Application Handbook ad Thermal Analysis in Practice Collected Applications | METTLER , TOLEDOCollected Applications Thermal Analysis Thermal Analysis in Practice Dr. Matthias Wagner The information contained in this handbeok is based on the eurent Knowle al experience of the authors. The hanchonk present a large numberof caellly selected application examples, The experimen were cule and evaluated wit the utmost eae using the instruments specified i the description of each application ‘This es nat however absolve you irom persnaly testing the suitability ofthe examples tor your own wock and purposes. Since the transfer and use ofan application is beyond our control, we cee of eourse acepl any rjonsity ‘When chemicals, solvents and gases are used, the general safety rules and the instructions given by the manufacturer ‘or supplier must be observed. {© PE All ames of commercial products car be registred trademaris, even i vot as wc METTLER TOLEDO Collected Applications Thermal Analysis in Practice Page 1Contents PREFACE ..... CONTENTS... 1 INTRODUCTION TO THERMAL ANALYSIS. 1.1 Dentwirions 112 ABlugrExpLaarion oF Iuron4er THER ANTS THOINTQUES 1.3. AerLicstion OvERview 14 Tae TeyPesaTuRe Proce [REFARENGES AND FURTHER READING 2 A BRIEF HISTORY OF THERMAL ANALYSIS ... 2.1 "THERMAL ANALISIS AT METTLER TOLEDO. REERENCHS AND FURTIIER READING 3 POLYMERS... 3.1 INTRODUCTION... 3.2. SywmiesisoF PocrweRs 3.3 TueRorlastics 3.4 THeRMosers. 3.5 ELasroMees ee 3.6 Potywte Appirives a 24 37 Useor Tuer AVYSS ro Cratcrenz Pours - i . m4 REFERENCES AND FURTHER READING 5 4 BASIC MEASUREMENT TECHNOLOG) 4.1 DEEN csc 42 SENSITEUTY 43° Noise 440 DETECTION LIMIT ne ue - scans AT 45° Dawe eG i q 46 Tove Const, LNG PRecuEN B 4.7 Diora. Resonwon av SameniNe TveRvaL » 4.8 CALIBRATION AND ADJUSTMENT OF SENSORS . 2 4.9 Tue Most IMpowtt ELECTRICAL TEMPERNTIRE SENSORS... 31 4.10 ‘TrareRsroRes 1 THERMAL ANALASI 5 GENERAL THERMAL ANALYSIS EVALUATIONS: 5.1 THe Oprim Coono iste SisTEN 5.2. Bormine Discs, 5.3 _DIeLANc INFOMATION 0H THE DATASE 54. Oprivtcase 1: PRESENTATION OF 4 Disciast 5.5 NORMALIZING MEASUREMENT CURVES TO SAMPLE MIS, 5.6 DispLavte Cueets wr RESPACT 70 TIME, REFERENCE TEMPERATURE OR Sar TeMPERATERE | 5,7 Saweie TEMPERATURE AS 4 FUNCTION OF TIME, jc 38 5.8 CURVE CoRRECTION USING a BASELINE Samer ene 38 5.9 Mameseattca EXAnustions sea . 395.10 Conve Courstiso.. 5.11 NuMeRIcs. Besic4TION.. 6 GENERAL MEASUREMENT METHODOLOGY. 6.1 USUAL CoORDINATE SESTENS OF DIAGRAMS... 6.2 THEATMOSPHERE IN THE MEASURING Cl... 6.3 CRUHBLES TV THERMAL ANALISIS.. 64 Ovexvten ov Texan Berio 6.5 CaLianarioy aN AMjUsT MEN REFERENCES AND) FURTHER READING ‘7 DIFFERENTIAL SCANNING CALORIMETRY 71 Ierkonieion... 72 Deseo DSC MescxemEsr PNP 73. Saupue PaxPARavion i 74 PERFORMING MENSUREMENTS ou 7.5. INTERPRERATIONOF DSC CURVES ws 7.6 DSCBYALUATIONS es 77 sowe Seri D8 wrssteENE 7.8 DSC ArpLicaTion OvERVEW... 79° Counesrion xo ats. 7.10. Aveeno Asses miz PeREORNUNCE OF DSC MESSING CHL SINGS ERENCES AND FLRTIER READING sees ‘8 DIFFERENTIAL THERMAL ANALYSIS 8.1 Twe DTAMzxcuReMenr Peinciet, 82 ‘TyPica, DTACuRvEs peat 83 THE CALCULATION OF Mie DSC CuRvE FROM! SDT. . $4 Tue SDTA Evacuations 19 THe STAR SOFTWARE... : es ‘REFERENCES aN FURTHER READING... 9 THERMOGRAVIMETRIC ANALYSIS..... 9.1 Iwreopucrion, 92. Drsiox ns Messe Pace. 93 Swe Pesraeamon. 94 PERFORMING MEASUREMENT 95 Ivmarnerine TGA Cuan, 9.6 TGAEVaLwxtions 97 TyPICsL APPLICXTION: RUBBER ANAS 98 Arrucarion OvenviEw.. 9:9 SToIciHIOMETRIC CONSIDERATIONS, 9.10 Cussasrion avn ADjestvent REFERENCES xD FURTHER REAING.. 10 THERMOMECHANICAL ANALYSIS. 101 Isreopucrn.. 102 Tw Desiev ap Messner PeNcPs OF ATMA 103. Savrte PrePaxamion. 104 TeurrsaTine Procaa . FEASIIREMERTS pnnenno nnn 13610.5 _yToRngetstion oF TMA Ceves 178 106 TMAEWUATIOS 183 10.7 Appuicarioy ovERVIEW FoR TMA, 191 10.8 CAUBRATION AND ADjUSTMENT OF ATMAMSDTA 192 REFERENCES 44D FURTHER READING 193 11 DYNAMIC MECHANICAL ANALYSIS .. 194 111 Iytropuerion aR 194 11.2 MeAsunrueNr Pervcrrie ap Desion 11.3. SAMPLE PREPARATION. : eet 114 Praroring Measttesers 115 _ INTaReRETATION oF DMA CuRVEs i 11,6 DMAEVs.varioNs 1L7 DMA APpLicaTion OvERMEW 11.8 Catineariow oF tue DMA/SDTASGT’ RREEPRENCES x0 FURTHER READING 12. THE GLASS TRANSITION. GLasses ano THE GLASS TRANSITION. 12.2 MEASUREMENT OF THE GLASS TRANSITION BY DSC, 12.3 DETERMINATION OF THE DSC Gass TRANSITION TEMPERATCRE, 12.4 Patent AGing ann Exrvatoy Resanation, 12.5 ‘The Guass TRaNsiTioN POR MATERIALS CLARACTERIZATION.. 12.6 Orman TERNAL AwRiSIS TECHNIQUES POR MEASURING THE GLa -REEERENCES AND FURTIER READING 13 BINARY PHASE DIAGRAMS AND PURITY DETERMLNATIO! 13.1 Ixtxopuction . 1.2 TWE Most MpoRrANT Bgy Pase DAGRANS... 13.3 Tae Use OF THE TiE-LINE TO PREDICT DSC CURVES enc 134 CONSTRICTING PHASE Dincleans mROM DSC MEASUREWENTS 13.3 DSC Pormry DeteRMiNarion ‘REFERENHS AND FLRTHER READING 14 POLYMORPHISM. 14.1 Tyrronccrton ann TeRws. 267 14.2. DeTecriov or PowywarPis.. 268 M3, TuEDSC bnsriaxioy oF PoLt\ORDUIS OF SULENMRIDNE so o 270 -RRPERENCHS AND RURTHER READING ....onnsnnnnnen 270 15. TEMPERATURE-MODULATED Dst 15.1. Ivrwonverion 15.2. Isoster”. 153 Amid 154 TOPEN™, [REPERENCES AND FURTIER READING wrnsnsninnnnnnnn om 16 EVOLVED GAS ANALYSIS. 16.1 Baise EvrRopucTION To Mass Srrcrnow16.2 Bales INTRODUCTION To FOLRLER TeaNstonu IureaReD SprctRoMEniy 284 16.3 CouPLinG THE TGA T0 A Gus ANALAZER, 284 164 EXAMPLES ons mi 285 REFERENCES AND FUITHER READING... 287 17 TGA SORPTION ANALYSIS 17.1. Baise INTRODUCTION To TGA SoRPTION ANSI. 288 17.2 Exawpurs. on 089 173 Cuseantoy.. eon DD A Arc APAAT i AAS, 293, -ReFeRENGES AND FLRTIER REATING 298 18 THERMOPTOMETRY ... nearneenes 294 18.1 _Iwronccriox 294 182 Theamowicresoory 183, ConuuinanesceNce y THERMAL ANALSI, 184 CONCLUSION... ‘ReveRENGES AND FURTHER READING.. 19. METHOD DEVELOPMENT .. 19.1 IntRonvcriox.. Srv 1: Choon RIGHT MessoREMENT Tec. Sree 2: SuaPLinc avo PREPARATION F THE Test SPECIME Srup 3: Choosine THe CRvcrBts (DSC AND TGA)... Sree 4: Cxoosine Te THMPERATURE PROGRAM SreP 5: ChoosiNG THE ArNSPHERE, : ‘rep 6; EXAMINING THE TEST SPECIMEN AFTER MOSORESENT SEP 7: EVALIATION... Sreo 8: VATION... AREFERENGHS av FURTHER READING, 20 OVERVIEW OF STANDARD METHODS FOR THERMAL ANALYSIS. ‘21 INDEX. ‘METTLER TOLEDO Collected Applications Thermal Analysis in Practice Page?——— 1 Introduction to Thermal Analysis LI Denrrioxs . 1.2 ABRiee EXPLANATION oF IMPORTANT THERMAL ANALYSIS TECANIQUES — - 9 1.3 APPLICATION OVERVIEW... ul 14 THe TeWPeRsTURE PRocRaNt re : on 2 [REPFRENCES AND FURTHER READING... 13 1.1 Definitions sn earlier definition proposed by the TCTAG, the International Confederation for Thermal Analysis and Calorimetry, vas “thermal analysis covers a group of techniques in which a property of the sample is monitored against tiaxe or temperature while the temperature ofthe sanaple is programuned. The sample is kept in a specified atmosphere. The temperature program may invole esting or coling at a fied rate of temperature change, or holding the deaperatare constant, or any sequence ofthese.” ‘Various objections were later raised and various: recommendations put forward to clarity certain points. For example: The distinction between a thermoanalstcal technique and a thermoanalstial procedure. Thermal analysis means the whole thermoanalytical method. It covers both the thermoanalytical technique (measurement of a change in sample propery) andthe theaoanalytica investigation procedure (evaluation and interpretation of the measured values) ‘© Analysis therefore means more than just monitoring In many cases, the change in the sample property is monitored and not the sample property tel {In most cases, the temperature of the enviconment is programmed rather than the temperature ofthe sample. Atmosphere is an operational parameter and is not essential fr the definition “This finally led to the most recent ICTAC definition of thermal analysis put forward in 2004. This defines thermal analysis simply as ‘Thermal Analysis sa group of techniques tha study the relationship between a stmple property and its emperatre.” “The definition clarifies key words used inthis definition as follows: Group ~ thermal analysis comprises a diverse range of techniques and experimental types that may be considered collectively if the total measurement conforms tothe total definition # Techniques ~ « technique is characterized bythe property that fs under investigation. ‘© Study — implies that time isan integral part ofthe thermal analysis experiment and the total experiment, and the interpretation and discussion of the measured data are included. «Relationship — implies that ether the sample property can be measured as a function of temperature (controlled- termperatore program), or the temperature can be meastired asa function ofthe sample's property (sample controlled heating) Sample — the material under study during the entire experiment (starting material, intermediates and final produs) and its close atmosphere. This is equivalent to the thermexlyanui stem. ‘¢ Property — any physical or chemical property of the sample. «Temperature — which can be diretly programmed bythe user or controlled by a property ofthe sample. The program may include an increase or decrease in temperatute a peciodic change, ora constant temperature or ay combination of these.“The data produced in a thermal analysis experiment is displayed as a thermoanalytical enree in a thermoanalytical “Sagram. Frequently several different measured signals are displayed at the same time (referred to as simultaneous measurement). “The thermoanalyst is usually interested in so-called thermal effects in which the messured signal changes more or Hess abruptly. “Often the objective isto measure physical quantities outside thermal effects, for example the specific heat capacity, the ‘expansion coefficient or the elastic modulus ‘Note: The term “thermogram” is dated and should not he used. It is nowadays reserved for the graphical representation ‘of te surface temperature distribution of objets. ‘The terms currently used are thermoanalytical curve or diagram, measurement curve, for example a DSC curve, a TMA diagram, etc. 1.2 A Brief Explanation of Important Thermal Analysis Techniques vo Lcomparson ot Toe estas Dsc Intogal 412.23.) lerumt 17080" RightLme 23665 °C LeblLimt 17050" Right Umit 23665°C Heating Rate 10.00 “Cin ‘aceline Type tne | ete 0.1282 mg Onset 21.71% step -96,2201 1 “6.3803 ng Residue 1.7537 % 0.4199 mg mmo 10 zoo 25000 go ago a0 «800 Fgup 1.1. Tha a kennquis used io mensute patyariGe & show dee enna! efeets, 052 Iain peak oe ctaline por TEA: yng and Gecomposron sip IA: scnenirg unt eas. DTA, Differential Thermal Analysis. In DTA the temperature difference herween the sample and an inert reference ‘substance is measured as a function of temperature, The DTA signal is °C or &. Previously, only the thermocouple voltage in my or wV was displayed. 'SDTA, Single DIA. his term was patented by METTLER TOLEDO and is a variation of classical DTA that is particularly ‘useful when used simultaneously with thermogravimetric analysis. The measnrement signal represents the temperature 20 Kin. V\/ Figue 14, Dynamic ating, lowed ty ccctng and second hating secon {Sot er erring recs carest Faw 1.5, Dynamic rsasuemts ot front hetng rots fo sve tine. Atore:A DSC inl chck Bal: MoxRes” used Wh TGA Tesolin amis goo he heaing refs aunty deceased poral toto nreasng reocon rte fe camo, Rs caan 0S area ee sis, te aig oe rcwases agi TH 2a gure 16, The sober! dyna seohamaleraportu program & mary usec fer the mare t ‘ne speci hc eapaty wih DSC and for Step” Th a www Eat Fou 1.7. Perodetargarau programs Top: A sees of tema saps, maint fr sty iesigatons ff emia ents or equine, and fr oS" ‘Mele: Te cow tom erogrom ius @vision of erating 09¢(ADSC). ‘oti: The srsoil madison (alo) is Me cre vatsion ou! ADSC technique. ADSC can separate cain ‘ets, Th shaoe shit at oces token the retng re and ho hoa! fois on odtona| ise of nfxnatin, The heating rate chosen applies to so-called reference position) because the real sample can exhibit first onder phase changes (e.g. melting) in which the heating rate cannot be controlled. This type of temperature control of the sample eavitonment is known as isoperibolie. In fact, the temperature of the sample advances compared to that of the reference during exothermic processes and lags bebind in endothermic processes. Depending on the thermal contact of the sample (thermal resistance), the sample temperature can deviate irom the relerence temperature by several tenths to several °C (K). *9 In the case of TGA/SDTA, DNA/SDTA und TMAVSDTA, this eoresponds to a fictive inert sample during the blank run, whereas in DSC it is a reference crucible during the measurement. References and Further Reading (1) ictac ¢ Nomenclature of Thermal Analysis, at http://w ietac org (METTLER TOLEDO Collected Applications Thermal Analysis In Practice Page 132. ABrief History of Thermal Analysis 2.1 ThbRMsl ANALYSIS aT METTLER TOLEDO... ReFeRENGES AD FURTHER READING 5 16 ‘Thermal methods were used long before anyone talked about chemistry ot material sciences, Even in antiguity, many substances were subjected to a “heat test” to verify their identity and authenticity. Nowadays, most of thee tests have been replacal by other physcal-chernical test methods In the 18" century, thermometers and temperature scales were developed, for example that of Anders Celsius in 1742. The thermal expansion of materials was used to indicate the temperature. Conversely, measurements of the change in length of materials asa function of temperature (dilatomety) had already been perfocmel at this time ‘The manufaeture of ceramic procicts i closely connected with the development of thermal analysis. For example, even today, so-called Seger cones are still used to indicate the temperature reached atthe position in the furnace where they are placed. A set of these consists of several triangular pyramids cones made of clay and oxide mixtures of increasing softening temperature. When they reach a particular temperature, they deform under the influence oftheir weight. In the 19” century, it became possible to measure heat quantities. This was after the difference between temperature and heat content or enthalpy ad been clarified using thermodynamic principles Tn 1887, Le Chateier [1] performed the first actual thermo: thermocouple in a sample of clay and heating the s photographic plate using a mirzor galvanometer Itical measurements as we understand them by putting @ mple in a furnace. The heating curve was recorded on a In 1899, Roberts Austen [2] significantly improved the sensitivity and meaningfulness of such measurements through the introduction of two differential thermocouples connected. in oppcsition. This allowed him to measure the leinperature difference between the sample and an inert reference. He is therefore regarded 2s the inventor of differential thermal analysis, DTA. Jn 1915, Honda [3] published the first thermogravimetric measurements in which the mass of the sample was almost continuously measured, Before this, it had only been possible to measute the mass difference after a thermal experiment by back-weighing, In 1955, Boersma [4] invented the present-day heat flow DSC with his idea of placing the thermal resistance outside the crucible, The development of power-compersated DSC was first described in a publication by Watson etal. [5] in 1964. Dynamic mechanical measurements with a constant selectable frequency have only recently become possible. Before this, similar measurements were performed using a torsion pendulum at resonance frequencies, Robert MacKenzie [6] studied the history of thermal analysis in great depth and published many articles on this subject In more recent years, thermal analysis has profited greatly from the availability of powerful computer hardware and software. This has had an enormous influence on the development of thermoanalytical methods. Before 1980, measurement curves were recorded with linear pen reconlers and evaluatsl manually. Selection of the wrong measurement range meart that the measurement had to be repeated using a more suitable range. Nowadays, the reasured curve is saved, then displayed in an optimal fashion for interpretation and if necessary automaticaly evaluated.2.1 Thermal Analysis at METTLER TOLEDO Echand Mettler, the founder of the compan ances, Jn 1960, he began looking for possible alditid ‘TAL: Hans-Georg Wiedemann came forwac with his ideas for the commercialization of his “thermobalance” just atthe right time, 4 development geoup was quickly set up to modify the purely mechanical semi-micro halanee using electromagnetic force compensation so that the balance signal could be recorded graphically on a pen recorder. At the same time, furnaces, temperature sensom and controllers were developed to enable temperature programs to he performed. Vacuum technology also had to be developed before in 1964 the first TAL “recording yacuum thermoanalyzer” was introduced. The TAL Thermal Analysis System could simultaneously perform TGA, DTG and DTA ‘measurements, Soon any reputable laboratory engaged in materials science research had to have aTAI even though the cost of such a system at that time was very high, about 120,000 to 200,000 Swiss Francs, depending on the particular version. The most important application areas of the TAL were inorganic compounds and ceramic materials had been very successful with his analytical and precision ba tu expand the range of products, FPI; At about the same time, an instrument for the automatic determination of the melting point of organic substances sas developed. An additional measuring cel for the dropping point of edible oils and lubricant greases and a hot stage for the observation of sarmples under the microscope completed the system, ‘TAZOOO: In. 1971, an instrument followed for quantitative differential thermal analysis, as the earlier versions of the present-day heat flow DSC were called at the time, Soon a temperature range of -170 to +500 °C was available, which ‘was ideal for the investigation of organic compounds and the increasingly important polymer plastic materials. At the same time, the first successful trials began with computer (PDP11) evaluation techniques following digitization of the analog measurement data. From 1973 onward, the first programmable desk top computers appeared on the macket and automatic evaluations became an economical proposition for normal customers TA3OOO: 1981 saw the introduction of the TA3DOO System with its new method concept. This was the first commercial instrument for the automatic measurement and evaluation of thermoanalytical data. Routine measurements enabled efficient quality control tasks to be performed. In addition to the DSG and TGA measuring cells, the TMA4O Thermomtechanical Analyzer was introduced, This was revolutionary atthe time because its programmable sample load enabled dynamic load TMA (DL-TMA) tobe performed for the first time, STAR* concept: ‘Thermal analysis has henefited enormously from the availablity of powerful but inexpensive personal ‘computers. The STAR" System was the result of the development of the TA4000 (1987) and then the completely new TASOOO (1992). Some of its most important features were its ‘* modular design, ‘© excellent measurement performance, ‘© unique calibration with FlexCal!™, ‘© fully automatic hough to result asessinent, and ‘© integrated relational database ‘The innovative DMA/SDTA861® Dynamic Mectiznical Analyzer was introduced in 2002. its modular design means that additional measuring modules can be developed in the future as required and integrated into the system using the STAR® sofware Finally, 2007 saw the introduction of the hermil Analysis Excellence line compeising the TGA/DSC 1 with its innovative SDTA, DTA and DSC sensors sample holders and the DSC 1 Both the perfomance and the operating convenience ofthe instruments were greatly improved. High priority was given to ergonomics and optisum ease of use. For example, the instruments can be operated using the intuitive touchscreen display or by actuating the hands-free SmartSens infrared sensors. The new TGA and USC models offer a choice of several different detectors. The TGA-DSC sensors simultaneously detect both weight and enthalpy changes with great accuracy. They differ in their maximum sample size and performance. The unique TGA-DSC sensor measures heat flow using six thermocouples and is very sensitive. METTLER TOLEDO Collected Applications Thermal Analysis in Practice Page 15‘The instruments can be connected to a computer directly via TCPA, ot via a network, This is an important advantage in a larger laboratory where several analysts operate different instruments and computers Finally, in 2009 the HP DSC 1 replaced the HP DSC 827°, It can he equipped with a PC10 gas controller which allows the cell pressure to be controlled. DSC microscopy is aow available for both high-pressure DSC and standard DSC. References and Further Reading [1] HL. Le Chatelier, C. R. Acad. Sci., Paris, 104 (1887), 1443. [2] W.C. Roberts-Austen, Proc, Inst, Mech. Eng,, (1899), 35. [3] K Honda, Sci. Repts. Téhoku Imp. Univ, Ser.t¥, (1915) [4] 8. L. Boersma, J. Am. Ceram. Soc., 38 (1955), 281 [5] ES. Watson, M.J, O'Neil J. Justin and N. Brenner, Anal. Chem., 36 (1964) 1233, [6] R.C. Mackenzie, Thermochim. Acta, 73 (1984), 249.3 Polymers 3.1. IxtRoouCTION... 