You are on page 1of 360



Edited by

Bruce A. Moyer
Oak Ridge National Laboratory
Oak Ridge, Tennessee


Raj P. Singh
Osram Sylvania
Towanda. Pennsylvania

Kluwer Academic / Plenum Publishers

New York, Boston, Dordrecht, London, Moscow
Library of Congress Cataloging-in-Publication Data

Fundamentals and applications of anion separations/[ed ited by] Bruce A. Moyer and Raj
P. Singh.
p. em.
Includes bibliographical references and index.
ISBN 0-306-47911-7
I. Anion separation-Congresses. 2. Anions-Analysis-Congresses. I. Moyer, Bruce A.
II. Singh, Raj Pal. III. American Chemical Society Meeting (2001: Chicago, Ill.)

QD562 .A54F86 2003

541' .3722-dc22

Proceedings of an American Chemical Society (ACS) Symposium, held August 26-31, 2001, in Chicago , IL.
ISBN 0-306-47911-7
©2004 Kluwer Academic/Plenum Publishers, New York
233 Spring Street, New York, New York 10013

http://www.kluweronline .com

A C.I.P. record for this book is available from the Library of Congress
All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose of
being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Permissions for books published in Europe:
Permissions for books published in the United States of America: permissions@wkap. com
Printed in the United States of America

This book documents the proceedings of the symposium "Fundamentals and

Applications of Anion Separations" held during American Chemical Society National
Meeting in Chicago, Illinois, August 25-30 , 200 I. Nearly 40 papers devoted to
discussions on anion separation related to fundamental research and applications were
presented. The symposium, sponsored by Osram Sylvania, BetzDearbom, and the
Separation Science & Technology Subdivision of the Industrial & Engineering
Chemistry Division of the American Chemical Society was organized by Bruce A.
Moyer, Chemical Sciences Division , Oak Ridge Nat ional Laboratory, P.O. Box 2008 ,
Building. 4500S , Oak Ridge , TN 37831-6119, and Raj P. Singh , Chemicals and
Powders R&D, Osram Sylvania, Chemical and Metallurgical Products Division,
Towanda, PA 18848. It drew presenters from Australia, the Czech Republic, France,
Germany, Japan , South Africa , Thailand , the United Kingdom, and the United States .
Separations constitute an integral part of chemical industry. Chemical products
typically originate in resources that must be concentrated and purified, chemically
transformed, and subjected to fmal purification . Effluent streams from the processes
must be treated to recycle reusable components and to remove environmentally harmful
species. Some industrial processes are devoted to environmental cleanup after pollution
has occurred . In add ition, many analytical methods require a separation for
preconcentration, or a separation may be an inherent part of the analysis itself. Micro-
separations occurring at membranes or interfaces are also related phenomena employed for
ion sensing. Many species targeted for separation are naturally anionic. Although the
standard separations techniques of extraction, ion exchange, adsorption, precipitation, etc.
apply, the chemical approaches required must be tailored specifically to accommodate the
negative charge and electron-rich nature of anionic species. Hence, the topic of anion
separations occupies an important area of interest within the broader scope of separation
science and technology.
The subject of anion recognition and separation has in fact been growing rapidly in
recent years, with exciting discoveries at the fundamental level and economically
significant applications at the industrial level. This book, the first of its kind, covers the
breadth of knowledge being pursued at both ends of the basic-applied spectrum of anion
separation research and development. The need for the symposium and corresponding
book was anticipated in that basic discoveries in chemistry often encounter delay in
reaching the application stage, basic researchers are often unaware of practical


applications, and applied chemists are often unaware of basic advances. Thus, reflecting
the objectives of the symposium, the main objectives of the book are (I) to disseminate
the high-impact opportunities for practical applications that can motivate basic research in
anion separations , (2) to identify applications for exciting discoveries being made at basic
research level toward developing innovative products and processes, and (3) to target
fundamental research for understanding and improving selective separation of anions.
Chapters presented in the book focus primarily on chemical approaches for
manipulating selectivity in anion separations. In so doing, the chapters deal with the
following subtopics: design and synthesis of novel anion receptors, selective separation
and recovery of anions by ion exchange, selective separation and recovery of anions by
liquid-liquid extraction, anion-selective sorbents and inorganic ion exchangers,
environmental remediation and waste treatment, and separation and purification of anions
via precipitation and crystallization . The subject matter of the various chapters pertains
to applications in nuclear-waste remediation, hydrometallurgy, ion sensing,
separation/preparation, and new anion sieves for environmental and other applications.
By way of presentation, more fundamental papers are collected at the front of the book,
becoming more applied and technique-oriented in later chapters. Approximately a third
of the chapters deal with the design and use of anion receptors, compounds that display a
special affinity for anions and function by multi-point interactions. A basic
understanding of such compounds is only now emerging, and they are gradually
beginning to find their way into applications. Eventually, specific anion receptors have
the potential to enhance standard separation techniques such as liquid-liquid extraction,
and several chapters in fact deal with such enhancements. The lead chapter summarizes
progress in the design and synthesis of anion receptors for separations and sensing.
Following are two chapters discussing respectively physical and theoretical aspects of
anion binding. Three subsequent chapters describe different families of anion receptors
and their chemical behavior. To begin to shift the focus of the book toward applications
and methods, specific needs for development of methods for anion separations in the
U.S. Department of Energy complex are then described in Chapter 7. Chapters 8-10
present specific approaches to anion extraction using receptors to enhance affinity and
selectivity. Anion exchange, a work-horse separation technique for half a century, still
represents vastly fertile ground for selectivity enhancements, improvements, and novel
applications, as discussed in Chapters 11-13. Materials derived from biological sources
represent an attractive area of development from the "green" perspective. Biosorbents in
particular are the topic of the next three chapters. A subsequent chapter deals with a
novel concept in activated-carbon materials, one of the most versatile chemical platforms
for sorbtion of many types of chemical species. Finally, the book is capped off by two
chapters on progress in inorganic sorbents and a chapter on a problem in anion
Although a large number of reviews and books are published every year on
separations of cationic species, a book on anion separation containing chapters from the
leaders in the field is welcome. This collection of chapters can be used as a guide for
graduate students who desire exposure to recent research results in anion binding and
separation . The book will also be a useful tool to scientists and engineers involved in
research and development in both academic and industrial settings.

Bruce A. Moyer
Raj P. Singh
August 2003

We express our appreciation to the participants in the Chicago symposium. Our

special thanks are extended to the tutorial speakers, Spiro Alexandratos and Jonathan
Sessler, and to the many session chairs . We thank the authors for their hard work in
preparing the respective chapters of the book. Financial support for the symposium was
provided by Osram Sylvania, BetzDearbom, and the Separation Science & Technology
Subdivision of the Industrial & Engineering Chemistry Division of the American
Chemical Society and is gratefully acknowledged. Thanks are also due to Kathy
Hammerly (Osram Sylvania, Towanda), Deborah A. Smith , Judith H. Moyer, Becky
Maggard, and Deborah L. Goddard (Oak Ridge National Laboratory) for editorial and
secretarial assistance. Bruce A. Moyer was supported by the Division of Chemical
Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences, U .S .
Department of Energy , under contract DE-AC05-000R22725 with Oak Ridge National
Laboratory, managed and operated by UT-Battelle, LLC.



Binding, sensing, and assembly .

Philip A. Gale

4. ASSEMBLy ............ 21


From halides to nucleotides 31

Hans-Jorg Schneider

2.1. Ion Pairing 31
2.2. Salt Effects on Ion Pairing 34
2.3. Ion Pairing and Additional Lipophilic Effects 35
2.4. Hydrogen Bond-Based Anion Receptors 38
2.5. Simple or Highly Preorganized Receptors? 40



Design criteria for polyalcohol hosts 43

Benjamin P. Hay, David A. Dixon, Gregg 1. Lumett a, Rubiceli a Vargas , and

Jorge Garza

3.1. N0 3- Complexes 45
3.2. S042- Complexes 50


Suzanne L. Tobey and Eric V. Anslyn



Jonathan L. Sessler, Bruno Andrioletti, Pavel Anzenbacher, Jr., Chris Black,

Leah Eller, Hiroyuki Furuta , Karolina Jurs lkova , Hiromit su Maeda ,
Manuel Marquez, Toshihisa Mizuno , and Andrew Try

5.1. Quinoxpyrroles.......................................... ..................................... 80

5.2. Quinoxphyrins................................................... ............................. 79

5.3. Quinoxaline-Br idged Schiff-Base Porphyrinoids........................... 81



K. Travis Holman and Jerry L. Atwood



Gregg J. Lumetta

2.1. Tank Wastes 108
2.2. The Environment .... 109
3.1. Removing Problematic Anions from Tank Wastes 110
3.2. Removing Problematic Anions from Ground water 112



Joseph M. Mahoney, John P. Davis, and Bradley D. Smith

4. SUMMAR Y 122



Konstantinos Kavallieratos, Jeffrey C. Bryan, Richard A. Sachleben,

Gary J. Van Berkel, Oscar D. Espetia, Melissa A. Kelly, Andrew Danby,
Kristin Bowman-James, and Bruce A. Moyer

2.1 . Cesium Hosts 128
2.2. Nitrate Hosts 129



Holger Stephan, Karsten Gloe , Werner Kraus, Hartmut Spies, Bernd Johannsen.
Kathrin Wichmann, GUnter Reck, Dillip K. Chand, Pari mal K. Bharadwa] .
Ute MUller, Walter M. MUller, and Fritz Vogtle

2.1. Liquid-Liquid Extraction Experiments 153
2.2. Structural Considerations 158
4.1. Synthesis 164
4.2. Liquid-Liquid Extraction Procedure 165
4.3. X-ray Crystallography............................................................ ....... 165



Spiro D. Alexandrato s

4. SUMMARY 176



Grant C. Lukey and Jannie S. J. van Deventer

2.1. Cyanidat ion 181
2.2. The Resin-In-Pulp (RIP) Process 181
2.3. Chemi stry 182



Minoru Isshiki, Kouji Mimura, and Tamas Kekesi

2.1. Equilibrium Tests 209
2.2. Separation Experiments 211
3.1. Anion Exchange Equilibrium 213

3.2. Anion Exchange Separation Examples 215




Eric Guibal, Javier Guzman. Ricardo Navarro. Montserrat Ruiz, and Ana Sastre

2.1. Materials 227
2.2. Chitosan Modification 227
2.3. Methods 228
2.4. Distribution of Metal Ion Species .. 228
3.1. Effect of pH 229
3.2. Molybdenum Species 230
4.1. Sorpt ion Isotherms 233
4.2. Vanadium Species 233
5.1. Sorption Isotherms in HCI and H2S04 Media 235
5.2. Effect of Chitosan Modification 238



Zdenek Matejka, Helena Parschova, Pavia Ruszova, Ludek Jelinek,

Pavia Houserova, Eva Mistova, Milan Benes. and Martin Hruby

4.1. Synthetic Sorbents 252
4.2. Polysaccharide-Based Biopolymer Sorbents 252

6.1. Uptake on Sorbent with l-deoxy-rnethyl-amino-glucitol Moiety

Immobilized in StyreneIDVB Matrix 252
6.2. Uptake on Sorbent with Diethanolamine (DEA) Moiety
Immobilized in StyreneIDVB Matrix 254
6.3. Uptake on Crosslinked Chitosan (poly-D-glucosamine) Beads 256
6.4. Uptake on Crosslinked Bead Cellulose
(without Functionalization) 257
6.5. Uptake on Brown Algea Seaweed (Ascophyllum Nodosum) 258



Katsutoshi Inoue, Kedar Nath Ghimire, Rabindra Prasad Dhakal ,

Kenjiro Makino, and Tohru Miyajima

2.1. Materials 264
2.2. Methods 265
3.1. Batch Experiment 266



Jan D. Miller, Gustavo A. Munoz , and Saskia Duyvesteyn

5.1. SynthesisofMACs 283
5.2. Characterization 284



Masamichi Tsuji

2.1. Inorganic Solids and Modes of Anion Uptake 294
2.2. Method of Anion Uptake Evaluat ion 295
3.1. Crystal Structure and Ion Selectivity ............................................ 299
3.2. Design of Se032. Sieves 300



Eiji Kanezaki

3.1. NMR Analysis 314
3.2. Thermal Analysis 314
3.3. FTIR Spectroscopy ...................................... ................................. 316
3.4. Powder X-ray Diffraction (PXRD) 317
3.5. Regularity of Interlayer Ag+-Thiosulfate Complex 320


Raj P. Singh

1.1. Industrial Significance of Precip itation 325
1.2. Separation of Anions via Precipitation/Crystallization 325
1.3. Carbonate Precipitation 326


2.1. Sample Collection and Analysis 326
2.2. Calculationof Supersaturation of Water Samples with
Respect to Calcite 327
2.3. Precipitationof Calcium Carbonate from Groundwaters 327
2.4. Scanning Electron Microscopic (SEM) and Energy Dispersive
Spectrometric (EDS) Analyses 328
2.5. X-ray Diffraction Analysis 328
2.6. X-ray Fluorescence Analysis 329
8. SUMMARY 337


INDEX................................................................................................................... 343
Binding, sensing, and assembly

Philip A. Gale"


Recent developments in the area of anion recognition and sensing have produced a
variety of new selective receptors for anions . I However, the great variety of anionic
species and their importance in the environment (pollutant anions from over use of
agricultural fertilizers cause eutrophication of lakes and inland waterways'), biological
systems (mis-regulation of anion transport is responsible for a number of medical
conditions including cystic fibrosis') and in medicine (the maintenance of sulfate anion
concentration in dialysis patients continues to be problematic") presents a continuing
challenge to the suprarnolecular chemist to design selective receptors for specific guest
spec ies. This chapter will discuss the efforts of our group and our collaborators in the
areas of recognition and sensing of anionic species and of the roles of anions in assembly
processes in solut ion and in the solid state . As such , this review does not provide a
comprehensive overview of this area of supramolecular chemistry .'


Calix[4]pyrroles are polypyrrolic macrocycles cons isting of pyrrole rings linked via
the 2- and 5- positions by Sp3 hybridized carbon atoms functionalized with two alkyl or
aryl groups . Although these molecules were first synthesized in the latter 19th century by
Baeyer, ' their utility as anion-binding agents was not recognized until the mid-1990s
when Sessler and coworkers reported the formation of complexes of meso-
octamethylcalix[4]pyrrole 1 and meso-tetraspirocyclohexylcalix[4]pyrrole 2 with
fluoride. chloride and d ihydrogen phosphate anions." These receptors are
conformationally flexible in solution. In the solid state , the receptors adopt the so-called
' I ,3-alternate' conformation wherein adjacent rings are oriented in opposite directions.
However, in the presence of fluoride or chloride, the receptor crystallizes in a 'cone'

Philip A. Gale, Department of Chemistry, University of Southampton, Southampton. SO 17 1Bl. UK.


Fundamentals and Applications of AnionSeparations, edited by

Moyer and Singh, K1uwer Academic/Plenum Publishers, New York, 2004 1

conformation thereby donating four hydrogen bonds to the guest halide.t" This discovery
opened the door to a variety of pyrrole-based macrocyclic receptors and sensors for

(a) (b)

Figure 1. The crystal structures of the chloride (a) and fluoride (b) complexes of meso-
octamethylcalix[4]pyrrole 1 (counter cation, solvent. and most hydrogen atoms omitted for clarity) ,

In 1999 Sessler and coworkers" and Floriani and coworkers'? both reported the
synthesis of calix[4]pyrrole macrocycles (e.g.i S) from 4-hydroxyacetophenone and
pyrrole. This reaction produces a number of configurational isomers of compound 3 as
shown in Scheme 1. As noted by Floriani and coworkers, the aaaa-isomer of
compound 3 in which all the aryl groups are directed onto one face of the macrocycle,
resembles p-tert-butylcalix[4]arene (the crystal structure of which was elucidated in the
1970s by Ungaro and coworkers)" in that it contains a bucket-shaped cavity formed by
phenol rings (Figure 2). By using synthetic methodology first applied to calixarenes,"
we have synthesized a variety of 'super-extended cavity' calix[4]pyrroles and
investigated their anion-binding properties.

(al (bl

o 0 0 0
o o o o
Figure 2. The crystal structure of the (a) aaaa-isomer of 3 and (b) p-terl-butylcalix[4jarene. Both compounds
possess a bucket-like cavity composed of phenol rings. Hydrogen atoms omitted for clarity.



0 MeS03H

Ar,•. Me



Scheme 1.

Compound 4a, a tetraester functionalized calix[4]pyrrole, was synthesized in 77%

yield by reaction of the parent extended-cavity macrocycle aaaa-3 with
ethylbromoacetate in the presence of potassium carbonate in dry acetone with heating at
reflux for 5 days. Similarly, a tetraamide derivative 4b has been prepared in 50% yield
by reaction of a a a a-3 with 2-chloro-N.N-diethylacetamide and potassium
iodide/potassium carbonate in dry acetone with stirring for 5 days ." Crystals of
compound 4a suitable for X-ray crystallographic analysis, were grown from a DMSO
solution of the receptor. The structure (shown in Figure 3) reveals a DMSO solvent
molecule occupying the phenolic cavity of the receptor, hydrogen bonded to the
calix[4]pyrrole core.




4a R = CH2COOEt
RO 4b R = CH2CONEt2

0 11
()Q ,


Figure 3. The DMSO complex of compound 4a.

The interaction of 4a and b with anions was studied via 'H NMR titration
techniques. Perhaps surprisingly, in DMSO-d6 solution, no interaction (i.e., no change in
the lH NMR spectrum) was observed upon addition of chloride or dihydrogen phosphate
(added as tetrabutylammonium salts) to solutions of the receptors. However , upon
addition of fluoride, new peaks were seen to evolve in the 'H spectra of both compounds
4a and 4b (Figure 4). The new resonances correspond to the fluoride complex that is
slowly exchanging with the free receptor on the NMR timescale. The NH resonance of
the fluoride complex appears as a doublet. This is caused by coupling to the bound 19p
nucleus through the hydrogen-bonding interact ions, as confirmed by the 19p NMR
spectrum of the complex (JNH-F= 47Hz for both complexes; see Figure 5).

NH protons
in F complex

showing coupling
to 19F
J} Excess solid
TBAF added
to NMR tube

-M 10F"


k 1.9F

LL '2F
1 lL O.77P
11 11
ArH pyH
13.0 12.0 11.0 10.0 9.0 8.0 7.0 6.0 5.0

Figure 4. NMR spectra of co mpou nd 4b in deute riated DMSO solution on add ition of aliquots of
tetrabutylammonium fluor ide (TBAF). Proton resonances corresponding to the free receptor and fluor ide
complex can be seen in addition to coupling between the NH proton s and 19F nucleus. Reproduced with
permission from Chem.Commun . 2000. 1129-1130. Copyright 2000. The Royal Society of Chemistry .

86.8 86.4 86

Figure 5. '''FNMRspectrum of the fluoride complex of 4b showing coupling to four NH groups.

It is thus possible to use compounds 4a and b as NMR probes for fluoride anions in
the presence of large excesses of other anions such as chloride. Modeling studies on
analogous systems (e.g., compound 5) using Monte Carlo free-energy perturbation
simulations, Poisson calculations and adaptive umbrella WHAM protocol" suggest that
this selectivity is due to the presence of a small positive electrostatic pocket at the
calixpyrrole core . According to these results, the smaller fluoride sitting lower in the
cavity (Figure 6) than the larger chloride results in more favorable electrostatic
interactions with the smaller halide.

Figure 6. The X-ray crystal structureof the tetrabutylammonium fluoride complex of a tetra-acetyl substituted
extended-cavity calix(4)pyrrole 5.

The calixpyrroles are only one class of a much larger set of anion receptors
containing pyrrole. In the biological arena, it has been discovered recently that
prodigiosins (e.g., 6,7, and 8), a class of tripyrrolic pigment molecules produced by
microorganisms such as Streptom yces and Serratia, uncouple vacuolar-type ATPase
through promotion of H+/Cl" symport and therefore have potential as new pharmaceutical
agents. IS

OMe Me


Other synthetic pyrrole-containing anion receptors include the expanded

porphyrins," calixphyrins." and quinoxalines." There also exists a small subset of
receptors containing a pyrrole ring linked to an amide moiety in the 2-position.1 9 Many of
these systems are quite complex, for example, Sessler and Vogtle's elegant catenane for
oxo-anion complexation." Recently, Schmuck has shown that guanidinium groups that
contain an appended pyrrole-amide moiety are useful in the selective complexation of
amino acids." We decided to 'extract' the pyrrole-amide diad from these receptors and
study the anion-recognition properties of this unit alone.22•23 We therefore synthesized the
mono- and bis-arnide pyrroles, 9, 10, 11, and 12 and studied their hydrogen-bonding and
anion-recognition properties.f

Crystals of compounds 9 and 10 were obtained by slow evaporation of

dichloromethane/methanol and dichloromethane/acetonitrile solutions of the
8 P. A. GALE

receptors, respectively. Both compounds dimerize in the solid state via pyrrole NH-
amide 0 hydrogen-bonding interactions (Figure 7). Single crystals of the bis-amide
derivatives 11 and 12 were obtained by slow evaporation of CH2CliEtOH and
acetonitrile solutions of these receptors. In the case of compound 11, the crystal structure
shown in Figure 8.(a) reveals that one amide moiety has twisted out of the plane of the
pyrrole ring and is forming an amide-amide hydrogen-bonding network. Two
independent molecules of compound 11 in the solid state differ only in the conformation
of the butyl chains. The amide oxygen atoms each accept a second pyrrole NH hydrogen
bond to bridge the dimers into chains that extend along the c axis. Both independent
molecules in the crystal structure of 12 form centrosymmetric dimers (Figure 8.[b)) via
both N-H"""O hydrogen bonds, and C-H""O hydrogen bonds.

Figure 'I X-ray crystal structures of (a) compound 9 (dimerization via amide C=o-HN pyrrole hydrogen
bonds NI-OI, distance 2.766(6)A) and (b) compound 10 similarly showing dimerization in the solid state via
amide C=O-HN pyrrole (NI·OI , 2.875(5) A).



Figure 8. X-ray crystal structures of (a) compound 11 (hydrogen-bonding distances : N6-04, 2.982(4)A; N5-
02, 2.938(4)A ; N3-02, 3.060(4)A; N2-04 2.984(4)A) and (b) compound 12 showing dimerization via amide
C=O 'HN pyrrole and amide C=O"'HC aromatic hydrogen bonds (N2-02, 3.238(4); N5-04, 3.I27(4)A).


Figure 9. In the presence of DMSO, compound 12 crystallizes in a ' semi-cleft' conformation (NI-03,
2.831(4) ; N2-03. 2.757(4)A).

Interestingly, when compound 12 was crystallized from DMSO solution, a different

conformation was adopted by the receptor, wherein two hydrogen bonds are donated to a
bound DMSO guest (Figure 9).
The results of anion-binding studies conducted by IH NMR titration techniques are
presented in Table 1. Unfortunately, solubility problems precluded quantitative studies
being conducted with receptor 10. Binding studies were conducted in acetonitrile-d) for
compounds 9 and 11 ; however, solubility problems forced the use of DMSO/water
(0.5%) mixtures for the studies with compound 12. Direct comparisons may therefore
only be drawn between the association constants calculated for compounds 9 and 11.
Both the mono-amide 9 and the bis-amide systems 11 and 12 proved to be selective for
oxo-anions over halides . Both receptors 9 and 11 show selectivity for benzoate in
acetonitrile-d) solution, binding this anion with association constants of 202 and 2500 M-
I, respectively . Dihydrogen phosphate is also complexed by these receptors, albeit with
smaller association constants (89 and 357 M·I , respectively) . There is little
discrimination between the halides with compound 11 binding chloride most strongly (K.
= 138 M·I ) whilst compound 9 binds fluoride most strongly from this subgroup of guests
with an association constant of 134 M· I . Compound 12 in wet DMSO is selective for
dihydrogen phosphate (K. = 1450 M-I ) whilst also binding benzoate reasonably strongly
(K. = 560 M-I ) in this polar solvent medium. The difference in the benzoate and
dihydrogen phosphate association constants for compounds 9 and 11 suggests that a
different binding mode may be operating in each case and therefore suggests that
compound 11 is forming a cleft-like conformation (i.e., donating three hydrogen bonds)
when binding carboxylate in solution . Interestingly, similar bis-amide cleft anion

receptors synthesized by Crabtree and coworkers based upon an isoph thalic acid core
show sele ctivity for hal ide anion s over oxo -anions: however, these studies were
conducted in less polar solvent media."

Table 1. Association constants of 9, 11, and 12 (M·I ) with various anionic guest species
at 25°ca.

Compound Anion" Solvent Ka(M·1 )b

9 fluoride CD 3CN 134

9 chloride CD 3CN 28
9 bromide CD 3CN <10

9 dihydrogen phosphate CD 3CN 89

9 benzoate CD 3CN 202
11 fluorid e" CD 3CN 85
11 chloride' CD 3CN 138
11 bromide CD 3C N <10

11 dihydrogen phosphate CD 3C N 357

11 benzoate CD3C N 2500
12 fluoride DMSO/H20 0.5% 74
12 chloride DMSO/H2OO.5 % 11
12 bromide DMSO/HP 0.5% <10

12 dihydrogen phosphate DMSO/H2OO .5% 1450

12 benzoate DMSO/H2OO.5 % 560
•Anions were added as tetrabutylammonium salts dried under high vacuum with heating at 70 ' C
for 24 hours prior to use. Aceton itrile water content = 0.03% . Solub ility problem s prevented
studies on compound 12 to be conducted in dry acetonitrile solution.
bErrors estimated to be < 15%.
'The amo unt of water present in the acetonitrile can have a dramatic effect on fluoride/chloride
selectivity. In the presence of 0.5% water. fluoride is bound with an association constant of 37.5
M" I. whereas chloride is bound more weakly (K. = 12.5 M" I) .

A number of anion-binding agents are based upon organic fram eworks that are
designed to hold anion-binding groups at certain positions in space. One of main classes
of organic scaffolds used in this manner are the .caIixarenes , and a number of anion
receptors based upon these polyphenolic macrocycles have been synthes ized." Although
many effective and selective anion receptor s have been synt hesized based on these and
other organic scaffolds , the synthetic accessi bility of these systems can often be a
12 P. A. GALE

problem. In order to circumvent this, we decided, in collaboration with Professor

Stephen J. Loeb (University of Windsor), to seek a simple inorganic scaffold upon which
to arrange anion-coordinating groups," The [PtL41 2+ cation (L nicotinamide) is an ideal
candidate to act as an anion receptor as it provides both hydrogen-bond donating amides
and an electrostatic contribution from the platinum centre. As a first test of this strategy,
the complex [Pt(L)41 2+ (L = n-butyl-nicotinamide), 13, has been synthesized as the PF6'
salt and shown to act as a polydentate anion receptor both in solution and in the solid
state. The Pt(II) complex was prepared by reacting one equivalent of PtCliEtCNh with
four equivalents of ligand and two equivalents of AgPF 6 in MeCN solution." After
filtration to ' remove AgCI and evaporation of the solvent, crude 13[PF6h was
recrystallized from MeCN/E~O and isolated in 87% yield. Hexafluorophosphate is
commonly used as a counteranion for positively charged anion-receptor molecules, as it
is fairly "innocent" (i.e., it is a poor acceptor of hydrogen bonds). Crystals of the
hexafluorophosphate salt were obtained from a CH2CI2hPr20 solvent mixture . Adopting
caIixarene nomenclature, the complex adopts a '1 ,2-altemate' conformation in the solid
state, with the amide CO groups accepting hydrogen bonds from cocrystallized
dichloromethane (Figure 10).

o I

BU 1~

f _ N-Pt'~B=a
~ ", 2+ -


Bu 13

The anion complexation properties of 13 were studied using IH NMR titration

techniques (Table 2). Unfortunately, a variety of solvent conditions were required due to
solubility problems during the titrations . Thus, the association constants determined in
different solvent mixtures are not directly comparable. Interestingly, it was found that
tetrahedral or pseudo-tetrahedral anions such as Re04', CF 3S03' , and HS04' form
exclusively 1:1 receptor:anion complexes with compound 13, whilst 'flat' anions such as
acetate and nitrate form 1:2 receptor:anion complexes (a possible structure of a 1:2
complex is shown schematically in Figure II). Of particular note is the fact that acetate
anions are relatively strongly bound even in the very polar 9: I DMSO/acetonitrile solvent
mixture . The fact that K 2 is greater than K, infers that binding of the first anion has a
positive allosteric effect, which favours binding of the second. The I:1 receptor:anion
binding observed for the weakly coordinated Re0 4', CF 3S03' and HS0 4' anions is
supported by the X-ray structure of 13[Re04h. In particular, in order to maximize their
hydrogen bonding, the relatively acidic nicotinamide CHs from the ligands on the
opposite side of the square plane form hydrogen bonds to the perrhenate anion. To do
this, the complex must significantly distort from centrosymmetry, a fact that presumably

disfavours the interaction with a second anion, resulting in the observation of I:I binding
in solution.

Figure 10. The X-ray crystal structure of 132'[PF..h2(CH2Cl2) showing the basic numbering scheme. C-H"""O
distances (A) and angles ("): H(50A) "0(4), 2.40; C(50)-H(50A) 0 (4), 113; H(508)"""O(l), 2.43; C(50)-
H(508) 0 ( 1), 118; H(60A) 0(2), 2.42; C(60)-H(60A).. .0 (2), 115; H(608 )""0 (3), 2.47; C(60)·H(608)" 0 (3),
125. PtF distances (A) and angles (0): Pt(I ) "F(I ), 3.42; Pt(I )""F(I )-P(I ), 147; Pt(I )"F(8), 3.28; PI(I )"F(I )-
P( I), 155.

Table 2. Association constants, K. , for 132+ with various oxo-

Anion" Solvent K. (M· I )

CF3S03' CD3CN 129

Re04 ' CD3CN 150
N0 3' CD3CN KJ =562, K] =132
HS0 4 ' CD3CNIDMSO-d6 3:1 v/v 149
CH3COZ' CD3CNIDMSO-d6 3:1 v/v precipitate"
HZP04 ' CD3CNIDMSO-d6 1:9 v/v precipitate"
CH3CO Z' CD 3CN/DMSO-d6 1:9 v/v KJ =230, K] =491
• Anion added as the tetrabutylammonium salt.
b Precipitation occurred during the titration. The precipitate redissolved upon

further addition of anions: however, the titration profile could not be fitted

Moving on from anion receptors, the next section of this review will cover anion
sensor systems.
14 P. A. GALE

Figure 11. Possible binding modeof nitrateto 13l+form ing a 1:2receptor: anion complex.


The synthesis of new molecular devices designed to sense and report the presence of
a particular anion is an area of supramolecular analytical chemistry that is attracting
intense current interest." Two approaches have been used in the production of
calix[4]pyrrole based optical anion sensors. The first of these is based on the covalent
attachment of a colorimetric" or fluorescent" reporter group to the calix [4]pyrrole.
Perturbation of the electronic properties of the reporter group upon guest complexation
produces a visual or fluorescence response indicative of the presence of an anionic
species. The second approach is the use of a so-called "displacement assay,'?' This type
of sensor consists of a guest species (e.g., a colored anion) that is bound in a receptor
(e.g., a calix[4]pyrrole). Upon addition of the analyte anion, the original guest is
displaced from the receptor undergoing a change in its optical properties. We have
investigated this approach in collaboration with Dr. Lance Twyman (University of
Sheffleld)." Specifically, it was discovered that the 4-nitrophenolate anion 14 loses its
intense yellow color when bound to meso-octamethylcalix[4]pyrrole 1 (Figure 12).
Therefore, the calix[4]pyrrole-4-nitrophenolate complex 1'14 may be used as a
colorimetric displacement assay (Scheme 2). Competing anions, such as fluoride,
chloride, or dihydrogen phosphate displace the 4-nitrophenolate anion from the complex
so regenerating the original absorbance of 4-nitrophenolate observable as a colorless to
yellow color change. In fact, the intensity of the recovered yellow color may be used as
both a relative and absolute measure of the affinity of calix[4]pyrrole for the analyte
anion in that more strongly bound anions (e.g., fluoride) cause a greater recovery in the
yellow color than more weakly coordinating guests (e.g., bromide) (Figure 13).

anaIyte anion
Me (e.g. F·)



yellow colourless yellow

Scheme 2.

Electrochemical molecular recognition is also a fast expanding research area at the

interface of electrochemistry and supramolecular chemistry. " Here again, two main
strategies have been applied to the electrochemical detection of host-guest complex
formation; extraction of a charged guest into a membrane and detection of the resultant
membrane potential , or detection of a perturbation of the host's properties on complex
formation. Ion-selective electrodes comprised of calix[4]pyrroles enmeshed within a
PVC membrane have been used by Kral and coworkers as anion sensors." An alternative
route to the production of a modified electrode is by the production of an
electropolymerized film formed from a receptor bound to an electropolymerizable
moiety." a-Free pyrrole rings are known to form conducting electropolymerized films
via either potential cycling or chronoamperometry methods. We therefore synthesized
two calix[4]pyrroles 15 and 16 containing pendant pyrrole groups by condensation of the
relevant calix[4]pyrrole mono-acid with 3-aminopropylpyrrole using the BOP amide
coupling agent."

g li~ ,ncreasing caI,xpyrrole


j 0.5


400 450 500

wavelengl/1 (rvn)

Figure 12. Decreasing absorbance of the 4-nitrophenolate anion 14 (3.6 x IO"M) upon addition of
calix[4]pyrrole I (5 x 10" M) in dichloromethane at 25 'c. Reproduced with permission from Chem.Commun .
1999, 1851-1852. Copyright 1999, The Royal Society of Chemistry.


.e 0.6
~ 0.4


Figure 13. Relative absorbance of solutions containing 1 (5.0 x 10 M), 14 (6.0 x 10" M), and various anions
(1.6 x 10" M) in acetonitrile at 25 'c.
16 P. A. GALE

N h

Initial attempts at producing conducting electropolymerized films were made using

potential cycling methods. Unfortunately, this method produced a nonconducting coating
on the electrode, as judged from the fact that current waves prior to monomer oxidation
were not seen to increase with the number of cycles. Attempts at using controlled
potential electrolysis to generate the desired film also failed. Presumably, unfavorable
steric interactions, associated with N-substitution, are serving to preclude film growth as
has been observed previously."
In order to reduce any potential steric interactions, copolymerization of the
calixpyrrole monomers was attempted in the presence of pyrrole. Under these conditions,
the pyrrole subunits may act as spacers, increasing the distance between the
calix[4]pyrrole macrocycles in the film, and thereby allowing polymerization to occur
(Scheme 3). The cyclic volatmmograms (CVs) recorded during potential cycling in a
mixed solution of 15 + pyrrole and 16 + pyrrole are shown in Figures 15a and 15b,
respectively . These CVs are indicative of the in situ generation of a conducting polymer
matrix (an increasing current is observed for each consecutive scan).

calixpyrrole anionbinding sites

potential cycling
+ or chronoamperornatry



Scheme 3.

(a) - I 5.OJ!A

, , , , I , • , , , , I I , , , I ,

0.8 0.6 0.4 0.2 0 -02 · 0.4 -0.6 -0.8 ·1.0 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 · 1.0
Potential I (V) vs. Ar;iAg' Potential! (V) vs. Ag/Ag'

Figure 14. Consecutive cyclic voltammograms of (a) 15 (5 x 10"" M) and pyrrole (5.0 x 10"" M) and (b) 16 (5 x
10"" M) and pyrrole (5.0 x 10"" M) in CH,eN, scan rate 50 mvs' .

Consecutive cyclic voltammograms of (a) 15 (5 x 10,4 M) and pyrrole (5.0 x 10,4 M)

and (b) 16 (5 x 10-4 M) and pyrrole (5.0 x 10.4 M) in CH3CN , scan rate 50 mvs' .
To confirm that co-polymerization of pyrrole and calixpyrrole had indeed occurred, CVs
of the presumed co-polymer were compared to those obtained from a solution containing
an equivalent concentration of pyrrole. It was found that both a higher initial oxidation
current and a faster current increase with the number of cycles were observed in the case
of the mixed polymer system, results that are consistent with the participation of the
calix[4]pyrrole subunits in the polymerization process.
To investigate the use of this copolymer film in sensing anions, modified electrodes
coated with either simple polypyrrole or the copolymer were immersed in solutions of
supporting electrolyte (0.1 M) containing tetrabutylammonium salts of various anions at
concentrations at least an order of magnitude lower than the supporting electrolyte. It
was found that for polypyrrole, when the electrode was immersed in a solution of the
electrolyte containing tetrabutylammonium fluoride (5 xlO·4 M), the polymer redox
waves shifted continuously to more positive potentials with the number of cycles . The
same anion effect was observed in the copolymer films of 15-pyrrole and 16-pyrrole.
This effect masks any perturbation due to the formation of calix[4]pyrrole-anion
complexes on the cyclic voltammogram. Work is planned to study the co-polymer films
by other methods including spectroscopic, potentiometric and impedance techniques.
The second approach, that is, the production of discrete molecular redox-active
receptors for anions, has been the goal of a number of research groups world-wide.P'"
Early attempts at producing redox-active calixpyrroles by Sessler and coworkers"
involved the attachment of ferrocene amide moieties linked by alkyl chains to the
macrocycles. The perturbations of the redox properties of these first-generation receptors
(e.g., 17 and 18) upon addition of anions were rather difficult to predict, giving both
18 P. A. GALE

Fe 18


~Me H
+ 6° methanol
~ + ()



Figure 15.The X-ray crystal structure of compound 19.


.. . . . ~ l~~.- T~.~.~. r. .

38 . . .. .
. : . . :

3. ~~=r=·=~·= ~

002 003 0 04

CcnoentratlOn aI telraDut)18.mmon""m et11onC1e 1M)

Figure 16. (a) Shifts of the ferrocene CH protons of compound 19 upon addition of chloride anions to an
acetonitrilelDMSO 9 1 solutionof the receptorand (b) a schematic representation of an anioncomplex.
20 P. A. GALE

Table 3. Association constants and Fc/Fc+ redox potentials for compound 19 with
various anionic guest species.

Anion Ka (M· I)" £1/2 (FclFc) (mV) 6E(mV)

vs. Ag/AgClb
No anion n/a +444 n/a
Fluoride 3375 +368 -76
Chloride 3190 +408 -36
Bromide 50 +432 -12
Dihydrogen phosphate 304c +350 (approx.) - 100
Hydrogen sulfate d +436 < 10

'Errors estimated< 20%. Measured in acetonitrile·diDMSO·d6 9:1 (v/v).

b E I12values obtainedfromSWVs.
c The CH protonresonance was partiallyobscuredduringthis titration; therefore this value should be treated
with caution.
d No shiftswere observed.

cathodic and anodic shifts of the ferrocene/ferrocenium redox couple. In collaboration

with Prof. Jonathan Sessler (The University of Texas at Austin), we sought to involve the
ferrocene group in a more intimate role in the anion-recognition process and to synthesize
a calix[4]pyrrole in which there was at least a possibility of the anion forming a hydrogen
bond directly to the ferrocene group. This was achieved by cocondensing
acetylferrocene, cyclohexanone, and pyrrole to produce the monoferrocene
calix[4]pyrrole 19 in 30% yield (Scheme 4).38 The structure of this material was
confirmed by X-ray crystallographic analysis of single crystals obtained from acetonitrile
solution (Figure 16). As is usual in the cases of uncomplexed calix[ 4]pyrroles, the
macrocyc1e adopts a 1,3-alternate conformation in the solid state, with alternate rings
oriented in opposite directions.
Association constants with a variety of anions were determined by 'H NMR titration
techniques in acetonitrile-d/DMSO-d~ (9: 1) solution. As with previous calix[4]pyrrole
derivatives, this receptor forms complexes with fluoride, chloride , and dihydrogen
phosphate. Interestingly, in this solvent mixture the pyrrole NH resonance broadens
during the titration. However, association constants may be calculated by following the
shift of one of the ferrocene CH protons (Table 3). The large downfield shift of this
resonance may be indicative of the formation of a hydrogen bond between this proton and
the anion. The shifts of all the ferrocene protons during the titration with chloride are
shown in Figure 16.
The electrochemical properties of 19 in the absence and presence of anions were
studied by cyclic voltammetric and square-wave (SWV) voltammetric techniques.
Amongst the halides, the magnitude of the cathodic shift of the femocene/ferrocenium
couple increases from bromide to fluoride. Dihydrogen phosphate causes a large shift in
the calix[4]pyrrole oxidation processes that move them into the Fe/Fe" oxidation window.

o free recepto<
c nuoride
e cIlIoride
X dihydrogen phosphale

, !
, t " I. , ,' , I , t I

500 450 400 350 300 250 200

Polential (mV) vs AgJAgCI

Figure 17. Square-wave voltammograms of 19 (5 x 10.4 M) recorded in CH3CNIDMSO 9: I v/v containing 0.1
M tetrabutylammonium hexafluorophosphate using a PI disk working electrode,PI gauze counterelectrode, and
AgI AgCIreference electrode. Voltammetry carriedout undera dry argonatomosphere.

