lamella and anchor into an adjacent one.1 must be considered with care. Barton3 has discussed in detail the theory and application of the cohesion (solubility) parameters including HSP for man y systems.6dimethyl-1. and e xpected crazing and cracking does not occur .1. Higher molecular weight enhances polymer chain entanglements. PVC. On the other hand. Those with free ends (cilia) extending into the amorphous phase. It is beyond the scope of this chapter to discuss models for the fracture mechanics mechanisms of ESC. such as with acids and bases. polyvinylchloride (PVC). polytetrafluoroet ylene (PTFE). In a sense these act lik e strong ph ysical cross-links. those with chains e xtending into the amorphous phase b ut that loop back into the same lamella (loose loops). polyether ether k etone (PEEK). ESC data are discussed for man y other polymers including polyeth ylene (PE). Figure 14. and thus reduces the tendency for ESC when solv ent is absorbed. En vironment-induced de gradation is also discussed with v arious plots for poly(2. sho ws two essentially concentric HSP spheres resulting from this work. The HSP correlations for ESC reported in Table 14. Hansen and Just 6 studied ESC in the COC-type polymer called Topas® 6013 from Ticona. and polysulfone. Lustiger 2 lucidly describes the mechanism of ESC in partly crystalline polymers with polyeth ylene as an e xample.1. If a chemical reaction is in volved. Additional HSP correlations for ESC in PET . ESC INTERPRETED USING HSP Wright1 has discussed ESC data in terms of HSP for polycarbonate (PC). Wright indicates that about 90% of ESC f ailures involve amorphous thermoplastic polymers with the remaining being found with partly crystalline thermoplastic polymers. and PC were reported for ESC (critical strain) data from Moskala and Jones 7. PCTG. This is presumably for a more or less massi ve uptake but not enough to completely dissolv e the polymer. There are man y practical examples of how not to do things. as witnessed by the 739 pages in his book and hundreds of references. The latter three are found originally in Vincent and Raha4.6 using HSP. The same solv ent polymer pair will ha ve dif ferent RED numbers for a HSP correlation of true . The critical strain is that minimum v alue of strain at which the given challenge chemical will cause crazing/cracking in the given polymer. Any strain le vels abo ve this will result in ESC on contact with the gi ven chemical. and polystyrene (PS). then the phenomenon is more properly called stress corrosion cracking (SCC). Similar considerations are v alid for amorphous thermoplastic polymers. The inner sphere encompasses those solv ents that dissolv e the polymer . PS. In the firs place there are other factors than HSP that are important. including the size and shape of the gi ven challenge molecule in addition to the state of stress/strain. The number of chemicals gi ving ESC in a gi ven polymer increases with increases in the le vel of stress/strain. Wyzgoski5 constructed several plots to help interpret ESC data for nylon (PA) 6.4-phenylene oxide). There are three types of polymer chains to consider in this case.

δP 10 5 0 0 10 Hydrogen Bonding Solubility Parameter.6% b PVC crit.0 12.0 2. Ticona.9 8.000 0. + cracks PC critical strain <0. J. <0.2 6. Mater.0 9.6% c PET crit..-W.6% PCTG crit. <0.. <0.967 0.1 HSP Correlations for ESC in Polymers Polymer Topas 6013 Solubility Topas 6013 Sol. b Data from Reference 10 through Reference 14..0 17.str.952 G/Ta 8/43 15/43 18/47 5/16 12/19 8/19 6/19 9/18 3/20 4/20 9/20 8/45 Note: Units are MP a1/2.31% PEI Ultem ® 1000 600 psi PEI Ultem 1000 1200 psi PEI Ultem 1000 2500 psi PEI Ultem 1000 solubility δD 18.5 9..0 FIT 1. str.0 4.3 11. str.0 3.5 16. See also Chapter 12. str. Eng. 904–916.1 5. 21–25.1 9. (Reprinted with permission from Hansen. a G/T is the number of “good” or attacking solv ents relative to the total number of solvent data points used in this correlation. str.0 7. Chem. Y. Sci.000 1.0 6.989 1.0 9. Topas 6013.6 δP 3. Ultem is a re gistered trademark of the General Electric Company.3 10.1 Three-dimensional HSP plot sho wing those solv ents that dissolv e the COC polymer .8 9.000 1.1.5 10.9 10. All correlations from Reference 6.0 5. With permission.Applications — Environmental Stress Cracking in Polymers 261 Dispersion Solubility Parameter.3 4. The two HSP spheres are almost concentric as can be seen from the data in Table 14.000 1. δH 15 FIGURE 14. Ind.000 1. . <0. in the shaded re gion.857 1.4 12.0 4.0 4.M.1 5.50% PC crit. 21. c Data from Mai. 2001. Cop yright 2001 American Chemical Society.3 6.000 0.7 13.3 18.8 6.0 3. δD 40 30 20 15 10 Polar 5 Solubility Parameter.7 4. <0. Res.3 17. L.0 17.0 R0 5. e xcept as noted.3 8. 1986.55% PCTG crit.974 0. and Just.3 18. C.0 17.9 10.0 7.3 21.4 19.000 0.0 21.) TABLE 14.8 18.6 δH 2. 40(1).6 7. and those that induce ESC in the clear shell.

n-butyl acetate. HSP correlations can be made with man y dif ferent types of data (barrier properties. acetophenone.6%. Delayed absorption can also potentially lead to postponing a catastrophic ESC f ailure be yond normal testing times. for correlating ESC in the polymers PC and PVC. The data used to construct these were tak en from references 10–13. and the benzene rings in acetophenone. Molecules that are too lar ge can simply not enter the polymer . are e vidently sufficient to pr vide the steric hindrance that prevents absorption.14 The methyl “side-group” on the methyl isobutyl ketone. All of the test liquids causing ESC f ailure in the immersed COC c ylinders had measurable surface resistances retarding absorption. it can be seen that these include meth yl isobutyl ketone. which all have very similar total (Hildebrand) or Hansen dispersion solubility parameters. say .12 These data clearly reinforce the need for consideration of shape. b ut they did absorb. There is a re gion at lo w RED including those challenge liquids that dissolv e the polymer or are very aggressive.10. ethyl acetate. these isomers ha ve critical strains that increase from 0. or else the absorption rate is slo w enough to allow relaxation of the polymer in preference to ESC. and nitrobenzene.10–13 to construct a ne w type of plot to correlate ESC data. Figure 14. the rate of motion into the b ulk is dependent on the local diffusion coefficient. The ESC region on these plots increases in size with increased tensile stress and/or higher v alues of strain. RED vs. and especially emphasizes the frequently overlooked surface resistance.) as discussed else where in this book. injection-molded cylinders of a COC-type polymer calledTopas® 6013 from Ticona. that is.2. V. ESC can occur in an intermediate re gion where there is some absorption of challenge liquid. It has been determined that 1.262 Hansen Solubility Parameters: A User’s Handbook solubility than it will have for a correlation for ESC where the “good” solvents have critical strains less than. Hansen9 used data from v arious sources 6. V. and phen yl acetate. and phen yl acetate do not absorb into this polymer at room temperature. There is a region at high RED where the absorption is zero or not great enough to matter . Methyl isobutyl ketone does give ESC at higher stress le vels.4 use the same parameters. For PC. Once an adsorbed molecule locates in a suitable hole. from the older literature.4-dioxane. Equation 1. etc. swelling. wetting behavior. and diethyl ether.6.14 Surface resistance phenomena may play an important part in the ESC process itself.2 sho ws that there are se veral liquids that do not gi ve ESC whereas this would normally be e xpected from their position on this plot. whereas the three not gi ving ESC apparently do not absorb at all under the test conditions.3 and Figure 14. so the sur ace transport coef ficient is i versely proportional to the molecular cross-section. Larger and more structurally complicated molecules ha ve much more difficult finding such a suitable hole. Figure 14. The explanation lies in the fact that the four liquids giving ESC have measurable absorption.2 is such a plot for a lar ge number of liquids in contact with small. meth yl isob utyl k etone. acetophenone. The term RED number is intimately connected with a gi ven HSP correlation. solubility. of the test chemicals. and the molar volume. It can be seen that additional liquids will gi ve ESC as the critical stress levels become higher. This surface resistance is thought to originate primarily from the rate at which adsorbing molecules can locate a hole in the polymer surf ace large enough to accommodate them. The RED numbers and molecular v olumes for all se ven of these liquids are comparable.85% to 1. although e xamples are gi ven below for other polymers where ESC tak es place without measurable absorption for good matches in HSP . 0. respecti vely. sedimentation rates. See the discussion in Chapter 16 for more details and xamples. and ESC is not found as such. nitrobenzene. Therefore the surface transport coefficient is directly proportional to th diffusion coefficient. From Table 14. . Closer study of Figure 14. one should be reminded that there are lar ge v ariations in the critical strain e ven for the he xane isomers. Chapter 16 in this book is therefore dedicated to absorption and dif fusion in polymers.68% as branching increases. These are surrounded on the plot by liquids that do result in ESC including n-hexane. This plot uses the RED number.9 There are three distinct regions on this plot. Before proceeding to perhaps still more complicated theories and situations in the next section.

M. an intermediate region for ESC. Polym.6 2.15–17 It is presumed that there are man y additional cases of this kind.2 C ESC ᭪ No ESC D Severe deformation DMF NEE No problem at stress level ACI CHK DAA NMP 2. in some cases v ery rapidly.2 1.0)..0 ● Soluble 3. (Reproduced from Hansen.0 RO 5.8 Solubility 0. and a re gion at higher RED where ESC is not found. Data from Reference 6.0 2.Applications — Environmental Stress Cracking in Polymers 263 δD TOPAS 6013 solubility 18.6 0. Ticona) using RED number vs. An e xample of surf ace mobility is contact with w ater converting surf aces lik e those of peat moss from h ydrophobic to h ydrophilic.8 2. Red number for true solubility 2.4 0. One possible e xplanation for this is HSP-induced motion of the polymer chain segments in the surf ace. There is a re gion of solubility (RED less than 1. C. This phenomenon is called surface mobility here and has also been discussed in Chapter 15 and Chapter 18.2 Plot of ESC and solubility data for injection-molded cylinders of a COC polymer (Topas 6013.4 HSP.2 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 Size parameter. 2002. 77. There are other e xamples given in Chapter 18. Degradation Stability.0 0. 43–53.2.0 1.0 δH 2.4 1. An applied w ater droplet initially beads up but soon soaks readily in.0 δP 3. Symbols are e xplained in Table 14.6 1. This might be considered as Nature’ s valve to conserv e water within a system.8 1.) ESC WITH NONABSORBING STRESS CRACKING INITIATORS Recent research has shown that even nonabsorbing chemicals can induce ESC. With permission from Else vier Science. The surface again becomes hydrophobic when the water is gone. molar v olume. Rotation or other form of polymer chain se gment . but they have not been reported as such. V MCHL ESC possible 2NP THF D EET ETA C C NTB ISOPH HXA C C DOX EDC ACETOPH MIK C BCN C ●● ● ● ● ● ● ● FIGURE 14. The ESC solvents all have linear molecular structure.

it is still capable of inducing motion in the polymer chains at the surf ace. when they were applied as films. Group A included those that spontaneously spread when applied as a droplet. b utter. A large number of challenge chemicals were divided into three groups. Contamination of the test surf ace may also lead to retraction of an applied film. Retraction of an applied film is not an indication tha ESC will not occur . These include contact with chemicals in hair spray . although the B types had higher critical strains in general. hand creams. . In all cases. Group B included those that w ould not spontaneously spread. but surface entry or lack of same (see Chapter 16) and surf ace mobility of polymer molecules may both be in volved. ESC w as found for PC polymers with polydimeth ylsiloxanes having molecular weights of 340 and belo w but not for those with 410 and abo ve. it is thought that the af finity of the act ve chemical must be high or moderately high.264 Hansen Solubility Parameters: A User’s Handbook TABLE 14. Exactly wh y this happens is not known. where this w ould not occur otherwise. The basis of the failures is the pulling of polymer chains from each other in all cases.2-Dioxane Ethylene dichloride Diethyl ether Ethyl acetate n-Hexane Methylene dichloride Methyl isobutyl ketone 2-Nitropropane Nitroethane N-methyl-2-pyrrolidone Nitrobenzene Tetrahydrofuran motion may also be suf ficient to start the cracking process. COC. and ABS type polymers. DISCUSSION HSP correlations of ESC phenomena ha ve been presented for a number of common polymers. A related study of surf ace phenomena by Nielsen and Hansen 18 explored whether ESC could be predicted by the wetting beha vior of the challenge chemicals on PC. As stated above.2 Symbol ACI ACETOPH BCN CHK DAA DMF DOX EDC EET ETA HXA MCI MIK 2NP NEE NMP NTB THF Compound Acetone Acetophenone Butyl acetate Cyclohexanone Diacetone alcohol Dimethyl formamide 1. There have been many other undocumented cases of ESC where one w ould not suspect absorption of the ESC initiator . It as found that all A and B liquids gave ESC.2 Symbols Used in Figure 14. deodorants. All of the ESC cases kno wn to the author where absorption has not been measurable have had relatively high stress/strain imposed on the samples. but which would not retract either. and this test is only an indication of a potential problem. Care must be tak en. such that e ven though it cannot absorb. and v arious kinds of oils including essential oils. Grou C liquids did not spontaneously spread and retracted when they were applied as films. for xample. the mechanisms for crazing and crack initiation and gro wth have not been discussed in detail here. Some type C liquids g ave ESC and others did not.

C. 1. just as liquids with critical strains abo ve this should ha ve RED larger than 1.55 ‫؀؀‬ ϫ ‫؀‬ ◆ ▲ ● ϫ ‫؀‬ ‫ ؀‬Cyclohexanol 0. The limit for critical strain w as chosen as 0.5 2 ● ● ● ▲ Glycerol 0.M.2-0. Specific data and a discussion r garding the correlation are included in the te xt. Higher critical strains lead to lar ger ESC regions on this plot. The critical strains are indicated in the caption.77.4-0.6%.6-0.48 ● ◆ 0.5 RED NO. 43–53.Applications — Environmental Stress Cracking in Polymers 265 CALCULATED ESC FOR PC AT CRITICAL STRAIN = 0.8 ϫ 0.98.2 clearly sho ws that the molecular shape of the challenge molecules is important in addition to V. 2002.2‫؀‬0. With permission from Else vier Science.0. The data for this figure are found in 6 Figure 14.) however.6 ▲ 0.3 Plot of ESC data for polycarbonate (PC) using data from Mai 10 and others 11–13 to establish a HSP correlation based on critical strain since true solubility data w as lacking.1 shows that the ESC solvents have RED numbers slightly larger than those required for solubility. Polym. ϫ ● 2-Propanol 1. 1..5 0 0 50 100 150 200 250 300 350 MOL VOLUME .02 ϫ ◆ ϫ ◆ ▲ 1 ◆ ▲ ◆◆ ● ϫ ϫ ϫ ● ϫ ϫ ϫ ϫ ϫ ● ● ◆● ϫ ◆ ● ‫؀ ؀‬ Dimethylformamide 1. 77.8-1 ●>1 FIGURE 14. A more detailed discussion of the phenomena in volved is considered be yond the scope of this chapter .CC/MOL ϫ 0-0.99 ● 1. necessarily with other parameters such as molecular size and shape of the challenge chemical and the stress/strain condition of the samples.0.4 ◆ 0. Figure 14. to pro vide improved predictability. HSP has been used as a correlating parameter . ESC may occur for a challenge chemical with a linear molecular structure.6 2. (Reproduced from Hansen. and to ask ne w questions re garding ESC. Degradation Stability. Liquids with critical strains belo w this value should ha ve RED less than 1. b ut not .

Polym.3 clearly shows larger ESC regions for higher critical strain limits for polycarbonate. The critical strains are indicated in the caption.266 Hansen Solubility Parameters: A User’s Handbook HSP CORRELATION FOR ESC IN PVC AT CRITICAL STRAIN = 0. The absorption rate of the b ulkier molecules is too slo w or the y may not e ven be able to absorb altogether . The samples providing the data for this figure were small ylinders.6 2.5 RED NO.4 ◆ 0. so it w as impossible to determine critical strains for the gi ven liquids with these samples. C. (Reproduced from Hansen.10–14 . M. With permission from Else vier Science.2-0.0 ● >1 FIGURE 14.5 ● 2 ● ▲ 1.0.8 ϫ 0. Solvents with critical strains below this v alue should ha ve RED less than 1. but with a c yclic or more branched structure. ▲ PENTANE CYCLOHEXANE ▲ ▲ ▲ METHYL CYCLOHEXANE HEPTANE 1 ϫ NITROMETHANE ● ϫ ● ϫ ▲ ● ‫؀‬ ● ● ETHYL BENZENE ‫؀‬ ‫؀‬ 0.4-0.5 ● ● 0 0 20 40 60 80 100 120 140 160 MOL VOLUME .6 ▲ 0. 43–53. ESC data were used for the correlation since specific solubility data o these samples were not available.4 Plot of ESC data for polyvin ylchloride (PVC) using data from Mai 10 to establish a HSP correlation based on critical strain. More solvents logically gi ve ESC as the strain increases. The data used in this figure were foun in.6-0. Degradation Stability. Higher critical strains lead to lar ger ESC re gions on this plot.) for one with the same RED and V. just as solv ents with critical strains abo ve this should ha ve RED larger than 1.6%.8-1. Figure 14.0.CC/MOL ● DISSOLVES ϫ 0-0. 77. The limit for critical strain was chosen as 0.. 2002.2 ‫ ם‬0.

Wright.. 13. Polymer. Vincent. 5. Eng.. 2001. 14(March). Ind. and Raha.Applications — Environmental Stress Cracking in Polymers 267 Figure 14. Ultem Resin Design Guide. Anonymous. the PC data in Figure 14. Understanding En vironmental Stress Cracking in Polyeth ylene. A. such as PC.. 4. Mater. Bisphenol-A polycarbonate immersed in or ganic media. C. 14. May. Sci. Macromolecules.J. Kambour. 77.-W. Polym.html. CRC Press. E. 1972.3 and the PVC data in Figure 14. C. FL. in Macromolecular Solutions.. 2002.. and Romagosa. M. Environmental stress cracking of glassy polymers and solubility parameters. 10 This figur confirms the ability to use this type of plot to correlate and predict ESC in polymers It is not unusual for suppliers of polymers to change compositions without indicating what has been done and to continue to use the general indication of polymer type. 1979. Hansen. and Hansen. 1974. MPB Article Inde x (originally published July 1996).. 1974. Boca Raton. 40(1). U. 6. 13. pp. Pittsfield. 2005. pp. Barton. New York. 43–53. MA.. Jacques. 1998. Mai. the information may not apply directly in practice because the practical materials may have added components. 8.M.. REFERENCES 1. 44(11). L. Thus. Res. The role of solubility in stress cracking on n ylon 6. 1989 9. Rapra Technology Limited. Res.. Eng. L. Care is especially required whene ver the HSP of the polymer gets too close to the HSP of potential challenge chemicals. It is usually the une xpected events that cause the catastrophic f ailures. In the latter case. 11. 2nd ed. C. 10. Researchers either must w ork with commercial materials the y do not ha ve full knowledge of.M. Y.M. http://www . 167–176. Eds. Swelling and response to stress. R. . 21–25. Medical Plastics and Biomaterials Magazine. Hansen.. C.4 shows the ESC behavior of PVC using data tak en from the literature. Nielsen. R.4 may not be v alid for all polymers of these nominal types. Polym.. This is a significant problem in further research and d veloping fundamental understanding. Polym. 23.H. Evaluating Environmental Stress Cracking of Medical Plastics. GE Plastics. J. Influence of ydrogen bonding on crazing and cracking of amorphous thermoplastics. This dilemma apparently has no fully satisfactory solution. 12. Sci. Medical Plastics and Biomaterials Magazine.I.K. Failure of Plastics and Rubber Products. 1986.. Gruner. Sci. M. A.. Significance of sur ace resistance in absorption by polymers. Lustiger. C. 1153–1166.. Prediction and evaluation of the susceptibilities of glassy thermoplastics to environmental stress cracking. and Stahl. 904–916. Chem. 1991. Chem.L. On Predicting En vironmental Stress Cracking in Polymers.E. and Just.6. but designs or changes in designs that lead to increased tensile stress ha ve also been the causes of ESC in practice.. D. or else tak e the path of making test materials to be certain of what is being dealt with. 3959–3965. and Wyzgoski.. Shrewsbury. It w ould appear that testing is still required to ascertain what beha vior is to be e xpected. Henry. 1982. T. Wyzgoski. Prediction of en vironmental stress cracking of polycarbonate from solubility considerations. G. Appl. 7..B.devicelink. Seymour. Degradation Stability.F..G.M. S...F. b ut these correlations clearly indicate where the problems will be greatest. 34–45. Prediction of en vironmental stress cracking in plastics with Hansen solubility parameters. most often with the simultaneous need to consider the size and shape of the acti ve chemical and the stress/strain condition of the polymer . 248–253. Eng. P. 2. CONCLUSION HSP have been used to de velop correlations of ESC. E. 7. Handbook of Solubility Parameters and Other Cohesion Parameters. Shropshire. 3.B. and Jones.G.com/mpb/archive/96/07/001. The required af finity between act ve chemical and polymer allows correlations of ESC with HSP .M. Ind. M. 2001.P.. 283–287. Moskala. Shawbury. 41–60. J. Per gamon Press. 21.A....

C. L. and Almdal.. En vironmental stress cracking resistance.268 Hansen Solubility Parameters: A User’s Handbook 15.M. 451–461. 17. 18.K. . Publications being prepared. Al-Saidi. C. 2002. Beha viour of polycarbonate in different chemicals by determination of the time-dependence of stress at constant strains.. Degradation Stability.F. Nielsen. and Hansen. Kjellander.. C. 89.. T. Polym. K. Environmental stress cracking of PTFE in k erosene. K. Polym. Degradation Stability. 2005. Hansen. Mortensen.B. Polym.. 82. 2003. 77.. 16.M. Surf ace wetting and the prediction of en vironmental stress cracking (ESC) in polymers. 511–513. 513–516. Degradation Stability.

The key to the importance of nonco valent interaction is that such interactions can continually be brok en and reformed under physiological conditions. Controlled changes in solv ent quality can lead to controlled changes in conformation.δH v alues equal to 19. lard. The three HSP parameters. representing only about 14% of the cohesi ve energy involved (using Chapter 1. polypeptides (proteins). δD. Biological materials such as proteins and DN A have well defined structures in a g ven environment. chloroph yll. The ultrastructure of wood is another example of Nature’s way of establishing order in complex systems. The ability of HSP to correlate surf ace phenomena has made this change mandatory.δP. Poulsen ABSTRACT The Hansen solubility parameters (HSP) of man y biological materials can be found from correlations of ho w the y interact with well-defined liquids. by solv ent quality. and in ma y cases controlled. The often-quoted “lik e dissolves like” has been expanded to “like seeks like” (self-association) to discuss the implications of these correlations. Examples of HSP correlations included in this chapter are DN A.0. implying that this is the dominating interaction. but also induce chiral rotation. These numbers clearly sho w that h ydrogen bonding pro vides by f ar the smallest contribution of the three types of interaction. whereas proteins consist of a combined shape of the secondary. dispersion type inter actions (D). The HSP correlation based on solv ent interactions with DN A resulted in δD. dipolar interactions (P). The term hydrogen bonding is generally used to describe the noncovalent interactions in DNA. proteins. The portion of the molecule with ener gy properties most similar to the surrounding liquid will be oriented toward the liquid (“lik e seeks lik e”). and in some cases quaternary structure that together determine the conformation of the protein. The proper function of a protein requires that certain functional groups are at precise locations within its tertiary and/or in some cases quaternary structure. Solv ents can change not only the ability of noncovalent interactions such as van der Waals. nicotine. The term hydrogen bonding must be considered as an insufficient description of the interactions that determine the structure in such molecules 269 .0. and molecular. and other biological molecules. cholesterol.Solubility 15 Hansen Parameters — Biological Materials Charles M.11. molecular. and ionic bonding. DNA adopts double helices.6–1. The conformation of proteins and DN A can be influenced. The solvent quality in a gi ven environment is e xpected to determine whether a protein is dissolv ed or not. hydrogen bonding. and also to control the w ay it adsorbs onto other materials or interacts with itself.8).20. human skin. w ood chemicals. and δH quantitatively account for the cohesion ener gy density arising from atomic. Hansen and Tim S. δP . and urea. hydrogen bonding interactions (H). Equations 1.0. tertiary.

There are man y simple e xperimental methods to determine the HSP for biological materials.1 for lignin are sometimes used to interpret relations among dif ferent materials. δP . 1 The good solv ents are located within the sphere which is based on Chapter 1. The HSP for simpler compounds can be calculated according to the methods gi ven in Chapter 1. Equation 1.” In the follo wing discussion. It is con venient to print the solv ent database in order from best solv ent to worst solvent to aid in finding alternat ves. coupled to a hydrophobic tail – usually h ydrocarbon in nature. This is a quantitati ve application of the generally used statement “Lik e Dissolv es Lik e. Table 15.. Equation 1.000 is perfect as discussed in Chapter 1). Such amphipathic molecules ha ve a head group that is strongly h ydrophilic. this equation is in agreement with the Prigogine corresponding states theory of polymer solutions as discussed in Chapter 2. significant surace adsorption or not. A most interesting and important class of molecules are called amphipathic. Figure 15. An example from biology is the amphipathic molecules (lipids) that form the basis of the biological membrane bilayers that surround cells. Furthermore. This may result in solutions or in selecti ve orientation of se gments of molecules in more complicated systems. These are discussed in the following in more detail with an indication of ho w such data may be used. HSP v alues for nicotine. 1–7 Most of the materials discussed in these references are also included in the present discussion. The output is often arranged with the best solv ent (lowest RED number) at the top of the list. The statistical thermodynamics approach presented in Chapter 3 also sho ws agreement with the concepts to which this book is dedicated. Ro.10). rapid permeation or not. as described in Chapter 10. These may be monolayers on the w ater surface. The data included in this table are the δD. differences in degree of equilibrium swelling. and molecular hydrogen bonding (H) interactions. as stated in pre vious chapters. The basis of the division of the cohesive energy density into three parts accounting for the atomic dispersion (D). this equation has also been sho wn to be correct for such comple x materials as asphalt and bitumen. The f act of solubility. is given in detail in Chapter 1. their general affinity for ater (hydrophilicity) and/or their general a version to w ater (hydrophobicity). micelles (spherical structures stabilized by a single layer of molecules at the w ater interface) or bilayer v esicles may form. the data fit (where a fit of 1. Alternatively.270 Hansen Solubility Parameters: A User’s Handbook INTRODUCTION HSP have been used to characterize man y biological materials. respecti vely. of course. or other measurable quantity significantly influenced by ysical affinity relation can be observ ed and used to find the HSP for a material being studied. with only the head groups immersed. These are by nature not only dif ferent in size and shape. The units for the solubility parameters and Ro are MP a1/2. b ut many more can be added by experiment or calculation. and δH parameters. Again. These involve contacting a material of interest with a series of well-chosen liquids. skatole. the radius of interaction for the HSP correlation. The native conformation of proteins is a strong . When one attempts to dissolv e these molecules in water. this concept is e xpanded to “Lik e Seeks Lik e” (self-association). RED numbers indicate solv ent quality with lo wer values being indicati ve of better solv ents (see Chapter 1. if appropriate.1 contains HSP data for se veral biologically interesting materials. as described in Chapter 9. molecular dipolar (P). These methods have been discussed in more detail in earlier chapters. G is number “good” solvents and the total number of solvents in a given correlation is T. Plots of the kind gi ven in Figure 15. that are discussed in this chapter were calculated using these methods. These exhibit both hydrophilic and h ydrophobic properties simultaneously .9. Another example is amino acid side chains. etc. w ood chemicals.1 sho ws a typical HSP sphere correlating e xperimental solubility data for lignin. A HSP correlation can. This implies that se gments of molecules seek re gions of similar HSP if this is possible. they form special structures. but also in the char ge they carry. be used to predict the beha vior of solv ents not included in the experimental work. and carbon dioxide solubility in solv ents. if the mixture is vigorously stirred. The correlations reported here are a result of data processing with the SPHERE program described in Chapter 1.

This is also highly dependent on the interaction that tak es place with w ater. A. the associating se gments may be characterized by high δH. because of the presence of an alcohol. When it is immersed in w ater a polypeptide chain will not stay in an elong ated form. With permission. δP Dispersion Solubility Parameter.M.1 HSP correlation showing the solubility of lignin. HYDROPHOBIC BONDING AND HYDROPHILIC BONDING (SELF-ASSOCIATION) The concept of “lik e seeks lik e” offers a general e xplanation of h ydrophobic bonding. is not soluble in w ater and ultimately finds anothe aliphatic h ydrocarbon chain with which to associate. and Björkman. Hydrophilic bonding is formed when the HSP for the associating se gments are too high to allo w solubility in the continuous phase. or amide group.1 δH 16.9 RO 13. acid. for e xample. An aliphatic hydrocarbon chain on a protein. as proteins e xist in an aqueous environment. . These general concepts of hydrophilic and hydrophobic entities can also be quantifie using HSP. 339.Hansen Solubility Parameters — Biological Materials 271 Hansen Solubility Parameters δD Lignin 21. If the continuous phase is a h ydrocarbon liquid. (From Hansen. Units are MP a1/2. δH FIGURE 15. 1998. for e xample.7 40 30 20 15 Polar Solubility 10 Parameter. Good solv ents are located within the sphere. This can be called hydrophilic bonding..9 δP 14. This same type of process leads to micelle formation when the solubility limit of surface active agents is exceeded. 52. δD Lignin 15 10 5 0 0 5 10 Hydrogen Bonding Solubility Parameter. Holzforschung. C. It will instead fold up into secondary structures according to the polarity of the side chains it contains and the rotation of peptide backbone bond angles that are lar gely determined by Van der Waals radii of side chains. Hydrophobic bonding is found when the HSP for the associating se gments are too lo w to allow solubility in the continuous phase.) function of the interactions that tak e place within and between polypeptide chains.