3.2 SYNTHESIS OF POLYMERS. 3.3. TheRMoPiastics, 3.4 THERMoszTs. 3.5 ELAStoms.. 3.6 Pouywirk ADDITIVES. ea 3.7 USK OF THERMAL ANALYSIS TO CHiRacTiae POLIMERS. [REFERENCES AND FUWTIIER READING 3.1 Introduction Polymers (or macromolecules) are extremely Large organic molecules made up of very many smaller units (monomers). ‘They are widely used in materials such as rubber, plastics, and adhesives to name a few. The length of au individual macromolecule is typically 10 1m to 1000 nm and the molar mass is more than 10,000 g/mol, Polymers always consist of mixtures of macromolecules of different size and are therefore characterized by their average molar mass. ‘A low temperatures, polymers ace glassy solids. Ahove their glass transition temperature, they become more ar les sft and elastic, There are several ways to classy polymers, for example based on the polymerization process used to produce them, on their structure (linear, branched, or network) or as below on their properties (thermoplastics, elastomers or thermoset). ‘© Thermoplasties are linear or branche uncrosslinked molecules. The thread-like macromolecules are joined together through entanglement and intermolecular forces, Thermoplastics soften or melt on heating and can therefore be molded and reeycled. On cooling they may form a glass helow the glass transition temperature. Ifthe polymer chains ae uniformly built up and mostly free of sie chains, they may partially crystallize, giving rise to amorphous (non-crjtalline) and crystalline regions. Above the crystallite melting temperature they melt and are liquid. Many linear polymers are soluble in certain solvents and can be cat as films fom solution, ‘© Thermosets are network polymers that are heavily crosslinked to form a dense three-dimensional network. ‘Thermosels cannot melt on beating and decompose at higher temperatures. They are therefore normally rigid and cannot be plastically molded or dissolved. Their starting materials are more or less liquid and cure tothe finshed polymer during the molding process. Ahave the glass transition temperature, they become somewhat rubbery and soli. «Elastomers are network polymers that ar lightly crosslinked, On cooling, elastomers become glassy. On heating, they cannot melt or flow because of their-crosslinks, If their lass ‘emperature fs below room temperature, they are soft ane rubbery at normal temperatures, inder mechanical stress, elastomers undergo marked deformation, but regain ther original shape almost completely when the stes is removed. Since the polymer chains are chemically linked through crosslinking (vulcanization), elastomers cannot be molded or dissolved. Molding is therefore performed prior to vulcanization ofthe thermoplastic stating material. [METTLER TOLEDO Collected Applications ‘Thormal Analysis in Practice Page 17c d Figue 3.1. Sheotie dlogrons of deer poner mokeuks, br hropous amogiaste, The he moeaneeules oe snow in sient clas in trae snguen Pom oe es BSericystalinenameplasia ithe oar of he dagrom is chan ees eye Theorie of te mole ond ne fed colored melee a notable fo eta because oe encomly oeounng sia yOups € Flostomer Tm tivo masomieJes oe lad lwo pits coos ive) Thermos Tre ted mao ei) ar tre.cimersionly cressinked by the bie unm ooo 3.2 Synthesis of Polymers Polymers are formed when very many (up 10 several thousand) monomer units are linked together enil to end by covalent bonds, The monomer units are reactive molecules that possess at least one bond that ean be relatively easily cleaved, This allows the monomer units to be joined together through a chemical reaction Polymerization In polymerization, the macromolecules are produced through successive linking of the same or similar individu monomer molecules to form a chain molecule. If there is only one ype of chemical repeat unit (monomer) the corresponding polymer is a homopolymer; if more than one type of monomer is involved, itis a copolymer. 4 typical example is the formation of polyethylene, which has one of the simplest molecular structures. The basic monomer unit for polyethylene isthe ethylene molecule (CH), whose tro carbon atoms are joined through a covalent double bond Under favorable conditions of pressure and temperature and in the presence of a suitable free-radical initiator such as henaoyl peroxide, the double bond of the C atoms is transformed info a single bond, Teaving each ( stom with Lnpaired electron. As a free radical it can then form a bond with another ethylene molecule. W = CH > HiC-CH TWyC-CH; + H\C=CH» HC -CH;-CHyCly + As can be seen, the resulting dimer is also a free radical so that further monomers can become attached,Although the most important chain reactions are those involving free radicals, there are also other mechanisins. The reactive center at the growing end of « polymer can be ionic in character. Tonic polymerization is subdivided into cationic and anionic mechanisms. If the manomer has a non-organic atom (eg vinyl chloride CHl:=CHCI) or a side ‘group (e.g, propylene CHi;-CH=CH)), the side groups can occur randomly in the macromolecule (atactie polymer, litle tendency for cristallization) or stereoregular (Syndiotacic, on alternate sides; or isotactc, on the same side). ‘Copolymers: Tae properties of copolymer depend not only on the content ofthe individual monomer units but also on their distribution. A random copolymer exhibits only one glass transition, whereas block and especially grat copolymers show transitions that correspond tothe constituent homopolymers. — Monomer a Random copolymer = Monomer'b come Homopolymer: Homanohmera Graft copatymer Homopalymer b gue 3.2, The monorers canbe rordoriy dis bule in te copolymer moeoue 107 present nocks Sid chains co aso be gated era ha main chain Polyaddition In polyaddition polymerization reactions, mactomolecules are produced through the chemical reaction of low molecular ‘weight compounds with reactive groups, such as hydroxyl, amino, acid, isocyanate or epoxy groups. The monomers are joined to each other by means of the oxygen or nitrogen atoms. For example, the reaction of an epoxy resin with an amine begins according to the following equation Ao 8 (a H.C-G-R-C—CH, + HN-RLNH, <= H,C—G—R—CH-CH,NH-R-NH, ‘The reaction continues without stopping due to the remaining reactive group of each monomer, Three-dimensional crosslinking to form a thermoset is only possible because the secondacy amine hydrogen can also react with an epoxy group. Fach molecule ofthe amine therefore has fous possible points of attachment, In general, molecules with two points of attachment form linear polymers, and those with three or more points of attachment, three-dimensional crosslinked polyauers, Polycondensation In polycondensation polymerization reactions the same or different types of monomer molecules are joined together with the elimination of a substance of low molecular mass (usually water). A well-known example i the polymerization reaction of hexamethylenediamine (1,6-diaminohexane) and adipic aid (hexanedioic acid) to form polyamide 66 (PA 66) oF nylon 66. 4s shown in Figure 33, an H atom of the hexamethylenediamine resets with an OH group of the adipte acid thereby eliminating a molecule of water. The reaction continues at both ends of the new molecule and leads to the formation of 2 long chain, ‘The numbers in the name polyamide 66 (nylon 66) refer to the aumaber of earhon atoms in the two monomers [METTLER TOLEDO Collected Applications Thermal Analysis in Practice Page 19H i Bye | H hexamethylenediamine J adipic acid 9 Beye nee y ge eM H Oo polyamide 66 water Figure 2.3, Poyeondensction ot exor=terenoine ond ase ac fo polorie 68, 3.3. Thermoplastics In thermoplastics, the polymer chains are held together by weak bonding forces (van der Waals forces) and entanglement; there are no crosslinks. The chains can therefore easily turn and stretch under load. Semicrystalline thermoplastics contain both amorphous and ersstalline regions, The later disappear on melting, The properties of thermoplastics are very temperature dependent. Below the glass transition temperature (7), thermoplastics are rigid glass-like materials. At the Z,, the thermoplastic becomes leery, at higher temperatures rubbery, and finally more or les fluid. For this reason, many thermoplastics ae easy to mold and can be recycled, The influence of temperature on the elastic modulus (Young's modulus) of an amorphous thermoplastic is shown schematically in Figure 3.4. The melting and glass transition temperatures of @ numberof diffeent thermoplastics are summa Modulus (log scale) Viscous Ty ‘Temperature Flqwe 3.4 unde of emptor on he Sete modulus and he behsor oo thamoplanie Polymer | Glass transition temperature | Melting temperature Polyetiylene (low density) 120°C | 115°C Polyettylene (high density) -120°6 137°C Polvviaylchloride (PVC) Ce 175-212 °C Polystyrene (PS) 85-125 °C | 240°C Polypropylene (PP) “16% 168-176 °C Polyester (PET) 75°C 5 °C Polyamide (P8.66) 50° 265°C Tele 3.1, Sloss tansiionlngeralue a ling pare of usious herreplass,Some semicrystalline thermoplastics are polymorghous and can exist indifferent crystalline forms (e.g, PYFE). The gree of enjtalinity of a thermoplastic depends on s number of factors. Simple polymers such a5 polyethylene cnstallize most easily because there are no bulky groups present to present regular arrangement ina lattice. The degree of cystaliniy ofa thermoplastic also depads on haw the material has heen cooled from the melt. Slow cooling allows the macromolecules sufficient time to form a erjstal lattice and leads to u high degree of erstallinity. On the other Jhand, shock-cooled polymers tend to retain an amorphous structure Deformation behavior of thermoplastics ‘Thermoplastics undergo both elastic and plastic deformation under the action of mechanical foee. The deformation depends on the duration of the stress ancl on the rate at which the stress is applied. Figure 3.5 shows a typical tensile ‘stress—strain curve for polyamide 66. Yield point Rupture Tensile stress in MPa Clas deformation Plastic deformation 20- \ insar elastic detormavon 0 100 200, 300 Strain in % Figute 3.5, Typeal wns stess-stoin cure of polyamide 66 In the region of elastic deformation, two main mechanisms are in effect. On the one hand, the entangled chains are clastically stretched and return to their original position again after the tensile stress has been removed. On the other Jhand, entire sections of the chains can be shifted with respect to one another. ‘These shifts are reversible in the elastic formation region but the characteristic ime constants forthe relaxation can be hour or even months. This behaslos, ‘which is known as viscoelasticity, determines the deformation in the non-linear region of elastic deformation. Ifthe polymer is subjected to a tensile stress above the vield point (elastic limit) the phenomenon of cold drawing occurs. The ‘molecular structure changes permanently and permanent plastic deformation occurs. ‘the chains are partially disentangled, stretched and simultaneously oriented parallel to each other. This process of cold drawing. leads to Jocalized lateral contraction or the formation of a neck (So-called necking). The neck region spreads until the entire specimen has been drawn into the new shape. Onee the polymer is fully drawn, it is suonger than during the necking propagation, The chains are now aligned and more densely packed. This leads to an increase ofthe elfective bonding forces between the chains and thus to a final upswing in the stres-strain curve, When the maximum tensile stength thatthe material can withstand is reached, itruptuces or breaks Viscous behavior and viscoelasticity Polymeric materials behave both as viscous fluids and clastic solids; they are viscoelastic materials and their mechanical properties depend on time and temperature. The extent to which mechanical stresses cause chain slippage and plastic deformation depends on the temperature and the rate at which the stress is applied. Ifthe stress increases slowly or at high temperatures, the chains react and adapt to the force exerted on them. Ifthe stress is exerted rapidly o¢ at low temperatures, the slipping and stretching process does not have sufficient time to adapt to the stress and the material is brittle and breaks, ‘he visccsity ofthe polymer is a measure ofthe slippage of its chains an is therefore property that characterizes deformation behavior of the material. In the ease of thermoplastics, a marked temperature dependence of the viscosity according to the equation 77= noexp(Qa/RT) is observed. Here Qa is the activation energy for the slipping process of the chains and yp a constant METTLER TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 21‘The time-dependent deformation of a material under an applied stress is called ereep [1]. A purely elastic material responds instantaneously to the stress and recover its initial shape when the stres is removed. A viscous liquid, however, will deform as long as the stress is applied. The response of viscoelastic material is in between the two. In amorphous thermoplastics, the activation energy of chain slippage and the viscosity are relatively stall; the polymer undergoes deformation even with low stresses. AL constant tensile stress, the polynner first of all reacts with rapid stretching. [n contrast to metals, te expansion does not attain a constant end value. Rather the polymer continues to stretch slowly “This creeping ofthe material increases with increasing stress and with increasing temperature Another phenomenon that is also due to the viscoelastic properties of polymers isthe stress relief in polymers that have heen stretched hy a fixed amount, For example, the stress in a rubber band that hes been placed around a pile of books used to hold them together decreases with time. ‘Thermoplastic elastomers ‘Thermoplastic elastomers are a subgroup of thermoplastics that have heen developed to combine the processing auvantages of thermoplastics with the properties of elastomers. The elastic behavior of thermoplastic elastomers is however not due to crosslinking (as with elastomers) but is a result of the special segmented chain structure of the macromolecule, This consists of alternate, mutually incompatible, hard and soft segments or blocks. The hard blacks tend to ageregate in domains that act as erosslinking points, The crosslinking usually takes place through thermally labile physical interactions. The result is that thermoplastic elastomers flow at elevated temperatures and can be processed and molded in the same way as thermoplsics Further information on thermoplastics can be found in the METTLER TOLEDO “thermoplastics” Handbook (2) 3.4 Thermosets Thermosels are heavily crosslinked, close-meshed, three-dimensional polymers. Because of thelr close-meshe crosslinking, thermosets as a whole resemble a single giant molecule rather than a material made up of individual macromolecules, In fact, the individual macromolecules can hardly move. The result is that thermosets are hard and brittle materials with great structural strength. fillers are often added to influence their mechanical properties ‘Thetmosets are insoluble but can swell. Once thermosets have been crosslinked (cured) they can no longer be thermally ‘molded. They do not melt and cannot be recycle Iniportant types of thermosets are ‘© phenolic resins fr electric insulation boards (printed circuit hoans) and tubing, melamine formaldehyde resins for furniture, © unsaturated polyester resins for boos, travel trailers, ateralt parts and car bodies, anc © eyoxy resins for molding and adhesive resins, as wel as for printed circuit boanis. ‘Thermosets are mostly used as composites, The audition of glass fibers, or even better, carbon fibers, yields very si components of low dersty suchas are used in Formula | racing cas. Further information on thermosets can he found in the METTLER TOLEDO “Thermosets” Handbook (3; 3.5 Elastomers Elastomers re Lightly crosslinked linear chain molecules that form a wide-meshed three-dimensional network Elastomers are also glass-like and britle at low temperatures (ie. inthe range -10°C to -80.°C). At higher temperatures, they largely retain their shape thanks to the crosslinking. They do aot melt but begin to degrade and decompose ifthe temperature is too high. For this reason they cannot be recycle. Elastomers are amorphous and show very little crystallization during processing. They can be elastically stretched without permanent plastic deformation oceurring,starting material for the production of elastomers is natural rubber (eaoutehouc) or synthetic rubber. By rubber, we ‘an unerossinked polymer that can be crosslinked in a process celled vulcanization and, which as certain rubbery fic properties and can undergo plastic deformation. Crosslinking is achieved using a vulcanizing agent. This reacts suitable positions on the inacromolecules and joins different chains together (See Figure 3.1c). The oldest and most dey used vulcanizing agent i sulfur. hardness or modulus of the elastomer can be influenced by the amount of vulcanizing agent used: small mounts of ‘colfur (iypially 11) lead to soft elastomers. Larger amounts of sulfur produce a hard elastomer. “The composition ofan individual elastomer is very complex and is matched tothe specific demands put on the material ‘Eesides the actual rubber, an elastomer contains numerous ingredients such a5 vuleanizing agents, vuleanizing fccclerators, activators, vulcanizing relarders, Elles, plasticizers, stabilizers, oxidation inhibitors, antiaging agents, ‘pigments, and so on “sn example showing the typical composition of the tread of an automobile tre is given in Table 3.2 Ingredient Content Natural rubber 39% Filler (carbon black) 35.1% Plasticizer (mineral oil) 19.4% Processing agent 1.2% Antiaging agent 15% YVuleanizing agent (sulfur) 08% Vulcanizing aocelerator 07% Dispersing agent (stearic acid) 0.8% ‘Yulcanization activator (zine oxide) 15% “fede 3.2, Typ! compan ofthe tond ofan cutter. ‘Natural rubber (caoutchoue) is obtained as the milky emulsion of rubber particles known as latex from the tropical rubber tre (Iievea brasiliensis). The other ingredients are produced synthetically, Synthetic rubbers ate produced from very different starting materials such as butadiene, styrene, acrylonitrile, chloroprene, ethylene, propylene and so on Elastomers are classified according to the type of rubber used ‘Some important elastomers are ‘¢-natural rubber, NR, for articles of daily use such as shoes, sponges, seals, automobile tres, tubing, ‘# syrene-butadiene-ubber, SBR, for automobile tires, . Perit, TIR, wherever low gas permeability and good heat and resistance to aging, are required (e.g. automobile ose), ‘¢ethylene-propylene rubber, EPSV/EPDM, seals ‘© acrylonitrile-butadiene rubber, NBR, seals, tubing, ‘fluorine elastomer, FPM, seals, moldal parts, cable insulation, amd ‘© chloresulfonated polyethylene elastomers, CSM, wherever good stability toward light (UV), ozone, weather and fire is required, Further information on elastomers can be found in the METTLER TOLEDO “Elastomers” Handbook [4] METTLER TOLEDO Collected Applications ‘Thermal Analysis In Practice Pago 233.6 Polymer Additives ‘ost polymers contain diferent types of additives that give them special properties. Some important addtives are summarized below. Fillers such as carbon black are added to rubber to inerease the strength and wear resistance of tees or shoe soles. Inorganic fillers in the form of flakes or short fibers improve the mechanical stability of polymers (eg, polyester mised ait glass fibers). Calelum carbonate silicate or clay i often used as an extender for large volume polymeric pars o relatively low polymer content. Pigments sere a5 addtves for coloring polymers. They are usually inthe form of fine particles that are disperse: uniformly throughout the polymer mass (c.g, TiO» partiles fora white color). Stabilizers counteract the decomposition of polymers under environmental influences (UV-radiation, oxygen, water heat), For example polyvinylchloride requires heat stabilizers, Otherwise it would lose hydrogen and chlorine atoms ever ‘at room temperature with the formation of hydrochloric acid and the polymer would become brittle Since most polymers ate poor electrical conductors, their surfaces can easily become charged with static electricity Antistatic agents bind moisture from the surroundings, which leads to an increase in the surface conductivity of the polymer. Mest polymers are flunmable becanse they are basicaly organic materials. Flame retardants usualy conta chlorine, bromine or metal salts They prevent the occurrence of the spread of polymer firs. Plasticizers are molecules of low molecular mass that lover the glass transition temperature. They act as lubricant inside the polymer and so improve its molding properties. Plasicizers are widely used with PVC products to make th PVC soft, for example for water hoses 3.7. Use of Thermal Analysis to Characterize Polymers Numerous important properties of polymers caa be quantitatively determined using thermoanalstical methods such « ‘BSC, TMA, TGA, DLTMA and DMA. Table 3.3 summarizes the diferent types of polymer, the thermoanalytieal effects, an the techniques that can be used to characterize them, | Polymer type | kffect and corresponding thermoanalytcal technique | Thermoplasties | © crystallinity (DSC) © glass transition (DSC, TMA) «melting behavior (056) | ‘© thermal sibility, oxidation stability (DSC, TGA) | elastic behavior (TMA, DLIMA, DMA) | | fillers and filer content (TG8-E68) a | Thermosets ‘glass transition (often lies in region of decomposition) (DSC, DMA) curing reaction and determination ofthe degre of cure (DSC) thermal expansion coefficients (TMA) gelation time (DLIMA) thermal stability, oxidation stability (DSC, TGA) fillers and filer content (TGA-EGA)Polymer type Effect and corresponding thermoanalytical technique ‘Elastomers ‘© viscoelastic behavior (TMA, DLYMA, DMA) thermal stability, oxidation stability (DSC, TGA) composition (16) vuleanization (DSC) fillers and filler content (TGA-EGA) 3.3. Polen Woes, termoanaicleecs andthe ecriqus at can be used le anes hem ences and Further Reading [Niing, Elastomer seals: Creep behavior and glass transition by TMA, METTLER TOLEDO ‘Thermal Analysis UserComn 28, 13-16. METTLER TOLEDO Collected Applications Handbook: “Thermoplasties” METTLER TOLEDO Collected Applications Handbook: “Thermosets” METTLER TOLEDO Collected Applications Handbook: “Elastomers” D_R. Askeland, The Science and Engineering of Materials, PWS Publishing Company, 1994. J.W. Nicholson, The Chemistry of Polymers, The Royal Society of Chemistry, 1997, J.M.G, Cowie, Polymers: Chemistry and Physics of Modem Materials, Nelson Thornes, 2001 1D. |, Bower, An Introduction to Polymer Physies, Cambridge University Press, 2002. LW. Hamley, Introduction to Solt Matter, Wiley, 2000. [20] Claus Wrana, Polymer Physics, LANXESS AG, Leverkusen, 2009 [METTLER TOLEDO Collected Applications ‘Thermal Anaiysis in Practice Page 254 Basic Measurement Technology 41 Detistmion. : pisetamnesta 3 sie 6 42° Sexsmvn %6 43 NOISE cms 26 44 Deteenoy Lin. of 45° Dwr ei . Ay 4,6 ‘THE CoNsTANT, LIMITING FREQUEKCY oe eat 28 4,7 DigitAaL RESOWZTION svn SAMPLING INTERVAL, pe Snesnencte Rrra 29 4.8 CALIBRATION AND ADJUSTMENT OF SeASORS a oi — ? 