It is therefore not possible to accurately measure the Fe/Fe" oxidation potential in the
presence of H 2P04'; however, it is estimated to be approximately 100 mV less than the
free receptor. The SWVs of the free receptor, and the receptor in the presence of
fluoride, chloride and dihydrogen phosphate are shown in Figure 17.
Moving on from sensing, the final section of this review will cover recent work on
the assembly properties of anions.


The use of anions to direct the assembly of new supermolecules is a relatively new
approach to the organization of supramolecular architectures." However, anions alone
have long been known to assemble in the solid state." We decided to investigate the
assembly properties of tetrabutylammonium oxo-anion salts in the solid state , as these
materials are commonly used in anion-complexation studies. Crystals of
tetrabutylammonium hydrogen sulfate were examined by single-crystal X-ray diffraction
techniques. In this salt, the anions dimerize in a manner similar to carboxylic acids"
(Figure l8(a» .42 The X-ray crystal structure of the hydrogen sulfate dimer is shown in
Figure 19. The material cr ystallizes in a large monoclinic cell with four
tetrabutylammonium ions and four hydrogen sulfate ions in the asymmetric unit. Each
hydrogen sulfate ion donates and accepts a hydrogen bond to form the two dimers present
in the asymmetric unit.

o O" '-H-Q 2-
".,/ " ,...0
O's..."o-H'" 0/5,0

" · ' H--O." {'-H-~ ,•..O·'-H--O.. /.o- . _.~ ..O.'-H--O. p··· ·H-Q _
---0', e H'-- 0
0-1+ 0
.. ,. "~- e'
-1+ •• (f
" ,'o"Ft'eA e
~ ",0·0
0-1+ "0 "o-H' " j"o-I+'"

Figure 18. (a) Hydrogen sulfatedimerization and (b) a dihydrogen phosphate chain.

Figure 19. A hydrogen sulfate dimer in tetrabutylammonium hydrogen sulfate. Reproduced with permission
from Acta Cryst. 2001.E57.0705,0707. Copyright 200\.IUCr.


~ .. x.:y:.. .
0.,.;>-..... .... .... ~.....
J \......
. ·~~·..·'\.f·....~·
X ::;!:: -! -......J\........f'o--.....J\
·V ...... - r


Figure 20. (a) One repeating unit of the infinite hydrogen-bonded dihydrogen phosphate chain and (b) packing
diagram showing the sheath-like nature of the tetrabutylammonium ions around the dihydrogen phosphate
chains. Reproduced with permission from Acta Cryst. 2001, E57, 0727-07297. Copyright2001, IUCr.

We also wished to examine the assembly propert ies of a diprotonated anion. To this
end, tetrabutylammonium dihydrogen phosphate was recrystal1ized from
dichloromethane. This salt crystallizes in a large triclinic cell with six
tetrabutylarnrnonium ions, six dihydrogen phosphate ions and eight dichloromethane
molecules in the asymmetric unit." The dihydrogen phosphate anions form an infinite
coordination polymer (Figure 18.[b]) with six independent ions in the asymmetric unit
and a chain repeat length of twelve (Figure 20.[a]). The tetrabutylammonium ions form
cationic channels around the infinite dihydrogen phosphate chains (Figure 20.[b]).
These structures suggest the intriguing possibility of using oxo -anions alone to
produce structures of controlled length in the solid state. We are currently investigating
whether mixing different oxo-anions in different proportions allows one to 'tune' the
length of the oxo-anion oligomer formed.
In collaboration with Dr. Mark Ogden (Curtin University , Perth) and Professor Allan
White and Dr. Brian Skelton (UWA), we have been investigating the assembly properties
of amid inium substituted calix[4]arenes with carboxylates as a route to noncovalently
linked macrocycle synthesis. Amidinium containing species have been utilized in a
variety of supramolecular systems including crystal engineered tapes," self-replicating
molecules ," and ordered three -dimensional aggregates." Recently, Diederich and
coworkers have studied the recognition properties of bisamidinium clefts and
tetraamidin ium substituted cresorcinarenes." Amidinium substituted calix[4]arenes 20
and 21 were synthesized from bisnitrile derivatized calix[4]arenes using Garigipati's
24 P. A. GALE

reagent." The hexafluorophosphate salts of these receptors were prepared via a

metathesis reaction with AgPF6 • Proton NMR titration studies with biscarboxylate salts
in DMSO-d6 solution suggest the formation of lower-rim bridged complexes with 20 and
longer dicarboxylates such as adipate." Titrations of receptor 21 with isophthalate also
suggest the formation of bridged complexes (Figure 21), as the shifts of aromatic protons
of the benzamidinium group change direction at one equivalent of the added


Figure 21. A lower-rim bridged complex (i.e.• a noncovalently linked macrocycle) formed from 21'+ and

The X-ray crystal structure of the picrate salt of 20 has been elucidated. Interatomic
distances are consistent with intramolecular hydrogen bonds between one hydrogen atom
of each of the amidinium groups and the phenol 0 atoms (Figure 22). The twist of the
amidinium group-phenolic oxygen hydrogen-bonding network relative to the calixarene
ring gives rise to enantiomers related by a centre of inversion. A picrate anion is linked
to each of the two amidinium groups via a hydrogen-bond network, similar to that
observed in the picrate salt of
The X-ray crystal structure of the malonate salt of 21 has also been elucidated. In
this system, the amidinium groups do not form intramolecular hydrogen bonds. Instead,
the cation, anion, and a water molecule are linked by a complex network of hydrogen
bonds (Figure 23.[b]). There are three modes of binding of the malonate to the calixarene
(Figure 23.[a]). Mode (a) is the expected carboxylate-amidinium hydrogen bond
interaction. Mode (b) involves the dicarboxylate anion spanning the two amidinium

Figure 22. The crystal structure of the picrate salt of 20.

moieties, with a single N ... O hydrogen bond linking each pa ir of functional groups and
mode (c), with a single am idinium group spanning the dic arboxylate, is unusual and is
presumably made possible by the short chain length of the dicarboxylate anion .
Upon one occasion during the preparation of the hexatlurophosphate salt of 20, an
old batch (ca . three years old ) of silver hexatluorophosphate was used ." It was
subsequently found by 31p NMR that the silver hexatluorophosphate was contaminated
with ditluorophosphinate (PFzOz') anions, a known hydrolysis product of
hexatluorophosphate. The ditluorophosphinate salt of 20 spontaneously crystallized from

methanol-a; The structure reveal s a complex network of hydrogen bonds between
amidinium NH groups and ditluorophosphinate and F atoms (Figure 24 ). The
calixarene is mono-deprotonated to balance the charge. To the best of our knowledge,
this compound is the first occurrence of PFzOz' in a complex hydrogen-bonding network
and suggests that this anion may be a useful synthon in the production of anion-directed
assemblies in the solid state.


.... . . . . . .~
~.. ".. . g-o... ..... .
\--; .~

Figure 23. (a) The crystal structure of the malonate salt of 21 showing three complexation modes of malonate
and (b) a packing diagram of the same complex.

Figure 24. The hydrogen bond ing network present in the difluorophosphinate complex of 20 (methanol


The fields of anion recognition , sensing, and assembly are exciting and yet to be
fully explored areas of supramolecular chemistry that continue to attract the attention of
our group and others. We are looking forward to the development of new more selective
receptors and sensors in the coming years and the further development of the
understanding of the roles anions play in assembly processes. The development of new
applications for anion complexation in separation science, pharmaceuticals , and catalysis
continues to drive our efforts in this area.


I would like to thank my coworkers Salvatore Camiolo, Colin N. Warriner,

Christopher P. Chapman, Ellen R. Bleasdale, and Graham J. Tizzard and my
collaborators Dr. Jonathan W. Essex, Chris Woods, Professor Steve Loeb, Chantelle
Bondy , Professor Jonathan L. Sessler, Rebecca Zimmerman , Dr. Mark I. Ogden, Dr.
28 P. A. GALE

George Z. Chen, Professor Mike Hursthouse, Dr. Simon Coles, Dr. Mark E. Light (to
whom special thanks must go not only for crystallography but also for preparing the
crystallographic figures for this review), Dr. Lance 1. Twyman, Dr. Brian W. Skelton,
and Professor Allan H. White , without whom none of the work described in this chapter
would have come to fruition. Additionally, I would like to thank the Royal Society for a
University Research Fellowship, the EPSRC for a project and quota studentships, and the
Royal Society of Chemistry for an International Journals Grant.


1. P.A. Gale, Coord. Chem. Rev. 199, 181-233 (2000).; P.A. Gale, Coord . Chem . Rev. 213, 79-128 (2001); P.O.
Beer and P.A. Gale, Angew. Chem . Int. Ed. Engl. 40, 486-516 (200 1); F.P. Schmidtchen and M. Berger,
Chem . Rev . 97, 1609-1646 (1997); Atwood, 1.L., Holman, K.T. and Steed, J.W, Chem . Commun.I996,
1401-1407; Davis , A.P.; Gilmer, J.F.; Perry U .; Angew. Chern. Int. Ed. Engl . 35 , 1312-1315 (1996) ;
Supramolecular Chem istry of Anions, A. Bianchi , K. Bowman-James, and E. Garcfa-Espafia Eds.,
(WILEY-VCH, New York, 1997). Sessler, 1. L.; Allen, W. E. Chemtech, 29,16-24 (1999).
2. B. Moss, Chern. & Ind. 407-411 (1996).
3. Chloride Transport Coupling in Biological Membranes and Epithelia, G. Gerenscer,; Ed. (Elsevier,
Amsterdam 1984); T.F. Boat, MJ. Welsh and A.L. Beaudet, in The Metabolic Basis ofInherited Disease,
6'" ed., C.R. Scriver, A.L. Beaudet, W.S. Sly, and D. Valle, Eds. (McGraw-Hill, New York, 1989), p2649.
4. M. Brunetti, L. Terracina, M. Timio, P. Saronio, E. Capodicasa, J Nephrol. 14,27-31 (2001); H. Kajiyarna,
Y. Nojima, H. Mitsuhashi, K. Ueki, S. Tamura, T. Sekihara R. Wakarnatsu, S. Yano, and T. Naruse J. Am.
Soc. Nephrol .H, 923-7 (2000).
5. A. Baeyer, Ber. Dtsch. Chem. Ges. 19,2184-2185 (1896) .
6. P. A. Gale, J.L. Sessler, V. Kral, and V. Lynch, J. Am. Chem. Soc.H8, 5140-5141. (1996).
7. S. Camiolo, SJ. Coles, P.A. Gale, M.B. Hursthouse, and J.L. Sessler, Acta Cryst. E57, 0816-0818 (2001)
8. J.L. Sessler , and P.A. Gale, In The Porphyrin Handbook ; Kadish, K. M., Smith, K. M., Guilard , R., Eds.;
(Academic Press: San Diego, CA and Burlington, MA, 2(00); Vol. 6, p 257-278; H. Miyaji, W. Sato, and
J.L. Sessler, Angew. Chem.slntl. Ed. Engl. 39, 1777-1780 (2000); P. A. Gale, J.L. Sessler , W.E. Allen,
N.A. Tvermoes, and V. Lynch, Chern. Commun . 665-666 (1997); 1.L Sessler, A. Andrievsky , P.A. Gale,
and V. Lynch, Angew. Chern. Inti. Ed. Engl. 35, 2782-2785 (1996)
9. P. Anzenbacher Jr., K. Jursfkova, V.M. Lynch, P.A. Gale, and J.L. Sessler, J. Am . Chern. Soc. 121, 11020-
11021 (1999).
10. L Bonomo, E. Solari, G. Toraman, R. Scopelliti, M. Latronico, and C. Floriani, Chem . Commun. 2413-2414
II. G.D. Andreetti, R. Ungaro, and A. Pochini J. Chem. Soc. Chern. Commun ., 1005-1007, (1979).
12. A. Arduini, A. Pochini, S. Reverberi, and R. Ungaro, J. Chem . Soc. Chem . Commun. 981-982 (1984) ; F.
Arnaud-Neu, E.M. Collins, M. Deassy, G. Ferguson, SJ. Harris, B. Kaitner, AJ. Lough, M.A. McKervey,
E. Marques, B.L. Ruhl, MJ. Schwing-Weill, and E.M. Seward, J. Am. Chem . Soc. HI, 8681-8691 (1989).
13. S. Camiolo, and P.A. Gale, Chem. Commun. 1129·1130 (2000).
14. CJ. Woods, S. Camiolo, M.E. Light, SJ.Coles, M.B. Hursthouse, M.A. King, P.A. Gale and J.W. Essex, J.
Am. Chem. Soc. 124, 8644-8652 (2002).
IS. S. Ohkuma, T. Sato, M. Okamoto, H. Matsuya, K. Arai, T. Kataoka, K. Nagai, H.H. Wasserman, Biochem.
J. 334, 731-741. (1998). .
16. J.L. Sessler, MJ. Cyr, V. Lynch, E. McGhee, and J.A. lbers, J. Am. Chern. Soc. H2, 2810-2813 (1990); M.
Shionoya, H. Furuta, V. Lynch, A. Harriman, and J.L. Sessler, J. Am . Chem. Soc. 114, 5714-5722 (1992);
V. Kral, H. Furuta, K. Shreder, V. Lynch, and 1.L. Sessler, J. Am. Chern. Soc.H8, 1595-1607 (1996) ; V.
Kral, A. Andrie vsky, and J.L. Sessler, J. Arn. Chem . Soc. 117,2953-2954 (1995) ; J.L. Sessler , A.
Andrievsky, V. Kral, V. Lynch, J. Am. Chem. Soc. H9, 9385-9392 (1997).; H. Furuta, MJ. Cyr, and J.L.
Sessler, J. Am. Chem . Soc. H3, 6677-6678 (1991); J.L Sessler, V. Kral, 1.W. Genge, R.E. Thomas, and
B.L. Iverson, Anal. Chern. 70, 2516-2522 (1998); J.L. Sessler, J.W. Genge, V. Kral, and B.L. Iverson,
Supramol. Chern. 8, 45-52 (1996); B.L. Iverson, K. Shreder, V. Kral, and J.L. Sessler, J. Am. Chem. Soc .
H5, 11022-11023 (1993); J.L. Sessler, and SJ Weghorn, Expanded , Contracted and Isomeric Porphyrins;
(Elsevier: Oxford, 1997).

17. V. Kral, J.L. Sessler. R.S. Zimmerman, D. Seidel. V. Lynch. and B. Andrioletti, Angew. Chem.. Inr. Ed.
Engl. 39. 1055 (2000)•. C. Bucher. RS. Zimmerman, V. Lynch. V. Kral. and J.L. Sessler. J. Am. Chem .
Soc. 123.2099-2100(2001).
18. P. AnzenbacherJr. A.C. Try. K. Jursfkova, V.M. Lynch. M. Marquez. and J.L. Sessler. J. Am. Chem . Soc.
122. 10268-10272 (2000);. C.B. Black. B. Andrioletti, A.C. Try. C. Ruiperez, and J.L. Sessler. J. Am .
Chem. Soc. 121. 10438-10439 (1999).
19. M. Scherer. J.L. Sessler. A. Gebauer. and V. Lynch Chem . Commun. 85-86 (1998); M. Scherer. J.L.
Sessler. M. Moini, A. Gebauer.and V. Lynch. Chem. Eur. J. 4.152-158 (1998).
20. A. Andrievsky, F. Ahuis, J.L. Sessler. F. Vogtle, D. Gudat, M. Moini, JAm . Chem. Soc. 120. 9712-9713
21. C. Schmuck. Chem.Eur. J. 6. 709-718(2000); C. Schmuck. Chem. Commun .• 843-844 (1999).
22. P.A. Gale. S. Camiolo, C.P. Chapman. M.E. Light. M.B. Hursthouse, Tetrahedron Lett . 42. 5095-5097
23. P.A. Gale. S. Camiolo, GJ. Tizzard, c.r. Chapman. M.E. Light. SJ. Coles. and M.B. Hursthouse, J. Org.
Chem.• 66. 7849-7853(200I).
24. K. Kavallleratos, S.R de Gala. DJ . Austin.and R.H. Crabtree.J. Am. Chem. Soc. 119.2325-2326 (1997).
25. B.R. Cameron. SJ. Loeb. Chem. Commun . 573-574 (1997); P.D. Beer. M.G.B. Drew. C. Hazlewood. D.
Hesek. J. Hodacova, and S.E. Stokes. J. Chem. Soc. Chem. Commun . 229-231 (1993); I. Stibor, D.S.M.
Hafeed, P. Lhotak, J. Hodacova, J. Koca, and M. Cajan, Gazz. Chim. /tal. 127.673-685 (1997); N.
Pelizzi, A. Casnati, A. Friggeri, R. Ungaro. J. Chern. Soc. Perkin Trans. 2. 1307-1311 (1998); P.D. Beer.
P.A. Gale, and D. Hesek, Tetrahedron Lett., 36, 767·770 (1995).
26. C. Bondy,P.A. Gale. and SJ. Loeb, Chem Commun . 729-730(2001).
27. V. Yu Kukushkin, A. Oskarsson, and L.1. Elding,lnorg. Synth, 31, 279-284(1997).
28. P.D. Beer, P.A. Gale, and G.Z. Chen, J. Chem . Soc., Dalton Trans.• 1897-1909 (1999).
29. H. Miyaji, W. Sato, and J.L. Sessler, Angew . Chem. Int. Ed.• 39. 1777-1780(2000); H. Miyaji, W. Sato,
and J.L. Sessler, and V.M. Lynch. Tetrahedron Lett.• 41, 1369-1373(2000).
30. H. Miyaji, P. AnzenbacherJr, J.L. Sessler, E.R. Bleasdale, and P.A. Gale, Chem . Commun. 1723-1724
(1999); P. Anzenbacher Jr, K. Jursfkova, and J.L. Sessler. J. Am. Chem. Soc. 122,9340-9351 (2000).
31. A. Metzgerand E.V. Anslyn, Angew . Chem.lnt. Ed.• 37. 649-652. (1998).
32. P.A. Gale. LJ. Twyman. C.1.Handlin, and J.L. Sessler. Chem . Commun., 1851 -1852 (1999).
33. V. Kral,J.L. Sessler.T.V. Shishkanova, PA Gale. R Volf, J. Am. Chem. Soc.• 121. 8771-8775 (1999).
34. Z. Chen. P.A. Gale. and P.D. Beer. J. Electroanal. Chem.• 393. 113-117 (1995).
35. P.A. Gale. E.R. Bleasdale, and G.Z. Chen. Suprmol. Chem . 13.557-564 (2001).
36. S.R. Collinson, T. Gelbrich, M.B. Hursthouse and J.H.R. Tucker, Chem . Commun . 555-556 (200 1). H.
Plenio, and C. Aberle. Angew. Chem . Int. Ed. 37. 1397-1399 (1998); J.e. Medina, T.T. Goodnow, M.T.
Rojas,J.L. Atwood, B.C. Lynn. A.E. Kaifer, GW . Gokel. J. Am. Chem. Soc. 114. 10583-10595 (1992).
37. J.L. Sessler, A. Gebauer, and PA Gale. Gau. Chim.ltal.127, 723-726(1997).
38. P.A. Gale. M.B. Hursthouse, M.E. Light, J.L. Sessler. C.N. Warriner. and R.S. Zimmerman. Tetrahedron
Lett. 42,6759-6762 (2001 ).
39. B. Hasenknopf, J.-M. Lehn, B.O. Kneisel. G. Baum, and D. Fenske. Angew. Chem . Int. Ed. Eng135 . 1838
(1996); B. Hasenknopf, J.-M. Lehn, N. Boumediene, A. DupontGervais, A. VanDorsselaer, B. Kneisel,
and D. Fenske, J. Am. Chern. Soc. 119. 10956-10962 (1997); R. Vilar, D.M.P. Mingos, A.J.P. White. and
DJ. Williams, Angew. Chem . Int. Ed. Engl. 37, 1258-1261 (1998); J.S. Flemming, K.L.V. Mann. C.-A.
Carraz, E. Psillakis, J.c. Jeffery, J.A. McCleverty, and M.D. Ward, Angew . Chem .Tnt , Ed. Engl. 37. 1279-
1281(1998); R.Y.Wang, Z.P. Zheng, T.Z. Jin, and R.I. Staples. Angew. Chem . Int. Ed. Engl. 38, 1813-
40. For selectedexamples see: M. Malchus, and M. Jansen. Acta Cryst. B54, 494-502 (1998); P.H. Toma, M.P.
Kelley. T.B. Borchardt, S.R. Byrn, and B. Kahr, Chem. Mater . 6.1317-1 324 (1994); J.M. Karle, and I.L.
Karle, Acta Cryst .• C44, 1605-1608. (1988); N. Ohama, M. Machida, T. Nakamura. Y. Kunifuji, Y. Acta
Cryst., C43.962-964 (1987).
41. G. Bruno, and L. Randaccio, Acta Cryst. B36. 1711-1712 (1980).
42. M.E. Light , P.A. Gale, and M.B. Hursthouse, Acta Cr yst . E57 . 0705-0706 (2001); (see:
43. M.E. Light. S. Camiolo, P.A. Gale, and M.B. Hursthouse, Acta Cryst , E57 , 0727-0729 (2001) (see:
hltp:/Ijoumals.iucr.orgJelissues/200 1/08/00lbt6060/index.html)
44. M.W. Hosseini, R. Ruppert , P. Schaeffer, A. De CianoN. Kyritsaka J. Fischer. J. Chem. Soc.• Chem .
Commun .. 2135-2136(1994); O. Felix, M.W. Hosseini, A. De Cian and J. Fischer Chem . Commun. , 281-
282 (2000); O. Felix. M.W. Hosseini, A. De Cian, and J. Fischer, Angew Chem .• Int. Ed. Engl., 36. 102-

45. A. Terfort,andG. vonKiedrowski Angew. Chern. Int. Ed. Engl. 31, 654-656 (1992).
46. J. Yang,R. Melendez. SJ. Geib,and A.D. Hamilton, Struct. Chem., 10, 221-228 (1999).
47. L. Sebo, B. Schweizerand F. Diederich, F. He/v. Chim. Acta, 83, 80-92 (2000); L. Sebo, F. Diederich and
V. Gramlich, Helv. Chim. Acta, 83, 93·113 (2000); and references cited therein.
48. P.A. Gale, Tetrahedron Lett; 39, 3873-3876. (1998).
49. S. Camiolo, P.A. Gale, M.l. Ogden,B.W. Skelton, and A.H. White, J. Chern. Soc. Perkin Trans. 2, 1294-
50. H. Nagata, Y. In, K. Tomoo, M. Doi, T. Ishida, and A. Wakahara, Chern. Pharm. Bull.• 43, 1836-1843
51. S. Camiolo, P.A. Gale,M.E.Light,M.B. Hursthouse, Supramol. Chern. 13,613·618 (2001).
From halides to nucleotides

Hans-Jorg Schneider"


Anion recogn ition by artificial receptors has received increasing attention, whereas
the earl y focus in the development of host-guest ionophores was mainly on cation
binding. Several books and reviews are ava ilable on the top ic of anion complexation,'
which allows us to concentrate in the present review on mechanistic aspects of anion
reco gn ition as studied in our laboratory. Emphasis is on the quantitative evaluation of
noncovalent interactions, which can lead to a better under standing of anion complexation
in synthetic) and natural' systems. It is also hoped that such analyses can help the rational
design of efficient anion-binding agents of industrial, anal ytical, or environmental
importance. Systematic comparison of binding energies in geometrically well-defined
host- guest complexes can provide insight into the underlying mechanisms and can
provide reliable data for the prediction of affin ities in synthetic and natural systems.'


2.1. Ion Pairing

Binding of anions in aqueous media relies most often on the formation of salt
bridges, although other effects such as hydrogen bonding or van der Waals interactions
can also play an important role , in particular with respect to more selective associations.
Th e efficiency of artificial host compounds that provide man y binding sites towards
anions and thus a chelate effect is still best illustrated with the azacrown-triphosphate
complex published already in 1984,4 which reaches association constants around 1011
M- 1• In this complex (Figure I), 10 to 12 charge-charge interactions can materialize,
which agrees with the additive MG increments for Coulombic forces discussed below . In

• Hans-Jorg Schneider, Universitat des Saarlandes, 0-66041 Saarbrucken, Germany.

Fundamentals and Applications ofAnion Separations, edited by

Moyerand Singh,KluwerAcademiclPlenum Publishers, NewYork,2004 31

accordance with this observation, hydrogen bonding contributes little to anion binding in
water, which is a too strongly competitive solvent. Further evidence for the negligible
role of hydrogen bonding with open receptors in aqueous systems is the generally
observed similarity of complexation strength with either alkylated or protonated
ammonium ions.3•s

Figure 1. The efficiency of ion pairing. Complex between triphosphate and a protonated azacrown ether (see
Ref. 4) with K = 1011 M- J• Only 7 out of the possible 10 to 12 charge-charge bridges are shown by dashed lines.

The classical Bjerrum equation (1), as discussed in electrochemical textbooks.i

describes the association constant K as function of the ion charges ZA and ZB, the
dielectric constant E, and a factor Q(b), which depends on E, Z, and on the distance a
between cation and anion.


Differentiation of K by the temperature T on the basis of the Bjerrum equation, or of

the more simple Fuoss equation' (see below), suggests that in water the enthalpic binding
contribution should be smaller than the entropic factor, the latter being due to liberation
of water molecules from the ion hydration sphere. Thermodynamic studies indeed show
generally TtS is major driving force, however with large differences between different
systems," Although considerable progress was made with recent extensions of the
Poisson equation," quantitative predictions remain difficult, in particular for ions with
large anisotropies and ill-defined charge distribution and polarizibility. Nevertheless, for
a large number of mostly inorganic ions, quite linear correlations were found between log
K and the charge product ZAZB.9 This result is in line with the Fuoss equation" for contact
ion pairs:


Systematic comparison of many, mostly organic ion pairs in aqueous media, has
even shown a surprising additivity of log K or tJ.G values. Accordingly, if one compares
the experimental tJ.G value with the number n of possible charge-charge interactions in a
large series of ion pairs, one observes a linear correlation," although the underlying ions
differ very much in shape, in polarizibility, and in charge delocalization (Figure 2). The
additivity of tJ.tJ.G holds equally well for anions like halides, carboxylates, sulfonates,
phenolates, or phosphates, as for cations as in tetraalkylammonium, pyridinium,
protonated ammonium, sulfonium, or phosphonium salts. Due to the "soft" Coulombic
energy, which falls off only with the inverse distance (r- I ) between the charges,
geometric matching is of lesser importance for this intermolecular force than for any
other one ; in addition, longer distances with large ions can be "paid off' by larger
polarizibility and higher charge dispersion. Even if an additional kind of binding unit,
such as stacking between aromatic surfaces (see below), is present, one can describe the
additivity by the simple equation (3); there tJ.GN stands for one kind of interaction site
(e.g., salt bridges), n for the number of such sites, and tJ.GM for another kind of interaction
with the number m:



o 3 6 9 12

Figure 2. Experimental complexation free energies (-AG, kl/mol) in water vs number n of salt bridges in ion
pairs . Cations include, for example, alkali, ammonium, and sulfonium; anions include sulfate, sulfonate,
phosphate, carboxylate, and phenolate; A,B and C,C' are complexes of a tetraphenolate cyclophane (4-) with
Me.N+ and an azoniacyclophane (4+) with mono- and dianionic naphthalene derivatives; D denotes anionic
(sulphonate or carboxylate) with cationic (ammonio) triphenylmethane derivatives ; E denotes organic dianions
with organic dications ; F denotes cationic azamacrocyle (6+ charges) with aliphatic dicarboxylates; G denotes
cationic azacrowns with adenosine mono-, di- and triphosphates [H.·J . Schneider and A. Yatsimirsky,
Principles and Methods in Supramolecular Chemistry (Wiley, Chichester, 2000). Reproduced with permission .
Copyright 2000, Wiley .].

Open macrocyles discriminate anions mainly by different charges, I as exemplified

with an azacrown ([18]aneN6). In water this macrocycle binds cr
with log K = 1.8 and
801- with log K = 4.1,11 with the doubling for 801- expected on the basis of Eq. (3) .
Remarkable selectivity with anions of equal charge is however observed with several
well-known macrobicyclic hosts, which are discussed in several reviews.v'? They
discriminate anions quite efficiently by shape but also by hydrogen bonds.v' The latter is
also responsible for the spectacular efficiency of porphyrin analogs for fluoride binding ,"
as well as for anion binding by amide-type receptors (see section below).

2.2. Salt Effects on Ion Pairing

The change of activity coefficients of ions by added salts is traditionally described by

the Debye-Hiickel equation (3) as a function of ionic strength 1 and z:

log K = log Ko + ZAZB vI/ (l + vI) (4)

Even though Eq. (3) is derived for simple spherical ions, it has been shown that it holds
for quite anisotropic large organic ions bearing multiple, highly dispersed charges; the
observed plots (Figure 3)15 not only are linear but also have slopes close to the theoretical
value predicted by the coeffic ients in Eq. (3).

5.0 5.0

4.8 4.8

4.6 4.6
log K log K
4.4 4.4 +
4.2 4.2

4.0 4.0

3.8 3.8

3.6 3.6

3.4 3.4

0.0 0.1 0.2 0.3 0.4 0.5

/1/2 (1 + /112)

Figure 3. Oebye-Hiickelcorrelations of salt effects (solid squares: tetraphenolate resorcareneas anion + E4NBr
as cation, NaOO as added salt; open squares: tetraazonia cyclophaneas cation, dinitrophenolate as anion, NaCI
as added salt [H.-J. Schneider, R. Kramer, S. Simova, and U. Schneider, J. Am. Chern. Soc. 110,6442 (1988).
Reproducedwith permission. Copyright1988, American Chemical Society.].

The average free energy increment derived from the plot of experimental b.G vs. the
number n of charge-charge interactions accord ing to Eq. (3) (Figure 2) yields as
discussed above a value of b.b.G = 5 ± I kl/mol for one charge-charge interaction.l'" if
the ionic strength of the medium is around 0.1 M, as typical of many buffer solutions .
Recent extrapolation to zero ionic strength, again with linear curves and slopes close to
the theoretical value, shows an increased increment of b.b. G = 8 ± I kllmol. 16 These
correlations allow the design of anion receptors in aqueous solutions based on ionic
interactions by rules of a thumb. It should be borne in mind, however, that simple ion
pairs are quite limited in terms of selectivity. As demonstrated in Figure 2, the observed
affinities are rather independent of size or polarizability of the anion.

2.3. Ion Pairing and Additional Lipophilic Effects

Selective anion complexation in water beyond charge effects can be achieved if

additional interactions are present, as typical for many organic anions. In aqueous media,
such additional contributions are mainly van der Waals effects, which in contrast to for
instance hydrogen bonds are perfectly maintained in water due the extremely low
polarizibility of this medium. In a series of aromatic ion pairs , stacking contributions
have been found to be again essentially additive according to Eq. (3) .17 As shown in
Figure 4, one can assume 5 kl/mol for each charge-charge interaction in these complexes
and plot the remaining free binding energies against the number m of paired benzene



0 2 3 4 5 6 7 8
Figure 4. Van der Waals (stacking) contributions in aromatic ion pairs. Exper imentalliG values are corrected
by 4.8 kJ/mol for each charge-charge interaction, and the resulting IiIiG values are plotted against the number m
of phenyl rings present in a given complex . Complexes H and I are off-line due to geometric mismatch . [H.-J.
Schneider, T. Schiestel, and P. Zimmermann, J. Am. Chern. Soc. 114 , 7698 (1992). Reproduced with
permission. Copyright 1992, American Chemical Society.].

units present in a given complex, resulting in a linear correlation with a sensitivity (slope)
of around 1.5 kJ/mol per phenyl-phenyl interaction. It should be noted that charge-
transfer interactions do not contribute significantly to the observed affinities, although
these complexes do show typical charge-transfer bands at higher wavelength in the UV-
vis spectra.
A combination of ion pairing and lipophilic interactions leads to the most efficient
artificial receptors for nucleotide anions in the form of aminocyclodextrins (Scheme 1).18

~"OHOH .S\

o HO



Scheme 1. Aminocyclodextrins for nucleotide anion complexation. Host I: RA = RB =
-NH2CH3, other R = -DH. Host II: all R = -NH2CH3•

Again, the binding contributions seem to be largely additive, with the major driving
force obtained from the salt bridges. One observes the predicted increase with the
increasing number of protonated aminogroups in the host compound, and from AMP over
ADP to ATP, the latter being complexed with K = 3.25 x 106 in water. Discrimination
between different nucleobases and sugar moieties (2'-hydroxy vs 2'-deoxy riboses) is
brought about by the cyclodextrin macrocycle.
Examples for even larger contributions of lipophilic effects are positively charged
porphyrin host compounds, which bind electroneutral nucleosides as well as nucleotide
anions with affinities as high as (Figure 5).19 Inorganic HPol- anions already have
sizeable affinity for the multiple binding sites offered by the porphyrin moiety, bearing
four positive, significantly delocalized charges." With several diazapyrenium host
compounds containing large aromatic surfaces one finds about the same affinities of
uncharged nucleosides as with anionic nucleotides."
For a stacking unit in nucleotide anion receptors one can also use, for example,
pyrene units instead of porphyrins. A corresponding complex is shown in Figure 6, with
complexation constants in water around K = 4000 ~l , as for the nucleotide anion
TMP.22 It should be noted that these affinities are reached without any salt bridges. In
DMSO as medium, the affinities were only around K = 20 to 200 ~l, in line with
stacking being the major driving force of the nucleotide binding, which can fully
materialize only in water. The practical advantage of aqueous media is also the
convenient detection by fluorescence changes.21,22

Figure S. Complex of a bis-porphyrinium host with AMP. CHARMm gas-phase simulation; hydrogen atoms
omitted, porphyrin-nucleot ide complex. 1M. Sirish and H.-J. Schneider, J. Am. Chern. Soc. 122, 5881 (2000) .
Reproduced with permission . Copyright 2000, American Chemical Soc iety.].

Ar =

Figure 6. A complex between a pyrene-containing anion receptor and AMp 2-


2.4. Hydrogen Bond-Based Anion Receptors

Hydrogen bonding is most frequently the basis of anion recognition in proteins, for
example, complexing phosphateor sulfate anions." Synthetic receptors on this basis have
been aptly described by Reinhoudt" and others.' With most artificial host compounds
based on hydrogen bonding, one is still restricted here to less competitive solvents than
water, such as to chloroform, acetonitrile, or DMSO. Nevertheless, such studies can shed
light on the mechanism of anion binding in natural receptors and lead to new host
compounds that may used for examplefor anion transportor for anion extraction.
A systematic exploration of halide association with mono-, di-, and tri-dentate
amides (Scheme 2) shows a distinct advantage for chloride, in line with its better acceptor
quality compared to bromide or iodide (fluoride could not be measured due to
hygroscopicity problems with the tetraalkylammonium salts used here).22 This sequence
of hydrogen-bond strength is also visible in the increasing NMR deshielding of the
-CONH amide proton (Scheme 3; a similar sequence was also reported for sugar models
as anion receptors)."

~ ~ f
8 Me MerN-(CI+,),-N
(CHz)z 8

C6HsCOO- 6.6 14.4 13.1

cr 5.7 11.6 14.2
Br- 4.6 7.2 12.0
r 4.1 9.0 9.8

Additive AG (kl/mol): cr r
AAG.v per -NHCo-, ca. ( ! ) : 5 4 3.5
NMR CIS (ppm) ofN-H 5 3 2

Scheme 2. Amide 6.G functions for anion binding. Measurements in CDC!), as BU4N+
salts; 6.Gin kl/mol, [F. Wernerand H.-J. Schneider, Helv. Chim. Acta 83, 465 (2000).].

(R = p-CI-Phenyl)
cr 9.2 21.3
3.4 10.2
<3 <3
(T in CDCh with R = l-Adamantyl : 20.4 K = 44.000 ~1)

Additive !:1G (kl/mol, in DMSO): cr

!:1!:1Gav per urea unit, ca. (!) : 8 3.5 <3

Scheme 3. Urea !:1G functions for anion binding. Measurements in ~-DMSO; !:1G in
kJ/mol. [F. Werner and H.-J. Schneider, He/v. Chim. Acta 83, 465 (2000)].

It is gratifying, and useful for future design of such receptors, that the affinity
comparison of all nine complexes studied shows again additive binding free-energy
increments !:1!:1G, in analogy to our observations with ion-pair based anion receptors and
to earlier studies on hydrogen-bond-based host-guest complexes.f These increments are
always relative scales, very much in analogy to a-values in Hammett or Taft equations.
In particular, they will change with the solvent; thus, one finds with carbon tetrachloride
instead of chloroform for hydrogen-bond association about twice as large association
energies for amide-type of aggregates." An exception of the regular increase seen with
the halides and the tridentate acceptor is obviously carboxylate as the anion component.
This is the consequence of the steric orientation of the two carboxylate oxygen atoms
towards two properly spaced amide protons in the complex, as illustrated in Figure 7.
The directional forces brought about by the orientation of hydrogen bonds in contrast to
other intermolecular interactions is the selectivity basis of the several carboxylate or
phosphate complexing agents described in the literature.'
Significant increase of affinities toward anions can be reached with the larger
hydrogen-bond-donating capacity of ureas instead of amides. 1d,22 Association constants of

a 0 0

~\ 0 0/ .,rI~

Figure 7. Hydrogenbondingfor anion complexation, based on polydentateurea receptors. (a) Complex with a
dicarboxylate. (b) Complexwith chloride, including space-filling stucture(CHARMm forcefield simulations).

increments are largest for chloride compared to bromide and iodide, for the latter only an
upper limit can be given (Scheme 3).

2.5. Simple or Highly Preorganized Receptors?

Obviously, one can obtain high affinities already by combination of the largest
number of usable acceptor functions within a quite flexible host compound, as visible for
example by comparison of the mono-and tridentate amide or urea derivatives (Schemes 2
and 3). There seems to be no particular advantage in preorganizing all possible functions
in the conformation required in the final complex, as long as the functions can adopt a
suitable convergence without building up strain. In line with this, Davis et al. have
reported remarkably high anion affinities also with hosts based on anion-binding
functions assembled on a steroid matrix instead of being implemented within a large
cavity made by multistep synthesis." The possible disadvantage of flexible anion
receptors with the functions aligned along a simple alkane chain instead of receptors
where the functions are conformationally preorganized in a convergent manner has been
studied with a the series of anion-host compounds shown in Scheme 4.27 The host and
guest derivatives were so chosen that strainless matching between the amide acceptor

functions and the opposing carboxylate anions was always possible, as controlled by
computer-aided molecular modeling. The observed binding-energy disadvantage due to
the presence of more single bonds was much less than usually expected on the basis of
entropy costs due to limitation of free rotation around single bonds in such complexes . If
one plots the binding free energy flG against the number n of single bonds present in the
complexes, one obtains a surprisingly linear decrease of flG with increasing n, but a slope
that corresponds to a disadvantage of only 1.3 kJ/mol per single bond. Similar results
were obtained with related cation-anion combinations in water, where with increasing
number of single bonds the disadvantage was only around 1 kJ/moI. 16 Literature data on
the entropic disadvantage of one single bond were until now derived largely on the basis
of theoretical considerations; they vary between over the large range 2-<i.4 kl/mol,"
translating into uncertainties of predicted variation of association constants by orders of
magnitude. Now for the first time experimental data show that the need for
conformational preorganization for effective recognition has been overestimated. This is
of obvious importance in particular for industrial applications, for which many anion
receptors reported in the literature may be just too expensive to synthesize. It also should
be kept in mind that anion receptors in nature are based on peptides containing a large
degree of conformational freedom, and yet have an unsurpassed efficiency.

sa ( in kJ/mol ) -12.0 -10.9

Scheme 4. The small influence of flexibility on complexation free energy flG (in kJ/mol,
measured in CDC b) illustrated with two examples . For data, see F. Eblinger and H.-J.
Schneider, Angew. Chern. Int. Ed Engl., 37, 826 (1998).


Our work is supported by the Deutsche Forschungsgemeinschaft, Bonn, and the

Fonds der Chemischen Industrie, Frankfurt.