4 5.9 20.0 6.0 δH 11.0 5.2 2.272 Hansen Solubility Parameters: A User’s Handbook TABLE 15.2 demonstrates ho w hydrophilic bonding between v ersamid polymer blocks reacted into an alkyd (polyester) polymer gives a thixotropic alkyd paint with its special nondrip properties.8 11.1 10.4 16.0 9. Here the nonpolar residues stick into the h ydrophobic part of the lipid membrane and the h ydrophilic residues form part of the channel interior responsible for the passage of small molecules across the membrane.5 16. G/T represents the number of good liquids (G) and the total number of liquids (T) in the correlation.9 15. b ut they reform quickly ag ain after application.6 7.2 18.4 4. beta-sheet backbones are stabilized by hydrogen bonds between tw o beta sheets.0 12.8 1.8 14.864 — — — 0.0 — — 11.000 0.990 0.1 19.4 15.856 0.1 16. air–w ater interf aces.9 11.5 22. This plays a crucial role in helix–helix interactions and in the interaction of small peptides that ha ve a helical conformation with membranes.3 Ro 11.980 0. Agitation of the paint is enough to break the h ydrophilic bonds allowing easy spreading.7 1.2 22.1 15.9 17.000 1.2 18.2 16.3 17.3 9.9 42.3 23.927 1.9 18.1 19.5 20.1 21.3 7.6 17.4 17.4 16. but the bonds occur between neighboring strands.1 15. MPa1/2 Material DNA Cholesterol solubility Lard 37°C solubility Lard 23°C solubility Olive oil solubility Psoriasis scales swelling Human skin — permeation Nicotine — calculation Skatole — calculation Chlorophyll — solubility Sinapyl alcohol calculation Coniferyl alcohol calculation p-Coumaryl alcohol calculation Lignin — solubility Dextran C (= amorphous cellulose) See Chapter 5 Sucrose solubility N-methyl-morpholine-N-oxide calculation Blood serum — swelling Zein — solubility Urea — solubility Water — >1% soluble in Water — totally miscible Water — single molecule δD 19.9 24. which disrupts lattices of w ater molecules.9 16. .000 — — 0.000 1.0 8.1 16.8 20.0 20.3 16. Lik e alpha-helices.000 1.5 7.0 25.4 22.0 19. An alpha helix can tak e the form of an amphipathic helix with a polar and a nonpolar side.0 19.7 15.4 12.999 0.0 12. This distrib ution of h ydrophilic and h ydrophobic residues has been observ ed in the membrane protein porin that forms a beta-barrel structure.0 20.2 11.9 12. Hydrophobic sidechains are oriented to minimize the ener gy lost by the intrusion of amino acids into the w ater solvent.4 19.8 7.6 18. Hydrophobic bonding forms an interior .964 0.000 1.1 Hansen Solubility Parameter Correlations for Biologically Interesting Materials.5 δP 20.2 19.981 0.880 — G/T 6/12 25/41 29/50 21/50 29/50 35/50 4/13 — — 7/35 — — — 16/82 5/50 6/50 4/51 4/50 14/50 88/167 47/166 — Note: The units for the solubility parameters and Ro are MP a1/2.0 14. The most common secondary structures are alpha helices and beta sheets that are stabilized by local inter-residue interactions mediated by h ydrogen bonds. Figure 15.4 5.9 24. Hydrophobic bonding is a major ef fect that dri ves proper protein folding.1 — — — 13.0 2.9 22.4 6.8 10.0 7.7 17.4 21.6 12. If the beta−strand contains alternating polar and non-polar residues it forms an amphipathic beta sheet.3 16. Beta sheets are alternati ve secondary structure to the alpha-helix in proteins.0 17. and self-assembly processes.0 — FIT 1.984 0.1 13.1 18.2 19.

Addition of n-butanol destroys the thixotropic effect. which in f act means that these se gments are dissolv ed in a good solv ent. In the Watson–Crick model the base pairs are stack ed on one another with their planes nearly perpendicular to the helix axis where the hydrophilic phosphate–deoxyribose backbones are on the outside. e. because the HSP of these se gments is too high. w ater. The Versamid segments associate because they are not soluble in mineral spirits. Additions of urea. DN A replication. but more research is required to quantify them. This complementary base pairing (hybridization) is central to all processes in volving nucleic acids. Polymer A (alkyd) δH FIGURE 15. increase the HSP of the continuous media to such an extent that it can no w dissolve the “hydrogen bonded” segments. DNA The double helix structure of DN A suggested by Watson and Crick is stabilized by h ydrogen bonding between bases on opposite strands when the bases are paired in one particular w ay (A+T or G+C).. These phenomena must also ha ve their e xplanation in the “lik e seeks/dissolves like” phenomena. as discussed later in more detail. . by ef fectively dissolving them in a solv ent mixture that is better than water itself.Hansen Solubility Parameters — Biological Materials 273 A Mineral spirits B A δP B regions are not soluble and “Precipitate” together B B A Polymer B (versamid) Butanol M. in contact with the aqueous environment. This denatures them. In cells it occurs in.S. Additions of salts can also improve solvency for a gi ven material or se gments of materials. where most h ydrophobic sidechains can closely associate and are shielded from interactions with solv ent water.g. Similar relations e xist for the true solution of some proteins by additions of urea to w ater.2 HSP relations for establishing thixotrop y in an alk yd-type paint. since the solvent then becomes too good. Additions of salts can also reduce solvency. The solid circle represents the solubility of the alkyd (A) and the dotted circle that of the Versamid (B). hydrophobic protein core. The protein is denatured. Such mechanisms of controlling solvent quality can be e xpected to be used by Nature in man y biological systems to control adsorption and/or transport of v arious types of materials as in self-association. Formation of “h ydrogen bonds” within proteins is based on the lack of solv ency in the continuous media.

2 13. adenosine.2 9.5 11. is found when a population of particular DN A sequences is at a midpoint between fully double-stranded and single-strand.0 o RED 0. meth yl riboside (both negligible) < cyclohexanol < phenol.5 20. microarrays. Fortunately.353 0.1 26. Formamide and dimethyl sulfoxide were also considered as “good” and added to the data for the correlation reported in Table 15.8 For in situ h ybridization this implies that microscopic preparations must be hybridized at 65–75° for prolonged periods. thymidine < purine.8 36. These HSP data were di vided into tw o arbitrary groups of “good” and “bad” with a dividing line between purine as good and pyrimidine as bad. Urea w as inef fective with poly A and only slightly ef fective with DNA.142 2.5 18. P. so that h ybridization can be performed at lo wer temperatures in the presence of formamide.8 78. inosine. coumarin.4 156. V is in cc/mole. but it will also be a good solv ent for phosphate salts.3 100. sequencing.0 12. ” Even urea.7 12.2 14. Formamide is often used in connection with DN A.9 26.0 19. and tRNA structure.1 12. uridine < c ytidine.65 °C for each v olume percent of the solv ent that is present. b ut it is also used in laboratories in RN A and DNA gel blots.2.1 D = 19.2 M Na+.3 14.2 6.9 17. This may lead to deterioration of morphology. Formamide is not at the top of the list.7 14.8 106.651 0.5 45.0 R = 11. An article in the older literature 9 reports aspects of the interaction of dif ferent low molecular weight materials with DNA.0 δP 16.7 4. The melting temperature.8 87.5 17.853 0.0 157. which may also contrib ute some ef fect. in situ hybridization.2 ranks the gi ven solvents in approximately the same order as that gi ven in Reference 9.9 18. Tm.4 17.3M.0 δH 10. F ormamide reduces the Tm of DN A-DNA and DNA-RNA duplexes in a linear f ashion by about 0.923 0.2 Hansen Solubility Parameter Correlation for DNA Solvent Dimethyl sulfoxide 2.0 18.3 162.0 11.1-0.342 1.807 0.492 2.4 11.6-dichloro-7-methylpurine. PCR.4 20.447 1.0 20.000 NO = 12 Note: Units of D.4 5. genotyping.0 117.5 19. All the solv ents from p yrimidine and lo wer were considered as being “bad” and all those abo ve this were considered as being “good.393 V 71. p yrimidine. At a concentration of 0. where performance may be af fected significantly by the presence of ater.0 20.0 P = 20.2 20.0 FIT = 1. transcription. Other common solv ents can also reduce Tm. including dimeth yl sulfoxide.4 13. The compounds intermediate in the above list were structurally too complicated to allow a reliable calculation.002 1.6-Dichloro-7-methyl purine Coumarin Purine Caffeine Formamide Pyrimidine Phenol Urea Cyclohexanol Methyl riboside Adonitol DNA δD 18.274 Hansen Solubility Parameters: A User’s Handbook TABLE 15. The encouraging correlation reported in Table 15. for e xample. H and Ro are MP a1/2. etc.7 10. DNA melts (denatures) at 90-100 °C in 0. purine lo wered the Tm of DNA by about 9 °C. 2. The effectiveness of formamide is primarily because of its low molecular volume. Dimeth yl . rRNA.9 39. The order in the table is from e xpected best at the top to e xpected worst at the bottom. The summary of this article states that the order of increasing acti vity was found to be: adonitol.2 95. seems to be placed correctly. deoxyguanosine < caf feine.0 H = 11. The HSP for se veral of these ha ving reasonably simple structures were estimated by the methods of Chapter 1. b ut is the preferred solv ent of use today in man y cases. organic solvents reduce the thermal stability of double-stranded polynucleotides.977 1.5 32.

as the small molecules can reach the critical sites more readily than lar ger ones. such as with the h ydrocarbon solvents.20. This has also been reported in Reference 10.0 MP a1/2 compared with δD equal to 19. Other supporting e vidence that the correlation is reasonable can be found in the estimated HSP for adenine and th ymine. The δH parameter of alcohol solv ents decreases relati vely more rapidly with increasing temperature than for solv ents where the δH parameter is lo w (or zero). then the D. This brings the HSP of the alcohol solvent closer to the HSP of the hydrocarbon solvents.2. Figure 15. . the estimated v alues for DNA based on its interaction with a number of solv ents as reported in Table 15. Both of these are reasonably close to HSP equal to 19.2.000.δP.0.20. This was not w arranted under the present circumstances. Many mixtures of tw o nonsolvents which dissolv e polymers when admix ed have been reported in the literature.14. it also became ob vious that the solubility of cholesterol in hydrocarbons was limited and quite temperature dependent.8.11. but is recommended if more e xtensive use of these data is planned.5 g of cholesterol in test tubes together with 5 ml of each of 41 dif ferent solvents.The limited solubility in h ydrocarbon media and v ery low solubility in w ater favors a location at an aqueous interf ace with the alcohol group of the cholesterol molecule oriented toward the high energy aqueous phase.000. and H could be maintained with a slightly larger Ro. Continued testing with additional test solv ents located in the boundary re gion of the sphere is possible to define it more precisel . is still considered reasonably reliable because of the essentially correct ranking. and δH and Ro for cholesterol solubility were found as 20. CHOLESTEROL Cholesterol has been characterized with HSP based on its solubility in a lar ge number of solvents. These can be considered as single rele vant portions of DNA.0 for th ymine. Changes to ward lo wer temperature will tend to force more cholesterol out of a h ydrocarbon matrix.Hansen Solubility Parameters — Biological Materials 275 sulfoxide will also be a reasonably good solvent for phosphate salts. The 50:50 vol mixtures of n-hexane with 2-nitropropane and n-hexane with ethanol predictably dissolv ed cholesterol at 0.0. A general confirmation of the HSP correlation for cholesterol as done by studying mixtures of nonsolvents.4.9.3 shows this HSP correlation for cholesterol.0. all v alues having units of MPa1/2.6. δP. The cohesive energy derived from hydrogen bonding is about 14% of the total using Chapter 1.16. combinations possible when the data fit is 1. There are many different D.13.4 and 12.0. If p yrimidine had been considered as being good. The data fit of 1. The temperature was 23°C. These data were analyzed by the SPHERE computer program described in Chapter 1 to find the HSP for cholesterol. and δP equal to 20.5. Low molecular volume is very conducive to dissolving polymers with structure or crystallinity .6 to 1.0. 1 Such synergistic mixtures can be predictably found when the y are pairwise on opposite sides of an HSP sphere.0. The 25 “good” solv ents dissolv ed the entire amount of cholesterol added.9 for adenine and 19. The HSP are δD. Equations 1. This behavior in hydrocarbon solvents relates to the interactions of cholesterol in the hydrocarbon (hydrophobic) portions of lipid layers. δD. All units are MP a1/2. and miscibility improves markedly as temperature increases. in spite of the v ery few solvents. This figure also includes s veral solv ents that are discussed in the follo wing. The test method involved placing 0. The radius is arbitrary and depends on the criterion used for good and bad.0. The δH value for DN A is only 11. P .8. ut the present correlation. This clearly sho ws that the h ydrogen-bonding interactions are f ar less important than the other two types of interaction. During the course of this study . and the h ydrocarbon portions oriented into the lipid layer .0. all else being equal.δH equal to 20. and H. Total solution or not at this concentration w as e valuated visually . and the data fit ould still be 1.5 g/5 ml. P. being considerably higher at slightly elevated temperatures. Smaller molecules are also predicted to be thermodynamically better. where it is more compatible.0 indicates that there are other sets of parameters for spheres which can b expected to give a perfect separation of the good solv ents from the bad ones by a “spherical” HSP correlation.

8 δH 9.6 DISSOLVING MIXTURES OF NON–SOLVENTS 2–NITRO– PROPANE 15 10 ETHANOL δP 5 (9. The presence of ethanol or other organic solvent in the lipid layer allows greater cholesterol miscibility in its hydrocarbon portions. This. The simple e xperiments described pre viously indicate that the cholesterol uptak e in h ydrocarbon portions of a lipid layer will be greatly enhanced when ethanol is present. The solubility of cholesterol in an essentially nonsolv ent such as w ater can be enhanced by additions of a solv ent improver such as ethanol. Or ganic solv ents with HSP resembling those of the lipid layer may be found due to occupational e xposure or for other reasons. preferentially remo ves some of the cholesterol from the blood stream. These can predictably be found by selecting pairs located on opposite sides of the HSP solubility parameter sphere. Nonsolvents which synergistically interact to become impro ved solvents when mix ed are indicated.1 0.4.3 HSP sphere correlating the solubility of cholesterol. 20.4) 0 20 15 δH 10 5 0 HEXANE FIGURE 15. of course.8. Therefore.9 15.8 δP 2.1 0. The average HSP for these mixtures are closer to those of cholesterol itself.0 19. such as drinking alcohol-containing be verages. The reason for this is the synergistic effect of ethanol and hydrocarbon segments described earlier. 2.2 14.8 12.4 RO 12. Units are MP a1/2.4 16.4 4. One can also surmise what might happen when ethanol or other or ganic solvent is present in the body.276 Hansen Solubility Parameters: A User’s Handbook MATERIAL CHOLESTEROL 2 – NITROPROPANE HEXANE ETHANOL δD 20. those persons with alcohol in their blood can anticipate a .0 8.

This is reported in Table 15. The cyclic nature of the solv ents. and sulfolane have. for example.927) indicates that a reasonabl reliable correlation for swelling of the psoriasis scales (k eratin) has been found. and methyl methacrylate. The units for these parameters are MP a1/2. The concentrations chosen were 10%. There are too fe w data to establish a reliable correlation. Permeation data generated in an e xtensive study allowed placement of the tested solv ents into groups according to actual permeation rates through viable human skin.0. gamma-butyrolactone. 12.1. which is v ery nearly a liquid at this temperature. such as the presence of w ater and lipids. LARD Experimental data and HSP correlations for the solubility of refined lard at 23°C and 37°C h ve been reported. however. Ne vertheless. as toluene. as the same solv ents that dissolve it at room temperature also dissolv e lard at 37°C. n-butyl acetate and toluene are also in this group. b ut the y are placed in the lo w permeation rate group for actual permeation through viable human skin.4 graphically shows the HSP correlation that resulted. high affinit for psoriasis scales. n-octyl acetate has a near zero permeation rate. so the conclusions drawn for the solubility of lard will also be generally v alid for depot f at. i fact. It w as anticipated that the solubility parameter correlation for the psoriasis scales (k eratin) w ould to some e xtent reflec permeation in human skin b ut that other f actors. This reflects the compl xity of actual skin permeation and the importance of using viable skin for testing. but a sphere with center at δD. Esters are among the best solv ents for lard at 37°C.1. This correlation cannot be considered precise because of insuf ficient data. 4 Figure 15. toluene. The refined lard is a semisolid with a melting point of 42°C The composition of refined lard is ery similar to that of human depot f at. Octyl alcohol does not ha ve a strong af finity for lard at room temperatur with a RED number (see Chapters 1 and 2) of 0. Ho wever. This effect and that discussed earlier should help to reduce cholesterol levels in the blood and blood v essels of those who ingest small to moderate amounts of alcohol on a re gular basis.0. which has a radius of 5. and there are. would also be important. 2 The criterion for a good solv ent is that it totally dissolv es the sample at the gi ven temperature. is also e xpected to slow the rate . whether it be an expanded correlation or formulation of products designed for a prescribed compatibility with human skin.96. Lik ewise. and 11. δP . or are predicted to ha ve. HUMAN SKIN A first attempt to characterize human skin with HSP as made by visually evaluating the swelling of psoriasis scales immersed for a prolonged time in dif ferent solvents.6. The good solvents reflect the crystalline natur of the lard. the c yclic solv ents propylene carbonate. The best room temperature solv ents for lard include trichloroeth ylene. there is a good guideline for future w ork. encompasses the parameters for the four solvents with the highest permeation rates while excluding the others.Hansen Solubility Parameters — Biological Materials 277 slightly higher solubility of cholesterol in their blood because the continuous phase has solubility parameters closer to those of cholesterol.5. for e xample. numerous spheres with some what similar b ut different combinations of the parameters that also can accomplish this. The results of the correlations are gi ven in Table 15.2 Uptake could clearly be seen by dimensional changes and a mark ed enhancement of clarity . Calculations for skatole and nicotine predict that moderate rates of skin permeation can also be expected for these. Olive oil is a con venient material to use at room temperature to study the beha vior of depot f at (lard). and δH of 17. styrene. These all have high δP and low δH. The data fit for this correlation (0. reflecting the presence of the este groups in the lard. the δD parameter is thought to be too high. is an e xcellent swelling solvent for partly crystalline polyethylene. It might be noted that the four solv ents with high permeation rates also ha ve very high affinit for psoriasis scales according to the correlation pre viously noted.

1 19.4 5.0 8.2 7.1 5.0 7.8 δP 16.0 18.0 PERMEATION RATE l HIGH l MODERATE ؋ LOW Ã “0” 20 HANSEN POLAR SOLUBILITY PARAMETER.4 6.6 2.8 3.0 15.0 18.0 PPC؋ BTA ؋ ؋ SUL lDMSO lDMF 15 lNMP 10 lDMAC l MEK l MCL l ETH 5 TOL؋ 0 0 BAC؋ Ã OAC 5 10 15 20 HANSEN HYDROGEN BONDING PARAMETER.9 δH 10.8 95.3 4.5 132.4 15.2 6.3 9.3 6. δ H = 11. .3 10.4 16.2 11.0 18. δ H FIGURE 15.0 106.0 92.0 19. RO = 5.5 96.5.8 18.4 17.3 196.278 Hansen Solubility Parameters: A User’s Handbook SOLUBILITY PARAMETER PLOT FOR SKIN PERMEATION RATE δD DMSO DMF DMAC NMP MCL MEK ETH BAC PPC TOL BTA SUL OAC 18.7 11. δ H CIRCLE: δ P = 12.2 16.5 12.4 7.5 85.7 18.8 76.8 15. Units are MP a1/2.0.8 20.9 90.5 63.1 58.4 16.0 1.6 16.1 2.3 77.1 MV 71.4 16.4 Permeation rates of selected solv ents through viable human skin sho w a correlation with the HSP4 although the data are not e xtensive.

the swelling of the psoriasis scales in volved equilibrium swelling of the individual systems. A major difference is that chlorophyll is soluble in ethanol.1. It is also common . These results are part of a study considering the ultrastructure of w ood from a solubility parameter point of vie w. The HSP parameters for blood serum and zein are not too dif ferent. The saturated solution of urea and w ater is also a (predictably) good solv ent in that it swells blood serum and dissolv es zein. This indicates a higher h ydrophilicity.3.1. whereas lignin is not. These HSP parameters are found by a form of e xtrapolation. however. a protein deri ved from corn. which are apparently not always predictable by these simple considerations. and that the HSP parameters for the proteins are much higher than for any liquid which can be used in such testing. It is note worthy that there are only four good solv ents in the data set reported in Table 15. Factors affecting permeation ha ve been discussed at length in Chapter 13. of course. Finally .4A and Table 15. Mixtures of solv ents. while the zein data are for visual observ ation of true solution. It can be seen that these indeed ha ve high affinity/p ysical resemblance to each other . The HSP for De xtran are higher than those of sucrose (which v alues are similar to the other sug ars as well).1. Solvents with the lo west RED numbers in the correlation for the solubility of zein are listed in Table 15. the presence of w ater and/or other skin components can also have an effect on the permeation rate. are included in Table 15. N-methyl-morpholine-N-oxide is an e xample. b ut it w as not included as a data point in the correlations as such. Of course. Here. The data used in these correlations are found in Reference 3. It is common for polymers to ha ve higher HSP than the monomers from which the y are made. British Drug Houses). It can be presumed that the HSP for these materials are the result of natural selection by nature for optimum compatibility relations with immediate surroundings and function. whereas the permeation rate studies did not have this uniformity.Hansen Solubility Parameters — Biological Materials 279 of permeation relati ve to linear solv ents of comparable af finit . and mixtures of solv ents with w ater have been avoided as test solvents to the extent possible because of too man y interactions. The HSP for amorphous cellulose are presumed to be similar to those of De xtran (Dextran C.6 The study is based on the principle of “lik e seeks lik e” and leads to a proposed configuration of the ultrastructure. WOOD CHEMICALS AND POLYMERS The results of HSP calculations and correlations for se veral w ood chemicals and polymers are given in Table 15. The values are very much dependent on the mathematical model which includes the coef ficient “4” (see Chapte 1 and Chapter 2). but this point has been studied in more detail for the relations among w ood chemicals and w ood polymers as outlined in the ne xt section. PROTEINS — BLOOD SERUM AND ZEIN HSP correlations for the swelling of blood serum and for the solubility of zein. More specific information on the lignin correlation is found in Table 15. The general prediction that additions of urea to water will impro ve solvency of proteins is discussed belo w. where all of the good solv ents are located in the boundary re gion of the respecti ve spheres. water. the HSP for lignin have a demonstrated clear importance with re gard to compatibility relations. and gi ves a higher δH parameter to chlorophyll compared with lignin.3. CHLOROPHYLL AND LIGNIN5 The results of HSP correlations of solubility for lignin and chloroph yll are gi ven in Table 15. The blood serum data are based on visual observ ation of swelling. The crystallinity in cellulose will require that good solv ents have higher af finity/HSP than most of those dissolvin Dextran.4B. A discussion of this is beyond the scope of this work. with the HSP values not being too different. Concentration gradients are required for permeation to occur .

9 87.4 18.9 13.101 1.0 114.4 18.8 15.997 0.3 9.8 79.2 16.6 94.0 71.2 18.8 9.2 11.001 1.1 5.4 6.8 76.068 1.7 5.5 8.165 1.5 86.0 13.991 0.226 1.1.0 16.4 8.0 100.8 16.5 14.7 19.073 1.0 18.7 17.5 4.3 10.2 17.4 δP 8.0 130.099 1.4 17.136 1.6 5.0 75.1 11.1 97.1.3 7.8 21.130 1.4 17.0 120.5 7.087 1.155 1.3 14.137 1.2 4.7 91.6 12.7 10.198 1.2 16.9 12.976 0.2.4 10.9 92.0 15.893 0.0 18.6 18.3 10.9 73.4 17.N-Dimethylacetamide Bromoform δD 18.108 1.0 18.876 0.5 175.8 6.8 4.112 1.0 9.6 11.9 9.0 17.999 1.910 0.2 13.4 12.5 109.9 116.2 14.8 66.0 108.933 0.1 7.2-Tetrabromoethane Ethanolamine Succinic anhydride 2-Pyrolidone Allyl alcohol Ethylene glycol Ethylene glycol monometh yl ether Cyclohexanol Diethylenetriamine Benzoic acid Triethyleneglycol 1.068 1.2 17.6 19.2 17.8 5.6 8.950 0.0 29.0 143.4 15.1 106.5 0* 0* 1 1* 0 0 0 1* 1* 0 0 0 0 0 0 .2 21.117 1.0 16.5 5.2-Tetrachloroethane Ethanol 1-Propanol Morpholine Ethylene glycol monoeth yl ether Dimethylformamide Propylene glycol monophen yl ether Quinoline Hexylene glycol Dimethyl sulfone Dimethyl sulfoxide Ethylene cyanohydrin 1-Butanol 2-Propanol Ethylene dibromide Tetramethylurea Glycerol Diethylene glycol monometh yl ether Diethylene glycol monoeth yl ether N.7 6.4 17.4 13.7 17.6 15.1 16.127 1.5 87.6 17.179 1.8 11.2 118.2 6.021 1.761 0.128 1.180 1.5 17.0 8.200 1.198 1.3 118.3 6.1 15.8 16.9 10.0 18.3 68.1 8.3 16.4 19.1 3.4 16.061 1.090 1.6 9.4 16.221 1.3 104.8 16.3 Calculated Solubility Sphere for Zein Solvent 1.2 12.4 16.5 76.2 17.4 55.7 84.6 17.1 10.9 8.5 103.2 87.2 10.0 105.2 7.2 14.3 91.043 1.7 8.2 17.8 59.3 12.4 16.8 19.0 123.004 1.4 68.3 20.2.5 23.3 15.228 V 87.891 0.2 89.169 1.8 26.8 5.4 19.0 8.1 δH 21.5 19.3-Butanediol Propylene glycol Diethylene glycol Ethylenediamine m-Cresol Aniline Dipropylene glycol 1.0 18.0 19.4 73.004 1.3 11.2 16.0 12.166 1.8 18.0 18.3 11.140 1.8 12.8 16.9 67.5 16.037 1.5 75.7 21.7 21.1 13.6 5.4 16.0 9.5 22.3-Benzenediol Benzyl alcohol Diethanolamine Phenol o-Methoxyphenol Furfuryl alcohol Hexamethylphosphoramide 3-Chloro-1-propanol 1.8 5.5 130.998 0.7 18.2 17.6 16.8 77.4 9.2 58.2 16.1 SOLUB RED 0.8 87.2 16.6 16.280 Hansen Solubility Parameters: A User’s Handbook TABLE 15.6 11.4 19.6 95.

one can deduce which chemicals are most prone to penetrate directly through w ood.236 1. for e xample. This generality then suggests special care is in order when handling w oodimpregnating chemicals. For example. with some side groups favoring the cellulose environment and others favoring the lignin environment. In addition to those pre viously mentioned.5. sinap yl alcohol.and ether -containing side groups on the hemicelluloses may not penetrate lignin as such b ut prefer to remain on its surface.0 D = 22. It is kno wn that pentachlorophenol.5 11. and some destruction of ultrastructure is possible.4 RAD. and p-coumaryl alcohol all have HSP which are on the boundary of the solubility sphere for solubility of De xtran (amorphous cellulose). The protective clothing chosen should ha ve HSP quite dif ferent from the HSP of the chemical in volved.7 115.11 (See also Chapter 18.8 H = 19.0 157. A sketch of these predicted relations is found in Figure 15.) Such a salt is hydrophilic and will collect w ater. The hemicellulose side groups with acetyl and ether linkages can be e xpected to orient toward the lower HSP lignin.4 P = 9.3 (CONTINUED) Calculated Solubility Sphere for Zein Solvent 2-Butanol 1-Octanol Ethyl lactate Methyl salicylate Zein δD 15. and untested solvents with a blank. coniferyl alcohol. If one considers the HSP for higher k etones. The same principles are valid. The relati vely high HSP for cellulose.0 δH 14.0 16. An important ef fect that may ha ve been o verlooked in the solubility of w ood and w ood components is that there are acid groups present in hemicelluloses.0 δP 5. as well as those of their side groups which contain –OH groups. for e xample. and ethers in Table 15. Hemicelluloses act lik e surfactants. readily diffuses into and through w ood specimens.5 12.0 16.6 8. b ut the hemicelluloses can form oriented config urations in connection with cellulose and with lignin. This indicates that the acetyl. The “*” points out those solv ents that do not conform e xactly with the correlation. Neither lignin nor hemicelluloses are compatible with cellulose in the usual sense.3 7. provides a proper ener getic environment for the backbones of hemicelluloses. that the solubility of crystalline polymers requires good solvents to have higher HSP than otherwise expected and that smaller molecular v olume is an adv antage. so their af finities indicate th y will seek the lower HSP domain of the lignin. probably findin a local (interf ace) site with closest possible HSP . Other types of predictions are possible from comparisons of the HSP correlations inTable 15.3 SOLUB RED 1. This gives an organic salt with high HSP. = 11.4.8 17.964 NO = 50 Note: Units are MPa1/2. This may lead to phase separation. and these can be neutralized by bases. bad solv ents with a 0. esters. it can be determined that all the solv ents dissolving lignin are also predicted to swell psoriasis scales.Hansen Solubility Parameters — Biological Materials 281 TABLE 15. A preferred pathway is where HSP are similar .9 12.0 129.1. This can be made possible by molecular rotation and orientation. These will dissolv e lignin. This can perhaps change with time and local en vironment. which also includes a lar ge number of –OH groups. This table contains the first entries in a much la ger database to show which solvents are most likely to affect proteins. The SOLUB column indicates good solv ents with a 1.9 FIT = 0.7 3. Still another question is ho w wood transports its o wn chemicals at various stages of the life of a tree. as discussed in Chapter 13.239 V 92.232 1. This is a clear e xample of self-association in nature.233 1. . it can be seen that none of these simple liquids will dissolve lignin. Included are chlorinated phenols and other w ood impre gnation materials. The monomers for lignin. This is an ef fect which is kno wn to ha ve caused blistering in coatings.