2 ‘Temperature scales iis : mere) 4.9 ‘Tw Mosr DaroxTaNT ELECTRICAL TEMPERATURE SENSORS. Seasick PRO 4.10 TEMPERATURES IN THERMAL ANALISI cc nnnesnnn i aoe ‘The aim of this section is to introduce and explain different terms and expressions that a newcomer to thermal ‘might encounter. 4.1 Del ion Sensors transform the physical or chemical property being measured into an electrical signal. The signal is usual xznalog, The term sensor covers a wide range of diferent measuring device. Ideally, the measurement signal produce by the sensor should be a unique function of the propery i is measuring. Quite often, the function is non-linear (e thermocouple voltage as a function of temperature). Ifthe non-linearity of «sensor is known and is reproducible, it ca be easily mathematically modeled using appropriate software, 4.2 Sensitivity very sensor has certain sensitivity, This is defined asthe siz ofthe electrical signal per unit ofthe measured quanti For example, a copper-constantan thermocouple at room temperature has a sensitivity of about 42 nV7K, See alt detection limit. ‘The behavior of sensors is normally described using polynomial mathematical models A+Bx+ Cr GD ‘where y isthe quantity effectively measured (e.g, the electrical resistance of a resistance thermometer) ‘Ais the onlinate intercept, B the slope (sensitivity ofthe sensor). C and possibly ational terms are needed to descr the non-linearity of the function. x is the true physical quantity, 4.3 Noise Signal noise is more important than the sensitivity because modern electronics nowadays allows even very weak sign {a be amplified. The noise is however also amplified. There are three main contributions to noise: 1. Real random fluctuations ofthe quantity (eg small flichuations in temperature), 2. Noise occurring in the sensor (statistical measurement errors), and 43, Amplifier and analog-cigital converter noise. Noise can often be reduced by contzolling the environment, For example, with a balance, the first two contributions noise can be diminished by using a weighing table (dampens building and floor vibrations) and by weighing in closroom (suppresses air turbulence). A noisy weighing signal can also he smoothed (averaged) in onder to obtain precise weight value. Weighing of course takes longer because ofthe time delay before the display stabilize corresponds to an alternating voltage of different frequencies superimposed on the signal. For this reason, as ‘alternating voltages, the noise is given as the root mean square value (rms) or the peak-to-peak value (pp). The atio is 2-2 = 2.82 for a sinusoidal oscillation, and about 4 to 5 for random noise The noise of a temperature measurement device with a copper-constantan thermocouple is 0.5 pY pp (ie. us), or 0.01 °C pp (ue. 0.002°C ms). ce Signal Noise uw 0.6 {Noise pp = 0.59 yi Noise ms = 0.12 4 “08 Jppims = 49 a4 Time /s a 10 20 30 40 50 Fue 6. csivoton of ptoko-penk ep) andro meen gue (ms) vues fom random nia cffeetve rms value can be ealeulated from the equation: (4.2) ‘mis the number of values, x the individual signal values, and ¥ the mean value of the signal. Detection Limit detection limit (often incorrectly called the “sensitivity") refer tothe smallest change in the measurement signal can be detected with reasonable certainty. The detection limit must of course be cleerly larger than the background ise, for example 10 times the rms value (equal to about twice the pp nose), See also TAWN sensitivity 45° Drift ‘When measurements are performed over long periods of time, the slow drift of the mensurement signal becomes ‘goporlant, not just the stalistical noise, This drift is given in units ofthe measurement quantity per hour or day. For a ‘balance, the drift can be significantly reduced by thermostating ‘the drift of a measurement curve is reproducible, the curve can he saved as a “blank” curve and subtracted from the ‘measurements that follow. METTLER TOLEDO Collected Applications ‘Thermal Analysis n Practice Page 274.6 Time Constant, Limiting Frequency In thermal analysis, physical quantities are usualy displayed as measurement curves. The signal produced by a sensor cannot follow changes in the measurement quantity infinitely quickly. For example, any thermocouple has a certain heat capacity, C, and is connected to the medium to be measured via a thermal rexstance, Ry. The product Ra C corresponds to the ime constant, 7 (tau), ofthis sensor RyC 43) ‘The thermal resistance is given in K/mW and the heat capacity in m)/K in seconds. The time constant is sometimes called the response time. t mW's/K) in order to obtain the time constan Output que 42, Attamacoule atest macsure the to fempertu of wor bo. The het flows eer the war ates 0 fhemal estan Ky toe $oldoed jin, wach has o cea net enpacy, As the following figure shows, the measured signal approaches the tre value asymptotically, provided the value remat constant Ifthe true value increases Vinearly, Uke measured signal lags behind by an amount given by the time cons (to lag” means to fall behind.) a wt nt” « rca 3 ign =3 so] rr i i en, eC Faure 4.3. Tho ne signa te Ingus shown ray Qe rectanqur, righ hanger, both wih fopirad ol 60 te cup egret (a> ow RC element cmon smred wih fhe consert e' .The ng equaney nis sup 0.05 Hz. the ting pring 19. ‘The reciprocal value ofthe time constant i called the limiting frequency, ay (angular frequency,» = 2279) =¥e or f, = : aA)thatthe limiting period is given by P= 2nr 5) ‘expression “limiting frequency” does not mean that a signal is completely damped above this frequency or thatthe is not deformed below this frequeney. Higher frequency signal changes are increasingly damped and are therefore Jonger “resolved”, This means that close-lying events ace not properly separate: 100 Siew! 0 sain 0 Hema 3 7” 0 50 0 0 0 0 ee ee ee ee) Figuie 44, Theo angular input sina of 0 fo 40 covespenng te mng aio ot ‘hs soncer, Ty oe wel esohe art hotcy damped. aca wh sgnfeanty shore pods {ond | 9) 99 Sergiy cared and tne arpa auc To about 8%, 4.7 Digital Resolution and Sampling Interval “Analog sensor signals are digitized so that they can be numerically displayed and electronically processed, The digital ‘resolution of the ordinate is chosen so that the last decimal place displayed is somewhat noisy. ‘The user can then ‘monitor whether the sensor is functioning properly (e.g. no noise a all or excessive noise are important alarm signals) Jn the case of the copper-constantan thermocouple, a sensible resolution would be 0.01 K. It would, for example, be ‘nonsense to resolve the noise 100 times forthe sole purpose of obtaining impressive values (0.1 mK!) for technical data forthe digital resolution ofthe instrument. ‘The analog signal is usually sampled at equidistant time intervals. The shorter the time constant of a sensor, the shorter the sampling interval must be to prevent the loss of information. An interval that i 3 to 10 times shorter than the time ‘constant of the measurement setup is optimal. Shorter intervals result in unnecessarily large data files. If no abrupt changes of the measured quantity are expected, the sampling interval can be increased without losing information {especially with very long measurements). 4.8 Calibration and Adjustment of Sensors Sensors must be calibrated st regular intervals, The calibration procedure checks whether the measurement deviation or ‘measurement error is within acceptable, individually specified error limits. If the exror is larger, the measurement system must be adjusted, that is, instrument parameters must be changed so that the eror is smaller or eliminated, Calibration requies reference materials with accurately known properties, that is either ‘© a property that defines the scale concerned (e.g. according to the Lnternational Temperature Scale, 17590, pure indium melts at 156.5985 *C, or the water-ice equilibrium at 0 °C) or ‘© scented reference substance (e.g. a mass standard of 100 mg + 5 jg). If no such reference material is available, other possibly les accurate “standards” recommended by experts in the field concerned are usd. METTLER TOLEDO Collected Applications ‘Thetmal Analysis in Practice Page 29Temperature scales ‘The temperature is 2 measure of the mean kinetic energy of molecules, atoms or ions. It follows from this that there i an absolute zero temperature below which itis impossible to go and at which the kinetie energy of molecules, atom aud ions isa minima, Since all physical and chemical processes are more or less temperature dependent, temperature is & very importa measurement quantity. For practical reasons, temperature measurement is based on comparison with a define temperature scale ‘The International Temperature Seale of 1990 (IS-90) is based on 14 primary fixed points. ‘these include fo example 4 the triple point of water 0.01 °C), ‘© the melting point of indium (156.5985 °C), ‘© the melting point of aluminum (660.323 °C) and ‘¢ the melting point of gold (1064.18 °C). ‘The two temperature scales in common use in the SI system differ in their zero point. + ‘The Kelvin scale, known as the absolute temperature seale or he thermodynamic scale of temperature, begins at 0 K, zero Kelvin. The unit is the Kelvin, ‘The temperature of the triple point of water is assigned to the value 273.16 K. 1 K is the 273.16" pact of the triple point temperature of water. the Celsius scale begins atthe melting point of water at normal pressure (275.15 K) and has the same scale division as the Kelvin scale, ie. a | K rise in temperature is the same as a1 °C rise in temperature. Two-point and multi-point calibrations are particularly recommended because they improve the modeling of th sensitivity function of the sensor. If all the measured values deviate from the reference value by about the same amount, itis sufficient to shift th ondinate intereept of the function (one-point adjustment). If the deviation increases with increasing values, the slo must also be adjusted. If sufficient calibration points are available (if possible distributed over the whole measureme range), non-linearity can also be adjusted. In the sensor polynomial, A+ BrtCx2 a) 1y is the effectively measured quantity (electrical value) or the physical quantity of temperature) still subject to erors terest (e.g. the measur A,B, C and possibly other terms are sensor parameters. xs the tre physical quantity (@.. the melting temperature a reference substance) For example, we want to calibrate an electronic thermometer using a thermocouple as a sensor. The reference substan is disiled water in atest tube. For the first measurement, the water contains ice cnstals (0 °C) and for the second vvater is boiling (at normal pressure 190 °C). We hold the thermocouple in the middle of the iee-water mixture an rea off the temperature as son asi is constant. Afterward ve boil the water above a Bunsen burner using boiling stones promote boiling, When the water boils, we hold the thermocouple slightly above the boiling water in the vapor phase x read off the temperature as soon as itis constant. Ideally, the measured temperatures are 0,0 and 100.0 °C, as in Case 1 in the diagram. In practice, Case 2 with values 1.6 and 102.2 °C or Case 3 with -2.5 °C and 103.7 °C are more likely to occur. The observed deviations are then plotte Against temperature. .——— "enperatare dviion read mins useage iC ‘Trae epee ia °C 2» 6 2 mo gua £5, Poof 7a msosuement ero! heme 1 Tho hemomoer sus the expect lpetosus. No odjsen s necessary 2: The masued vas re choul2°C to gh Hie vol al he ordre lop, is odusad by about 1.9, raw meceuemer should gle coceabevoluas of -0:3 "Cand 100.3 °C 3 Inti cose, both ond Me sope, 4, Pave To beads A es measure aid gies the teat ea shown 8 | ‘Ths means hat efter ogksmen, 2 naw choral sarc curve is obsined. To acta oust ment move neurite sotworatroauomataly elculaes Me pew poraraes. 1 ‘nd (and posily chars), The asu of re odjustmert shoud enekod by parorring 0 ne calor (¢ feasts one-pot measuremer?. 4.9 The Most Important Electrical Temperature Sensors Jn thermal analysis instruments, temperatures are nearly always measured with resistance thermometers and thermocouples. Resistance thermometers depend on the temperature dependence of the resistance of electrical ‘conductors or semiconductors. Very often the Pt100 sensor is used whose resistance at 0 °C is about 100 $2. In the range 180 °C to +700 °C, its electrical resistance, R, is given by the equation R= 1000 + 0.3908QK' T+ -580210°OK? TF (4.6) ‘The PL1O0 sensor mate of coiled platinum wire exhibits excellent long-term stability. Above 700 °C, there isthe risk of ‘ecrystallization processes occurring which could change the resistance at 0.°C. Thermocouples consist of tvo different metal wires that are joined together (soldered or welded) at both ends to form circuit. If the two junctions are at different temperatures, a continuous current flows in the thermoelectric circuit, If the cireut is broken at the center, an electrical voltage can be measured that is proportional to the temperature diference and the nature ofthe two metas, Thermocouples are therefore ideal for measuring temperature differences, for example, ata thermal resistance to measure the heat flow through the resistance, Actual temperatures are measured by holding the second junction at a constant temperature (reference temperature or comparison temperature or also ‘electronically generated). | Ftace with hot soldered junetion Thermoclectric voltage Ice water with cold soldered junction Fou 46. Meosureman ote a tempiraue Fa fine using thormpeoupes. ha plein tes. ae raw back tnd fre ptm asian wie gry METTLER TOLEDO Collected Applications ‘Thetmal Analysis in Practice Page 1‘The sensitivity or slope, S, of the thermocouple is expressed in yiV/K, Sis temperature dependent and in genet increases with increasing temperature. Lnportant types of thermocouple are for example: ‘© Copper-Constantan (Type T) from -250 °C to +400 °C, Soyo tore ~ 42 N/K. 4 ron-Constantan (Type J) from -250 °C to 700 °C, ion rusts easily, So ore ~ 54 RA. «© Nickel-Chromium, practically identical to the Caromel-Alumel (Type K) from -200 °G to 1300 °C (in non-oxidizin: stmospheres) or to 600 °C in oxygen-containing atmospheres. © Sow ineec ~ AL BWR, © Platinum-Platinum 10% Rhodium (type 8) from 0 to 1600 °C. Note that certain substances, so-called platinum. poisons, change the thermoelectric properties, especially in ralucing atmospheres. © Sprouse ~ 6.4 BWAK, ‘© Gold-Gold Palladium from -200 °C to 750 °C, Sisc-e ~ 9 HWA 4.10 Temperatures in Thermal Analysis A temperatuce difference, A; that depends on the thermal resistance is necessary for heat to flow from the furnace the reference point and to the sample, In METTLER TOLEDO designs, this is achieved by imereasing the furna temperature by the same value of A. The time difference between 7. and 7; is equal to the time constant, + independent of the heating rate. The heating rate, A is equal to the slope ofthe triangle shoven in the enlarged cise section of the diagram in Figure 47 BHAT Ita From this, it follows that AT =f tux 4 At the beginning of each heating segment, the calculatel temperature increase, A, is added to the set value of fumave temperature Temperature _ Time = Pras Fipuse 47. Te tree impotent emacs atthe race lernpertue, 7, he rperciue thereerence port, 7; andthe soma lroerlu, 7 Ica ferel avn accu (meting). The enlogad sede shows the w'stonshe bets he Neaing hs log a constr cri 37, tne tempers prcgam show et the eon neeoead tempera rc exay corpaneces A7 1s apie a he aging o the dare sear “The dium on he rght 2.6 he Usa emperor progr who the eMPTOE INCENSE yy aston peed hat st rom hm ta 0 ‘emoeroie progam unt eyrorvcequllorum eoches (he DSC signal has sola).do we use the increased temperature? shortens the transient period (or setting time), pean that is, the time taken to reach dynamic equilibrium (the refecence and sample have assumed the heating rate ofthe furnace). i linearies the temperatnee progeam ofthe reference point, Since Aes depends on the temperature, oly the Furnawe temperature is linear in the usual temperature program. ‘It makes the reference temperature independent of the heating rate. ‘hag and its temperature function (nae = A +B7'+...) is usualy taken into aecount using Total Calibration [METTLER TOLEDO Collected Applications, ‘Thetmal Analysis in Practice Page 335 General Thermal Analysis Evaluations 5.1 Tus Ovrtwust CoonDinsre SisTeM, a4 5.2 Rowe Dacraus. a -asareypmmaacinenereionlinne th 5.3, DIsPLaving [NFOSMIATION FROMTHE DATABASE Se 35 5.4 Opmivizine’tws PRESENT®TION OF & DNGkAM ao scaciccrcoonlll 5.5. NowsausanG MEasumenenr Cres 10 Saurus Mass esi soracmaese 5.6 DISPLAYING CuwQes WIT Restuc7 70 Thue, REFERENCH TENPERATURE OR SAUPHE TEMPERATURE. 7 5.7 SaMPLe TrvPeRATURE AS AFUNCTION OF TAKE... arti 10:0 BB 5.8 CURKE ConRNCTIOV USING 9 BASELINE SEGMENT sonia are 38 5.9 TILEMATICAL BYALINTIONS sen : aii 39 5.10 Cums Costasison a : . singgaciAl 5.10.1. In One Single Coordinates 4 5.10.2 Multi-Coordinae Systems a 5.11 Nowenica Excusrions ernie - 7 % SALT Onset saree 5 SAL2 Elstner ‘ z 5 5.13 Onset and Ensen = - “6 S114 Logarithmic Onse, Eadset, Pes, Lagarihie Step Horizontal/Tangential 46 SAILS Peak bs " SL Tables. anette Sr is 31,7 Minirnum-Masimemn. : = : ‘8 3.118 Signal Vaine is This chapter deals with the evaluation of curves obtained from the diferent thermal analysis measurement techniques Specific evaluations can be found inthe sections dealing with the particular measurement technighe ‘The first part discusses various graphical display possibilities, he section on cur topic o the simultaneous presentation of several curves. ‘comparison focuses on the important ‘The second part covers generally applicable evaluations that give numerical results, 5.1 The Optimum Coordinate System Itpays to conser in advance which ordinate and abscissa is best for a particular task. For example, it makes no sense evaluate DSC curves that are displayed with respect to time and then alterwand decide to change the abscissa to tht reference temperature and thezeby lose all previous evaluations. 