I. a) A. Bianchi. K. Bowman-James, E. Garcia-Espana (Eds.), Supramolecular Chemistry ofAnions, (Wiley-

VCH, New York, 1997). b) P. D. Beer and P. A. Gale, Angew. Chem.. Int. Ed. Engl. 40,487 (2001). P. D.
Beer and P. Schmitt, Current Opin. Chem. Bioi. 1,475 (1997). P. D. Beer and D. K. Smith, Progr. Inorg.

Chem. 46, I (1997). c) A. Casnati, L. Pirondini, N. Pelizzi, and R. Ungaro, Supramolecular Chemistry 12,
53 (2000). d) J. Scheerder, J .F. J. Engbersen, and D. N. Reinhoudt, Reel. Trav. Chim. Pays-Bas 11,307
(1996). M. M. G. Antonisse and D. N. Reinhoudt, Chem. Commun. 443 (1998). e) P. D. Beer and D. K.
Smith, Progr. Inorg. Chem. 46, I (1997). t) P. A. Gale, Coord. Chem. Rev. 213, 79 (2001). g) F. P.
Schmidtchen and M. Berger, Chem. Rev. 97, 1609 (1997). h) T. S. Snowden and E. V. Anslyn, Curro
Opin. Chem. BioI. 3,740 (1999). See also J. J. Lavigne and E. V. Anslyn, Angew. Chem.. Int. Ed. Engl.
40,3118 (2001). i) J. L. Sesslerand J. M. Davis, Acc. Chem. Res. 34,989 (2001). j) A. P. Davis and A. S.
Wareham, Angew. Chem. Int. Ed. Engl. 38,2978 (1999).
2. S. Mangani and M. Ferraroni, in: Supramolecular Chemistry ofAnions. ref. la, (1997),p. 63.
3. H.-J. Schneider, Angew. Chem.. Int. Ed. Engl. 30, 1417(1991). H.-J. Schneider, Chem. Soc. Rev. 22, 227
4. M. W. Hosseini, J. M. Lehn, and M. P. Mertes, Helv. Chim. Acta 66,2454 (1983).
5. See, for example, H.-J. Schneider and T. Blatter,Angew. Chem.. Int. Ed. Engl. 31, 1207(1992).
6. See, for example, a) J. O'M. Bockrisand A. K. N. Reddy, Modern Electrochemistry, (Plenum, New York,
1970). b) B. E. Conway, in: Comprehensive Treatise on Electrochemistry, Vol. 5, Eds. B. E. Conway, J.
O'M. Bockris,and E. Yeager(PlenumPress, New York, 1983).
7. For data see ref. 3b, and H.·J. Schneider and A. Yatsimirsky, Principles and Methods in Supramolecular
Chemistry. (Wiley,Chichester, New York,Weinheim, Brisbane, Singapore, Toronto,2000), p. 70 ff.
8. See, for example, B. Honig and A. Nicholls, Science 268,1144 (1995). B. Honig, K. Sharp, and A. Yang, J.
Phys. Chem. 97, 1\01 (1993).
9. See C. De Stefano, C. Foti, O. Giuffre, and S. Sammartano, Talanta43, 707 (1996), and references cited
\0. H.-J. Schneiderand I. Theis, Angew. Chem.. Int. Ed. Engl. 28, 753 (1989).
II. R.1. Gelb, L. B. Schwartz, and L. J. Zompa, Inorg. Chem. 25, 1527(1986), and referencescited therein.
12. E. Kimuraand T. Koike, Chem. Commun., 1495 (1998). E. Kimura, Top. CurroChem. 128, 113(1985).
13. E. Graf and J. M. Lehn, J. Am. Chem. Soc. 98,6403 (1976). B. Dietrich, B. Dilworth, J. M. Lehn, J. P.
Souchez,M. Cesario, J. Guilhem, and C. Pascard, Helv. Chim. Acta 79, 569 (1996).
14. M. Shionoya, H. Furuta, V. Lynch, A. Harriman, and J. L. Sessler, J. Am. Chem. Soc. 114,5714 (1992). V.
Kral, A. Andrievsky, and J. P. Sessler, Chem. Commun. , 2349(1995), and references cited therein.
15.H.-J. Schneider, R. Kramer, S. Simova, and U. Schneider, J. Am. Chem. Soc. 110,6442 (1998).
16. A. Md. Hossain and H.-J.Schneider, Chem. Eur. J. 5, 1284(1999).
17. H.-J. Schneider, T. Schiestel, and P. Zimmermann, J. Am. Chem. Soc. 114,7698 (1992).
18. A. V. Eliseevand H.-J. Schneider,J. Am. Chem. Soc. 116,6081 (1994).
19. M. Sirish and H.-J. Schneider,J. Am. Chem. Soc. 122,5881 (2000).
20. M. Sirish and H.-J. Schneider, Chem. Commun.. 23, (2000). The binding constantsfor buffer-freesolutions
reported in this preliminary paper were later found to be too high, due to difficulties of pH control in pure
water(T. Liu and H.-J. Schneider, unpublished data.)
21. I. Piantanida, V. Tomisic, and M. Zinic, Perkin Trans. 2, 375 (2000), and references cited therein.
22. F. Wernerand H.-J. Schneider, Helv. Chim. Acta 83, 465 (2000).
23. H. Luecke and F. A. Quiocho, Nature 347, 402 (1990). J. J. He and F. A. Quiocho, Science 251, 1497
(1991). F. A. Quiocho and J. W. Ptlugrath, J. Mol. BioI. 200, 163(1988), and references cited therein.
24. J. M. Coteron, F. Hacket, and H.-J. Schneider,J. Org. Chem. 61, 1429(1996).
25. H.·J. Schneider, R. K. Juneja,and S. Simova,Chem. Ber. 112, 1211 (1989).
26. A. P. Davis, J. J. Perry,and R. P. Williams, J. Am. Chem. Soc. 119, 1793 (1997).
27. F. Eblingerand H.-J. Schneider, Angew. Chem.. Int. Ed. Engl. 37, 826(1998).
28. a) M. L. Page, The Chemistry of Enzyme Action, Ed. M. L. Page, (Elsevier, Amsterdam, 1984),p.
I. b) W. P. Jencks, Proc. Natl. Acad. Sci. USA 78,4046 (1978). c) D. H. Williams, M. S. Searle,
M. S. Westwell, V. Mackay, P. Groves, and D. A. Beauregard, CHEMTRACTS - Organic
Chemistry 7, 133 (1994). d) M. S. Searle, D. H. Williams, and U. Gerhard, J. Am. Chem. Soc.
114, \0697 (1992), and references cited therein.
Design criteria for polyalcohol hosts

Benjamin P. Hay, David A. Dixon, Gregg J. Lumetta, Rubicelia Vargas,

and Jorge Garza"


Anion recognition by synthetic hosts is a rapidly emerging field. l -4 One of the

challenges is the development of receptors that selectively bind polyatomic oxyanions,
such as N0 3- and SO/-, which are important targets for environmental, industrial, and
health-related applications. A successful approach for preparing molecules that
coordinate with anions has been to add hydrogen-bond donors to an organic scaffold to
yield charge-neutral receptors that interact with anions through hydrogen bonding.5- 12
Because hydrogen bonds are directional, it should be possible to design hosts with shaped
cavities that are capable of differentiating between anionic guests with different
geometries . The deliberate design of host architecture requires knowledge of the
structural aspects of hydrogen bonding with the guest anion. Yet, surprisingly little
attention has been given to this critical facet of anion host design.
Studies of hydrogen bonding in a wide variety of systems suggest that certain
geometric features must be present to attain the optimal hydrogen-bonding interaction
between a generic donor, D-H , and the oxygen atom acceptor of an oxyanion, o-A.13•14
First, for any given hydrogen bond, there is an H···O distance that will give the strongest
interaction. This distance will depend on the identity of the D-H,o-A pair and can show
significant variation. Second, the D-H vector should point toward the acceptor atom
yielding D-HoooO angles near 180°. Third, there may be some directionality at the
oxygen acceptor defined by preferred H",o-A angles and H···o-A-O dihedral angles.
Studies of weak-to-moderate hydrogen bonds with organic oxygen acceptors, that is,

• Benjamin P. Hay and David A. Dixon, W. R. WileyEnvironmental Molecular Sciences Laboratory, Pacific
Northwest National Laboratory, PO Box 999, Richland, Washington, 99301-9904, USA. GreggJ. Lumetta,
PacificNorthwest National Laboratory, PO Box 999. Richland, Washington, 99301-9904, USA. Rubicelia
Vargas and Jorge Garza, Departamento de Quimica, Division de Ciencias Basicas e Ingenieria, Universidad
Autonorna Metropotitana-Iztapalapa, A.P. 55-534,Mexico Distrito Federal 09340, Mexico.

Fundamentals andApplications of AnionSeparations, edited by

Moyerand Singh,Kluwer AcademiclPlenum Publishers, NewYork,2004 43
44 B. P. HAY ET AL.

ketones, aldehydes, ethers, epoxides, and alcohols, show that there is a weak, but definite,
preference for the D-H vector to point toward the oxygen atom in directions traditionally
associated with the location of electron lone pairs based on Sp2 or Sp3 hybridization.1s- 19
In a recent communication, we presented crystallographic evidence for the existence
of significant and general oxygen acceptor directionalities in both trigonal planar and
tetrahedral oxyanions.i" We now examine in detail the structural aspects of hydrogen-
bonding interactions between a single alcohol donor, methanol (MeOH), with the trigonal
planar oxyanion, N03-, and the tetrahedral oxyanion, so,". Electronic-structure
calculations are used to determine stable geometries, interaction energies, and potential-
energy surfaces for selected structural distortions in one-to-one complexes. Comparison
with experimental data both verifies the theoretical results and firmly establishes the
nature of the structural requirements for optimal interaction between the alcohol O-H
donor group and oxyanion receptor. Using this information to generate input structures
for higher-order complexes, additional calculations yield possible configurations for the
placement of three alcohol donors about N03- and four alcohol donors about S04 2-.


Possible geometries for complexes of MeOH and methane with N03- and sol- were
explored with semi-empirical calculations using the PM3 module" of the MacSpartan
program.f The PM3 geometries were further optimized using density functional theory
with the generalized gradient approximation in which non-local corrections are
introduced perturbatively.23,24 Initial optimizations of these structures were done with the
BP86 functionae s.26 and the DN" polarized double numerical basis set using the
MacSpartan program?2 For all one-to-one complexes, further optimization of the
BP86IDN** geometries was done with the B3LyP functionae S,27 and a polarized triple-t;
basis set (TZVP)28 using the NWChem program.29.30 Frequency calculations were done
at the B3L yprrzvp level to verify that the optimized geometries were stable points on
the potential energy surface.
Potent ial energy surfaces for distortions of'selected structural parameters in one
[N03(MeOH)r complex, 2, and one [S04(MeOH)f- complex, 7, were obtained using
single-point energies at the BP86IDN" level of theory. The structural parameters were
the O"'H distance, the O-H"'O angle, the A-o"·H angle, and the O-A-o"'H dihedral
angle (A = central atom of the anion). For each potential energy surface, geometries
were generated by variation of the specified structural parameter, keeping all other
structural features constant.
Experimental structural parameters for intermolecular hydrogen bonds between
aliphatic alcohols and the N03- or sol- anions were retrieved from the April 2001
release (Version 5.21) of the Cambridge Structural Database using the QUEST and
VISTA programs." For each anion, the search was subject to the following constraints:
H"'O distance between 1.0 and 2.0 A, anion not coordinated to a metal ion, R-factor less
than 0.10, no disorder, and error free. This search yielded 80 examples for N0 3- and 25
examples for so,", These structures were corrected for systematic errors in O-H bond
length by normalization of all O-H bond lengths prior to plotting and statistical

analysis." This involved moving the hydrogen position along the observed O-H bond
direction until the O-H distance was equal to a standard value (0.95 A).32


3.1. N03- Complexes

3./ ./. Structures and Energies/or [NOJ(MeOH)] - and [NOJ(methane)] - Complexes

Two stable geometries located for [N03(MeOH)r, 1 and 2, are shown in Figure 1.
Electronic binding energies, tlE, at both levels of DFT theory are given in Table 1, where
it can be seen that , relative to the B3LYP/TZVP method, the BP86IDN** method
underestimates the strength of the interaction by 1.5 kcal/mol (10%) for 1 and 1.1
kcal/mol (8%) for 2. The!!.E values are in the range of -13.5 to -15.7 kcal/mol,
revealing that these interactions should be viewed as "strong" (> 10 kcal/mol) hydrogen
bonds . 13 For comparison , normal hydrogen bonds, generally formed between neutral
donor and acceptor groups, exhibit !!.E values within the range of 3 to 10 kcal/mol.

" ~
2 3

Figure I. Opt imized geometries (B3LYPffZVP) of two [NOJ(MeOH)r complexes, 1 and 2 , and the
[NOJ(methane))- complex, 3.

Table 1. Electronic binding energies (!!.E) for N03- complexes 1 and 2 at the
BP86IDN** and B3LYP/TZVP levels of theory and comparison of calculated structural
parameters to average values observed in X-ray crystal structures.·
Complex I Complex 2

BP86 B3LYP BP86 B3LYP X-ray

,t;"E -14.3 - 15.7 - 13.5 -14.6
O"'H 1.73 1.75 1.72 1.76 1.84 ± 0.08
O"'H-0 171 176 163 162 164±9
N-0···H 113 116 118 118 113 ± 8
O\-N-0"'H 0 4 0 0 0 ±22
0 2-N-0' ''H 180 176 180 180 180 ± 21
a Units: energ ies in kcal/mol, distances in A, angles in degrees . Uncertainties in X-ray values represent one
standard deviation from the mean. 6E = Ed ee( complex) - E.",,(NOJ-) - E.",,(ligand).

The DFT structural parameters are compared with average X-ray values in Table 1.
Both levels of theory give similar geometries and, with the exception of O···H distance,
the structural parameters for the o-HoooO hydrogen bond are in good agreement with the
average experimental values . In all cases, the o-H'''O angles are near linear, 160 to
180°; the N-o"'H angles are acute, 110 to 120°; and the o-N-OoooH dihedral angles are
near 0 or 180°, indicating a preference for the hydrogen atom to lie within the NO)-
The calculated Os--H distances are roughly 0.1 A shorter than the X-ray values.
This difference is not unexpected. "Gas phase" calculations on complexes between an
ion and a single ligand generally yield tighter complexes than observed in condensed
phases, because (i) a lone ligand does not polarize and redistribute the charge on the ion
to the extent that a full coordination shell of ligands does, and (ii) hydrogen bonds, which
are predominantly electrostatic, are weakened when placed in a dielectric medium. Later
(Section 3.1.4), it is shown that the experimental Os--H distance is obtained in the
calculation when three MeOH ligands are present.
Complex 1 is more stable than 2 by 0.8 kcallmol at the BP86IDN** level and 1.2
kcallmol at the B3LYPITZVP level. This added stability is likely due to the additional
C-HoooO hydrogen bonding interaction that is present in 1, where one of the methyl
group hydrogen atoms contacts an adjacent NO)- oxygen atom with the structural
parameters (B3LYP/TZVP): HoooO, 2.65 A; c-a-oo, 125°; N-OoooH, 118° ;
o-N-OoooH, 12 and 168°. In order to test this assumption, calculations were performed
to obtain the structure and binding energy of the NO)- complex with methane.
Only one stable [NO)(methane)r complex, 3, was located (Figure 1), with binding
energies of -1.7 kcallmol with BP86/DN** and -2.6 kcal/mol with B3LYP/TZVP.
Structural parameters (B3LYPITZVP) for the single C-HoooO hydrogen bond in 3 are:
HoooO, 2.40 A; C-HoooO, 167°; N-ooooH, 105°; o-x-o-ei, 0 and 180°. With the
exception of the C-HoooO angle, which is much more acute in 1, the structural parameters
for the C-HoooO interaction are similar to those observed in 1. The magnitudes of the
binding energies for 3 are fully consistent with the calculated differences in binding
energy between 1 and 2 if we note that the C-HoooO interaction in 1 is weakened by
failure to attain the optimal geometry due to the presence of the much stronger o-H"'O
hydrogen bond and to the sharing of the negative charge between two hydrogen-bond
protons . Finally, we observe that, although the C-HoooO distance is significantly longer
than the o-H"'O distance, the methane C-H donor group exhibits an orientation with
respect to the NO)- acceptor that is remarkably similar to the alcohol o-H donor group.

3.1.2. Potential-Energy Surfaces for Selected Structural Distortions ofComplex 2

The potential-energy surfaces (PESs) for distortions of OoooH distance, o-HoooO

angle, N-ooooH angle, and o-N-o'''H dihedral angle provides valuable information for
host design by showing the extent of destabilization to the Ro-H"'ON~- hydrogen
bond that would result from variation in the different structural parameters. Therefore,
we performed a series of calculations to generate these PESs for distortions of the
structural parameters of the hydrogen bond in 2, where we choose 2 over 1 in order to
minimize the influence of C-HoooO interactions. Plots of relative energy (BP86IDN**)
versus structural distortion are presented in Figure 2. The Y-axes run from 0 to 5

kcal/mol in each case to allow for a direct visual comparison of the different PESs. A
useful point of comparison between the PESs is the extent of distortion required to give a
1.0 kcal/mol decrease in binding energy for the complex. Examination of each PES
yields the following results (low 1 kcal/mol threshold S minimum S high 1 kcal/mol
threshold): O···H distance, 1.54 S 1.72 S 1.97 A; O-H·"O angle, 148 S 163 S 180°;
N-o···H angle, 100 S 118 S 138°; G-N-o"·H dihedral angle, -42 S 0 S 42° or 138 S

10 10

8 8
"'B 6 "'B 6
1: 1:
4 .* 4

2 2
0 0
5 5
4 4
0 0
E 3 E 3
-;; -;;
'" 2
... J:l 2
t.i t.i

0 0
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 90 120 150 180
O' ''H distance, A O.H" 'O angle, "

10 10

8 8

"'B 6 "'B6
1: 1:
.* 4 .*4

2 2

0 0
5 5

4 4
0 0
E3 E3
-;; -;;
J:l 2 J:l 2
t.i t.i

120 150 180 ·120 ·60 0 60 120 180

N.O" ' H angle, · O· .O...H dihedral angle, ·

Figure 2. A compar ison of experimental distributions of geometric parameters with potential energy surfaces
(PES) obtained at the BP86IDN" level of theory for the [NOl(MeOH)r complex, 2: O"'H distance (top left),
o-H"'O angle (top right), N-Q" 'H angle (bottom left), o-N-Q"'H dihedral angle (bottom right).
48 B. P. HAY ETAL.

180:5 222°. Thus, the variation in bond distance is 0.20 to 0.25 A, the variation in the 0-
H"'O angle is -15°, and the variation in the N-o"'H angle is -20°.
Histograms of the distributions of geometric parameters observed in the X-ray data
are shown above each PES in Figure 2. With the exception of the O"'H distances, the
distributions are clustered in the bottom of the potential wells. In all cases, the bulk of
the data lies within 1.0 kcallmol from the minima. This result provides additional
verification of the theoretical model and further validates the comparison of "gas- phase"
calculations to condensed-phase results. The agreement suggests that the hydrogen-
bonding geometries observed in the crystal structures are the result of some intrinsic
geometric preferences that are also captured by the electronic-structure (OFT)
calculations on isolated one-to-one complexes.

3./.3. Intrinsic Hydrogen Bonding Directionality ofN0 3

It is possible to rationalize the observed geometric features of the hydrogen bonds in

1-3 by examination of the attractive portion of the electrostatic potential for the NOJ -
anion." A contour map of the electrostatic potential surface in the plane ofNO J - (Figure
3, left) reveals the location of minima for the placement of positive charge. There are
two minima on either side of the each oxygenatom within the planeof the anion, in other
words, at approximately the positions expected for Sp2 hybrid lone pairs. These minima
closely correspond to the observed locations of the hydrogen atoms in the complexes.
The electrostatic potential surface reveals that the NOJ - anion has an intrinsic
hydrogen-bonding topography (Figure 3, right) in which a hydrogen-bond donor must
approach along one of six vectors in order to form the most stable complex. Whereas the
O···H distance will depend on the nature of the donor group and the number of donor
groups, the geometric features that define the orientation of each donor group with
respect to the NOJ - oxygen atom will be relatively constant. In the optimal complex with
a given donor group, the D-H, the D-H vector will point towards the oxygen atom, the
N-o"'H angle will be -120°, and the hydrogen atom will lie in the NOJ - plane giving
O-N-O···H dihedral angles of 0 and 180°.

Figure 3. (left) Contour map of the electrostatic potential surface for a positive point charge in the plane of the
NO,- anion generated at the BP86/DN·· level. The six equivalent low-energy regions are bound by contours at
-165 kcal/mol. Contour lines are spaced at 20 kcal/mol intervals. (right) The most stable complexes are formed
when the hydrogen atom approache s along one of the two vectors attached to each oxygen atom .

3.1.4. Possible Arrangements ofMeOHGroups in [NOj(MeOHhl - Complexes

The preceding sections have identified structural features needed to achieve the most
stable one-to-one complexes between MeOH and N0 3-. It is desirable to have multiple
binding sites in a polyalcohol host with a cavity designed for binding N0 3-. One possible
design strategy would be to provide one hydrogen-bonding site for each oxygen atom
acceptor. In order to identify stable arrangements of three alcohol donor groups about the
N0 3- anion, we optimized two possible geometries of [N03(MeOH)3r at the BP86/DN**
level. The structures are shown in Figure 4.
The more stable complex, 4, was obtained after optimization of a starting geometry
built with three MeOH in the configuration observed in 1. The total interaction energy of
4 is -34.9 kcal/mol. This translates into -11 .6 kcal/mol per donor group, a value that is
decreased from the -14.3 kcallmol observed for 1. The three OH groups lie in the nitrate
plane, in an approximate D3h symmetry. Average structural parameters for the stronger
O-H···O hydrogen bond in 4 are: H...O, 1.85 A; O-H"'O, 173°; N-O···H, 111°;
O-N-o···H, 0 and 180°. The only significant difference between the O-H···O hydrogen
bond geometry in 1 and 4 is that the H...O distance has increased by 0.12 A, now in good
agreement with the experimental average of 1.84 A. The C-H···O contacts present in 4
are quite long, 2.99 A, suggest ing a decreased role of the C-H···O interactions in
stabilizing this complex.
The other complex,S, was obtained after optimization of a starting geometry built
with three MeOH in the configuration observed in 2. This complex can be obtained from
4 by 180° rotations of each methyl group about the OH bond. The total interaction energy
of 5 is -33.6 kcallmol. This translates into -11 .2 kcallmol per donor group, a value that
is decreased from the -13.5 kcal/mol observed for 2. As with 4, the three OH groups lie
in the nitrate plane, in an approximate D 3h symmetry. Average structural parameters for
5 are: H...O, 1.84 A; O-H"'O, 162°; N-O···H, 117°; O-N-o...H, 0 and 180°. As with

Figure 4. Optimizedgeometries (BP86/DN**) for [NO)(MeOHhr complexes, 4 and S.

50 B. P. HAY ETAL.

the comparison of 1 and 4, the only significant difference between the O-HoooO
hydrogen-bond geometry in 2 and 5 is that the He--O distance again has increased by
0.12 A.
Other geometries for the [N03(MeOH)3r complex are possible. However, because
there are no close contacts between any of the MeOH groups in 4 and 5, it is likely that
they represent the two lowest-energy geometries. The complexes 4 and 5 provide
information about the optimal dimensions of a tri-alcohol host cavity. An optimal D 3h
array of three alcohol donor groups would require a host cavity with a radius of 2.59 A to
2.66 A defined by hydrogen atom location in 4 and 5, respectively, and 3.55 A to 3.66 A
defined by oxygen atom location in 4 and 5, respectively.

3.2. 801- Complexes

3.2.1. Structures and Energies/or (S04(MeOH)/- and (S04(methane)/- Complexes

Only one stable geometry was located for [S04(MeOH)]2-, 6, as shown in Figure 5.
As for N0 3-, this complex contains a short O-H---O hydrogen bond and a longer
C-H---O hydrogen bond . Unlike N0 3-, however, the C-H---O interaction is strong
enough to cause the MeOH to adopt a conformation in which the 0-H bond is almost
eclipsed with one of the C-H bonds. Despite starting from several likely geometries, for
example 7 and 8, efforts to locate a stable point for [S04(MeOH)f- in which the MeOH
interaction with SO/- was limited to a single 0-H...O hydrogen bond failed. However,
geometry optimizations with C, symmetry imposed did yield the two structures, 7 and 8,
in which the influence of C-HoooO interactions is minimized. Vibrational analysis
indicates 7 (one negative frequency at -22 cm') and 8 (two negative frequencies at -17
and -38 cm") not to be true minima at the B3LYP/TZVP level of theory, and on
breaking symmetry , both structures collapse to yield 6. liE values, at both levels ofDFT
theory are given in Table 2, where it can be seen that, relative to the B3LYP/TZVP
method, the BP86IDNu method slightly overestimates the strength of the interaction by

'" .,.'lJ- "~

aio .'.'

6 7 8 9

Figure 5. Optimized geometries (B3LYPffZVP) for three [SO.(MeOH)]2-complexes, 6-8 , and one
[SO.(methane)]2- complex, 9.

Table 2. Electronic association energies for sol- complexes 6, 7, and 8 at the

BP86IDN** and B3LyprrZVp levels of theory and comparison of calculated structural
parameters to average values observed in X-ray crystal structures.·

Complex 6 Complex 7 Complex 8

BP86 B3LYP BP86 B3LYP BP86 B3LYP X-ray
ss -31.1 -30 .0 -28.9 -27.4 -28.7 -27.3
O···H 1.50 1.57 1.43 1.51 1.42 1.52 1.73 ± 0.08
O"·H-0 177 178 169 165 162 164 165± 11
S-O···H 128 128 130 132 126 132 123 ± 7
Ol-S-O···H 10 13 60 60 0 0
02-S-O"· H 130 132 180 180 120 120
OrS-O···H 250 253 300 300 240 240
• Units: energies in kcal/mol, distances in A, angles in degrees. Uncertainties in X-ray values represent one
standard deviation from the mean. I!.E = E,I,.(complex) - E,Iee(SO/j - E,I,.(ligand).
b X-ray data does not show a distinct o-S-o."H preference (see Figure 6).

1.1 kcal/mol (4%) for 6, and 1.5 kcal/mol (5%) for 7 and 8. The AE values, from -27.3
to -31.1, are twice as strong as those observed for NO)- consistent with the additional
negative charge in so,".
The DFT structural parameters are compared with average X-ray values in Table 2.
Both levels of theory give similar geometries and, with the exception of O···H distance,
the structural parameters for the Q-H"'O hydrogen bond are in good agreement with the
average experimental values. In all cases, the Q-H"'O angles are near linear, 160 to
180°, and the S-o···H angles are acute, 120 to 135°. It is not meaningful to make a
comparison of Q-S-o···H dihedral angles, because a distinct orientation is not observed
in the X-ray data (Section 3.2.2). The calculated O"'H distances are 0.2 to 0.3 A shorter
than the average experimental value . Possible reasons for such differences were
discussed in Section 3.1.1, and it will be shown (Section 3.2.4) that the calculated O"'H
distance significantly increases when four MeOH ligands are present.
Complex 6 is more stable than 7 and 8, which are roughly the same energy, by 2.3
kcal/mol at the BP86IDN** level and 2.7 kcal/mol at the B3Lyprrzvp level. As with
NO)-, the added stability is due to the C-H"'O hydrogen-bonding interaction clearly
present in 6, where one of the methyl group hydrogen atoms contacts an adjacent sol-
oxygen atom with the structural parameters (B3LYP/TZVP): H...O, 2.26 A; C-H"'O,
156°; S-o...H, III 0; O-S-O"'H, 5, 123 and 246°. For comparison, a single stable
[S04(methane)f- complex, 9, was located (see Figure 5) with binding energies of-6.5
kcal/mol with BP86IDN** and -6.4 kcal/mol with B3LYP/TZVP. Structural parameters
(B3LYP/ TZVP) for the C-H"'O hydrogen bond in 9 are: H'''O, 2.05 A; C-H"'O,
173°; S-o"'H, 124°; O-S-o···H, 0, 120, and 240°. The structural parameters for the
C-H"'O interaction are similar those observed in 6. Although the C-H"'O distance is
significantly longer than the Q-H···O distance, the methane C-H donor group exhibits
an orientation with respect to the S04 2- acceptor that is remarkably similar to the alcohol
Q-H donor group.

3.2.2. Potential-Energy Surfaces for Selected Structural Distortions ofComplex 7

PESs for O"·H distance, o-H"'O angle, S-o"'H angle, and o-S-o"'H dihedral
angle were generated for structural distortions in 7, with 7 chosen over 6 in order to
minimize the influence of C-H"'O interactions. Plots of relative energy (BP86IDN**)
versus structural distortion are presented in Figure 6. The Y-axes run from 0 to 10

O-+--...,....~----,-­ O+-----.----&.f........
10 + - -- - ' - - - " - - - ' - ---'-----+ 1 0 + - - - - - ' - -r-r- --'-----+
8 8
]6 ]6
Oi ;;
.1:4 .1:4
2 2

01-- - - -.-- - - --r- .,......;:::v;---r

1~ lA 1~ 1~ 2.0 90 120 150 180
0 ···11distance. A 0 .11" '0 angle, 0

5 +-~---'-........------'-- - - __t 5 +-~"'----.........- .......- -''--- - ' ---t

4 4

1l 3
] 2

0-+-----....., o
101r---- -'-- - - -'-- - - -+ 10 +---'-- -'-- -'-- -'---'---+
8 8
E 6 ]6
~ .1:4
"'" 4 ..,..
2 2

120 150 180 -180 · 120 ·60 0 60 120 180
S·O" ·II angle, • O-S.O" 'II dihedral angle. 0

Figure 6. A comparison of experimental distributions of geometric features with potential-energy surfaces

obtained at the BP86IDN·· level of theory for the [SO.(MeOH)]2- complex, 7: O"·H distance (top left),
o-Ho··O angle (top right), S--o"'H angle (bottom left), o-S--O"'H dihedral angle (bottom right).

kcal/mol in each case to allow for a direct visual comparison of the different PESs. A
useful point of comparison between the PESs is the extent of distortion required to give a
1.0 kcallmol decrease in the binding energy of the complex . Examination of each PES
yields the following results (low I kcal/mol threshold :s minimum :s high I kcallmol
threshold) : O"'H distance , 1.19 s 1.43 s 1.63 A; D-H···O angle, 159 s 169 s 177°; S-
O···H angle, 112 s 130:s 143°. The variation in the D-H···O distance is 0.20 to 0.25 A,
in the D-H···O angle is _10° , and in the S-o···H angle is -15°. These variations are
very similar to those found in the NO J - complex . The PES for the O-S-o···H dihedral
angle is remarkably flat and does not exceed 0.4 kcallmol over the entire range of 0 to
Histograms of the distributions of geometric parameters observed in the X-ray data
are shown above each PES in Figure 6. Although the correspondence between the PES
and the distribution of X-ray values is not as clear as with NO J - (see Figure 2), the
experimental data for the D-H·"O and S-o···H angles are clearly distributed in the
bottom of the potential wells. Although the O-S-o···H dihedral angle distribution could
suggest a slight preference for 0 and ± 120°, in other words an eclipsed conformation,
there is a significant scatter across the range, consistent with the absence of any
significant barrier to rotation.

3.2.3. Intrinsic Hydrogen Bonding Directionality a/SOl

It is possible to rationalize the observed geometric features of the hydrogen bonds in

6-9 by examination of the attractive portion of the electrostatic potential for the SO/-
anion." A contour map of the electrostatic potential surface in one of the O-S-o planes
of S04 2- (Figure 7, left) reveals the location of minima for the placement of positive
charge . As with N0 3- , there are two minima on either side of the each oxygen atom, but
in the case of SO/ -, the minimum-energy region that lies between the two coplanar
oxygen atoms is larger. Considering that each oxygen atom resides in three O-S-o

Figure 7. (left) Contour map of the electrostatic potential surface for a positive point charge in one of the 0-5-
o planes of 50/ - anion generated at the BP86IDN** level. The four low-energy regions are bound by
contours at -275 kcal/rnol. Contour lines are spaced at 30 kcal/mol intervals. (right) The most stable
complexes are formed when the hydrogen atom lies on the surface of one of the four cones attached to the
oxygen atoms of the 50 / - anion
54 B. P. HAY ETAL.

planes and each plane exhibits an identical electrostatic potential surface, the minima in
the tetrahedral S04 2- can be visualized as a torus about each oxygen atom with bulges
that correspond to o-S-O···H dihedral angles of 0 and ± 120°.
The electrostatic potential surface reveals that the S04 2- anion has an intrinsic
hydrogen bonding topography (Figure 7, right), in which a hydrogen-bond donor must
approach on the surface of one of four cones in order to form the most stable complex.
Whereas the O···H distance will depend on the nature of the donor group and the number
of donor groups, the geometric features that define the orientation of each donor group
with respect to the S04 2- oxygen atom will be relatively constant. In other words, in the
optimal complex with a given donor group, D-H, the D-H vector will point towards the
oxygen atom, and the S-o···H angle will be -130°.

3.2.4. Possible Arrangements ofMeOH Groups in [SOlMeOHhf- Complexes

The preceding sections have identified the structural features needed to achieve the
most stable one-to-one complexes between MeOH and so,", It is desirable to have
multiple binding sites in a polyalcohol host with a cavity designed for binding so,",
One possible design strategy would be to provide one hydrogen-bonding site for each
oxygen-atom acceptor. In order to identify stable arrangements for four alcohol donor
groups about a S04 2- anion, we optimized two possible geometries of [S04(MeOH)4]2- at
the BP861 DN** level of theory. The structures are shown in Figure 8.


Figure 8. A top and sideviewof the optimized geometries (BP86IDN**) for [SO.(McOH).]2- complexes, 10
and 11.
The more stable complex, 10, was obtained after optimization of a starting geometry
built with four MeOH in the configuration observed in 6. The total interaction energy of
10 is -92.9 kcal/mol. This translates into -23.2 kcal/mol per MeOH , a value that is
decreased by - 9 kcallmol from the -31.1 kcallmol observed for 6. The four MeOH
ligands yield a complex with approximately D 2d symmetry. Average structural
parameters for 10 are: H·"O, 1.67 A; D-H"'O, 177°; S-o···H, 132°; O-S-o···H, 12,
107, and 228° . The only significant difference between the D-H"'O hydrogen-bond
geometry in 6 and 10 is that the H"'O distance has increased by 0.17 A, in much better
agreement with the experimental average of 1.73 A. The C-H"'O contacts present in 10
are 2.48 A, 0.22 A longer than those observed in 6.
The other complex, 11, was obtained after optimization of a starting geometry built
with four MeOH in the configuration observed in 7. The total interaction energy of 11 is
-89.5 kcallmol. This translates into - 22.4 kcal/mol per donor group, a value that is
decreased by 6.5 kcallmol from the -28.9 kcallmol observed for 7. As with 10, the four
O-H groups yield a complex with approximately D2d symmetry. Average structural
parameters for 11 are: H"'O, 1.64 A; D-H"'O, 165°; S-o"'H, 130°; D-S-o...H, 60,
180, and 300°. The only significant difference between the D-H"'O hydrogen-bond
geometry in 7 and 11 is that the H"'O distance is 0.21 A longer in 11.
Other geometries for the [S04(MeOH)4]2- complex are possible . However, because
there are no close contacts between any of the MeOH groups in 10 and 11, it is likely that
these structures are representative of the lowest-energy forms. The complexes 10 and 11
provide some information about the optimal dimensions of a tetra-alcohol host cavity for
SO/-. The structures suggest a host cavity with a radius of 2.85 A to 2.92 A as defined
by hydrogen atom locations in 11 and 10, respectively, and 3.85 A to 3.88 A as defined
by oxygen atom locations in 10 and 11, respectively. Finally, we note that, although
SO/- is a tetrahedral anion with Td symmetry, it does not follow that the alcohol O-H
donor groups must be spatially organized in Td symmetry. Because the S-o···H angles
are acute, 10 and 11 have approximately D 2d symmetry. In the case of 11, the D 2d
symmetry is compressed such that the four D-H groups are actually closer to being
distributed in a plane, rather than in a tetrahedral, array.


Electronic-structure calculations at the DFT level on [N03(MeOH)r and

[S04(MeOH)f- have identified the optimal geometries for O-H"'O hydrogen bonding
between these anions and methanol, taken as a representat ive of aliphatic alcohols. The
results are fully consistent with experimental distributions of structural parameters
obtained from analysis of X-ray crystal-structure data. In addition, the results are
rationalized by examination of the electrostatic potential of the anions . Hydrogen-bond
geometries have been discussed in terms of the O"·H distance, D-H"'O angle, A-o"'H
angle, and O-A-o"'H dihedral angle, where A is the central atom of the anion, and
optimal values for these parameters have been presented. The results confirm the
presence of definite oxygen-atom acceptor directionality that is energetically significant.
The calculations on the MeOH complexes in some cases revealed the presence of
C-H"'O hydrogen bonds . DFT calculations on [N03(methane)r and [S04(methane)f-
56 B. P. HAY ET AL.

established that C-H"'O interactions are roughly 20% as strong as the O-H"'O
interactions and, therefore, playa significant role in stabilizing the complexes.
The results presented herein provide criteria for the deliberate design of polyalcohol
host cavities that structurally complement a given oxyanion structure. DFT calculations
on [N03(MeOHhr and [S04(MeOH)4]2- yield some examples of complementary
orientations for sets of alcohol donor groups and provide a basis for assigning optimal
cavity radii for these species.


This work was supported in part by the Pacific Northwest National Laboratory
(PNNL) Laboratory Directed Research and Development program and in part by the
Division of Chemical Sciences, Office of Science, U. S. Department of Energy (DOE) .
Additional support for J. Garza was provided by CONACYT under project 133153. The
work was performed in the William R. Wiley Environmental Molecular Sciences
Laboratory (EMSL) at PNNL (managed and operated under DOE contract
DE-AC06-76RLQ-1830 by Battelle Memorial Institute). The EMSL is a national user
facility funded by the Office of Biological and Environmental Research in the DOE.


I. Supramolecular Chemistryof Anions, edited by A. Bianchi, K. Bowman-James, and E. Garcia-Espana

(Wiley-VHC, New York, 1997).
2. P. A. Gale, Anion coordination and anion-directed assembly: highlights from 1997 and 1998, Coord. Chem.
Rev. 199, 181-233 (2000).
3. P. A. Gale, Anion receptor chemistry: highlights from 1999, Coord. Chem. Rev. 213,79-128 (2001).
4. P. D. Beer and P. A. Gale, Anion recognition and sensing: the state of the an and future perspectives, Angew.
Chem.lnt. Ed. 40,486-516 (2001).
5. W. B. Farnham, D. C. Roe, D. A. Dixon, R. L. Harlow, and J. C. Calabrese, Fluorinated macrocyclic ethers
as flouride-ion hosts - novel structures and dynamic propert ies, J. Am. Chem. Soc. 112, 7707-7718
6. A. P. Bisson, V. M. Lynch, M. K. C. Monahan, and E. V. Anslyn, Recognition of anions through NH-pi
hydrogen bonds in a bicyclic cyclophane - selectivity for nitrate, Angew. Chem. Int. Ed. 36, 2340-2342,
7. A. P. Davis, J. J. Perry, and R. P. Williams, Anion recognition by tripodal receptors derived from cholic acid,
1. Am. Chem. Soc. 119, 1793-1794 (1997).
8. K. Kavallieratos, S. R. de Gala, D. J. Austin, and R. H. Crabtree , A readily available non-preorganized
neutral acyclic halide receptor with an unusual non-planar binding conformation, J. Am. Chem. Soc. 119,
2325-2326 (1997).
9. P. Anzenbacher, Jr., K, Jurslkova, and J. L. Sessler, Second generation calixpyrrole anion sensors, J. Am.
Chem. Soc. 122,9350-9351 (2000).
10. K. Kavallieratos and B. A. Moyer, Attenuation of Hofroeister bias in ion-p air extraction by a disulfonamide
anion host used in strikingly effective synergistic combination with a calix-crown Cs' host, Chem .
Commun. 1620-1621 (2001).
II. P. A. Gale, S. Camiolo, C. P. Chapman, M. E. Light, and M. B. Hursthouse, Hydrogen-bonding pyrrolic
amide cleft anion receptors, Tetrahedron Lett. 42,5095-5097 (2001).
12. K. Choi and A. D. Hamilton, Selective Anion binding by a macrocycle with convergent hydrogen bonding
functionality, J. Am. Chem. Soc. 123,2456-2457 (2001).