3 16.4 18.3 10.7 19.2 13.5 132.506 1.1 123.7 6.2 7.3 15.683 1.5 5.543 1.6 16.7 10.0 7.615 1.2 6.0 5.293 0.5 59.4-Dioxane Dipropylamine Dipropylene glycol Ethanol Ethanolamine Ethyl acetate Ethylbenzene 2-Ethyl-1-butanol Ethylene glycol Ethylene glycol monob utyl ether Ethylene glycol monoeth yl ether Ethylene glycol monoeth yl ether acetate δD 14.1 104.774 0.5 15.5 123.925 1.3 5.3 16.9 130.9 58.897 1.369 1.3 11.0 4.2 5.4 91.2 18.9 δP 8.0 15.4 18.0 16.096 0.0 17.3 0.9 91.3 9.5 74.158 1.0 16.212 1.3 60.3 2.9 0.8 15.5 3.0 10.6 4.6 19.831 0.0 16.0 2.8 14.8 3.7 106.0 97.4 0.2 10.0 0.169 1.7 13.5 15.5 16.0 5.5 5.5 17.5 15.3 5.582 1.6 2.4 21.0 8.833 1.2-Dichlorodiethyl ether Diethylamine Diethylene glycol Diethylene glycol monob utyl ether Diethylene glycol monometh yl ether Diethyl ether Diethyl sulfid Di(isobutyl) ketone Dimethylformamide Dimethyl sulfoxide 1.403 1.1 16.788 1.7 6.0 76.1 21.7 108.6 7.0 7.2 7.1 .2 7.0 104.0 118.761 1.3-Butanediol 1-Butanol Butyl acetate Butyl lactate Butyric acid gamma-Butyrolactone Butyronitrile Carbon disulfid Carbon tetrachloride Chlorobenzene 1-Chlorobutane Chloroform m-Cresol Cyclohexane Cyclohexanol Cyclohexanone Cyclohexylchloride Diacetone alcohol o-Dichlorobenzene 2.8 17.0 16.6 0.424 1.5 16.1 11.211 1.4 13.8 6.105 1.013 1.1 7.9 170.8 87.1 0.6 117.6 12.0 6.0 4.6 124.5 101.9 19.586 1.9 16.0 15.5 80.3 12.8 17.193 1.254 0.001 1.8 117.0 104.7 10.0 4.2 1.006 1.8 16.0 2.0 3.2 103.4 19.306 1.7 10.1 77.8 98.337 0.8 131.8 136.6 5.480 0.2 14.0 11.0 15.8 17.4 5.1 20.0 16.9 3.277 1.3 10.5 10.4 20.7 3.134 0.4 177.836 1.1 17.085 1.552 0.8 2.605 1.6 118.895 1.3 10.8 17.8 107.5 89.5 26.1 9.4 17.6 16.4 18.5 4.2 112.8 19.1 3.3 19.7 δH 13.1 102.0 89.0 110.0 71.0 15.2 17.3 20.002 1.6 8.5 8.7 136.060 1.5 149.1 0.3 14.631 0.6 12.044 1.7 104.1 3.210 1.8 15.8 18.917 1.4A Calculated Solubility Sphere for Lignin Solubility Solvent Acetic acid Acetic anhydride Acetone Acetonitrile Acetophenone Aniline Benzaldehyde Benzene 1-Bromonaphthalene 1.8 19.8 14.282 Hansen Solubility Parameters: A User’s Handbook TABLE 15.006 1.2 4.4 4.4 19.0 12.204 V 57.4 140.4 1.6 97.988 0.6 5.8 17.1 5.0 5.1 10.4 10.3 85.3 5.1 2.1 6.298 1.8 1.0 4.195 1.7 12.2 94.727 1.7 16.1 94.0 52.8 15.2 15.4 0.6 SOLUB 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 1* 0 0 0 1 1 0 0 1 1 1 0 0 0 1* 0 1 0 RED 1.1 2.0 15.2 55.

2 123.3 90.6 81.447 1.987 1.9 115.254 1.1 1.2 15.0 5.9 2.7 71.392 1.9 8.8 99.372 1. The data fit is ery good.0 16.3 D = 21.076 1.9 16.6 16.0 73.0 9.5 54.9 R = 13.9 12.500 1.2 125.3 0.9 7.8 1.7 The saturated solution of urea in water has found particular successes in the following examples.0 5.479 1.3 6.524 V 79.3 86.8 19.7 FIT = 0.0 δH 16.2 6. Perhaps the most interesting thing about this correlation is that it clearly sho ws that additions of urea to w ater will improve solubility for a variety of materials including proteins.1 9.1 H = 16.0 20.0 106.3 5.8 12.6 148.513 0.906 1.1 4.990 NO = 82 UREA Data for the HSP correlation for urea solubility in or ganic solvents are given in Table 15.5 115.8 9.0 16.1 13.1 4.9 4.6 δP 9.1 0.125 1. The saturated solution of urea and w ater is also the best physically acting solvent for whole.5 6.421 1. The f act of high HSP for urea/w ater mixtures has led to its use in man y v aried types of products.7 136.1 o SOLUB 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0* 1 0 0 0 0 0 0 0 RED 0.1 14.8 18.3 8.163 1.2 1.0 3.4 90.986 1.2 85.5 5.4 16.6 15.1 2.4 6.4 4.1-Trichloroethane Trichloroethylene Xylene Lignin δD 16.6 40.0 18.6 16.8 20.1 8.298 1.0 2.8 87.1 142.1 4.8 16. dried blood that the author could locate in a pre vious (unpublished) study.8 163.0 16.3 4.7 3.4 5.9 17.143 1.0 1.3 131.1 4.4 15.1 22. which is characteristic of a lo w molecular weight solid.5 18.3 3.0 7.944 0.117 1.2 4.7 169.0 16.5 73.3 8.4 15.237 1.9 7.4 4.6 75.1 12. This is the reason for the impro ved solubility discussed previously in connection with the destruction of h ydrophilic bonding in proteins.8 19.260 1.3 18.8 17.128 1.054 1.1 102.6 150.538 1.1 12.904 1.9 108.7 2.9 15.3 29.279 1.6 80.268 1.0 5.0 156.0 15. All of the parameters are rather high.8 12.1.1 15.516 1.0 15.0 2.Hansen Solubility Parameters — Biological Materials 283 TABLE 15. .6 18.8 18.4 8.464 0.8 127.1 23.0 5.0 17.3 15.274 1.6 5.2 17.3 1.3 3.411 1.1 4.9 P = 14.1.1 72.0 16.4 14.286 1.4A (CONTINUED) Calculated Solubility Sphere for Lignin Solubility Solvent Ethylene glycol monometh yl ether Furan Glycerol Hexane Isoamyl acetate Isobutyl isobutyrate Isooctyl alcohol Isophorone Mesityl oxide Methanol Methylal Methyl ethyl ketone Methyl isoamyl k etone Methyl isobutyl carbinol Methyl isobutyl ketone Morpholine Nitrobenzene Nitroethane Nitromethane 2-Nitropropane 1-Pentanol 1-Propanol Propylene carbonate Propylene glycol Pyridine Styrene Tetrahydrofuran Tetrahydronaphthalene Toluene 1.

8 10.5 19.013 V 76.9 17.9 76.2 16.0 16.3 13.6 16.0 18.7 18.4 5.599 0.3 7.2 18.6 16.0 16.2 9.7 10.0 18.918 0.5 115.5 116.727 0.2-Tetrachloroethane Tetramethylurea Diethylene glycol monoeth yl ether Morpholine Pyridine Ethanol Furfural Hexylene glycol Propylene glycol monophen yl ether Diethylene glycol monometh yl ether Ethylene glycol 2.821 0.0 22.6 8.8 16.1 17.3 9.6 171.7 16.4 16.0 5.766 0.2 19.0 55.2 9.9 8.8 12.4 16.8 19.0 16.878 0.001 1.988 0.7 17.2 5.3 175.1.0 106.2-Dichlorodiethyl ether Acetic anhydride Tricresyl phosphate Cyclohexanol δD 19.7 18.6 23.2 11.7 17.7 14.1 80.8 18.0 1 1 1 1 1 1 1 1 1 1 1 0* 1 1 1 1* 1 0 0 0 .4 11.2 19.986 0.4 18.7 13.0 59.0 11.2.609 0.0 5.2 13.0 17.2 9.5 12.8 118.2 17.0 17.0 89.2 123.0 18.6 86.6 19.7 109.006 1.4 16.9 58.5 130.0 71.2 18.2 94.9 8.1 5.006 1.5 84.1 8.5 16.8 117.2 9.8 16.9 7.4 20.774 0.866 0.0 105.4B Calculated Solubility Sphere for Lignin Solubility Solvent 2-Pyrolidone Succinic anhydride Dimethyl sulfone Dimethyl sulfoxide Hexamethylphosphoramide o-Methoxyphenol 1.2-Tetrabromoethane Ethylene glycol monoeth yl ether Quinoline Propylene glycol Triethylphosphate 1.8 11.0 19.8 11.758 0.0 104.5 15.7 9.3 21.7 15.836 0.788 0.8 18.4 8.6 18.6 26.4 18.6 10.965 0.925 0.6 15.7 7.000 1.7 96.6 12.5 10.5 SOLUB RED 0.2 11.902 0.8 16.906 0.9 87.1 17.2 17.8 12.5 13.5 316.5 83.3 7.284 Hansen Solubility Parameters: A User’s Handbook TABLE 15.761 0.919 0.3 5.3 4.944 0.6 12.1.008 1.839 0.913 0.833 0.865 0.2 79.9 6.665 0.4 17.792 0.1 9.0 9.8 100.3 15.0 143.4 16.1 12.6 19.3 7.8 6.831 0.3 14.895 0.4 δP 17.8 94.N-Dimethyl acetamide Trimethylphosphate Benzoic acid m-Cresol Methyl-2-pyrrolidone 1.2 4.6 16.2 21.2 118.4 10.0 108.9 103.3 77.4 66.6 21.0 18.3 10.987 0.5 67.1 18.800 0.769 0.6 11.998 1.968 0.0 17.0 5.2 10.4 16.1 92.3 21.4 19.8 8.002 1.2.5 10.1 δH 11.7 16.5 18.4 11.2 4.4 5.2 16.3 68.8 95.785 0.4 5.0 9.2 120.9 91.911 0.4 130.8 75.0 73.4 17.3-Butanediol Ethylene cyanohydrin Dimethyl formamide Diethylenetriamine Ethanolamine Diethanolamine Benzyl alcohol Furfuryl alcohol Dipropylene glycol gamma-Butyrolactone Diethylene glycol Phenol Ethylenediamine Allyl alcohol Triethyleneglycol 1.973 0.9 5.2 8.0 15.897 0.2 14.0 12.9 87.5 87.981 0.7 12.0 16.8 97.4 114.8 16.8 14.8 13.2 8.917 0.3 16.4 18.915 0.0 17.2 7.929 0.3-Butanediol Aniline 3-Chloro-1-propanol Ethylene glycol monometh yl ether N.9 19.989 0.2 14.9 8.5 10.5 68.

0 18.8 18.3 6.4 17.125 1.060 1.8 92.1 8.1 6.8 16.0 81.3 39.9 306.1 4.5 5.066 1.5 157.1 37.1 3.9 10.0 6.0 117.1 11.077 1.6 δH 17.149 1.7 16.5 5.3 5.5 124.1 7.0 163.3 12.0 11.171 V 75.4 19.7 73.112 1.034 1.158 1.8 150.3 8.0 17.160 1.6 16.0 204.6 3.5 5.7 6.9 5.4 4.9 4.7 6.1 8.2 146.2 17.6 17.7 15.9 8.8 191.028 1.013 1.5 8.169 1.5 6.5 12.096 1.4 12.1 13.4 15.0 16.8 40.8 7.0 101.7 10.149 1.113 1.3 16.4 4.0 131.2 5.160 1.136 1.0 18.021 1.121 1.6 4.0 129.2 19.5 76.3 121.0 20.4 14.163 1.2 10.4 11.4B (CONTINUED) Calculated Solubility Sphere for Lignin Solubility Solvent 1-Propanol Propylene carbonate Triethyanolamine Nonyl phenoxy ethanol Methyl salicylate Dimethyl phthalate Ethyl lactate Benzaldehyde Trifluoroacetic aci Di-(2-Chloro-isopropyl) ether Nitrobenzene Ethylene dibromide 1-Butanol 2-Propanol Methanol Bromoform Diacetone alcohol Ethylene carbonate Epichlorohydrin 2-Butanol Acetophenone Diethylene glycol monob utyl ether Methylene dichloride Benzyl butyl phthalate Acrylonitrile Formic acid Isophorone 1-Octanol Glycerol Formamide Ethylene glycol monob utyl ether Ethylene dichloride 1-Pentanol 1-Nitropropane Bromobenzene Ethylene glycol monometh yl ether acetate Ethyl cinnamate Propylene glycol monometh yl ether Diethyl phthalate Diethyl sulfate Butyl lactate Diethylene glycol he xyl ether Propylene glycol monoeth yl ether Propylamine Tetrahydrofuran 2-Octanol 2-Ethyl-1-butanol Isobutyl alcohol 2-Methyl-1-propanol 1-Decanol δD 16.0 6.3 19.0 22.105 1.044 1.3 4.0 67.2 8.0 17.7 4.9 92.3 4.128 1.9 5.1 12.8 15.6 7.0 10.0 15.6 88.6 5.125 1.4 170.4 15.9 8.1 10.8 19.5 4.090 1.6 166.9 12.8 16.085 1.1 17.9 16.7 159.088 1.7 2.144 1.7 14.0 102.154 1.5 149.4 11.116 1.076 1.1 3.044 1.0 20.6 83.6 63.9 15.144 1.4 22.8 16.3 11.4 108.2 5.1 26.8 198.0 91.2 3.8 5.1 15.0 15.018 1.0 115.1 23.2 16.7 8.0 10.149 1.8 131.6 15.1 15.5 δP 6.7 5.Hansen Solubility Parameters — Biological Materials 285 TABLE 15.6 6.0 15.6 79.1 123.054 1.2 66.5 7.5 8.015 1.9 29.6 20.026 1.8 93.6 14.169 1.8 16.6 8.059 1.9 18.4 15.5 74.0 12.052 1.2 6.5 4.9 16.129 1.3 115.4 105.2 16.134 1.0 19.6 7.143 1.3 16.3 4.3 11.1 11.8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .4 10.145 1.3 9.3 5.6 16.5 15.2 8.0 SOLUB 0 0 RED 1.0 13.7 87.7 87.6 17.6 4.1 21.6 7.6 15.2 85.169 1.8 19.6 19.5 15.0 133.7 16.095 1.0 15.2 21.162 1.5 3.7 10.2 275.2 92.4 9.163 1.8 15.0 79.1 10.0 19.0 19.

0 8.210 1.1 9.274 1.3 90.5 5.8 7.1 5.279 1.0 3.8 5.5 5.3 7.1 8.7 15.4 19.5 54.2 113.7 8.0 57.3 5.7 7.0 5.9 4.8 17.7 74.9 79.295 1.0 8.0 2.0 18.3 9.3 7.286 1.0 15.0 3.3 11.3 6.4 7.3 266.8 17.2 SOLUB RED 1.1 4.4 104.8 14.1 6.3 80.1 13.8 14.2 87.301 1.1 6.7 3.3 δH 9.257 1.224 1.8 85.7 17.8 17.5 5.195 1.0 6.1 9.9 3.6 3.247 1.5 15.8 15.1 80.6 5.0 20.5 86.1 133.3 8.5 8.236 1.0 15.5 18.4 17.4 15.0 231.254 1.1 136.9 90.2 17.1 4.9 71.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .8 16.9 2.1 11.211 1.1 6.1 7.3-pentanediol monoisobutyrate Methylene diiodide Butylamine Mesityl oxide 1.212 1.212 1.298 1.3 11.226 1.6 17.1 112.5 3.254 1.7 208.250 1.200 1.4B (CONTINUED) Calculated Solubility Sphere for Lignin Solubility Solvent Propylene glycol monoprop yl ether Dibutyl phthalate Dipropylene glycol meth yl ether Cyclohexanone Acetic acid Ethyl formate Trichlorobiphenyl Anisole Ethylene glycol monoeth yl ether acetate o-Dichlorobenzene 1.0 119.7 6.5 140.3 18.0 103.0 15.8 3.9 18.4 16.202 1.6 79.267 1.3 106.2 16.6 3.2 115.5 5.3 19.3 6.0 12.0 373.0 17.3 15.1 4.9 7.8 12.3 6.1 12.3 16.3 4.267 1.1-Dichloroethylene Propionitrile Methyl ethyl ketone Acetonitrile Methyl isobutyl carbinol Ethanethiol Methyl methacrylate Nitromethane Chloroform Diethylene glycol b utyl ether acetate Butyronitrile Trichloroethylene Cyclohexylamine Methyl acrylate Ethyl acetate δD 15.1 7.4 80.8 98.193 1.8 10.5 4.3 15.232 1.5 6.0 70.1 6.212 1.2 5.1 2.5 15.1 3.0 111.5 5.1 9.5 14.204 1.8 δP 7.2 11.8 113.1-Dimethyl hydrazine bis-(m-Phenoxyphenyl) ether 2.293 1.286 1.2 16.3 1.7 156.1 4.1 4.5 16.4 4.247 1.9 19.1 52.8 16.260 1.4-Dioxane Acetone Nonyl phenol Acetaldehyde 1.4-Trimethyl 1.2 187.6 156.8 15.2 6.2 15.196 1.1 17.241 1.1 76.260 1.0 15.3 7.1 3.5 99.9 12.9 4.0 115.0 57.277 1.3 5.2 4.4-Pentanedione Ethyl chloroformate Dibenzyl ether 2-Ethyl hexanol Isooctyl alcohol Tetrachloroethylene 2-(Diethylamino) ethanol Benzyl chloride Benzonitrile Ethyl bromide Nitroethane 1-Bromonaphthalene Naphthalene 2-Nitropropane Methyl acetate 2.5 16.271 1.263 1.1 5.0 102.1 5.0 9.302 1.2 74.4 4.5 17.0 157.1 15.1 4.232 1.3 19.5 15.0 5.6 76.192 1.2 5.0 14.286 Hansen Solubility Parameters: A User’s Handbook TABLE 15.1 95.180 1.6 127.6 9.2 19.1 12.4 5.237 1.4 3.213 1.5 19.3 5.0 15.268 1.6 101.306 V 130.280 1.6 192.3 15.7 227.174 1.9 14.7 10.237 1.2.0 10.0 8.7 6.5 17.298 1.273 1.2 15.5 8.

1 7.6 1.0 5.7 3.5 3.7 14.486 1.3 8.1 6.0 142.336 1.1 3.0 15.0 110.392 1.4 133.535 1.3 16.5 10.330 1.1 4.3 84.8 139.8 3.1 7.3 15.6 148.0 5.0 99.1 4.318 1.5 16.465 1.6 2.1 7.1 15.5 377.3 3.356 1.1 2.8 167.7 4.0 171.432 1.5 4.7 4.0 2.9 4.0 242.0 16.366 1.3 1.0 6.450 1.8 83.1 7.6 156.3 15.0 2.3 6.5 0.1 5.428 1.8 4.1-Trichloroethane 1.5 55.7 5.1 4.1 75.317 1.524 1.7 2.8 104.1 15.9 15.403 1.3 3.0 3.3 20.470 1.0 2.8 16.6 14.1 2.4 δH 9.1 4.6 125.512 1.8 3.2 17.1 6.1 2.1 1.4 88.1 3.346 1.0 3.0 3.7 4.0 143.5 1.4 169.0 17.0 16.5 326.506 1.7 15.1 3.1 123.417 1.480 1.8 345.3 1.0 16.2 132.8 15.1 70.502 1.0 4.447 1.6 15.424 1.0 2.0 123.7 5.2 5.6 7.464 1.8 6.9 15.485 1.500 1.8 7.9 9.427 1.313 1.411 1.9 8.2 15.6 4.2 6.Hansen Solubility Parameters — Biological Materials 287 TABLE 15.7 3.8 16.1 10.451 1.2 15.0 15.9 6.8 65.516 1.1 8.1 72.3 5.1 17.1 4.4 177.5 6.1 10.0 0.0 159.8 132.8 16.3 14.0 88.6 15.4 108.2 4.374 1.372 1.0 6.538 V 132.8 154.408 1.8 16.2 138.8 2.0 115.8 9.369 1.3 16.5 16.0 2.5 133.331 1.8 14.4B (CONTINUED) Calculated Solubility Sphere for Lignin Solubility Solvent Propylene glycol monoisob utyl ether Propylene glycol monob utyl ether Di(2-Methoxyethyl) ether Ethylene glycol b utyl ether acetate 1-Methyl naphthalene Bromochloromethane Butyric acid Diethyl ketone Ethyl acrylate Tributyl phosphate Diethyl carbonate Chlorobenzene Furan Dioctyl phthalate Di-iso-butyl carbinol Methacrylonitrile Tetrahydronaphthalene Butyl acrylate Butyl acetate Stearic acid Methyl isoamyl k etone Ethyl butyl ketone Octanoic acid Styrene Cyclohexylchloride Amyl acetate Butyraldehyde sec-Butyl acetate Ethyl amyl k etone Isoamyl acetate Biphenyl Dichloromonoflouromethan Propyl chloride Methyl butyl ketone Methyl isobutyl ketone Methyl amyl acetate Isobutyl acetate Methyl chloride Methylal Di(isobutyl) ketone Ethyl chloride Tridecyl alcohol 1.1 6.0 142.372 1.395 1.8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .9 1.0 102.1 18.3 15.0 106.421 1.5 121.3 16.5 8.351 1.9 3.0 177.4 2.3 21.3 5.3 18.2 163.389 1.6 17.2 15.8 6.5 3.464 1.3 15.434 1.3 3.1 1.462 1.337 1.6 4.3 19.6 14.7 6.418 1.7 15.0 3.0 δP 4.6 19.3 15.3 5.2 3.0 106.0 SOLUB RED 1.8 5.3 316.8 15.0 5.1 8.473 1.1-Dichlorethane 1-Chlorobutane o-Xylene Isobutyl isobutyrate Xylene Oleyl alcohol Toluene δD 15.6 4.1.0 7.0 148.6 17.6 118.9 136.7 6.479 1.0 121.5 4.3 3.4 15.


Hansen Solubility Parameters: A User’s Handbook

TABLE 15.4B (CONTINUED) Calculated Solubility Sphere for Lignin Solubility
Solvent Diethyl sulfid Diethyl amine Benzene Naphtha.high-flas Carbon disulfid Oleic acid Diethyl ether Triethylene glycol monoole yl ether Ethylbenzene Methyl oleate Dipropylamine Dibutyl stearate Triethylamine Trimethylbenzene Isopropyl palmitate Dibutyl sebacate cis-Decahydronaphthalene para-Diethyl benzene Mesitylene Carbon tetrachloride trans-Decahydronaphthalene Chlorodiflouromethan Cyclohexane Methyl cyclohexane Eicosane Trichlorofluoromethan Hexadecane Dodecane Mineral spirits Decane Nonane Octane 1,1,2-Trichlorotrifluoroethan Heptane Hexane Pentane Tetraethylorthosilicate Butane 2,2,2,4-Trimethylpentane Isopentane 1,2-Dichlorotetrafluoroethan Dichlorodifluoromethan Water Perfluoro(dimet ylcyclohexane) Perfluoromet ylcyclohexane Perfluoroheptan Bromotrifluoromethan Lignin δD 16.8 14.9 18.4 17.9 20.5 14.3 14.5 13.3 17.8 14.5 5.3 14.5 17.8 17.8 14.3 13.9 18.8 18.0 18.0 17.8 18.0 12.3 16.8 16.0 16.5 15.3 16.3 16.0 15.8 15.7 15.7 15.5 14.7 15.3 14.9 14.5 13.9 14.1 14.1 13.7 12.6 12.3 15.5 12.4 12.4 12.0 9.6 δP 3.1 2.3 0.0 0.7 0.0 3.1 2.9 3.1 0.6 3.9 1.4 3.7 0.4 0.4 3.9 4.5 0.0 0.0 0.0 0.0 0.0 6.3 0.0 0.0 0.0 2.0 0.0 0.0 0.1 0.0 0.0 0.0 1.6 0.0 0.0 0.0 0.4 0.0 0.0 0.0 1.8 2.0 16.0 0.0 0.0 0.0 2.4 δH 2.0 6.1 2.0 1.8 0.6 5.5 5.1 8.4 1.4 3.7 4.1 3.5 1.0 1.0 3.7 4.1 0.0 0.6 0.6 0.6 0.0 5.7 0.2 1.0 0.0 0.0 0.0 0.0 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0.0 0.0 0.0 0.0 0.0 42.3 0.0 0.0 0.0 0.0 SOLUB 0 0 0 0 0 0 0 RED 1.543 1.552 1.582 1.585 1.586 1.603 1.605 1.614 1.615 1.628 1.631 1.643 1.645 1.645 1.647 1.652 1.669 1.673 1.673 1.683 1.704 1.719 1.761 1.774 1.790 1.797 1.803 1.823 1.823 1.844 1.844 1.858 1.860 1.873 1.904 1.936 1.944 1.969 1.969 2.003 2.042 2.065 2.081 2.122 2.122 2.161 2.340 V 107.4 103.2 89.4 181.8 60.0 320.0 104.8 418.5 123.1 340.0 136.9 382.0 138.6 133.6 330.0 339.0 156.9 156.9 139.8 97.1 156.9 72.9 108.7 128.3 359.8 92.8 294.1 228.6 125.0 195.9 179.7 163.5 119.2 147.4 131.6 116.2 224.0 101.4 166.1 117.4 117.6 92.3 18.0 217.4 196.0 227.3 97.0




D = 21.9 P = 14.1 H = 16.9 RAD. = 13.7 FIT = 0.990 NO = 82

Hansen Solubility Parameters — Biological Materials


--M1β4M1β4M1β4G1β4M1β4G1β4M1β4M1β4G– 2 332 3 6 Ac Ac β Ac Ac α (LIGN) 1 (LIGN) 1 M Ga (CELL) (CELL)
FIGURE 15.5 Expected generalized sk etch of the configuration of cellulose, hemicelluloses, and lignin i wood cell w alls. See te xt or Reference 6 for further details. The sketch is for glucomannan. M is mannose monomer; G is glucose monomer; Ga is galactose monomer; Ac is an acetyl group; (LIGN) is a region similar in HSP to lignin (or acetal etc.); (CELL) is a re gion similar in HSP to cellulose, being an y of cellulose, hemicellulose backbone, or hemicellulose side chain with an alcohol group (M, Ga).

1. Lithographic stones were previously conditioned to make them more receptive to ink by application of this liquid to change wetting beha vior. 2. The saturated solution of urea and w ater, which swells and softens w ood, has been used to give wood fl xibility so that it can easily be formed. 3. It has been used by Eskimos to soften seal skins by swelling and softening them. A similar application in Me xico involves curing leather . This application probably originated in prehistoric times. 4. It has been used to impro ve the fl w of house paints on cold days or when no other source of liquid has been available (such as on a scaffold), as it is a good solvent miscible in many paints. 5. It is reported to ha ve been used to set hair , as it also softens and swells it. 6. It w as used in the early manuf acture of gunpo wder as a dispersion medium during grinding because of impro ved wetting for the po wder. 7. Amazonian Indians used this liquid to coagulate late x prior to sale and shipment. This was practiced particularly during World War II. Other unspecified and undocumented uses include those possible because the liquid has th ability to soften human skin, thus allo wing easier transport of medicinal chemicals into the body . Urea itself has HSP v ery close to those of sug ar and proteins. As all of these are biocompatible materials, it is clear that the incorporation of significant numbers of urea groups in, for xample, polyurethane polymers or other products, can greatly enhance biocompatibility .

Water has been discussed in detail in Chapter 1. Briefly stated, one can use the HSP for ater or a correlation for w ater solubility to get a general e xplanation for observ ed phenomena. Accurate calculation of the HSP for solv ent–water mixtures cannot be expected because of the irregularities of w ater associating with itself, the solv ent, and a potential solute. Lindenfors 12 described the association of two molecules of water with one molecule of dimethyl sulfoxide, a solvent frequently mentioned in connection with biological systems. A simplistic approach based on the ratio of δH for w ater as a single molecule vs. that in the correlation(s) for w ater solubility suggests that (42.3/16.5)2 or about six water molecules are linked by hydrogen bonding into some type of entity. Various structures for assemblies of w ater molecules ha ve been discussed in the literature. The clusters with six water molecules are among the more probable ones.13 The data on water solubility used in the HSP correlations are reported by Wallström and Sv enningsen.14


Hansen Solubility Parameters: A User’s Handbook

Surface mobility allows given segments of molecules to orient at surfaces in a direction where their HSP match more closely . The surfaces of h ydrophobic polymers (peat moss) can become h ydrophilic when contacted with w ater. One can speculate as to wh y this occurs. One possibility is that this phenomenon conserves water within the structure. Whenever water is present on an otherwise hydrophobic surf ace, it can become h ydrophilic if the surf ace molecules can rotate or mo ve hydrophilic entities toward the water. This allows the water to spontaneously spread and potentially enter the structure if there are suitable passages. When this is accomplished, and contact with water ceases, the surface dries and becomes h ydrophobic once more. The molecules rotate with a lo wer energy moiety toward the air. This hydrophobic surface helps prevent evaporation of water, as water is not particularly soluble in it, and the h ydrophilic segments oriented to ward the interior of the structure will help bind the w ater where it is. The basis of the orientation ef fects described earlier for hemicelluloses is another e xample of orientation to ward re gions where HSP matches better . These phenomena are also discussed in Chapter 18. It is also appropriate to repeat that solv ent quality has a great deal to do with pigment dispersion stability , in that the adsorbed stabilizing polymer should remain on the pigment surf ace. A solvent which is too good can remo ve it. This is discussed in detail in Chapter 5. The implication of these e xamples is that solv ent quality is v ery important for the orientation of molecules at interf aces. A change in solv ent quality can easily lead to a change in the configu ration of molecules at surfaces. It is not surprising that Nature has used this to advantage in various ways.