5.2 Editing Diagrams ‘This includes cutting out, copying and inserting text, changing the sample size (correcting. weighing errors @ referencing the evaluations tothe active or dry mass of the sample), the sample name (e.g, typing mistakes), the color the type of line and the font If necessary, lines or arrows can be inserted to clarify particular features (Figure 5.1)Somple: Gyoerd tripalmtate, 3.2800 m5 Now name Now sale Figue 51. A ping aor nth sample name hs bron cord and sme moss alr roster Ge. winout he vote cova) has been ete fr ier clslaons, Atows and lines are crm ilesoe dleent pois. Imperial fete hgbigiied by using loge etexocas, The OSG uve sows fhe polymorpic beter a aft ocorsed a Kin, Displaying Information from the Database ples: Name of the method, the sample, the customer, type of gas used, adjustment parameters, temperature {shown graphically). ‘em trate von tates ‘Curve !2-Nivophenol, 13.10-2006 0843.07 wn ‘Sample: 2-Nirophenal, 1.704919 6 en EET Ear aatraaE Te zo 260" 280" 900” 20" 340” 960” 360 400 420" Fo 200: bom mm m8 34 35 38 40 mn Figue 52. Dien nmin om tha datobase,Abova. Th cue nor {poss wih symbols the ircae changes: for exam [] means farts 2 sacton ofthe cue and | manrs some mass ‘station win a dob of tho measursret ond sample ome vith Some weigh Baio: The ‘emperaue program. The cuve stave the them) dacomposion of 2-etepherel OSD ah pressiecrucle (aang fle 10K) [METTLER TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 355.4 Optimizing the Presentation of a Diagram ‘This includes the following: © Automatic scaling (displays the entire curve in the diagram). Zooming (displays a desired section of the curve on an expanded scale) 4 Displaying more than one coordinate sjstem in the diagram, either as desired or exactly superimpose. Configuring the coordinate system linear or logarithmic axes, with or without gridines, entry of numerical limi cexaetly like a template, so that diagram can be displayed ‘© Displaying the axis relatively as inthe upper curve of Figure 5.3 or absolutly, © Defining units of time (, min, h) and temperature (°C, K) 00 apn Options Saree, Hh oni PE Roe 267 Relaive ordinate scale: the curve ean be shied] verti or easy une camarison, Absolute ordinate sone tosee te eract ‘rd voles Flow 5.3. Ihe wo cooirase syste or ret csplaye ntl with so ha hee oom i etnias on The it The two abscissas auloratialy corescond excl. The oper coors system shows. ‘he DSC eure inne oun ha ova halved wi espe e comp mass in Wg, Me use cordnete sistem has dines, The Agu shows the DSC maiing cunes of aeent po}ahyene samples 5.5 Normalizing Measurement Curves to Sample Mass When measurement curves are normalized, the ordinate unit changes from mW to W/g (Figure 5.3), fom mg | (TGA) and from wm to % (YMA). Normalization allows you to more easily compare curves recorded! using dilfe: amounts of sample. The curves are not identical because larger samples produce broader effects.Normalized TWA Curve of Quartz 96] Semple Quart, 4.3887 mm 100, Mettes My 100,00 % ae a 4505°C Max. 100 91 % at 594.30 °C roo. r002. 000: sq Too 10 Pop 250" 80” 380 400460" 600" 580600" "C Fours 54. Th exursion cuve of oso pece of quart wh he display nowelze wih resedl 0 lengh. Te poo ison al 577 °C is us fo he hanson ef «quate fo f-quaz. Displaying Curves with Respect to Time, Reference Temperature or Sample Temperature ‘The reference tempersture is normally used for the abscissa (default). In certain cases, the other pssibilities are also ‘sluable especially for comparing curves, ‘Note: DSC curves displayed with respect to sample temperature can sometimes be “non-monotonic”, i. they can show snore than one ordinate value ata particular abscissa value. Such curves cannot be directly evaluated ‘wo Non Monotone Cvs tswatr,10:12.2002 29 5418 Water, 9000 mg ig Meting of sine at tee diferent etn aa fates and rsializaton rns m between 40 MC comang 30 0 wo Fique 5.5, Two exomplas of OSC cues dsloyed wih respect fo samol emcee. Above lf: The Insitod clagam shows fhe cystlizan of vt on colng at 10 Kr. Cystaizctlon doos not ‘ogi url about 15°C due fo marked supsccoling, The enhapy of cylalizaion veloc or Me ‘aud etlzaton recut in Momertary inciease in he samp taper. Tiss yi peok ‘as an unusual slope, Th main diagrams show ho repectes ming (heating rales 2, ant 10 Kin) ‘on erjtilioghon of zine which dows nol show any apaeschle suparaating, In tis casa, the ‘ntalzaton peck does ni opm. [METTLER TOLEDO Collected Applications Thermal Analysis in Proctice Page 375.7 Sample Temperature as a Function of Time If the thermal effects are strong, the sample temperature does not exactly follow the temperature program. This evident from the display of sample temperature versus time (Figure 5.6). Samp Temperaure as Funston cf Tine ater, 10000 mg so] Seeing at Kinin mv 3 1 58 0 $10 45 20 25 0 | ee 8 min bie 3 song ‘tow “10 teneeroure ‘tee “tes ‘2000 “toa i ton sa? 42200 “1070 — ‘aaoo i161 20+ 42400 1254 Figue 6.8, Kove: The OSC cooing curve ct wot Balow: Tha sample mgarsue slay using te ‘gme cbscssa Thy sample of 18 mg vate! enclosed In 0 fwmisoaly seoled eondard aumru fuse wos coled at 5 Wen ond cystiza at chou -15 °C (ue fo supecootng). The ently of (tlizaton eaves the sample tempeckae To iaraae, The ena cut cf sample isnot suet feeb reach he meting port of °C. Nevetnaess, the sop femparctu Ioteases fo 10.7 °C see 7b one). 5.8 Curve Correction Using a Baseline Segment An artifact in a measurement curve can be simply replaced by a segment ofthe baseline. The “Spline” baseline type most often used Advanced users use this possibility to discuss hypotheses when reactions overlap.soo aoveiot 2.Nlvopheno. 17040 mg Heating rate 10 Kirin FrstPeak ee Imegra) 75 28m Totals 450094 4 if a zo” 260 280 900 20-30” 380” 30” a0 4B Figure 6.7. Agplootons shoxng he replacement cf par of « DSC cue bya baseline: Ire bck cure shows the decomposton of 2-nfrophencl in a hghpessue cule. Te macsuemert wos dstubed {arow af 13 °O) by seeing o craven ho ftorrory Donen. Inod a opcting the measure, ‘aff! we cu! or rploced using tha Spiro baseline eg he red cune. Tho fst emall ak fhe red exra wat the removed un alaoad ty 0 Spina bascine oobiin Fe bly curve. The Bue curve wos then sustaced om the fad cue ode To obtain the dfeence pak (Fist Pea, aber en, Mathematical Evaluations include the following ‘The calculation of the integral curve (Figure 5.8). ‘The calculation of the first and second derivatives (each of these curves can be evaluated or a further derivative calculated (Figure 5.8). Cutting out the curve outside the range of interest to omit unimportant details. Smoothing of curves using smoothing polynomial with a selectahle width of window and onder. This allows you to select the degree and type of smoothing (Figure 5.9). ‘Adding « constant to the selected curve in onder to shift the zero point of the curve Multiplying and dividing a selected curve by @ constant, Adding, subtracting, multiplying or dividing curves. Separating measurement curves of complicated temperature programs into the individual segments, Calculating the upper and lower envelope curves of periodic curves. Subtracting a hand-drawn Line (straight Jine or polygonal line) irom a measurement curve (Figure 5.10). Polynomial fit: for example when validating 2 method, the results can be entered as a function ofthe actual values in onder o display them graphically and to fit them. Te polynomial ft is also used to calculate « DSC curve from the DTA curve. Fast Fourie transformation for periodic curves such as occur in ADSC measurements [METTLER TOLEDO Collected Applications Thermal Analysis in Practice Page 39see itecenteon an migration ‘Semple: High oriented PE fer, 282 ma ros a Bm ‘st devatne of OSC ave Lo ib wo a0 250 SOD aD AUO abo SOD” 580 BOD OS i Intagral of O50 curve 7g ew OO ta TO HOO Figure 6.8. Top The D90 curve ofa Dynoema ter (op flawed ty tots! dete afte DSC cure (edie) tnd te integral cane (bom). Te ist dete cave ofen shows smal changes inthe ‘haosurenest signa more cea than he criginal curve. The gral crtsponds te aru of heat Transported he somgl ossuming Mate DS cure on whi Ils bosed has ben ion cones nexo ‘Smeething Sample: 25 mg Azoxydarisole Rate 0.1 Kinin Celt DSC8228 Smocthed using z5pans Ts Die tao se 1500” 88S Figue 5. Cure smoothing: Ihe ture shows te nsmaiciso!Ope tension of o sal sample of 0 Tiulf ental meosizes a @ very 1 Pong flo. Unda es conte, the Fanston paok 9 very ‘anal and the Pgh expanded DSC carve 1 ney (Ove). Hist orter smmoathing Is cpl to he ue wih a raow of 25 pons, ha ighaquency nose dsappcor. Ta peck shape 1, weve otic delaSubsacon of Bo ‘Same LOIR) MKPOG, 4.3900 mg Costing ate: 2Kimin 002 Integral 0.46 md normalized 0.11331) we en Ne ne ao Wh az as ee gue 8.10, Poygon subracton: The figure shows th cooling cuve of aig crys Assuming at the Fre sharp O30 peos eon the shoulder of te as nook, a sald con a down 98 apo gonal lin fed dashes). Te peygon fs new subrated om he OSC ove and ciara ur (ue) is ‘bioina, This con then be negates using a sraight baseline 10 Curve Comparison shermal analysis one often wants to display several curves in one diagram. 10.1 In One Single Coordinate System with identical ondinate and abscissa units are automatically displaved in a coordinate sjstem (Figure 5.11). The can be moved verticully with respec: to one another so that they do not ross (Figure 5. exo ree n re Coordin Sym ‘Sample Epoxy UHU rapid, 3.3000 mg eating cate 2 Komin Heating rate $Kimin Hosting rato 10 Kim Figur 5 1 tse dame 9G cues or opened ther ascer ina single coorcinae system. The ‘exemple shows tne epoxy curing recin of UU Rapid METTLER TOLEDO Collected Applications ‘Thetmal Analysis in Practice Page 41a0 Sopraton! Cuwes ie ‘Extrapo| Peak 3230°C ‘Sample’ Epony UHU rapid, 8.3800 mg Extapol Peak 4827°C Heating rate 2 Kinin exrapol Pook 873°C) Heating ate 5 iin Heating rate 10 Kin 7 mB que 5.12. Ta tv cares fom Fig 511 con bo sparc veteay using the mouse. The miv-aHs Ch te reved ony fe poston fo he daa. Futhemars, he ecopoiak ps fares oF ‘lu fe shaw tom he pook ep cf eacon seak depends on ra Metin oe, Itis also possible to zoom and display particular details of a surv y curve (Figure 5.13). si Zoned Woo . - zs DSC af sufpyeine shocecooled alr meting ae Glass transtion een rset 54.190 joo mae é Integral 65992 m) ormaled 4.11391 wo wo tO 18) gue 6.13. The upper SC oun Is suey tun. Below: The sections of ket ae zoomed 20 Is init crane destin. In his tramps he sects oe Ihe gloss ronson and fhe manaopc oi {Did otic, both of wc oon roxy bo seen In te suey cure. Prior fo rneasuomet, the ‘topyrn was enledarthen shock ened 90a soli fo form @ glass The esting oe woe xin 5.10.2 Multi-Coordinate Systems Jn multi-coordinate systems several coordinate systems of course appeas. This occurs «if curves from different measurement techaigues are compared, eg. TGMDTA (Figure 5.14) or ‘© if curve and its derivative are displayed (Figure 5.15).‘want to view both the complete survey curve and tvo-dimensionally zoomed details in the same diagram, you frst it twice, After you have cut out the desired region, you can change the onfinate or abscissa units ofthe zoomed (ordinate normalized or the abscissa displayed against time, see Figur 5.16). The two coordinate systems can. also displayed one on top ofthe other (Figure 5.15) and next to one another (Figure 5.17). soto {Comparten of Throa Teawiauet negra 412.230 Utuimt — 700% Right Lm 23666 °C uatblume 37080 °C Aight Lat 2366 °C eating Rate 30.00 °cmina Sasol Tpe ine ‘Step -1.8857 9% “1262 me TMA mount Onset 211.71 «C ‘ ma sep 96.2021 "| 1000 6.5803 mg Residue 1.7537 % 0199 ma joa" "Ye0"” "200" "280" 300” 360” 400480 S00 Figure 8.14, Ihe gue shows tee arent reasurmort curve deployed wi roped the rarance Jagatoiuve Te mart ies of soo ouNeHow otendy bee ssced. Th crnla seas for this ‘ype of comparison ae na 20 inerian! anchors ra shan ‘TGA and DTG Curves of PAS % | TGAcuve yoo Te 200 B50 a0 380 ao aS SOU 091 wee] DTG cuve Sample PA, 6.8386 mg Woo Te 20250 90D seo aoa OO Figue 6.16. Tis isan lag’ way vtomalinay spl fe ago. You con dois wih any ruber of cordnnre syste, METTLER TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 43‘ayo wpe 0 eu patois #9 286 UBD QL arEly UDN SN OH Bu CLS NBS woyeraesAdou3 pw o 2.9498 wodown Scenes 198 touwues, see btm 00 ops ss joes pawooz onsen, Fa RRR aeRO ‘swaps omy yun jor wo saa 9 andy pepwonaio sauna Woop au SMAIAS OUIOES BU! 0.79) 4HA “eKOH "WOIBE Sr swapts acuiptoo yo quid ey fg OU st oxy “cuud uy Ak cus oy Ur paKOS (Sy gamed jo ubo sumo yO ¥ "=U09 4m JAN) AU U FOIE 8} ORI aBLCO euL—pyoNION® ‘uo patooe 3, (UojeuD4 SO} Aad) waa JO od EUs UO ‘SKuoHB sy wl G ary 9 os = o FE a tee vogexea Aolemua 2.8195 odo Slebes ieu0 Cons S56 Sys s9ys paween-nous bn ‘supuddeyrs 950, 00 uousuen se evar powarz Tore aeRO1 Numerical Evaluations ‘Area determination (integration) is used mostly for DSC and is therefore covered in Section 7.6. The step ination is also described in the same section. }1.1 Onset ‘of the most frequent evaluations made in thermal analysis is to determine the extrapolated starting temperature or ©. This is the point of intersection of the haseline before a thermal effect and a tangent (often the inflectional at the steepest part of the curve). More information about the onset can be found in Section 7.6. nase and endnote THA Cave Um) sample: Enarnal on copper wire = Conse 251 86°C 00- 260. 260. 240. 220. yao” 10” 180” 200” zk” 240” 260” 280” 900” 20 a gue 5.18. An nlf coppar wa isc inthe TMA at 1D Kinin in an a amas in his ‘arpa tre dacompeson region fh insulin yer sO eres You can set se nd se valuation fi. 11.2 Endset the word suggests, it means the extrapolated end of a reaction of the sample, that is, the point of intersection of the with the baseline ater the effec. Te endset i the mirror image ofthe onset. [METTLER TOLEDO Collected Applications Thermal Analysis in Practice Page 4S5.11.3 Onset and Endset ‘Stop-like effets can be elegantly evaluated as shown in Figure 5.19. ConeetvEndsot Exaluation um] TMA. ‘Sample: PTFE radial, 24200 mm 2458: aso: aes: Onset gas9"C : Enisst_ 280°C 2a: bree PL 2475°C Ilse Sip, 187 uno Lak ume 1788" 2435: Rattume 3871°C HestingRate 5.00 "Cron Resut Mode Sarple Temp ao ° ie 20 30 a Figure 5.19. Te TMA measuretert shows tne symone ston of elon ne rx don ot foom tempat. The slandare ooaul reus ora ons ond ih ast, The mere oh pont UGeitodtar the sop al the pot ot lon, he evlueton iss, te heng re ond the rest Mode ‘19 aval oe optons. 5.11.4 Logarithmic Onset, Endset, Peak, Logarithmic Step Horizontal/Tangential ‘these evaluations all have in common that the tangents or baselines are drawn to a measurement curve and poin intersection are evaluated, In logatithmic display, which for example is often used for the elastic meadulus, a st fine of constant slope would be curved. These special evaluation yield straight line tangents in logarithmic displays, The results of the Kinear and logarithmic presentations are of course not identical. Por exanuple, the onset ofthe transition of an epoxy resin in a linear presentation is 104.5 °C, and in a logarithmic presentation 110.7 °C.available are ight evaluation limits, rate, Jimits that are not quite right. ‘ordinate value atthe extreme point, and (sample temperature, segment time or abscissa unit) 5.20, left), Peak Evaluation Evatiation Frame Mode: Automate Line vrapet Peak 112.02°C Peak vse 2828 mW nomralnd Let Lime Roh umt 11215" Heating Rate 2.00 "Grant Pea Tse" Rewut Mode Sampie Teme Ba 6 8 40 42 4 48 48 Evatation Fore Moce: Marcel EWapol Peak 135.97 °C Peak Vale -3.78 ni Peak yasate Sample DMI with 11 mols SA. 8.7480 ma 20” 22mm ges 520 Lat Th ‘orvorinfad h 17 mol sale oot (38), omic Lis mee delerines the slopes ong in he working for of fro ama. Rig In tho Nanval Lis ode (igh, langrts cro crown (al shown He) aay a ho ome Imi, hich of oouse influences the vou oe exoglaies fens AIRough the ogra ‘Sepujed us 0 fureton ct te, sample tearchee oppecr ce ‘nuts becouse the Raeut Mode "ae ssn appropri elsod. The diagram stows he DSO meng peck ef dy fophibalts (CNT) Applications ‘Thermal Analysis in Practice ddravn to the measurement curve at the evaluation limits. Their point of intersection is the extrapolated sand abscissa value atthe extreme point (peak value and peak). These results are relatively insensitive to © mode: the evaluation limits define the range for the automatic search routine ta locate the steepest Page 475.11.6 Tables In certain cases, one wants to display values from a curve in the form of «table. Some important tables are ‘© enthalpy-temperature functions, © cp temperature functions, and ‘© conversion as @ funetion of time or temperature. ‘Temperature functions are usually shown with the same temperature increments. With conversion curves, a equidistant conversion increments is prefered, Finally, acconding to ASTM D3850, decomposition temperatures at 1 30, $0, and 75% conversion are calculated. ‘Table optons veces sina stm 09950 ‘uidstant cuidstant Table Teble Tale c “ % cm c Conversion curva 30000 Tozze03 0.00 25855 1000 41605, asad on te 32500 018 $0go 41605 2000 «2083 ‘tesnal 350.00 ot 2000 420.33 30000 43812 ‘decomposition Braco 122 Sooo 49812 S000 44976 efEA SiO Ki 400.00 434 4000 446.43 7500 46156 sco tse2 $000 44078 asoco So4s 60.00 45481, 47500 gags 70.00 45942 S000 100.02 8000 44.8 9000 47133 ‘0 a Figus 8.21. Lt The conversion curve coculed fom 0 TGA cune of Ine themal desorption of PRG. Rig: The comasion dls cspiaves mumercay in hee aes rable. 5.11.7 Minimum-Maximum Minfax displays the minimum and maximum ordinate values of an entire curve or that within the selected we range (Figure 5.4). The abscissa values are also displayed, Optional results: evaluation limits, 5.11.8 Signal Value Signal Value displays the ordinate value of a curve at any desited point on the curve (Figure 5.4). Optional rsu abscissa value corresponding to the ordinate valueral Measurement Methodology LUstAL CooRnINATE SHSTENS OF DIAGRAMS... 1 Ascisa 2 Ordinate: DEATOSPIENE THE MEAS CE non 621 low Rateand Purity ofthe Amospher ec nsaenmssnnon $22 How are Low Oxygen Conditions Achieved? 525 Commonly Used Parge Gases. G24 Reduced Pressure and Overpressure ‘CRUCILS IN THEROAL ANILIS.. 631 Contact between the Sample and the Atmosphere of the Measuring Cell 56 OveRvnew oF Tua EFFECTS CALIBRATION AND ANJUSTMENT 651 Some Definitions 652 Purpose of Calihration 65.5 Requlrements for Reference Substances... 654 Propenties Requiring Callration in Ther Analy... 655 Procedures in STAR" 656 FlexCal™ ENCES AND FURTEER REABING, 1 Usual Coordinate Systems of Diagrams 1.1 Abscissa: analytical measurement data can be plotted against time, the temperature of the reference point or the sample ature, Each type of abscissa presentation has is advantages and disadvantages: ‘Time: suitable for mixed (dynamic and isothermal segments) and simple orientation (especially with inserted temperature program), The newest values are always tothe right ofthe older data. 1 only makes sense to overlay ‘curves recorded with the same temperature program. In this respect, comparison of the first and second measurement runs is often very informative Te: Temperature isthe most important thermoanalytical physical quantity; curves measured using different temperature programs are always correctly overlaid, With cooling segments, the (time) display is from right to left Isothermal segments practcaly disappear (the measured values are plotted vertically over the temperature). Measurement curves with just one dynamic segment look the same as display proportional to time (7 is proportional to time), Tr: One suight think that the sample temperature is the best type of display because the sample temperature is usually of interest. However, the display of measurement curves during a first order transition is distorted (7/.is not proportional to time). METTLER TOLEDO Collected Applications ‘Thermal Analysis in Practice Paye 49LS we 68 00 62 64 66 68 70 72 74 76min ‘igue 6.1, Roove Two deren cresenaios of the som crstatizgon cue of wate: mecsue el & ‘cooing ro of SAinin, The curve (Ole) potsd ogo the reference Yep 7. which is froponcnal oe, shoes th ual rslaton peok Howe. wn fh cuNe plied ogo he ete empertue 7. ti ‘natrmoneonc, fx example af -12C tere oe tyoe orice voles (Bloc). Balew: To pion this es, Me sample forpraise is dpioyed aso funcon oft (ed Give) 15°, tho wate Dega oarsolz. The esaizaon enh of he | 8 mg same Is not ‘kffcen! fo heat he sample cre ula fo O°, bul neneteess 10.7“ acre 6.1.2 Ordinate Possibilities for normalized presentation: = psc Normalized to sample mass: Ordinate in W/g,for curve comparison Normalize to rate: Ordinate in J/K (= heat capacity) as well as sample mass and rate: Ordinate in Jg 8 pt heat capacity), forthe corect comparison of curves meastred a diferent rates with respect to area (Figure 6 = TGA Normalized to sample mass: Ordinate in %, DIG in % per abscissa unit, that is, ®K for the correct eomparisor curves measured at heating rates # 0, or Ye/min for isothermal measurements *T™A Normalizad to the original length (thickness), ordinate in % for the comparison of curves. 