13. G. A. Jeffrey, in: An Introduction to Hydrogen Bonding, edited by D. G. Truhlar (Oxford University Press,
Oxford, 1997).
14. G. R. Desiraju and T. Steiner, in: The Weak Hydrogen Bond in Structural Chemistry and Biology (Oxford
University Press, Oxford, 1999).
15. R. Taylor and O. Kennard, Crystallographic evidence for the existence of C-H"'O, C-H"'N, and
C-H"'CI hydrogen bonds,J. Am. Chem. Soc. 104,5063-5070 (1982).
16. R. Taylor, O. Kennard, and W. Versichel, Geometry of the N-H"'O=C hydrogen bond. I. Lone-pair
directionality,J. Am. Chem. Soc. lOS, 5761-5766 (1983).
17. P. Murray-Rust and J. P. G1usker, Directional hydrogen bonding to Sp2_ and sp3-hybridzed oxygen atoms
and its relevance to ligand-macrocycle interactions,.!. Am. Chem. Soc. 106, 1018-1025 (1984).
18. T. Steiner, J. A. Kanters, and J. Kroon, Acceptor directionality of sterically unhindered C-H"'O=C
hydrogen bonds donated by acidic C-H groups, Chem. Commun. 1277-1278 (1996).
19. T. Steiner, Unrolling the hydrogen bond properties of C-H'''O interactions, Chem. Commun.727-734
20. B. P. Hay, D. A. Dixon, J. C. Bryan, and B. A. Moyer, Crystallographic evidence for oxygen acceptor
directionality in oxyanion hydrogen bonds, J. Am. Chem. Soc. 124, 182-183 (2002).
21. J. J. P. Stewart, Optimization of parameters for semiempirical methods, J. Compo Chem. 10, 209-220
22. MacSpartan Pro User's Guide (WaveFunction, Inc., Irvine, California, 2000);
23. R. G. Parr and W. Yang, in: DensityFunctional Theory ofAtoms and Molecules (Oxford University Press,
Oxford, 1989).
24. Density Functional Methods in Chemistry, edited by J. K. Labanowski and J. W. Andzelm
(Springer-Verlag, New York, 1991).
25. A. D. Becke, Density-functional exchange-energy approximation with correct asymptotic behavior, Phys.
Rev. A 38,3098-3100 (1988).
26. J. P. Perdew, Density-functional approximation for the correlation-energy of the inhomogeneous
electron-gas, PhysRev.B 33, 8822-8824 (1986).
27. C. T. Lee, W. T. Yang, and R. G. Parr, Development of the Colle-Salvetti correlation-energy formula into a
functional of the electron-density, Phys. Rev. B 37,785-789 (1988).
28. N. Godbout, D. R. Salahub, J. Andzelm, and E. Wimmer, Optimization of Gaussian-type basis-sets for
local spin-density functional calculations. I. Boron through neon, optimization technique and validation,
Can. J. Chem. 70,560-571 (1992).
29. D. E. Bemholdt, E. Apra, H. A. Fruchtl, M. F. Guest, R. 1. Harrison, R. A. Kendall, R. A. Kutteh, X. Long,
J. B. Nicholas, J. A. Nichols, H. L. Taylor, A. T. Wong, G. 1. Fan, R. J. Littlefield, and J. Nieplocha,
Parallel computational chemistry made easier - the development of NWChem, Int. J. Quantum Chem.
Symp. 29,475-483 (1995).
30. R. J. Harrison, J. A. Nichols, T. P. Straatsma, M. Dupuis, E. J. Bylaska, G. 1. Fann, T. L. Windus, E. April,
J. Anchell, D. Bemholdt, P. Borowski, T. Clark, D. Clerc, H. Dachsel, B. de Jong, M. Deegan, K. Dyall,
D. Elwood, H. Fruchtl, E. Glendenning, M. Gutowski, A. C. Hess, J. Jaffe, B. Johnson, J. Ju, R. A.
Kendall, R. Kobayashi, R. Kutteh, Z. Lin, R. Littlefield, X. Long, B. Meng, J. Nieplocha, S. Niu, M.
Rosing, G. Sandrone, M. Stave, H. Taylor, G. Thomas, J. van Lenthe, K. Wolinski, A. Wong, and Z
Zhang, NWChem. A Computational Chemistry Packagefor Parallel Computers, version 4.0.1, William
R. Wiley Environmental Molecular Sciences Laboratory , Pacific Northwest National Laboratory,
Richland, Washington 99352--{)999, USA, (2001).
31. F. H. Allen and O. Kennard, 3D search and research using the Cambridge Structural Database, Chemical
DesignAutomation News 8(1),31-37 (1993).
32. N. L. Allinger, M. Rahman, and J.-H . Lii, A molecular mechanics force field (MM3) for alcohols and
ethers,.!. Am. Chem. Soc. 112, 8293-8307 (1990).

Suzanne L. Tobey and Eric V. Anslyn'


The recognition of anions using abiotic hosts is an area of interest in the field of
molecular recognition, which recently has been rapidly expanding. The variety of shapes
and sizes of anionic targets has resulted in rather elaborate receptor designs that have
made it possible to investigate the propert ies of anion recognition. These efforts are
directed towards modeling the molecular-recognition characteristics of enzymes and
focused on industrial applications such as sensing and separation. Within this context,
this paper focuses on research done within the Anslyn group directed at physical studies
of recognition systems and the use of these data in other fields.
The ability to design and study synthetic abiot ic receptors for the recognition of
anions in aqueous and organ ic media furthers our understanding of several physical
phenomena. For example, the recognition or encapsulation of anionic targets using
abiotic receptors serves to advance our understanding of binding forces involved in the
interactions between the guest and the host.' Ion-pairing behavior, hydrogen bonding,
geometric considerations, the interplay of the microenvironment with pK. values, and the
lowering of activation barriers to reactions are areas of particular interest.' Also pertinent
to anion recognition is the development of techniques that signal the binding events
spectroscopically.' Specifically, in our group, the knowledge garnered has been applied
to waste-water analysis, blood analysis, design of artificial enzymes , food and beverage
quality control, and RNA hydrolysis.
Advances in the field of molecular recognition have been well documented through
several review articles.' The structural considerations that are used in the design of an
anion-complexing agent are the size, the charge, and the shape of the anion . Inorganic
an ions, organic anions, and large biological anions are common targets. The
architectures of the receptors can be described as polyfunctionalized macrocycles,"
clefts," tweezers," or bowls," with organic or metallic recognition sites.

• Suzanne L. Tobey and Eric V. Anslyn, Department of Chemistry and Biochemistry, The University of Texas
at Austin, Austin, Texas, USA78712-1167.

Fundamentals andApplU:ations of AnionSeparations, edited by

Moyerand Singh,KIuwer AcademicIPlenum Publishers, NewYork,2004 59

The purpose of this manuscript is to provide a review of the advances made in our
own research group . We focus on the development of anion-recognition ensembles
containing guanidinium groups" with specific applications in RNA hydrolysis, pK.
determinations, inorganic-anion recognition, and selective anion recognition in complex


The mechanism of RNA and DNA hydrolysis by enzymes continues to attract much
attention." The ability of an enzyme to promote RNA hydrolysis relies in part on the
recognition of the phosphodiester linkages and the ribonucleotides in the RNA backbone.
In a similar fashion, abiotic receptors designed to recognize phosphodiesters can aid in
the elucidation of a mechanism for the hydrolysis, with the eventual goal of incorporating
the design into functional artificial enzymes .
Specific to our group is the research performed using crescent-shaped polyaza clefts .
Early investigations involved the synthesis of several polyazaclefts characterized by a
tricyclic core with appendant aminoimidazoline groups (1 and 2). The incorporation of
the guanidinium groups served to mimic the arginine residues that reside in the active site
of staphylococcal nuclease and promote RNA and DNA hydrolysis." 31p NMR titrations
in different DMSOIH20 mixtures indicated that 2 bound dibenzyl phosphate with both


N H eN H
1 2

1:1 and 2: 1 stoichiometries giving affinity constants on the order of 102 M- 1•12 Similar
studies with 3 and 4 were pursued with a focus on the effect of dibenzylphosphoric acid
dimerization on host-guest complexation in chloroform. Although the hydrogen bond-
donor groups on 2 and 3 are different, it was confirmed in both cases that the clefts bound
dibenzyl phosphate more strongly than a control host containing one binding site . An
investigation into the ability of these binding motifs to complex phosphodiesters and
phosphomonoesters offered affinity constants on the order of 103 ~l and 102 M- 1 for 1
Modifications to the ring size of the spacer units and relative stereochemistry
through synthetic means led to an array of clefts, rendering them more rigid or more



3 4

flexible, more open or more closed. These modifications resulted in manipulation of the
guanidinium recognition units such that four-point hydrogen bonding with dibenzyl
phosphate as a surrogate was optimized (1 optimal over 2 and 4 optimal over 3). The
abilities of the clefts to complex dibenzyl phosphate in DMSO/H 20 mixtures were
enhanced by the presence of chloride and reduced by tetraphenylborate counterions."
The trends observed in the affinity constants were attributed to solvation effects of the
complexes in the presence of these counterions. These observations were further
confirmed based on a crystal structure of meso-l with bound dibenzyl phosphate, where a
participating chloride was found in the co-crystal. The physical studies indicated that 2
would a good candidate for RNA hydrolysis investigations. Indeed, 2 was found to give
a 20.7 rate enhancement in the hydrolysis of mRNA when compared to imidazole alone
. a k"metre assay. 16. 17

Current research is focused on elucidating the mechanism of phosphodiester

hydrolysis using a polyaza cleft with two guanidinium moieties and an added metal
center (5) . This will offer insight into any cooperativity, negative or positive, between
the metal center and the guanidinium units. This research is aimed at mimicking the
active sites of enzymes that have both metal centers and arginine residues.



Of significant interest in the arena of molecular recognition is the selective binding

of anions in complex aqueous mixtures. Towards this end, a series of receptors have

been developed for the detection of carboxylate and phosphate-containing natural

products. The design of these hosts is derived from a 1,3,5-trisubstituted-2,4,6-
triethylbenzene scaffold. The alternating ethyl groups on the benzene ring serve to orient
the binding functionalities at the I, 3, and 5 positions to one face of the plane. These
binding sites are chosen to complement functionalities of the intended analyte, while the
preorganization of the binding sites around the scaffold serves to maximize the binding
interactions between host and guest.
In addition to the design of the cavities for anion binding , we sought to develop a
dye-displacement assay to sufficiently and conveniently quantify the association of the
guest to the host. This involves the noncovalent binding of an indicator to a host, such
that the optical properties of the indicator are altered . Upon addition of an analyte, the
indicator is displaced from the binding cavity into solution, where its native optical
properties are restored. Thus, these assays produce color changes, allowing binding
events to be signaled colorimetrically and monitored using UV-vis or fluorescence
Substitution of the aforementioned scaffold with aminodihydroimidazolium groups
in the 1,3,5 positions yielded 6.'8 The binding constants of 6 with several carboxylate
containing analytes were determined by IH NMR, and 6 was found to be selective for
citrate (K. = 6.9 x 103 M- 1 in 0 20), a tricarboxylate. A crystal structure of 6 with
tricarballate bound to the cavity was reported, verifying the orientation of the
guanidinium groups to one face of the benzene plane. Development of a competition
assay in which citrate displaced a cavity-bound dye, 5-carboxyfluorescein, was used to
quantify the binding of citrate to the hOSt.1 9 The assay was successfully used to
determine the citrate content of eight commercial beverages.

6 7

Modifications in the synthesis of receptor 6 led to the replacement of one of the

guanidinium containing groups with an ortho-amino methyl boronic acid to yield 7.2°
The combination of the guanidinium groups for ion-pairing interactions and the boronic
acid for complexing diols makes this receptor amenable to the recognition of poly-
functional analytes. The cavity of 7 was found to be effective in the recognition of
tartrate (K. = 5.5 x 104 M- 1 in H20/MeOH 1:3) and malate (K. = 4.8 x 104 M"l). The
magnitude of the binding constants obtained suggests that the guanidinium and boronic
acid recognition units work cooperatively to effect binding of the analyte. The K. values
were obtained by using a dye-displacement assay employing alizarin complexone in
conjunction with 7. Several wines were tested using this assay, and the tartaric and malic

acid concentrations were accurately determined, as found by comparison to a control

method in which IH NMR studies were used.
In keeping with the substitutions of boronic acid functional groups on the scaffold,
compound 8 has been synthesized and found to be effective in binding phenolic acids."
A competition assay using pyrocatechol violet was used to determine the binding affinity
4 1
of gallate to 8 (K. = 1.0 x 10 M- ) . This receptor scheme was used in the analysis of
scotch whiskies. The collective response of the host to the phenolic acid components of
the whiskies was correlated to their ages.
An alternate cavity design involved the C3-symmetric receptor characterized by the
substitution of three meta-amino methyl boronic acids around the triethylbenzene
scaffold (9).22 31p NMR titrations confirmed complexation of glucose-6-phosphate to 9
in 70:30 MeOH/H 20 with K. = 1.6 x 103 M- 1• Complexation is thought to occur through
reversible covalent bond formation between the alcohols and the boronic acids.


I h

q;f~H ® He..))H
HH- ® H
H ~~, H~ (±)
8 9

Application of a signal-transduction assay for UV-vis spectroscopy served to signal

the binding of glucose-6-phosphate with a high affinity constant (K. = 2.2 x 10 M- 1) in
The triethylbenzene scaffold was used as the basic building block for azacalixarene
10, a form of an expanded calixarene." The dynamic nature of 10 warranted extensive
conformational analyses, the NMR studies of which have been reported. We attribute the
conformational freedom to flexibility and to the lack of intramolecular hydrogen bonds in
aqueous med ia. The predisposed arrangement of four amines on the periphery of the
cavity and the four amines within the interior of the cavity was expected to serve as
binding sites for phosphorylated carbohydrates. There was a significant response of 10 to
the complexation of four different analytes when monitored using dye-displacement
assays. The most significant result was the association of inositol triphosphate (K. = 2.4
x 105 M- 1 in methanol) to the cavity .
An alternate synthetic route yielded a bowl shaped cavity with six recognition
elements on the periphery (11).24 Two compounds were obtained, one in which the R
groups were amines and one in which the R groups were imidazolines. The
64 S. 1.. TOBEY ET AL.

./ R
R!' R_
-' ~;; ",F\.

,y '<:: I

10 11

preorganization imparted by the triethyl benzene scaffold aids in the orientation of the
recognition units towards the interior of the cavity. Fluorescent and UV-vis methods
were used to monitor the displacement of 5-carboxyflourescein from 11 (R =
guanidinium) upon addition of inositol -l,4,5-trisphosphate (IP3 ) . This secondary
messenger bound within the cavity with K.= 4.7 x lOs M- 1 in water. The electrostatic
interactions of the guanidinium groups with the phosphate groups on IP3 are not
significantly altered by the presence of high salt concentrations.
The hexa-substituted benzene 12 incorporates a phenanthroline bound Cu(II) moiety
that provides an internal signal transduction scheme." The complexation of 12 with
citrate is thought to occur through charge-pairing of two carboxylates with the
guanidinium functional groups and coordination of one carboxylate to the copper center.
Photophysical studies on this system and one lacking the guanidinium groups using
fluorescence to monitor the modulation of the metal quenching indicate that the
guanidinium groups and the Cu(II) center work cooperatively in the binding of citrate.
Analysis of commercially available beverages using this sensing ensemble was successful
in quantifying the citrate content.


In recent work, a 2,2':6,2"-terpyridine cleft characterized by a Zn(II) center and two

guanidinium groups (5) was synthesized ." A dye-displacement assay using
pyrocatecholviolet was used to determine the binding affinities of several amino acids to
5. The highest affinity, K. = 1.5 x to' M- 1, was reported for aspartate in 1:1 H20/MeOH .
Similar studies with control hosts confirm that the complexation of aspartate is due to the
cooperativity of the guanidinium groups and the metal center.



Literature precedent describes the stabilization of enolate intermediates by enzymes,

resulting from the formation of hydrogen bonds." An understanding of the intricate
interactions and ramifications of enolate stabilization through hydrogen-bonding
interactions provides the impetus for the research done within our group towards the
complexation of enolates using neutral abiotic receptors.
The design and synthesis of a preorganized polyaza cleft (13) was accomplished."
The cleft is characterized by convergent amide-like hydrogen-bond donors to
complement 1,3-diketo-like enolates. The binding constants (range 1.35 x 104 ~I to
1.75 X 102 M- 1) of 15-crown-5 sodium enolates complexed to 13 in acetonitrile were
determined through I H NMR studies." Additionally, pK. determinations for the
conjugate acids of several enolates (range 20.7-15.9) in acetonitrile were done using
nonaqueous titration methods. Results reveal that more basic enolates have higher
binding affinities to 13. The affinites of the complexes did not rely primarily on the
basicity of the enolates, but on the arrangement of the hydrogen-bond donors and
acceptors . The strength of the binding interactions was attributed to conformational
freedom of the anions, hydrogen-bonding abilities to heteroatoms, and the presence of
counterions. It is also interesting to note that the hydrogen-bonding moieties work
cooperatively in binding, as the affinity constants of the analytes to a control host
containing fewer hydrogen-bond donors were lower. Stabilization of the enolate is
thought to occur through the formation of hydrogen bonds to the lone pairs of the enolate.


Further, pH titration experiments revealed that there was a pK. shift in the diketone
of about 1.0 pK. unit while in the presence of the abiotic polyaza cleft. Additional

titrations involving the addition of a base to a 13/guest solution resulted in a 13/enolate

complex. The presence of the host shifts the equilibrium for deprotonation, thereby
effecting enolate-host complexation in the presence of a base.
An extension of this work involved studies into the stabilization of active methylene
compounds through bonding of the 1t-system of the enolate, rather than through binding
of the lone pair electrons. For this purpose, cavity 14 was developed, taking advantage of
the fact that the six amide protons directed into the interior of the cavity are preorganized
to form hydrogen-bonds with the 1t-system of an enolate as it lies parallel to the benzene
rings of the cavity. Association constants for a series of I5-crown-5 sodium enolate
guests to 14 in 95:5 CD 3CN/CD2Cl2 were determined using IH NMR titrations.'" The
pK. of 2-acetylcyclopentanone in the presence of 14 and 2,4,6-trichlorophenoxide was
2.9 units lower than in the absence of 14. This pK. shift is larger than that determined for
diketones in the presence of 13 . The physical studies indicate that NH-1t hydrogen
bonding is more effective in lowering the acidity of ex-carbon acids than hydrogen
bonding to the carbonyl lone pairs .
Combined, these results highlight the ability to use abiotic receptors to differentiate
the modes by which enolates can be bound using hydrogen bonds. The contrast between
stabilization through lone pairs or through the 1t-system may offer insight into the
behavior of enolate substrates within the active sites of enzymes.25


We have further reported the use of 14 for nitrate recognition." The notable
properties of this particular cyclophane include a rigid structure with six amide protons
directed towards the interior of the cavity. This cyclophane was shown to have utility in
the recognition of anions AcO- > N03- > CN- > Cl"> H 2P04- > Br" > HS0 4- with
affinity constants ranging from 5-770 ~I . The solution-phase stoichiometries of the
host-guest complexes were determined from Job plots using IH NMR data. The
association constants of these anions to the cyclophane were determined through IH
NMR titrations in a 25% CD 2Ch/CD3CN solvent mixture. Also reported in the paper are
the solid -state structures of the cavity, the cavity with chloride bound, and the cavity with
acetate bound .
The ability of the cavity to bind nitrate was incorporated into a dye-displacement
assay to develop a chemosensor for nitrate." A competition assay using resorufin with
14 was used in a 1:1 mixture of MeOH/CH 2Ch and a 3:1 mixture of CH 3CN/CH2Ch.
The assay was monitored for several analytes by UV-vis with nitrate having affinities of
380 ~l and 500 M- 1 respectively.
The results indicate that the cyclophane is selective for acetate and nitrate on the
basis of size, for the larger, more basic anions bind with significantly lower affinities.
The most intriguing result is that of the encapsulation of nitrate, which is believed to
reside within the cavity with six hydrogen-bonds to the amide protons. Nitrate is
geometrically complementary to the interior of the cav ity, and the ability to utilize six
hydrogen bonds is sufficient to overcome the poor coordinating ability of the anion .


In summary, the design, development, and eventual application of macrocyclic

receptors for the recognition of anions provides the opportunity to investigate several
facets of molecular recognition . The above examples demonstrate the successes of
exploiting specific interactions for pK. determinations, phosphodiester cleavage, selective
recognition in complex mixtures, and signal-transduction schemes. A variety of different
scaffolds have been exploited, wherein a main design theme was convergence and
preorganization. These are enthalpy-based considerations. Given the recent discoveries
derived from ITC data that much of ion-pairing molecular recognition in high-dielectric
solvents is driven by entropy, our group now focuses upon both enthalpy and entropy as
important concepts during our initial design stages. Given the new discoveries and host
designs that are increasing in number in the literature for anion recognition, it is clear that
the field is alive and vigorous . Hence, there remains a great deal of host-guest chemistry
to be further explored as more elaborate host designs and signaling methods are pursued.


We would like to gratefully acknowledge the NSF, the NIH, and the Welch
Foundation for continued support of anion recognition in the Anslyn group over the last


I. K. A. Conners in: Binding Constants, The Measurement of Molecular Complex Stability, (John Wiley &
Sons,New York, 1987), pp. &-15.
2. A. J. Kresge, What makes proton transfer fast, Ace. Chem. Res. 8(10),354-360 (1975). R. Kluger, Ionic
intermediates in enzyme-catalyzed carbon-carbon bond formation: patterns, prototypes, probes, and
proposals, Chem . Rev. 90(7), 1151-1169 (1990). W. W. Cleland and M. M. Kreevoy, Low-barrier
hydrogen bonds and enzymic catalysis, Science 264(5167), 1887-1890 (1994). J. A. Gerlt and P. G.
Gassman, Understanding the rates of certain enzyme-catalyzed reactions: Proton abstraction from carbon
acids, acyl transfer reactions, and displacement reactions of phosphodiesters, Biochemistry 32(45),
11943-11952(1993). I. P. Petrounia and R. M. Pollack, Substituent effects on the binding of phenols to
the D38N mutant of 3-0xot.5-stero id isomerase. A probe for the nature of hydrogen bonding to the
intermediate, Biochemistry 37(2), 700-705 (1998). E. Kimura, H. Kitamura, T. Koike, and M. Shiro,
Facile and selective electrostatic stabilization of uracil N(I)-anion by a proximate protonated amine: A
chemical implication for why uracil N(I) is chosen for glycosylation site, J. Am. Chem. Soc. 119(45),
10909-10919(1997). G. A. Jeffery in: An Introduction 10 Hydrogen Bonding, (Oxford University Press,
New York, 1997).
3. T. S. Snowdenand E. V. Anslyn, Anion recognition: Synthetic receptors for anions and their application in
sensors, Current Opinion in Chemical Biology 3, 740-746 (1999).
4. J. H. Hartley, T. James, and C. J. Ward, Synthetic receptors, J. Chem. Soc.. Perkin Trans. 119, 3155-3184
(2000). M. M. G. Antonisse, D. N. Reinhoudt, Neutral anion receptors: Design and application, Chem .
Comm. 4, 443--448 (1998). P. Beer and P. A. Gale, Anion recognition and sensing: The state of the art
and future perspectives, Angew. Chem. Int. Ed. 40(3),48&-516 (2001). F. P. Schmidtchen and M. Berger,
ArtificialOrganic Host Molecules for Anions, Chem. Rev. 97(5), 1609-1646(1997). T. Schrader,Chelate
complexes with the P=O double bond i\ a New Concept for Molecular Recognition, J. of Inclusion
Phenom. Macro. Chem. 34(2), 119-131 (1999).
5. J.-M. Lehn, E. Sonveaux and A. K. Willard, Molecular recognition. Anion cryptates of a macrobicyclic
receptor molecule for lineartriatomicspecies,J. Am. Chem. Soc. 100(15),4914--4916(1978). P. D. Beer,

F. Szemes, V. Balzani, C. M. Sala, M. G. B. Drew, S. W. Dent, and M. Maestri, Anion selective

recogn ition and sensing by novel macrocyclic transition metal receptor systems. IH NMR,
electrochemical, and photophysical investigations,J. Am. Chem. Soc. 119(49), 11864-11875 (1997). A.
Andrievsky, F. Ahuis, J. L. Sessler, F. Vogtle, D. Gudat, and M. Moini, Bipyrrole-Based[2)Catenane: A
new type of anion receptor, J. Am. Chem. Soc. 120(37),9712-9713 (1998). Y. Inoue, T. Hakushi, Y. Liu,
L.-H. Tong, B.-J. Shen, and D.-S. Jin, Thermodynamics of molecular recognition by cyclodextrins. I.
Calorimetric titration of inclusioncomplexation of napthalenesulfonates with c- , ~-, and y-cyclodextrins:
enthalpy-entropy compensation,J. Am. Chem. Soc. 115(2), 47~81 (1993).
6. C. Rapso, N. Perez, M. Almaraz, M. L. Mussons, M. C. Caballero, and J. R. Moran, A cyclohexane spacer
for phosphate, Tetrahedron Leu. 36(18),3255-3258 (1995). V. Hedge, P. Madhukar, 1. D. Madura, and
R. P. Thummel, Fischerroute to pyrido[3,2-g)indoles. A novel receptor for urea derivatives, J. Am. Chem.
Soc. 112(11), 454~550 (1990). V. Hedge, C. Y. Hung, R. Cummingham, T. Hopfiner, and R. P.
Thummel, Design of receptorsfor urea derivatives based on the pyrido[3,2-g)indolesubunit, J. Am. Chem.
Soc. 115(3),872-878 (1993). T. W. Bell, Z. Hou, S. C. Zimmerman, and P. A. Thiessen, Highly effective
hydrogen-bonding receptors for guanine derivatives, Angew. Chem. In/I. Ed. Engl. 34(19), 2163-2165
(1995). T. W. Bell, N. M. Hext, and A. B. Khasanov, Binding biomolecules with designed hydrogen-
bonding receptors, Pure Appl. Chem. 70(12), 2317-2377 (1998). J. Cuntze, L. Owens, V. Alcazar, P.
Seiler, and F. Diederich, Molecular clefts derived from 9,9'-spirobi-9H-fluorene for enantioselective
complexation ofpyranosides and dicarboxylic acids, Helv. Chim. Acta. 78(2),367-390 (1995).
7. M. Mei and S. Wu, Fluorescentsensor for a , n-dicarboxylate anions, New J. Chem. 25(3),471-475 (2001).
H. Wennemers, M. Conza, M. Nold, and P. Krattiger, Diketopiperazine receptors: A novel class of highly
selective receptors for binding small peptides, Chem Eur. J. 7(15), 3342-3347 (2001). R. 1. Jansen, R.
Gelder, A. E. Rowan, H. W. Scheeren, and R. J. M. Nolte, Molecular clips based on propanediurea.
exceptionally high bindingaffinities for resorcinol guests, J. Org. Chem. 66(8), 2643-2653 (200I). S. K.
Lee, S. Y. Shin, S. Lee, C. Lee, and J. W. Park, Conformation and binding properties of polymethylene
linked bis-viologens-2-napthol complexes, J. Chem. Soc., PerkinTrans. 2 10, 983-1988 (2001). B. H. M.
Snellink-Ruel, M. M. G. Antonisse, J. F. J. Engbersen, P. Timmerman, and D. N. Reinhoudt, Neutral
anion receptorswith multipleurea-binding sites, Eur. J. Org. Chem. 2000(I), 165-170 (2000).
8. L. Sebo, F. Diederich, and U. Gramlich,Tetrakis(phenylamidium)-substituted resorcin[4)arene receptors for
the complexation of dicarboxylates and phosphates in protic solvents, Helv. Chim. Acta. 83( I), 93-113
9. For a few other guanidinium based receptors see: R. P. Dixon, SJ . Geib, and A. D. Hamilton, Molecular
recognition: Bis-acylguanidiniums providea simple family of receptorsfor phosphodiesters. J. Am. Chem.
Soc. 114,365-366 (1992). M. Berger, F. P. Schmidtchen, Zwiterionic guanidiniumcompounds serve as
electroneutralanion hosts. J. Am. Chem. Soc. 121(43),9986--9993 (1999). F. Cuevas, S. D. Stefano, J. O.
Magrans, P. Prados, L. Mandolini, and J. de Mendoza, Toward an artificial cetylcholinesterase. Chem.
Eur. J., 6(17), 3228-3234 (2000). Y.Kato, M. Conn, and J. Rebek,Jr. Water-soluble receptors for cyclic-
AMP and their use for evaluating phosphate-guanidinium interactions, J. Am. Chem. Soc. 116(4),
3279-3284 (1994). L. J. Lawless, A. G. Blackburn, A. J. Ayling, M. Perez-Payan, and A. P. Davis,
Steroidal guanidines as enantioselective receptors for N-acyl a-amino acids. Part I. 3a-Guanylated
carbamates derived from cholic acid. J. Chem. Soc., Perkin Trans. I , 1329-1341 (2001). M. Bonnat, M.
Bradley, and J. D. Kilburn, The solid phase synthesis of a guanidinium based 'tweezer ' receptor.
Tetrahedron Lell. 37(30) 5409-5412 (1996). K. Ariga, A. Kamino, X. Cha, and T. Kunitake, Multisite
recognition of aqueous dipeptides by oligoglycine arrays mixed with guanidinium and other receptor units
at the air-water interface, Langmuir 15(11),3875-3885 (1999). T. Kajiki, H. Moriya, K. Hoshino, T.
Kuroi, S. Kondo, T. Nabshima, and Y. Yano, Functionalized flavin receptors, regulation of redox
properties of 6-azaflavin via hydrogen bondingswith melaminederivativesbearing guanidinium ion(s) in
organicsolvents. J. Org. Chem. 64(26),9679-9689(1999).
10. C. B. Anfinsen,Principlesthat govern the foldingof proteinchains, Science 181(4096),223-230 (1973). F.
A. Cotton, V. W. Day, and E. E. Hazen, Jr., Structure of bis(methylguanidinium) monohydrogen
orthophosphate. Model for the arginine-phosphate interactions at the active site of staphylococcal nuclease
and other phosphohydrolytic enzymes, J. Am. Chem. Soc. 96(14), 4471-4478 (1974). J. K. Bashkin,
Introduction to RNAIDNA cleavage, Chem. Rev. 98(3), 937-938 (1998).
II. F. A. Cotton, E. E. Hazen, Jr., and M. J. Legg, Staphylococcal nuclease: Proposed mechanism of action
based on structureof enzyme-thymidine 3',5'-bisphosphate-calcium ion complexat 1.5 A resolution, Proc.
Natl. Acad. Sci. U. S.A. 76(6),2551-2555 (1979).

12. K. Ariga and E. V. Anslyn, Manipulating the stoichiometry and strength of phosphodiester binding to a
bisguanidine cleft in DMSO/water solutions,J. Org. Chem. 52(7),417-419 (1992).
13. F. Chu, L. S. Flatt, and E. V. Anslyn, Complexation of phosphoric acid diesters with polyaza-c1efts in
chloroform: Effects of phosphodiester dimerization, changing cavity size, and preorganizing amine
recognition units,J. Am. Chem. Soc. 116(10),4194-4204 (1994).
14. D. M. Perreault, X. Chen, and E. V. Anslyn, The advantages of using rigid polyaza-clefts for hydrogen-
bondingmolecularrecognition, Tetrahedron 51(2), 353-362 (1995).
15. D. M. Kneeland, K. Ariga, V. M. Lynch, C. Huang, and E. V. Anslyn, Bis(alkylguanidinium) receptors for
phosphodiesters: Effects of counterions, solvent mixtures, and cavity flexibility on complexation, J. Am.
Chem. Soc. 115(22), 10042-10055(1993).
16. E. V. Anslyn, J. Smith, D. M. Kneeland, K. Ariga, and F. Chu, Strategies for phosphodiestercomplexation
and cleavage,Supramol. Chem. 1,201-208 (1993).
17. D. M. Perreault, L. A. Cabell, and E. V. Anslyn, Usingguanidinium groups for the recognitionof RNA and
as catalysts for the hydrolysis of RNA. Bioorg. Med. Chem. 5(6), 1209-1220(1997).
18. A. Metzger, V. M. Lynch, and E. V. Anslyn, A synthetic receptor selective for citrate, Angew. Chem. Int.
Ed. Engl. 36(8), 862-864 (1997).
19. A. Metzger and E. V. Anslyn, A chemosensor for citrate in beverages, Angew. Chem. Int. Ed. 37(5)
649-652 (1998).
20. J. J. Lavigne and E.V. Anslyn, Teaching old indicatorsnew tricks: A colorimatericchemosensingensemble
for tartrate/malate in beverages, Angew. Chem.Jnt. Ed. 38(24), 3666--3669 (1999).
21. S. L. Wiskur and E. V. Anslyn, Using a synthetic receptor to create an optical sensing ensemble for a class
of analytes: A colorimetric assay for the aging of scotch, J. Am. Chem. Soc. 123(41), 10109-10110
22. L. A.Cabell, M. C. Monahan,and E. V. Anslyn, A competition assay for determining glucose-e-phosphate
concentration with a tris-boronic acid receptor, Tetrahedron. Lell. 40, 7753-7756 (1999).
23. K. Niikura and E. V. Anslyn, Azacalixarene: synthesis,conformationalanalysis, and recognition behavior
toward anions,J. Chem. Soc.. PerkinTrans. 2. 2769-2775 (1999).
24. K. Niikura, A. Metzger, and E. V. Anslyn, Chemosensor ensemblewith selectivity for inositol-triphosphate,
J. Am. Chem. Soc. 120(33),8533-8534 (1998).
25. L. A. Cabell, M. D. Best,J. J. Lavigne, S. E. Schneider, D. M. Perreault, M. C. Monohan,and E. V. Anslyn,
Metal triggered sensing of citrate using a synthetic receptor, J. Chem. Soc.• Perkin Trans. 2. 315-323
26. H. Att-Haddou,S. L. Wiskur, V. M. Lynch,and E. V. Anslyn, Achieving large color changes in response to
the presence of amino acids: A molecular sensing ensemble with selectivity for aspartate, J. Am. Chem.
Soc.. 123(45), 11296--11297 (200I).
27. J. D. Robertus, J. Kraut, R. A. Alden, and J.J . Birkroft, Subtilisin. Stereochemical mechanism
involvingtransition-state stabilization, Biochemistry 11(23),4293-4303 (1972). J. A. Gerlt and P. G.
Gassman, Understanding enzyme-catalyzed proton abstraction from carbon acids: details of stepwise
mechanism for ~oelimination reactions, J. Am. ChemSoc. 114(15),5928-5934 (1992). M. M. Benning,
K. L. Taylor, R.-Q. Liu, G. Yang, X. Hong, G. Wesenberg, D. Dunaway-Mariano, and H. M. Holden,
Structure of 4-chlorobenzoyl coenzyme A dehalogenase determined to 1.8 A resolution: An enzyme
catalyst generatedvia adaptivemutation, Biochemistry 35(25), 8103-8109 (1996).
28. A. M. Kelly-Rowley, L. A. Cabell, and E. V. Anslyn, Enolate complexation in Acetonitrile with a neutral
polyaza-c1eft, J. Am. Chem. Soc.. 113(25),9687-9688 (1991).
29. A. M. Kelly-Rowley, V. M. Lynch, and E. V. Anslyn, Molecular recognition of enolates of active
methylene compoundsin acetonitrile. The interplaybetweencomplementarity and basicity, and the use of
hydrogenbondingto lowerguest pKa's, J. Am. Chem. Soc.• 117(12), 3438-3447 (1995).
30. T. S. Snowden, A. P. Bisson, and E. V. Anslyn, A comparisonofNH-n versus lone pair hydrogen bonding
effects on carbon acid pKa shifts, J. Am. Chem. Soc. 121(26),6324-6325 (1999).
31. A. P. Bisson, V. M. Lynch, M. C. Monohan, and E. V. Anslyn, Recognition of anions through NH-n
hydrogen bonds in a bicyclic cyclophane-Selectivity for nitrate, Angew. Chem. Int. Ed.• 36(21)
2340-2342 (1997).
32. K. Niikura, A. Bisson, and E. V. Anslyn, Optical sensing of inorganicanions employinga synthetic receptor
and ioniccolorimetricdyes,J. Am. Chem. Soc.. PerkinTrans. 2. 111-1114(1999).

Jonathan L. Sessler," Bruno Andrioletti, Pavel Anzenbacher, Jr.,

Chris Black, Leah Eller, Hiroyuki Furuta, KarolinaJursikova,
Hiromitsu Maeda, Manuel Marquez, Toshihisa Mizuno, and AndrewTry


Related to the general problem of anion separation is the underlying challenge of

anion recognition and sensing.l" In recent years increasing attention has been devoted to
this aspect of supramolecular chemistry, and a number of elegant approaches have been
described.t'" However, in spite of considerable effort, simple molecular systems capable
of sensing effectively and selectively such classic inorganic species as fluoride, chloride,
and phosphate anions remain rare. 27-40 Part of the problem is that many of the recognition
motifs currently being studied in the context of anion sensing are either highly charged,
require pre-protonation, or are based on substructures that contain not just hydrogen bond
donor elements but also interfering hydrogen bond acceptor functionality. Additionally,
efforts to incorporate a read-out element into the overall sensor design have usually come
at the cost of increased structural complexity and reduced synthetic accessibility.
Because of this, we have focused our attention on pyrrole-based receptor systems33-46 and
have recently found that 2,3-dipyrrol-2'-ylquinoxaline (DPQ; 1; Scheme 1) acts as an
effective "naked eye" detectable sensor for fluoride anions in nonpolar solvents.35•39 In
this chapter, we review our efforts to develop DPQs as anion sensors giving special
emphasis to recent work designed to fine-tune the inherent specificity and selectivity of
this easy-to-prepare receptor system .47•50

• Jonathan L. Sessler, Leah Eller. Bruno Andrioletti, Chris Black, Toshihisa Mizuno, and Andrew Try,
Departmentof Chemistry and Biochemistry, Institute for Cellular and Molecular Biology, The University of
Texas, Austin, TX 78712. Pavel Anzenbacher, Jr., and Karolina Jurslkova, Department of Chemistry and
Center for Photochemical Sciences, Bowling Green State University, Bowling Green, OH 43403. Manuel
Marquez, Kraft Foods Research and Development Center, The Nanotechnology Laboratory, 801 Waukegan
Road, Glenview, Illinois 60025. HiroyukiFuruta and Hiromitsu Maeda, Departmentof Chemistry, Graduate
Schoolof Science, KyotoUniversity, Kyoto606·8502, Japan.

Fundmrrmt41s andApplications ofAnionSeparations, edited by

Moyer and Singh, KluwerAcademic/PlenumPublishers,New York, 2004 71


2,3-Bis(lH-pyrrol-2-yl)quinoxaline 1 is not a new compound. It was first reported

by Oddo in 1911 51 and was prepared subsequently by Fischer52 and Behr.53 However, its
utility as a possible anion sensor was not recognized. We were attracted to this system
because it contains two pyrrolic entities that could serve as potential NH hydrogen bond
donors (i.e., anion recognition motifs) and a bridging quinoxaline subunit that could serve
as a built-in colorimetric sensing core. The fact that DPQ itself, and potentially a range
of derivatives, could be made in two steps from a variety of readily accessible,
commercially available precursors as shown in Scheme I provided an added incentive to
study DPQs as potential anion sensors."

.. :JicR----.. . .. R-R.
9. R=H

10. R=F

, NH2

I ... NH2 1. Rl = A2=R3=R.=Rs=H

2. R, = F.R 2= R3 = R.. = Rs =H
3. R, = F,R 2= N0 2• R3 = R.. = Rs= H
4. R, = ~ = R3 =R. = H, Rs= Br
7. R, = A2= R3 =Rs = H,R.= N(h
8. R1 = ~= R. = Rs = H,R 2= N0 2
9. R1 =R..=Rs=H.R2=~=N~

5 8

Scheme 1. Synthesis of dipyrrolylquinoxaline (DPQ) derivatives.