It has been found that anthracene units appended to a single scre w-sense helical polyguanidine changed orientation when the temperature was increased beyond 38.5°C.15 The configuration foun above 38.5 °C was the same as that found in tetrah ydrofurane. At temperatures lower than this, the orientation of the appended anthracene was that found in toluene.A mixture of tetrahydrofurane/toluene equal to 90/10 v ol% approximated the conditions at the critical temperature. F or those who have diligently read this handbook, it w ould appear ob vious that it is the cohesi ve energy density just above or just belo w the critical temperature that controls the structure. More specifically it i the set of HSP values that do this, as these reflect the mix of sources of the cohes ve energy density according to Equation 1.6 to Equation 1.8. There is also massi ve e vidence sho wing that the interactions can be interpreted as the dif ference in HSP using Equation 1.9. It is well kno wn that solubility limits can be passed by lo wering the temperature in some cases and by increasing it in other cases. When the cohesi ve energy density of the solv ent is higher than that of the polymer , solvency increases with increased temperature. When the cohesive energy density of the solvent is lower than that of the polymer, solvency decreases with increases in temperature. This is discussed in Chapter 2 and has been thoroughly treated by P atterson.16,17 In the present case the HSP of toluene are comparable to those of anthracene whereas tetrah ydrofuran has much higher v alues. δD, δP, and δH equal to 18.7, 4.1, and 3.3 for anthracene have been reported by a multiple regression technique based on its solubility in a lar ge number of solv ents.18 The corresponding v alues are 18.0, 1.4, and 2.0 for toluene and 16.8, 5.7, and 8.0 for tetrah ydrofurane. Thus increasing the temperature will increase the solv ency in tetrah ydrofurane to the point where it becomes able to cause the appended anthracene to adopt the same orientation as it has in toluene. As toluene has lower HSP than anthracene, the solv ency will decrease with increases in temperature. Extending this w ay of thinking to the problem of mo ving very large biological molecules — while they are being assembled, for e xample — is presumably one of controlling the local solubility . When the molecule is locally able to reside in the surrounding fluid, it can m ve much more readily

Hansen Solubility Parameters — Biological Materials


than when it is not. An insoluble molecule or molecular se gment will adsorb at a location where the ener gies (HSP) match, and where the geometry is also accommodating. This is most often called hydrogen bonding, b ut it must be all three (or more) types of cohesi ve ener gy that are collectively acti ve. The molecule or molecular se gment can be remo ved ag ain when it and the surrounding liquid ha ve a f avorable energy relation.

Many materials of biological significance h ve been assigned HSP based on their interaction with a large number of solv ents whose HSP are kno wn. A correlation for solv ent effects on DN A has ranked the extent of these effects for different solvents in essentially the same order as that reported in an older study .9 This correlation for DN A can presumably be impro ved by additional data, b ut still reflects the magnitudes of the types of ene gies that are in volved in forming/destro ying the double helices. The δD;δP;δH found for DN A are 19.0;20.0;11.0, all in MP a1/2. This clearly sho ws that hydrogen bonding is by far the smallest of the energies involved in the noncovalent interactions that determine the DN A structure. A HSP correlation has been used to find predictably syne gistic solvent mixtures where tw o nonsolvents dissolve cholesterol when mix ed. The ethanol/aliphatic h ydrocarbon synergistic mixture is discussed as being of particular interest to the f ate of cholesterol in lipid layers. The HSP of chlorophyll and lignin are quite similar , indicating they will be compatible with v ery much the same kind of surroundings. The physical interrelationships for wood chemicals and wood polymers (lignin, hemicelluloses, and cellulose) are discussed. The side chains on hemicelluloses which contain alcohol groups and the hemicellulose backbone will be most compatible with cellulose and will orient to ward this. The hemicellulose side chains without alcohol groups (acetal, acid) are closer in HSP to lignin and will orient in this direction. The acetal side chains actually ha ve lower HSP than will dissolve lignin, for which reason they are expected to lie on the surface of the lignin or perhaps penetrate slightly into the lignin at v ery special local points where the HSP match is better than the a verage values seen o ver the lignin molecule as a whole. Molecular design of molecules or structures that change conformation with slight changes in the cohesive energy characteristics of given continuous media seems possible using HSP concepts. The changes are caused by preferred orientation of se gments of one conformation to ward the continuous phase, where its HSP match better , thus reducing the free ener gy of the system. If the cohesive energy characteristics of the continuous media change in a direction that no longer f avors this orientation, the molecule will change configuration to one where the free ene gy is lower. The attraction of the se gments not oriented to ward the continuous phase to neighboring molecules is commonly called hydrogen bonding in proteins and similar materials. This attraction is caused collectively by all the types of ener gy involved through the pre vailing difference in HSP and is a result of insolubility (rejection by) the continuous media. Geometrical considerations are clearly also a major f actor in addition to the cohesi ve energy density focused upon here. HSP analyses of relative affinities can be applied to a la ge number of other biological materials and may pro vide insights into relationships which are not readily ob vious or cannot be studied otherwise. The best situation is where the materials in question can be tested directly , otherwise the calculation procedures described in Chapter 1 can be used with some loss of reliability in the predictions.

1. Hansen, C.M., The three dimensional solubility parameter — k ey to paint component af finities I Solvents, plasticizers, polymers, and resins, J. Paint Technol., 39(505), 104–117, 1967.


Hansen Solubility Parameters: A User’s Handbook

2. Hansen, C.M. and Andersen, B.H., The affinities of o ganic solvents in biological systems, Am. Ind. Hyg. Assoc. J., 49(6), 301–308, 1988. 3. Hansen, C.M., The universality of the solubility parameter , Ind. Eng. Chem. Prod. Res. Dev., 8(1), 2–11, 1969. 4. Ursin, C., Hansen, C.M., Van Dyk, J.W., Jensen, P.O., Christensen, I.J., and Ebbehoej, J., Permeability of commercial solv ents through living human skin, Am. Ind. Hyg. Assoc. J., 56, 651–660, 1995. 5. Hansen, C.M., 25 years with the solubility parameter (25 År med Opløselighedsparametrene, in Danish), Dan. Kemi, 73(8), 18–22, 1992. 6. Hansen, C.M. and Björkman, A., The ultrastructure of w ood from a solubility parameter point of view, Holzforschung, 52(4), 335–344, 1998. 7. Hansen, C.M., Solv ents for coatings, Chem. Technol., 2(9), 547–553, 1972. 8. Blake, R.D. and Delcourt, S.G., Thermodynamic effects of formamide on DN A stability, Nucl. Acid Res., 24, 2095–2103, 1996. 9. Ts’o, P.O.P., Helmkamp, G.K., and Sander, C., Interaction of nucleosides and related compounds with nucleic acids as indicated by the change of helix-coil transition temperature, Proc. Natl. Acad. Sci. U S A, 48, 686–698, 1962. 10. Hansen, C.M., Cohesion energy parameters applied to surface phenomena, Handbook of Surface and Colloid Chemistry, Birdi, K.S., Ed., CRC Press, Boca Raton, FL, 1997, pp. 313–332. 11. Hansen, C.M., Some aspects of acid/base interactions (Einige Aspekte der Säure/Base-W echselwirkung, in German), Farbe und Lack, 7, 595–598, 1977. 12. Lindenfors, S., Solubility of cellulose ethers (Löslichkeit der Celluloseeäther, in German), Das Papier, 21, 65–69, 1967. 13. Gregory, J.K., Clary, D.C., Liu, K., Bro wn, M.G., and Saykally , R.J., The Water Dipole Moment in Water Clusters, Science, 275, 1997, pp. 814–817. 14. Wallström, E. and Svenningsen, I., Handbook of Solvent Properties, Report T1-84, Scandinavian Paint and Printing Ink Research Institute, Hoersholm, Denmark, 1984. 15. Tang, H.-Z., No vak, B.M., He, J., and Pola varapu, P.L., A thermal and solv ocontrollable cylindrical nanoshutter based on a single scre w-sense helical polyguanidine, Angew. Chem. Int. Ed., 44, 7298–7301, 2005. 16. Patterson, D. and Delmas, G., Ne w aspects of polymer solution thermodynamics, Off. Dig. Fed. Soc. Paint Technol., 34(450), 677–692, 1962. 17. Delmas, D., Patterson, D., and Somcynsky, T., Thermodynamics of polyisobutylene-n-alkane systems, J. Polym. Sci., 57, 79–98, 1962. 18. Wu, P.L., Beerbower, A., and Martin, A., Extended Hansen approach: calculating partial solubility parameters of solid solutes, J. Pharm. Sci., 71(11), 1285–1287 (1982).

and Diffusion in 16 Absorption Polymers
Charles M. Hansen
Predicting whether or not a gi ven chemical will attack a gi ven polymer is important. Hansen solubility parameters (HSP) ha ve been used for this purpose as discussed else where in this book. Consideration of the absorption and dif fusion of the chemical in the polymer is often required in addition to HSP in order to mak e reliable predictions, however. This has been discussed in particular in Chapter 12 through Chapter 14, where chemical resistance, barrier properties, and en vironmental stress cracking are treated in detail. Chemicals with smaller and more linear molecules absorb and diffuse more readily than those with lar ger and more b ulky structures. Surf ace resistance to absorption is sometimes so dominating that absorption does not occur in some cases, e ven though this might be e xpected based on simple HSP considerations. This chapter e xamines surface resistances in connection with absorption and dif fusion in polymers in order to help impro ve understanding of these f actors and to emphasize the necessity of simultaneous consideration of surf ace resistance when absorption rates and dif fusion within the bulk of the polymer itself are of interest. Methods to measure surf ace resistance and concentration-dependent dif fusion coef ficients ar discussed. Solving the diffusion equation with simultaneous consideration of surface resistance and with a concentration dependent dif fusion correctly models absorption, desorption, film formatio by solvent evaporation, and v arious forms of so-called anomalous dif fusion such as “time-dependent,” Case II, and Super Case II. Surf ace phenomena such as surf ace resistance to absorption deserve far more attention than has been gi ven in the past.

(Please note that these are dif ferent from those used in the other chapters.) A B C CA Cs D D0 D1 D2 Dapp Dav Dmax Dv F FM FB Fa Minimum cross-sectional area of molecule in Equation 16.20 Ratio of diffusion resistance to surf ace resistance. See Equation 16.13 Dimensionless concentration. See Equation 16.5 Concentration at break in curv e in Figure 16.1 Dimensionless surface concentration Diffusion coefficient. Preferred units are c 2/s Diffusion coefficient at zero concentration or l west concentration in an e xperiment Diffusion coefficient on xposed side of fil Diffusion coefficient on xit side of fil Apparent diffusion coefficien Average diffusion coefficien Maximum diffusion coefficient in an xperiment Increase in the dif fusion coefficient ver zero conditions for a gi ven situation Mass flu Correction factor for concentration dependent dif fusion in Equation 16.11 Correction factor for surf ace resistance in Equation 16.11 Correction factor for concentration dependent dif fusion in absorption e xperiments


Hansen Solubility Parameters: A User’s Handbook

Fd L Mt M∞ P Papp P∞ R Rd Ri Rs T X Vf Vt V2 b c cs c0 c1 c2 c∞ h hav k k2 l t t1/2 w x Δp ΦR

Correction factor for concentration dependent dif fusion in desorption e xperiments Film thickness Absorbed mass at time t Mass at equilibrium conditions Permeation coefficien Apparent permeation coefficien True permeation coefficien Radius of c ylindrical sample in Equation 16.24 Resistance to mass transport from dif fusion Resistances to permeation from sources 1, 2, 3, etc., in Equation 16.17 to 16.19 Resistance to mass transport by surf ace resistance(s) Dimensionless time given by Equation 16.3 Dimensionless distance Volume fraction of dif fusing material Constant in Figure 16.4 used to find local di fusion coefficients at Vf greater than 0.20 Diffusion coefficients at Vf equal 0.20 relative to diffusion coefficient at zero concentra tion (Figure 16.1 and Figure 16.4) Thickness of c ylindrical sample in Equation 16.24 Concentration of diffusing material Surface concentration Initial concentration Concentration on e xposed side of sample Concentration on e xit side of sample Concentration at equilibrium conditions Surface mass transfer coef ficien Average surface mass transfer coef ficien Constant in Equation 16.6 Constant in Equation 16.20 Length of sample in Equation 16.23 Time Time required to absorb (or desorb) one half of the equilibrium amount Width of sample in Equation 16.23 Distance Pressure difference of diffusing material across membrane Local concentration is Figure 16.6

Chapters 12 through 14 ha ve dealt with chemical resistance, en vironmental stress cracking, and barrier polymers, respectively. Absorption and diffusion of the chemicals into polymers are impor tant in each of these. F actors af fecting absorption and dif fusion in polymers are therefore of considerable importance and must frequently be included along with the Hansen solubility parameters (HSP) to mak e correlations and predictions for these phenomena. This chapter emphasizes the importance of surface resistance for absorption in polymers as this has been largely overlooked in the literature. The focus of much of the rele vant literature has been on anomalous dif fusion, but even in this context the influence of sur ace resistance has largely been neglected. Surface resistance can delay or pre vent the absorption of solv ents that should absorb readily based on simple HSP considerations. This could lead to a f alse sense of security based on short time testing only . Absorption requires some de gree of solubility . Therefore, equilibrium absorption can be expected to correlate with HSP. In studies of the absorption of a chemical into a polymer , its surface concentration is usually assumed to reach the equilibrium v alue immediately. As discussed in the

Absorption and Diffusion in Polymers


following, this is not al ways true. Whatever the rate at which the surf ace concentration increases, absorption will proceed according to the la ws of dif fusion, Fick’s First La w, Equation 16.1, and Fick’s Second La w, Equation 16.2. The latter is often called the dif fusion equation. Its deri vation can be found in Crank. 1 For the sak e of simplicity, these equations are gi ven here for dif fusion in one dimension (x) only . For a constant dif fusion coefficient F = - D 0(∂c/∂x) ∂c/∂t = ∂/∂x (D 0∂c/∂x) (16.1) (16.2)

The diffusion equation is deri ved in a v ery general w ay and also accounts for concentration dependent diffusion coefficients when ver this is encountered. F is the mass flux, 0 is the (constant) diffusion coefficient or the di fusion coefficient at the l west concentration if there is concentration dependence (see below), c is the local concentration, x is the distance in the x dimension, and t is the time. The solutions to the diffusion equation given in this chapter use the dry film thickness a reference. This is because it is f ar simpler to k eep track of what is going on instead of continually adjusting local film thickness as a function of the amount of sol ent present. When a dif fusing solvent is present, for example, then the actual local thickness should be increased proportionately according to its local v olume fraction. The use of dimensionless v ariables to solv e these equations mak es the solutions more useful by making them applicable to all v alues of the combined v ariables. For this purpose the follo wing are defined Dimensionless time: T = D 0t/L2 Dimensionless distance: X = x/L Dimensionless concentration: C = (c – c 0)/(c∞ – c 0) (16.5) (16.4) (16.3)

L is the thickness of the plane sheet being considered. c 0 is the initial uniform concentration in the film. The (local) dimensionless concentrations rise from 0 to 1.0 in an absorption experiment where equilibrium absorption, c ∞, is finally obtained. For the sak e of completeness the predicted and e xperimentally confirmed xponential dependence of the dif fusion coefficient on concentration, D(c), will be introduced at this early stage. A more detailed discussion of its significance is g ven in a special section belo w. D(c) = D 0ekc = D 0Dv (16.6)

“k” is a constant that is v alid up to a gi ven concentration as discussed belo w. D v is the increase in the diffusion coefficient ver that at zero conditions for a gi ven concentration. At the maximum concentration this becomes D max, and for a constant dif fusion coefficient it is 1.0 Using these variables, Equation 16.2 can be re written in dimensionless form as Equation 16.7. The dimensionless diffusion equation for an e xponential diffusion coefficient is ∂C/∂T = ∂/∂X (D v∂C/∂X) (16.7)


Hansen Solubility Parameters: A User’s Handbook

The earliest edition of Crank’ s monumental work1 has the advantage of many graphical solutions to the dimensionless dif fusion equation being presented with plots that can be used with a high degree of accurac y. These plots include numerous reference lines not found in later editions.

In the absence of significant sur ace resistances, solving Equation 16.1 (or Equation 16.7) for steady state permeation in one direction only for a constant dif fusion coefficient g ves Equation 16.8. F = –D 0(c1 – c 2)L (16.8)

The surface concentration on the exposed side, c1, is usually assumed to be c∞, and c2 is usually assumed to be zero. Initially , c 1 will be less than c ∞ when there is a significant sur ace resistance on the e xposed side, as discussed in the follo wing. Likewise, c 2 will be greater than c 0 if there is a significant sur ace resistance on the lo w concentration side of the film. The preferred units for these quantities are F in g/(cm 2×s), L in cm, c in g/cm 3, and D 0 in cm 2/s. It can be seen from Equation 16.8 that the surface concentration on the exposed side determines the concentration gradient over the film, assuming 2 is zero. This is a situation that seems to prevail in general, b ut exceptions are discussed belo w, and there will be significant sur ace resistance in presumably all cases as film thickness approaches zero. The surface concentrations, c 1 (or c s), will be higher for closer matches in HSP between challenge chemicals and polymers. Therefore, it is not surprising that HSP correlations can be made when permeation rates for a lar ge number of chemicals have been measured for a gi ven polymer, as reported in Chapter 13. This is particularly true when the molecules involved are all relatively small. When the molecular sizes of the challenge chemicals become too lar ge and/or their shape becomes suf ficiently complicated with side group and cyclic structures, simple HSP correlations are no longer possible. The diffusion coefficient i affected by these size and shape f actors, and the HSP can no longer be used as a single correlating parameter. In addition, surf ace resistances can also become v ery significant, as discussed bel w and in Chapter 14. The molecular v olume, V, of the challenge chemical has been used with some success to account for size ef fects, b ut this does not directly account for dif ferences in shape. Examples of correlations for dif fusion through chemical protecti ve clothing, for e xample, demonstrated that molecular size had to be tak en into account 2 (see also Figure 13.2 and the discussion in Chapter 13). The most reliable HSP correlations in these cases do not immediately consider the solvents with smaller molecules that may permeate f aster than e xpected by comparison with all the others. Likewise, improved understanding and correlations are obtained by initially ne glecting the solvents with larger molecules that do not permeate as fast as expected, in spite of close matches in HSP with those solvents that do permeate rapidly. Once a reliable HSP correlation is established without these ob vious outliers, their beha vior is better understood. Predictions then often become possible for other v ery small and/or v ery large molecular species as well.

Relevant solutions to Equation 16.2 or Equation 16.7 are used to measure dif fusion coefficients The diffusion equation must be solved with two boundary conditions and an initial condition. This discussion will consider a plane film xposed to absorbing chemical on tw o sides. The initial condition is chosen as a uniform concentration within a film, so that C is 0 for all X, r gardless of whether the initial concentration is zero or not. Dif fusion is also usually assumed to tak e place in one direction only, but side effects can become important in thicker samples as discussed below. The chemical concentration at the e xposed surf ace(s) is assumed to immediately rise to the

Absorption and Diffusion in Polymers


equilibrium value. The second boundary condition is found at the middle of the free film xposed on two sides where there is no transfer in either direction. In other words Equation 16.1 is set equal to zero at this point. For a constant dif fusion coefficient, solving the di fusion equation with these conditions gi ves an initial straight-line absorption curv e as a function of the square root of time. This is generally called Fickian diffusion. The straight line passes through the origin, and the time required to absorb one-half of the equilibrium amount, t 1/2, can be used in Equation 16.9 to find 0. D0 = 0.049 L 2/t1/2 (16.9)

This equation is based on T having a value of 0.049 when half of the equilibrium amount has been absorbed. There is an identical result for desorption e xperiments where the time required for half of a uniformly absorbed material to lea ve the film also requires the same T value.

Solutions for the diffusion equation have also been generated for concentration dependent diffusion coefficients for the same boundary and initial conditions as described ab ve. In this case it can be shown that Equation 16.7 reduces to Equation 16.10. 3–5 ∂Dv/∂T = D v(∂2Dv/∂X2) (16.10)

This equation has been solv ed numerically man y times for dif ferent values of D max for both absorption and desorption with a uniform initial concentration. 3–5 Dmax is the ratio of the maximum diffusion coefficient found at the maximum concentration encountered in an xperiment to D0. The half-times calculated for both absorption and desorption were con verted to F a or F d, respectively, their ratio to the v alue 0.049. These values are given in Table 16.1. F a and F d are the F M for use in Equation 16.11. The correction f actors for desorption e xperiments were also found for the time required for only one fourth of the material to lea ve the film in desorption xperiments. This was necessary because of the extremely long experimental times (months) required for even this amount to leave, and also holds true for very thin films. These quarter-time correction factors for desorption were used to generate the results reported in Figure 16.1 and are discussed in the follo wing.6 The results reported in this figure confirm that a and F d are correct and useful. The same dif fusion coefficients were found by both absorption and quarte -time desorption measurements. The correction factors for the absorption measurements, Fa, were close to 2.0, whereas those for the quarter time desorption measurements, F d, were in the range of 40 to 144. D (c ) = FM × FB 0.049 × L2 tE (16.11)

The factor F B in Equation 16.11 is a related correction accounting for an y surface resistance as discussed belo w. F B will al ways be greater than 1 as a surf ace resistance slo ws the transport process and leads to an apparent dif fusion coefficient that is too l w. The exponential increase in dif fusion coefficient with concentration is xpected based on free The main volume theory.7 It is be yond the scope of this chapter to include this theory in detail. feature of diffusion in polymers is that the macromolecular chains are barriers to transport. F actors that either promote mobility of the chain se gments or increase the distance between them will enhance the movement of smaller molecules. The transport occurs as very small movements rather than larger jumps. An increase in concentration of plasticizing smaller molecules leads to an increase in the free volume of the system, as the smaller molecules ha ve more free volume associated with

has preferred units of cm/s.11 Desorption Dmax 100 2 5 101 102 103 104 105 106 107 108 (Fd)1/2 1.97 10.60 12.1 leads to e xactly the same result for the exponential dif fusion coef ficients of chloroform in polystyrene as as found by the method of integrals. The surface concentration at any given time is c s.5 290 506 Absorption (Fa)1/2 1. the absorption process then can be modeled with the follo wing boundary condition at the surf aces: F = h (c∞ − cs ) (16.7 443 1. (Fd)1/4 is for desorption quarter -times.84 10.14 7. This corresponds to slightly more than doubling the dif fusion coefficien for each added 0. and (Fa)1/2 is for absorption half-times.01 v olume fraction of solv ent.30 4.70 2.30 138.30 1.03 v olume fraction at lo wer solvent concentrations.370 4.1 and Figure 13.01 3.85 6.40 89. h can be estimated from this initial flux d vided by c ∞.00 13. Thus diffusion coefficients increase as sol ent concentration increases.670 (Fd)1/4 1.00 1. The limiting slope at time approaching zero is used to find .12) The surface mass transfer coefficient. the diffusion coefficient for common sol ents increases by a f actor of about 10 for an increase of solv ent concentration 0.0 160.300 13. As c s is zero at time equal to zero.55 2. them than do polymers.1 Treatment of the experimental data given by Crank with the correction f actors given in Table 16. SURFACE RESISTANCE MATHEMATICAL BACKGROUND The diffusion equation must be solv ed with the appropriate boundary conditions at the surf aces when significant sur ace resistances are encountered.1 Correction Factors for the Measurement of Concentration Dependent Diffusion Coefficients for Use in Equation 16.20 23. .10 47. Concentration dependence in measuring dif fusion coef ficients by absorption can also b accounted for by the method of integrals given by Crank. Both procedures require iterations as the true v alue of D max is not kno wn initially and must be estimated to find a (and a ne w D max) until convergence is obtained. h.10 Note: (Fd)1/2 is for desorption half-times.298 Hansen Solubility Parameters: A User’s Handbook TABLE 16.70 4.40 43.56 2.00 1.63 8. As can be seen in Figure 16. F or a film xposed on two sides.61 3. One estimate of h can be obtained by plotting the weight g ain against time.00 1.1.

the correction for surf ace resistance was a factor of 250. With permission from Else vier.1 Diffusion coef ficients for chlorobenzene in polyvi ylacetate measured by absorption halftimes.5 ∂Dv/∂X = BlnD v (16.25 = 1.3 x 1. desorption quarter-times.1 0. Vf. considering the x direction only is ∂C/∂X = B(1 – C s) (16. this boundary con dition can be used with Equation 16.15) . Prog.63 F = Fa x FB = 1.8 Fd = 144 10 F = Fa x FB = 1. Corrections for surface resistance. 2004.2 0.Absorption and Diffusion in Polymers 299 6 . R d = R s. Coat. In an e xtreme case for absorption with Vf near 0.6 It is useful to rewrite this boundary condition in dimensionless terms using the quantity B. The dimensionless boundary condition corresponding to Equation 16. Thus.10 as 3. F or B = 1.13) Large B is indicati ve of a dif fusion-controlled process. and for lo wer B surface resistance dominates.5.2 v olume fraction.6 The lower curve is based on dry polymer content. R d.) 0. 51(1). F B. are also required for Vf above about 0. cm²/sec Isotope 8 Absorption Fa = 1.14) For an e xponential concentration dependence of the dif fusion coefficient.12 for use with Equation 16.LOG diffusion coefficient at 20 °C.3 Vf FIGURE 16. Org. B = R d/Rs = (L/D 0)/(1/h) = hL/D 0 (16..5 0.4 0.7.. This is the ratio of dif fusion resistance.M. 55–66. The upper curve is based on total thickness. Surf ace resistance becomes increasingly important as the film thick ness decreases.21 (Reprinted from Hansen. C. and isotope experiments as a function of the v olume fraction of chlorobenzene.2 x 250 = 300 Desorption (to vacuum) 12 Fd = 40 14 0. to that of surf ace resistance R s.

when surf ace resistance can be ne glected.4-dioxane. Solv ent absorption w as follo wed in injection-molded samples for 13 solv ents in c yclic olefinic copolymer (COC). 0.3-dioxolane in the COC polymer . with a clear “S” shape. FB = (3.2. General Electric). Such systems included 1. At higher B v alues it is not e xact. n-hexane. the delayed uptak e can be seen as a form of time-lag phenomena.5 for dieth yl ether .95 6. is SURFACE RESISTANCE IN ABSORPTION EXPERIMENTS Solutions to the diffusion equation have been presented for v arious surface resistances (surface conditions) by Crank for a constant diffusion coefficient 1 FB has been evaluated from these graphical results as the ratio of the half-time for a gi ven B value to 0.16 can be used for B less than about 0. and 20 for prop yl amine.16) As stated abo ve.0 1.2 Correction Factors. Topas® 6013 from Ticona. There is no significant sur ace resistance to absorption in those cases where the absorbing molecules are smaller and linear such as for tetrahydrofurane. the square root of time initially is a straight line that passes through the origin.5 M Note: This means that F equal to 1.5. as can be seen by the S-shaped 8 The apparent B values for these cases are 10 for eth curves that do not pass through the origin.049. a plot of the (relati ve) uptake vs.0 in this case. Surface resistance is significant in all three cases sh wn.5 0. When surf ace resistance becomes important.75/B) +1 (16. and 1. FB. The factors leading to and controlling the pre vailing surface concentrations are of major interest and not necessarily the f act that these increase in an e xponential manner with time. 4 sol ents in tw o different grades of polycarbonate (PC).300 Hansen Solubility Parameters: A User’s Handbook TABLE 16. When the challenge molecules are too large or bulky.1 1/B 0 0. Approximate surf ace mass transfer coef ficients and approximate di fusion coef ficient were determined where possible.5 1 2 10 FB 1. Numerical solutions to the diffusion equation of the type described in Reference 3 and Reference 5 confirm that significant su ace resistance leads to an e xponential increase in the surf ace concentration. methyl isobutyl ketone. Equation 16.11 B ∞ 10 2 1 0. There . It was discovered that a surface resistance to absorption was significant in 19 of thes 23 cases. no absorption occurs. ylene dichloride.2 sho ws S-shaped curv es for absorption in the COC polymer . Figure 16. and b utyric acid in PC (Le xan® 104R.1 0. These results are reported in Table 16. It is also interesting to note at this point that solutions to the dif fusion equation with the boundary condition of an e xponential increase of the surf ace concentration with time give absorption curves with exactly the same “S” shapes. for Use with a Constant Diffusion Coefficient in Equation 16. and 2 solvents in the terpolymer acrylonitrile/butadiene/styrene (ABS). acetophenone.14 4. and phenyl acetate in the COC polymer.45 3.8 37.

C.B. at 23 °C. the apparent permeation coe ficient ould have been one-half that of the true permeation coef ficient The data in this type of figure can be interpreted using the foll wing set of equations: F = Δp/(L/Papp) = Δp/(L/P∞ + R 1 + R 2 + R 3 …. diethyl ether.. which is a normal film thickness. Equation 16.Absorption and Diffusion in Polymers 301 Ethylenedichloride in COC 300 Weight change [mg/g] 250 200 150 100 50 0 0 25 50 75 Diethylether in COC Propylamine in COC 100 125 Time [sqrt(min)] 150 200 200 FIGURE 16. Ticona. Ind. depending on the situation. and Hansen. and its resistance is therefore ef fectively infinitely la ge. and vice ersa.17) (16. Other extrapolations are possible to g ain more information.18) (16. 1/ ∞. Cop yright 2005 American Chemical Society. Topas® 6013. 9 If a single measurement had been made at a film thickness o 40 microns. e ven though HSP w ould predict this. for different film thicknesses and xtrapolates the inverse of the apparent transport coef ficient ersus the in verse of the film thickness to zero corresponding to infinite film thickness This is portrayed in Figure 16.2 Absorption of ethylene dichloride. (Reprinted with permission from Nielsen. 1/Papp = 1/P ∞ + (R 1 + R 2 + R 3 …. T. SURFACE RESISTANCE IN PERMEATION EXPERIMENTS As stated above.) was no weighable absorption. 3959–3965.13 clearly sho ws that the importance of surf ace resistance will increase as film thickness decreases. and the slope gi ves the sum of the resistances.3 for the permeation of water through an acrylic coating. Chem.19) The extrapolation to 1/L equal to zero gives the inverse of the true permeation coefficient. This can be used in permeation measurements to fin the true permeation coef ficient as well as the sum of all other resistances to the transport process One measures apparent permeation coefficients app. The molecules can simply not get through the surf ace layer . Eng.M. for xample. 2005. and the gi ven R represent dif ferent sources of resistance. Between these extremes are situations where surface resistance clearly affects the absorption process. and n-propylamine in a COC polymer. 44(11). Other examples of lack of absorption are oleic acid that simply does not absorb in the PC polymers or ABS. Surface resistance becomes significant when molecules can be transported way from the surface into the bulk of the polymer faster than they can be adsorbed/absorbed just at/in the surf ace. Research.) L/Papp = L/P ∞ + R 1 + R 2 + R 3 …. Δp is the o verall vapor pressure dif ference in the system. 10 One can find the apor diffusion coefficient (resistance). by varying the amount of liquid .)/L (16.