1 derivative ofthe TMA curve, ordinate in % (or ppm) per abscissa unit, that is, 8/K or ppaa/K (the expansior coefficient) for the correct comparison of curves measured at heating rates # 0, of %/min for isothermal 8Sc Cun Entrlby Crarze ‘Gene itega) 900g wes Ce ‘as “The figure shows DSC curves measured at 2, 5 and 10 K/min. The peak areas appeac to be quite different ecause visually you integrate the heat flow with ‘espect to temperature. The STAR" software of course ‘integrates the curve correctly with respect to time using ro 09, tas ane Rte Normand cp Curves EEninbey Change (Gompecatire integra) ‘300g, 800 Wao oC Fgwe 6.2. Comsorson ef measurement cuves ef chem eacion mecsiod at sift hetng Division by the heating rate yields the specific heat capacity. [n this presentation, the areas are identical. Integration with respect to the abscissa is also possible using the STAR® software “Mathematical Integration” program. “TA Integration”: AH = fe, aT 62) AH = fod 61) 6.2 The Atmosphere in the Measuring Cell {n practically all thermoanalytical measurements, itis necessary to have a defined atmesphere in the sample chamber. In most cases, this is achieved by purging the measuring cell with a purge gas at a particular flow rate, The lmosphere can be cither inert, reactive or corrosive. no reaction with the sample or the crucible chemical reaction with the sample is expected, eg air, 0, NH (flammable!) citemical rection withthe sample is expected, risk of reactions with the crucible and parts ofthe ‘measuring cel, eg. HCI, Cl, $0, The measuring cell may suffer damage ss tight measuring cell can be s in the range to 10.MPa are © Inert © Reactive © Corrosive: Most measurements are performed at constant pressure (atmospheric pressure). A operated at reduced pressure (partial vacuum) or at over pressures. Such applic possible with the high-pressure DSC The measurement curve is influenced by the type of gas, the pressure and the Flow rate ofthe gas in the measuring cell [METTLER TOLEDO Collected Applications ‘Thermal Analysis in Practice6.2.1 Flow Rate and Purity of the Atmosphere ‘The flow rate must of course be measured. This can be done using a flowmeter based on the rotameter priniple oF an electronic “mass flow” meter, This makes sure that the purge gas is flowing and prevents an excessively Large gas flow from blowing the sample out of the crucible or from cooling the measuring cell. ‘Typical flow rates are 20 to 100 mL/min. Flow rates in this range do not affect measurement as long as the flow remains constant ‘Thermogeavimetric measurements in particular are disturbed by flow rates that fluctuate, For example, pressure- reducing valve whose pressure slowly oscillates hetween two extreme values generates sinusoidal artifacts on the TGA curve, The pressure reducing valves used must therefore show no tendency to oscillate. 6.2.2 How are Low Oxygen Conditions Achieved? The rate at which residual ar is purged from the measuring cell depends on the flowrate. Exponential purging can be assumed if the cell is gas tight and if there is no dead volume (parts of the system separated from the sample chamber but not hermetically sealed, for example tubing or hore holes that arent purged) 63) exp -f 0 OP! AV Tar where cis the concentration; c; the initial concentration and ¢the purge time, V the volume being purged; AV/At the purge rate. Example: An air-filled furnace chamber of 50 mi volume is purged with nitrogen at 50 mL/min. How quickly does the ‘oxygen concentration decrease (eq = 20%): 1 [2 5 10min GI [27% 0.13% 0.0009% (9 ppm) | “This means that in the ideal case the oxygen concentration decreases to just afew ppm within 10 min, The situation is however adversely affected by ‘oxygen absorbed on pats ofthe measuring cel, ‘the existing oxygen concentration ofthe punge gas. Nitrogen of 99.99% purity can still contain up to 10 ppm ‘oxygen. This means that the 9 ppm obtained inthe calculation will never be achieved ‘© dead volumes fram which oxygen diffuses, ‘= small leaks, and ‘© Jong lengths of plastic tubing for gas supply (oxygen diffuses through plastic walls). Test for oxygen purity: ‘© TGA: Ater purging sufficiently long, maintain activated carbon isothermally at 700 °C. A maximum combustion rate of 10 min due to residual oxygen i reliable limiting value ‘© DSC: Heat several milligrams of unstabilized polyethylene (packaging film) in an open crucible at 10 K/min from 100 to 300 °C. Any oxygen present will give rise to an exothermic peak above 200 °C. A very desirable side-effect of the purge gas is that it protects the sensor and the measuring cell against corrosive decomposition products from the sample. Sensitive sensors in particular, such as a microbalance or the TMA measuring cell, require a separate supply of protective gas. This should aso flow between and after the measurements, Decomposition reactions with volatile reaction products procee differently depending on whether the volatile component i flushed away from the sample surface or remains in contact wlth the sample. In the latter case, the sample is almost in equilibrium with its decomposition products and « self-generated atmosphere is produced, Such conditions are most easily obtained using a hermetically sealed crucible with a pinhole (eg. 50 yrn) in the lid to reste diffusionCommonly Used Purge Gases phieres most frequently used in measuring cells are easionally also static air (stationary atmosphere). Air is often used for calibration. Since the main component of fpitrogen, its physical properties are very much the same as those of nitrogen. Air can be inert or reactive depending on the type of sample. It is inert toward most inorganic samples in the temperature range up to 300 °C, for example for the melting of indium or dehydration of calcium sulfate, In contrast, ai is reactive toward ‘malerials such as polyethylene, Furthermore, metals such as in or zine oxidize on melting in ait. This causes the citing peak to change noticeably in repeated| measurements. cases, ambient air can be used that is supplied using an squarium pump via a flowmeter. It then obviously ns a certain anount of moisture, In contest, “synthetic air” from a pressure bottle contains practically no water carton dioxide gen is used for measurements under oxyyen-free (actually low-oxygen) conditions. Purity requirements sum 10 ppm O:, Nitrogen isthe most frequently used “inert gas”. At high temperatures, nitrogen is however by no nee toward many metals (nitride formation) is used for the determination of the oxidation and combustion belavior, The purity requirements for oxygen lly not high, the cheapest quality is adequate for O1T measurements, is used as an inert purge gas for the TGA-MS combination if carbon monoxide is of interest. Nitrogen is fable in this case because it has the same molar mass (28 g/mol) hhas a much better thermal conductivity than the above gases. ‘This makes it interesting as a heat transfer am for TMA measurements and also for DSC measurements to reduce the signal time constant, Helium is also an ‘gas with no tendency to condense even below -180 °C. It is therefore often used for low temperature measurements. ‘thermal conductivity makes it difficult to reach temperatures above 1300 °C. on dioxide can be used for carboxylation reactions. on monoxide is not only flammable (see hydrogen) but also poisonous. The purge gas (and decompesition sof samples) must he trapped in cold traps or hy specific filters, For risk of explosion, see hydrogen. ized hydrogen is hydrogen that has been diluted to such an extent (for example with argon) that it cannot a explosive mistures with air, Argon can be obtained ready mixed with 4% hydrogen hy suppliers of compressed es. This minimizes the possible risk of an explosion. We strongly recommend that you do aot produce mixtures of and hyelrogen yourself by mixing the two guses on-line at corresponding flow rates. Applications: reactions in jucing atmospheres, for example to suppress the formation of oxide layers in dilztometric measurements, and for thermogravimetric reduction of metal oxides. hydrogen is very dangerous. When mixed with air it forms explosive mixtures over a wide range of acentrations. specialists with experience in the handling of flammable gases should work with hydrogen. This also applies to flammable or poisonous gases such as (ils, CO, Nj, pS, SO;). Additional requirements are ‘+ & gastight measuring cll, and ‘+ an automatic hydrogen detector close to the measuring cell, which sounds an alarm when a concentration of 0.1% I is reached in the laboratory air. [METTLER TOLEDO Collected Applications ‘Thermal Anaiysis in Practice Page $3The main dangers include for example 1. Leals in the supply tubing ¢rom the valve of the gas hotle tothe flowmeter), 2. Leaks at the connection to the measuring cell (lowmeter to connection nipple) 3. Leaks within the measuring cell leaky valve, nipple, weld seams, connecting holes, cracks, dirty O-rings, measuring cell not completely closed. 4. The formation of explosive mixtures inthe sample chamber. Explosive mistures can be prevented by evacuating t0 about 100 Pa, flooding with inet gas, e-evacuating to 100 Pa before and after filling with hydrogen. Instruments ‘vith a motor driven furnace opening should be protected against unintentional opening above 100 °C, for example ‘fault inthe electronics occus (risk of explesion following the ingress of oxygen!) 5. Purge gas outle, vacuum pump outlet (how to dispose ofthe hydrogen’). since hydrogen is avery small molecule, it diffuses through plastic tubing and plastic parts inthe gas switching device If all these safety measures are adhered to, itis possible to work with hydrogen. We have demonstrated this in application study ofthe hydrogenation of sunflower oil in the high-pressure DSC [1] METTLER TOLEDO however denies all responsibility if flammable, explost used. ,, corrosive or poisonous plirge gases & 6.2.4 Reduced Pressure and Overpressure In some cases, better separation of a physical transition from a chemical reaction can be achieved under reduc pressure (i.e. partial vacuum): For example, volatile plasticizers in plastics vaporize hefore the thermal decompositi Of the main components (at normal pressure both effects may possibly occur in the same temperature range Chapter 9, Thermogravimetrie Anais). High-pressure DSC: If an oxygen-free stmosphere is required, the information given shove under low-oxyt contlitions applies. In principle, the high-pressure DSC can also be operated at reduced pressure At the same time, the heat transfer fr the DSC sensor to the sample crucible is strongly pressure dependent below 1 kPa. This makes it difficult to interpret DDSG curves. Besides this, the diameter of the tubing is not optimal for vaccum operation; it can take several minute: reach the desired end pressure, The pressure meter should not be inthe suction line tothe vacuum pump but connec to a different connection in order to obtain realistic values. Page 54 Thermal Analysis in Practice [METTLER TOLEDO Collected Appliesa e 0 d Crucibles in Thermal Analysis serve as containers forthe samples during thermoanalyticsl measurements [2|. The type of crucible used for rement influences the results. A few considerations before the measurement often help to save time later ot juating the curve all DSC and TGA/SDTA experiments, the sample is measured in a crucible. crucible should be inert, that i, the crucible material should not react withthe sample inthe temperature range est Exceptions: The “copper stability” of polyolefines or lubricant greases and ols is determined in copper ls, and ectlain reactions are measured in catalytically active platinum crucibles. The crucible material should ‘undergo any physical transitions inthe temperature range used and the melting point must be sufficiently high. crucible protects the measuring cell against direct contact of the sample, which prevents contamination of the ‘measurement sensor or TGA/SDTA crucible holder. ‘ype of crucible (crucible shape, heat capacity) to some extent determines the specications ofthe measurement sich as For example the calorimetric sensitivity and the signal time constant, & short tinse constant yields DSC peaks and hence good resolution and separation of elose-lying effets ‘bles made of materials of high thermal conductivity and with a flat base enable optimum heat transfer between -sample holder and the sample with minimum temperature gradients, that remain solid over the entire temperature range ean be measured without using a crucible, for example, the temperatures of 4- to 5-mm disks of metals or ceramic materials. CG trbie mata) Cenc vourme >) mei ie eae peraurerange’) + Crucle Cem qu 6.3. Foctrs euancing he choo cre R TOLEDO Collected Applications Thermal Analysis in Practice Page $56.3.1 Contact between the Sample and the Atmosphere of the Measuring Cell Jn general, we distinguish between three different conditions: os (if necessary with auxiliary exteenal pressure in a high-pressure © Hermetically sealed, pressure-resistant crucibl DSC). #A“self- generated atmosphere” where volatile products from the sample remain in the rucible The sample i in equilibrium with its volatile preduets without a large increase in pressure occucring, The “diffusion bacrier” is achieved by piercing a very small hole in the crucible lid (crucible lids with a 50-uum hole are available). An even more effective diffusion barrier i obtained by scattering about 1 mg aluminum oxide powder on the rim the tandan! aluminum crucible before sealing (cold welding). This creates fine channels through which gas can diffuse Ina self-generated atmosphere, decomposition reactions are shifted to higher temperatures, resulting in better sepacation. The boiling point can also by easly determined in this way because premature evaporation is suppres 4+ Free access tothe fumace atmosphere in an open crucible: For practical reasons, ald with a ole is often used, f example to protect the measuring cell against samples that creep out or sputter out ofthe crucible. For this purpe the lid ofthe aluminum crucible is pierced several times with a needle on a clean rubber surface before sealing (1 diameter about 1 mm). te Intuorc of Gas Exchange 38 Sac suato al ccna, Ta catcum sufate cyerate, vn tn pinhole in 2oehmg a 00 120 140 100 180 00 e020 200 -C guia 64. DSC cures shawing he alminoton of woe of xsolizaton fom eas suse ahyeate (Grisum) measured nan open cube, Ina cuse wit mm hose i and in w ot generat fltossrete non open cual, the water vapor formod escapes ram the mansurng al, Ie he two messuremenis, ne wot var wma Tore oF fs Inside Iho cuctle. This cousss the ‘decompoetion ration ost ghar fempectues ond te suconsia loses of 14H, and 4 H:0 ‘0 Sopoeed naa goneckd anosprae, 6 tx example possible edema Smal mus of the cheatin he herby ey TEA [3 Page 56 Thermal Analysis in Practice [METTLER TOLEDO Collocted Applice‘Overview of Thermal Effects rement signal of «thermoanalstical curve changes more or less abruptly, itis referred to as a thermal eect sffects are caused by physical transitions or chemical reactions occurring in the sample. We distinguish peak-like and step-ike effets. The following tble summarizes the physical eects. For chemical reactions, see tion DSC, TGA. DSC (DTA) TGA TMA. t ake See oe a ads dumps iu is M ot dcptin pcs heel ran resto the sender Fock del sens tee prcare mar isenrny ‘ohanechang wong sah | ‘ Gosallztion ag 4 betas ital TOLEDO Collocted Applications Thermal Analysis in Proctice Page $7‘Thermal effect psc (D1) rea TMA Polymorphism Enantiotropic solid-solid transition of the low {temperature form Liguid-Liquid- solid transition of the metastable econ + 4 Fie ror te la ute as pert 5 atte neath heen extn Pp cxytalizes i mets 4 ao al oy (uur pass an samp : FT Glass transition th ny an ms TNO arene Bends 1,2 Data ding fi Sess sate campos a ese tinge ans Titaa mage Ferromagnetic Curie-tansiion Se es | apg ee Evaporation, 8 drying, = desorption, sublimation Tolan eee do ding Page $8 Thermal Analysis in Practice METTLER TOLEDO Collected Applicat(6g tg ojpnig uy sisdyouy jousoyL ‘swouauddy persion OG3104 peje eo (so) od Bajau po (Gow) paayou04) Lowy 6 AdoULe ALND A IK RUSUOURN 8 PR A= SHI NALGID JO paMNseeu SOARS 959 Sony su ou agua Ue 2 JO Sr ea Ts RUASIDD YG 8 8G CO OO 19800 pemjeuiou Pueeaai- — 1s8A rvaup, 99s sue Bune, 21991 WED iyi srae- peqewou ruueee- " (eBen aug 9g He se 3.085) fu coe np =1dweS rweroa vuwip, 002 seu Suyeor op unyumy Josey dues i 82- pacreuvou Zo 0 ag OSL 8 =| a ‘aygendaooeun st tp J ape suo are saunsnpy uouMA}sH; axp JO ays MHA aq MogH vorENZO;NY sapiHoud UO “swry sow ayqendaaoe ap uayptn aq ast uolDIqHID9 Jo esoding ‘anes aduarApas amp tuody paanseaLL ayp Jo AouAAYP (taMUNXEWI) ayqeidanoy SHU -suomeaqy7e9 05 pas sonuadand uno rp aomIsqNS Y :AaMASANS anes aouarayay aug (pion Some anya paunseaut ayy yex) os siayauNEsEd pans a SISTIpY a P24, pay ox st aunpooud si, anjes aouaajes& moyy ane posnseaut v Jo aLXAyIp a SALNLAAING suomuyed ewos juewysn{py pu uo}01q)}09 soe fous ashud voc sow ma (WL) os.6.5.3 Requirements for Reference Substances ‘The properties of the reference substances usedl must be known (reliable literature valles), Sources are. STAR® database for the reference substances suggested by METTLER TOLEDO Internet, eg. http:/Avebbook nist.gov The reference substance must not react with the erucible and/or the atmosphere. Unfortunately most metallic reference substances form a low melting eutectie with the aluminum pan when in intimate contact. Remedy: Do nat press the metal into the pan! + For heat flow calibrations there should be no change in heat capacity or only a very small change in heat capacity at the phase transition used © Practical information, such as stability (shelf life and thermal stability), toxieity, availability Certified reference samples are available from the Jaboratory ofthe Government Chemist, UK at hup:/Awww.lge.co.k or a the National Institute of Standards and Technology (NIST), USA at http://ts.nist.gov/sem METTLER TOLEDO markets the reference substances needed for calibration and adjustment. The following reference substances are traceable to the manufacturer: Substance r Order Number Indium In 1566°C ME 00 119 442 | Ta | sn 319°C ME 51 140 621 Teal Ph 327.5°C ME 00 650013, Zine Zn 419.656 ME 00 119 441 Aluminum Al 660.3 °C 397.0 Vg ME 51119 701 Gold Au 1064.2°C 87g ME 51 140.816 Palladium Pd 15s40°C [16202 —_—| ME51 140817 Table 62 Retsence subsonescwalabe tom METTLER TOLEDO, 6.5.4 Properties Requiring Calibration in Thermal Analysis The measured signal (ordinate) and the physical properties in conjunction with the abscissa of a diagram need to be calibrate, Ordinate © Heat flow, peak area (DSC) ‘© Mass (GA, automatically performed in the electronic microbabance) ‘© Length (displacement) and force (TMA and DMA). Abscissa © Temperatnres © yp which makes the temperature independent ofthe heating rate (at the sample crucible position) © Time (e.g. for isothermal measurements); since itis derived fom the quart clock of a microprocessor, itis extremely accurate. Page 60, Thermal Analysis in Practice METTLER TOLEDO Collected ApplicationsProcedures in STAR® at combinations ation of a measuring module, type of crucible and atmosphere can be used ta perform calibrations and The calibration parameters are stored in the database. Example methods are available inthe database for important standard combination (e.g. DSC, 40-p. Al crveble, air). IF DSC measurements are performed in a that has not been adjusted, the calorimetric results obtained are expected to be less accurate because the data of the standard combination is extrapolated using empirical factors, 1t Timits limits depend on the application, for example for the determination of onset temperatures, the ertor limits of ‘beat flow can be 10% o higher. Often, the goal of TGA is to determine the mass los step; the temperature range this occurs i less important. The error limits can for example be 5 K for such applications Jimits used in the ready-to-use check methods are based on the instrument specifications and are thus rather woe Imerval ‘ve suggest a calibration interval of once a month. If the results ane repeatediy within acceptable error limits, can be doubled, f several measurements with unacceptable results are obtained, the interval should he to half adjustment, one should always perform a calibration to verity that correct values are obtained. -io-use checks, for example the In and the Zo cheek for DSC, have the advantage of automatic evaluation the automatic validation ofthe results (pass or fal). The error limits used in these clue methods are based. iment specifications and are thus rather small methods can of couse also he used for this purpose ‘they show the measured results and it isthe user's “sbether to accept the results as they stand orto use them to adjust the instrument ifthe deviation is considered table [4] [5] ‘Spl: num #2800 mg be rece) 18133 Tomales “2864p ona e8re 2 mW] The D8¢ mec ie within specications! (2788, 28.05 ig, 156..1969°C) We ee se Te es Se Figure 6 6. Th enue of is DOC carton are win the permis recesony or ims. An ada not ons TOLEDO Collected Applications ‘Thetmal Analysis in Practice Page 61Adjustments The usuel way to perform adjustments i to vse the METTLER TOLEDO Total Calibration method. There are also specific ready-to-use calibration methods. Temperature calibration methods using one or several reference substances are available forall techniques. After a check or another measurement with a reference material, the data obta calibration menu of the Module Control Window to perform the adjustment ned can be entered manually into the Comment: Reference substances should be chosen in such a way that they cover the temperature range needed for ‘measurements. Extrapolation is less accurate than interpolation [4). 