As we originally envisioned it, interactions between an anionic substrate and the

pyrrolic portions of a DPQ derivative would serve to modify the orbital overlap between
these subunits and the quinoxaline core, thus changing the electronics of the system as a
whole as well as the color of this latter chromophore. While this mechanism of action,
shown schematically in Figure I, remains subject to debate,§ it is at least consistent with
the observat ion that the mono-SEM protected derivative of 1 (2-(lH-pyrrol-2-yl)-3-(l-

I Interestingly. X-raydiffraction analy... of the letrabulylammonium fluoride(TBAF)complexof DPQ. 1. reveal that, in the solid state, the
fluoride anionis boundbetween a pairofpyrroleresidues coming fromtwoseparate molecules of 1. A molecule of water i. also foundboundto
eachF anion. Thenet resultis an infinite chainin whichthe TBAcationsdo notparticipate in anionbinding.

trimethylsilylethoxymethylpyrrol-2-yl)quinoxaline) failed to act as an effective anion

sensor under conditions where DPQ itself does. Specifically, it was found that dilute
solutions of DPQ 1 in dichloromethane (ca. I mM), underwent a change in color from
yellow to orange when exposed to 10 molar equiv. of tetrabutylammonium fluoride
(TBAF) whereas its mono-SEM derivative did not, even in the presence of 100 molar
equiv . of fluoride anion ." Supporting the conclusion that interaction with both pyrrolic
NH groups was critical for effective interaction were optical-based measurements of
binding affinity; these latter revealed a two-order of magnitude difference in the K.
values in dry dichloromethane (for 1:1 F" binding, K. = 18200 M· I vs. 120 M· 1) .35

~ H
H -"
> ~.
,@ 'N-H
Conformation I
Electronic Change

Figure I. Proposed mechanism of anion-induced colorimetric response showing in schematic fashion the
changes in orbital overlap that couldresult from anion binding. Alternative mechanisms involving complete or
partial proton transfer are also possible.

The above findings led to the consideration that the fluoride anion affinity of DPQ 1
could be increased by appending electron withdrawing groups onto either the quinoxaline
core or the pyrrolic subunits, since such changes were expected to increase the acidity of
the pyrrole NH groups thus making them better hydrogen bond donors, In order to test
the validity of this hypothesis and to ascertain what effect, if any, such changes would
have on the inherent fluoride selectivity of 1, a range of DPQ derivatives with varying
substituents was prepared. Specifically, in accord with the general procedure of Scheme
I, 2,3-bis(3,4-difluoro-1 H-pyrrol-2-yl)quinoxaline 2,39 6-nitro-2,3-bis(3,4-difluoro-1 H-
pyrrol-2-yl)quinoxaline 3, 2,3-bis(l H-pyrrol-2-yl)benzo[g]quinoxaline 5, 2,3-bis(lH-
pyrrol-2-yl)dibenzo[f,h]quinoxaline 6, 5-nitro-2,3-bis( IH-pyrrol-2-yl)quinoxaline 7, 6-
nitro-2,3-bis(IH-pyrrol-2-yl)quinoxaline 8,39 and 6, 7-dinitro-2,3-bis(1 H -pyrrol-2-
yl)quinoxaline 9 were synthesized in one step from I ,2-di( IH-pyrrol-2-yl)ethane-1 ,2-
dione 10 or I ,2-bis(3,4-difluoro-1 H -pyrro l-2-yl)ethane- 1,2-dione 1 1 and the
corresponding commercially available 1,2-aromatic diamine derivative. The 2,3-bis(5-
bromo-I H-pyrrol-2-yl)-6-nitroquinoxaline 4 was prepared by treating 1 with NBS (N-
Once in hand, compounds 2-9 were evaluated as potential naked-eye anion sensors
by noting the color changes induced when fluoride, chloride, and dihydrogenphosphate
anions were added (10 molar equiv. in the form of their tetrabutylammonium salts) to 0.1
mM dichloromethane solutions of the receptors in question. Whereas under these
conditions, the parent system 1 displays a faint but noticeable colorimetric response in the
presence of fluoride anion, the nitro-derivatives 7-9 undergo significant yellow-to-deep-
purple color change in the presence of fluoride anion, and, to slightly lower extent also

with dihydrogenphosphate (Figure 2). The color itself, however, is not bright. By
contrast, compound 3 shows much brighter coloring compared to its non-halogenated
congener 8. In fact, all the halogenated derivatives, in particular 3 and 4, display a sharp
yellow-to-orange/red color change in the presence of both fluoride and dihydrogen-
phosphate anions. The brightness of color is a critical factor in the design of naked-eye
sensors where the visual aspect of the sensing process plays a crucial role.
More quantitative analyses of the anion-binding ability of several of these sensors
were made by observing the extent to which the fluorescence intensity of the sensors in
question was quenched in the presence of anions or by monitoring the change in
absorption induced by the addition of anions. The resulting binding affinities, K.,
representing the thermodynamic binding constants for I: I interaction (as supported by
Job plots), are listed in Table I ;t those for I , 2, and 8 have been published previously.35,39
The results illustrated in Figure 2 and summarized in Table 1 serve to highlight the
fact that appending fluoro- or bromo-substituents onto the pyrrolic portions of the DPQ
derivatives results in sensors that are both easier to visualize and which display increased
affinity for both fluoride and dihydrogenphosphate anions. However, the result of this
enhancement in affinity for both species is a net reduction in the inherent selectivity for
fluoride anion as compared to the parent system 1.

Table 1. Affinity constants (K. in M· l ) for dipyrrolylquinoxaline sensors and typical

anionic substrates as determined from the anion-induced change in fluorescence or
absorption as recorded in dry dichloromethane at 20 °C.a

1 2 3 4 5 8
18200 61600 >200000 60000 11700 118000
60 17300 150000 26000 2600 80
50 <500 2020 1400 1800 65
Anionswere added in the fonn of their corresponding tetrabutylammonium salts. The tabulatedaffinities
were derived from fits performed usingsingle reciprocal plots. Errors are generally<15%.

Increases in affinity can also be induced by introducing nitro substituents onto the
quinoxaline core. However, it is systems that combine both approaches, pyrrole halogen
substitution and quinoxaline nitro group functionalization (e.g., 3) that show both the best
visualization as well as the greatest increase in overall binding affinity. Interestingly,
annulated, polyaromatic systems, such as 5, a species that does not display greatly
increased affinities relative to I, appear very attractive as sensors. They clearly provide a
very bright and easily discernible colorimetric response in the presence of fluoride anion.

t While the K. values determinedfor the various DPQ derivatives were found to bereproduciblefrom run to run under a given set of experimental
conditions (e g. TBAf·JH,O, dryCH,CI,). theywereroundto varyas a functIon or solventpurity and salt hydration This proved particularly
true in the case of the nitre-substitutedsystems (eg.• 3 and I) ; here, large variations in therecorded K. values were found While the subject of
current Investigations, this observation is presently rationahzed In tenns of the higher NH acidity of these systems that could allow for color
changesdue coanion-induceddeprctonationof the CPQ core
Figure 2. Color changes induced by the addition of fluoride, chloride , and dihydrogenphosphate anions to
dichloromethane solutions (0.1 mM) containing various dipyrrolyquinoxaline (DPQ) derivatives . The anions
(10 molar equiv . in all cases) were added in the form of their tetrabuty lammonium salts. Within each group, the
vials are arranged in the order of blank, fluoride, chloride and dihydrogenphosphate (left to right), whereas the
overall sequence is (left to right) 1, 2, 3 (top row); 4, 5, 6 (middle row); and 7, 8, 9 (bottom row).

Table 2. Affinity constants (K, in Mol) for metal-containing

dipyrrolylquinoxaline sensors 15 and 16 and typical anionic
substrates as determined from absorption spectral changes
recorded in dry DMSO. a
1 2 12 15 16
P- <100 59000 440 12000 54000
cr NDc NDc NDc 10 20
c c
HZP0 4- ND 5000 ND 40 50
Anions wereused in the fonn of theirtetrabutylammonium (TBA) salts.
Estimate; clean binding profiles werenot observed.
C ND = not determined.
76 J. 1.. SESSLER ET AL.


Annulated systems such as 5 and 6, in addition to their utility as sensors per se,
provide a jumping-off point for further elaborations involving the DPQ core. While a
number of directions could be conceived within the context of this paradigm, one that
piqued our interest was the possibility of generating bipyridine- or phenanthroline-fused
DPQ derivatives that might show enhanced anion affinities as the result of metal
coordination. As a test of this hypothesis, the phenanthroline-containing dipyrrolyl-
quinoxaline ligand system 12 was prepared by reduction of the dinitro species 9 and
reaction of the resulting diamine 13 with 1,IO-phenanthroline-5,6-diketone 14 (cf.
Scheme 2).47 Metalation with bis-bipyridine ruthenium(II) dichloride monohydrate and
bis-phenanthroline cobalt(III) trichloride monohydrate then produced the corresponding
di- and tricationic Ru(II) and Co(III) complexes 15 and 16.
Based on simple electrostatic arguments, illustrated schematically in Figure 3, it was
expected that the Co(III) complex 16 would display a greater affinity for anions than
theRu(II) complex and that both would display affinities greater than either the parent
system 1 or the free ligand from which they were derived, 12. Quantitative studies,
carried out in DMSO for reason of solubility, confirmed this expectation. Indeed, as
shown in Table 2, both complexes bind fluoride anion in a 1:1 fashion and display
affinities that are not only substantially increased relative to these "controls" but also
quite high in absolute terms given the competitive nature of the solvent. Not surprisingly,
therefore, both complexes (as dilute solutions in DMSO) were found to undergo a color
change, from red-pink to pale purple, when treated with ca. 3 equiv. of TBAF. Of
interest, however, and as yet not fully understood, is the observation that systems 15 and
16 did not undergo a color change when treated with tetrabutylammonium
dihydrogenphosphate (even in large excess) and, indeed, displayed but poor affinities for
this particular anionic substrate. This high selectivity is noteworthy, as it could provide
the basis for generating highly effective fluoride-specific anion sensors.


An alternative means of augmenting the affinity of DPQ-type receptor systems has

recently been studied in our group. It involves the use of systems, such as 17 and 18, that
contain built-in pyrrolic "claws" (cf., Scheme 3).49 Based on the observation that mono-
SEM protection of the parent system 1 served to reduce the fluoride anion affinity
substantially, it was thought that providing additional pyrrolic NH hydrogen bond donors
might lead to specific anion "chelation" and hence greatly increased anion binding


) -(
Pd/e . r 9. R = NOz
NHzNH z t..- 13. R = NHz 12

Scheme 2. Synthesis of phenanthroline-fused dipyrrolylquinoxalines.


Ema"'.d Al finty

Figure 3. Schematic representation showing how metal complexation may be used to increase the anion
affinity of DPQ-type receptors and sensors.

1) NaBH..tEtOH

2) TFAlpyrroie

POCI3. (1. R= H
DMF 19. R = CHO


Scheme 3. Synthesis of quinoxalines bearing oligopyrrole substituents.

As in the case of the quinoxaline derivatives bearing only two pyrroles, treating
dilute dichloromethane solutions of 17 and 18 with tetrabutylammonium fluoride
(TBAF), caused the color of the solution to change from yellow to red in both cases.
However, in contrast to many of these simpler systems, color changes were also induced
upon the addition of chloride and dihydrogenphosphate anions (as the TBA salts). Good
isosbestic behavior was observed in all cases when the additions were carried out in an
incremental fashion. This observation, supported by UV-vis spectroscopic Job-plot
analyses in several cases, led to the conclusion that receptors 17 and 18 were binding
these anions in the form of I: 1 complexes, at least under the conditions of the UV-vis

Table 3. Affinity constants (KI in M·I ) for anion binding by the

quinoxaline derivatives 17 and 18, as determined from absorption
spectral changes seen in dry CH2Ch upon the addition ofTBA salts.
II 17 18
F' 18200 32000 >1000000
cr 50 550 5800
H2P0 4' 60 4300 300000
KF·/Kcl· 360 58 >170
KF"/KH2POf 300 7.4 >3.3
Measured via fluorescence quenching; seeTable I.


bindingmodi. binding modi b

Scheme 4. Proposed anion binding modes for receptor 18.

More detailed quantitative analyses of the binding, summarized in Table 3, supported

the conclusion that systems 17 and 18 could be used to complex both chloride and
dihydrogen phosphate anions and confirmed, as expected, that the hexapyrrolic system 18
was the far better receptor . However, these same analyses revealed that the relative
increase in binding was greater in the case of chloride and dihydrogenphosphate than in
the case of fluoride . Thus, while adding "extra" NH donor groups certainly serves to
increase the affinity for fluoride anion (18 is a stronger receptor for F- than 17, which is,
in tum, a stronger receptor for F" than 1), the enhancement caused by these groups is
particularly notable in the case of HZP04- and, to a lesser extent, cr. These findings were
interpreted in terms of these latter, larger anions being better able to fit within an NH-rich
"pocket" formed by the pyrrolic "claws". Specifically, it is proposed that these anions in
particular are bound via binding mode b in Scheme 4, rather than via some less-well
chelated motifs, as exemplified by binding mode a in this same scheme . Specific
support for these structural conclusions was provided by IH NMR spectroscopic analyses,
which revealed both a lack of pyrrole NH signal splittings and shifts in these resonances
to lower field. This is as expected for a highly symmetric, chelated binding arrangement.


Another time-honored approach to improving the selectivity and affinity of a given

receptor motif is to incorporate it into a macrocycle. In the case of
dipyrrolylquinoxalines, the resultant systems would be of intrinsic interest because they
would bear resemblance to various porphyrin analogues, such as expanded porphyrins,"
calixpyrroles.t'i" and calixphyrins40,ss-s7 that have been introduced into the literature
recently. With pairs ofpyrroles linked via what can be thought of as an annulated ethene
bridge, they would also bear an interesting resemblance to Vogel's porphycenes. s8-60
Accordingly, considerable effort is currently being devoted to generating macrocyclic
systems containing dipyrrolylquinoxaline subunits.

5.1. Quinoxpyrroles

The first approach to generating macrocycles containing DPQ subunits was inspired
by recent work with calixpyrroles and calixphyrins. These systems, which act as efficient
anion receptors, may be made readily by condensing pyrrole itself or appropriate
oligopyrrolic fragments (e.g., dipyrromethanes) with .acetone (or other ketones) to
generate macrocycles with one or more Sp3-hybridized bridging, meso-like carbon atoms.
Accordingly, dipyrrolylquinoxaline 1 was subject to condensation with acetone. This
produced a range of products, including the hexapyrrolic macrocycle 20 (Scheme 5).
This system, because it contains Sp3-hybridized, acetone-derived bridging elements, was
termed quinox[3]pyrrole to underscore its analogy to the calix[n]pyrroles.44,45 While
preliminary tests showed that 20 had promise as an anion receptor," all efforts to obtain
single crystals suitable for X-ray diffraction analysis failed. Thus, further effort, perhaps
employing different dipyrrolylquinoxaline or ketone precursors, will be required to
develop fully the promise embodied in this system.

5.2. Quinoxphyrins

An alternative approach to generating dipyrrolylquinoxaline-containing macrocycles

involves using pentafluorobenzaldehyde as the carbonyl source . This precursor has
proved highly effective in allowing the construction of several interesting porphyrinoids,
including corrolesf and calix[6]phyrins.4o It was thus condensed with 1 to provide, after
oxidation, macrocycle 21, as well as several other products. While not yet subject to X-
ray diffractions analysis, compound 21 is of interest because it contains only Sp2_
hybridized bridging elements. As such, it is expected to act as a ligand for cation
coordination, as well as, potentially, a highly effective, preorganized anion receptor.

acatone. TFA


Scheme 5. Synthesis of quinox[3]pyrrole, 20.


+ 2.DDQ,RT
+ products

f F


Scheme 6. Synthesis of quinoxphyrin, 21.

5.3. Quinoxaline-Bridged Schiff-Base Porphyrinoids

Schiff base-bridged systems have a storied history in macrocyclic chemistry . The

underlying synthetic strategy, involving the generation of imine bonds, is one that we and
others have used with success to generate a range of so-called expanded porphyrins ."
Viewing dipyrrolylquinoxalines moieties as "partial porphycene" fragments as discussed
above, led us to consider that application of this strategy in the present instance could
lead to a range of novel porphyrinoids. As an initial test of this hypothesis, the DPQ
dialdehydes 19 were condensed with 1,8-diaminoanthracene in the presence of an acid
catalyst as shown in Scheme 7.* The resulting macrocyclic products, 22, were


19•• R = H 22•. R . H
19b. R = OCH3 22b. R=OCH 3
tse, R = OCH2CH3 22e. R = OCH2CH3

Scheme 7. Synthesis of the quinoxaline-bridged porphyrinoid, 22.

t Corresponding condensations carried out using hydrazine instead of l,S·anthraccnediamine gave rise to productsthat, based on IH NMR
spectroscopic and elMS analyses. wereconsideredto be analogous [2+2] dieza-bridged macrocycles. Unfortunately. poorsolubility bas so far
precluded a definitivestructural assignment

characterized by X-ray diffraction analysis in the case of 22b and 22c and were found to
adopt 1t-stacked conformations in the solid state (cf., Figure 4) .48 Follow-up IH NMR
spectroscopic studies provided support for the conclusion that this organized structure,
with two potential anion binding pockets, was retained in CDC!) solution.
In accord with might be inferred from the "two-pocket" nature of its structure,
system 22 was found to act as a ditopic receptor, binding two molecular equivalents of
fluoride or dihydrogenphosphate anion when challenged with the TBA salts of these two
anions in dichloromethane. In the particular case of 22a, which bears the greatest
structural analogy to unsubstituted dipyrrolylquinoxaline, 1, Hill coefficients of ca. 2
were seen for both F and H2P04••48 The square roots of the associated Kequil. values,
corresponding to hypothetical 1:1 anion-receptor binding, were calculated to be 3 x io'
and 80 M· I , respectively. These values are increased relative to the simple control DPQ
system 1 (K. = 1.82 x 104 and 60 M'I for F and H2P04' , respectively"), as would be
expected given the preorganized nature of the pocket and the presence of multiple NH
donor groups. Not surprisingly, therefore, dilute solutions of22a in 10% DMSO/CH 2Ch
(v/v) could be induced to undergo a color change (from pale lemon-yellow to red) in the
presence of both F' and H2P04' , with greater concentrations of the latter anion being
needed to observe an equivalent effect «20 equiv. VS . >300 equiv.). This result
highlights the inherent fluoride selectivity of the dipyrrolylquinoxaline building block,
while illustrating ways wherein, through appropriate construction of pre-organized
structures, it could be potentially fine-tuned.

FlKUre 4. Viewof the structure of22b ·4CHCh as determined froma singlecrystal Xvray diffraction analysis.
Displacement ellipsoids are scaledto the 50%probability level. Mosthydrogen atomshave been removed for
clarity. The macrocycle lies around a crystallographic inversion center at 1/2, 1/2, 1/2. Open ellipsoids
represent atomsrelatedby l-x, I-y, l-z, Similar molecular structures were observed for 22c, as well as when a
second set of crystals of 22b grown under slightlydiffering conditions (22b'6CHClrC,Ha) were subject to
analysis. Significant differences in the intermolecular contacts were, however, observed between 22b (both
structures) and22c. The latterdisplays columnar structures characterized by a "molecular tube"motif.


In summary, the dipyrrolylquinoxalines prov ide an easy-to-make and highly

accessible set of anion sensors . They can be modified readily by taking advantage of a
wide range of commercially available precursors and, further, can serve as building
blocks for the elaboration of more elaborate receptors, such as metal complexes IS and
16, pyrrole-functionalized "chelates" 17 and 18, or macro cycles 20-22. At the present
time, the chemistry of this seemingly versatile colorimetric motif is still in its infancy and
it is to be expected that a wide range of new structures based on it will be obtained, just
as a wide range of applications based on its near -unique naked eye anion sensing
characteristics can be envisioned .


This work described in this review was supported by the National Institutes of Health
(grant GM 58907 to J.L.S.) , the National Science Foundation (grants CHE-9725399 and
CHE-OI07732 to J.L.S.), the Robert A. Welch Foundation (grant F-IOI8 to J.L.S .), and
Kraft Foods . H.M. expresses thanks to Prof. Atsuhiro Osuka of the Graduate School of
Sciences of Kyoto University for a study leave and the JSPS for a Young Scientist
Research Fellowsh ip.


J. F. P. Schmidtchen and M. Berger , Artific ial organic host molecules for anions, Chem. Rev. 97, 1609-1646
(I 997}.
2. Supramolecular Chemistry of Anions, edited by A. Bianchi , K. Bowman-James, and E. Garc ia-Espana (VCH
Verlag , Weinheim, 1997).
3. P. A. Gale, Anion receptor chem istry: highlights from 1999, Coord. Chem. Rev. 213, 79-128 (2ool ).
4. P. D. Beer and P. A. Gale , An ion recognition and sensing: The state of the art and future perspectives,
Angew. Chem. Int. Ed. 40,487-516 (2001).
5. P. D. Beer , Transition-metal receptor systems for the selective recognition and sensing of anionic guest
species, Ace. Chem. Res. 31, 71-80 (I 998}.
6. T. S. Snowden and E. V. Anslyn, An ion recognition: synthetic receptors for anions and the ir applicat ion in
sensors, CurroOpin. Chem. Bioi. 3, 740-746 (1999).
7. K. Ni ikura, A. P. Bisson, and E. V. Anslyn, Opt ical sensing of inorganic anions employing a synthetic
receptor and ionic colorimetric dyes, 1. Chem. Soc.• PerkinTrans 2,1111-1114 (1999) .
8. A. Metzger, E. V. Anslyn, A chemosensor for citrate in beverages, Angew. Chem. Int. Ed. 40, 649-652
(I 998}.
9. K. H. Choi and A. D. Hamilton, Selective anion binding by a macrocycle with convergent hydrogen bonding
functional ity, J. Am. Chem. Soc. 123,2456-2457 (2ool).
10. K. Choi and A. D. Hamilton, A Dual Channel Fluorescence Chemosens or for An ions Involving
Intermolecular Excited State Proton Transfer, Angew. Chem . 40, 3912-3915 (2001).
II . S. Nishizawa, H. Kaneda , T. Uchida, and N. Terarnae, Anion sensing by a donor-spacer-acceptor system: an
intramolecular exc iplex emission enhanced by hydrogen bond-mediated complexation, J. Chem. Soc.•
Perkin Trans 2, 2325-2327 (1998).
12. S. Nishizawa , R. Kato, T. Hayashita , and N. Teramae, Anion sens ing by a thiourea based chromo ionophore
via hydrogen bonding, Anal. Sci. 14,595-597 (1998).
13. R. Kato, S. Nishizawa, T. Hayashita, and N. Teramae, A thiourea-based chromoionophore for selective
binding and sensing of acetate, Tetrahedron Len. 42, 5053-5056 (2001).

14. Y. Kubo, M. Tsukahara, S. Ishihara, and S. Tokita, A simple anion chemosensor based on a naphthalene-
thiouronium dyad, Chem. Commun., 653-654 (2000).
IS. C. Lee, D. H. Lee, and J. I. Hong, Colorimetric anion sensing by porphyrin-based anion receptors,
Tetrahedron Lett. 42, 8665-8668 (2001).
16. D. H. Lee, H. Y. Lee, K. H. Lee, and 1. I. Hong, Selective anion sensing based on a dual-chromophore
approach, Chem. Commun., 1188-1189 (200 I).
17. D. H. Lee, K. H. Lee, and 1. I. Hong, An azophenol-based chromogenic anion sensor, Org. Lett. 3 , 5-8
18. V. Amendola, E. Bastianello, 1. Fabbrizzi, C. Mangano, P. Pallavicini, A. Perotti, A. M. Lanfredi, and F.
Ugozzoli, Halide-ion encapsulation by a flexible dicopper(lI) bis-tren cryptate, Angew. Chem. Int. Ed. 39,
2917-2920 (2000).
19. S. S. Sun and A. J. Lees, Anion recognition through hydrogen bonding: a simple, yet highly sensitive,
luminescent metal-complex receptor, Chem. Commun., 1687-1688 (2000).
20. S. Mason, J. M. L1inares,M. Morton, T. Clifford, and K. Bowman-James, Snapshots of Fluoride Binding in
a Aza Cryptand,1. Am. Chem. Soc. 122, 1814-1815 (2001).
21. M. Montalti, 1. Prodi, N. Zaccheroni, 1. Charbonniere, 1. Douce, and R. Ziessel, A Luminescent Anion
Sensor Based on a Europium Hybrid Complex, J. Am. Chem. Soc. 123, 12694-12695 (2001).
22. D. H. Lee, K. H. Lee, and J.-1. Hong, An Azophenol-Based Chromogenic Anion Sensor, Org. Lett. 3, 5-8
23. K. Odashima, R. Naganawa , H. Radecka, M. Kataoka, E. Kimura, T. Koike, K. Tohda, M. Tange, H.
Furuta, J. 1. Sessler, K. Vagi, and Y. Umezawa, Chemical Sensing Based on Membrane Potential Change
Induced by Host-Guest Complexation at Membrane Surface, Supramolec. Chem. 4,101-113 (1994).
24. X. M. Lin, K. Umezawa, K. Tohda, H. Furuta, J. 1. Sessler, and Y. Umezawa, Potentiometric Responses of
Expanded Porphyrin Incorporated Liquid Membrane Electrodes toward a Series of Inorganic and Organic
Anions, Analytical Sciences 14, 99-108 (1998).
25. V. Knil, J. 1. Sessler, T. V. Shishkanova, P. A. Gale, and R. Volf, Molecular recognition at an organic-
aqueous interface: Heterocalixarenes as anion binding agents in liquid polymeric membrane ion-selective
electrodes, 1. Am. Chem. Soc. 121,8771-8775 (1999).
26. H. Miyaji and 1. 1. Sessler, Off the Shelf Anion Sensors,Angew. Chem.Tnt. Ed.• Engl. 40,154-157 (2001).
27. K. P. Xiao, P. Buhlmann, S. Nishizawa, S. Amemiya, and Y. Umezawa, A chloride ion-select ive solvent
polymeric membrane electrode based on a hydrogen bond forming ionophore, Anal. Chem. 69,1038-1044
28. C. 1. Ward, P. Patel, and T. D. James, A molecular colour sensor for fluoride, Chem. Lett. 406-407 (2000).
29. S. Yamaguchi, S. Akiyama, and K. Tamao, Colorimetric fluoride ion sensing by boron-containing pi-
electron systems, J. Am. Chem. Soc. 123, 11372-11375 (2001).
30.1.1. Sessler, A. Gebauer, and P. A. Gale, Anion Binding and Electrochemical Properties ofCalix[4]pyrrole
Ferrocene Conjugates, Gazz. Chim. Ital. 127,723-726 (1997).
31. K. H. Lee, H. Y. Lee, D. H. Lee, and J. I. Hong, Fluoride-selective chromogenic sensors based on
azophenol, Tetrahedron Lett. 42, 5447-5449 (2001).
32. M. Takeuchi, T. Shioya, and T. M. Swager, Allosteric fluoride anion recognition by a doubly strapped
porphyrin, Angew. Chem.lnt. Ed. 40,3372-3376 (2001).
33. H. Miyaji, P. Anzenbacher, Jr., J. L. Sessler, E. R. Bleasdale , and P. A. Gale, Anthracene-linked
calix[4]pyrroles: fluorescent chemosensors for anions, Chem. Commun., 1723-1724 (1999).
34. P. A. Gale, 1. 1. Twyman, C. I. Handlin, and J. 1. Sessler, A colourimetric calix[4]pyrrole-4-nitrophenolate
based anion sensor, Chem. Commun., 1851-1852 (1999).
35. C. B. Black, B. Andrioletti, A. C. Try, C. Ruiperez, and J. L. Sessler, Dipyrrolylquionoxalines. Efficient
Sensors for Fluoride Anion in Organic Media,J. Am.Chem. Soc. 121, 10438-10439 (1999).
36. H. Miyaji, W. Sato, and J. 1. Sessler, Naked-Eye Detection of Anions in Dichloromethane: Colorimetric
Anion Sensors Based on Calix[4]pyrrole, Angew. Chem. Int. Ed. 39, 1777-1780 (2000).
37. P. Anzenbacher, Jr., K. Jurisikova, and J. 1. Sessler, Second Generation Calixpyrrole Anion Sensors, J. Am.
Chem. Soc. 122,9350-9351 (2000).
38. 1. 1. Sessler , P. Anzenbacher , Jr., H. Miyaji, K. Jurslkova, E. R. B1easdale, and P. A. Gale, Modified
Calix[4]pyrroles,lndustr. Engin. Chem. Res. 39,3471-3478 (2000).
39. P. Anzenbacher, Jr., A. C. Try, H. Miyaji, K. Jurisikova, V. M. Lynch, M. Marquez, and J. 1. Sessler,
Fluorinated Calix[4]pyrrole and Dipyrrolylquinoxaline. Neutral Anion Receptors with Augmented
Affinites and Enhanced Selectivities,J. Am. Chem. Soc. 122, 10268-10272 (2000).
40 . C. Bucher, R. S. Zimmerman, V. Lynch, V. Kral, and J. 1. Sessler, Synthes is of Novel Expanded
Calixphyrins: Anion Binding Properties of a Calix[6]phyrin with a Deep Cavity, 1. Am. Chem. Soc. 123,
2099-2100 (2001).

41. J. L. Sessler, P. I. Sansom, A. Andrievsky, and V. Kral, Application Aspects Involving the Supramolecular
Chemistry of Anions, in: Supramolecular Chemistry ofAnions, edited by A. Bianchi, K. Bowman-James,
and E. Garcia-Espana(VCH Verlag, Weinheim, 1997),pp. 355-419.
42. J. L. Sessler, A. Andrievsky, and J. W. Genge, Anion Binding by Sapphyrins, in: Advances in
Supramolecular Chemistry Vol. 4, edited by G. W. Gokel (JAI Press Inc., Greenwich, 1997), pp 97-142.
43. 1. L. Sessler and J. M. Davis, Sapphyrins: Versatile Anion-binding Agents, Ace. Chem. Res. 34, 989-997
44. P. A. Gale, J. L. Sessler, and V. Knil, Calixpyrroles, Chem. Commun., 1-8(1998).
45. P. A. Gale, P. Anzenbacher, Jr., and J. L. Sessler,CalixpyrrolesII, Coord. Chem. Rev. 222, 57-102 (2001).
46. S. Camiolo, P. A. Gale, and J. L. Sessler, Pyrrolicand PolypyrrolicAnion Binding Agents in: Encyclopedia
of Supramolecular Chemistry, edited by J. L. Atwood and 1. W. Steed (Marcel Dekker, New York), in
47. T. Mizuno, W.-H. Wei, L. R. Eller, and J. L. Sessler, Phenanthroline Complexes Bearing Fused
Dipyrrolylquinoxaline Anion RecognitionSites: Efficient Fluoride Anion Receptors, J. Am. Chem. Soc.,
in press.
48. J. L. Sessler, H. Maeda, T. Mizuno, V. M. Lynch, and H. Furuta, Quinoxaline-bridgedPorphyrinoids,
submitted for publication.
49. J. L. Sessler, H. Maeda, T. Mizuno, V. M. Lynch, and H. Furuta, Quinoxaline-oligopyrroles: Improved
colorimetricanion receptors, Chem. Commun. 862-863(2002).
50. U. Lucking, D. M. Rudkevich,and J. Rebek Jr., Deep cavitands for anion recognition, Tetrahedron Lett. 41,
9547-9551 (2000).
51. B. Oddo, Syntheses in the Pyrrole Group. Dipyrrolyl and its Derivatives, Gazz. Chim. Ital . 41, 248-255
52. H. Fischer and H. Orth, Die Chemie Des Pyrrols (Akademische Verlagsgesellschaft, Leipzip, 1934),p. 380.
53. D. Behr, S. Brandiinge, and B. Lindstrom, Synthesis of Some Aryl n-Diketones and Aryl Glyoxylic Acid
Derivativesby Acyclation of Electron-rich Aromatics,Acta Chem. Scand. 27,2411-2414 (1973).
54. 1. L. Sessler, A. Gebauer, and S. J. Weghom, ExpandedPorphyrins,in: The Porphyrin Handbook, edited by
K. M. Kadish, K. M. Smith, and R. Guilard (Academic Press: San Diego and Burlington, 2000), Chapter
55. J. L. Sessler, R. S. Zimmerman, C. Bucher, V. Kral, and B. Andioletti, Calixphyrins: Hybrid Macrocycles
at the Structural Crossroads between Porphyrins and Calixpyrroles, Pure Appl . Chem. 73, 1041-1057
56. V. Knil, J. L. Sessler, R. S. Zimmerman, D. Seidel, V. Lynch, and B. Andrioletti, Calixphyrins: Novel
Macrocycles at the Intersection Between Porphyrins and Calixpyrroles, Angew. Chem., Int. Ed. 39, 1055-
57. C. Bucher, D. Seidel, J. L. Sessler, V. Kral, and V. Lynch, Novel Synthesis of Hybrid Calixphyrin
Macrocycles, Org. u« 2, 3103-3106(2000).
58. E. Vogel, M. Kocher, H. Schmickler, and J. Lex, Porphycene-ANovel Porphin Isomer, Angew. Chem. Int.
Ed. Engl. 25, 257-259 (1986).
59. E. Vogel, M. Brooring, J. Fink, D. Rosen, H. Schmickler,J. Lex, K. W. K. Chan, Y. D. Wu, D. A. Plattner,
M. Nendel, and K. N. Houk, From Porphyrin Isomers to Octapyrrolic Figure-S Macrocycles, Angew.
Chem.lnt. Ed. Engl. 34, 2511-2514(1995).
60. J. L. Sessler, A. Gebauer. and E. Vogel, Porphyrin Isomers, in The Porphyrin Handbook, edited by K. M.
Kadish, K. M. Smith, and R. Guilard (AcademicPress: San Diego and Burlington,2000), Chapter 8.
61. J. L. Sessler, N. A. Tvermoes, J. Davis, P. Anzenbacher, Jr., K. Jursikova, W. Sato, D. Seidel, V. Lynch, C.
B. Black, A. Try, B. Andrioletti,G. Hemmi, T. D. Mody, D. J. Magda, and V. Kral, Expanded Porphyrins.
SyntheticMaterialswith PotentialMedical Utility, Pure Appl. Chem. 71,2009-2018 (1999).
62. Z. Gross, N. Galili, L. Simkhovich, I. Saltsman, M. Botoshansky, D. Blaser, R. Boese, and I. Goldberg,
Solvent-free condensation of pyrrole and pentafluorobenzaldehyde: A novel synthetic pathway to corrole
and oligopyrromethenes, Org. Lett. I, 599·602 (1999).

K. Travis Holman and Jerry L. Atwood


Anion recognition plays a central role in the chemistry of natural processes .' This is
evidenc~d by the fact that the substrates of a majority of characterized enzymes are
anionic . Despite the prominence of anion-recognition chemistry in biological systems,
the design of supramolecular anion receptors was slow to develop with respect to the
analogous chemistry of cations. This discrepancy may be traced to a number of inherent
difficulties in designing appropriate anion hosts. ' Foremost are the synthetic challenges
associated with designing hosts for the diverse range of shapes, charges, and typically
larger sizes of anions (compared to simple metal cations) .7 Furthermore , in comparison
to cations of similar size, anions usually have higher free energies of solvation and,
hence, hosts are often required to compete more effectively with the surrounding
Progress in anion recognition chemistry has been rapid in recent years. • It is
important to note, however , that although early stu<lies adopted the electrostatic/
hydrogen-bonding approach utilized by Nature, the search for anion-selective receptors
has not been limited to the mimicry of natural systems. Anion recognition chemistry has
also been driven by the potential for applications as anion sensors,IO·'2 ion-transport
agents, environmental-cleanup agents, biomedical applicat ions, and enzyme mimics/
. cata Iysts. I

Corresponding author : Jerry L. Atwood, Department of Chemistry , Universiry of Missouri -Columbia,

Columb ia, MO 65211 USA; E-mail, atwoodj, TEL : (573) 882-8374 . FAX : (573) 882-2754.
Present address : K. Travis Holman , Department of Chemistry, Georgetown University, Washington, DC.

Fundamentals andApplications of AnionSeparations, edited by

Moyerand Singh,KluwerAcademic/Plenum Publishers, NewYork,2004 87

The molecular-recognition chemistry of calixarene cavitands has focused on the

binding of substrates at any of three sites according to Figure I, principally: i)
intracavity, withinthe bowl of the host, ii) at a synthetically modified lower rim, or iii) at
a modified upper rim. The electron-rich nature of the arene-lined cavity, combined with
the cyclic array of oxygen donors located at the lower rim, has essentially limited
intracavity substrate complexation to either neutral or cationic guests. · \7 The cation
binding properties of the electron-rich bowl site are exemplified by the ability of p-tert-
butylcalix[4]arene, t, to selectively transport Cs+cations through an organic membrane
(CH2CI2/CCI4) . 18 The first report concerning anion inclusion within the cavity of a
calixarene involved the methylsulfate anion within the cavity of [p-sulfonato-
calix[4]arene - Ht. This inclusion of an anion-by-an-anion, however, is based solely
on hydrophobic considerations and does not represent any significant binding of the
anionic portionof the guest by the bowl cavity.


(b) (c)

Figure 1. (a) The three sites for substrate recognition by bowl-shaped cavitands: i) intracavity, ii) lower rim,
and. iii) upper rim. (b) Crystal structure of [p·rerr-butylcalix(4)arene - HI" containing a Cs' cation. (c) Crystal
structure depicting the intracavity complex between [P-sulfonatocalix(4)arenet and the trimethylanilinium

The inability of the calixarene cavity to accommodate anionic substrates has limited
their use, and indeed the use of polyaromatic cavitands in general, as anion-binding host
molecules. Some anion-binding calixarenes have been reported,21.22but these hosts make
use of substantial upper or lower rim functionalization to mediate binding at these sites.
Thus, in these cases, the calixarene component represents a molecular platform on to
which known anion-binding functionality can be attached.

o NH

2 3 4
t -B -Bu -Bu -Bu

~ I~
I v

( OH OH )

~?P 5 6

7 8
Figure 2. Beer's anion hosts derived from the calix[4]arene platform.

Beer et al. have extensively exploited calixarenes as such a platform." Indeed,

upper and lower rim functionalized calix[4]arenes, such as the cobaltocenium containing
2, 3, and 4;24 the tris(bipyridyl) ruthenium(lI) molecules 5, 6, and 7;25 and the ferrocene-
based 8;26 couple the electrostatic utility of cationic redox-active transition-metal centers
with the directional hydrogen bonds of the amide functionality to recognize and sense
certain anionic substrates (Fig. 2). Again, in all cases anion complexation takes place at
upper-or-Iower rim sites, and inclusion within the calixarene bowl is not observed .
The use of transition-metal fragments (MLn) for the activation of nonnall~
unreactive It-hydrocarbons toward nucleophilic attack has been known for decades. 7
For example, the coordination of arenes by [Cr(COh],28 [CpFe+],29 and [Mn(COh+]30 is
known to inhibit their nucleophilic properties and induce electrophilic ones.
We, therefore, reasoned that coordination of such metal fragments to the outer arene
faces of calixarenes, and other bowl-shaped polyaromat ic macrocycles, may be sufficient
to convert the ostensibly electron-rich cavities of these host molecules into electron-
deficient sites for anion complexation (Fig. 3). Furthermore, such activated organo-
metallic calixarenes would be expected to provide synthetic access to new calixarene
hosts, whose synthesis would otherwise be difficult, ifnot impossible, by current organic
methodologies .

Figure 3. Modifying an electron-rich host cavity by coordination of the arenes to transition-metal fragments
(ML.) . The result is an electron-deficient, anion-binding molecular cavity.

The development of receptor molecules for the specific complexation and/or

detection of anions has, over the years, proven to be a formidable task. With the
exception of Lewis acidic hosts, an overwhelming majority of synthetic anion receptors
rely on complementary recognition through hydrogen bonds either exclusively, as with
amide containing neutral hosts, or in conjunction with sites of positive charge, as with
protonated cyclic polyamines and guanidinium-based hosts, for example. The utility of
this approach can hardly be overemphasized since hydrogen-bonding topology often
constitutes the basis for anion selectivity. For the complexation of some anions,
however, the synthetic incorporation of hydrogen-bonding groups (e.g., amide,
guanidinium) could actually be detrimental to selectivity. Indeed, for the selective
binding of target anions that are not especially susceptible to hydrogen-bond formation
(with respect to competitor anions in a system), it may be more useful to employ cationic
hosts with only appropriately preorganized sites of positive charge, those which lack the
ability to partake in significant hydrogen-bonding interactions. In this respect,

1t-metallated polyaromatic cavitands (e.g., calixarene, cyclotriveratrylene) may possess

the ability to discriminate between anionic substrates based upon parameters such as
charge topology, the flexibility/rigidity of the host system, and size/symmetry
complementarity between host and guest.
Prior to our investigations, the only reports of n-metallated calixarenes were
synthetic studies involv ing [Cr(CO)3] complexes of the various rigidified
tetrapropoxycalix[4]arene conformers."