) in a cup-type e xperiment. where the cooling rate (and orientation) is dif ferent. McDonald et al. 1985. In addition to the discussion and literature cited abo ve. also arri ved at the conclusion that surf ace flux limited di fusion of solv ent into polymer could e xplain observ ed beha vior such as Case II and transitions between Fickian . The following also deal with surface effects in the absorption of solv ents into polymers. The rapid surface cooling in the injection molding process gi ves a surf ace different from the interior ..10 SURFACE RESISTANCE — A DISCUSSION In the literature surface resistance has alternatively been called skin layer effect. and Hansen. 13(3/4). surface effects ha ve been noted in man y studies of dif fusion in polymers. surface condition. M. ood in the fiber direction. or boundary layer resistance. It has been found that these surf ace and vapor diffusion effects must be accounted for in cup-type and related e xperiments with thinner polymer films and porous types o materials. This effect is presumably found with man y injection molded polymers. (Reprinted from Huldén. Extrapolation to 1/L to zero g ves the inverse of the true permeability coef ficient.12 NMR microimaging was used to confirm that carbon tetrachloride absorption as retarded by this layer . and an estimate of the surf ace resistances on the tw o sides of the paper . Prog.3 A plot of the in verse of the apparent permeability coef ficient ( app in kg P a1 s1 m1) vs. for example. With permission from Else vier. Org. These include thinner paint films.9–11 This type of e xperiment using paper as a film separated resistances for the permeation of th paper by water as well as for dif fusion of water in air. the inverse of the film thickness (L in m) for an acrylic coating.M.302 Hansen Solubility Parameters: A User’s Handbook 1 x 10-12 Papp 20 15 10 5 P∝ 05 10 15 20 25 1 x 10-3 L FIGURE 16. Coat. C. 171–194. and pape . (heat transfer) evaporation of the water. and leads to what is often called an induction time and “time-dependent” diffusion.. interfacial resistance. A skin layer was found on the surface of injection molded polypropylene.

This was about 180 microns or higher for the linear aliphatic molecules.5(10) -6. and increased for more complicated structures.05 cm at 35°C. the surf ace concentration only slo wly rises to the equilibrium v alue. 16 Water permeation through Nylon-6 films as studied for dif ferent film thick ness. The significance of the sur ace resistances can be demonstrated by the film thickness at which the resistance from di fusion within the film is equal to the sum of the t o surface resistances. Hwang and Kammerme yer showed significant sur ace resistance by studying permeation as a function of film thickness for ater in acetyl cellulose acetate and Nylon-6. and n-pentane was studied.21) . This is discussed in more detail belo w. Dav can be found from Equation 16. and for w ater in cellulose acetate. It li ewise indicates that the greater the cross-sectional area of the molecule. ethyl acetate. Characteristic S-shaped absorption curves were found in the methylene chloride-PEEK system by Gryson et al. n-butanol. cellulose acetate. The reason for this is thought to be that the orientation of molecules lea ving the film is directe more toward the e xit surface.20) hav (in cm/s) is the average surface mass transfer coefficient. benzene.21 as the log arithmic mean where D 1 is the dif fusion coefficient on the xposed side and D 2 is the dif fusion coefficient on the xit side. Surf ace resistance was found to be particularly important in thinner sheets. 15 The initial rate of absorption w as found to be strongly dependent on the surf ace condition but the equilibrium v alues were not. h on the e xposed side of the films as approximately 1/3 of that on the e xit side. This equation indicates that a molecule in the surf ace of a polymer will proceed inw ard in direct proportion to the diffusion coefficient. The right orientation at the right place allo ws adsorption/absorption. and that some of the observed features of anomalous sorption can be e xplained with the help of this model. benzyl alcohol. an A (in cm 2) is the (minimum) cross-sectional area of the molecule in question. Surface resistance w as shown to af fect permeation through polymer films by Kim and Kam mermeyer who measured actual concentration profiles in Nylon-6. When there is a surf ace resistance. k 2 is a constant with the v alue approximately 4. Case II diffusion involves a linear absorption curv e when time is used rather than the customary square root of time.17 Skaarup18 performed man y permeation e xperiments on Polyamide 6 (P A 6) (B ASF Ultramid B4) and polyvin ylacetate (Hoechst. Skaarup arri ved at the following general relation for the permeation of these liquids through P A hav = k 2(Dav/A1/2) (16. h ydrogen through stainless steel.13 Toluene diffusion in polystyrene is discussed in detail.4-dimeth yl-2pentanol. Similar studies sho wed that there were also significant sur ace resistances for dioxane.21). 2. Shankar studied interf acial resistance in absorption e xperiments. Mo wilith 50) where surf ace resistance w as significant. Surface resistance in the meth yl iodide-cellulose acetate system produced the characteristic Sshaped absorption curv e. and polyeth ylene membranes. Dav = (D 1/D2 – 1)/ln(D 1/D2) (16. Systematic variations in absorption phenomena with film thickness are a clear indication of a significant su ace resistance. and n-hexane in polyeth ylene. A molecule landing “sideways” and hitting a polymer segment will be rejected.Absorption and Diffusion in Polymers 303 diffusion and Case II dif fusion. the more difficulty it will h ve to reside at the surface in a condition where it can take a small jump into the b ulk. O particular interest are the studies on P A 6 where the permeation of eth yl laurate. av (in cm 2/s) is the average diffusion coefficient in the film (see Equation 16. whereas the orientation of molecules approaching the entry side is more random.14 The conclusions were that non-Fickian characteristics can result from slo w transfer to the surf ace layer. and p-dioxane through Nylon-6 and polyeth ylene. It w as clearly sho wn that the surf ace concentration of w ater did not reach the equilibrium value at the equilibrium permeation rate unless the film thickness as greater than 0.

These equations are based on the initial uptake (or loss) to e valuate D app.24.9 times lar ger than w ould ha ve been measured for an e xtremely lo w concentration dif ference across the membrane. This value for t 1/2 can be used in Equation 16.20 volume fraction where the system is abo ve its glass transition temperature. have significant side e fects when they are used to measure diffusion coefficients. For c ylindrical geometry a deri vation similar to that used to find Equation 16. The uptake is more rapid than otherwise anticipated.44. It can be seen that D 0 and D app are equal to the e xtent that the second and third terms in Equation 16. Tensile bars that are 4 mm thick and 10 mm wide will require corrections that are at least as great as a f actor of 1.3.23 for dif fusion in plane samples of isotropic media can be used for this purpose.9 if there is a linear absorption curv e as f ar as t 1/2 when the square root of time is used. A square sample that is 10 mm on each side and 1 mm thick requires a correction for end ef fects equal to 1. (c1 – c2).20 This w as done by comparison of e xperimental results with solutions to the dif fusion equation accounting for both .24) R is the radius of the c ylinder and b is its thickness. an easy procedure is to e xtrapolate the initial slope on a plot of uptak e versus the square root of time to a fict ve t 1/2 found when the relati ve uptake is equal to 0. 19 D0 = D app/(1 + L/l + L/w) 2 (16.9 to find app.96. These same considerations are v alid for permeation coef ficients since P = DS (16. If the ratio D 1/D2 is 10 then D av is 3.304 Hansen Solubility Parameters: A User’s Handbook This equation shows how much the average diffusion coefficient di fers from that at essentially zero concentration corresponding to D 2. as judged from the dif fusion coefficient data reported in Figure 16.15 volume fraction. Here ag ain.18 are not significant. and S is the solubility coefficient. SIDE EFFECTS Corrections for absorption or desorption from the sides of thick er samples should be applied to measured diffusion coefficients if there is significant d fusion through the sides of thicker samples. respectively. Co ventional tensile bars.22) P is the permeation coef ficient measured at the g ven steady state concentration dif ference.5.23) Dapp is the apparent dif fusion coefficient found from initial slope measurement 1 or Equation 16.03 v olume fraction across the membrane when rigid polymers are involved. D 1/D2 equal to 10 implies that the concentration dif ference is only about 0. This implies that concentration dependent diffusion can lead to significant errors when difusion and permeation coefficients at ery low concentrations are needed. The length and width of the film are l and .23 results i Equation 16. with L being retained as the thickness. so the apparent diffusion coefficients must be reduced accordingl . for example. D0 = D app/(1 + b/R) 2 (16. If 1 is 10 times greater than D2 then P is approximately 3.9 times lar ger than D1. Equation 16. For elastomers this same ratio implies a concentration dif ference near 0.1 at concen trations above about 0. MEASURING DIFFUSION COEFFICIENTS CONCENTRATION DEPENDENCE WITH SURFACE RESISTANCE AND It was necessary to measure diffusion coefficients ver as wide a range as possible in order to solve the dif fusion equation to simulate film formation by sol ent e vaporation.

1. at low concentrations were close to 1. and air v elocity past the surf ace.21 The technique used w as to interpret the e xperiments initially assuming that the diffusion coefficient as a constant at the concentration assigned to the e xperiment. Absorption e xperiments only were rele vant abo ve about 20%vol because there is a marked change in diffusion behavior at concentrations above and below this value (for the system chlorobenzene and polyvin ylacetate).0 (corresponding to increases of from 50 to 100%) for the usual step-wise absorption e xperiments from one concentration to a slightly higher one.3–5. As can be seen in Figure 16. F B. Vt is a coefficient from 0 concentration up to A.3 The isotope experiments are for very high solvent concentrations that correspond to liquid lacquer formulations.1 in Equation 16. They can become as high as 100 or more. The correction f actors. In this situation h is af fected by factors such as vapor pressure and latent heat of the solv ent.11. Surf ace resistance has been present in many studies reported in the literature and clearly af fects many results involving diffusion in polymers. 21 The total dif fusion coef ficient cur e from 0 to 100% solv ent (chlorobenzene) was completed with a self-dif fusion coefficient and s veral isotope experiments. whereas the second stage is controlled by the rate at which sol ent molecules can diffuse to the film/air sur ace in order to e vaporate.20 Surface resistance was significant at concentrations ab ve about 0. where the self-dif fusion coefficient is 1. where dif fusion coefficients are reported for the same system. This situation naturally led to solving the dif fusion equation with rele vant values for dif fusion coefficients an surface resistances for additional situations of interest. The overall result was a simple explanation for the various types of so-called anomalous diffusion. Loss of solv ent was followed experimentally for about 2 years from an initial Vf of 0. When concentrations reach higher than about 20%v ol. These coordinated absorption and desorption experiments were necessary to determine the true diffusion coefficients at all concentrations so that the solutions to the di fusion equation for fil drying by solvent evaporation were correct.1 sho w that the same dif fusion coefficients were found in both cases.3. These techniques are reported elsewhere.Absorption and Diffusion in Polymers 305 concentration dependence and surf ace resistance at the same time. F a. Surf ace resistance controls the first stage. increase rapidly . Fd.2 v olume fraction solv ent in the system that w as studied most e xtensively (chlorobenzene in polyvinylacetate). This was also confirmed by the calculations since the sur ace concentration fell to essentially zero when this amount was reached. this is the xtent of the v ariation of the dif fusion e.20 It is also necessary to consider concentration dependence simultaneously with surf ace resistance in man y cases of practical importance.4. The solvent-in-polymer diffusion coefficient for chlorobenzene in polyvi ylacetate increases more than 9 decades as the solvent concentration increases from essentially zero to 100%. It is the balance between the (concentrationdependent) diffusion resistance and the surf ace resistance that determines whether the dif fusion is "Fickian" or whether it is presumed to be anomalous. heat transfer to the surf ace. The experimental and calculated curv es are reported in Figure 16. The correction factors. were well above 100 for desorption measurements from an equilibrium state near 15–20%v ol solv ent to v acuum. The results in Figure 16.5–2.1.65(10 5 cm 2/s. Corrections to this estimate (multiplying f actors) were then made by comparison with suitable solutions to the diffusion equation using the data in Table 16. This is discussed in the ne xt sections. The quantity V2 in Figure 16. FILM FORMATION BY SOLVENT EVAPORATION The process of film formation by sol ent evaporation has been fully described using the dif fusion equation with a significant sur ace resistance and local dif fusion coefficients found from the dat in Figure 16. The process of solv ent loss tak es place in tw o distinct stages.75 until it w as totally lost. b ut only rarely has there been mention of this f act. 3. . and the continued loss of solvent was controlled by internal diffusion to the surface. The local diffusion coefficient changed from 5(10 –7 cm 2/s initially to 1(10) –14 cm 2/s at zero concentration. the correction f actors for surf ace resistance. the concentration at the break in the curv fictitious di fusion coef ficient to calculate the di fusion coef ficients at concentrations ab ve C A.4 is gi ven as 10 6.

for example. SUPER CASE II) For absorption with a constant dif fusion coefficient one normally finds that the initial weight ain is linear with the square root of time as stated abo ve.306 Hansen Solubility Parameters: A User’s Handbook 101 Volume Solvent / Volume Polymer V2= 106 Vt= 1010 CA= 0. because the distances in this figure are based on dry film thicknes When the concentration at the center of the film starts to increase ab ve zero. Absorbed w ater plasticizes the film. This is called Fickian or normal dif fusion. There is an adv ancing front into the film that is som what more pronounced than that sho wn in Figure 16. The films were d vided into a suf ficient number of finite d ference elements to assure correct results. It only mak es certain that the surf ace concentration is zero. This can be seen in Figure 16. This is largely a result of concentration dependent diffusion coefficients with a la ge jump in concentration. B=106 7 22 microns B=10 ~MO –CA –CA CS= O CS= O For B=107 For B=106 5 10 CS= O For B=105 Calculated Experimental Effect of water . the weight increases linearly with time.6. The calculated and measured desorption curves under vacuum coincided at long times (several months). The chronology of the e vents is that surf ace concentration increases as time goes on. 5 Uptake is linear with time early in the process. 165 microns B=10 Exptl.20 See the discussion in the accompan ying text. Film drying in a climatized room w as f aster than film drying in a acuum apparatus. The diffusion resistance . where diffusion coef ficients were measured. Diffusion coefficients were calculated for each local site in the films The iterative procedure at each successi ve time interv al w as that described in Crank 1 as the Crank–Nicolson method. Vacuum does not hasten release of solv ent by dif fusion at longer times.2 B as indicated Exptl. This change occurs when the dif fusing material reaches the middle of a sheet e xposed on both sides. See Figure 16. In Case II diffusion. (L)2 10-3 10-2 FIGURE 16. which it will be in almost all cases an yway.5. This is true e ven when there is v ery little surf ace resistance of significance 5 Super Case II emer ges as the surf ace resistance becomes more significant relat ve to the dif fusion resistance.6 where concentration gradients ha ve been calculated for Super Case II behavior. b ut at some longer time the rate of g ain increases mark edly. the rate of absorption also begins to increase. ANOMALOUS DIFFUSION (CASE II. Simultaneous diffusion and surface resistance combine in a special way to produce this beha vior. The concentration profiles ultimately become flat at moderate elapsed tim as diffusion within the film is n w rapid compared to the surface resistance.3.a steeper slope One day L=30 microns 10-1 10 -2 10-7 10-6 10-5 10-4 Dot Dimensionsless T.4 Calculated and experimental drying curves for the evaporation of chlorobenzene from polyvin yl acetate.

8 6 T x 10 1. and the rate of uptae increases..1). The absolute value of B depends on the de gree of concentration dependence o ver the selected concentration interv al .250 0. Polym. (Reproduced from Hansen.0 0. The rate of weight g ain decreases just before the equilibrium v alue is attained.6 ØR 0. This is a typical Super Case II type curv e with initial absorption linear with time follo wed by a sudden increase in the absorption rate as the solv ent reaches the center of a film xposed on tw o sides. There is still another ef fect at very long times where the surf ace resistance begins to be more important ag ain relative to the dif fusion resistance.125 0. 20(4).M.0 0. 20(4). Sci. 252–258..4 0.2 0.216 0. Eng.386 0.319 0.0 0.037 0. Eng. C. ∅R is the local concentration.146 0. Polym. 252–258.4 0.6 Concentration gradients for conditions corresponding to Super Case II type (anomalous) diffusion. B is the ratio of dif fusion resistance to surf ace resistance.Absorption and Diffusion in Polymers 307 1.8 0. 1980.6 0.4 0. 1980. (Reproduced from Hansen.. The concentration gradients in Figure 16. The final l veling off at very long times is also characteristic of Super Case II.625 0.875 1.467 0.5 Solvent uptake curves for v arious values of B and concentration dependent dif fusion coeffi cients equal to those reported for chlorobenzene in polyvin ylacetate (See Figure 16. Sci.869 0.5 for B = 10 7.6 are for the curv e in Figure 16.6 0.8 Mt / M∝ 0.2 106 0..M.) becomes less and less as the concentration in the middle of the film increases.0 0.0 FIGURE 16.500 0.0 0.0 1. including concentration dependence. Diffusion in Polymers. With permission from the American Chemical Society.562 0.2 1.With permission from theAmerican Chemical Society.098 0.4 B: 109 108 107 FIGURE 16.2 0.375 X 0.) 1. Values on the curv es are relati ve weight g ains.750 0. C.0 0.

When the challenge molecules are too lar ge or b ulky. absorption into coatings. Additional material on dif fusion in polymers including simple mathematical descriptions of film drying by sol ent evaporation and so-called anomalous dif fusion have been included to present several aspects of the importance of surf ace resistance to contribute to its better understanding. etc.308 Hansen Solubility Parameters: A User’s Handbook in the given case. Predicting aggressive behavior is important. The measurement of surf ace resistances and concentration-dependent dif fusion coefficients did not requir use of an y mathematical tools or e xplanations other than the dif fusion equation solv ed with appropriate initial and boundary conditions. no absorption occurs.22 and on explanations based on stress relaxation phenomena at the head of this adv ancing front. they can only be convincing when the v erifiable sur ace resistances and the v erifiable xponential concentration dependence of the dif fusion coefficient are also ta en into account. 23. This absorption can be strongly af fected by both surf ace phenomena and the rate of diffusion within the bulk of the polymer. The simple approach of solving the dif fusion equation with a v erifiable ( xponential) concentration-dependent dif fusion coefficient. and its resistance is therefore ef fectively infinitely la ge. The molecules can simply not get through the surf ace layer.24 Whereas these may ha ve aspects of v alidity. There were adv ancing fronts in volved both in the mathematics and in the samples in the e xperiments. e ven though Hansen solubility parameter considerations lead one to predict that this is clearly e xpected. This is true with and without significant sur ace resistance. impro ved prediction of the beha vior of chemical protecti ve clothing. is dependent on the increase in free v olume brought locally by the solv ent molecules themselv es. This chapter has emphasized the importance of surface resistance in this process. CONCLUSION Many organic liquids are aggressive with respect to many types of polymers. Between these e xtremes are situations where surf ace resistance necessarily af fects the absorption process. Surf ace resistance becomes significant when molecules can be transporte away from the surf ace into the bulk of the polymer f aster than they can be adsorbed/absorbed just at/in the surf ace. release of sterilization g as. and the absorption of the or ganic liquids into the polymers is a k ey factor in this respect. absorption and transport through polymeric packaging of v arious kinds. The methodology and concepts presented in this chapter on dif fusion in polymers combined with the methodology and concepts in the rest of this handbook should pro vide insight into man y situations of industrial and theoretical interest. There is no surf ace resistance to absorption where the absorbing molecules are small enough and/or linear. These discussions are focused on time-dependent replication of the observ ed dif fusion phenomena. release of absorbed chemicals from plastics. It is not yet possible to predict . The parameters chosen to calculate this curv e are realistic. xplains and replicates both the absorption and the desorption of solvents in polymers over an extremely large concentration range. It w ould appear that enhanced stress relaxation. The simple approach of solving the dif fusion equation with appropriate and v erifiable param eters for the boundary conditions e xplains and replicates both the absorption and desorption of solvents in polymers o ver an e xtremely large concentration range. See Reference 5 for greater detail. such as an adv ancing front. along wit appropriate and verifiable parameters for the boundary conditions. GENERAL COMMENTS Discussion of the v arious other e xplanations for anomalous dif fusion is be yond the scope of this chapter. like the increase of dif fusion coefficient with sol ent concentration. These include controlled release of drugs.

Aspects of solubility . 9. 20. 1956.J.. . Technol. pp.M. and Hansen.J. C.. Solute and penetrant dif fusion in swellable polymers. The role of w ater transport in composite materials. Phys.. Polym. R.A. Macromolecules. Ed. C. Phys. C. J. J.M. Ed.. Sci. Copenhagen. Else vier. J. and Hansen. Sager. J. N.. 6(4). The Mathematics of Diffusion. Mansdorf. H.. Ind.. N. Solubility parameter prediction of the barrier properties of chemical protective clothing. Surf ace Resistance in Dif fusion Processes in Polymers (in English). G. 5. 25.. K.. J. Hansen. C.M. Eds. Polym. Interpretation of anomalous sorption kinetics in polymer -penetrant systems in terms of a time-dependent solubility coef ficient. and dif fusion in polymers.. 1988. C. R. R.. 39(3)... 12. 51(1). A. Performance of Protective Clothing: Second Symposium. 21.. 1981. 3. Hansen. 8. 22. chap.M...M. 1968. C.. 3959–3965. K.. J. R. Ed. 2005. 23. Phys. 1987. R. Coat. 252–258. Skin layer ef fects on the dif fusion of carbon tetrachloride into injection moulded polyprop ylene studied by 1H NMR microimaging. 3. J. Polymer. Sep. 1984. and Peppas. ASTM STP 989. Huldén. Lecture at Nordic Polymer Days. American Society for Testing and Materials.M. 1988. Surface flux limited di fusion of solvent into polymer. 197–205. Hansen.K. Doctoral dissertation. Org. Coat. 34. C.: Part B: Polym. pp. 24. and Sear.M. Fundam..A. 11. 4.A. Oil Colour Chem. Prog. Marom. J.. 6.. Eng. Windle. The measurement of concentration-dependent diffusion coefficients — the xponential case.. in Polymer Permeability. 1981. 22. Nielsen. but some similarity in HSP must be present for absorption to occur in an y event. Comyn. 1985.J. M. 27–43.. Von Meerwall. P. M. S. Experimental studies are still required. Hunter . 18. S. C. Polym.B. 1967. Chem.. Ind. Assn. Ed. J. Hwang..H. Sci. Sci. Transport of meth ylene chloride in poly(aryl-ether -etherketone) (PEEK).. Coat. 1980.. Polym.Absorption and Diffusion in Polymers 309 which systems will gi ve significant sur ace resistance to absorption. Oxford. 197–208. in Permeability of Plastic Films and Coatings. Grayson.. and Kammerme yer. 51(1). Sci. 13. Case II sorption. Abstract. Hansen. 2001.. Their Importance in Surface Coating Formulation. E. Comyn. Polym. Abbott. Plenum... Verification of theoretical models. Eng.M. 1985... V. K. 7. surf aces.. Danmarks Ingeniør Akadami..T... 171–194. Sci. 1981. 5(6). 3311–3315. The Three Dimensional Solubility Parameter and Solvent Diffusion Coefficient. J.. Hansen. Significance of sur ace resistance in absorption by polymers.. 15. chap.. Lett.W. London. 2004. 1967. Ev aporation and vapor diffusion resistance in permeation measurements by the cup method..M.: Polym.. L. 409–434. and Wolf. Chudek. Sci. J. 2. Dif fusion coef ficient measurements by sol ent absorption in concentrated polymer solutions. 679–697. K. C. 19. 935–945. Sci. C. 1108–1110. Hopfenberg. Skaarup. P ao.H. Hansen. Org. Nilsson.A. 16. Appl. Crank. Hansen.. and Kammerme yer. 1985. 55–66. E. 17. Eng. 1974. 14. II. Danish Technical Press.. and Hansen. P. 1885–1900.. Sackin. G. Influence of inter acial resistance on kinetics of sorption.. 57–60. Godward. Färg och Lack.. Philadelphia. Elsevier. 15. K emiafdelingen (in Danish) 1981.Z. Oxford University Press. Poly. 748–752. Actual concentration profiles in membrane permeation. REFERENCES 1. 1048–1057. and Squires. London. Water permeation in coatings. 61–64. Grænseflademodstand ved Diffusionsprocesser I Polymerer. 53(680).B. Potential errors in water/water vapor permeation measurements using the cup method.M.. Ed. Chem. 20(4). 13(3/4). J. J.P. 26. Shankar. C.S.. A mathematical description of film drying by sol ent evaporation. 1996. and Hansen. Ne w York. Res. Dif fusion in polymers. 9. 1993. McDonald.M.M. Kim. Petropoulos... T. Prog. 609–614. C. 1986. Korsmeyer..P. Ef fect of thickness on permeability . in Polymer Permeability.. PA. Hansen. 1970. and Nielsen. 22. Mater.. 10.. 44(11).

. J. 2001. M.310 Hansen Solubility Parameters: A User’s Handbook 24. Systematic analysis and model interpretation of micromolecular non-fickian sorption kinetics in polymer films Macromolecules. 1400–1410.H. .. and Petropolous. 34. Sanopoulou.

products. The emphasis here is on the use of sound formulating principles to reduce the potential hazard in terms of reformulation or substitution. SUBSTITUTION Substitution involves the replacement of a potentially dangerous process or chemical with a ne w process or chemical ha ving less hazardous properties. INTRODUCTION Many organic materials are potential safety hazards. labeling requirements. thus pr viding information on the means to impro ve handling of the problem. All of these situations are discussed in more detail in the follo wing. b ut a list of this type is not the purpose of this chapter. biode gradability. whereas the y were used in massi ve quantities earlier . Hansen ABSTRACT Hansen solubility parameters (HSP) can be used to gain insight into many safety and environmental issues. Evaluating other forms of en vironmental risks can be aided by using HSP . Ev aluating risks for inadv ertent chemical uptak e in plastic can be helped by HSP correlations of chemical resistance and/or permeation phenomena. where a listing of solv ents having HSP similarity to the one(s) to be substituted pro vides an o verview of the potential choices for impro vement. HSP can indicate which chemicals can do this. weight g ain. These same approaches can be used to e valuate the potential for uptake of chemicals through human skin. The problem of replacing ozone-depleting chemicals is a case involving the e xternal environment. This type of information can be generated for any polymer where HSP correlations of chemical resistance. it is often a matter of e xperience before the risks are unco vered because of damage being done. These are lik ely to dif fuse into the plastic container w all itself. The v olatility of products is also a significan 311 .— Safety and 17 Applications Environment Charles M. They can also be harmful to the en vironment. The hazards can be judged using accepted approaches — for e xample. can be generated. Similarity of HSP suggests which chemicals are most lik ely to be rapidly absorbed into given plastic types. ag ain. and processes. there ha ve been a series of substitutions with or without the aid of HSP to eliminate or to at least reduce such problems.. toxicology assessments. HSP can help. personal protection of one type or another may be required. Here. Selection of suitable chemical protecti ve clothing can be improved by considering HSP correlations of breakthrough time. etc. An example is the lack of emphasis on the use of some eth ylene glycol ethers as solv ents because of their teratogenic ef fects. and physical properties for the chemical or products. Other lar ge-scale substitutions can also be cited where HSP can aid. making customary washing insufficient. These include substitution to more desirable materials. When a satisfactory substitution cannot be found. Unfortunately. An e xample is the occasional misuse of plastic containers normally used for soft drinks to store chemicals such as herbicides and pesticides. Thus. This is discussed a great deal in Chapter 11. over the years.

and other w aterreducible products. 1 Olsen.5–8 Some general principles of solv ent selection have been discussed in Chapter 8 and earlier .. The currently used HSP techniques and correlations can aid in some aspects of substitution. cleaners of arious types. Many cases of substitutions in practice have been listed by Goldschmidt. One must then sort throug these potential replacement candidates using other information to arri ve at a better alternati ve. On the other hand. The use of solv ents in alternati ve coatings systems has been the topic of se veral pre vious publications by the author . 3 and Filskov et al. to predict polymer/filler interactions. 2 Soerensen and Petersen.e. ho wever. Examples of preferred coatings technologies are the use of po wders which fl w at higher temperatures or polymerization by radiation. Examples of such systems are coatings with higher solids. chemicals which prefer entially distribute to a nonaqueous phase to a v ery high de gree Efforts should be made to de velop products with the lo west possible hazard. toxic. adhesives. sealers. Those who ha ve read the earlier chapters in this book will immediately recognize HSP as a tool to aid in the substitution and systematic formulation for reduced safety and en vironmental risks.9 as well as else where more recently. which emphasizes HSP only . A key element in this is a listing of solv ents where those most resembling the candidate for substitution are at the top of the list. as demonstrated in earlier chapters. The use of technologies in volving w ater or mechanical methods. 4 A long list of references for the Danish experience with occupational risks and solutions is given by Soerensen and Petersen. and fillers In general. It might appear that solvent technology and use of HSP will not be as important as it has been in the past. Ho wever. teratogenic. to impr ve barrier polymers. po wder coatings. 3 ALTERNATIVE SYSTEMS Alternative systems with less solv ent or no solv ent ha ve been focused on by the coatings and printing ink industries for man y years. F or example. radiation-curable inks and coatings.312 Hansen Solubility Parameters: A User’s Handbook factor with lo wer volatility being preferred due to reduced w orkplace concentrations and reduced replacement requirements for cleaners and the like which often recirculate in nearly closed systems. both of which use solvent-free base products to pro vide the coating.10 . Other product types which may be tar geted include cutting fluids. It is clear that much more data than HSP are required to mak e the desired substitutions. carcinogenic.10) at the top of the list. Equations 1. such as mechanical joints rather than the use of solv ents. The program described in Chapter 1 can do this by entering the HSP for the solv ent to be replaced and requesting a listing with those solv ents most similar to it (i. HSP principles can be used to aid in impro ved stability and adhesion. one primarily wishes to substitute for • • • • • Carcinogens or suspected carcinogens Substances with risk phrases for being very toxic. the lo west RED numbers as defined in Chapter 1. electrodeposition coatings. are preferred. This is not the case. allergenic. the problem should not just be transferred from the air e xhaust system to the sewer. and to aid in understanding some biological phenomena. or causing cumulati ve or irreversible effects Substances with moderate or serious aquatic toxicity Nonbiodegradable substances Substances with high predicted aquatic ef fects — for e xample. and it is anticipated that future correlations will help in this endea vor. mutagenic. a further discussion of this is be yond the scope of this chapter .