6.5.6 FlexCal™ Conventional thermal analysis instruments are adjusted for specific contitions only, Whenever the experimental conditions change, the instrument has to be readjusted. The STAR' FlexCal!™ system keeps the instrument properly adjusted under all conditions. For example for: All heating rates With FlexCal”, the temperature at the sample position becomes independent of the heating rate when tas has been properly adjusted. All types of crucible Details ofthe various crucibles with ther different geometry and mass are stored in the database. This information fs automatically taken into gecount All atmospheres Details of the different gases and their different heat conductivities are stored in the database. The information is automatically taken into account swe {ltraton ar Tom Act Incium (186 8 °C, 28.45.) TH@319°6, 80.1419) — — Integral 176 73 md nomalized 2876 1g%1 Onset 15650°C Hating Rate 20 "Cina ResuMede Sample Temp 201 tnoral 17808 ms fr | "romaized 28644 Onvet 166500 Heating Rata. “Grant Resutods Sample Temp Integral 178.63 ms nomalzed 2873 jg\1 Onset 169.97°C Healing Rate 10.99 "cman Rosutade Sample Teme Imtegral 375.47 ms rotralized 8015 1 Onset 23104 - Heating Rete 200 “Grin ResutMede Samale Temp Integral 37467 md hormatzed 60.02 1 Oneet 231886 Heating Rata 5.00 “Cmin-t Resut ode Sample Temp Iniggral 374.85 my malized 60.05 tt Onser23195-C Heating Rate 10.00 “Crit ReeultMode Sample Temp [ann nny SRERPT Senne eeneees renee eT io "v0 "140 "wo go ol Figwe 67. Ate austmont he DSC module ges cone! regurs fr tum and, Bln the enolpy of fuser (nomalze ia) ard te rofira poi! nat of gum (156.6 °) and ot (231.8 °C) ‘xg wi tou vals and cre dopant oe Nang vad (2 Sand 10 Km). Page «2 ‘Thermal Analysis in Practice METTLER TOLEDO Collected Applicationsne or al FEE SBA SOT ESTEE OT TOOT and Further Reading METTLER TOLEDO Collected Applications Handbook: “Food”, Melting Behavior and Hydrogenation, p. 38. METTLER TOLEDO Brochure M. Schubnell, Determination of calcium sulfate dihydrate and hemihydrate in cement, METTLER TOLEDO Thermal Analysis UserCom 26, 16-17. Calibration, METTLER TOLEDO Thermal Analysis UserCom 6, 1-5. Low-temperature calibration, METTLER TOLEDO Thermal Analysis UserCom 9, 14, ‘rucibles for Thermal Analysis Many of the application examples in this chapter have been taken from UsesCom, the METTLER TOLEDO jal customer joucnal that is published twice a year. Back issue of UserCom can he downloaded as PDES from the at al wwW.mn.com/ta-usercoms . {ER TOLEDO Collected Applicata ‘Thormat Analysis n Practice Page 037 Differential Scanning Calorimetry ZL INTRODUCTION ne a 65 7.2. DESIGN AND DSC MEASITREMENT “ vCIPLE =i i ws atest TLL How Is the Heat How Measured? E 68 722 How Is the Sample Temperature Seas? wide ae) 723 the Shape ofthe Melting and Crystallization Ped... : si 7 7.3. Saunne PREPSRATION, B 7.4 PERFORMING MEASUREMENTS pane = A. ssancaucasinn TS 7.4.1 The Porge Gas in DSC Measurements... a ee 742 Crucibles for DSC MessureMents 0 ache eames 743 Procedure with Unknown Samples a eae 76 7.5__ Ixmmmpwecine DSC CuunEs W 75.1. Interpreting Dynamic DSC Cures. : — 7 DSC: Cures Mat Show No Thermal Lets n DSC Cures That Show Thesmal Beets 0 7 Physical Transitions é 8 estng, Cstalization snd Sesophase Ta ae “#0 Transitions with Siguficant Los of Mass . 81 The Glas Transition 82 Lams Transitions sn 8 lemieal Resetions - 83 Wentying Arias “ 84 Interpreting Isothermal DSC Cures... i sb Physical Trastias.no : 87 (hemical reactions 7 Final Comuments on Interpreting DSC Curves, - 90 7.6 DSCEYALiAnioxs, 90 76 Characeristie Tenyeratures enna oe GD 7611 Onset an 91 THL2 Onset ith thesold Yalu 0 7013 Glas Transition os os 92 762. nthalpy Change by Integration ofthe DSC Gunn a - %% 7621 Baselines . = 95 7622 Content Determination : ” 71623 Determination o the Degree of Crystalliiy 2 " = 100 763. Conversion. ai a wi 764 lp sc dea 3 7465 Specific heat enpacity si Everees reese Heat capacity . 3 14 ‘The Specific Heat Capacity . = ise 104 7651 Glsing Sapphive 2 17 7.66 DSC Purity Determination ce 108 7.67 nth Onler Kinetics. Fes zs m0 KINETICS ecm se ea er il trot. _— E a . m Kinetic modeling in peactice th 112 767 Choining the Baseline and Lvaluainn Range col 5.672 Important Evaluation Settings 15 3673. Aeplicaions kinetic Daa 153 2 Proiction of Conversion asa Function of Reaction Tio 15 relition ofthe Reaction emperatore Needel to Reach a Parieular Conversion ina eran Tine eeeralll Samuating DSC Curves, 7 Isthermal Neasurements : ug Kinetis Acconing ta ASTM E698. ee eb 100 7469 Kinetics Secording to ASTM E1641 m1 746.10 Model Pree Kinetics, MEK m “640.1 aggletions of Mole Fre Kiet. : . moet us TAL NRSEA USHER ETHERS NSIT. ce age 64 Thermal Analysis In Practice [METTLER TOLEDO Collected Applicat!s9teq 9 o0id uy sishlouy youzouL ‘suojooiddy pev201109 043101 st ‘soanyeuxlo sous Sqvsoprsu00 ze sino90 sjuauodwoo ayisjos Jo uonzuodia ayy pu AfluEatuls sseasoUL suoRIOKaL snoauAONaIaU JO ayes atp ‘aunssaxd sapug “9¢jaunssoud-yty x wy ose apgeduuun'y 0 stiouestod yu 40 ansseud Japun sjuauounsvaKe pu onposd t jo saypreq wasayyp Jo uostredusoa (110) Auqqras toeprxo “uontsoduoaap aarepixo ‘surononas jromttiay jo stioneSysanu| Sages: ‘suoyavas jo auto 2yp ytoge suoyopad pure so2u1y uoNDeaL Jo uorERN soa aM ‘soydqexqua uonaeas “uoneatsacusjod 20 vonysodtwooap yeuoip se Yans suonatau yeoqwi2g> “uonwanausjodap pur sisjoaid “sqeparew snordzorr yo suorustren ssa an ‘wstyduoussjod pu suonsiren prjos—pijos “Suyjooorodns pur sojatyaq wonezy Tes “gots yo Adyerqua pure quod ua 2g ‘ue pamnseacu \nuenb ag ue stowjaeas peonuars pur suoRstNA) Pr! aw yey sossapord pw sonyadaud atnog “poumeuarap jaa “iapedeo way oyoads euyuerp pute Siayp sompo so yead aig azuapeseyo wip sameaduay auMAIAP “(soulyryqua woroeas pr wopisue) seam: yrad ean ‘agp otaYjOXD pue onWAYROpUD 2219p nod swoye 1-964 ‘Savsoye> Bunume>s eaUasaylp pajeo ose si anbyuysey ayy “uTyesedtL2gyuE,SUCD yep jot 40 ‘poyooo ‘parzay si AF uaysA afduews UF sINOIO JEU MOY WO oy] sUNseDKE JoFUUTTOTD SuTMUKDS peRUAZANP YOHOMPOMU] Ls ~ DNIAYIY MILLAN ANY SIONS sa 130 AnsmIstRS, ay OTL uauanseay 9SQ 1 JO MONN|ONSH AL COL yea BunjoH wimpy ap Koy sHapLEEIPY TELEdY yo UOT — TOLL Suyavetvay SanIs Ds Tk ONAN ¥ 0 OHMMEOHEL| ML XISISY HONTAAY OT oneagrg amending apni FKL suoneaqig aus EGE suo HHUA —~TEGL sutoursnlpy pa saneaqipey MOY EL agp a ONGC TL satmasaipe pay suotmesqq Tod 30 TEL ~ woesqyp9 oda srs uorgHPO TUL GL “INIKISTIOY ONY NOLIN. 6L) NSIAHANO NOWNONTOAWOSO SE suoneiisaul does €LL amssaig sapun sjueumamseayy 9s LL uy uow>npo] vonepXo) 44030 donENTERAT ML TLL SINARANISVWY JSG EAS AROS LL vonNjoNUCRy] | TTL aR SOHOUIY 94 PONY PROWEAY — TT'9L suing JSq 6 UOT FOT'SL um Mew o zanuoy pag Yoeay cn aon, oN AMA VOTING FOILOverview Zoomed aes of ‘las wanton ray : | srneromeion | 0 100 130 % % 7” e gue 7.1. & tpl! O80 ouve, Samp. 8 mg ofan ernie substan, heating fate 5 K/min. Lo Suey un fom 40 0 200° showing diferent efits fi: Tho ghae horsHon wth crires and tlscssa cele exzansion | a 7.2 Design and DSC Measurement Principle {In 1955, 8.1. Boersma introduced a quantitative DTA cell, which thereby led to the development of present-day heat flow Ds The current METILER TOLEDO heat flow DSC measuring cell with ceramic sensors exhibits the following features [1), ‘Very small furnace made of pure silver with electrical flat heater * P2100 temperature sensor with excellent long-term stability © Exchangeable FRSS and {1557 DSC sensors with star-shaped arrangement of thermocouples underneath the crucible positions that measures the difference between the two heat flows. Connection ofthe thermocouples in series results in high calorimetric sensitivity. Recess ground into the underside ofthe sensor disk provide the necessary thermal resistane. The thermal "esistance i very small and the heat capacity beneath the crucible is low because much ofthe material has been removal inthe grinding process. The resulting signal time. constant is therefore also very smal The disk-shaped sensor is connected vertically from below therehy minimizing horizontal temperature gradients ‘© Various cooling options (air cooling, citculator cryostat, IntraCoole, liquid nitrogen). © The same farace and DSC sensor is incorporated in «high-pressure DSC system, the HP DSC high-pressure DSC cell, usable up to 10 MPa [2] Page 66 Thermal Analysis in Practice METTLER TOLEDO Collected Applicationsw9steg oqo UI sieAlouy yowsouL ep aoa uy 0 oy 084 Bo LORE MIA arimodway eu fox. aus 4) peneoau st ues ‘8414016 HGH 94 uD 90 fd ay UO Moy 94 84 3, Inuowodod 5) ub s 1 pamsceu su ‘ous ey cu e500 sia a uno pu (4 ~ SOOR ar oH wy, sdreoowey pafioun App Guo) oss Ke "ABP cova coin so a, Sea cooUr ay 0B IE 33, at Bayo or pao 0 oy ney a A ueNoy Syd au, 2 Lamb oes dune ou uoRes FapuEGKS Cc NEL! i ES 4 4 | fe ne | / REE L i nas gga 4 9pUN 60 4 4 p99] B04) 208 ‘ind au p40 sum xfs ose 00 on ou sa BY ‘meq wes SOHO B09 8,“ OOU ay orp ea seme 44 #04 mo} a LI LIAL 9 Bazouuno sob li a. epouny 240 Sue 4) ayy oss ay sawn (omy) ue duo 3016 J x9 wy Yep 5¥e ay OU UND sass0M) sno yan Ase Aone ea coun 4 pu0 duos uy asus Ca ue yp Padre |= uns 330 0 ues sie ping 22 nb uouypen Huyjoo3, Sous Puls ogy 5194] yt st 5st ued omy unrayduey IL ae ‘suoqcotiddy pav2ati09 043101quo 7.4. The MuSTAR™ HS anc HSS7 DSC sensors 7.2.1 How Is the Heat Flow Measured? The heat flow, ®, flows radially through thermal resistance Ris, of the FRS5 and HSS? sensors. The thermal resistance is in the form of a ring under each of the two cruetble positions. as already mentioned, the temperature difference across this thermal resistance is measured by the radially arranged thermocouples From Ohm's law it follows thatthe heat flow on the left side (compose of the hea flow to the sample erueible and to the sample) is given by T-t aay 2) erence etween the two heat flows (73) Page 68 ‘Thetmal Analysis n Practice METTLER TOLEDO Collected Applicationsthermal resistances on the [eft and right sides aze identical due tothe symmetrical arangement, The same is true The equation forthe determination ofthe DSC signal ean therefore be simplified to ay the temperature differences are measured by thermocouples, we sill need the equation that defines the sensitivity thermocouple, $= V/AT, where Vis the thermoelectric voltage. From this, it follows that viv eee os V is the seasor signal, The product Ry, S is called the calorimetric sensitivity £ of the sensor. Ry, and S are ture depenilemt. The temperature dependence of A is described by means of a mathematic model, curves, a peak area for example, is the integral of the beat flow over time and corresponds to the change in ipy. AH, of the sample How Is the Sample Temperature Measured? 7.2 shows that the fumace temperature, 7., is measured using a PLLOO sensor. Basically, the Pt100 sensor is a ce made of platinum wire that hasan electrical resistance of 100.2 a 0 °C. "The relationship between resistance ‘emperature T is scribed by a polynomial: ReA+BT+CP (76) ‘DSC measurement, the heating rate selected refers to the reference temperature because the sample can undergo ‘order phase transitions during which the heating rate eannot be controlled ature difference, AT, that depends on the thermal resistance is necessary for heat to flow from the furnace to ‘erence crucible. In METTLER TOLEDO instruments, this is achieved by inereasing the furnace temperature same value of AT. Independent of the heating rate, the time difference between 7, and 7; is equal to the time , tay The heating rate, is equal tothe slope ofthe triangle shown in the expanded scale section in Figure Infollows that BRAT /tmy ot AT =f nw 2 the dynamic segment, the calculated temperature increase, AT, is added to the set value of the furnace ure s0 that the reference temperature exactly follows the temperature program. TOLEDO Collected Applications ‘Thermal Analysis In Practice Page 68a al ‘Temperature a Time Faure 7.8 The thee imporontlengeruhes at tumacetempertu 7, he lrenoe lampsctute 7 and ha sop larperiue 7. A tarnol act occu ¢sapia mateg a The enlarged sason ofthe dagram soa ne witonshp beeen te hecng (loo) 2h ng re sonst, x4, nd he toga odvorce, AT ‘The fumace temperature is increased hy AT atthe beginning ofthe dynaunic segment As shown in Figure 7.5, there is a difference between T, and T;, especially during thermal effec. 11 corresponds to the sensor signal in eq 7.1, which can be solved for 7: Tea T, + OR, = T+ OE ce ‘This is how the software calculates the sample temperature. Note: Strictly speaking, there is a small difference between the temperature within the sample and the measured temperature of the sample crucible, This difference is largely compensated through the right choice of he thermocouple sensitivity, 5, in the software Cotta aa Corot sient er 498.42m ‘Sample: num. 6.265 mg riommalzed 28 40.191 Sample hot Aluminu 43 ul One ys8sb"C Gas! air, S0mtinn Heating Rata. 2.00Gmin-t Intggre! 18837 md omalized 28.4591 Creat 5681" Hating Rate $09 "Crintt Intagral 1985 Tomales 28.441 Oneet 15885"C Healing Rate 10.00°Cmnint Lee aaa anna eared qe se iho he 1 ee wee Oe ae 68 We gue 7.6, OSC cues of he etn ot a pu mel (6 225 mg rium) mansiged of cieert hating ‘nae. The onsat rpc ar independant of heating rte. The onset iste tempera ef the ope fe ntsc Tha nga fre The ec “selina” wy he anger crown the sooo rm oak Page 70 ‘Thermal Analysis in Proctice METTLER TOLEDO Collected Applicationsra 3 The Shape of the Melting and Crystallization Peak us consider an imaginary experiment with a sample (a pure material, non-polymeric, not polymorphous), for le, indium, water, or dimethyl terephthalate ‘sothermal segment below the melting point of the sample is for temperature equilibration of J, and T,, We now ad mic heating segment. Because of the temperature advance (not shown for simplicity), 7; immediately increases iy wit time, 7; lags somewhat behind due tothe heat capacity ofthe sample. When the sample reaches its melting ture, Tj, the temperature remains constant until the sample has completely melted, At this point, no more iy of fusion has to be supplied and the temperature inereases rapidly until it once again lags slightly behind 7; short isothermal segment is included to achieve temperature equilibration. This is followed by 2 cooling in which 7, once again lags slightly behind 7; The sample shows 2 certain degree of supercooling and ouly to ersstallize below T;, The enthalpy of crystallization associated with this proces causes the sample temperature ‘eerease and in this particular case to reach the melting point. After complete crystallization, T. again lags slightly ind 7, ‘sensor signal is equal to T,— Tp. The shape of the melting peak is triangular and finally approaches the baseline ically. The crystallization peak resembles a trapezium that begins almost vertically and ends asymptotically Temperatures “Crystalization with supercooling FRSS Signals 7, - 7, exo Temperature (7) >" Time () Figue 7.7. Top: The upper agro shows the couse of 7, and 7, on hesing a sample with a ratieg point 7 (la, orc on cooing the some oa), The led sore ces nol stale on eoching a elg pot but axis suprcodkng. A Soon as ctysallaton boys, he Jeroerue Iereosts ona reacts ha main got he somple moss is slcisty xy. Wi sel samples, ely ot eysuzation St stile fo ea! he sample a he erie ,. Botan, The esig snc signals T.~ 7, ote plc accord wih ICTAC Uk, 7, ~ Ts poste er excel processes (xc) ond nage forename) presses » the sensor signal is converted to the DSC signal using eq 7.1. At the same time, the sign is set correctly (ICTAC: jc in the upward direction, or anti-ICTAC: endothermic upward). Although the calorimetric sensitivity is ure dependent, the appearance of the curve hardly changes over the small temperature region observe so We without another diagram with the DSC signal ‘The DSC curves in this handbook are displayed with exothermic changes the upward direction. DSC measurements are usually plotted against temperature rather than against time, If the reference ure Is chosen as the abscissa, the curve remains linear with time snd does not change in appearance, DSC ‘ae, however, distorted if they are ploted with respect to sample temperature ‘TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 71ot exo Cooling Cooling ee 0 ss => Heating t . + — 7, _ Reference temperature 7, Sample temperature Figue 7.8. Lt: Te two OSC curse ae pot oganst he tence fone, Oomarison wih Fgue 7.7 immedi stows tht {i sarpl emeetekucs on he dashes in must b deel, To stops ois nes ~1/R, gh. Te DSC caves re sled agoist amps tmpartra ese pak 80s donot eareepors io quarts of hes, Ctalizaon curs i pact ok rth uns In Figure 78, lines showing the same sample temperature have been draven. In the right diagram, the line is vertical, but not in the left diagram. ‘This has to do with the fact that the reference temperature continues to inerease during iscthermal melting of te sample, Let us assume that T, — 7; is just 1 K. In this case, according to eq7.1, Ru Ga the heat flow is —1 K/Rip. The slope of the line of constant sample temperature (the slope of the pure melting peak) is therefore -1/Re, An important evaluation procedure is derived from this line of constant sample temperature, namely the “extrapolated starting temperature”, more usually known as the “onset”. ‘This method evaluates sample temperature at the intersection ofthe tangent before the effect (i, the “baseline” and the tangent drawn to the peak. se Constant Temseatrs Dung Mating ‘Sonal Vaue 185 mW ‘Signal Valve 4.90 mW a 18860°C. a aa8'C Siena value 1122 mW | = 15868°C nat Valve 18.7 Py Signal Value. -20.69 mW Hone ye ae ay te e2a2-c Signal Value 36.54 mW a eas0"c 15:28mgin, 10 Kinin 9.999% puny 6.281mg venting, 10Kimin 98.9% purty wo wo ts tao 50160 Figs 79. The sorleempartug of the mating pack of dium renoins consort ol 186 8° (oath tle to treo ‘arparaure hereses fom 166. to 167-4 °C (rot vstle fom the gue). Only ‘atone pute ron pole motetels ait a consort (Sarg! tampertre curing meting. i freasng mouily ovo, fe ‘erpsanue oreases mete ond more diring malig. AS the vain ‘hing peak (le snows, a inseace ct 982% purty is akeaoy 025 K. Page 72 ‘Thotmal Analysis in Practice [METTLER TOLEDO Collected Applications3 Sample Preparation ple preparation is of utmost importance for achieving optimum measurement quality. Besides the right crucible, tion must also be given to i good thermal contact between the sample and crucible so that thermal effects are not smeared, ‘the prevention of contamination of the outer surfaces of the crucible either with sample or with its decomposition products, and ‘the influence ofthe atmosphere surrounding the sample Fgurs 710, (1); Standort 40-palurninur cube wy et bso sci. (2) 10 6): ‘te heretical sling tough cold welding (2) Liuid er powder saree fia (g20raty. (3: Puncio ou sample wi he fers focng cownwar. (4) Bod example, ‘arn requ 80 feng dowmard (6). Ongrally to lai fi hor curves une on Peutry. (6: Sompla wih dtomad I pressed dowrword 7) ond (8: iht aluminum cruciie win a sample of nega shape, or exnpla a so ir, tere sing (7) ant to salng (), Te ubbor punch adap hw i ete shape otha sap dl contact: thermal contact results in large temperature gradients in the sample. Hfects that are in reality sharp become | Small temperature gradients give sharp effects. This increases the repeatability of results and improves the tion of neighboring peaks. temperature gradients in the sample crucible are obtained by using samples of low mass and through good contact between the sample and crucible: disks, dense powers, and liquids are ideal arly shaped samples, for example plastic parts, are optimized by sawing and grinding flat the side in contact ‘the bottom of the crucible. ite substances are ground to a fine powder in x mortar, The powder is then added to the crucible using a funnel ‘compacted by means of a Teflon rod. Even paste-like samples can be pressed down into the crucible with a Teflon Deformation of the bottom of the crucible can be prevented by placing the crucible on a flat surface (if necessary a hole in it 10 accommodate the center pin). ples that have been punched out: any bures (ragged or rough edges) should be removed or the sample is placed the crucible with the burrs facing upward ids: Dip a spatula into the sample. The drop on its end is tansferred by touching the sample erucibe. natively a small syringe can be usa (be careful because plastic parts may be attacked by sobvents) «Ifthe fibers are suficently thick, they can be cut into small lengths that Hie flat in the crucible (covering with “ayer of heat conducting power is sometimes advantageous). in fibers can be packed ito a bundle in a small piece of degreased aluminum foil th: “Teflon rod. The packet is then placed in the crucible with the flat side facing downward, films that often curve upward during melting can be held down flat against the bottom ofthe tana num crucible using a light aluminum lid. ly exothermic reactions can he measured by mixing (diluting) the sample with relatively coarse alumniaum ‘or glass powder. This also facilitate the diffusion of gases from the sample, The substance used to dilute the Je must be dry and must not react with the sample: ‘samples sublime, the crucible volume should be aluminum crucible as a filler") then pressed lat using low as possible (use light aluminum crucibles or the lid of a ase ofthe crucible should always he lat, nat indented or bow-shape - otherwise heat transfer is poor TOLEDO Collected Applications Thermal Analysis in Practice Page 73External contamination of the crucible: The crucible must not be contaminated with any sample residues once it has been prepared! This could prevent proper operation of the sample changer. Furthermore, sample material should never come into direct contact with the DSC sensor. A contaminated sensor can produce artifacts (eifects due to the contaminant) and cause poor heat transfer Onganie sensor contamination can be removed by heat cleaning (at 600 °C for 10 min using air as purge ges). Water- soluble contamination ean be carefully removed using a moistened cotton bud followed by heat cleaning, Certain samples tend to creep up the walls of an open crucible during the measurement and contaminate the BSC sensor. This can be prevented by using a crucible lt with a hole), Influence of the atmosphere: «An open crucible without a lid allows the atmosphere of the measuring cell to come into contact withthe sample {free exchange of gas). The measurement is then performed under isobaric condition (i.e. under the practically constant pressure of the surrounding atmosphere). There is of course the danger that substances that creep out of the crucible or tha sputter ean damage the measuring cell, This can be prevented by covering the crucible with a lid with a hole, Restricted gas exchange (self-generated atmosphere) is necessary t determine the boiling point of liquid. 