In 1993, Atwood and coworkers initiated a study directed at the synthesis of n-

metallated calixarenes and related macrocycles.r' Studies focused on cationic fragments
of the second- and third-row late transition metals (i.e., Ru, Rh, Ir), as there is ample
33 36
literature precedence for the facile synthesis such of 116-arene compounds. - Indeed , in
many cases , complexes can be formed with phenol," suggesting that the analogous
chemistry would be feasible for calixarenes. Furthermore, these complexes exhibit a
significantly greater stability toward air and moisture than [Cr(COh] or [(arene)Fef+
metallated arenes . Indeed , some [(arenehRuf+ complexes even have appreciable water
solubility, an important feature for the potential use of supramolecular anion hosts in
waste remediation or other environmental applications.
Thus , reactions of 1 with the solvated [(11 5-Cp*)M(Sh]2 + (M = Rh lll , Ir'", S =
acetone) species, derived from pretreatment of the chloride-bridged dimers [{(11 5-
Cp*)MCI(Il-C1)b] (M = Rh 9, Ir 10) with silver salts , Ag[X], were investigated (Scheme
I) . The strategy was immediately successful in as much as the exclusively distal
dimetallic hosts [{(115-Cp*)Mh(116:116-p-tert-butylcalix[4]arene)][X]4 (M = Rh lla, Ir
l lb, X = BF4-) were produced in nearly quantitative yields ."

~M" /MIfJ
cr ~2'
U, /CL, ~I Ag[X], acteone I
I CI - AgCI{s) • III

S = acetone
M = Rh 9, Ir '0

1 ..
M111 +
s "s
s= acetone

R1 = ISU, R2 = H 11
1=H 2
R ,R =H 13
1=H 2=Pr
R ,R 15
Scheme 1. General synthesis of [(11 5-Cp*)Mt (M = Rh, Ir) dimetallated

No monometallated or higher metallated products were observed. Previous reports

have illustrated that n-phenolic complexes of rhodium, namely [(TJs-Cp*)Rh(TJ6-
C6HsOH)f\37b have a strong tendency to deprotonates, and this enhanced acidity is also
exhibited by 11, as the two deprotonated forms can also be prepared. However, the X-
ray crystal structures of the isostructural monodeprotonated species [tt-H]
[BF4kEt20'Me N0 2 did not display anion-inclusion properties, and instead the hosts
exhibit the more classical type of calixarene inclusion chemistry with a molecule of
diethyl ether occupying the host cavity. This result is not particularly surprising,
however, given that the host cavity still consists of two unmodified, and therefore
electron-rich, arene walls. The same synthetic methodology, applied to de-tert-butylated
calix[4]arene, 12, resulted in the clean formation of either the bimetallated or
tetrametallated calixarenes [{(TJ s-Cp*)Mh(TJ6:TJ 6-calix[4]arene)][BF4]4 (M = Rh 13a, Ir
13b) and [{(TJs-Cp*)lr}4(TJ6: TJ 6:TJ 6:TJ6-calix[4]arene-ZH)][BF4]6' 14, depending on the
solvent conditions employed (acetone/CF 3COOH VS. neat CF3COOH).38 Again, no
monometallated or trimetallated products were observed.

,......,CI.<, I
Ru -,,/R
Ag[X), acteone
<RLu'7 R

I<, C(.... 1'? -

AgCI(s) S ....S" "-S
CI ~
S = acetone

s= acetone

17 18
Scheme 2. Synthesis of[(TJ 6-arene)Ru]2+ metallated calix[4]arenes.

Bis(arene) ruthenium dications, consisting of two different arenes, are synthetically

accessible by a similar reaction scheme. The general procedure involves the similar
cleav-age of a related chloride-bridged dimer [{(TJ6-arene)RuCI(ll-CI) h ], 16, derived
from re-duction of the corresponding 1,3- or l,4-cyclohexadiene by RuCI 3 in ethanol,
with a silver salt to create the tris(solvatedj, [(TJ 6-arene)Ru(Shf+species that will further
react in acidic media to coordinate arenes. 4 Rybinskaya et al. have shown that, in some
cases, 16 may react directly with arenes to form the bis(arene) complexes:" These
proceduresapplied to calix[4]arenes yield metallated products(Scheme 2).
The Rybinskaya method (16a) applied to 12 yields cleanly the deprotonated di-
ruthenium species [{(TJ6-p-MeC6H4CHMe2)Ruh(TJ6:TJ6-calix[4]arene-H)]3+, 17, as its bis-
(trifluoroacetate) salt. Similarly, only dimetallated products were obtained when silver
salts, AgX (X = [H2P04'], [WOl]05' [CH3C02'D, were employed according to the
general procedure. Analogous pretreatment of 16a with AgX salts, (X = [BF4'],
[CF3S03'], [SO/"]05' or [PF6-D, followed by reaction with 0.5 equivalents of
calix[4]arene, 12, in refluxing trifluoroacetic acid, resulted in the clean formation of the
corresponding tetra-metallated derivatives [(TJ6-p-MeC6H4CHMe2)Ru l4(TJ6:TJ6-
calix[4]arene-2H)][X]6, 18, (X = [BF4'] , [CF3S03'], ([HS0 4'Msol D, [PF6 Again, n.
these complexes are strongly acidic, and the tetrametallated calix[4]arenes, 18 could only
be isolated as the twice deprotonated species. Notably, these tetrametallated derivatives
could only be synthesized in the presence of relatively non-coordinating anions, implying
that some sort of anion-metal coordination inhibits complete reaction in the case of 17.
Indeed, there is ample evidence for the coordination of carboxylate and halide ions to
ruthenium centers, whereas [BF4'] , [CF3S03' ] , and even [SOl] are found to be only
weakly coordinati
The crystal structure determination of 18, [BF4]dH 20)15. clearly establishes the
ability of metal coordinated calixarene hosts to complex anionic guests within their bowl
cavities. Indeed, the structure reveals that one of the tetrafluoroborate anions is situated
centrally within the bowl cavity of 18 (Fig. 4a). The host cavity and the guest anion are
apparently compatible in size, as one fluorine atom is directed essentially straight down
in such a way as to match the pseudo-conical surfaces of the host and the substrate. A
distance of nearly an Angstrom (ca. 0.9 A) highlights the depth of anion penetration
within the host cavity between the penetratingfluorine atom and the plane defined by the
upper-rim carbon atoms of the host. The closest anion-cation inter-molecular contacts
involve this fluorine atom and the lowest (Cl) carbon atom of the calix-arene (2.85 A).
The complementary host-guest fit is further established by the relatively small thermal
parameters of the cavity-bound anion with respect to the others in the lattice. Thus, this
crystal structure represents the first real inclusion of an anion within the arene-linedbowl
of a calixarenehost.
Anion metathesis of 18[BF4]6 with excess [NBu4][1] in nitromethane affords the
analogous iodide salt, 18[1]6, as a precipitate.Y Importantly, the metathesis reaction is
accompanied by a marked color change of the solution, from pale yellow to orange,
likely as a result of second-sphere charge transfer between the iodide and the host. The
perrhenate and dihydrogen phosphate salts of 18 can also be prepared in this fashion.
The X-ray crystal structure of this iodide salt, 18[l]dH 20h '(MeN0 2h , again
demonstrates the inclusion of one of the anions deep within the calixarene cavity (Fig.
4b). Similar to the tetrafluoroborate structure, the cavity bound ion exhibits the smallest

thermal parameters, but in this case the iodide lies ca. 0.1 A lower in the cavity than the
boron atom of thetetra- fluoroborate salt. Since the initial report of the tetra [(Tl s-
Cp*)Irf+ metallated calix[4]- arene 14,41 a crystal structure of the hydrogen sulfate salt
of the monodeprotonated form (14 - H) has been determined.Y In this structure, one of
the [HS04-] anions is included within the host cavity (Fig. 5).

(a) (b)

Figure 4. X-ray crystal structures of (a) 18c[BF.] and (b) 18cL

Figure S. Side view, derived from a crystal structure determination , of the inclusion ofa hydrogen sulfate
anion within the bowl cavity of[14 - H]" .


The cone conformation of calixarenes is not limited to calix[4]arene and its specific
derivatives. Although higher calixarenes typically display a dynamic range of
conformational isomers, calix[5Jarene is known also to prefer the cone conformation in
solution and in the solid state. Clearly, in order to achieve anion recognition based

upon size and shape, the binding properties of a range of calixarene bowl sizes should be
investigated. It was anticipated that larger anionic guests could be accommodated by the
larger host cavities of metallated calix[5]arene derivatives. Reaction of p-tert-
butylcalix[5]arene, 19, with excess iridium chloride dimer, 10, yielded a product which
showed two Cp" resonances (2:1 ratio) in its ' H NMR spectrum, indicative of metallation
of either rings A, B, and C or rings A, B, and D. Mass-spectral characterization
confirmed the trimetallated formulation with a peak at m/z 1964 (M3+ + 2[BF4'])' A
single-crystal structure determination of [{(TJS -Cp*)lr h(r{r{r{p-tert-
butylcalix[5]arene-H)][BF4k(MeN02)odEt20h, Figure 6, established not only the A, B,
D metallation pattern, but also the fact that anions may be included in the cavities of
hosts that have not been fully metallated. The anion inclusion is, however, somewhat
different in this case, as it appears that the included anion is too small to optimally
occupy the host cavity. Whereas the guest anions were located centrally within 14 and
18, the tetratluoroborate anion in the crystal structure occupies an unsymmetrical
position closest to metallated rings A and B.

Figure 6. X-ray crystal structure of a metallated calix[5]arene complex with [SF.] .


The general successes involving both the synthesis and anion-inclusion properties of
n-metallated calixarene derivatives have prompted investigation into metallating other
bowl-shaped polyaromatic macrocyles." In particular, the readily available trimeric
macrocycle cyclotriveratrylene (CTV), 20, possesses a saucer-like cavity, that is slightly
more shallow, although wider, than those provided by the calix[4]arenes. Moreover, the
electron-rich veratryl rings can be expected to act as good ligands for transition metals.
Thus, it was expected that, due to the differences in cavity shape and size, metallated
CTVs might display different anion affinities than metallated calixarenes. Furthermore,
inclusion of guests within the cavity of this potential host molecule is extremely rare as a
consequence of its typical bowl-in-bowl solid-state packing. Therefore, it was also
reasoned that the attachment of transition metals to the outer-ring faces would disturb
this packing mode and could therefore modify the solid-state inclusion properties of this

shallow cavitand. A schematic of CTV, which depicts the carbon atom number scheme
adoptedin the reported crystal-structure determinations," is shownas 20.

The synthesis of organometallic [(T{Cp*)Irf+ and [(1l6-arene)Ru]2+ derivatives of

cy-clotriveratrylene was accomplished by the same general methodology applied to
metallated calixarene hosts. Schemes 3 and 4 outline the procedures and the products
formed. Reaction of the solvated precursors, derived from silver salt treatment of 10 or
16a-c with CTV in retluxingtritluoroacetic acid (tfa) gave the following range ofmono-,
di-, or trimetallated compounds in good to excellentyields.

Ag[X). acteone
< '7

cw . tfa reflux
- AgCI(s) 8"''/ -,
8 8
8 =acetone
arene = p-MeC6H 4CHMe2 16a
arene = C6H6 16b
arene = C6Me6 16c

~ M

~ ~ ,A-O~ 0-

o O



I~ s > 1) 0 0
arene = p-MeC6H,CHMe2 21 arene = p-MeC6H,CHMe2 24
arene C6H6 22
arene C6Me6 23

Scheme 3. Synthesis of [(T}6-arene)Ruf+metallated cyclotriveratrylenes.


~ /¥lf2 >$<
U, /Ck., I Ag[X), acteone CTV, tfa reflux


Ir<, CI -AgCI(s) Ir
s = acetone

,A- 0- ~
~ M

-O0- ~
' I
0 '¥'? ~~II"
\ I

27 28

Scheme 4. Synthesis of[(T}5-Cp*)Ir]2+ metallated cyclotriveratrylenes.

In general, the formation of mono-, di- or trimetallic products was found to depend
simply upon the ratio of the solvated species, [Ru(T}6-arene)(ShJ2+or [Ir(T}5-Cp*)(ShJ 2\
to CTV in the reaction mixture. Exceptions to this includethe synthesis of the trimetallic
species 2Sa in which at least a 1.5-fold excessof[Ru(T}6-arene)(ShJ2+is requiredin order
for the reaction to proceed to the desired trimetallated product. Even under these
conditions, however, the trimetalIated derivative obtained can be contaminated by up to

ca. 10% of the dimetallated compound . In this situation, the compound can be purified
by fractional crystallization when diethyl ether is diffused into an acteone solution of the
product. Also, despite attempts using a large excess of starting material, complex 2Sb
could not be synthesized without significant contamination by the dimetallated complex.
The two products could again be easily separated, however, by fractional crystallization
from Me N021Et20. Notably, under the conditions employed, it proved impossible to
generate any multimetallated CTV derivatives from the [(C6Me6)Ru(Shf+ precursor,
and only the monometallic 23 was obtained. In this case it seems likely that addition of
more than one [(116-C6Me6)Ruf+ moiety to the outer surface of the CTV molecule is
precluded by excessive steric interactions between the methyl substituents of the adjacent
hexamethylbenzene rings.
Interestingly, the second and third metallation steps were found to be inhibited by
the presence of certain counter ions. For example, direct reaction of a large excess of 16a
with CTV, according to the method of Rybinskaya," results in only the monometallic
complex , namely, 21a . This effect seemingly parallels that observed in the synthesis of
metallated calix[4]arenes and may well be attributable to ruthenium coordination of the
more strongly ligating anions. Indeed, even the [PF6'] anion is known to undergo partial
hydrolysis at room temperature in the presence of the tris(acetone) complex [(11
arenejkutacetonejj]" to give mainly mesityl oxide and [RU2(~-P02F2h(116-arenehr via a
coordinated ketol intermediate."
Salts of anions whose conjugate acids have a pK. greater than that of CF3C02H are
not directly available via the route outlined in Scheme 3 because of protonation by this
solvent medium. A number of complexes are obtainable , however, by simple anion
metathesis. Compounds [{Ru(p-cymene) h(CTV)][Re04h[CF3S03] 24c, [{Ru(P-
cymene)h(CTV)]Xt (X = Br' 24d, r 24e) and various phosphate containing salts can be
obtained this way.
The solid-state structures of the new host-guest systems presented here were probed,
whenever possible, by X-ray crystallography . Crystals of the hydrogen ditrifluoroacetate
salt lIb of the same cation, isolated from the direct reaction of excess chloride dimer 16a
with CTV, were also obtained . The cation adopts essentially the same conformation as in
2la with inter-ring centroid separations falling in the range 4.62-4.86 A. The anions
consist of strongly interacting [H(CF3C0 2)2'] dimers in a transoid conformation with the
oxygen atoms approximately coplanar . The O(H)"'O hydrogen bonding distances in the
two anionic dimers are 2.47(1) A and 2.45(1) A. The host cavity, in this case, is
occupied by the hydro-phobic trifluoromethyl substituent of one of the [H(CF 3C0 2h']
anions (Fig. 7). This facilitates an interesting form of anion-cation interaction in which
the carbonyl oxygen atom points down toward the metallated ring of the macrocycle .
The O' ''C(CTV) nonbonded contacts [O(l2)" 'C(4A) = 2.89(2) A, O(12)"'C(5A) =
3.05(2) A] suggest that this electro-static interaction, along with a hydrophobic van der
Waals type interaction involving the trifluoromethyl moiety, is a likely driving force for
this type of inclusion. Other anion-cation contacts are above 3.11 A.
The binuclear ruthenium complex 24b '(H 20) was also characterized by X-ray
crystallography. The dimensions of the molecular cavity are similar to those of the

Figure 7. X-ray crystal structure of21b depicting thecavity included [H(CF,COz12'1anion,

monometallated hosts with a relatively short centroid-centroid separation of 4.69 A

between the two metallated rings and longer separations (4.72, 4.89 A) to the
uncoordinated aromatic residue. As noted for the monometallated hosts, both metal
centers exhibit longer bonds to the carbon atoms bearing methoxy substituents than the
remaining four (2.353(8) AV5 . 2.200(7) A av.). Most importantly, however, the structure
determinat ion demonstrates that, like calixarenes with two or more adjacent metallated
rings, one of the triflate anions is located within the cavity of the macrocycle (Fig. 8).
The sulfonato head group of the anion points directly between the two metallated rings
whilst the hydrophobic -CF3 substituent occupies the center of the cavity . The closest
sulfonate oxygen atom-to-host contact is 2.95(1) A. The sterically restricted nature of
the intracavity anion is clearly reflected, however, in the crystallographic thermal
parameters , which are all markedly smaller than those for the other three triflate anions.
In the crystal, the host molecules pack in a columnar array separated by the included
triflate anion.

Figure 8. X-ray crystal structureof24b'HzO highlighting the cavity-includedtriflateanion.


Slow crystallization of 24b (MeNOidiethyl ether, vapor diffusion) in the presence

of a two-fold excess of [NB14][Re04l resulted in the isolation of crystals of a mixed
perrhenate/triflate salt [{RU(TJ 6-p-cymene) h (TJ6:TJ6-CTV)][Re04h [CF3S0 3]-MeN0 224c
which has also been characterized by X-ray crystallography , JR, and IH NMR.44b The JR
spectrum of24c displays an absorption at 900 ern", assigned to the [Re04"] anion (cf 907
em" for [NB14][Re04]).47 The X-ray crystal structure clearly demonstrates the inclusion
of a perrhenate anion, rather than a triflate anion, within the cavity of the host (Fig. 9).
The included anion is situated slightly off center, in the direction of the metallated rings,
with three of its oxygen atoms directed into the cavity such that the three-fold axis of the
guest and the pseudo three-fold axis of the hosts are matched. Again, the thermal
parameters of the cavity-included anion are significantly smaller than those of the lattice
anions. The inclusion of this particular anion is of exceptional interest since it is a
structural model for the environmental contaminant [99Tc04"J. Preliminary radiotracer
studies have shown that 24b demonstrates a remarkable ability to extract this anion from
aqueous into organic solvents .

Figure 9. Two views of the 24c[ReO'J complex highlighting size/shape complementarity between host and

The iridium-containing, dimetallic host 28 was crystallized as a diethyl ether solvate

by vapor diffusion of diethyl ether into a nitromethane solution of the salt. The presence
of diethyl ether was confirmed by the I H NMR spectrum of the complex, which
suggested ca. two molecules of solvent per host molecule, in agreement with the
crystallographically determined value. The general conformation and structure of the
host 28 is similar to its ruthenium analogue 24 with similar intracentroid separations
between arene rings and relatively long bonds from the metal center to the methoxy
substituted carbon atoms. As anticipated, one of the four [BF4-] anions occupies the
intracavity site and resides closest to the metallated rings (Fig. J0). The anion is situated
such that one of the fluorine atoms points between the two coordinated decks of the CTV
ligand, and the other three fluorine atoms are disordered around this B-Faxis, suggesting
that the included anion is free to rotate within the cavity about this axis. This result is
consistent with the poor size match between the relatively small [BF4-] anion (vs.
[CF3SOd , [Re04-]) and the wide CTV cavity. For 28 the shortest anion-host contacts
involve the included anion.

Figure 10. X-raycrystalstructure OJ me ZSc(BF4) species. For clarity, the anron IS depicted in only one


(b) (c)

Figure 11. Views of (a) the Cl synunetrical host conformation of25 in crystals of25a'(H20h and (b) the quasi,
C. confonnation of 25 in crystals of 25a ·(MeN02h. (c) Anion location in the 2k[BF'l species observed in
crystals of 25a'(Et20)'5' For clarity. only one anion orientation is shown.

Single crystals of compound 25a can be obtained from a variety of solvent systems.
The first structure determination, of crystals derived from an acetone-water solvent
system 25a'(H 20h, highlights the three-fold symmetry of the host, as it adopts a cubic
packing mode (space group Pa_3).44a The host cation is situated upon a threefold rotation
axis such that only one third of the molecule is unique (Fig. lla). This is in itself
remarkable since all previous structures of native CTV, or even metallated CTV, indicate
that the host always deviates substantially from idealized Cl v symmetry and typically
exhibits quasi Cs or even asymmetric conformations. Some investigators have attempted
to rationalize this behavior as a consequence of the interannular repulsion between the
three carbocyclic rings and the absence of homoaromaticity." In this structure, the
crystallographically imposed C3 symmetry results in equivalent intracentroid separations
between the CTV rings which measure 4.84 A. In similarity to 28, the cavity-included
[BF4' ] anion appears extremely disordered over three sites, again highlighting the size
mismatch between the CTV host cavity and this anion.

Second and third crystal structure determinations of the 25a salt demonstrate that the
trimetallated host need not adopt the C3 conformation. Indeed, in crystals of
25a '(MeN0 2h the host conformation is more reminiscent of other CTV crystal
structures, and it adopts a quasi-C. conformation with CTV ring centroid separations of
4.63 A, 4.63 A, and 4.97 A. In the diethyl ether solvate 25a '(Et20>u, the host exhibits
an asymmteric conformation as seen from the orientation of the cymene ligands and the
intracentroid separations of 4.69 A, 4.75 A, and 4.83 A. In both structures, the
intracavity anion again appears orientationally disordered, but adopts a definable position
significantly closer to one of the three metallated rings . Figure I 1 shows the C3 host
conformation, the C. conformation, and the offset position of the included [BF4' ] anion in
25a ·(Et20 )1.5.
Crystals of the analogous triirid ium compound 29a·(MeN0 2)o.s ·(H 20) were also
obtained. The host cation does not hold C3 symmetry as seen in crystals of 25a '(H 20h,
but is asymmetric with intracentroid separations of 4.60 A, 4.80 A, and 4.90 A. The
intracavity anion resides between two of the metallated rings in a fashion similar to that
seen in 28, although it is closer to one ring than the other (Fig. 12). Short anion-host
contacts of 2.92 A and 2.78(3) A highlight the electrostatic interactions between this
anion and the metallated arene rings of the host cation.

Figure 12. X,ray crystal structure ofthe 29c[BF,] species.


It has been demonstrated that the appendage of cationic , electron-withdrawing

transition metal fragments (i.e., [Cp*Rhf+, [Cp*lrf+' [(arene)Ruf+) to the outer aromatic
ring faces of polyaromatic bowl-shaped macrocycles (i.e., calixarenes, CTV) sufficiently
modifies the electron density of these ostensibly electron-rich cavities such that anion
inclusion may be observed, at least two adjacent arene rings of a particular host must be
metallated . Metallated molecules that do not satisfy this condition are typically seen to
include neutral molecules within their host cavities . These are the first real examples of
anion inclusion within the bowl cavities of these types of macrocycles, which are

typically more suited for cation or neutral-molecule inclusion. Solid-state studies show
that the anions located within the bowl-shaped cavities of the host are undoubtedly
engaged in interactions with the faces of the electron-deficient metal-coordinated arene


I. 1. W. Steed and 1. L. Atwood. Supramolecular Chemistry (Wiley. New York. 2000).

2. (a) L. G. Lange 1II. J. F. Riordan, and B. L. Vallee, Biochem. 13, 4361 (1974). (b) F. P. Schmidtchen
Nachr . Chem . Tech. Lab. 36, 8 (1988).
3. Cation Binding by Macrocycles, edited by Y.lnoue . and G. W. Gokel, (Dekker, New York. 1990).
4. (a) B. Dietrich, in: Inclusi on Compounds. edited by 1. L. Atwood, J. E. D. Davies, and D. D. Macblicol,
(Oxford University Press, Oxford, 1984). Vol. 2. Chapter 10. pp. 373-405. (b) B. Dietrich. Pure
Appl . Chem . 7.1457 (1993).
5. (a) M. W. Hosseini and J.-M. Lehn, Helv. Chim. Acta 70,1312 (1987). (b) J. L. Sessler. H. Furuta. and V.
Kral, Supramol. Chem . 1. 209 (1993).
6. D. Kaufmannand A. Otten, Angew. Chem .lnt. Ed. Engl. 33.1832 (1994).
7. J. L. Atwood, K. T. Holman. and J. W. Steed, Chem . Comm . 1401 (1996).
8. (a) F. Vogtle, H. Sieger, and W. M. Muller. Top. Curro Chem . 98, 107 (1981). (b) H. E. Katz, in:
Inclusion Compounds, edited by J. L. Atwood, 1. E. D. Davies. and D. D. MacNicol, (Oxford
University Press, Oxford, 1991). Vol. 4. Chapter 9, pp. 391-405. (c) R. M. Izatt, K. Pawlak, 1. S.
Bradshaw,and R. L. Bruening, Chem . Rev. 91, 1721 (1991). (d) L..L. Pierre,and J. Baret, Bull. Soc.
Chim. Fr. 367 (1983). (e) K. T. Holman, J. L. Atwood, and J. W. Steed, in Advances in
Supramolecular Chemistry. edited by G. W. Gokel, (JAI Press, Greenwich,CT, 1997).Vol. 4.
9. Supramolecular Chemistry of Anions . edited by A. Bianchi, K. Bowman-James, and E. Garcia-Espana,
(Wiley-VCH, New York, 1997).
10. (a) J. J. Lavigneand E. V. Anslyn, Angew. Chem .lnt. Ed. Engl. 40.3119 (2001). (b) T. S. Snowden and
E. V. Anslyn, Curro Opin. Chem . Bioi. 3, 740 (1999). (c) K. Niikuraand E. V. Anslyn,J. Chem . Soc.
Perkin Trans 2 2769 (1999). (d) K. Niikura, A. P. Bisson, and E. V. Anslyn, J. Chem . Soc. Perkin
Trans 2,1111 (1999).
II. (a) 1. A. Wisner, P. D. Beer, and M. G. B. Drew, Angew. Chem .lnt. Ed. Engl. 40,3606 (2001). (b) P. D.
Beer and P. V. Bernhardt, J. Chem . Soc . Dalton Trans. 1428 (2001). (c) P. D. Beer and P. A. Gale,
Angew. Chem . Int. Ed. Eng/. 40, 487 (200 I ).
12. (a) P. A. Gale, M. B. Hursthouse, M. E. Light, 1. L. Sessler, C. N. Warriner. and R. S. Zimmerman, Tet.
Lett. 42,6759 (2001). (b) C. Rucher, R. S. Zimmerman, V. Lynch. V. Kral. and 1. L. Sessler, J. Am.
Chem. Soc. 123.2099 (2001). (c) H. Miyaji and 1. L. Sessler, Angew. Chem . Int. Ed. Engl . 40.154
13. S. G. Bon, A. W. Coleman,and 1. L. Atwood,J. Am. Chem. Soc. 108, 1709(1986).
14. J. L. Atwood, G. W. Orr, F. Hamada. R. L. Vincent, S. G. Bon, and K. D. Robinson,J. Am . Chem . Soc.
113,2760 (1991).
15. J. L. Atwood, F. Hamada,K. D. Robinson,G. W. Orr, and R. L. Vincent, Nature 349,683 (1991).
16. V. Bohmer, Angew. Chem.lnt. Ed. Engl. 34, 713 (1995).
17. Comprehensive Supramolecular Chemistry, edited by J. L. Atwood, J. E. D. Davies, D. D. MacNicol, and
F. Vogtle, (Pergamon, Oxford, 1996).
18. S. R. Izatt, R. T. Hawkins, J. J. Christensen,and R. M. Izatt,J. Am . Chem . Soc. 107,63 (1985).
19. S. G. Bon, A. W. Coleman, and 1. L. Atwood, J. Am . Chem . Soc. 110,610 (1988).
20. S. Shinkai, K. Araki, T. Matsuda. N. Nishiyama, H. Ikeda, 1. Takasu, and M. Iwamoto,J. Am. Chem . Soc.,
112,9053 (1990).
21. J. Scheerder,M. Fochi,J. F. J. Engberson, and D. N. Reinhoudt,J. Org. Chem. 59,7815 (1994).
22. B. R. Cameronand S. J. Loeb, Chem. Comm. 573 (1997).
23. (a) P. D. Beer, Chem . Comm . 689 (1996). (b) P. D. Beer, Ace. Chem . Res. 31, 7, (1998).

24. (a) P. D. Beer, D. Hesek, J. E. Kingston. D. K. Smith, S. E. Stokes. and M. G. B. Drew, Organometallics,
14,3288 (1995) . (b) P. D. Beer. M. G. B. Drew, C. Hazelwood, D. Hesek, J. Hodacova, and S. E.
Stokes,J. Chem. Soc.. Chem. Comm . 229 (1993) . (c) P. D. Beer, M. G. B. Drew, D. Hesek, and K. C.
Nam, Chem . Comm . 107 ( 1997).
25. (a) P. D. Beer, Z. Chen, A. J. Goulden , A. Grieve, D. Hesek, F. Szemes, and T. Wear, J. Chem. Soc..
Chem. Comm . 1269 ( 1994). (b) F. Szemes, D. Hesek, Z. Chen, S. W. Dent, M. G. B. Drew. A. 1.
Goulden , A. R. Graydon, A. Grieve, R. 1. Mortimer, T. Wear. 1. S. Weightman, and P. D. Beer,
Inorg. Chem . 35, 5868 (1996) . (c) P. D. Beer. R. J. Mortimer, N. R. Stradiotto, F. Szernes, and J. S.
Weightman. Anal. Commun . 32,419 (1995) .
26. P. D. Beer, Z. Chen. A. 1. Goulden, A. Graydon. S. E. Stokes. and T. Wear, 1. Chem . Soc.. Chem. Comm .
1834 (1993).
27. D. A. White. Organomet. Chem. Rev.. Sect . A, 3, 497 (1968) .
28 . (a) M. F. Semmelhack, Annals N. Y. Acad. Sci. 295, 36 (1977) . (b) G. Jaouen, Annals N. Y. Acad . Sci.
295,59 (1977). (c) A . Solladie-Cavallo, Polyhedron 4, 901 (1985) .
29. A. N. Nesrneyanov, N. A. Vol'kenau, and I. N. Bolesova, Dokl. Akad. Nauk SSSR 175,606 (1967).
30. L. A. P. Kane-Maguire, E. D. Honig. and D. A. Sweigart. Chem. Rev. 84, 525 (1984) .
31. (a) H. Iki. T. Kikuchi, and S. Shinkai, J. Chem Soc.. Perkin Trans. I 69 (1992) . (b) T. Kikuchi, H. Iki, H.
Tsuzuki, and S. Shinkai, Supramol. Chem . I, 103 (1993) . (c) H. lki, T. Kikuchi. and S. Shinkai, 1.
Chem Soc.. Perkin Trans. I 205 (1993) .
32 . 1. L. Atwood, K. T. Holman, and J. W. Steed. Chem Comm . 1401 (1996).
33. (a) R. A. Zelenka and M. C. Baird. J. Organomet. Chem . 44 , 383 (1972) . (b) R. A. Zelenka and M. C.
Baird, Can. 1. Chem . 50, 3063 (1972).
34. (a) M. A. Bennett and T. W. Matheson, J. Organomet. Chem. 175,87 (1979) . (b) M. A. Bennett, T. T.
Matheson, G. B. Robertson , W. L. Steffen, and T. W. Tumey , J. Chem Soc.. Chem Comm. 32
35. (a) 1. W. Kang, K. Moseley, and P. M. Maitlis, J. Am Chem. Soc. 91, 5970 (1969). (b) C. White and P.
M. Maitlis, 1. Chem Soc. A 3322 (1971 ).
36. P. J. Fagan, M. D. Ward, and 1. C. Calabrese , J. Am. Chem. Soc. Ill, 1698 ( 1989 ).
37. (a) D. 1. Cole-Hamilton , R. 1. Young, and G. J. Wilkinson, Chem. Soc.. Dalton Trans. 1995 ( 1976). (b) C.
While, S. J. Thompson. and P. M. Maitlis, 1. Organomet. Chem . 127,415 (1977). (c) M. Green and
T. A. Kuc,J. Chem. Soc.. Dalton Trans. 832 (1972).
38. J. W. Steed, R. K. Juneja, R. S. Burkhalter, and 1. L. AlWOOd, 1. Chem Soc.. Chem. Comm . 2205 (1994 ).
39. M.1. Rybinskaya, A. R. Kudinov, and V. S. Kaganovich ,J. Organomet. Chem . 246 , 279 (1983) .
40 . (a) E. L. Muneries , 1. R. Bleeke, E. 1. Wucther, and T. A. Albright, Chem Rev. 82, 499 (1982) . (b) M. A.
Bennett, in: Comprehensive Organometallic Chemistry, (Pergamon , Oxford, U.K., 1982), Vol. 4,
Chapter 32. (c) R. O. Gould, C. L. Jones, D. R. Robertson, T. A. Stephenson. and D. A. Tocher, 1.
Organomet. Chem . 226. 199 ( 1982).
41. 1. W. Steed, R. K. Juneja , and 1. L. Atwood, Angew. Chem.. Int. Ed. Engl. 33, 2456 (1994) .
42 . M. Staffilani. K. S. B. Hancock. J. W. Steed. K. T. Holman, J. L. Atwood, R. K. Juneja, and R. S.
Burkhalter . 1. Am. Chem . Soc. 119, 6324 (1997) .
43 . 1. W. Steed, C. P. Johnson , R. K. Juneja, and J. L. Atwood, Supramol. Chem. 6, 235 (1996) .
44. (a) 1. W. Steed, P. C. Junk, and 1. L. Atwood, J. Am. Chem . Soc . 116, 10346 (1994) . (b) K. T. Holman,
M. M. Halihan, J. W. Steed, S. S. Jurisson , and 1. L. Atwood, 1. Am. Chem . Soc . 117.788 (1995) .
(c) K. T. Holman. M. M. Halihan, S. S. Jurisson, J. L. Atwood, R. S. Burkhalter, A. R. Mitchell, and
J. W. Steed,J. Am. Chem. Soc . 118,9567 (1996).
45 . (a) 1. W. Steed, H. Zhang, and 1. L. Atwood, Supramol. Chem . 7, 27 (1996) . (b) H. Zhang and J. L.
Atwood, 1. Cryst. Spec. Res. 20, 465 (1990) . (c) G. I. Birnvaum, D. D. Klug, 1. A. Ripmeester , and
J. S. Tse. Can. J. Chem , 63. 3258 (1985) .
46 . M. A. Bennett, T. W. Matheson, G. B. Robertson , W. L. Steffen, and T. W. Tumey , J. Chem. Soc.. Chem.
Comm . 32 (1979 ).
47 . K. Nakamoto. Infrared and Raman Spectra oflnoganic and Coordination Compounds, (John Wiley and
Sons, New York, 1970), pp. 142.
48. G. I. Birnbaum, D. D. Klug, J. A. Ripmeester. and 1. S. Tse, Can. J. Chem. 63. 325 (1985).

Gregg 1. Lumetta'


Since the days of the Manhattan Project during World War II, the U.S. Department
of Energy (DOE) and its predecessors have performed a variety of research-and-
development as well as production activities. These activities have occurred at numerous
sites throughout the country. One of the legacies of these past practices is an enormous
quantity of radioactive waste at several contaminated sites. These wastes and the
associated environmental contamination pose a serious threat to human and ecological
health. Consequently, environmental restoration is a high priority for DOE.
Probably the most vexing challenge is the 3.4 x 105 m3 of high-level wastes stored in
tanks at the Savannah River, Hanford, Idaho National Engineering and Environmental
Laboratory (INEEL), and Oak Ridge sites.' The tank wastes must be retrieved, converted
into stable waste forms, and safely disposed of. This process is estimated to take 70 years
and cost $227 billion .' Given the enormity of this endeavor, it is important to understand
the fundamental chemistry of the wastes so as to minimize technical uncertainties. It is
also important to develop separation technologies that could reduce the cost and
accelerate remediation.
Although not always explicitly recognized, anions playa central role in many of the
environmental problems faced by DOE . Anions can be a hazard in and of themselves
(e.g., CrO/', Tc04'), or they can interfere with waste process ing (e.g ., sol', Pol '). In
this paper, I will present an overview of how anions manifest themselves in the DOE
environmental problems . I will also present a brief review of work that has been done to
address some of the problems caused by anions in the DOE complex. Emphasis will be
placed on issues relating to the Hanford Site in the State of Wash ington, but similar
issues are often encountered at other DOE sites.

Gregg J. Lumetta, PacificNorthwestNational Laboratory, Richland, Washington. PNNL is operated for the
U.S. Department of Energyby BattelleunderContract DE-AC06-76RLOI830.

FundmnentDls and Applications ofAnion Separations, edited by

Moyerand Singh, KluwerAcademicIPlenum Publishers. New York,2004 107


Table I presents a list of anions that are important to DOE environmental problems.
This is not meant to be a comprehensive list, but it highlights the major species of
interest. These anions can be conveniently divided into those affecting tank-waste
processing (i.e., vitrification) and those affecting the environment, although one category
does not precludethe other.

Table 1. Anions of Importance to DOE.

Anion Significance
sol Limiting component in formulation of
low-activity waste (LAW) glass waste form
pol Problematic component in vitrification of high-
level waste (HLW)
Present in large quantities in tank wastes-
major waste component
Added to tank wastes for corrosion control-
major waste component
AI(OHk Present in highly alkaline tank-waste solutions
and sludge leachingsolutions
Mobile, long-lived (t1/2 = 2.13 x lOs y)
radionuclide-poses significantlong-term risk
c-o," Mobile, toxic contaminantin the environment
U0 2(C03h2., U02(C03h4· Mobile contaminant in the environment

2.1. Tank Wastes

Sulfate and phosphate affect waste immobilization because the vitrification processes
do not tolerate these components even at relatively low concentrations. Phosphate may
limit waste loading in immobilizing Hanford HLWand calcine wastes stored at INEEL.
An immiscible alkali-phosphate phase can form in the glass melter at P20S concentrations
~2.5 wt %.3 The immiscible phosphate melts are intermixed with the silicate glass melt
and often crystallize during cooling of the glass. The resulting alkali-phosphate glass or
crystals contain high concentrations of radionuclides in a water-soluble form, thus
adversely affecting the performance of the final waste form.'
Similarly, sulfate salts can segregate as a separate phase in glass melters. This is
mainly an issue for Hanford LAW and INEEL sodium-bearing waste. Sulfate
segregation can occur at sulfate concentrations well below their solubility in the glass
melt (which is typically -I wt % as S03).4 Once segregated, redissolution into the glass
melt is very slow. Segregated molten salt is undesirable because it I) is very corrosive to
melter construction materials, 2) increases the risk of steam explosions in the liquid-fed
melters, and 3) may increase the rate of radionuclidevolatility,"

Aluminate, nitrate, and nitrite are important because of the great mass of these
components present in DOE tank wastes. If not destroyed before vitrification, nitrate and
nitrite must be dealt with in terms of the off-gas (as NO.) from the LAW and HLW glass
melters . Nitrate leaching from cementitious grout waste forms is also a potent ial ground-
water contamination issue. The presence of aluminate in tank supernates and sludge
washing/leaching solutions calls for careful management of such solutions so that
precipitation or gelation does not occur at inopportune times. Chromate and
pertechnetate ions can also be present the tank supernates and sludge washing/leaching
solutions. Chromate is preferably routed to LAW treatment systems, whereas it might be
necessary to remove pertechnetate.
Carboxylates form another class of anions that are significant to the DOE
environmental mission . A variety of such materials are present in certain Hanford tank
wastes. At the high pH values encountered in these tanks, the carboxyl groups are
deprotonated, forming anionic species that can bind metal ions present in the waste . This
provides a mechanism for enhanced dissolution of radionuclides (e.g ., 90Sr and the
transuranic elements) that would otherwise be poorly soluble under the alkaline tank

2.2. The Environment

Chromate, pertechnetate, and uranyl carbonate complexes are mobile species, and
the uncontrolled spread of these contaminates poses risks to human health and the
environment. A well-characterized example of chromate migration is the chromium
plume at the Hanford 100H Area.' This site is immediately adjacent to the Columbia
River. Furthermore, this stretch of the river is a major spawning area for salmon.
Chromium levels in pore water from the spawning gravel near the 100H Area were
measured to be as high as 130 11g/L. Although this poses no immediate threat to human
health , it is of concern because of the toxicity of Cr(VI) to young salmon. The Fresh
Water Chronic Toxicity criterion set by the U.S. Environmental Protection Agency
(EPA) for Cr is II 11g!L.s With Cr levels entering the river at -10 times above this
criterion, there is obviously reason for concern.
Similar plumes for technetium and uranium contamination at Hanford have been
characterized. A prime example is the U and Tc plumes near the U Plant in the Hanford
200 West Area.' Fortunately, these plumes are located several miles from the Columbia
River . But nonetheless, such contamination spread is undesirable. The average uranium
concentration detected in a well at this site was 1900 I1g/L (-650 pCi/L, assuming
activity due to 238 U) in FY 2000.5 Average 99Tc concentrations of up to 19,400 pCi/L
were also detected in wells near the U Plant.' These are well above the EPA drinking-
water standards of 900 pCi/L and 30 pCi/L for 99Tc and U, respectively. ' These plumes
are the result of previous releases into the environment. Of additional concern, especially
for technetium, is the potential future release caused by wastes leaking from the
underground waste storage tanks . Removal of technetium from the tank wastes is
discussed in more detail below.