Spraying product in a smaller room with limited entilation v requires much more protection than applying paint with a brush outdoors. HSP can be used in this type of endea vor. cleaners. Such a change allo ws the use of higher solids at the original viscosity . fresh air mask. 11 A key element in these tables is the labeling code de veloped in Denmark according to the MAL (in Danish: Maleteknisk e Arbejshygieniske Luftbehov) system. A Danish publication comparing se veral of these is a vailable. by primarily professional users. or for an y product containing significant amounts o solvent or hazardous chemical. etc. Materials with least potential risk are to be used in the Nordic countries where ver possible. printing inks. Other numbers in addition to the TLV/GV/OEL (occupational exposure limit) and FAN have been generated to help e valuate risks by inclusion of e vaporation rate/vapor pressure considerations. perhaps more specificall . However. the simple MAL = F AN has been tab ulated in Table 17. A number of the general formulation principles can be briefly stated for th sake of completeness. 4. The surf ace tension of w ater-reducible coatings can often be significantly reduced b relatively small additions of ethanol or other alcohol-type solv ent. these may evaporate more rapidly and can be e xpected to ha ve a lower flash point 2. For present purposes. the quality of the w orking environment must be considered in all cases where organic solvents are being used. The risk must be indicated by the seller/producer in terms of a labeling code.Applications — Safety and Environment 313 SOLVENT FORMULATION AND PERSONAL PROTECTION FOR LEAST RISK Solubility parameter principles have been used in formulating alternati ve. dust mask. Solvents with linear and smaller structures diffuse more rapidly than those with branched and larger structures. 3. Tables ha ve been published which gi ve the protection required (glo ves. The labeling code dictates the personal protection required for the product. Solvents with lower viscosity most often lead to polymer solutions with lo wer viscosity. Such labeling is required on paints. body suit. These can. depending on the e vaporation rate (or v apor pressure). so the multiplier is lar ger. also be used in conjunction with other surf ace active materials.1 for se veral solvents.) for a gi ven set of application conditions for a wide v ariety of products from paints and printing inks through cleaners. this is translated as the F AN (fresh air number). Two (or more) mix ed solvents with lower labeling requirements may be able to replace a single solv ent. The Danish MAL system or other labeling system can be systematically used for this purpose. Higher MAL/F AN dictate that more extensive personal protection is required. low VOC (volatile organic compound) products. depending on the w ay it is used.4 To demonstrate the principle. The vapor pressure di vided by the TLV is called the vapor hazard ratio (VHR) and the actual calculated vapor composition (using activity coefficients) d vided by the TLV has been called SUBFAC. more linear solv ents can hasten the through-drying of a coating. The Danish MAL reflects the cubic meters of fresh air require for ventilation of 1 l of product to belo w the threshold limit value (TLV). of course. This number is modifie by a constant. The concept behind the MAL system can be better understood in English by translating the MAL number as the FAN. Inclusion of slo wer evaporating. The risk can then be assessed by users or . These include the follo wing: 1. Higher e vaporation rates imply greater hazard. . THE DANISH MAL SYSTEM — THE FAN12 As indicated pre viously.

amines at moderate concentrations. 3. and industrial coating (spray box es. T. 5. 4. e ye contact. to 6 as the potential hazard increases. a fresh air supplied mask. a combination filter mask. Xn. This second number will not be less than 1 if or ganic solvents are included in the product in significant amounts. surface preparation. This number is a summation of the hazards possible for the components which are considered potentially hazardous. C. including interior and e xterior painting and gluing. whether or not lar ge surfaces are involved. all common eth ylene glycol based ethers and their acetates (including dieth ylene glycol monob utyl ether . and by ingestion. and the most common chlorinated solvents. 11 The protective measure required may be a f ace guard. for e xample). in order of increasing requirements. a tw o-digit MAL code defines which safety precautions are required for each o a lar ge number of processing operations and conditions. One must also consider R (risk) and S (protecti ve measure) labeling requirements. each of the printing processes. a body suit. painting of ships. cabinets.1. breathing system. respiratory system contact. As indicated. Se veral of the solv ents require such labeling as well. eyes. The second number varies similarly and relates to the potential hazard from direct contact with the skin. 2.The following is a list of se veral solvents which are considered less desirable.314 Hansen Solubility Parameters: A User’s Handbook TABLE 17. b Estimated from composition of the mix ed solvent.1-Trichloroethane C>9 Aromatics Tetrahydrofurane Acetone FAN/MAL 20 19 17 15 14 14 13 13 13 12 7 6 5 4b 0 0 Solvent n-Propanol Propylene glycol monomethyl ether acetate Propyl acetate Propylene glycol monoprop yl ether Mineral spirits/white spirit Butyl acetates Ethyl acetate Cyclohexane Benzin/petroleum ether (as heptane) Heptane Ethanol Propylene glycol monob utyl ether Dipropylene glycol monomethyl ether DBE (dibasic esters) Ethylene glycol Propylene glycol These numbers are de veloped primarily with re gard to health hazards from v apors. . etc. and/or ingestion. monomers at rather low concentrations. based on this second number being 3 (or higher for higher concentrations in some cases): Toluene and Xylene at >10%.). larger construction sites. etc. the European Union requires use of Xi. The first number relates to the potential hazard from th vapors and will v ary from 00 through 0.15 Each product containing solv ent is assigned a tw o-digit number to place it into a potential hazard category. In addition to these. e ye protection. The second number in the FAN code is added for hazard for skin contact. terpenes.1 Fresh Air Numbers (FAN/MAL) for Selected Solvents from the Danish MAL Labeling Systema FAN/MAL 1400 1100 880 110 88 78 74 58 54 48 46 29 28 26 25 24 23 a Solvent Chloroform Tetrachloromethane Benzene Dichloromethane Trichloroethylene n-Hexane Toluene C9 Aromatics Methanol Methyl ethyl ketone Xylene 2-Propanol Propylene glycol monomethyl ether 1. 1. the quality of v entilation provided. a dust mask. A “3” in this cate gory places the protection required in a significantly highe category with requirements for glo ves as a minimum and frequently fresh air masks as well. a g as filter mask.

They have many advantages (which will not be discussed here). the molecular size of the chemical in volved is important in these e valuations. If this number is less than 1. A simple w ashing operation cannot be e xpected to remo ve all of the absorbed material. This is the f act that plastic materials are able to absorb v arious liquids to some e xtent. a plastic container can do this. Whereas the earlier glass or metal containers could not absorb potentially dangerous materials.10). One can usually divide the results into groups of clearly not acceptable. the system in not expected to be suitable for use. The RED number is discussed in Chapter 1 (Equation 1. There ha ve been recent attempts to impro ve on the direct correlation of breakthrough times and permeation rates with HSP by trying to estimate the solubility and dif fusion coef ficient separately using HSP. This is not the point of the present discussion. The major use of such correlations is to e valuate potential barrier types for chemicals where test results are not available.1. in an attempt to find a substitute with a l wer labeling requirement. and w orthy of further consideration. This is happening all o ver the world. Earlier publications also include HSP correlations of barrier properties of chemical protective clothing. including barrier properties relati ve to the contents. SELECTION OF CHEMICAL PROTECTIVE CLOTHING HSP correlations for barrier properties of some types of chemical protecti ve clothing are gi ven in Chapter 13. such as in Table 17. Containers in contact with food ha ve been tested well for suitability for this purpose. One then e valuates the RED number for the situation of interest. These correlations are based on data presented by F orsberg and Kieth. RED numbers significantly greater than 1. The extent of absorption clearly depends on the HSP of the plastic used in the container compared with the HSP of the liquid which is in contact with it. questionable. but there is also one disadvantage that HSP can shed more light on. A problem e xists with the inadv ertent storage of hazardous liquids in the plastic container prior to its e xpected recycling as a container for a food or be verage. there may be some doubt about the recommendation.0 can be considered for use. Washing . These may be found in a suitable table or can be calculated according to the procedures outlined in Chapter 1. 18–21 The procedure for using these correlations requires kno wledge of the HSP of the chemicals involved.1. Systematic consideration of labeling requirements is becoming a significant parameter i commercial applications of solv ents and products containing solv ents. Many types of liquids can be temporarily stored in such containers. The lo w label requirements of DBE (dibasic esters) ha ve been emphasized in comparison with other solv ents.22–25 These efforts have been discussed in Chapter 2 and Chapter 13.0. Such a procedure has been used to arri ve at optimum commercially useful solv ent compositions with the lo west possible risk for w orkers in the serigraphic printing industry as described in Danish patents DK 153797B (1989) and DK 160883 (1991) which correspond to European patents EP 0 205 505 B1 and EP 0 270 654 B 13.14 The preferred compositions reduce the MAL number to a minimum and also consider lo west possible internationally required labels as a requirement. UPTAKE OF CONTENTS BY A PLASTIC CONTAINER Plastic containers ha ve become increasingly popular in recent years. both with regard to worker safety as well as to the e xternal environment. If the RED number is close to 1. 17 Other e xamples of HSP correlations of barrier properties of protecti ve clothing are discussed in Chapter 13. As discussed in Chapter 13. Table 13.Applications — Safety and Environment 315 Examples of complete labeling of products and solv ents are beyond the scope of this chapter .15 Systematic solvent selection procedures ha ve also been strongly suggested for use in the selection of solvents for restoring older paintings. 16 This is discussed in Chapter 5. The purpose of the discussion is to suggest that possible substituting solv ents can be listed.

Among these is the presence in artic re gions of chemicals that do not readily break do wn. A skin penetration w arning has been attached to man y liquids taken up in the lists of limiting values for workplaces which are published in dif ferent countries. most notably the combination of dose and toxicity. HSP concepts can focus attention on the types of chemicals that can absorb into a gi ven type of plastic container. or aquatic species after the y have taken them up.5. TBBPA has a distance from the center of the spherical HSP correlation for Lignin solubility of only 6. The size and shape of the penetrating liquid molecules must also be considered.1 and Chapter 15. There may be other w ays to reduce the problem. Table 13. This is very close and means that where PCP is soluble. The HSP of given chemicals can gi ve a clue as to whether or not the y can follo w the same pathw ays in the environment.e. It was found earlier based on the HSP correlation with the swelling of psoriasis scales that this practice could be misleading. This is useful information in terms of what analyses should be performed prior to recycling. Predictions of the barrier properties of viable human skin should recei ve more attention. which presumably tak es place at some higher temperature.1. It is recognized that the δD parameter is dif ferent. as HSP predicted man y liquids without this w arning also swelled psoriasis scales (k eratin) and could therefore be e xpected to penetrate the skin. The finding that comparabl δP and δH are found from correlating the permeation rates of solv ents through living skin is a ne w input into this discussion. it has been possible to characterize some aspects of the beha vior of human skin by HSP . The HSP for these correlations are included in Chapter 13. Ne vertheless.2 where the HSP for tetrabromobisphenolA (TBBPA) are reported along with the similarity of these with other rele vant materials. TBBPA is readily soluble in lignin (w ood). TRANSPORT PHENOMENA Many chemicals ha ve been the subject for concern in the past for v arious environmental reasons. and to be transported by animals. In addition. but reasons for this are not clear. An example is given in Table 17. i. The bad experience giving the warning includes a combination of all effects.. there is some discussion of the use of HSP in this respect in Chapter 15. birds. TBBPA has a distance from Pentachlorophenol (PCP) of only 3. rather than the potential dose ef fect only which is indicated by similarity of HSP .28 liquids in contact with viable skin This has been discussed in Chapter 15 in more detail. Some chlorinated materials. 26 is perhaps required to give improved predictions in marginal cases. Table 15. Earlier discussions also led to the impression that those in volved in this area did not consider the swelling of psoriasis skin as having relevance to the permeation of li ving skin. 27. such as pentachlorophenol. those near the boundaries of the HSP sphere describing the situation.5 units. The HSP found in a correlation of permeation rates of 26 are similar to those found for the swelling of psoriasis scales. ha ve the ability to penetrate skin and w ood.316 Hansen Solubility Parameters: A User’s Handbook only removes what is on the surface or what can diffuse to the surface during the washing process.8 units. 27 The lack of a skin penetration w arning for these liquids is partly attributable to the fact that this warning is based on experience. . b ut also relates to w orker safety. SKIN PENETRATION Human skin is a complicated system. TBBPA will also be soluble. An improved HSP correlation of the permeation rates of solv ents through living skin based on a lar ger number of solv ents than the 13 included in the w ork of Ursin et al. Dividing this by the radius of the sphere to find the RED number (relat ve energy difference) shows that it is well within the solubility re gion with a RED of 0. The principles discussed here can possibly contrib ute in other w ays to impro ve the recycling process based on the increased le vel of kno wledge.

bisphenol A. mono 2-ethylhexyl phthalate. The rate of absorption is dictated by the size and shape of the molecule.1 (RED = 0.R.0 17. Pharmaceutical and Medical Packaging 2001. misdirect. H.8 12. Copenhagen.1–20. The major problem with this collection is that the central nervous system is almost completely f atty in nature. HSP are a vailable in Appendix Table A. If the chemical stability is comparable or better . and an e xcess of foreign materials can shortcircuit. 2001.8 16.6 24.1 Ro — — 13.Applications — Safety and Environment 317 TABLE 17. These data confirm that TBBPA is readily soluble where pentachlorophenol is soluble. In principle an y compound of interest can be assigned HSP for use in predicting behavior in connection with a lar ge number of en vironmental subjects. Similar analyses can be done with other chemicals. TBBPA will readily absorb into the outer layer of the skin (k eratin) here described by swelling of psoriasis scales. HSP pro vides other insights with re gard to . The same is true of related brominated compounds in general. the solubility of a protein (human blood serum).5 11. and v arious glycol ethers based on ethylene oxide.8 Dist.2 10. or stop messages. CONCLUSION In conclusion. The size and shape of the molecule will dictate the rate of permeation. to TBBPA 0.5 6.36) 9. One can also use HSP to more rapidly arrive at an optimum choice of chemical protecti ve clothing. pp. and not by the solubility relations. They will also be readily taken up at lower concentrations without delay by human skin as shown by the correlations of rapid skin penetration and the swelling of psoriasis scales. δD 20..3 (RED) = 1.5 (RED = 1.8 (RED = 0. Source: Reprinted with permission from Hansen. b ut that it will ha ve some solubility in the f atty tissues as well.7 5.. then it can be presumed to appear at the same places if it gets into the en vironment. C. Sk ov. A simple rule of thumb is that lack of w ater solubility will encourage collection in the f atty tissues. This is just outside the region for rapid permeation.5 δP 9.9 17.9 11.1 6.1 12. Both pentachlorophenol and TBBPA can readily penetrate w ood and w ood products. 20.73) TBBPA has a distance from the center of the spherical correlation for rapid skin permeation of 6.3 (RED = 0. etc.4 22. leading to memory problems. not the solubility relations.3 δH 13.. and swelling of k eratin (psoriasis scales swelling) indicate that TBBPA prefers to reside in the nonf atty tissue.5 21.1 for man y phthalate plasticizers. Conference Proceedings. The collection of chemicals in f atty tissue is another topic that could be e xplored. N-methyl-2-pyrrolidone.M.0 12.2 The Affinities of Tetrabromobisphenol A (TBBPA) for Selected Biological Materials Material TBBPA Pentachlorophenol Lignin solubility correlation Rapid skin penetration correlation Swelling of psoriasis scales correlation Depot fat (37 °C) total solubility Blood serum Note: Units are MP a1/2.9 5.0 19.49) 14. but means that permeation will certainly take place at a moderate rate.0 3. He xagon Holding. The correlations of the solubility of depot f at at 37 °C. but this could be given more precision with HSP.10.9 14.2 21. Ed.9 25.50) 6.5.6 15. it can be noted that HSP pro vides a tool to aid in substitution and in systematic formulation of less hazardous products and processes.9 1.20 13.0 12.

C. 301 af 13 maj 1993. and Petersen. Dev.A.M. Assigned to CPS K emi Aps (Now a part of the Autotype/MacDermid Concern). 1983. Madsen. Chem. Systematic Ev aluation of Risk Potentials of Solv ents. 1997. and Hansen. Holten-Andersen... pp. C. and Hansen.L. January 1993. Copenhagen.. Altnau. Nordic Conserv ation Congress Preprints. EP 0 270 654 B1. 1989. J. 185. C. L. Filskov.. 34–37.. 9. C. 20–23. REFERENCES 1. C. Dan. Forsberg. C.. Substitution in Practice — Experience from BST (Substitution i Praksis — Erf aringer fra BST). 16. C. Prog. 11(3).H. Ev aporation from high solids coatings (A vdunstningen Från LFFårger). S.. Anonymous. M. 12. A... Coat..M. Org. Madsen. 1994. G. Copenhagen.M... Goldschmidt. and Hansen..M. 54. Kemi.. Eds. Hansen. Hyg. G. V. Progress in Chemical and Biological Research 220.M. ASTM STP 989. 1. C. 6/98. maj 1993. 16(3). Färg och Lack.. 1982. Conserv ation and Solubility P arameters. 266–268. 10(3). in Danish). 103(9). and Hansen. Austin. 36–43. C. H.. Johansen. K. Res.M.. TX. K. Hansen.. Coat. Assigned to CPS K emi Aps (Now a part of the Autotype/MacDermid Concern). Hansen. 5. Substitution: a method to fulfill the orking environment law for air quality (Substitution: En Metode til at Overholde Arbejdsmiljoelovens Krav til Luftkvaliteten. 15. and Hansen. 10. Liss. Saarnak. Directi ve on Determination of Codes ( Bekendtgørelse om fastsættelse af kodenumre. 11.P. J. in Danish). Anonymous. Directive on Work with Products Having Codes (Bekendtgørelse om arbejde med kodenummererede produkter. 1988. K.. 1987.J. 1995. C. Solv ent and w ater e vaporation from coatings..M. Solv ents in w ater-borne coatings. 100–105.. 547–553. in Danish). L. E. 1986. in Danish).. Arbejdstilsynets bekendtgørelse nr. Wu. 13. 4th ed. 454–457. Solubility parameter prediction of the barrier properties of chemical protective clothing...K. Arbejsmiljøfondet.. American Society for Testing and Materials. Risik opotentiale v on Lösemitteln systematisch be werten. Pedersen. Soerensen. Ne w York. Hansen. 219–240. Occup. D. and Nielsen. J.. 2(9). Riihimäki. 67(5). G. Solv ents for coatings. Farbe+Lack. 4. pp.H. Färg och Lack. Coat. Polym. 19. Hansen. Inc. 197–208. F ormulating solv ents to remo ve hazardous air pollutants.. L. Danish Directorate for Labor Inspection. Prod.. Instant Reference Sources. Substitution of hazardous chemicals and the Danish e xperience. 17.Z. 261–278. 1998. The systematic choice of material in personal protection equipment for or ganic solvents in safety and health aspects of or ganic solvents.M. Am.M. Eur. 297–302. Olsen. Paint Colour J. and Wibroe.S.M. 14. 3. Chem. 1991. F. Hansen. R. 1977. 7. . Danish Directorate for Labor Inspection. 6. An analytical approach for reducing workplace health hazards through substitution... Ind. 18.. Goldschmidt. and Kieth. P.. 8. C. Arbejdstilsynets Bekendtgørelse nr..M. 1–13. Hyg. Eds.. Organic solvents in high solids and w ater-reducible coatings. Technol.M.H.302 af 13. 2. Which glo ves should I put on? (Hvilk e Handsk er Skal Je g Tage På?. December 1995. Assoc.. in Danish). Hansen. and perhaps e ven in a wide variety of other materials such as those found in nature. EP 0 205 505 B1. 45–49. 20.. Performance of Protective Clothing: Second Symposium. and Hansen. Chemical Protective Clothing Performance Index. in packaging materials. T. 146–153. 1991.. 1972.T. Ind.. 1988. Hansen. Sager. Alan R. Prog. 33(3). pp.M. Höglund.. C. and Ulfv arson. Mansdorf. Philadelphia. in Swedish. 30(5).. 331–352. Eng. C. U.318 Hansen Solubility Parameters: A User’s Handbook uptake of undesirable chemicals in the human skin. PA. 1984. 1986. Org. C.

Impro ved methods for the determination of Hansen’s solubility parameters and the estimation of solv ent uptake for lightly crosslinked polymers. and Andersen. Hoersholm. C. P. Anna. Van Dyk. D.T . 1996.. Polym.P. Appl. Scandina vian Paint and Printing Ink Research Institute.. X. Critical analysis of the graphical determination of Hansen’s solubility parameters for lightly crosslink ed polymers. Appl.M. Polym. 25.O.H. The Performance of Protective Clothing. breakthrough times.. 1993. Report No. . 50.T. Ursin.. 27. 56. J. Am. 301–308. 28.. C. R... C. 2069–2080. 49(6). E.. 531–540.. C. 1988.W. Sulewski.T. ASTM STP 1133.. J..B.. 1992.P. J. Appl.M.H. 1995.T. J. E. Three-dimensional solubility parameters and chemical protecti ve clothing permeation. 23.. American Society for Testing and Materials. Polym.. C. 513–530.M. Jensen. 651–660. N. J.. Hansen. R.M.. Modeling the solubility of or ganic solvents in Viton® gloves. J. Anna. G. E. Am. J. C. Zellers. Hyg. 24. McBriarty. Hyg. 22. D. and Wei. and steady-state perme ation rates of or ganic solvents in Viton® gloves. A. X. and Zhang. Sci. Sule wski.. 2081–2096.. Eds.J. 26. Ind.Applications — Safety and Environment 319 21.H. Three-dimensional solubility parameters and chemical protecti ve clothing permeation. Selection and use of molecular parameters to predict permeation through fluoropolyme -based protective clothing materials. Denmark. 50. Zellers. 62. T 3-82.W.-Z... Hansen.. Zellers. 1982. Sci. B. Assoc. and Bentz.. I. Ind.. Philadelphia. Assoc. J.. II. Appl. Fourth Volume. Hansen. 62. E.. Zellers. 1996.. Sci. and Ebbehoej. Permeability of commercial solv ents through living human skin. and Henry.. Modeling diffusion coefficients. and Wei. 1993.. Sci. 894–907. Hansen. Billing. The affinities of o ganic solvents in biological systems. Christensen. P A.. The Absorption of Liquids into the Skin. J. I. pp. Polym.


The second matter of concern is ho w reliable the HSP data are and ho w accurately the correlations can predict the beha vior of untested systems. This handbook attempts to help impro ve this situation by publishing HSP for a lar ger number of liquids.2. INTRODUCTION There are many matters related to HSP which still need clarification and xpansion. Problems and situations clearly needing further attention are discussed in the follo wing. This is important. The first matter of concern is the vailability of data. as sh wn in Chapter 2. Table A.18 The Future Charles M. In the future. This handbook also contains ne w HSP correlations not present in the firs edition.1. as it confirms that the HSP correlations must continue to include a constant not too di ferent from the currently used “4” (or 0. These findings h ve been supported more recently by the statistical thermodynamics approach of P anayiotou and co workers summarized in Chapter 3. Hansen ABSTRACT Hansen solubility parameters (HSP) help to quantify the statements “lik e dissolves like” and “lik e seeks like.25). Other sources are discussed belo w. about 1200. Ho wever. Qualitati ve indications of this for the 321 .” These parameters have found use in man y fields of research and practice. HSP ha ve extended the original Hildebrand single solubility parameter approach by quantitati vely taking into account the molecular permanent dipole–permanent dipole and molecular h ydrogen bonding (electron inter change) interactions. but others are not so clear . One need not necessarily extend its theoretical scope to accomplish this. Many of these are gi ven as examples in the te xt. swelling. spontaneous spreading). Based on the large number of current uses of HSP. The general predictions possible for these ph ysical interactions ha ve been demonstrated for both b ulk phenomena (solubility . Neglecting this differentiation will lead to misinter pretations. the following discussions will start with more practical topics. dewetting. the conclusion must be that the geometric mean can be used to average not only dispersion interactions b ut also those attrib utable to permanent dipoles and to hydrogen bonding. the theory should be e xplored and used with this general applicability in mind. a glimpse has been gi ven of a v ery general ener getic approach to systematically predict and control molecular interactions among man y materials of widely dif ferent composition. As the agreement is general. Table A. This same mean must be used in the Prigogine corresponding states theory if agreement is to be found with the HSP correlations presented in this book. As this book is primarily directed to ward the practitioner. primarily in the Appendix. HSP and the Prigogine corresponding states theory of polymer solutions are mutually confirming with r gard to treatment of specific interactions. compatibility) and surf ace phenomena (adsorption. and others are included in the Appendix. The existing data can be used in a strictly empirical manner if so desired. Some limitations are clear. primarily becaus their unique predictive capabilities are based on sound theoretical principles. This is necessary to dif ferentiate between the atomic ( δD) and the molecular (specific) interactions δP and δH). one can easily suppose that there are man y more practical uses which remain to be discovered and developed. This approach allo ws independent calculation of each of the three parameters. The geometric mean a verage for the interaction of unlik e molecules is inherently used in the Hildebrand approach and in the HSP approach as well.

Another concern related to reliable HSP v alues is based on the f act that most chemicals in the intermediate molecular weight range. the problem is similar to measuring the surface tension of a surface which has such a high v alue that even water spontaneously spreads on it. for e xample. It is impossible to establish the three HSP based on such data. for the reliability of the “original Hansen” approach does not correspond to this e xperience. then the number of “errors” will increase. A situation is often met where only a few solvents having high solubility parameters dissolv e a polymer which has still higher HSP . for e xample. and these usually ha ve another explanation for their behavior.0. Logical applications for HSP will be found in self-assembling systems and in what is called nanotechnology . The solubility of De xtran C (British Drug Houses) 1 is an e xample of this as sho wn in Table 18. See Chapter 5 and Chapter 7 for further discussion of this problem which is present for both polymers and particulate matter. engineering approximations leading to a systematic course of action ha ve been possible using data which is currently a vailable. This was the case for the correlations presented in Table 5. so there are no pure liquids which can be used to test the v ery high ener gy materials.4A. such that the interactions among still more materials can be predicted. such as that characteristic of plasticizers. Normally there are perhaps 5 or 6 boundary solvents that are not predicted correctly out of about 100 test solv ents in a correlation of experimental data. One can only conclude that its surf ace tension is greater than that of w ater.9 which includes the constant “4. Equation 1. In man y cases. Ne w thinking and ne w techniques are required to accurately characterize such high energy materials. or ganometallic materials. such as very large molecular size. The reason for this discrepancy is not kno wn.1. One e xample is the self-stratifying paints discussed in Chapter 8. Similarly. and he xane. These surfaces are clearly wet because of the lo wer surface tension of all of the liquids. Very high cohesi ve energies lead to the formation of solids. In the present case. A third point which is sometimes irritating is that the scope of the characterizations possible is limited to the cohesi ve energy spectrum of the test liquids.322 Hansen Solubility Parameters: A User’s Handbook data generated by the author are gi ven in the rele vant tables.2. The given good solv ents are often in the boundary re gion of the HSP spherical characterizations. w ater. only a fe w solv ents may interact intimately with a surf ace which has v ery high HSP . very small molecular size. Another is the ultrastructure of cell w alls in w ood discussed in Chapter 15. decisions and ideas are needed on ho w the data should be used. glycerin. One can often arri ve at a prediction for e xpected behavior using the “lik e seeks lik e” principle. b ut only a few with high HSP prolong suspension of finer particles. There are very rarely major outliers. In a sense.1 and Table 18. e ven for these v ery high ener gy characterizations. for xample. and salt solutions might be helpful in these situations (see the following corresponding sections). A full understanding of the beha vior of w ater. are soluble in almost all of the test liquids. It is hoped that this book will aid in the generation of still more HSP data ha ving a uniformly high quality . as can be seen from those included in this handbook. . especially for polymers. In general correlation coef ficient approach 1. reactions. One generally has to rely on group contrib ution methods or other calculations or comparisons. It is here that the e xisting theory and future extensions of it are most important. rather than the ratio of 99 correct answers with 23 incorrect answers indicated in Table 4. Table 4. e ven though accurate numbers and an appropriate detailed theory may be lacking. there is a model to e xtrapolate HSP to higher v alues than can be measured directly . The energy characteristics of such surf aces are apparently higher than those of an y liquids which can be used to study them by these techniques.4A. e xcept for . but if group contrib ution or other estimates are in volved. or the lik e. The current practice is to extrapolate into the region of v ery high HSP using Chapter 1. ” It is assumed that this constant is still v alid. The experience reported in Chapter 4. This indicates perhaps only a fe w minor outliers in the correlations. Once the necessary reliable HSP data are a vailable. and there will be some uncertainty in volved with this.

8 12.7 104.346 1.0 16.172 1.7 20.8 19.7 106.1 11.2 18.795 1.1 0.6 14.2 26.6 0.3 6.5 19.8 73.3 7.0 16.3 131.8 98.0 10.8 131.6 97.5 17.3 6.7 16.4 22.211 1.5 17.9 104.003 1.1 6.4 14.2 15.2 7.7 63.0 71.1 4.8 15.1 18.1 18.0 9.4 19.144 1.7 19.454 1.3 2.082 1.336 1.3 8.4 55.6 19.756 1.8 3.1 102.3 3.4 8.4 11.0 17.2 17.4 δH 7.4 0.4 21.0 73.517 1.080 1.0 15.2 4.6 150.1 10.2 5.2 20.000 1.4 19.6 0.0 16.5 76.371 1.1 4.559 1.8 117.5 15.2 12.8 102.3 3.406 1.474 1.9 13.0 97.9 91.3 7.5 59.0 5.0 16.1 23.601 1.0 15.3 16.8 6.9 58.5 79.9 127.0 117.5 40.2 17.2-Dichlorodiethyl ether Diethylene glycol Diethyl ether Dimethyl formamide Dimethyl sulfoxide 1.2 7.0 15.4 17.8 79.9 16.1 Calculated Solubility Sphere for Dextran C Solubility The Solvents with Their Parameters Solvent Acetone Acetophenone Aniline Benzaldehyde Benzene 1.053 V 74.0 5.8 15.7 130.3 14.241 1.991 2.1 5.077 1.3 20.7 5.0 5.8 16.4 1.6 18.915 0.1 39.2 6.2 112.2 16.0 8.0 29.3-Butanediol 1-Butanol Butyl acetate gamma-Butyrolactone Carbon disulfid Carbon tetrachloride Chlorobenzene Chloroform m-Cresol Cyclohexanol Cyclohexanone Diacetone alcohol o-Dichlorobenzene 2.880 1.0 17.2 9.9 85.180 0.3 10.640 1.479 1.246 1.8 60.313 1.7 3.485 1.8 10.4 89.7 16.037 1.0 3.8 19.7 54.4 17.8 16.1 9.6 2.8 18.9 13.0 20.6 16.3 SOLUB 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1* 0 0 0 1 0 0 1* 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 RED 1.1 4.4 16.3 9.411 1.5 15.5 5.0 104.6 15.054 1.3 86.0 17.0 124.6 .4 18.556 1.3 85.8 19.313 1.The Future 323 TABLE 18.4 18.3 6.1 5.8 77.5 101.6 5.7 12.5 5.8 17.2 12.4 19.0 12.0 2.416 1.399 1.1 7.8 δP 10.0 4.3 5.0 21.312 1.410 1.3 4.6 8.2 125.4 0.2 94.776 1.5 132.3 0.4 91.858 1.8 19.0 7.679 1.473 1.363 1.7 10.4 15.4-Dioxane Dipropylene glycol Ethanol Ethanolamine Ethyl acetate Ethylene dichloride Ethylene glycol Ethylene glycol monob utyl ether Ethylene glycol monoethyl ether Ethylene glycol monomethyl ether Formamide Glycerol Hexane Isophorone Methanol Methylene dichloride Methyl isobutyl carbinol Methyl isobutyl ketone Nitrobenzene Nitromethane 2-Nitropropane Propylene carbonate Propylene glycol δD 15.5 89.1 80.1 26.0 12.1 8.170 0.0 0.1 12.1 4.000 1.