18 prevents the sample from prematurely evaporating The self-generated atmosphere is obtained by sealing the crucible using a lid with a small hole, The hole is made by placing the lid on a relatively hard surface, for example, the ‘erucible box, and piercing it with a shaep acedle, If possible, the diameter of the hole (20 t0 100 ym) should be examined under a microscope. To check whether in fact there isa hol, the lid can be held infront of a source of light. We recommend the use of lids with the pre-punched 50-y hole. If sample is hermetically sealed in the crucible, no work of expansion oceurs (e.g. endothermic evaporation). Since the sample is subjected to increasing pressure from its decomposition proslucts, the onset of decomposition shifts to higher temperatures. This isochoric type of measurement is limited by the pressure Tints of the crucible (aluminum standard crucible withstands about 200 kPa overpressure), The high-pressure crucibles have proven useful for such measurements 0 086 of Water Stand A Cucios ‘Sample! Water in sealed crucible, 3.3000 mg ‘Samal: Water $0 Um hola ie, 4100079 Onaat 8.89°C. Beitng goin 100 mw Sample: Water, 60 i 1.2400 mg ae dana eT oe em 7 BO oa tad ao. 18 gue 711. DSC ouves of wei. Above: In a harnsiccy seed rub, hes a no boing point The ‘ate boul 125°C 1s due oth eveibe bung. Middle: In 0 sel-geneted atrosphee Fe boing pit con be mecsued os the ose. Below nan open excl, wate woporles teote Me bong ait seco, ‘ Page 74 Thermal Analysis in Practice METTLER TOLEDO Collected Applications4 Pertorming Measurements 14.1 The Purge Gas in DSC Measurements protect the DSC measuring cell and to achieve good reproducibility, we recommend purging the cell using a gas flow ‘of about 50 mivinin. This applies to all meastiements. When open crucibles ar erncibles with « hole in the lid are the sample is exposed tothe atmosphere ofthe measuring cel. gen is inert in the temperature range up to about 600 °C and is therefore the standard atmos surements. Many measurements are performed in air because most samples do not react with the oxy 100 to 200 °C. Oxygen is normally used for studies involving oxidative behavior. jum is completely inert and has excellent thermal conductivity. This property lowers the D it's sometimes used instead of nitrogen to obtain better separation of close-Lying peaks. ime constant, which notes on atmosphere can be found in Chapter 6, General Measurement Methodology Crucibles for DSC Measurements | information on erucibles can be found in Chapter 6, General Measurement Methodology. Crucibles serve as containers and protect the DSC sensor against contact with the sample, They should siormally be inert with 10 the sample, that is, they should not react with it in any way, ‘most important crucibles used for DSC are: Aluminum crucibles. sluminum is largely inert. I is however attacked (issolved) by sodium hydroxide and ‘many acids. In some cases, metal samples can form low melting alloys with aluminum, Solution: heat the aluminum crucible at 400 °C for 10 minutes in air to enhance the protective oxide layer. ‘Under pressure, uluminum crucibles made of pure aluminum can be cotd-velded and hermetically sealed. “The METTLER TOLEDO standard 40-pL aluminum crucible with lid isthe crucible most often used. There are “also special lids that are pierced with a fine needle by the sample changer directly before measurement. This prevents ‘gas exchange occurring before analysis while the sample waits on the sample turntable. he light aluminum crucible fo films and powdered samples gives improved peak separation de to its very time constant, Liguid samples should not he measured with this crucible because they are often squeezed out ‘shen the lid is closed. -pressure erucibles are mostly used for safety investigations of chemicals and reaction mixtures. The tage ofthese crucibles i thatthe suuple remains completely inside the crucible and really does reach the tion temperature. In an aluminum crucible, it would evaporate (depending on the vapar pressure) and be swept ‘of the measuring cel by the purge gas without undergoing the reaction, ER TOLEDO offers pressure crucibles of different types. They are sealed by pressing or screwing the lid onto the ble. ‘made of platinum, gol, copper, sapphire or glass are also available for special purpases but are less frequently TOLEDO Collected Applications ‘Thormal Analysis In Practice Page 757.4.3 Procedure with Unknown Samples The temperature range you choose for the frst trial measurement is hased on any physical-chemical information ‘you have about the particular sample. The temperature range should be sufficiently large to make sure all the effets have been detected. Net too much time is lost at a heating rate of 20 K/min if the range first chosen turns out to be 100 K reaver than necessary. ‘The following information summarizes a number of basic rules when measuring an unknown sample for the first time Sample size and temperature range: Organic substances: 1 to 10 mg in aluminum cruci 20 °¢/min, mosphere Nz (“low rate about 50 mL/min) with pierced lids. Temperature range 25 to 350 °C, heating rate With inorganic substances, a sample mass of 10 to 30 mg is used and a higher final temperature, for example 600 °C. Mass loss: Its good practice to weigh the sample and crucible before the measurement, You ean then determine a possible loss of mass of the sample by back-weighing after the measurement. Losses up to about 30 pg can be attributed to the ‘evaporation of surface moisture from the crucible, whereas larger losses indicate the loss o! volatile substances from the sample, Examining the measured sample: © Does it look as if it has melted? Can you identify a melting peak in the DSC curve? If recrystallization is of interest, you can measure a new sample by adding a cooling segment at 10 K/min dircty after the melting peak. Do not forget tat samples can supercool by 1 to 50 X! Many substances often do not crystallize from the melt but instea form a glass. © Ist colored? Organi substances tien brown when they decompose Are any gas bubbles visible or ae there sign of foam formation? This indicates decomposition accompanied by a significant los of mass Has reaction occurred with the material of the crucible? If the crucible is not inet, it might dissolve or be 30 x), you would of course like to assign the mass loss particular peak. The effec: is usually endothermic because of the work of expansion due to gas formation. In this heat a new sample step by step across the individual pecks and weigh the cructhle helore and after each effect (at -R TOLEDO, this is called “offline thermogravimetry”). If x TGA instrument is available, you can measure a new if possible, in the same type of crucible used for the DSC measurement, -shape of the DSG curve of an effect is usually very characteristic and helps you to ide fy it. following sections discuss the mest important effets and describe the typical curve shapes obtained TOLEDO Collected Applications Page 777.5.1.3 Physical Transitions In principle, a physical transition can be measured as many times as desired provided that 1, The sample returns tothe original state on cooling. This is however not always the case and depends on the sample and the cooling rate. At high cooling rates, many substances solidify from the melt as amorphous glasses, which is ‘why no melting peak is observed inthe second heating run, Some metastable erstal modifications only erystallize ‘out in the presence of particular solvents. The sample does not escape from the crucible through evaporation, sublimation, or (chemical) decomposition, or does riot undergo a transition. Any sample lst by evaporation cannot crystallize in the sarap crucible on cooling because the purge gas has alrealy removed it from the measuring cel. 7.5.1.3.1 Melting, Crystallizotion and Mesophase Transitions Evaluation ofa melting peak yields the enthalpy of fusion and the melting point or melting range. The low-temperature side of the melting peak of a pure substance is practically a straight line (Figure 7.12a) and the melting point corresponds to the onset temperature. The low-temperature side of melting eurves of impure of polymeric samples with concave sides are characterized by their pesk temperatures (Figure 7,12b and c). Semicrystalline polymers exhibit particularly broad melting peaks due to the size distribution ofthe eystallite (Figure 7.12), Many organic compounds undergo decomposition on melting. The decomposition reaction can be exothermic or endothermic in nature (Figure 7.12d and e. An endothermic peak on DSC heating curve isa melting peak if ‘the sample mass does not significantly deerease during the measurement, Some substances, however, exhibit marked sublimation in the region of the melting temperature. The NSC curve is not affected by sublimation or evaporation if hermetically sealed crucibles are usd visual inspection after the measurement shows thatthe sample has melted. Powdered organic substances, in pasticular, form a pool on melting, On cooling, this solidifies to a glass (no exothermic crystallization peak) or crpstallies (exothermic crystallization peak) Note: The surface of many metals is covered with a high-melting point oxide layer. The oxide layer remains behind asa rigid envelope which is not deformed when the metal melts. On opening the crucible, the sample sil looks the same sit did before it melted. Noble metals form no oxide layer and produce spherical droplets on melting ‘the surface ara isin the range 10 to 400 J/g, The enthalpy of fusion of organi non-polymere substances is uswally between 120 and 170 Wg. the half-width is signiicantly less than 10 K. Semicrystalline polymers may in some cases melt aver a broader range. ‘The purer and lighter the sample, the narrower the peak. Very small and pure samples give peaks with half-widths of less than 1K. mpure samples, mixtures and blends often exhibit several peaks. Substances with eutectic impurities show two peaks (Gigure 7.12). The first is the eutectic peak, whose size inereases with increasing impurity, followed by the main ‘melting peak. Sometimes the eutectic Is amorphous, in which case the first peak is missing. Liquid-crstalline substances remain anisotropic even after the melting peak. The melt only becomes isotropic after one or more small sharp peaks due to mesophase transitions have oceusred (Figure 7.121. Page 78 Thermal Analysis in Practice [METTLER TOLEDO Collected Applications1 a b im d e f 9 FA «<—_ gue 7.12, Motng Feque 7.13. cyelizson ‘0: Nor polymer pure suostanoe 1 Bure stance (7; tha mating pi). be: Sampo wih outa: Imp, Separate droplets Soy iniduly vt trent ©: Sanlr/staiin pai {dys95 of suprcoting ating min decommnstion Mel aol orp 6 Moting wih decomposition. ‘Sorghum. ‘Fuqua ent, Shas cook mal ashlee heting cove the {oss tansiion erecta 20 eysalzaicn) f, Somicretotine clase, Lig en ‘exothermic peak on the cooling curve isa crystallization peak if the peak area is about the same as that ofthe melting peak. Since the enthalpy of fusion is temperature dependent, a deviation of up to 20% can however accur on erjstalization depending on how much the sample supercools. the degree of supercooling (Le. the difference between the onset temperatures af melting and crystallization) is Fpetween 1 and about 50 K. "The peak of a substance that erstallizes rapidly after nucleation has an almost vertical sie up until the point at winich the melting temperature is reached (provided the sample is lange enough, Figure 7.13a and g), the liquid phase is present as several individual drops, each drop supercools toa different extent and several peaks are (Figure 7.13b). R TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 79Organic and other compounds that crystallize poorly form a solid glass on cooling (Figure 7.130), Such amomphous samples can then. crystallize on heating to temperatures above the glass transition temperature. This process is known as cold erystallization. On further heating, several possible polymorphic transitions may occur before the solid phase just formed finally melts (Figuce 7.13e). ‘When the melt of sample contaminated with a eutectic impurity is cooled, the main component usually crystallizes out Figure 7.134) but can also solidify as a glass (Figure 7.13c). If the eutectic remains amorphous, no eutectic peak is observed. A polymer melt supezeools by about 30 K before rapid cooling (Figure 7.13e).. crystallizes (Figure 7.138). Many polymers solidify to form a glass on ‘When a liguid crystal melt is cooled, the mesophase transitions occur first (often without any supercooling!). The subsequent crystallization process exhibits the usual supercooling process (Figure 7.13g) 7.5.1.3.2 Solid-Solid Transitions and Polymorphism ‘The characteristic feature common to all solid-solid transitions is that a powder sample remains a powder even after the transition, ‘The monotropic solid-solid transition of metastable crystals (marked a’ in Figure 7.14) to the stable et-form is frequently observed with organic compounds and is exothermic (Figure 7.14a). As the name implies, monotropic ‘ransitions proceed in only one direction (they ae irreversible). Monotropic transitions are generally slow. They are fastest « few degrees K below the melting point of the metastable phase. In spite of this, the peak height is usually less than 0.5 mW and can easily be overlooked next to the melting peak that follows, which is about 10 mW (gray arrow in Figure 7. 14b). Often, iti best to wait for the monotropic transition to occur isothermally. At heating rates above 5 K/min, it is quite possible that the rate is too high to observe the slaw transition (Figure 7.14b) and the melting temperature of the metastable form i reached. The monotropic solid-solid transition is ten either invisible or it could be misinterpreted as a slight exothermic baseline shift before the melting pesk If a few stable crstals are present that can serve as nucleation pons for the crystallization of the liguid phase formed, the melting peak is immediately followed and by the exothermic crstallization peak. This case is referred to as transition via the liguld phase - on immediate cooling to room temperature the sample would have appearel to huve visibly melted. Finally, the melting temperature of the stable modification is reached. If no cenuclei are present, no ceenstallzation peak and of course no a-melting peak occurs (Figure 7.L4e). tn contrast, if the entire sample is present as the stable form, only the ex-melting peak appears and polymonphic eects are not observed (Figure 7.144). Depending on the substance, the cr'form melts at a temperature | to 40 K lower than the stable modification The enantiotropie solid-solid transition, which occurs less often, is reversible. The 2B transition starting from the Jow temperature form o. tothe high temperature form Bis endothermic, The enantiotropic transition gives rise to peaks of different shape depending on the particle sizeof the sample because the nucleation rate of each erstal is different. For statistical reasons, finely crystalline samples produce more or les bell-shaped (Gaussian) peaks (Figure 7.15a and ¢). A small number of large erjstals can give rise to peaks with very bizarre shapes (Figure 7.15b and d), especially with the reverse transition Ba. ‘The half-vidths of peas of enantiatropic transitions are typically about 10 K, Page 80 ‘Thermal Analysis n Proctice METTLER TOLEDO Collected Apalicationsous vas just out cis on ‘he vie ale Fle 7.1. Wore tons. ak 1: The crtow sions sositon of ho sald sols to "oon ahd Sm duo Sed sora so stow he fom ie ches sing gr een esi es ») «Bue fom ma a ue meth gum 7.15 Reval enostotepehostons le a: Fie pwd © cone tts % 6: Rose nel he Me poe A 4 Remreatanston one ce ays al ‘x and Savi rama en rim z 1.3.3 Transitions with Significant Loss of Mass a tions ofthis type can of course only he observed in an open crucible, that i, either without a lid or with a Hid with hoe to protet the measuring cell against substances that tend to creep out or sputer, The examples shown ae a ” Seep out or sp ‘evaporation of «Liquid sample (Figure 7.11, below and Figure 7.16a), * ‘deving (desorption of adsorbed moisture or solvent, Figure 7.16) limation of solids (Figure 7.16b), and the ¢ imposition of hydrates (or solvates) with the elimination of water of crystallization. In an open crucible, the of the curve corresponds to that shown in Figure 7.16b, and ina self-generated atmosphere to thal in Figure bc. speaks have a half-width of 220 K (excopt in a self-generated atmosphere); the pesk shape is similar to thet by chemical reactions. The decomposition of solvates is known us pseudo-polymonphism (probably because in cally sealed crucible, a new melting point occurs wien the sample melts in its own water of crystallization) ymorphism can also be thought of as a chemical reaction je-generated atmosphere (with a 50-pm hole inthe crucible lid), the evaporation of liguids is strongly impeded. sssually narrow boiling point peak is not observed until the boiling point is reached (Figure 7.11, middle and Figure TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 81The loss of mass common to these reactions causes the haseline to shift in the exothermic direction, This is a direct consequence ofthe decreased heat capacity of the remaining sample, bf Fg 716, Tenstins with mass os, Fgua 717. Sepwice arsine 9: Goss ronson, {Evaporation in on open cute be Desapton, sublimation, ©: Dehra, 6 Baling na. cruckle wih o small roe isla 7, he lie po cane ranstion 7.5.1.3.4 The Glass Transition ‘tthe glass transition, the specific heat of an amorphous material inereases by about 0.1 0 0.5 gk. AS a result, the DSC curve shows a characteristic shift in the endothermic direction (Figure 7.1, right and Figure 7.174). Typically, 4 the radius of curvatute at the onset is significantly greater than atthe endst, and ‘© the curve slopes in the endothermic direction hefore the transition but is almost horizontal afterward. Ifthe sample has been stored fora long time below the glass transition temperature, 7, an endothermic relaxation peak often occurs with an area of 1 to maximum 10g (Figure 7.17b). This peak is not observed on cooling (Figure 7.17¢) of when the sample is heated a second time immediately afterward. The glass transition region usually covers a temperature range of 10 to 30K. ‘You can identify an effect as a glass transition by checking whether the sample is