It should be noted that the U plume near the Hanford U Plant has not been definitively attributed to anionic
uranyl carbonate complexes. Other mechanisms are possible.


3.1. Removing Problematic Anions from Tank Wastes

In terms of tank-waste issues, we focus our discussion here on removing

pertechnetate and sulfate from the Hanford LAW stream. Investigations to date suggest
that phosphate can be adequately removed from the HLW sludge solids by caustic
leaching, which results in metathesis of phosphates to sodium phosphate and insoluble
metal hydroxldes.i Aluminate ion can largely be managed by controlling solution
temperature and caustic concentration to avoid unwanted precipitation. There are
currently no plans to remove or destroy N0 3' or N0 2' from DOE tank wastes. However,
Hobbs has previously demonstrated electrochemical methods for destroying the latter two
waste components," and Herting has demonstrated that NaN03 can be removed from tank
wastes by fractional crystallization. 10

3././. Pertechnetate

Both ion-exchange and liquid-liquid extraction methods have been investigated for
removing pertechnetate ion from Hanford tank waste. The primary liquid-liquid
extraction methods investigated for removing Tc04 ' from alkaline wastes are
1) extraction with tetraalkylammonium salts and 2) extraction with a crown ether. These
two approaches rely on fundamentally different chemistries as indicated by the following


---- (2)

where N~+ is a lipophilic tetraalkylammonium ion and CE is a crown ether.

In Eq . 1, the pertechnetate is extracted as the counter anion to the
tetraalkylammonium cation. The transfer of Tc04' to the organic phase is mainly driven
by the low hydration energy of Tc04' relative to that of N0 3•• It should be noted that the
Hanford tank wastes contain high concentrations of N0 3 ' , which works against the
process described by Eq. 1. Nonetheless, Tc04' can be extracted from alkaline Hanford
tank wastes with tetraalkylammonium salts. II The disadvantage of using
tetraalkylammonium salts to extract Tc04' is that it is difficult to strip and recover the
Tc04' from the loaded organic phase. Stripping requires the use of very high nitric acid
concentrations or the addition of reductants and complexants. These chemical additions
are undesirable in waste-processing tlowsheets.
The applicability of crown ethers (Eq. 2) overcomes the stripping problems . In the
case of crown ethers, a sodium ion' is complexed by the crown ether and transported to
the organic phase. To maintain electroneutrality, an anion must also be transferred to the
organic phase. Again, because of its relatively low hydration energy, Tc04' is the anion

• Other alkali or alkal ine-earth metal ions may also be complexed by the crown ether.

preferably transferred to the organic phase. The beauty of this method is that contacting
the loaded solvent with water easily reverses the reaction depicted in Eq. 2. This affords
a product Tc stream free of added chemicals. This chemistry forms the basis for the
Strontium and Technetium from Alkaline Waste (SRTALK) process.V The SRTALK
process uses di-t-butylcyclohexano-18-crown-6 to separate both 90Sr and 99Tc from
highly alkaline waste solut ions. Batch contacts with Hanford double-shell slurry feed
demonstrated the potential of the SRTALK process for treating Hanford wastes .'?
However, the liquid-liquid extraction methods have not been pursued further, because of
a decision to focus on chromatographic separations .
Although a number of ion exchangers' have been investigated for separating Tc04'
from tank wastes,'! '!' recent work has focused primarily on using Superl.ig" 639 (IBC
Advanced Technologies, American Fork, Utah) . The chemistry underlying Tc04'
removal with Superl.igf 639 is based on Eq. 2. Superl.ig'" 639 contains a dicyclohexano-
18-crown-6 moiety bound to a polystyrene bead." A number of tests have been
conducted in support of the Hanford Waste Treatment Plant, including those with liquid
waste from tanks AW-IOI 15 and AN-107 . 16 The decontamination factors' were 30 to 35
for AW-10 I but <2 for AN-I 07. The low decontamination factor for AN-I07 is
attributed to the fact that most of the Tc in that particular waste is not in the form of
Tc04 '.
Figure I presents the loading and elution curve for the AW-lO I test. This test
employed two columns connected in series . The actual AW-IOI waste solution was
spiked with 95mTc04' so that the Tc behavior could be monitored in near real time by
gamma spectroscopy. The agreement between the 95mTc and 99Tc results indicates that
the Tc in the AW-10 I waste was indeed mostly Tc04', but there was indication of
breakthrough on non-pertechnetate species early in the experiment. The figure shows
that for this two-column system, in excess of 300 bed volumes of waste solution can be
processed before the Tc limit for the low-activity waste is reached.

3.1,2, Sulfate

It was originally planned to use another Superl.ig" resin for removing from the sot
Hanford LAW stream . However, testing Superl.ig" 655 with waste from Tank AN-I07
showed that this material had little affinity for S04 under those conditions ."
Alternat ive methods were pursued, including precipitation as BaS04 (preceded by
treatment with Ca2+ to remove excess carbonate ion)." This approach proved inadequate,
because large volumes of CaC0 3 were produced, and this material solidified into a
concrete-like solid after standing for a few days . Another approach invest igated was
selective crystallization by evaporating the solution." This approach failed because
Na2S04 did not selectively precipitate. It has recently been suggested that sulfate might
be removed by precipitation as burkeite [N~S04(C03h] , but testing of this idea has not

Forthe purposes of thisdiscussion, we usethe term ionexchange to referto a chromatographic separation

using a solidseparations medium. In many cases, the material investigated operated by mechanisms other
than ionexchange.
t The decontamination factor is defined as the concentration of the contaminantbeforetreatment dividedby
the concentration after treatment.
I D. J. Geniesse, personal communication, COGEMA, Richland, Washington (2001).

3.2. Removing Problematic Anions from Groundwater

There are two basic strategies for removing contam inants from groundwater- pump
and treat or in situ remediation. The pump-an d-treat methodology is gene rally very
expen sive to implement and requires a sustained operational effort. For these reasons,
the DOE has pursued in situ methods for mitigating groundwater contamination. One
such technology- In Situ REDOX Manipulation-aims to reduce the spread of Cr(VI) to
the Columbia River." This method relies on the reduct ion of Cr(VI) to Cr(lll), which is
less toxic and much less mobile in the environment. To achieve this, sodium dithion ite is
inj ected into the aquifer . The dithionite ion reduces ferric iron to ferrous iron according
to Eq. 3. The resulting ferrous hydroxide can then react with chromate ion to form water-
insoluble Cr(lll) hydroxide (Eq. 4).

£0 = +O.56V (3)


70 tI"c.9U.... 2.U I
, I
T c·95m 1". 213 I I 00

60 lAW rbu limit. 0.1 Ci/m
(0.027 Ci/m' in reod) """0 0 •
40 • o •
Tem.per::lture= 21.5 °c
Feed role = 6 BV l lu
e 0

Bed wlume 4.7 mL 0
AW· IO\, (Nal = 4.6 M

s 20 SL-639 balCh # = 980624001DC

<, 99Tc C, = 64.1 uCi/L (3.75 me /L) 0
10 •0 v
5 0

~ • v

Lead column. "Tc
Lead co lumn. '"-reO'-
0.5 P V
• Lag Culumn. "Tc
0.2 Iv V LaC Culumn.

10 100
Bed Volumes
Figur e 1. Breakthrough curve for removal of Tc from Hanford tank AW-IOI supernate with Superl.ig" 639.
Taken from Ref. 15.

Crol + 3Fe(OH)2 + 4H 20 ~ Cr(OHh + 3Fe(OHh + 20H- EO = +0.43V (4)

In this manner, the spread of Cr(VI) in the environment can be reduced. The In Situ
REDOX Manipulation technology was recently demonstrated at the Hanford IOOH
Area." Chromate concentrations in the test zone were 60 ppb before the test, but
dropped to less than 8 ppb after deployment of the In Situ REDOX Manipulation
technology. Core analyses showed that 60 to IOO% of the available reactive iron was
reduced within a 25-ft radius of the injection well. Furthermore, the treatment zone
rema ined anoxic, and the chromate concentrations were still below detection limit 2 years
after the dithionite injection.
A similar methodology has been proposed for controlling technetium migration. In
this case, a barrier would be constructed containing iron filings or zero-valent iron
dispersed into a gel matrix . . The iron metal would reduce Tc04- to a lower oxidation
state, thereby reduc ing Tc migrat ion.


Anion separations are critical to the success of DOE 's environmental remediation
mission. Problems involving anions manifest themselves in a variety of matrices, ranging
from groundwater (dilute) to tank waste (high salt). Technologies are being developed
and implemented to address some of the issues involving anions. For example, in situ
redox technology has recently been deployed to reduce the migration of Cr(VI) in
Hanford groundwater. Promising liquid-liquid extraction and ion-exchange technologies
exist for pertechnetate separations. On the other hand , technologies are lacking for
removing sulfate from highly alkaline, high-salt, tank liquids. There are other areas
involving anions where technological solutions might be important, for example,
removing nitrate from tank liquids (or destroying them) and in situ methods to prevent
the migration of uranyl carbonate complexes in the environment.


The preparation of this paper was supported by DOE through the Laboratory
Directed Research and Development Program at the Pacific Northwest National
Laboratory. The author thanks W. C. Cosby and B. M. Rapko for reviewing this
document and John Vienna and Douglas Sherwood (U .S. Environmental Protection
Agency) for their helpful discussions.


I. T. L. Stewart, J. A. Frey, D. W. Geiser, and K. L. Manke, Overview of U.S. Radioactive Tank Problem,
in: Science and Technology for Disposal of Radioactive Tank Waste, edited by W. W. Schulz and N. J.
Lombardo (Plenum Press, New York, 1998), pp. 3-13.
2. U.S. Department of Energy, 1996 Baseline Environmental Management Report, available at

3. C. M. Jantzen , K. G. Brown, and J. B. Pickett, Impact of Phase Separation on Durability in Phosphate

Containing Borosilicate Waste Glasses : Relevance to Vitrification of INEEL High Level Waste (U),
Report WSRC-MS-20oo-00307 (Westinghouse Savannah River Company, Aiken, South Carolina, 2000)
presented at the 102""American Ceramic Society Annual Meeting in St. Louis, Missouri, May 2, 2000.
4. H. Li, P. Hrma, and J. D. Vienna, Sulfate Retention and Segregation in Simulated Radioactive Waste
Borosilicate Glass Ceram. Trans., 119,237-246 (2001).
5. M. J. Hartman, L. F. Morasch, and W. D. Webber, Hanford Site Groundwater Monitoring for Fiscal Year
1999, Report PNNL-13 I 16 (Pacific Northwest National Laboratory, Richland, Washington, 2000).
6. S. J. Hope and R. E. Peterson, Pore Water Chromium Concentration at loo-H Reactor Area Adjacent to
Fall Chinook Salmon Spawning Habitat of the Hanford Reach, Columbia River, Report BHI-00345,
Rev. I, (Bechtel Hanford, Inc., Richland, Washington, 1996).
7. 40 CFR 141,2000, U.S Environmental Protection Agency, National Primary Drinking Water Regulations,
"U.S. Code of Federal Regulations:'
8. G. 1. Lumetta, B. M. Rapko, 1. Liu, and D. J. Temer, Enhanced Sludge Washing for Pretreating Hanford
Tank Sludges, in Science and Technology for Disposal of Radioactive Tank Wastes, edited by W. W.
Schulz and N. 1. Lombardo (Plenum Press, New York, 1998), pp. 203-218.
9. D. T. Hobbs, Electrolytic Denitration of Radioactive Savannah River Site Waste (U), Report WSRC·TR-
95-0176, (Savannah River Technology Center, Aiken, South Carolina, 1995).
10. D. L. Herting, Clean Salt Process Final Report, Report WHC-EP-0915 (Westinghouse Hanford Company,
Richland, Washington, 1996).
II. G. J. Lumetta, L. A. Bray, D. E. Kurath, 1. R. Morrey, J. L. Swanson, and D. W. Wester, Exploratory
Study of Complexant Concentrate Waste Processing, Report PNL·8438 (Pacific Northwest Laboratory,
Richland Washington, 1993).
12. R. A. Leonard, C. Conner, M. W. Liberatore , P. V. Bonnesen, D. J. Presley, B. A. Moyer, and G. J.
Lumetta, Developing and Testing an Alkaline-Side Solvent Extraction Process for Technetium Separation
from Tank Waste, Sep. Sci. Technol. 34(6&7),1043-1068 (1999).
13. D. L. Blanchard, Jr., G. N. Brown, S. D. Conradson, S. K. Fadeff, G. R. Golcar, N. J. Hess, G. S. Klinger,
and D. E. Kurath, Technetium in Alkaline, High-Salt, Radioactive Tank Waste Supernate: Preliminary
Characterization and Removal, Report PNNL-11386 (Pacific Northwest National Laboratory, Richland,
Washington 1997).
14. N. M. Hassan, W. D. King, D. 1. McCabe, L. L. Hamm, and M. E. Johnson, SuperLig® 639 Equilibrium
Sorption Data for Technetium From Hanford Tank Waste Supernates, Solvent Extr. Ion Exch., 20,211-
226 (2002).
IS. D. L. Blanchard, Jr., D. E. Kurath, and J. R. Bontha, Small Column Testing of SuperLig® 639 For
Removing ~c From Hanford Tank Waste Envelope A (Tank 241-AW -IOI), Report PNWD-3004
(Battelle, Pacific Northwest Division, Richland, Washington, 2000).
16. D. L. Blanchard, Jr., D. E. Kurath, and B. M. Rapko, Small Column Testing of SuperLig® 639 For
Removing ~c From Hanford Tank Waste Envelope C (Tank 241.AN.I07), Report PNWD- 3004
(Battelle, Pacific Northwest Division, Richland, Washington, 2000).
17. D. E. Kurath, J. R. Bontha, D. L. Blanchard, Jr., S. K. Fiskum, B. M. Rapko, Ion Exchange Studies for
Removal of Sulfate from Hanford Tank Waste Envelope C (241-AN-107) Using SuperLig® 655 Resin,
Report PNWD·3053 (Battelle, Pacific Northwest Division, Richland, Washington, 2000).
18. S. K. Fiskum, D. E. Kurath, and B. M. Rapko, Development and Demonstration ofa Sulfate Precipitation
Process for Hanford Tank Waste 241·AN.I07, Report PNNL-3050 (Battelle, Pacific Northwest Division,
Richland, Washington, 2000).
19. G. J. Lumetta, G. S. Klinger, D. E. Kurath, R. L. Sell, L. P. Darnell, L. R. Greenwood, C. Z. Soderquist ,
M. J. Steele, M. W. Urie, and J. J. Wagner, Removal of Sulfate Ion From AN·107 by Evaporation, Report
PNWD·3036 (Battelle, Pacific Northwest Division, Richland, Washington, 2000).
20. U.S. Department of Energy, In Situ Redox Manipulation, DOE/EM·0499 (2000), available at
21. N. E. Korte, L. Liang, B. Gu, M. T. Muck, J. L. Zutman, R. M. Schlosser, R. L. Siegrist, T. C. Houk, and
Q. Fernando, In Situ Treatment of Mixed Contaminants in Groundwater : Application of Zero-Valence
Iron and Palladized Iron for Treatment of Groundwater Contaminated with Trichloroethene and
Technetium-99, Report ORNLITM·13530 (Oak Ridge National Laboratory, Oak Ridge, Tennessee, 1998).
22. S. M. Ponder and T. E. Mollouk, Powerful Reductant for Decontamination of Groundwater and Surface
Streams, ENVR-149, U.S. Patent 6,242,663, (American Chemical Society, Washington, D.C., 2001).

JosephM. Mahoney, John P. Davis, and Bradley D. Smith'


Two potentially useful anion-separation methods are liquid extraction and Iiquid-
membrane transport . These two related methods employ an organic-soluble receptor
molecule to selectively extract an otherwise insoluble target anion into an organic phase.
If the receptor is a neutral compound, then a cation must accompany the extracted anion,
which means that the anion-separation problem is actually a salt-separation problem.
Over the past ten years, a large number of neutral receptors have been prepared and
evaluated for their ability to associate with inorganic anions in organic solvents.' Almost
all of these binding studies have used guest anions with noncompeting
tetralkylammonium counter-cations . However, in many anion-separation situations, the
luxury of noncompeting counter-cations is not available, and ion-pairing of the target
anion with its counter-cation can become problematic . In particular, NMR titration
experiments show that the presence of alkali cations can reduce receptor/anion
association constants. For example, titration of N-phenyl-N' -octyl urea, 1 (Scheme 1), in
DMSO-d6 with tetrabutylammonium acetate leads to Kacetate = 310 ~l. 2 Repeating the
titration experiment in the presence of one molar equivalent of metal tetraphenylborate
changes Kacetate in the following way: Cs+ (340 ~I); K+ (200 ~l); Na+ (90 ~I). A more
dramatic example is reflected by the lH NMR titration curves shown in Figure 1.
Titration of 1 with tetrabutylammonium dihydrogenphosphate in CD3CN results in the
expected 1:1 binding isotherm, and an association constant of 400 ~l can be extracted
by curve-fitting methods.' If the titration is repeated with the initial host solution
containing one molar equivalent of potassium tetraphenylborate, then a very different
isotherm is observed . The first one-molar equivalent of dihydrogenphosphate has no
affinity for the urea NH residues, but subsequent additions produce a "normal looking"

Department of Chemistry andBiochemistry. University of NotreDame. NotreDame. IN 46556-5670. USA.

Phone: 219-631-8632, Fax: 219-631-6652, E-mail:

Fundamentals and Applications ofAnion Separations, edited by

Moyerand Singh,Kluwer AcademicIP1enum Publishers, NewYork,2004 115

titration curve. Apparently, the added K+ is able to stoichiometrically sequester the

dihydrogenphosphate, which prevents host/anion binding.


0 9.0 C •
(ppm) C



0 10 20 30 40
[dihydrogenphosphate1 (mM)

Figure I. NH chemical shift (5) for 1 (initially 10 mM) in CD3CN and 295 K upon addition of
tetrabutylammonium dihydrogenphosphate; • presence and 0 absence of potassium tetraphenylborate
(initially 10 mM). (Reprinted withpermission from Ref. 9. Copyright 2000American ChemicalSociety.)

In general, receptor/anion association constants are likely to be reduced in organic

solvents because of the ion-pairing shown in Figure 2. In most organic solvents,
inorganic salts do not exist as free ions; instead they are present as solvent-separated ion-
pairs, contact ion-pairs, and/or aggregated contact ion-pairs.' Ion-pairing with a
competing cation diminishes anion basicity, and can lower host/anion association
constants by two ways. (i) The metal cation can sterically hinder the host/anion
interaction. This steric effect increases with the degree of ion-pair aggregation. (ii) The
associated cation lowers the anion 's effective charge by either a polar ization or a
shielding effect. This electrostatic effect also increases with the degree of ion-pair

M+ + A- = M+A- = [M+A-j

Figure 2. Saltscan exist as free ions, ion-pairs and/oraggregated ion-pairs in organic solvents. (Reprinted with
permissionfromRef.9. Copyright 2000American Chemical Society.)

One way to alleviate this cation-induced inhibition of host/anion binding is to use a

dual-receptor strategy, that is, a binary mixture of anion and cation receptors," The dual-
receptor strategies reported so far have used either a crown ether or a calixarene as the
cation-binding compound and a tripodal amide, cyclic peptide, thiourea or zinc porphyrin

as the anion-binding compound." These dual-receptor systems have been shown to

promote solid/liquid extraction and membrane transport of salts.
An alternative approach to salt binding is to design a single ditopic receptor that can
simultaneously bind both the cation and the anion.5•6.7 Covalently linking the anion and
the cation binding sites raises the possibility of bindingcooperativity. That is, the binding
of the first ion can alter the ditopic receptor structurally (induced fit)7. or electronically
(through-bond electrostatic effects) 6t.7i.8 in a way that the second ion binds more tightly.
Most ditopic salt receptors described in the literature have spatially separated cation- and
anion-binding sites, so that the anion and the cation are not in contact with each other
when they are boundto the receptor. As a result, binding cooperativities are often low. In
our laboratory, the aim is to develop ditopic receptors with juxtaposed binding sites that
allowthe salt to bindto the receptor as an associated ion-pair (Figure3). In the following
sections we describe our salt receptor designs and summarize their complexation
properties. We finish with a discussion of the potential utility of ditopic salt receptors in
liquid extraction and membrane separation processes.

("0] ff r0p
c O

m ···Pt'

Scheme 1: Monotopic and ditopic receptors.

Figure 3. Ion-pair recognition using heteroditopic salt-binding host. (Reprinted with permission from Ref. 9.
Copyright 2000AmericanChemical Society.)


Initially, we tested the simple idea of using a flexible linker to connect a cation-
binding site to an anion binding site," We found that this approach did not produce highly
cooperative salt binding, which we define operationally as the enhancement of anion
binding constant induced by the presence of one molar equivalent of metal cation. For
example, with compound 2 the presence of sodium or potassium cations hardly enhances
or sometimes even lowers anion-association constants. We reasoned that receptor
flexibility produced entropic penalties and also the exposed binding sites were probably
strongly solvated. Thus, we designed more preorganized second-generation ditopic
macrobicycles with the expectation that they would form complexes like that shown
generically as structure 3. We prepared recegtor 4 in six steps and found that it bound
alkali halides in solution and solid-state.' Moreover, the presence of sodium or
potassium cation enhanced chloride binding by about a factor of ten. An X-ray crystal
structure of a 40NaCI
complex (Figure 4) showed that the salt is bound as a solvent
separated ion-pair. The bridging solvent molecule is either a chloroform (62% relative
occupancy) or two water molecules (38% relative occupancy). The salt binding
cooperativity for this receptor appears to be due to the anion interacting with the cation
via the bridging solvent molecule(s).
We felt that anion-binding enhancement could be further improved if the receptor
bound the salt as a contact ion pair. This would increase the anion 's electrostatic
attraction to the receptor/cation complex. Thus, we prepared third-generation
macrobicycle S with a smaller cavity and found that it could slowly extract solid alkali
halides into chloroform solution. 11 The extractions could be easily monitored by IH NMR
because the receptor/salt complex was in slow exchange with the free receptor. The
receptor/salt complexes are stable enough to survive column chromatography using silica
gel and weakly polar solvents. We were pleased to learn that the presence of alkali metal
cation greatly enhanced halide-binding constants. For example, the s/cr association
constant in CDCi):DMSO-d6 (85:15) was increased from 80~) to 2.5 X 104 ~) by the
presence of one molar equivalent of potassium tetraphenylborate. This remarkable 300-
fold enhancement in association constant is considerably higher than that observed with
most, if not all, other salt binding systems. NMR titration experiments show that receptor
5 binds KCl better than NaCI. X-ray crystal structures of the SoNaCi and SoKCl
complexes confirm that the salts are bound as contact ion pairs (Figure 5), and that the
MX bond is shorter than that for solid MX. An explanation of why K+ is better than Na+
at enhancing cr binding is provided by a comparison of the corresponding crystal
structures. For NaCI the diazacrown alters its conformation and reduces its size to
accommodate the smaller Na+ ion which is only coordinated by the four crown oxygens
and one of the two nitrogens. The average distance from the four diazacrown oxygens to
the cr is 4.20 A, which is significantly shorter than the 4.71 A distance observed in the
KCI complex crystal structure. This is likely the major reason why K+ enhances s/cr
binding better than Na+. After binding a K+ ion, receptor S is able to nicely accommodate
a cr and form a favorable contact-ion pair. However, in the case ofNa+, the advantage of
forming a contact-ion pair is offset by increased ion-dipole repulsions between the cr
and the diazacrown oxygens that closely coordinate the Na+. In other words, the
difference in salt affinities seems to be due to subtle changes in cation solvation by the
crown heteroatoms.

Figun 4. X-ray crystal structure of 4oNaCIoCHCk (Reprinted with permission from Ref 10. Copyright 2000
American Chemical Society.)

Figure S. X-ray crystal structures Top SoKCI, Bottom: SoNaCi (Reprinted with permission from Ref 11.
Copyright 200I American Chemical Society)



If both target ions are very hydrophilic, then any separation process that uses liquid-
extraction or membrane-transport methods will be improved with salt-binding receptors .
The subsequent question is whether to use a binary mixture of anion and cation receptors
(dual-receptor strategy) or a ditopic receptor. Generally, the monotopic ion receptors are
easier to prepare; thus the dual receptor system will likely have an economic advantage.
So what is gained by using a ditopic salt-binding receptor? With simple liquid extraction,
the major benefit is that a ditopic receptor can selectively bind and extract salts with very
high affinities. Therefore, ditopic receptors may be appropriate for trace removal
In the case of liquid-membrane transport, the receptors extract a target salt from an
aqueous source phase into a liquid membrane, then carry it across the membrane, and
release it into a separate aqueous receiving phase. Typically, the rate-determining step in
the transport process is diffusion of the carrier/salt complex across the unstirred boundary
layers at the membrane interfaces. On the basis of the Stokes-Einstein relationship (D =
kT/61t11T, where D is the diffusion constant, k is the Boltzmann constant, T is temperature,
TI is the viscosity, and r is the radius of the complex), the diffusional velocity of two large
carriers should be lower than one ditopic carrier; however, there is literature evidence
suggesting that this may not be the case. In 1999, Reinhoudt and coworkers reported a
detailed study that compared the CsCI transport ability of ditopic carrier 6 (Figure 6) to a
binary mixtures of monotopic carriers 7 and 8, and the KCI transport ability of ditopic
carrier 9 to a mixture of monotopic carriers 7 and lO.4e By measuring flux as a function of
source phase salt concentration , they were able to derive carrier saturation curves and
compute diffusion constants for the saturated carriers. Counter-intuitively they found, in
both cases, that the diffusion constant for the ditopic carrier was significantly lower than
that for the corresponding mixture of monotopic carriers (Table I). The authors did not
offer any detailed explanation for their observations, but one possibility is the saturated
ditopic carriers are quite polar and that their diffusional motions are slowed by
carrier/solvent interactions. This raises the idea that zwitterionic receptor/salt complexes

9 n= 1, R=2-e thyhexyl 1 0 n =1 7
6 n=2, R=2-e thylhexyl 8 n=2

Figure 6. Reinhoudt's carriers for salt transport . (Reprinted with permission from Ref. 4e. Copyright 1999
American Chemical Society.)

Table 1. Diffusion results for ditopic carriers and

mixtures of anion and cation carriers. Data from Ref.
CsCI 4.2
_ __ _ _ _' ''''''' NH._ N

CsCI 5.2
KCI 3.9
KCI 6.8

with large molecular dipoles have inherently slow diffusion constants. To test this
hypothesis, we prepared the known diastereoisomeric polyether Ca2+ complexes cis-11
and trans-Ll and measuredtheir self-diffusion constants(Table 2). These complexes have
the same molecular formula, but their X-ray structures indicate that they have very
different moleculardipoles. 12 We used Pulsed Field GradientNMRI3 to measure the self-
diffusion constants for cis-11 and trans-11 in methanol at 24°C and found that they are
essentially identical. We conclude that large differences in molecular dipole do not
necessarily produce large differences in diffusion constants in a polar solvent. At present
there appears to be no clear explanation for the difference in diffusion constants for
Reinhoudt's transport carriers. One possibility is that the saturated ditopic carriers self-
aggregate in the weakly polar organic solvents used for liquid membrane transport, and
so diffusemore slowly.However, additional studiesare neededto test this hypothesis.

Table 2. Self-Diffusion Constants of 11 at 24°C in Methanol-d,


cis-11 0.Ql5 0.65

/rans-11 0.Ql5 0.59

cisltrans 0.00510.010 0.62 0.61


ooe ~o coo"
cis-11 IrWls-11
122 J. M. MAHONEY st AL.


We have demonstrated that a highly effective way to extract hydrophilic salts into
organic phases is to use ditopic receptors that bind the salts as their contact ion-pairs. Our
results suggest that ditopic salt receptors offer real advantages for anion separation using
simple liquid extraction. Contrary to expectations, there is literature evidence indicating
that saturated ditopic salt receptors have unusually slow diffusional velocities. The reason
for this is not clear, but it is does not appear to be due to stronger carrier/solvent
interactions. Based on the current evidence, it is reasonable to conclude that no obvious
benefit is gained by using a ditopic salt receptor instead of a binary mixture of monotopic
cation and anion receptors (dual-receptor strategy) for salt transport across liquid organic


This work was supported by the University of Notre Dame and the National Science


I. The Supramolecular Chemistry of Anions, edited by A. Bianci, K. Bowman-James, and E. Garcia-Espana

(VCH,Weinheim, 1997).
2. M. P. Hughes and B. D. Smith,J. Org. Chem. 62,4492-4501 (1997).
3. (a) M. Swarc in Vol. I, Ch. I. of Ionsand Ion Pairsin Organic Reactions, edited by M. Swarc (Wiley, New
York, 1972)(b) J. Smid, Angew. Chem. Inti. Ed. 11, 112-127(1972). (c) F. G. Bordwell, J. C. Branca,D.
L. Hughes, and W. N Olmstead, J. Am. Chem. Soc. 45, 3305-3313 (1980). (d) M. J. Kaufman and A.
Streiwieser,J. Am. Chem. Soc. 109, 6092--{)097 (1987). (e) J. A. Kromand A. Streitwieser, J. Org. Chem.
61, 635<Hi359 (1996).
4. a) Q. Qian, G. S. Wilson, K. Bowman-James and H. H. Girault, Anal. Chem. 73,497-503 (2001). b) K.
Kaval1ieratos, R. A. Sachleben, G. J. Van Berkel, and B. A. Moyer, J. Chem. Soc.. Chem. Comm.,
187-188 (2000). c) K. Kavallieratos, A. Danby, G. J. Van Berkel, M. A. Kelly, R. A. Sachleben, B. A.
Moyer, and K. Bowman-James, Anal. Chem. 72,5258-5264 (2000). d) S. Kubik, J. Am. Chem. Soc 121,
5846-5855 (1999). e) L. A. J. Chrisstoffels, F. De Jong, D. N. Reinhoudt, S. Sivelli, L. Gazzola, A.
Casnati and R. Ungaro, J. Am. Chem. Soc 121, 10142-10151 (1999). f) D. M. Rudkevich, A. N.
Shivanyuk, Z. Brzozka, W. Verboom, and D. N. Reinhoudt, Angew. Chem. Int. Ed. Engl. 34,2124-2126
5. Reviews that discuss salt-binding hosts: (a) P. A. Gale and P. D. Beer, Angew. Chem. Int. Ed. Engl. 40,
486-516 (2001). (b) M. M. G. Antonisse and D. N. Reinhoudt, J. Chem. Soc.. Chem. Comm., 443-447
(1998). (c) E. Kimuraand T. Koike, J. Chem . Soc. Chem. Comm. 1495-1500(1998). (d) 1. Stibor, D. S.
M. Hafeed, P. Lhotak, J. Hodacova, J. Koca, and M. Cajan, GazellaChimica Italiano 127, 673--{)85
(1997). (e) M. T. Reetz, in: Simultaneous Binding of Cations and Anions, edited by 1. L. Atwwod, F.
Vogtle, J. Davies, E. D. MacNicol, J.-M. Lehn, and G. W. Gokel, Vol. 2, 553-562 (Elsevier Science,
6. Earlierexamples of salt-binding receptors: (a) H. Tsukube, Tet. Lett. 22,3981-3984 (1981). (b) H. Tsukube,
J. Chem Soc., Perkin Trans. J, 2359-2363 (1982). (c) M. Zinic, L. Frankec, V. Skaric, J. Trafton, and G.
W. Gokel, J. Chem. Soc. Chem. Commun. 1726-1728 (1990). (d) U. Olsher, F. Frolow,N. K. Dalley, J.
Weiming, Z. Y. Yu, J. M. Knoleoch, and R. A. Bartsch, J. Am. Chem. Soc. 113, 657<Hi574 (1991). (e) M.
T. Reetz, C. M. Niemeyer, and K. Harms, Angew. Chem. Int. Ed. Engl. 30, 1474-1476 (1991). (f) E. A.
Arafa, K. 1. Kinnear, and 1. C. Lockhart, J. Chem. Soc.• Chem. Comm., 61--{)4 (1992). (g) W. Petit, Y.
Iwai, C. F. Barfknecht, and D. C. Swenson, J. Heterocyclic Chem. 29, 877-881 (1992). (h) S. S. Flack, J.-
L. Chaumette, 1. Kilburn, G. J. Langley, and M. Webster, J. Chem. Soc.• Chem. Comm., 399-400 (1993).
(i) S. Badouche, L. Jaquet, M. A. Lobo-Recio, C. Marzin, and G. Tarrago,J. Inclus. Phenom 138, 81-89

(1993) . (j) K. I. Kinnear, D. P. Mousley, E. A. Arafa,and J. C. Lockhart, J. Chem. Soc., Dalton Trans.,
3637-3643 (1994). (k) D. M. Rudkevich, Z. Brzozka, M. Palys, H. C. Visser, W. Verboom, and D. N.
Reinhoudt, Angew. Chem. Int. Ed. Engl 33,467-468 (1994). (I) P. B. Savage, S. K. Holmgren, and S. H.
Gellman, J. Am. Chem. Soc 116, 4069-4070 (1994). (m) T. Nagasaki, H. Fujishima, M. Takeuchi, and S.
Shinkai,J. Chem. Soc.. Perkin Trans. 1,1883-1888 (1995). (n) D. M. Rudkevich, J. D. Mercer-Chalmers,
W. Verboom, R. Ungaro, F. de Jong, and D. N. Reinhoudt, J. Am. Chem. Soc 117, 6124-{)125 (1995). (0)
Y. H. Kim, J. Calabrese, and C. McEwen, JAm. Chem. Soc. 118, 154.5-1546 (1996). (p) N. Voyerand B.
Guerin, J Chem. Soc. Chem. Comm. 2329-2330 (1997). (q) T. Nabeshima, A. Hasiguchi, S. Yazawa, T.
Haruyama, and Y. Yano, J. Org. Chem. 63,2728-2789 (1998). (r) P. D. Beer and S. W. Dent, J. Chem.
Soc., Chem. Comm., 825-826 (1998) . (s) P. D. Beer and 1. B. Cooper, J. Chem. Soc., Chem. Comm.,
129-130 (1998). (t) N. Pelizzi, A. Casnati, A. Frigerri, and R. Ungaro, J. Chem. Soc., Perkin Trans. 2,
1307-1311 (1998). (u) J. E. Redman, P. D. Beer, S. W. Dent, and M. G. B. Drew,J. Chem. Soc.. Chem.
Comm., 231-232 (1998). (v) J. C. Bryan, R. A. Sachleben, J. M. Lavis, M. C. Davis, 1. H. Bums,and B. P.
Hay,lnorg. Chem. 37,2749-2755 (1998).
7. Recent examples of salt-binding receptors: (a) P. D. Beer, P. K. Hopkins, and J. D. McKinney, J Chem. Soc.,
Chem. Comm., 1253-1254 (1999). (b) M. D. Barboiu, N. D. Hovnanian, C. Luca,and L. Cot, Tetrahedron
55,9221-9232 (1999). (c) D. 1. White, N. Laing, H. Miller, S. Parsons, S. Coles, and P. A. Tasker, J.
Chem. Soc., Chem. Comm., 2077-2078 (1999) . (d) A. Lutzen, A. R. Renslo, C. A. Schalley , B. M.
O'Leary and, 1. Rebek, JAm. Chem. Soc 121,7455-7456 (1999). (e) S. Kubik and R. Goddard, J. Org.
Chem. 64, 9475-9486 (1999). (f) S. L. De Wall, L. J. Barbour, and G. W. Gokel, J. Am. Chem. Soc 121,
8405-8406 (1999). (g) E. S. Meadows, S. L. De Wall, L. J. Barbour, and G. W. Gokel, J. Chem. Soc.,
Chem. Commun. 1557-1558 (1999). (h) S. Bartoli and S. Roelens, J. Am. Chem. Soc. 121, 11908-11909
(1999) . (i) S. Nishizawa, K. Shigemori, and N. Teramae, Chem. Lett., 1185-1186 (1999). U) V.
Andrisano, G. Gottarelli, S. Masiero, E. H. Heijne, S. Pieraccini, and G. P.Spada, Angew. Chem. Int. Ed.
38,2386-2388 (1999). (k) T. G. Levitskaia, J. C. Bryan, R. A. Sachleben, J. D. Lamb, and B. A. Moyer, J.
Am. Chem. Soc. 122,545-562 (2000). (I) K. Kavallieratos andB. A. Moyer, J Chem. Soc.. Chem. Comm.,
1620-1621 (2001) . (m) S. Memon and M. Yilmaz, J. Mol. Struct. 595, 101-109 (2001). (n) T.
Nabeshima, T. Hanami, S. Akineand T. Saiki, Chem., 560-561 (2001). (0) X. Shi, J. C. Pettinger,
and J. T. Davis, Angew. Chem. Int. Ed. 40, 2827-2830 (2001). (p) T. Tuntulani, S. Poompradub, P.
Thavomyutikam, N. Jaiboon, V. Ruangpomvisuti, N. Chaicit, Z. Asfari, and J. Vicens, Tetrahedron Leu.
42,5541-5544 (2001) . (q) L. H. Uppadine, 1. E. Redman, S. W. Dent, M. G. B. Drew, and P. D. Beer,
Inorg. Chem. 40, 2860-2869 (2001).
8. T. Tozawa, Y. Misawa, S. Tokita, andY. Kubo, Tetrahedron Lett. 41, 5219-5223 (2000).
9. R. Shukla, T. Kida, and B. D. Smith, Org. Lell. 2, 3099-3102 (2000).
10. M. 1. Deetz, M. Shang, andB. D. Smith, J. Am. Chem. Soc 122,6201-6207 (2000).
II. J. M. Mahoney, A. M. Beatty, andB. D. Smith, JAm. Chem. Soc 123, 5847-5848 (2001).
12. J. P. Behr, 1. M. Lehn, D. Moras, and J. C. Thierry, JAm. Chem. Soc. 103,701-703 (1981).
13. O. Mayzal and Y. Cohen, J. Chem. Soc., Chem. Comm., 1901-1902 (1994).

Konstantinos Kavallieratos.l" Jeffrey C. Bryan.' Richard A. Sachleben.'

Gary J. Van Berkel. ' Oscar D. Espetia,' Melissa A. Kelly ,2
Andrew Danby.' Kristin Bowman-James.l" and Bruce A. Moyer'"


Extraction of alkali meta l salts by designed cation hosts, such as crown ethers I and
calixarenes,' has been widely investigated in recent years, as driven , to an extent , by the
importance of separations of certain cationic contaminants such as Cs + and Sr 2 + in
nuclear- fuel reprocessing and waste remediation. ' New cation hosts used as extractants
for these and other metals have reached impressive levels of selectivity for such
demanding "needle -in-the-haystack " applications. The success of neutral cat ion hosts
used in ion-pair extract ion systems stems in part from the practical advantage of stripping
extracted salts with water, resulting in processes that yield a pure salt product stream with
little secondary waste. Examples of such practical ion-pair extraction applications include
processes for nuclear-waste treatment that produc e aqueous streams of relatively pure
cesium nitrate/" or sodium pertechnetate,' suitable for vitrification and geologic disposal.
These processes employ cal ix-crown and crown ether extractants, such as those shown in
Figure I for cesium. Although ion-pair extraction involves anion co-extraction, ligand
design has focused primarily on the design of the cation host, typically with little or no
attention to accommodat ing the anion.

To whom correspondence should be addressed.