4.3 SOLUB 0 0 0 0 RED 1. The first edition of this handbook contained ver 800 HSP values for chemicals. Table A. HANSEN SOLUBILITY PARAMETER DATA AND DATA QUALITY The author and others including most solv ent suppliers and some paint companies (at least) ha ve databases including HSP data for solv ents and HSP correlations for polymer solubility etc.1 δH 8.0 5. as not only is there a lack of possible data. has indicated a computer program based on 6. See Chapter 1 for details of the calculations.3 δP = 19.0 106.5 Ro = 17. The problem of estimating a sphere based on limited data which do not experimentally define the whole sphere becomes more acute.11. δP.9 δH = 22.9 Some of the v alues have been revised o ver the years.2–5 There is still a tendency to regard the contents of such databases as proprietary information for the benefit of th owner and/or his/her customers.999 NO = 50. The cohesion energies are so high that no liquids are possible and only solids are present. The calculated values have been checked against performance data reported in the literature where this has been possible. HSP correlations in addition to those given in connection with examples in the text are included in Appendix.8.12 but many correlations based on these data are included for reference anyway.2.0 δP 5. smaller molecules whose HSP are kno wn can be used to derive group contributions for estimating the HSP of lar ger molecules wherein they appear as a part.1 (CONTINUED) Calculated Solubility Sphere for Dextran C Solubility Solvent Tetrahydrofuran Tetrahydronaphthalene Toluene Trichloroethylene δD 16. In some cases data for whole. 2. The methods described in Chapter 1 were used.9 2. Table A. Exxon. the δD.324 Hansen Solubility Parameters: A User’s Handbook TABLE 18. The materials gi ven in dark type ha ve had some de gree of e xperimental verification.450 1. An example of a good HSP correlation from the solvent range studies is that of polyeth ylene sulfide.2. for e xample. This has been e xpanded to about 1200 v alues in the second edition. This is sho wn in Tables 18. b ut much of the v olume of the HSP sphere is located where there are no liquids. and 500 pesticides are included. The solvent range chosen for the studies does not completely fill out the possibilities selection o different liquids w ould have allowed.2 Note: δD = 24. The use of these parameters is becoming so commonplace that. Table A.4. The solvent listing in the Appendix.618 1.4 3.3 and 18.6 18.5. An example is the solubility data reported for poly(vinylidene chloride) (PVDC).8 90.8 19. All the others are based on calculations only .10 Appendix. although in man y cases data w as lacking to such an e xtent that group contrib utions were used.0 1.4 FIT = 0. A comparison of .744 1. There are many additions to the original set of data. There are limitations on the accuracy of the HSP data deri ved from Burrell’s solvent range studies reported in standard reference works.560 V 81. This polymer has relati vely low HSP.7 2. This seems to be satisf actory. Lo w molecular weight solids with relati vely low melting points ha ve been treated as if the y were liquids for e xtrapolation of latent heats to 25°C. Tables of HSP data for man y materials are also included in standard reference w orks. and the solvents in the test series pro vide nonsolvents at higher HSP than those of the polymer to locate the boundaries with suf ficient accura y. Only data judged (reasonably) reliable are reported.0 18.1 also includes HSP for a number of lo w molecular weight solids. 450 resins.7 136. and δH parameters often appear without an y specific reference to where th y came from or what the y actually represent. This problem is greatest for polymers with high HSP .7 HSP where data for over 500 solvents and plasticizers. in man y studies. and it is consistent with the treatment of high boiling liquids.0 2.1 includes the pre viously published set of some 240 solvents which ha ve appeared earlier in se veral sources.

0 96.053 1.2 100.6 13.5 84.136 1.4 9.4 18.0 18.3 94.2.2-Tetrabromoethane Ethylene glycol monoethyl ether N.068 1.1 10.8 92.6 23.4 7.172 1.7 8.258 1.3 40.0 8.4 16.5 104.2 12.3 7.0 δP 19.220 1.054 1.8 16.3 11.0 130.2 13.8 17.4 16.3 21.219 1.0 29.121 1.8 9.6 16.000 1.9 108.3 10.3 17.180 1.2 8.9 55.5 175.2 18.7 86.915 0.4 12.3 21.9 87.2 Calculated Solubility Sphere for Dextran C Solubility The Solvents with Their Parameters Solvent Succinic anhydride Triethanolamine Dimethyl sulfone Ethylene cyanohydrin 2-Pyrolidone Ethanolamine Formamide Diethanolamine 1.5 83.2 20.8 5.4 59.9 16.7 26.1 85.3 12.1 10.5 79.2 18.4 16.7 5.0 16.152 1.0 58.7 18.3 21.132 1.216 1.0 68.2 21.2 SOLUB RED 0.2 17.8 95.1 16.6 11.6 12.0 21.0 16.3-Butanediol Glycerol Dimethyl sulfoxide Diethylene glycol Ethylene glycol Propylene glycol 1.2 123.0 18.6 17.250 1.0 114.846 0.8 17.972 0.0 68.8 97.147 1.3 19.7 δH 16.0 171.082 1.1 14.2 11.5 5.0 13.8 39.8 7.277 V 66.7 19.0 16.867 0.4 15.9 17.0 11.5 17.8 73.000 1.167 1.265 1.0 16.3 15.984 0.8 8.0 17.4 18.152 1.4 19.0 87.1 9.2 75.077 1.N-Dimethyl acetamide 3-Chloro-1-propanol Hexylene glycol Methyl-2-pyrrolidone Furfural Aniline m-Cresol 1-Propanol Benzoic acid Triethylphosphate Quinoline Diethylene glycol monoethyl ether Acetic anhydride δD 18.272 1.4 4.246 1.1 16.259 1.0 9.2 11.4 12.5 16.5 115.0 23.0 9.7 75.0 17.080 1.5 14.241 1.2 19.1 17.1 6.2 20.880 0.8 133.063 1.1 5.3 12.0 22.4 17.5 73.8 6.8 103.2 19.4 109.4 10.230 1.2 14.8 12.4 18.144 1.0 17.2 91.6 15.7 17.4 19.8 13.3 14.2 7.2 10.9 77.144 1.3 10.3 76.0 118.4 12.215 1.6 66.4 17.211 1.0 15.6 15.2 5.6 16.7 16.8 130.7 5.3 16.6 19.3 18.6 17.2 18.3 71.2 16.2 14.739 0.866 0.991 1.0 76.5 116.5 1 1 1 1 0 1* 0 0 0 0 0 0 0 0 0 0 0 0 .9 11.3 17.6 8.4 16.170 1.1 19.The Future 325 TABLE 18.6 89.2 10.5 16.199 1.5 15.8 22.7 67.0 19.2 16.9 6.9 5.8 8.6 10.9 94.7 11.7 18.4 8.5 26.7 10.6 7.4 18.3-Butanediol Diethylenetriamine Triethyleneglycol gamma-Butyrolactone Dipropylene glycol Dimethyl formamide Allyl alcohol o-Methoxyphenol Hexamethylphosphoramide Ethylenediamine Methanol Furfuryl alcohol Trimethylphosphate Benzyl alcohol Ethylene carbonate Phenol Ethylene glycol monomethyl ether Propylene carbonate Ethanol 1.9 9.117 1.819 0.0 18.003 1.1.2 22.

14 are not compatible with those of the author . and ha ve lo wer predicti ve ability than had w ater been ne glected. The small molecular v olume. An example of a poor correlation using solv ent range data is that of the solubility of polyvinyl alcohol. . such that it can be included in studies leading to HSP correlations.0 14.0 37. ethanol and 2-propanol.1 tak es full coverage into account.0.9 δH = 22.2 (CONTINUED) Calculated Solubility Sphere for Dextran C Solubility Solvent Tricresyl phosphate Formic acid Tetramethylurea δD 19.0.7 δP 12. A qualitative indication of the reliability of the correlations is gi ven for this reason. the group contrib ution procedure presented by van Krevelen and Hoftyzer 15 does not give satisfactory agreement with the procedures given in Chapter 1. This does not pre vent use of such data as demonstrated else where in this book. the three partial solubility parameters tabulated by Hoy13. some of these liquids must be handled with care for reasons of toxicity.18. A goal of future w ork should be to be able to account for the beha vior of w ater in a reliable manner .8 120. the solvents included in Table 18. Data for chemical resistance.δP. permeation. Reliable HSP data for man y polymers of practical importance are not a vailable at this time.1 and the follo wing discussion).4.4 FIT = 0. Finally . One of the problems with some of the reported correlations in the Appendix.0.9.0 SOLUB RED 1.6 11. exceptionally high δH parameter. Including them in a possible future edition of this book may also be a possibility . The use of these data is not recommended.9 8. For the sak e of completeness.5 Ro = 17. This leads to a correlation with the following data: δD.3 with those in Table 18.3) appear promising for some applications where the HSP v alues for water as a single molecule are clearly not applicable. There are shortcomings in terms of lack of full co verage of the HSP space as well as in the total number of liquids for which there are data. As discussed in the ne xt section.285 V 316.999 NO = 50. Table A. Only tw o of the solv ents. Ho wever. This results in HSP values which are less reliable.284 1.3 16.0 in MPa1/2 with a perfect fit for t o good solv ents out of 56 in the set of data.3 δP = 19. b ut it does place some limitations on the reliability of the predictions obtainable from the correlations.δH.2 is that the data on which the y are based were not generated for this purpose.277 1.1 sho ws which ones are lacking in the high HSP range. Note that a standard set of test solv ents such as that used in Table 18.Ro equal to 17. and other phenomena related to solubility which can be correlated with HSP are practically ne ver accumulated with an HSP correlation in mind. As noted in Chapter 1.2 δH 4. Mixtures of or ganic solvents with w ater are still more problematic when used as test liquids (see Figure 18. dissolv e it.3 11.4 Note: δD = 24. and tendency to self-associate depending on the local environment all lead to the lik ely result that w ater will be an outlier for the correlation. It would seem advisable for ra w material suppliers to determine the HSP for their rele vant products in a reliable manner and to publish these data on their product data sheets or else where. Ro is clearly too small by comparison with Ro found in HSP correlations for solubility for other water-soluble polymers.326 Hansen Solubility Parameters: A User’s Handbook TABLE 18.5 16. water (or its mixtures) should not be included currently in any HSP correlations without a v ery careful analysis of the results. The HSP v alues for w ater found from the correlation for total w ater solubility reported in Chapter 1 (T able 1. a couple of w arnings are appropriate before proceeding to the next section.

352 2.2 3.0 85.1 4.0 3.1 0.0 13.600 0.6 91.308 1.699 2.3 0.077 2.0 19.480 4.3 5.3 9.1 5.8 5.3 7.2 16.3 5.2 22.3 3.1 18.6 97.3 2.7 15.4 17.7 90.2 92.0 2.1 15.993 2.869 2.551 V 57.898 2.0 3.3 21.0 5.697 1.2 5.9 12.947 1.1 102.006 0.919 1.1 74.7 1.2 3.4-Dioxane Ethanol Ethyl acetate 2-Ethyl hexanol Ethylene carbonate Ethylene glycol Ethylene glycol monob utyl ether Ethylene glycol monoethyl ether Furfural Furfuryl alcohol Glycerol Isoamyl acetate Isoamyl alcohol Isopropyl acetate Methanol Methyl acetate Methyl ethyl k etone Methyl n-amyl k etone Nitroethane Nitromethane Octane δD 14.4 17.155 2.1 26.0 6.8 15.8 17.8 109.771 1.9 8.6 66.4 103.0 5.4 16.2 14.7 89.9 19.3 15.7 108.286 6.0 SOLUB 0 0 0 0 1 0 0 0 0 1 0* 0 0 0 1* 1 0 0 0 0 0 0 0* 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 RED 3.2 9.5 16.8 83.491 6.825 3.4 18.1 75.1 10.3 75.5 156.1 4.0 15.4 19.0 15.5 7.7 58.604 2.0 15.0 14.2 4.5 δP 8.8 12.5 98.8 17.9 0.744 1.5 97.0 16.6 7.7 17.7 5.3 15.462 1.1 40.483 1.2 18.0 16.0 4.5 112.6 12.5 15.0 6.8 5.8 14.1 2.019 3.970 0.954 1.1 0.3 19.5 73.2 2.325 2.5 54.3 7.7 11.9 7.5 12.973 3.8 117.4 15.2 11.2-Dichlorodiethyl ether Dichlorodifluoromethane (Freon 12 Dichloromonofluoromethan Diethyl amine Diethyl ether Diethylene glycol Di-isobutyl ketone N.4 19.5 133.286 1.5 89.6 17.0 55.123 0.0 5.1 9.772 4.0 0.3 8.2 18.5 16.3 104.8 71.9 177.1 29.3 163.8 14.7 6.3 7.3 148.916 1.1 5.9 15.0 12.0 5.8 15.1 118.107 1.1 3.1 10.2 5.9 17.1 139.3 Calculated Solubility Sphere for Polyethylenesulfide The Solvents with Their Parameters Solvent Acetic acid Acetone Acetonitrile Aniline Benzene 1-Butanol sec-Butyl acetate Butyraldehyde Carbon tetrachloride Chlorobenzene p-Chlorotoluene m-Cresol Cyclohexane Cyclopentanone 1.125 0.2 11.752 3.1 15.4 18.634 3.6 5.5 5.8 94.605 2.1 92.3 6.0 52.7 0.N-Dimethyl acetamide Dimethyl formamide 1.8 0.0 15.0 2.8 15.0 16.3 7.476 1.0 11.6 2.5 15.0 16.0 16.1 20.4 91.043 5.521 4.631 1.898 1.2-Dichloro ethylene (cis) o-Dichlorobenzene 2.7 79.6 88.852 1.8 8.6 12.7 11.5 13.0 10.5 77.The Future 327 TABLE 18.0 19.7 4.3 14.5 18.9 14.2 104.1 0.7 3.038 3.0 5.4 117.8 7.2 86.6 16.5 .8 131.0 δH 13.4 7.285 3.793 2.8 19.

0 5.0 1.3 0. The use of various predictive methods which arri ve at dif ferent results has al ways been a problem.5 4. δH = 2.0 123. This radius is larger for smaller molar volume in this inverted approach.724 V 157.8 70.3 (CONTINUED) Calculated Solubility Sphere for Polyethylenesulfide Solvent 1-Octanol Pentane 1-Pentanol 2-Propanol Propionitrile Propylene carbonate Styrene t-Butyl alcohol Tetrahydronaphthalene Xylene δD 17. although man y other authors ha ve chosen to do so.0 14.6 76. Values added o ver the years are appended to the original table which w as attributable to Beerbo wer.6 95.3 18.913 3.21 but it has never been fully explored.2.9 15. GROUP CONTRIBUTION METHODS Suggested calculation procedures to arri ve at the HSP for solv ents are gi ven in Chapter 1.7 2.8 15.0 1. NO = 56.8.005 3. This idea was presented many years ago.0 13. The group contrib ution v alues consistently used by the author are reported in Chapter 1.933 3. which was of primary interest.0 δH 11. K oenhen and Smolders 19 evaluated various equations for predicting HSP . New group contributions should be checked for reliability of the predictions in some w ay. Ro = 4.655 0.317 0.1 SOLUB 0 0 0 0 0 0 1 0 1 1 RED 2. This is discussed in Chapter 5.1 14.1 2.328 Hansen Solubility Parameters: A User’s Handbook TABLE 20. The δD parameter.6 15. POLYMERS AS POINTS — SOLVENTS AS SPHERES One way to possibly improve predicting the behavior of polymers is to consider them as points (or more accurately.6 17.5 6.17.981. is not predicted well.9 0.996 0. 2 As stated pre viously.18 Barton has also collected man y tables of group contrib utions for various purposes. in particular .642 2. Methods for reliable a priori calculation of the HSP for polymers are not a vailable.948 3. δP = 3. as is presently done. The author chose not to use these at an early date. Calculation of the radius of interaction is a particular problem in this respect.0 18. This is a serious shortcoming. A given solvent is assigned a rather large radius of interaction.9 3. which is not al ways possible within the timeframe of most projects.4 5. Chapter 2 discusses some of the actors which must be taken into account when calculating the radius of interaction.2 19.401 1.1 4.8 136. spheres. it can be assumed that some of the inherent errors will not af fect relative evaluations.5 15.6 δP 3.1. Utracki and co workers20 estimated HSP for a number of polymers assuming their δD parameters were not different and group contributions for the δP and δH parameters. If one consistently uses the same method of estimating HSP.1 14. The group contribution methods need e xpansion with new groups. FIT = 0.9 85. so there was no need to start .2 108. The first indication were that there seemed to be no real benefit in terms of impr ved reliability of predictions for polymer solubility in organic solvents.7 116. with v ery small radii of interaction that depend on molecular weight) rather than as spheres with lar ge radii.8.0 115. the group contrib utions tabulated by v an Kevelen15 have not been found reliable. and almost e very time has ultimately resorted to experiment. This is definitely an area requiring attention.9 16.3 Note: δD = 17. The author has tried several times to calculate the HSP for indi vidual polymers using the same group contributions suggested for the liquids.0 4.

0 5.9 12.0 15.8 75.0 9.6 15.2 14.2 5.1 5.6 0.2 17.6 4.2 5.0 15.6 71.4 17.0 3.1 2.8 18.2 10.7 116.2-Dichlorodiethyl ether Methyl ethyl k etone Isoamyl acetate Diethyl amine Diethyl ether sec-Butyl acetate Methyl acetate Pentane Butyraldehyde Isopropyl acetate Cyclopentanone Aniline Acetone 1-Octanol 2-Ethyl hexanol m-Cresol Ethylene glycol monob utyl ether N.3 2.973 0.9 108.0 97.5 74.0 6.5 95.019 1.8 16.5 14.352 3.0 16.6 5.3 15.3 4.0 4.1 14.1 3.2 88.0 δP 4.0 14.4 16.3 15.7 131.9 14.3 3.7 8.0 3.5 76.3 10.5 2.898 2.3 13.0 15.724 0.7 156.933 1.993 0.286 3.2 0.8 57.8 12.0 16.2 109.631 2.0 5.825 2.043 2.The Future 329 TABLE 18.2 14.1 91.3 18.8 15.9 18.5 117.948 3.9 19.2 8.9 5.7 75.1 SOLUB 1 1 0* 1 1 1 0* 1 0 0 1* 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 RED 0.521 2.744 1.8 12.4 Calculated Solubility Sphere for Polyethylenesulfide Solvent Chlorobenzene Xylene p-Chlorotoluene Styrene o-Dichlorobenzene Benzene Di-isobutyl ketone Tetrahydronaphthalene Carbon tetrachloride Methyl n-amyl k etone 1.0 .9 4.947 2.1 10.7 7.107 2.2 18.038 3.3 4.5 92.325 3.8 16.5 13.8 103.1 13.480 1.2 1.0 86.6 104.3 118.5 98.7 2.285 2.6 19.1 7.N-Dimethyl acetamide Dichlorodifluoromethane (Freon 12 Furfural Isoamyl alcohol Propionitrile 1-Pentanol Nitroethane Dimethyl formamide Furfuryl alcohol t-Butyl alcohol Ethylene glycol monoethyl ether Acetic acid 1-Butanol 2-Propanol Propylene carbonate δD 19.401 2.0 5.1 9.7 5.605 1.4-Dioxane Octane Ethyl acetate 2.3 2.9 11.2 0.0 7.5 15.462 3.0 11.0 6.3 1.308 1.600 0.9 3.9 5.5 17.1 2.1 18.1 136.1 3.7 14.7 163.5 77.752 2.0 157.7 6.8 16.697 1.3 115.005 3.286 3.5 15.2 16.772 1.1 1.5 16.8 15.5 11.0 15.6 19.919 1.5 13.869 0.8 133.0 16.1 18.0 0.3 5.1 123.0 17.0 3.642 3.2 5.898 1.0 15.0 17.8 89.8 20.1 2.0 5.3 0.4 4.4 15.5 15.8 19.0 δH 2.4 3.551 1.1 11.6 4.2 7.2-Dichloro ethylene (cis) Cyclohexane Dichloromonofluoromethan 1.5 14.4 177.0 3.5 11.2 0.1 139.604 1.6 79.655 V 102.4 15.6 7.5 15.954 0.3 14.1 91.0 19.2 90.0 5.9 0.771 2.8 0.006 1.4 70.4 0.6 92.634 2.7 7.699 1.3 83.1 148.7 7.155 1.2 104.1 89.8 97.3 9.9 17.0 8.125 2.3 5.5 117.1 5.8 85.913 0.0 6.7 14.4 85.6 17.1 7.123 1.317 3.9 16.6 112.996 1.5 108.0 0.0 7.

This is primarily because such data are lacking b ut also because surf ace cohesion parameters may not be reflected by nominal ulk composition.3 Note: δD = 17.7 55. This was not explored.793 3. in addition to that in Chapter 6 and Chapter 7.8 19. comparing distances between small spheres is relatively easy.970 5. These include selection of an improved solvent.6 15. reduction of polymer molecular weight.3 15.8 21. ho wever.0 12. which in turn allows systematic study and design of surfaces for desired beha vior.3 58.0 17. there may be adv antages in terms of more reliable prediction of polymer–polymer miscibility.2. and modification of a polymer s HSP in a desired direction based on the HSP group contrib utions of its repeating unit or comonomers. all over again with this in verted system. depending on how it has been surface treated. for example.4 δP 18. This possibility was demonstrated many years ago.981.1 δH 6.3 26.4 5.22 As sho wn in Chapters 6 and 7.1 17. NO = 56.0 94. can have widely different surface cohesion parameters.8.476 4. for e xample.6 54.7 22. Ro = 4.330 Hansen Solubility Parameters: A User’s Handbook TABLE 18.1 19. One current problem is that blindly entering wetting or spontaneous spreading data into the usual computer routine for finding the HSP alues often leads to ne gative . On the other hand.0 12. Data on surf ace characterizations. CHARACTERIZING SURFACES The characterization of surfaces with HSP.8. δH = 2. However. guidelines for impro ving polymer–polymer miscibility are easily found in the present approach.077 6. a surf ace characteristic for a polyvinyl chloride or a polyethylene cannot be expected to be valid for all polymers normally said to be of these compositions.8 73. FIT = 0.8 15.1 20. Li ewise.8 8. The requirement of polymer miscibility will be that the respecti ve points (v ery small spheres) for the polymers must be very close to each other.3 11. There may also be additi ves which ha ve dif ferent compositions and which may ha ve migrated to the surf aces.4 16.483 6.1. This is practical and almost necessary from an experimental point of vie w as most solv ents are so miscible as to not allo w an y e xperimental characterization in terms of a solubility sphere.3. δP = 3. An exception to this is the data for w ater reported in Table 1.3 SOLUB 0 0 0 0 0 0 0 0 RED 3.1 5.0 18. Neither has the ef fect of adsorbed w ater been clarified.7 12. It appears that the relati ve simplicity of the surf ace characterizations discussed in this book would lead to their wider use. this type of approach can lead to a ne w understanding of surface phenomena. This type of comparison is sometimes dif ficult to ma e in the present approach where the degree of overlapping of rather lar ge spheres is used to estimate polymer–polymer miscibility . is still in its inf ancy. The HSP reported here are the center points of HSP spheres where the good solv ents are either those that are completely miscible or those that are miscible to only 1% or more. solvents are considered as points. or perhaps more correctly cohesion parameters (ha ving exactly the same numerical v alues).491 4. Traditionally.9 40.916 V 52.4 (CONTINUED) Calculated Solubility Sphere for Polyethylenesulfide Solvent Acetonitrile Nitromethane Ethanol Ethylene carbonate Diethylene glycol Methanol Ethylene glycol Glycerol δD 15. for example. are not provided here.0 29.5 66. The same basic pigment or fille .852 4. No fi ed rules of thumb have been established to estimate how much overlap is required for miscibility.

G. 281–287. With permission.) .The Future 331 40 30 δh/MPa½ 20 10 0 0 10 δp/MPa½ 20 FIGURE 18. Basic Violet 10) showing potential problems with incorporation of w ater mixtures as test solv ents (see te xt for discussion). 1976.I. Farbe und Lack.. (From Riedel. 82(4).1 HSP plot of characterization of Rhodamin FB (C.

Currently .17. Some special types of materials are singled out here as w orthy of still more attention in the near future. and sulfolane. 8. and aromatic (fragrances) materials. The cohesion ener gy parameters characteristic of fluorocarbons are too l w. Each surfactant must be assigned three sets of HSP .17. and the third one is required for predictions when the whole molecule is soluble in the system. The HSP for the h ydrophobic end can be estimated by the methods discussed in Chapter 1. they point to the possibility of quantifying af finities of sur ace active materials in terms of the cohesion energy parameters of their respective end groups. discussed later in the Or ganometallic Compounds section (Figure 18. one anticipates some degree of orientation of the ends of the surf actant molecule to ward regions of similar HSP . This is reflected by their cohesion ene gy parameters. This same reasoning applies to other polymers. Figure 18. Pigments and fillers need to be characterized. MATERIALS AND PROCESSES SUGGESTED FOR FURTHER ATTENTION Examples of the use of HSP for man y types of materials and phenomena ha ve been presented in earlier chapters.332 Hansen Solubility Parameters: A User’s Handbook numbers for one or more of them. It seems unlik ely that this is a coincidence. An example to help illustrate the type of prediction possible is to try to answer the question of where the h ydrophobic end of a gi ven surfactant will tend to preferentially reside. A surfactant with a fluorinated end wil not dissolv e in man y polymers where a h ydrocarbon end will. These include surf ace active agents.23 The statement “lik e seeks lik e” indicates that surface active agents should be extensively treated in terms of HSP. dimeth yl acetamide. although Beerbo wer has developed some aspects of a theory for gi ven situations. the best approach is to compare plots or e ven to just compare tab ulated data for the test solv ents to determine where two surfaces differ in af finities. as demonstrated by the e xample of lithium stearate. Surface active materials have remained essentially untouched in terms of HSP . The first is for the ydrophobic end.3). or by simple comparison with similar (usually) h ydrocarbon molecules of the same size. An aliphatic end group w ould have lower affinity for polystyrene. Those familiar with cohesion energy parameters can already discern dif ferences that may impro ve the chances of success. HSP for surfactants can be assigned by e xperiment or by the methods described in the ne xt paragraph. organic and inor ganic salts. Each end of such molecules will require its o wn set of HSP . The uptake of potentially dangerous chemicals in recyclable packaging also needs attention.3 confirm the need for the first t o characterizations. The data in Chapter 11 confirm that di ferences in HSP are critical if soils are to be removed effectively.23 SURFACE ACTIVE AGENTS Surface active agents have not been systematically characterized by HSP yet. An additional area of interest may be found in that many commonly used reaction solvents have similar HSP. although Beerbower started on this man y years ago. for xample. Although these e xamples are ob vious to those skilled in the science of surf aces. Barton has collected group contrib utions that can also be used in connection with the . whereas toluene will. This is also still in its inf ancy. Still another area of major interest is the systematic formulation of filled systems using HS . Even in a completely dissolved condition. Octane will not dissolv e polystyrene. for e xample. Guides for action can also be found by simple comparison of th HSP of those solv ents which sho w a dif ference in beha vior. This was discussed in Chapter 6. These include dimethyl sulfoxide. It could be that the solubility of an acti vated species ha ving high polarity ( δP) and moderate hydrogen bonding (δH) is determining the reaction rate(s). the second is for the hydrophilic end. S veral of these applications are discussed in more detail below. organometallic materials. than an aromatic one. with group contributions. A more systematic approach for the use of cohesion parameters to describe surf ace phenomena w ould be desirable. dimethyl formamide. gases. and the third is for the molecule as a whole. 8. w ater.