noticeably soft, almost liquid or rubbery elastic above the 7. Ifa TMA or DMA is not available, you ean do this by heating a sample isothermally at a temperature of 7, + 20 K in a crucible without a lid. After several minutes, open the furnace lid and check whether the sample is oft by pressing it with a spatula or piece of wire. (is, however, difficult to detect the softening of highly filled plastics in this way. i ‘These second order solid-soid transitions exhibit A-shaped ¢p temperature functions, The most important ofthese isthe ferromagnetic Curie transition, which was previously used for TGA temperature calibration. The DSC effet is however very weak (Figure 7.17d). To make sure, you can check that there is no ferromagnetism above the Curie temperature using a small magnet, : -3.5 Lambda Transitions Page 82 ‘Thermal Analysis in Practice (METTLER TOLEDO Collected Applications1.4 Chemical Reactions ical reactions can generally only be messurel in the first heating run. On cooling to the start temperature, the products remain chemically stabo, so that on heating a second time no further reaction gecurs “In some cases, ‘action is incomplete after the first heating run and a weak reaction is observed in the second run (e.g. curing and wring of epoxy resins). ‘width at half height of a chemical reaction peak is about 10 to 70 K (usually about 50 K ata heating rate of 10 to in). dons that take place without significant loss of mass are ususlly exothermic (about 1 to 20,000 J, Figure 7.188 5). The others tend to be endothermic because the work of expaasion predominates DSC cures obtained from a chemical reaction show just one single smowth peak (Figure 7.182). In practice, the ofthe peak fs ften distorted by overapping reactions, suc asthe melting of aditnes (Figure 7.18), secondary 3, or decomposition reactions (Figure 7.180). Jes of reactions with significant loss of mass are al decomposition (pyrolysis in an inert gas atmosphere), frequently with gaseous pyrolysis products such as CO, chain alkanes, 1,0, and Ns, Jymeriation with move or less quantitative formation of the monomer, and ‘ondensation for example in the curing of phenolic and melamine resins * 's accompanied by a significant inerease in mass are nearly always reactions with oxygen and are strongly ic. Examiple ofthis are corrosion of metals such as ito, and ‘uptake of oxygen at the beginning of the oxidation of organic compounds, In the course ofthe reaction, volatile ton products such as carhoxylic acids, CO, and 1,0 are formed so that finally a los of mass occurs. The initial snorease can hest be observed in «TGA curve). of reactions with no significant change in mass are ion and polyaddition reactions, curing of epoxy resins, erization, dimerization, sngements, and vak ‘oxidation of organic samples (e.g, polyethylene) in hermetically sealed crucibles with just the approximately °, sug of residual oxygen from the air (Figure 7.184). ta the cure sae very few exceptions Wo this ule, One example isthe polymervation that occurs on heating sulfur at about 15D °C. Thi fevers to about 130°C. tly exothermie reactions ar offen measur in hig-pressuze cries in onde o suppress the endothermic vaporization peak of side-produets sons ar often performed in hermetically ell alumnus evethles wo prone the release of lightly volatile compo ts ‘ons TOLEDO Collected Applications ‘Thermal Analysis in Practice Page 83owe 7.16, Chemica weacns 6 deol cure shops fon exami ection >: Reason with meting pric! norton nd fs hegening of decompoaron. « Chemion easton wth secandy reaction Portal oxidaton of an oxuans sap veh resol oxgin ina hari sane cre 7.5.1.5 Identifying Artifacts ‘The curve should be first checked for any artifacts in order to eliminate the possibility of false interpretation, Artifacts are effects that are not diteetly caused by the sample, but have some other origin that does not relate to the sample Properties you want to measure. Some examples are shown schematically in Figure 7.19- a) Abrupt change ofthe hest transfer from the sample tothe crucible due to ~ irregularly shaped samples that topple over in the crucible or ~ plastic films that have not been pressed down flat on the bottom ofthe erucible. On heating forthe first time, they often twist and warp before they melt After meting, they make goed thermal coniaet again (Figure 7.20). Abrupt change ofthe heat transfer from the crucible to the DSC sensor: 1) Deformation of a hermetically sealed altminum crucible due tothe vapor pressure ofthe sample. 2) Slight shift of the postion of an aluminum crucible on the DSC sensor in a dynamic temperature program du to the different expansion coefficients of aluminum and the DSC sensor (Al ~24 ppm/K, DSC sensor ~9 ppm/K, see also Figure 7.20). The artifact is about 10 WW and is only visible at high ordinate scale expansion (ordinate scale <1 mW). This effect does not ocenr with platinum crucibles (~$ ppm/K) 3) The measuring cell suffers a mechanical shock: the erucible jumps around on the sensor and may move horizontally fit does not have a center locating pin. ‘The inflow of cold air into the measuring cell due to a poorly adjusted furnace lid leads to temperature fluctuations and generates a noisy signal Electrical influences: 1) Discharge of static electricity in a metal part ofthe system or power supply disturbances (spikes) 2) Radio transmitters, mobile phones and other sources of high frequency interference. ‘Thermal Analysis in Practice METTLER TOLEDO Collected Applicationsle ©) Sudden change in room temperature, for example through direct sunlight. 4) The crucible lid bursts due to an increase in vapor pressure ofthe sample, Depending on the amount of gas or vapor that escapes, an endothermic peak with a height of 0.1 to 100 mW can occur 2) Intermittent blockage (often periodic) ofthe hole inthe crucible Lid caused by droplets of sample that condense or foam. ) Contamination of the sensor due to residues of sample material from earlier experiments. Effects occur reproducibly al the same temperature and are characteristic of the substances involved. Solution: Cleaning by heating out in air or ‘oxygen. This type of artifact depends to a large extent on the sample material involved. eo endo Figura 7.19, 086 ats (etal ao given ins a) Allects con ete rape he ‘mecsurement using 0 rau sano th same subionce ans observing vee he sl! accu ‘gain af he someplace eat aaa lace one curve, Exceptions br anh, wih con be ey renrocuatle ‘can interfere with automatic evaluations (EvalMacro), especially evaluations that use automatic limits (artifacts that have been clearly identified as such can be removed from the measured curve using suitable so 0 race 121g PE Fim nested at 20 K/min PE Filmin ight luce Ui pressed down Us rim above fim {amg PS, fst heating at 10 Kin Gass Traction Onset Ba 13°C Mipint 05 °C a ata, ang Rate 10.00 mint Heght 7 ww eo 7 be io ao 8 Figue 7.20. nov: The dhe! ne stows th C86 cur of «PE fn hat had not men pressed down fim in the crcive, The att (ano) arises becouse he fis Was tos 0 "us! and warp win te ‘evel. Tre eon nv cuve shows ha “conecmeting ce. oboe om the some PE Aim afer had been crssed don on th Boe ef he cris using tid of aig auminum ube, Bet: DSC hotiny cave of 1.92 mg poytyens. Tre arc! ot oro" 78-°C ie coused by tho thermal ‘exgonson ee uminu cube TOLEDO Collected Applications Thermal Analysis in Practice Page 857.5.2 Interpreting Isothermal DSC Curves le AA [th | Fue 7.2) laoronnal piel ranesons «: Crystalzaten of o por, 0.9. poiypropyee coke lam he mt, Tins 130°C (aten wi a shoul), ya nen ofa pe mea © Enaniogl eve Narsion of a igh femporaue oh low taparture mostnsion (ho cystliznion of a melt eosin o separa crops of a pure sutstarce eeu ok ike Mis) «Evaporation oo sche au? 10K tay the balirg port ro custo who Tran oa ne i {x cansnt epee he erylizaton roto sutsanes hot rysaiaes wal ¢b) and he evaporation toe (6) eon proce conte une and oh 085s. Isothermal DSC is used inthe following fields © The study of exystallization processes including polymorphism, © Desomption, evaporation and drying ‘© Chemical reactions such as autoxidation, polymerization or thermal decomposition. Isothermal DSC measurement curves are usually easier to interpret than dynamic measurement curves An important advantage of isothermal measurements is the fact that an effect can be measured almost free of any interferences (other effects occur at other temperatures). Changes in the heat capacity of the sample are of course not detected unless quasi-isothetmal methods are use such as the isothermal step method [3], of temperature-modulated DSC (ADSC), in which the temperature fluctuates slightly round a mean value [4] or IsoStep® [5], Because changes in the heat capacity ofa sample are not visible in ordinary isothermal curves, baselines are exactly horizontal (except in the initial transition region fom dynamic to isothermal). All isothermal DSC curves approach 0 miW asymptotically at the end of the reaction. Strictly speaking, only the DSC furnace is isothermal, the sample however “isuperibolie” because it is coupled to the isothermal furnace via the thermal resistance of the DSC sensor. For example, if the thermal resistance is about 0.04 K mW", the sample temperature deviates by about 0.4 K from the temperature of the furnace at a heat flow of 10 mW. When the DSC signal reaches zero during the course ofthe effect studied, the sample temperature equals the furnace temperature Isothermal measurements are often performed rather differently to dynamic measurements so we will now summarize the main points Page 86 ‘Thermal Analysis in Practice [METTLER TOLEDO Collected Applicationsare two possible ways by means of which you can raise the temperature of the sample as quickly as possible to the temperature for isothermal measurements: Preheat the measuring cell for several ruinutes al the desired temperature. Insert the sample crucibles using the automatic sample robot, Using this method, the sample reaches the programmed isothermal temperature ‘eproduciby to within 0.1 K in about half a minute. This applies tothe light aluminum crucibles and the standard aluminum crucible. Equilibration and thus the transition peri take longer with heavier crucibles such as the high- pressure crucible, if sample robot is not available, you can introduce the crucible manually even more rapidly (with abit of practice). This manual method allows you to thermally pretreat the sample. For example, you can shock-cool a sample toa gla in liquid nitrogen and then allow it to crystallize isothermally inthe DSC. Altematively, you can fpremelt the sample at 200 °C and then allow ito crystallize in the DSC, for example at 130 °C. Inthe market Suppor laboratory, we use an old DSC20 measuring cell as avery accurate furnace for thermal pretreatment, Baise the temperature of the measuring cell with the sample already inserted tothe desired isothermal temperature ‘sing a dynamic segment. The advantage ofthis method is that almost any thermal history can be reprodi bly ‘preprogrammed (an advantage with routine measurements). The disadvantage is that it can take several mites to seach the isothermal temperature and to stabilize (i.e. the transition period is longer). This method is limited by the maximum possible heating and cooling rates of the measuring, cel. ‘want to evaluate the measured reaction fre of any interference in the transition region from the dynamic to the al transition region, you can correct the measurement curve by deconvolution. 4 hetter way is to subtract the of an inert sample of similar heat capacity measured using the same method (or a second measurement of the sample) 1 Physical Transitions examples of physical transitions are: mal crystallization below the melting point, for example in Figure 7.212 polypropylene at 130°C, or in 7.22 (above right) indium at 155.9 °C. Compared with dynamic cooling, larger crystallites with few defects are melting in Figure 7.22 above left. Using several isothermal steps, you can try to carefully approach the ture of thermodynamic equilibrium of the liquid and solid phase (the melting and crystallization rate is 0, ic. heat flow is 0) -rmal monotrapic transition below the melting point of the metastable modification. You can transform the completely tothe sable form, for example if you want to determine its enthalpy of fusion, srmal enantiotropic reverse transition below the equilibrium temperature, This provides an insight into Dizarre kinetic helavior of the sample (Figure 7.23) srmal evaporation (Figure 7.21d) below the boiling temperature or sublimation below the melting ture, volatile component can be completely removed to investigate the residue in a dynamic measurement :R TOLEDO Collected Applications ‘Thetmal Analysis in Practice Page 87sora eg and Crystazton DSC Cune of197mg indium robermal epson Inge Soh rom romatres 2691 oni Pook TS6.ET°C Isothermal meting Irtgral 558 82 md ormaiead 28 76g? Peak 1667°C 187.0 157 0158.51580 1859 gua 7.22. Above: Th nds sop os inetd bo he prteated measuring cl ot 187.0°C. The ‘simple aos: mmadatay ogi 10 mel. fs Men cooled 158.9°C o1 08 Karn, tsomem Crytollzelen tears ofer obo. 4 min. Below. Tne cure clevove fe samelefomcerous. Buse thamalrasionoe batweon fhe DSC sensor ond the Indium sample Is not execty reprodusbe, the macsures meling tempartue 80.06 « higher than he soieRcaon temperate, KOIO4 fe estas, 289°C, 6.89 mg K0IO4 coarse crystals, 299°C, 641 Ng eo nS mn Fgu 7.23 The cranforope revere naraion of he hgh aperture fem ef plasium perce ‘onsufed of 7K nec the equisbaum Tempera. The kines shown by Me lrge numba trytls (obove) 6 qute det fo he Kisioe ofthe smal number af ecorse esos (xl), Ih iu, the trust ysis nave an induction pared ef almost on hour, Sempes coining a ogo uber o ety fre esis yl a” clos emcor ell shaped cue basowe of te lege Number af individu crysis (toss. Page 88 Thermal Anolysis in Practice [METTLER TOLEDO Collected A1.2.2 Chemical Reactions ee schema ChonkclRapoons ‘Normal Reaction ‘Smale: Dibenzoylperxie at 130°C, 1.9520™me Reacion wit Induction Period ‘Sample Ethy acylate st 120°C, 72700 m9 gure 7.24, above: The ‘nama’ ccuse othe dacompestion maton of dibenzot parade dlesohad Inout piholle i an clurinu crue win o -ym hol In he bd, The TO highest tthe ‘aging of Re eocion when the consetoton of he ureodes saring retrial scrote Aaa Ing eoclen rol decreases axympiotcaly Yo 280. Blow: An example of a teacien wih an icin tod of aro then sever haus of 110°C. During te nduin petcd, noting appears o Dapper to Ine etl aoyine (i fos, 0 sobilzar Is consumes). Aer fis, ne polymerization reacton roy rwochos ho makimum a led “normal” chemical reactions begin immediately on reaching the reaction temperature. The reaction rate then ily decreases as the concentration of starting materials decreases (Figure 7.24, abave). Autoaccelerating, reactions ic or inhibited through the addition of stabilizes) first have am induction period (Figure 7.24, below) in othng spear open (he DSC siguiente than bout 1.1 mf). The eatin ae hen cress ively rapidly to its maximum value after which it decreases in the same way as in a “normal” al measurements are very useful forthe detection of autoacoelerating processes that are not so easy to control that are difficult to recognize using a dynamic temperature program. Preliminary isothermal measurements are iy performed al a temperature about 40 K below the onset of the dynamic measurement, ‘sothermal measurement of the OIT (Oxidation Induction Time) i often used to compare the oxidation stability of is or petroleum oils. The measurement is usually terminated on reaching « threshold value of about 5 mW only the induction time (ie, the onset) is of interest. The measurement ean also be performed under incceased for example at 3MPa in a high-pressure DSC, in order to prevent sample material evaporating at the rement temperature of around 208 °C. ting reaction resins are often cured isothermally and the glass temperature determined afterward. ally measured reaction peaks provide a valuable direct insight into reaction TOLEDO Collected Applications Thermal Analysis in Practice Page 897.5.3 Final Comments on Interpreting DSC Curves If you are uncertain about the interpretation of the DSC curve, you should consider the use of other thermoanalytical methods, These include for example: ‘© Thermogravimetric analysis, ideally in combination with DIA of SOTA. ‘* ‘Thermomechanical and Dynamic Mechenical Analysi ved Gas Analysis) using mass spectrometry (MS) or Fourier ® Analysis of evolved gaseous substances (EGA, transform infrared spectroscopy (FTIR). ‘© Observing the sample using the hot-stage microscope (TOA, Thermo-Optical Analysis in the FP82 of using the FP84 with simultaneous DSC). ‘© other chemical or physical investigations depending on the type of sample, If necessary, this ean be done by quickly cooling the sample and performing the analysis each time a thermal effect occurs. Note: DTA and SDTA curves are interpreted in a similar way to DSC curves. There are of course Limitations due the lower sensitivity. 7.6 DSC Evaluations ‘Most evaluations involve © the determination of characteristic temperatures and © the calculation of enthalpy changes (heat conversion). Enthalpy changes correspond to areas, for example, peak areas under the DSC cure caleulated by integration with respect to tie Further evaluations coacem the calculation of ‘© the conversion curve (conversion as a function of temperature or time), enthalpy (enthalpy 28 a function ofthe temperature or time), the specific heat capacity, purity, as well as etic data and predictions of the rates of chernical reactions. 7.6.1 Characteristic Temperatures ‘The temperature values assigned to DSC eflvcts are usually determined as the points of intersection of lines drawn to a curve (baselines and tangents), Various different standards define how to do this, — extrapolated peak mardi \ Serre ent First tangent / Baseline \onwst Endset i cs ah a ewe ees ‘wh shal monsters). The franco blween the pak emperce ond he exopsitod oak ferpsrtu can semelmes ba qua age, 7 ve pts rection are nl dr. Page 99 ‘Thermal Analysis in Practice METTLER TOLEDO Collected ApplicationsFigue 7.26. Charts ompertures of OSC slp, for empl ds oa glass tositon. ‘The eitranen tween re con an andthe midpart eon e qe ge deers an the shape othe cave. The midpoe i cin nde! wos. he STAY cxout eau ol ha los ranion Uses fre poof nersecon ol he DSC cava wh Ia bis lhe angl batwsen ‘eo oases, whie ASTMETB6 uses Ire enn voi othe onset cite enc ‘onset or extrapolated starting temperature is often evaluated, Its defined as the point of intersection of a baseline the thermal effect with a tangent. For pure non-polymecic materials, ths value corresponds to the melting point. Onset 15804 °C neat 16882°C | Heating rate: 10 Kin | Sample ¢-Actaropera 3.8900 mg | “ie 140 146 NSO SHB tes NNTS Figure 7.27, Te sample is plymaphic and exhibits wo DSC melting panks, To cuolole the Mo ring femperaus, both ees howe "a be dstrmind, To cose ne used 19 donne he sacora ‘hue! the ene used fo avout the onse fe rs peak and noone own dung the exothermic rezyatlizaion hat 2009 beeen fe ating penis, R TOLEDO Collected Applications ‘Thermal Analysis n Practice Puget7.6.1.2 Onset with Threshold Value sexo Onset wih Threshold Semple: Hestalen, 123000 mg Onset 1707 min Slope mwa Tiwesnols 5.00 mW Lettie 13 Rightumt 1848 min 20 Hosting Rate 000 "Cmin™t \nw ‘Type tresrole Resui Mode Segment Tine 20 Wo ae ao mn Figue 728, Th DSC cune of a pojethyene sample mecsued isohwiraly of 220°C n oxygen Che esting ond empertre elon perods ae nol shown. Aron induction patie, the examine ‘xan reason begins. Tare are slams dead fo: the onset conor fo Wich re nga must be craw oF o cso sgn ig Ins exam, the fveshold value chosen wos fae oreo Grd), Tho sul tina (17.07 mn) dows nok ages wi the abscissa (but 27 min) because tha Induction ‘ima le mencizod ram to tr when he croephat is esaned Yo 04/0, At the onset, the optional results available are ‘the slope of the tangent in m/s, the threshold selected and the time taken to reach the threshold, ‘the evaluation Limits sed, ‘© the heating rate inthe evaluation range, the onset type, that is a tangent atthe threshold value or atthe point of inflcetion (steepest part), and © the results mode segment time Relative thresholds given in % of the signal size ean also be used 7.6.1.3 Glass Transition Amorphous materials do not have a melting point hut exhibit a glass transition, An introduction to the theory of glass transitions can be found in Chapter 12, The Glass Transition At the glass transition, the DSC curve shos a step due to the change of ¢, of the sample. In addition, an endothermic relaxation peak ean occur with physically aged samples. ‘The sample is usually measured at a heating rate of 10 K/min. Other heating rates (also cooling rates) can be used. The results are to some extent however influenced by the heating rate used Page 92 ‘Thermal Analysis in Practice METTLER TOLEDO Collected Applications