The kinetics of the situation may also be important. and the HSP of the hydrophilic end of the surf actant. Some types are also vulcanized to impro ve properties. and it can be expected that the orientations may be changed with the transport or presence of other materials such as w ater through a given local environment. the static contact angles with w ater disappeared and a coherent w ater film as obtained.25 After several hours of exposure of a fresh coating to w ater. It is thought that this is best done by a veraging using estimated molecular volumes for the respective ends relative to the molecular volume of the whole molecule. and polyurethanes combined with polyethers or polyesters (TPU). These predictions and inferences appear to agree with what is e xpected or has been established by independent measurement.The Future 333 different ends of surf ace acti ve agents. Experiments are preferred. In closing this section. function much like surfactants with the backbone and those side chains containing hydroxyl groups favoring placement to ward cellulose (or their o wn kind). b ut it is too early to say that confirmation has been obtained indepen dently. If the HSP do not match sufficiently well. there will be some similarity between the HSP of the protein or urea. or whate ver resembles this end best. There will also be some form of interf acial or boundary layer resistance influencing the adsorption and absorption of sur actants into soils. it is interfacial mobility that is in focus. The peat moss has become hydrophilic (but returns to the hydrophobic state on drying again). if this also denatures the protein (see Chapter 15). Hemicellulose side chains containing acetyl. as shown in Chapter 11. for e xample. A drop of water initially pearls on the surf ace but shortly thereafter disappears into the interior in a spontaneous manner . The association of blocks of polymers is also the basis for the thermoplastic elastomers (TPE). polyeth ylene oxide or polyeth ylene glycol. When a surfactant denatures a protein. organic salts (see below). The treatment of dif ferent segments of block copolymers as separate entities is a related endeavor where more quantitative predictions of compatibility should be possible. SURFACE MOBILITY (SELF-ASSEMBLY) The rule of thumb that “lik e seeks lik e” can be v ery useful in understanding the structure of complicated systems. for example. The estimation for the molecule as a whole in volves combining the two sets of HSP for the ends. and the estimation for gi ven inorganic salts is still uncertain. If enough data of this kind can be generated. . it can also be repeated that thermodynamic surf ace and interf acial phenomena correlate with HSP . Typical e xamples include the styrene/butadiene/styrene block copolymer (SBS). 2 The HSP of the h ydrophilic end can be estimated by comparison with existing HSP of. Chapter 16 discusses adsorption and absorption in polymer surfaces where there is a surf ace resistance. adsorption and absorption will presumably no occur as readily as when the HSP do match to within some required limit for the desired ef fect. Peat moss is an example. The rotation of some hydrophobic materials to become more hydrophilic when in contact with water is still another example of like seeking like. acid. sucrose or other w ater soluble entity. as demonstrated by the e xamples in Chapter 16. the polyether/polyamide block copolymer (PEBA). inorganic salts (see belo w). The size and shape of the adsorbing/absorbing entity is also presumed to be important from a kinetic point of vie w. It is kno wn that additions of a block of polymer C to both polymer A and polymer B impro ves their chances of compatibility (at some molecular level). of course. In this example. This has been amply sho wn in Chapter 6 and Chapter 7. the HSP of the surfactant can be e xperimentally confirmed. This phenomena was actually employed to develop an electrodeposition coating for an evaporator where film-wetting by ater was required for good e vaporation efficien y. That this type of consideration can lead to useful results can be seen in the way that the beha vior of w ood polymers and the ultrastructure of cell w alls in w ood was treated 24 Hemicelluloses appear to in Chapter 13 and in much more detail by Hansen and Björkman. or ether groups are expected to favor orientation toward lignin regions. These are made with a wide v ariety of dif ferent immiscible (hard and soft) blocks where the phase separation is critical to performance.

the advancing contact angle also started to fall. It is not yet advisable to include w ater in a standard set of test liquids for x eperimental evaluation of the HSP for polymers or other materials because of its tendenc y to be an outlier . it fell with e xposure time/aging at shorter times. The concept of hydrophilic (hyperphilic?) bonding. This means a challenge still exists to understand ho w to be able to incorporate w ater into a standard set of HSP test liquids without al ways being concerned about special interpretations for w ater. when the surface was oxygenated to a greater extent. The urea additions correct for this deficien y. These phenomena also point to the use of urea or urea groups to impro ve biocompatibility. Basic Violet 10). There may also be other possibilities. Hydrophilic bonding (usually referred to in the present context as intermolecular h ydrogen bonding) is responsible for the configurations of proteins i water. A characterization problem caused by nonideal mixtures with w ater is the interpretation of HSP correlations for materials such as the dye Rhodamin FB (C. are no longer “h ydrogen bonded” in the con ventional usage of the term. b ut can specific beh vior be predicted? More research is needed in this area. The HSP correlations for the solubility of solvents in water presented in Chapter 1. The orientation of molecules to allo w given segments to locate in re gions of similar HSP is presumed to be a general phenomenon. water.I.3 have not been tested extensively as yet. does not ha ve the correct set of HSP to truly dissolv e these se gments of the protein molecules. The use of the HSP for w ater found from the correlation of total w ater solubility appears to be the most promising set of v alues to work with at the present time. Systematic studies of these effects are badly needed. This is especially true for w ater in lower energy systems. As noted earlier on se veral occasions. At longer exposure times. which is the opposite of hydrophobic bonding. as the y are no w truly dissolved by an ef fectively good solv ent that can also dissolv e these se gments/bonding sites. 2. its v ery high δH parameter for a liquid. but whether or not these HSP data for w ater can be used in a lar ger context remains to be determined. Many of the concepts discussed here are directly applicable for self-assembling systems and to procedures and products within the concepts of nanotechnology . whereas the (static) adv ancing contact remained constant. and only for water. The HSP data leading to formulation of selfstratifying coatings also pro vides useful information 26 (see also Chapter 6). in the meantime.334 Hansen Solubility Parameters: A User’s Handbook Many surface phenomena can be understood from the preferences of gi ven segments or materials to seek out re gions of similar HSP . is discussed in more detail with e xamples in Chapter 15. They are clearly useful to mak e predictions for the solubility of untested solv ents in w ater. Aging can be expected to lead to increased oxygenation and perhaps also to a decrease in a verage molecular weight. Table 1. These ef fects both contrib ute to the tendenc y/ability for oxygenated species attached to an otherwise more h ydrophobic polymer to rotate into an applied w ater droplet. as described in Chapter 15. b ut. The proteins that can be dissolved in mixtures of water and urea or given salts. for example. Sometimes it acts like a single molecule. Surface mobility also has an important role in biological processes. and its tendenc y to selfassociate or to associate with other materials forming special structures. and sometimes it acts as a cluster of about six molecules (according to the HSP comparison. The usual solvent. Some inferences may be possible from the studies performed on compatible (or nearly compatible) polymers.16 Use of mixtures of solv ent and w ater as test solv ents led to a v ery nonspherical (noncircular) cohesion . When the (static) receding contact angle for w ater w as measured. General beha vior can be predicted. w ater can be considered as having (at least) duality. at least). b ut do seem promising. An example of how this can lead to oddities is discussed in the follo wing. WATER The current treatment of the HSP for w ater discussed in Chapter 1 and Chapter 15 needs confi mation and/or modification. and the protein is said to be denatured in the process. w ater is v ery special because of its lo w molecular v olume. One such study 27 confirmed that the rotation ability (mobility) o aging polymer surfaces could be followed by measuring the (static) receding contact angle of w ater.

7 106.-S. but the characterization of the dye in question is not useful in relation to prediction of interactions with other materials.7 16.8 20. Boca Raton. FIT = 0.7 17.6 8.294 0.295 1.487 0.5 δP 10.4 26.589 1.4 3.1 7.0 15.8 15. This is not true.0 15.2 17. FL. δP = 17.6 73. 82(4).1.0 8. It should be remembered that man y .934 0.7 12.3 40.1 156.5.6 16.902 1.2 18.4 0.8 6.8 9.2 94.7 1. 5 The HSP data for this correlation are gi ven in Table 18.0 15.0 5.707 1.4 16.930.3 11. Ed.5 108.4 0.7 12.5 Calculated Solubility Sphere for Rhodamin FB Solvent Acetone Benzene 1-Butanol Butyl acetate gamma-Butyrolactone Cyclohexane Diacetone alcohol Diethylene glycol Diethylene glycol monomethyl ether Diethylenetriamine Dimethyl formamide Dimethyl sulfoxide Dipropylene glycol Ethanol Ethylene glycol Ethylene glycol monoethyl ether 2-Ethyl hexanol Glycerol Methanol Methylisoamyl ketone Methyl-2-pyrrolidone 1-Pentanol 1-Propanol Propylene glycol Tetrahydrofuran Toluene Trichloroethylene Water δD 15.4 91. but for the present it seems to be the only procedure possible to maintain consistenc y in the HSP procedures de veloped.5 76. energy parameter plot (see Figure 18.4 23.4 11. G. 10. NO = 28.9 17.988 2.2 10.0 16. Ro = 12.8 13.3 Solubility 0 0 0* 0 0* 0 1* 1 1 1 1 1 1 1 1 1 0 1 1 0 1* 0 0* 1 0 0 0 0 RED 1.3 5.530 1.615 1.3 12. This procedure may be subject to revision at some future point in time.6 81.3 10.2.2 16.7 16.7.741 0.7 19. Handbook of Surface and Colloid Chemistry. The HSP correlations for w ater-soluble polymers and other high ener gy materials involved similar extrapolations into domains where there are no liquids.5.6 14.2 3. The data fit as 0.6 75.0 16.3 130.0 108.0 2. 281–287. Figure 18.9 (with the constant 4) w as used in this correlation.6 0.9 17. 1997. The irregular plot can presumably still be used as such. with a significan portion in the high ener gy re gion where there are no liquids. CRC Press.3 42. as it has been used in all the other HSP correlations in this book. 1976. Source: Riedel.2 73.4 5. The plot has given several individuals the impression that there are significan problems with the HSP approach when it is applied to this kind of material.3 8. Chapter 1.4 10.3 4.0 16.1).8 29.517 0. Birdi.3 4.0 18.732 0.5.7 124.8 16.0 δH 7.4 16.5 132.570 0.0 Note: δD = 16. Farbe and Lack.0 18.772 1.8 108.9 16.4 18.8 11.3 22.996 0.0 16.93 for 28 data points.815 0.8 17.7 3..2 13.138 0.7 142. δH = 18.486 0.076 1. chap.9 118.9 58.991 0.5 6. K.3 7.8 16.1 16.001 0.0 89.8 18.0 9.8 96.5 15.2 12..4 15.889 0.999 1.5 55.8 79.5 18.3 13.0 5.1 16.The Future 335 TABLE 18.677 0.1 clearly shows that this HSP sphere co vers more space than the data.0 2. A computer analysis based on the pure solvent data given by Riedel16 confirms that a good “spherical characterization of Rhodamin FB is possible using the same data otherwise used in Figure 18.8 19.2 15.. Equation 1.0 71.125 1.965 V 74.1 12.0 15.0 7.042 1.0 77.8 90.

9 should be increased. The HSP considerations discussed in this book pro vide help toward reaching this goal. HSP deri ved from these numbers seem to be in good qualitati ve agreement with e xpectations (even at 25°C). 2 Solubility parameter correlations for oxygen 28 (Chapter 13) and nitrogen 29 (Chapter 13) have been used as e xamples in this book.586) 5 ● N2 (. Figure 18.) (most) liquids with high HSP (w ater. Table 13. Figure 1. This same procedure is applicable to other g ases. The δP and δH parameters for these tw o gases are zero.639) ΔET KJ ∕ MOLE 15 Xe (.577) ✳ ● CH4 (.7 2H6 (.583) Ar (. Carbon dioxide is e xtensively discussed in Chapter 10 where the HSP are calculated for lar ge v ariations in pressure and temperature. A specifi example where this is not the case is carbon dioxide.569) ✳ N O E BL G ES AS ●C TR =0.624) ISOBUTANE (. This figure i worthy of some consideration from a theoretical point of vie w. The basis of the HSP is a corresponding states calculation for E D as the ener gy of v aporization of a corresponding h ydrocarbon solvent (same V and structure) at the same reduced temperature. it was found necessary to e xtrapolate the data in Chapter 1.6 ● C3H8 ● (.15 K di vided . of course.5 TR =0.2 Cohesion energy for various low molecular weight materials as a function of molecular v olume and reduced temperature (gi ven by curv es or in parentheses). Some examples are given by Barton. This is discussed more in Chapter 2.642) ● 20 TR =0.803) ✳ 30 40 50 60 MOLAR VOLUME CC/MOLE 70 80 90 100 FIGURE 18. Chapter 3 also treats a method to calculate the three partial solubility parameters for g ases.336 Hansen Solubility Parameters: A User’s Handbook 25 ENERGY OF VAPORIZATION FOR “NON POLAR” LIQUIDS BUTANE C4H10 (. and in man y cases quantitative agreement with physical behavior has also been found.578) ✳ O2 (. methanol. A unifying concept and procedure for the use of w ater in all testing is needed. HSP for man y gases where this is not true are reported in Chapter 13. Equation 1.1 to lo wer molar v olumes. (See te xt for discussion. This f act might gi ve the impression that the constant 4 in Chapter 1.6025) 10 Kr (. glycols) also ha ve lo w molecular v olumes (V). Solubility parameters are usually deri ved from data at the normal boiling points.This is.2 is deri ved from this.607) ✳ ● 0 0 10 20 H2 (. GASES HSP can also be used to impro ve understanding of the solubility beha vior of g ases.4.613) He (.613) Ne (. 298. In the process of calculating the HSP for g ases. This makes them “better” solvents than expected by comparison with all the other solvents (whose average V is closer to 100 cc/mol).

however. 30 The result w as that the or ganic salts al ways had considerably higher HSP than either of the components making them up. deserv e more attention (see Chapter 15 and Reference 24). such as in hemicelluloses.2. and Ro equal to 19.5.δP.5. why is oxygen among the noble g ases in cohesion behavior rather than rated with the other g ases? At the time of choice for a reference for the dispersion bonding ener gies.13. There was about 10% shrinkage in v olume compared with the original v olumes of the acids and bases. These values are high enough to mak e the salt entities insoluble in most polymers. The salting in and salting out of v arious polymers can perhaps pro vide clues to assign HSP in this connection. Questions can be raised as to wh y the noble g ases dif fer from the hydrocarbon solv ents and whether the h ydrocarbon solv ents were the best choice as reference materials. Some additional considerations may be found in further study of the relation of HSP to the corresponding states theory of Prigogine and co workers as discussed in Chapter 2. respecti vely.0.5. The δD parameter seems to be qualitatively capable of describing the beha vior of metals to some e xtent.δH.8. ORGANIC SALTS The HSP of se veral or ganic salts ha ve been compared with the HSP for the or ganic acids and organic bases from which the y were made.3.22.16. the cohesion ener gy of the salts is high enough to mak e them solids rather than liquids. lower δD for the nitrate portion of a salt than would be e xpected from the group contrib utions from a chloride.2.5 and 16. and the necessary data for a re vised reference are lacking.19. It may be possible to arrive at an approximate description of inor ganic salt solubility in or ganic media (perhaps w ater. The beha vior of the h ydrocarbon solvents appears to be included within the Prigogine parameter for dif ferences in size ( ρ). Finally . In some cases. and 13.21.δP.1. This study sho wed that or ganic salts can indeed be characterized by HSP .δH equal to 17. all in MP a1/2. . b ut the corrections w ould be minor. Also. there were ample data on latent heats for the y hdrocarbons and the aliphatic h ydrocarbons were considered as having δP and δH values equal to zero. with other types of salts.19.9. This general relationship was also found for other salts formed by combinations of or ganic acids with a v ariety of amines.δP. as there are e xperimental data to justify the numbers. Group contributions to the HSP from the nitrate group are e xpected to result in lo wer HSP. This would lead to greater solubility of the nitrates in or ganic solvents. Nitrates are kno wn to be among the most soluble of salts. and. in particular .9. and 16.δH are 14. whereas the δD.11. The HSP for the salts are generally close to those mentioned earlier . This may not quite be true.1 in dark type.22. which is indeed the case. the acid groups found in nature. Somewhat less soluble than the nitrates are chlorides. It appears that the currently chartered course of using h ydrocarbon solv ents as a basis will be maintained. 14.9. Their affinities for ater will be v ery high. In particular . More w ork is necessary.5 for acetic acid.8.1. The HSP derived from this are δD. These values are reported in Appendix Table A. both because of high HSP and also because of the char ges associated with the salt groups. the salts made from formic acid and acetic acid combined indi vidually with dimeth yl ethanolamine had δD. it should be noted that an e xcellent HSP correlation of the chemical resistance of an inor ganic zinc silicate coating is reported in Chapter 12. These are only partly soluble in a few organic liquids with very high HSP.15.2. As examples. INORGANIC SALTS The solubility of magnesium nitrate [Mg(NO 3)2·6H2O] was evaluated in a standard set of solv ents1 and later correlated more precisely with HSP.8.21. respecti vely.3 for dimeth yl ethanolamine. All of these values are in MPa1/2.1. The reduced temperature at the boiling point is indicated in parentheses in Figure 18.8.The Future 337 by the solv ent’s critical temperature.2. Table 12. too) using HSP or some modification/ xtension thereof. ho wever.6 for formic acid.

cosmetics. Surface contacts are clearly important. 33 Unfortunately. There is also a question.LITHIUM STEARATE ? ? 3 2 1 ? 1 ? 1 2 3 4 5 δP 6 7 8 9 FIGURE 18. A suspicion is that.7 HANSEN PLOT . chewing gum. which oscillate more slo wly when the y g ain weight.7 8.8 δH δP 1 3. at least in practice. for example.3 where Beerbower31 used data from Panzer32 to show that lithium stearate does indeed have two distinct regions of solvent uptake and that a HSP plot can show why. of whether metal atoms in the center of more complicated molecules are ef fectively shielded from an y (surface) contact with a solv ent.1 1 3. as least not to the author’ s knowledge. It is clear that HSP e xist for these materials.) ORGANOMETALLIC COMPOUNDS No systematic studies of the HSP of or ganometallic compounds have been made.3 HSP plot for solvent uptake by lithium stearate. we do not often deal with pure metals in this conte xt. b ut it appears that the nature of the central atom also has an ef fect. AROMAS AND FRAGRANCES Aromas and fragrances are important in connection with packaging materials. foodstuf fs. but rather metal oxides. One of the examples included in Reference 34 was the development of an artificial nose based on coate oscillating sensors. This example shows that one can calculate HSP values where the relevant data can be found in the literature and then test these with rele vant e xperiments. A recent report 34 discussed HSP in connection with fragrances and aromas. etc. the cohesion ener gy derived from the “metallic” bonding in organometallic compounds can be coupled with the dispersion parameter . An exception is perhaps that shown in Figure 18. Other . (From Alan Beerbower. it might be noted that Hildebrand and Scott presented a chapter on the solubility parameters of metals. Finally . Metallic bonds differ in nature from those usually discussed in connection with or ganic compounds. b ut very little work has been published in the area. Matching HSP for the coating and material to be detected leads to increased weight g ain and increased sensitivity. Group contrib utions w ould be v aluable.7 2. for which no HSP work has been reported. personal communication.32 Units for δP and δH are (cal/cm 3)1/2.338 Hansen Solubility Parameters: A User’s Handbook 11 10 ? 9 8 7 6 δH 5 4 ?? ? ? ? 2 cc ABSORBED ? > . 31.

See Chapter 16. b ut the sensitive areas are covered anyway. This was evidenced by x-ray diffraction and DSC measurement that confirmed the drugs were embedded completely amorphou (solid solution). An additional activity. pass through open passageways. coatings. Table 18. A closer match would presumably be required to do this for higher concentrations of the drugs. A relatively slow diffusion process is required to do this. De gussa Pharma Polymers. Such data will allo w greater use of the correlations since guidelines for potential improvements can be obtained. Darmstadt) that calculations of HSP for a number of drugs and for a series of EUDRA GIT® polymers provided a correlation confirming that the drugs are soluble in those polymers where the HS match closely enough. When the HSP relations between a drug and its surroundings are known. When there is a good match in the HSP v alues. Once a chemical has absorbed into a polymer . although h ydrochlorides and other salts . is to assign effective HSP to various test materials such as mustard. A poor match in HSP may leave holes and expose the drug for more rapid release. It is suggested that an e xtensive HSP analysis be done for those polymers where potential misuse or contamination of containers prior to rec ycling is a possibility . When the match is poor. it can be v ery difficult to get it out a ain.6 contains HSP for some materials whose solubility in a vrious media may have interest. and particularly if it is a rigid polymer with a relatively high glass transition temperature. CONTROLLED RELEASE HSP considerations can provide an extra formulating parameter for the controlled release of drugs.These shortcomings mean that HSP correlations of chemical resistance must be done with great care. Andreas Gryczk e. and uptak e of solv ents in many polymers. It is thought that the masking effect is the result of adsorption of the fragrance on the regions of the sensory system where HSP are similar to those of the undesired odor .. The recycling of polymeric containers has a potential problem in that the polymers used are able to readily absorb those chemicals whose HSP are not too dif ferent from their o wn. and other gi ven products that often appear in tests of chemical resistance. and this barrier pre vents the odor from arri ving at the given sites. k etchup. of course. which should be done for practical reasons. and this may slo w release more than desired. permeation phenomena. ABSORPTION OF CHEMICALS IN PLASTICS HSP correlations e xist for chemical resistance. etc. CHEMICAL RESISTANCE Chemical resistance studies ha ve generally been performed with too fe w liquids and without the necessary spread of HSP to allow the data to be correlated with confidence. b ut it can. attainmen of equilibrium is not usually confirmed. The key to interpretation is. the drug will be more soluble with the ability to mo ve at some rate within a polymer matrix. as usual.The Future 339 examples where HSP could be systematically used include: counteracting undesirable odors using fragrances that ha ve reasonably similar HSP. Drugs will also tend to adsorb at surf aces where there is a good HSP match. There may not be a complete match re garding steric adsorption. the drug will not be able to permeate through its polymeric surroundings. what appeared to be true solubility w as found. development of packaging with designed HSP to either function as a barrier or as a sink. predictions of its behavior can be made. This can point out which chemicals are most lik ely to present the greatest problems. This has been discussed in Chapter 12 in more detail.). On the other hand it can be surrounded by a matrix with similar HSP . In addition. When the drug and polymer HSP were within 10 MPa1/2 at the concentration studied (20% wgt. absorption of odors into plastics. It has been sho wn (personal communication. and an estimate of where a gi ven aromatic material is lik ely to reside. sealants. that similarity of HSP means higher affinit . It is possible to calculate HSP for essentially all drugs.

0 21.0 17.6 13. Pharmaceutical and Medical Packaging 2001.10.8 3. Derek L.3 20.0 5.4 18.2 10.0 V 154.5° in toluene.0 13. The following link reports what follo ws for studies of or ganically modified nanoclays (o ganoclay) in nanocomposites: http://www.8 14.M.7 δP 8.7 206.1–20.9 8.1 6.9 12.0 18. He xagon Holding.edu/ACA/ACA03/abstracts/text/W0383. More details and additional e xamples are found in Chapter 15.0 268.0 24..0 16. Skov..html Quote: W0383 Evaluating Or ganoclays for Nanocomposites by Small Angle Scattering .1 11.2 310.1 106.7 16.1 11.4 12. have high affinity for skin NANOTECHNOLOGY Controlling the orientation of molecules can be a k ey for switches and the lik e. but changed to a new orientation at temperatures above 38. C.6 13.6 4. H.3 141.0 19.6 should be measured rather than calculated only . The configuration at the higher temperatures as also that adopted in tetrah ydrofurane. There are correlations of HSP with skin permeation presented in Chapter 15 that should allo w an estimate of which drugs can enter via this route.1 9.340 Hansen Solubility Parameters: A User’s Handbook TABLE 18.0 8. as there is v ery little e xperimental data on which to base an estimate with group contrib utions.6 Calculated HSP for Various Materials Material Adrenaline 4-Aminopyridine Ascorbic Acid Caffeine Cycloheximide DDT Dopamine Ecstacy Meclofenoxate (base only) Norephedrin 2-Oxopyrrolidinacetamid Quinine Saccharin Serotonin Spermidin Note: Units are MP a1/2.8 144.6 10. National Institute of Standards and Technology. 20. .2 and Charles J. Simple calculations confirm that dopamine nicotine.1 19.5 10.9 198.7 15. for e xample. Copenhagen.2 18. pp. Ho 1.2 δH 19. R. This change of orientation is the result of a poorer solv ent at the higher temperatures compared with a good one at room temperature.9 171.5 18.1 12. An example has been given in Chapter 15 where anthracene units appended to a polymer molecule adopted one orientation in toluene at room temperature.5 87.9 27. 2001.hwi. Conference Proceedings.5 6. 1Center for Neutron Research. Glinka 1.5 19. and nitroglycerine.8 180. skatole.2 11.7 157.0 18. δD 20.0 16.5 20.0 18.2 11.4 155. Source: Hansen.9 116.3 5. Ed.0 202.buffalo..

The author’s current view as expressed in Chapter 2 is that the Flory approach is a special case of the more general Prigogine corresponding states theory . With Hansen’s solubility parameters. nanocomposites composed of clays and polymers ha ve been found to ha ve impro ved mechanical properties as well as enhanced thermal stability . Chapter 4 deals with theories of vapor–liquid equilibrium in particular and how these relate to HSP. 37–40 A major problem is the reliability of the chi parameter (and also HSP) v alues in the literature (see Chapter 2 for more details). Recently. it must include both the atomic/dispersion interactions as well as the specific interactions reflected by th δP and δH parameters. δo2 = δd2 + δp2 + δh2. leads to confidence in the HSP approach The geometric mean must be used in the Prigogine theory to arri ve at this similarity of treatment. and the apparent agreement in the treatment of specifi interactions by both the HSP and Prigogine treatments. The or ganically modified clays studied in this ork precipitated in an y solv ent with molecules with moderately strong h ydrogen-bonding groups. . trichloroeth ylene. reflected by δd. the correlation between the de gree of e xfoliation of or ganoclays and the solv ent in which the clay platelets are dispersed/mix ed has been analyzed. Zellers and coworkers have recently made an attempt to use this theory in conjunction with HSP studies. 2Dept. The improved properties are related to the degree of dispersal and e xfoliation of the clay platelets in the polymer matrix. THEORETICAL PROBLEMS AWAITING FUTURE RESOLUTION POLYMER SOLUBILITY The Flory chi parameter has been used to describe polymer–solvent interactions for many years. In order to understand and optimize potential processing conditions. College Park. These and the many other examples in this handbook related to self-assembly (see abo ve). More understanding is required before chi can be calculated with reasonable accuracy. molecular orientation. The statistical thermodynamic approach of P anayiotou and co workers thoroughly discussed in Chapter 3 provides convincing evidence that the division of the cohesion energy density into three parts (at least) is the correct procedure to understand af finities among ma y types of materials. Attempts to calculate chi using HSP are reported in Chapter 2. Univ. Furthermore. of Maryland. Finally.36 If this single parameter is to be complete in this function. of Materials and Nuclear Engineering. MD 20742. which completely exfoliates the or ganoclays studied in this w ork. but intensified e forts seem warranted. and af finity among molecules and molecular s gments should provide ample evidence that molecular guidance in nanotechnology endea vors can be found in the HSP concept. This is in agreement with the vie w of Patterson discussed in Chapter 2.The Future 341 Gaithersburg. Understanding the interaction between or ganically modified clay (o ganoclay) platelets and or ganic solvent molecules as well as the corresponding structure of or ganoclays in a suspension is a critical step toward tailoring and characterizing nanocomposites formed by or ganoclays in a polymer matrix. surface adsorption. is the principal factor determining whether the clay platelets remain suspended in the solv ent while the polar ( δp) and h ydrogen-bonding ( δh) forces af fect primarily the tactoid formation/structure of the suspended platelets. or ganoclays were dispersed in a number of or ganic solvents covering a range of solubility parameters and characterized using small-angle neutron scattering and wide-angle x-ray scattering techniques. the v ery general applications of the HSP approach demonstrated in this book and else where.35. It has been found that the dispersion force of the solvent. Chapter 4 is a length y discussion of the use of HSP in thermodynamic models for polymer solutions. The correlation found has been used to correctly identify a solv ent. MD 20899.

42 HSP characterizations of surf actants are also badly needed. Specific areas needing more theoretical ork related to HSP in the near future include better understanding of usage for predictions of beha vior for w ater. friction of polyethylene untreated and treated with sulfuric acid. Dev. the systematic use of HSP in nanotechnology to control the assembly and orientation of molecules or se gments of molecules. A good starting point is the Handbook of Surface and Colloid Chemistry. REFERENCES 1. etc. 8(1). It is the surfaces of molecules which interact with each other (also in b ulk and solution phenomena). edited by K. No other parameter can be assigned to such a range of materials spanning from g ases and liquids. Res. A formalized unifying theory linking HSP to both b ulk and surf ace phenomena is still lacking. gases. char acterizing the surfaces of polymers. o ver surf aces. . The universality of the solubility parameter .M. Much more research needs to be done with these relations. Handbook of Polymer-Liquid Interaction Parameters and Solubility Parameters. C. to inor ganic salts.. Chem. 3. 1991. inor ganic salts. mention should also be made of the e xtension of the three-parameter HSP approach to a fi e-parameter approach by Karger and coworkers.342 Hansen Solubility Parameters: A User’s Handbook SURFACE PHENOMENA Surfaces can clearly be characterized by HSP as demonstrated in Chapter 6 and Chapter 7. chara terizing organic salts and inor ganic salts. Eng.S. A. CRC Press. many of which apparently have HSP very much higher than any of the liquids. and the Jof fé ef fect — ef fect of liquid immersion on the fracture strength of soda-lime glass. 2.M. 1969.M.e. Boca Raton. Boca Raton. Most materials having HSP in the range of the customary test liquids can be studied using HSP with reasonable success. The extension of this. Extensions of practical applications related to chemical resistance and the uptake of potentially dangerous materials in polymers are also required. Handbook of Solubility Parameters and Other Cohesion Parameters... 1990. This theory should quantify the adage “like seeks like. and salts. CONCLUSION HSP have been sho wn useful in solv ent selection.F. This is not fully the case for g ases. FL. many of which have much lower HSP than the well-studied liquids. Hansen. so it is not surprising that cohesion parameters can be applied with success to both solubility and surface phenomena. Ind. CRC Press.. Barton. Presently. include surface phenomena as well as b ulk phenomena.. and more recently to the w ork of P anayiotou summarized in Chapter 3. predicting polymer–polymer miscibility. the Rehbinder ef fect — the crushing of aluminum oxide (Al 2O3) under v arious liquids. fillers. Prod. A. Barton. correlating permeation phenomena. FL. and fibers. 2–11.” also seems to ha ve been demonstrated. These results and the close relation with the Prigogine corresponding states theory of polymer solutions. Finally there is a great deal to be done in the areas of controlled drug release. indicate that a still more general theory exists. Here ag ain. and organometallic compounds. organic salts. the successful use of HSP to such applications might not have been anticipated had it been considered as a parameter for use in bulk systems only. impro ved understanding of some biological processes. and last b ut not least. 1983. 2nd ed. 41 If we consider chromatographic techniques as depending primarily on surf ace phenomena.F. Birdi. g as solubility.” i. “lik e seeks lik e. the best that can be said is that the generality “lik e dissolves like” can be quantified i many cases.. The work of Beerbo wer contained in Reference 17 and Reference 23 has also sho wn applications of HSP to such varied phenomena as the work of adhesion of liquids on mercury. Water remains special because of its lo w molecular v olume and high δH.

20. 1997. WV. 11. Assoc. Vol. July 1996. pp. 1163–1179. 1967. Ingeniørforening in Danmark. and Schuch.. in Kirk-Othmer Encyclopedia of Chemical Technology. 73(8). 26. Appl. König. C. 18. Solubility parameters.. in Interdisciplinary Approach to Liquid Lubricant Technology. J. Solubility parameter values.. J. Properties of Polymers: Their Estimation and Correlation with Chemical Structure. E. C. 301–308.. 660–664. Boundary Lubrication — Scientific and Technical Applications Forecas