Florida International University, Department of Chemistry, 11 200 SW 81h St., University Park, Miami, FL
33199-000I. Email:
Chemical Sciences Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 37831-6119.
, Department of Chemistry, Unive rsi ty of Kansas, Lawrence, KS 66045. Email :

Fundamentals and ApplicationsofAnion Separations, edited by

Moyerand Singh, K1uwer AcademiclPlenum Publishers, NewYork, 2004 125

Figure 1. Two effective Cs+host frameworks that havebeenusedfor cesium extraction studies: tetrabenzo-24-
crown-S (I) and calix[4)arene-bis(tert-octylbenzo-crown-6) (BOBCalixC6, 2), shown here as their Cs+
complexes. The lipophilic "tails"(suchas the tert-octyl one for the calixarene receptor) are addedfor increased
solubility in hydrocarbon solvents.

Extraction of salts from water to relatively less polar organic media normally suffers
from unfavorable dehydration-resolvation energetics.' In salt-separation systems
employing neutral cation receptors as extractants, the unfavorable energetics for
partitioning of the anion to the solvent phase contributes to weak extraction,
commonly referred to as the "anion problem." Additionally, the extraction selectivity
for the anion in the absence of an anion receptor generally follows the well-known
Hofmeister series," which Hofmeister originally associated with differences in anion
hydration. Since an ion hydration tends to dominate over anion solvation by organic
solvents, the Hofmeister series is the "normal" order observed in extraction phenomena.
The resultant anion-partitioning Gibbs energies are directly related to differences in
charge-to-size ratios for various anions" and may be collectively described by the
empirical observation that lor univalent anions, extractability tends to increase with
increasing anion thermochemicalradius. Selectivity trends that depend monotonically
on a single physical variable, such as anion radius, are said to exhibit "bias" versus
"recognltlon.v'" Although the bias phenomenon in anion separations has provided a
powerful tool for separating anions based on their size, the anion problem has led
researchers to investigate methods by which the partitioning of small inorganic
anions is avoided altogether. Accordingly, three different approaches have been
pursued: i) addition of hydrophobic anions to the aqueous phase," ii) addition of
synergistic organophilic acids to the organic phase," or iii) use of ionizable
macrocycles." Such methods have been successful but have not addressed the
scientifically interesting question related to facilitation of the extraction of small
anions. One might anticipate applications, for example, wherein it might be
desirable to remove a salt from an aqueous solution without otherwise changing the
composition of the aqueous solution. This cannot be the case with systems based
on ion exchange or the addition of more extractable anions. One approach to

facilitating anion partit ioning entails employing organic diluents!" or diluent

modifiers!' that prov ide enhanced solvation of the anion. In practical extraction
processes by cation hosts, a lipophilic alcohol is normally added to the hydrocarbon
organic phase as a diluent modifier with the dual purpose of: i) solvating the anion and ii)
increasing the solubility of the cation host in the diluent. This latter purpose, is also
served by the incorporation of lipophilic tails (such as tert-butyl or tert-octyl) to the
crown ether framework (e.g., 2 in Figure I). A variation on the use of an alcohol
diluent modifier is the incorporation of the hydroxy group in the cation host
itself. 12 In this case, the hydroxy group may potentially interact with the anion,
thereby providing additional stabilization for the extraction complex.
More recently, the ever-increasing sophistication in the design of anion receptors'?
and combined ditopic host molecules that can bind both an anion and a cation" have led
to increased interest in separations targeted at specific ion pairs. IS Some recent elegant
examples involve salt transport across supported liquid membranes'" by anion
carriers.l'" combinations of cation and anion carriers,17a.b or elegant ditopic
receptors combining both anion and cation recognition sites."
The utilization of anion receptors offers a potential means for gaining better control
of the anion co-extraction and for escaping from reliance solely upon ion-solvation
effects. The potential benefit of complexing the anion is two-fold: i) The use of an anion
host can potentially provide an additional favorable energetic term corresponding to the
I:iGO of anion binding, counterbalancing to an extent the unfavorable energetics for the
anion partitioning. ii) By using an anion host, opportunities arise for altering the anion
extraction selectivity according to principles based on the geometric and electronic
complementarity between the host and guest anion. In comparison to the ditopic salt-
receptor option, the dual-host approach presents the additional advantage of synthetic
simplicity and versatility. Given the present availability of multitudes of possible pairs of
extractants that can be chosen for their specificity for target salts, and providing that such
extract ants can be used without mutual interference, economical processes may be
envisioned in which targeted salts are extracted by design from a complex salt mixture.
Moreover, when both hosts are known and can be readily purchased or synthesized, dual-
host processes can be developed rapidly without necessarily involving lengthy and
laborious ligand design.
Herein, we discuss some recent developments of the dual-host strategy using
synergistic combinations of crown ether cation hosts and tripodal amide-type anion hosts
for extraction of cesium nitrate. The systems examined are based on three well-known
tripodal backbones: i) 1,3,5-benzenetricarboxylic acid (trimesic acid) , ii) tris(2 -
aminoethyl)amine (tren), and iii) 1,3,5-tris(aminomethyl)benzene. Anion binding was
further demonstrated by electrospray ionization mass spectrometry (ESI-MS) and
quantified by NMR titrations. Different structural features of simple anion hosts (i.e .,
number, type, and disposition of hydrogen-bond donor groups, skeleton rigidity,
and preorganization for binding) can affect their anion-binding properties and thus
affect their behavior as co-extractants. Therefore, ion-pair extraction investigations
in conjunction with homogeneous phase binding studies can give complementary
information. Such information may be used to ident ify selective co-extractants that
combine a large number of favorable structural characteristics for further
analytical, industrial, or environmental applications.



2.1. Cesium Hosts

The univalent cation Cs+ has been selected as the target cation in this work in
connection with our focus on principles of size selectivity in the binding and extraction of
spherically symmetric alkali metalcations by crowncompounds. In this regard, achieving
complementarity with exclusivity for Cs+ has proven particularly challenging, because of
its large size and the well-known flexibility of crownethers. The design of selective host
molecules for practical applications has been reviewed extensively in the past,' and a
detailed discussion is beyond the scope of this article. General requirements for practical
systems include strong, selective, and reversible cesium binding with rapid complexation-
decomplexation kinetics; simple extraction stoichiometry free of complicated aggregation
phenomena; acceptable solubility; high stability to severe conditions; and synthetic
accessibility. However, for the purposes of investigating and designing effective dual-
host systems, additional considerations arise. The effect of the added anion receptor in
the system should be maximized and at the same time also remain clearly defined, thus
resulting in simple behavior that will allow theoretical interpretation. Therefore, along
with the previous considerations, it is desirable that i) The interaction between the cation
host and the anion host is minimal. Such interactions are expected owing to the mutual
affinity of Lewis acids (anion host) and bases (cation host). If sufficiently strong, this
competing effectwould most likely result in weaker binding and even possible loss of the
extractants due to precipitation. ii) The abilityofthe cation and anionreceptors to extract
salts when used singly is weak. Naturally, strong extraction by the receptors used alone
would tend to mask the desired synergistic effect and possibly obviate the need to
consider a dual-host system. iii) The ion-pairing between the complexed cation and the
counteranion in the organic phase is relatively weak. Ion pairing provides an additional
driving force for extraction, but if this effect is important, it then becomes a design
consideration for the dual hosts. Although this is not necessarily undesirable, it reduces
the flexibility and ease of using the dual-host approach. Moreover, extensive ion pairing
can prevent effective anion binding and may also present difficulties in the analysis of the
synergistic effect. In this context, solvation is an additional consideration.P:" as the
choiceof solventallows a degree of control of ion pairing, of interaction between the two
hosts, and of solubility.
The crown ether tetrabenzo-24-crown-8 (1) presents us with the aforementioned
advantages in our efforts to gain an understanding at the initialstagesof our development
of the dual-host approach. Past distribution experiments using tert-octyl derivatives and
subsequent equilibrium modelling with the program SXLSQI had shown that the ion-
pairing between the complexed cationand the anion is weak. IS8Crystal structures 1Sa,19 of
tetrabenzo-24-crown-8 with CsN03 have shown no direct interaction between the cation
and the anion. The crownether does not complete the Cs+ coordination sphere but leaves
two V-shaped clefts available for additional ligation .i" l,2-Dichloroethane,ls8
dichlorornethane," or acetonitrile" solvent molecules tend to occupy those additional
coordination sites within the Cs-crown cavity, interacting directly with the complexed
C s + ion. As it can be seen from the crystal structure of [Cs(tetrabenzo-24-

crown-S)(NCMeh(TJ2-CH2Ch)][NOJ]19 (Figure 2) , the methylene chloride solvent

molecule is bound to Cs+ via an unusual TJ2-coordination pattern. The NO J' counteranion
remains outside the coordination sphere despite the potentially favorable Cs+-0
electrostatic interactions that would result from nitrate coordination. This behavior has
been observed consistently in various structures and has been ascribed to the unique
shape of the Cs+-macrocycl e cavity that creates a sterically and electronically
" incompatible" environment for nitrate.


Figure 2. PLUTO (left) and space filling (right) representations in two different orientations of the X-ray
crystal structure of [Cs(tetrabenzo-24-crown-8)(NCMel2(l]z.CHzCh»)(NO ,) (3 j. " The CHzCh l]z-coord ination
modeand the exclusion of nitratefrom the Cs coordinationsphereare apparent.

Despite the reduced selectivity for Cs+ compared to the more recently studied
calixarene bis-crown extractant BOBCalixC6 (2), we chose tetrabenzo-24-crown-S (1) in
1,2-dichloroethane for our initial studies on dual-host systems. Alkylated derivatives of 1
were shown to bind and extract Cs+ strongly and to ion- pair weakly, owing to the
exclusion of the anion from the inner coordination sphere of the Cs+ cation.l" It was also
anticipated that the reduced electron-pair basicity of the ether oxygen atoms and steric
hindrance, both due to the benzo substituents, would disfavor interaction with the
hydrogen-bond donating groups of added anion receptors. Hence, the synergistic effect of
an added anion receptor could be max imized, and the analys is of the extraction equil ibria
greatly simplified. The actual practical applicability of tetrabenzo-24-crown-S for cation
extraction to lipophilic diluents is limited due to its low solubility. However, more
lipophilic derivatives have been prepared and shown to be effective extractants for Cs+.18

2.2. Nitrate Hosts

Despite the abundance of nitrate in the environment and its presence as a major
component of nuclear waste streams,2f,J few designed nitrate rece~tors have been
reported." and examples of neutral lipophilic nitrate hosts are rare. la.b The generic
requirements for an efficient and practical anion host for extraction of ion pa irs are
in some ways not unlike the ones for cation hosts : Nitrate binding should be strong ,
selective, and reversible, with rapid complexation-decomplexation kinetics. The host
should be soluble in organic diluents of low polarity , preferably with minimal tendency to

aggregate, and it should be stable under the extraction conditions. Ideally such a host
should also be accessible in only a few synthetic steps from commercially available
starting materials . Ligand design for both cations and anions must take into account
distance and directionality of the donor group (either electron-pair donor or hydrogen-
bond donor). Alkali metal cations can be considered as positively charged spheres that in
themselves possess no electronic preference for specific coordination geometries. The
forces involved in binding these cations primarily involve electrostatic ion-dipole
interactions, which are particularly dependent on the distance between the cation and the
donor atom on the host backbone. The interaction strength also depends on the
orientation of the dipole-moment vector of the donor functionality with respect to the
cation." which imposes a directionality component. For most anion-host designs in
organic media, the primary binding force is hydrogen bonding, which is a directional
electrostatic interaction dependent on distances and angles that involve a three-atom
X-H... yo- configuration. This arrangement introduces additional constraints, especially
with relation to the X-Hoo . yo- angle, which for most hydrogen bonds lies in the range of
160-170° (from crystallographic data).22a Strong "hornonuclear" X-Hoo .X- hydrogen
bonds are almost linear, and it has been computed that the strength of an Q-Hoo.O
hydro~en bond would fall rapidly to 60% for a 149 angle and to only 10% for a 1100

angle. 2b Unlike the case for metal cations as guests, anionic guests of interest include
polyatomic species, which impose new challenges in achieving complementarity, both in
terms of guest geometry and size. With regard to nitrate, consideration should be given to
the fact that nitrate is planar, with its negative charge delocalized and concentrated
equally at the three oxygen atoms. Hydrogen-bond donor sites on the anion host should
then in principle be arranged to bind all three oxygen atoms of the nitrate anion.
Therefore, we were naturally led in our design efforts to consider practical tripodal
frameworks. Some tripodal frameworks that are commercially available or easily
obtainable in a few synthetic steps are shown in Figure 3.

Figure 3. Three tripodal carboxamide frameworks thoughtto have favorable orientation for nitrate binding: a)
1,3,S-benzenetricarboxamides (trimesic acid derivatives) (4), b) tris(2-aminoethyl)amine(tren) derivatives (5)
and c) 1,3,S-tris(aminomethyl)benzene derivatives (6). The trimesic acid framework allows straightforward
functionalization on the amino part, while the tren and tris(aminomethyl) benzene derivatives allow
functionalization of the carboxy part, instead. The R ancillary groups mayrepresent alkylor aryl functions.

Recent density functional theory calculations together with the crystallographic

record have provided a basis for design of receptors for oxyanions such as nitrate.22c
According to these results, the ideal arrangement for hydrogen-bond donor groups X-H

in relation to nitrate entails an in-plane approach of the X-H vectors approximately in

line with the Sp2 lone-pair orbitals of the nitrate oxygen atoms. This approach
corresponds to H···O-N-O dihedral angles of 0 and 180° and H.. ·O-N angles of 120°.
Such crystallographic and theoretical studies offer an invaluable tool for anion-
receptor design purposes. In the present case, it may be seen that the tripodal receptor
geometry that we initially proposed has merit, but now with special constraints for how
the three X-H vectors must be aligned. Subjectto these constraints, the most symmetrical
binding geometry has C3 symmetry (C3h if the three X-H vectors lie in the nitrate plane),
since the X-H vectors will not point toward the central N atom but toward the Sp2 lone
pairs on the 0 atoms. This effectively reduces the cavity size needed for nitrate, though it
may be noted that the tripodal designs shown in Figure 3 still lack sufficient cavity size
and could therefore at best perch on top of the nitrate in the most symmetrical (C3)
binding geometry.



To evaluate the difference between the dual-host strategy and the "classical"
approach of salt extraction based solely on cation complexation, it is instructive to
examine the extraction process in terms of its component thermochemical steps." The
overall Gibbs energy of extraction fJ.G ex° may be considered as the sum of Gibbs energy
changes associated with a) partitioning of the individual metal ion M+ from water to the
solvent phase [fJ.G pO(M+»), b) partitioning of the individual anion X- from water to the
solvent phase [fJ.GpO(X-»), c) complexation of the metal ion by the cation receptor CE in
the solvent phase [fJ.GcpxO(CEM+»), d) complexation of the anion by the anion receptor R
in the solvent phase [fJ.GcpxO(RX-»), and e) ion pairing in the solvent phase
[fJ.Gip·(CEM+RX-)) . Accordingly, the standard Gibbs energy change fJ.G exo associated
with the net extraction may be expressed as the sum of these five component steps:

It should be noted that the ability to express the partitioning of individual ions requires
the use of an extrathermodynamic assumption. Although the choice of assumption is
essentially arbitrary, it is preferable to employ a convention that is chemically reasonable.
The most popular convention, and the one used here, posits that
fJ.GpO(tetraphenylarsonium) = fJ.GpO(tetraphenylborate) and is referred to as the TATB
assumptlon.l'" The simplified thermochemical diagram shown in Figure 4 illustrates the
component steps that describe the dual-host concept for the case of the well-studied
cation host calix[4)arene-bis(tert-octylbenzo-crown-6) (2 or CE)24 and a general anion
receptorR (Figure4). From the diagram, it is clear that the unfavorable energetics for ion
partitioning from water to a less polar organic solvent are overcome by the strong
complexation of the cation by the crown ether host, plus a potential additional energetic
gain via anion complexation by receptor R. Ion pairing adds another gain, which in the
case shown is -16.4 kl lmol for the dialkylated analog of 1 used with no anion receptor. IS
This effect would likely be weakened by adding an anion receptor, if the ion pairing

X"= N0 3"
CE =Iis(tel1-Octybenzo)dibenzo-24-crown-a

R =Anion Receptor Ion pairing

CE + M+ + X- + R oc CEM+ + RX" Complexation

Figure 4. Simplified thermochemical scheme for the extraction of cesium nitrate by bis(tert-
octylbenzo)dibenzo-24-crown-6 (CE) and an anion receptor R in 1.2-dichloroethane at 25 ·C. The numbers
given for the corresponding equilibria are standard Gibbs energies in kJ/molas reported in Ref. 18a for the
extraction of cesium nitrate by thedialkylated derivative oftetrabenzo-24-crown-8 (CE) only.

effect is primarily coulombic in nature. Hence, it may be appreciated that the largest
enhancement upon adding R to CE is obtained in systems in which ion pairing is absent.
The natural question arising from a theoretical examination of a dual-host system
focuses on comparing directly the receptor-anion binding constants in the organic phases
with extraction enhancements from distribution studies . For example, is it possible to
predict the synergistic enhancement in the extraction of a given salt by CE upon addition
of R if one already knows the anion-binding constant? Anion-receptor binding constants
Kq,x(RX) are often obtained by IH-NMR titrations and in principle should directly
correspond to the binding constant operative in the extraction system, provided that
conditions in the two experiments are the same. Such comparisons require specific
assumptions, but they can still provide useful insight to the extraction process. If the
conditions are chosen so that ion pairing is minimal, then the extraction equilibria for the
crown-only (unsynergized) and the dual-host (synergized) systems can be written as Eqs.
(2) and (3), respectively:

.. (2)

.. (3)

Here, B is further shorthand for the cation host (same as CE earlier). Both B and R are
taken to be highly lipophilic, so much so that they and their complexes can effectively be
considered exclusively organic-phase species (negligible distribution to the aqueous
phase). This treatment also assumes no significant aggregation, crown-receptor

association, or higher comple xation, which would have to be dealt with explicitly if
needed. The equilibrium constants for the crown-only and synergized systems are defined

(4) K syn = (5)

[Cs](aq)[NO]'] (aq) [B]

For convenienc.e, these equilibrium constants are written as concentration constants in

which the activity coefficients are incorporated into the constants themselves; for
applications where activity effects are not constant, the derivation must be repeated with
explicit treatment of activity coefficients.
The experimentally accessible quantity in extraction is the cesium distribution ratio
.Q:a defined as the organic-phase cesium concentration [Cs]org divided by the aqueous-
r:hase cesium concentration [Cs]aq. With use of sensitive analytical methods, such as
37CS radiotracer counting, the "window" of measurement of Dcs may be up to ten orders
of magnitude, making extraction an attractive technique to assess anion binding . A simple
analysis of synergism follows if the only form of cesium in the organic phase is the Cs+·B
complex. This is true if cesium extraction is negligible in the absence of the cation
receptor . Then [Cs]org = [Cs+'B]org and Dcs = [Cs+oB]org![CsJaq. We wish to distinguish
between the cesium distr ibution ratio measured with B alone and with B plus R and will
refer to these measured values as DCsB and Dc s syn, respectively . From Eq. (4), D CsB may
be defined according to Eq. (6) , where the subscript notat ion "R=O" indicates that the
anion receptor is absent:


Noting that the total organic-phase cesium concentration must equal the total organic-
phase nitrate concentration, Eq. (5) then leads to:

[(Cs+oB)1(ag) [NO] ' ] (Org)IOtal [NO] 1018) + [(RoN 0] )1<0 18)

Dc• •yn = (7)
[Cs+](aq) [Cs+laq) [Cs+laq)

The synergistic factor (SF) is normally defined as the ratio of the distribution ratio
for the synergistic system to the sum of the distribution ratios of the individual extractants
used alone. Hence, SF corresponds in the present case to the ratio of the distribution ratio
of the dual-host system (D c s syn) to the sum of the distribution ratios for crown only
(D cs' B) and for amide only (Dc s.J0. If the amide anion receptor by itself is effectively
unable to extract cesium nitrate, then D Cs•B » Dc s' R' and we then have:

DCs syn Dc. syn

Synergistic factor(SF) ::: - - - (8)

Assuminga simple 1:1 nitratereceptorequilibrium in the organicphase,we have:


or (10)

From Eq. (7) and taking accountofEqs. (10) and (4), one obtains:

[N031(org) [N03.]~q) [B]

DCssyn = - - - - (1 + Kcpx [R](org.~ = Kex ( 1 + Kcpx [R](ag» (11)
[CSi(8lV [CsB](ag)

Recallingthe conditionset above that [Cs](org) = [Cs+B](org), the definitionof Dcs syn in Eq.
(7) and the mass-balance relation [CS];nit = <1>[Cs~](org) + [Cs](aq) give rise to the useful
relation, where <1> is the phaseratio:
[Cslnit. Dc s syn (12)

We then have from Eq. (11):

Dc s syn

Using the relation Eq. (12) again for the condition that R is absent and substituting Eq.
(3), one arrives at the following relation that allows one to predict the synergistic
enhancement in terms of the anion-receptor bindingconstant:

Under experimental conditions for which we have low extraction (<1> D cs « 1),
negligibleloadingof the crown,and equal phaseratios, Eq, (14) transforms into:

Dc s Syn2
= (1 +Kcpx [R](inil~ (15)

This theoretical treatment is presented here as an ideal model with which the
behavior of real dual-host synergistic extraction processes may be compared.
Accordingly, the limitations of this treatment should always be kept in mind, where

deviations can be taken as evidence for an effect that has not been taken into account. For
example, ion pairing in particular can be a very important factor, especially when the
countercation for the two systems in comparison is dramatically different. Usually a
tetraalkylammonium salt is used for the NMR titration experiments, while a significantly
bulkier Cs-crown complex takes its place in the distribution study. It may also be pointed
out that the organic solvent in a distribution study is saturated with water, whereas this is
typically not the case for the NMR experiments, in which the water content of the
solution is low but usually unknown. The role of water in anion-binding systems remains
to be understood, but it may be assumed that the free anion and anion receptor, as well as
the their complex, are hydrated to a degree in the organic phase." Despitesuch questions,
the outlined simple model allows guarded comparison of binding constants obtained from
NMR titrations with extraction data for the purpose of gaining deeper understanding.
Toward designing extractive separation systems for salts, the model also provides a
useful tool for forecasting extraction synergism from abundant binding constants for a
given anion receptor and whether the receptor could potentially be useful in augmenting
salt extraction processes based on a given cation receptor. The simple treatment
highlights two critical factors in the design of a dual-host system: strong anion binding
and high obtainable concentrations for R (which for most systems is limited by the anion-
host solubility). From a practical point of view, an inexpensive, soluble, and strong anion
host can then be used to reduce the amount of the selective cation host needed, which
may often be the single mostcostlycomponent of a salt-separation process.



The 1,3,5-benzenetricarboxamide framework was our initial choice for a designed

nitrate host, because it combines a C3 symmetry environment (suggesting a potential
formation of three hydrogen bonds to the three oxygens of the nitrate) with synthetic
flexibility and availability from inexpensive starting materials. Amides have been widely
employed recently in anion host-guest chemistry/7.29 and efficient tripodal anion
receptors have been synthesized." Isophthalamides have been proven to be strong anion
complexants" possessing a flexible framework that can adapt its geometry involving two
convergent hydrogen bonds to accommodate a variety of anions ranging from the small
and hard fluoride to the much larger and softer iodide. 29a,b Toward an effective tripodal
system, initial modeling suggested that three hydrogen bonds to the oxygen atoms of the
nitrate can potentially be formed, with distances and angles compatible with previous
crystal structures for other anions. Specifically, the dihedral angles between the
O=C-N-H plane and the plane of the central aromatic ring can be as large as 35 degrees
to accommodate the binding of the bulky Br" anion so that two linear hydrogen bonds can
be fonned.29a.b Nevertheless, the binding is relatively strong (7100 ~I for Br-)?9a,b
Finally, compared with other frameworks, tricarboxamides offer excellent synthetic
flexibility in the choice of the amine moiety, thus allowing control of the electronic and
steric contributions to the binding, as well as the amide solubility and hydrophobicity.

Such modifications allow "fine tuning" of the co-extractant behavior, for practical
The tricarboxamide 7 and the corresponding controls 8 and 9 (Figure 5) were
synthesized by standard synthetic methods from the corresponding 2,5-di-tert-butyl
anilines (for increased solubility) and acid chlorides 1,3,5-benzenetricarbonyl chloride (7
and 8) or benzoyl chloride (9). The amide receptors 7-9 were tested for their ability to

IB~~::O IB~~
/I ,
1 .--::
N ..... I·Bu
,\ #
, ,
0 'H H""'~
..... I-Bu

I'BU -8U N-H I·B I.B~trCH3 I-B~
~ I-Bu ~.I·Bu
7 8

<. HH~~ I.Bu


Figure S.Tricarboxamide (7) and the corresponding controls (8,9) usedas anion hosts.

synergize CsN03 extraction using tetrabenzo-24-crown-8 (1) as a cation host" at 5 mM

in 1,2-dichloroethane. Aqueous phases contained 0.1 mM RN0 3 and 10 mM CsN03,
wherein the slight acidity eliminates possible effects of trace acidic impurities. Control
organic phases containing only amides and no crown ether were also used in order to
determine if 7-9 are able to extract Cs+ into 1,2-dichloroethane when used alone. After
equilibration of the two phases and phase separation, the distribution of Cs+ between the
two phases was determined by gamma (y) counting using a small amount of added
I31CsN03 tracer. The results (Figure 6) show an extraction enhancement when amides 7
or 8 are present in a synergistic combination with the crown ether, as compared with the
extraction experiments using only the crown ether under the same conditions. For 10 mM
of the amide 7, the extraction is enhanced by a factor of 1.9 (Dcs = 6.3 x 10-3 for 7 versus
3.3 x 10- 3 for the crown-only experiment), which is significantly higher than the
enhancement observed for the same concentration of control monomethylated triamide 8
(Dcs= 4.2 x 10-3). Monoamide 9 failed to give any noticeable extraction enhancement,
suggesting negligible nitrate binding. As expected from Eq. (15), the enhancement for 7
and 8 is dependent on the amide concentration (Figure 6), and the attainment of much
higher enhancements, notwithstanding solubility limitations, is thereby indicated. Control
experiments with amide only (no crown present) did not give any observable extraction

- --T- -- -·...,-- - - - - , r - -- -,-- ----,

• 7
• 8
• 9


4.010- 3

3.0 10-3 ' - -_ _- ' -_ _---' - ' -_ _---' ...J

0.00 1.00 2.00 3.00 4.00 5.00

[R] I/[CE] t
Figure 6. Cs extraction enhancements for different ratiosofR = 7-9 to CE (tetrabenzo-24-crown-8) (I) .

of CsN03 (Dcs < 2 X 10-5) , indicating that the anion receptors are unable to extract
CsN03 in the absence of a cation extractant. The ordering of the anion receptors
according to ability to synergize CsN0 3 extraction by 1 suggests stronger nitrate
complexation by the triamide, thereby thermodynamically favoring the overall salt
transfer to the organic phase. The fact that 8 shows a smaller enhancement is expected
from its potential ability to form only two hydrogen bonds in S-N03- versus three in
lH-NMR titration experiments" with (BU4N)N03 in 1,2-dichloroethane-d4 provided
further insight into nitrate complexation by the host, from both qualitative and
quantitative aspects. For the receptor-nitrate complex (Figure 7), significant downfield
shift changes for the N-H protons were observed due to hydrogen bonding of the amide
N-H protons to the oxygen atom of the nitrate anion. Some aromatic C-H protons also
showed downfield chemical shift changes, related to their proximity to the oxygen atom
of the nitrate in the complex. Specifically, for the diamide 8, both the N-H and the 2-
C-H protons between the N-H groups (Figure 7) gave significant downfield shift
changes, with the 4-C-H chemical shift being virtually constant. This was taken to be
indicative of nitrate complexation by two hydrogen bonds to one anion in a fashion
similar to that demonstrated before for isophthalamide receptors and other anions.29a.b The
C-H protons of the tricarboxamide 7 instead exhibited a single resonance with a
downfield movement that is overall significantly smaller than for 8. This is consistent
either with a 1:1 7'N03- complexation pattern via three hydrogen bonds or alternatively



~ 0.80

CoQ 0.60


0.0 10° 1.0 10.2 2.0 10.2 3.0 10.2 4.0 10.2 5.0 10.2 6.0 10.2
[N03']t (M)
Figure 7. 'H-NMRtitration curves showing N-H and 2-C-H chemical shiftchanges for amides7 and 8 in 1,2-
dichloroethane-da uponaddition of tetrabutylammonium nitrate.

with a fast exchange among the possible bidentate configurations of the 1:1 complex,
each configuration binding in the same fashion as 8. No irregular chemical shift change
patterns were observed even after addition of 60 equivalents of N0 3-, and the binding
curves were perfectly fitted using the 1:1 binding isotherm.l" indicating that there is no
higher complexation.
The obtained binding constants of 250 ~t for 7.N03- versus 23 ~t for 8.N03-
(which correspond to t:..GO of complexation of -13.4 and -7.5 kJ/mol, respectively) can
be related to the stronger complexation of nitrate by the tripode due to the presence of an
additional hydrogen bonding site. The IH-NMR observations were confirmed by 1~_
NMR experiments in which ISN0 3- was titrated with increasing amounts of
tricarboxamide 7: Significant upfield shifts were observed for the ISN0 3- resonance upon
addition of tricarboxamide, while the benzanilide control gave smaller upfield shift
changes.'! Electrospray ionization mass spectrometry in the anion detection mode was
performed in the organic phases. The results verify the 1H-NMR evidence of a 1:1
7·N03- complexation stoichiometry, even when the concentration of 7 is many times
higher than that of N03- in the organic phase. For the same conditions, using the diamide
8 gave also a peak corresponding to 8·N03- with a much smaller intensity (as compared
to 7·N03l , and using the monoamide 9 did not give any peak corresponding to 9.N03- ,
in full agreement with the NMR and extraction results. Vapor phase osmometry as well
as NMR dilution experiments together indicate that these amides exist in solution almost
entirely as monomers.i! However, by using the simple theoretical model derived above

[Eq. (15)] assuming no aggregation, association with the crown, or dimerization of7 or 8,
these binding constants would predict respective extraction enhancements of 1.9 and 1.1
for 7 and 8. These values are in good agreement with the actual values of 1.9 and 1.3.
Based on our I H-NMR observations (vide infra), the two possible binding modes are
i) a symmetrical or unsymmetrical tridentate hydrogen-bond structure in which nitrate is
bound to the three amide N-H hydrogens; ii) a dynamic equilibrium between the three
possible bidentate structures which form only two hydrogen bonds to one nitrate as for
8·N03- ; and iii) any combination of i) and ii). The tenfold higher Kcpx for 7·N03- versus
8·N03- suggests cooperativity of the third N-H group that cannot be attributed to the
statistical factor only, which would be expected to increase the binding only by a factor
of 3, indicating a contribution of the tridentate mode to the overal1 structure of 7·N03- .
The free carboxamide has its C=O double bonds almost co-planar with the central
aromatic ring," while the nitrate-bound receptor is forced to adopt a nonplanar
conformation. Therefore, the energy gain by the formation of three hydrogen bonds
versus two could be partially negated by the loss of conjugation. X-ray structures of
anion-isophthalamide complexes showz9a,b that isophthalamides could adopt a nonplanar
binding conformation, apparently gaining the energy to do so from the subsequent
formation of the two hydrogen bonds to the same anion. Supporting a significant
contribution of the tridentate hydrogen-bond structure is the fact that the dBmax of the 2-
C-H chemical shift change is significantly larger for S·NOJ - than for 7·NO J - . This
suggests greater proximity of the 2-C-H protons to the anionic site in S·NOJ - or
alternatively a higher delocalization of the negative charge in 7·NOJ - . The upfield shifts
observed in the IsN_NMR titration experiments with tricarboxamide 7 or control
benzanilide are consistent with any of the hypotheses presented. A variable-temperature
NMR experiment showed resolution of the C-H and N-H peaks below -70°C, which is
consistent with either a nonsymmetrical I :I species or with the presence of multiple
species. The N-H and C-H resonances, which are singlets at room temperature, were
clearly resolved into six peaks at -90 °C. The integral ratios suggest the presence of both
a CJ symmetric complex and a Czv symmetric complex with only two hydrogen bonds to
the nitrate or alternatively a single unsymmetrical 1:1 complex."
To summarize, our investigation of the 1,3,5-benzenetricarboxamide derivative of
2,5-di-(tert-butyl)aniline (7) as a co-extractant with tetrabenzo-24-crown-8 (1) for CsNOJ
extraction into 1,2-dichloroethane gave a significant extraction enhancement, thus
demonstrating the first example of dual synergism as a result of ion-pair recognition in a
liquid-liquid cation extraction system, which can be ascribed to synergism via anion
complexation. NMR titrations, as well as electrospray ionization mass spectrometry
confirm the potency of this carboxamide derivative in enhancing CsNOJ extraction via
nitrate binding.



Amide derivatives of tris(2-aminoethyl)amine (tren), like the trimesic acid

derivatives, also provide a potential C3 symmetric environment for nitrate complexation"

through three hydrogen bonds. Tren was thus selected as the starting amine for the
synthesis of an anion host family for dual-host extraction studies. Moreover, being
commercially available, it provides a simple and readily available framework,
which-unlike the trimesic acid derivatives-allows variation of the carboxy "tail" and
therefore tuning of the receptor binding properties for the targeted anion in a very
different fashion. Figure 8 shows the triamides and monoamide controls used in this
study, illustrating such versatility. Tren has attracted widespread attention in the past as
an anion host: A positively charged bicyclic tren-derived polyammonium receptor was
found to bind two nitrate anions in the solid state.21g Reinhoudt and coworkers have
reported amide and sulfonamide tren-derived anion receptors, although nitrate binding
properties were not studied." Beer and coworkers recently combined tren-derived anion
binding sites and l5-crown-5 cation binding sites in the samemolecule and demonstrated
synergism for sodium pertechnetate extraction."

~c~ 14. R = CSH 11

H R 15. R = C 7H 1S
10. R= CSH 11
11. R = C7H 1S 14·15
12. R =CeH1e
13. R = C11H23

10 - 13

18 19

Figure 8. Tren-derived nitrate receptors (16-13, 18,and 19)andcontrols (14-17).

We decided to carry out a systematic studyoftrialkylamidederivatives with variable

alkyl chain length (Figure 8, 10-13). Monamide controls 14-17 were also synthesized to
compare the effects due to the presence of additional hydrogen bonds between the host
and the anion. Moreover, a triaryl-amide derived from (tert-butyl) benzoic acid (18) was
studied, which could allow better correlation with the tricarboxamide systems. Finally,
the NMR nitrate binding properties of the fluorescent tris(dansyl)amide (19)33.35 were

studied to determine if such a sulfonamide would be potentially more effective in an ion-

pair extraction system.
The tren derivatives 10-13 and 18 and the monoamides 14-17 (Figure 8) were tested
for CsN0 3 extract ion, again using tetrabenzo-24-crown-8 (1) as the cation host. The
conditions set for these experiments were identical to the tricarboxamide distribution
experiments (vide inf ra). Under these conditions , anion binding is independent of the
cation concentration, and therefore the analysis of extraction behavior is greatly
slmplifled." The results (Table I) show significant extraction enhancements when
amides 10, 11, and 18 are present together with the crown ether, in comparison with the
extraction experiments using only the crown ether under the same conditions. For 50 mM
solutions of the triamides 10 and 11, the extraction is enhanced by a factor of2.4 and 1.7,
respectively. These enhancements are significantly higher than the marginal enhancement
observed for the same concentration of the corresponding monoamides 14-17, ranging
between 1.2 and 1.3. As in the tricarboxamide case, the enhancement increases with the
amide concentration . The control experiments with amide only (no crown ether) did not
give any observable extraction of CsN03 (Des < 2 x 10-5) . The fact that the monoamides
14-17 give only a marginal enhancement undoubtedly follows from the presence of only
one hydrogen-bond-donor amide group in the monoamides versus three in the tren
derivatives. Compounds 12 and 13 failed to give any significant extraction enhancement
for their maximum obtainable concentration in 1,2-dichloroethane. This concentration
was limited by considerably lower solubilities (-10 mM for 12 and -2.5 mM for 13). For
the smaller-chain amide 10, the extraction enhancement is larger than for the longer-
chain amide 11, in the high concentration range of amide (>20 mM).

Table 1. Some representative distribution ratios Des, log Des values, and synergistic
factors for various concentrations of the different tren derivatives and controls. All values
are for 5 mM tetrabenzo-24-crown-8, 10 mM CsN0 3, and 0.1 mM HN0 3 • Uncertainties
in Des values are estimated to be <5%.
Amide [RI,(M) Synergistic Factor
None 0.000 2.6 X 10-3 -2.60 1.0
10 0.020 3.9 x 10-3 -2.41 1.5
10 0.050 6.2 x 10-3 -2.2\ 2.4
11 0.020 3.\ X 10-3 -2.5\ 1.2
11 0.050 4.4 X 10-3 -2.36 1.7
12 0.010 2.\ X 10-3 -2.67 <1.0
13 0.0025 2.3 X 10-3 -2.64 <1.0
14 0.050 3.2 X 10-3 -2.50 1.2
15 0.050 3.3 X 10-3 -2.48 1.3
16 0.062 3.3 X 10-3 -2.48 1.3
17 0.050 3.1 X 10-3 -2.5\ 1.2
18 0.020 5.7 X 10-3 -2.25 2.2
18 0.050 1.2 X 10-2 -1.94 4.6




~ -2.30

0 -2.40



-3.50 -3.00 -2.50 -2.00 -1.50 -1.00

Figure 9. Plotoflog De; vs log[Rl~ facilitating comparisons between the trimesic acidandthe trenderivatives.

Figure 9 illustrates the extraction results in a log Des versus 10g[R]eq. plot format, It
also facilitates a direct comparison between the previously described tricarboxamide and
the tren derivatives. Generally the synergistic effect for the trialkylamides 10-11 is weak,
while for the monoalkylamide controls 14-17 it is practically negligible. As it can be
seen, the only tren derivative that gives comparable extraction enhancement with the
tricarboxamide 7 is the aromatic derivative 18. These results are not surprising, since the
acidity of both those amides, which can be correlated directly to hydrogen-bond donor
strength, is comparable and apparently significantly higher than for the trialkyl

Table 2. Results of H-NMR titrations of tripodal amide receptors with
tetrabutylammonium nitrate in 1,2-dichloroethane-d4 at 25°C, 400 MHz.
Receptor Kcpx (RoNOll logKcpx AGO
(M') (kJ/mol)
10 52 (± 3) I. 72 (%0.02) -9.6
11 36 (± 2) 1.56(± 0.02) -8.8
12 41 (± I) 1.61 (± 0.01) -9.0
18 33 (± 2) 1.52(± 0.02) -8.5
19 60 (± I) 1.78(± 0.02) -10 .2

Nitrate complexation by compounds 10-12 , 18, and 19 was confirmed by NMR

titrations with tetrabutylammonium nitrate in dichloroethane-d4.36.37 Large downfield
chemical shift changes were observed as expected for a hydrogen-bonded amide proton.
The equilibrium constants at 25 °C obtained from the titration curves and the
corresponding I:1GO values of nitrate complexation by 10-12, 18, and 19 are given in
Table 2. Apparently, the dansylamide 19 is the strongest receptor. This is not unexpected,
taking into account the higher acidity of sulfonamides compared to amides. On the other
hand, it is unexpected that triamide 18 which gave a significant extraction enhancement
gave a lower binding constant than the alkylamide derivat ives by NMR. By using the
simple theoretical treatment outlined above and applying the observed binding constants
for the receptors 10, 11, and 18 (52, 36, and 33 ~l respectively) in Eq. (15) for 50 mM
receptor concentration, we get expected extraction enhancements of 1.9, 1.7, and 1.6,
respectively. Compared with the experimental values of 2.4, 1.7, and 4.4, reasonable
agreement is found with receptors 10 and 11, though the enhancement for the strongest
nitrate receptor 18 is clearly underpredicted. Different effects of the receptors on ion-
pairing under the extraction and NMR experimental conditions could be responsible for
such discrepancies. Another factor could be the formation of complexes with higher
receptor/nitrate stoichiometric ratios for the conditions of the extraction experiment.
While in the NMR experiment the nitrate concentration is in the same range or higher
than the receptor concentration, in the distribution experiment (due to the low distribution
ratio for nitrate by tetrabenzo-24-crown-8) the receptor/nitrate concentration ratio can be
as high as 900. Under these conditions, the formation of 2:1 complexes of receptor to
nitrate cannot be excluded and would introduce complications to a simple I: I equilibrium
treatment. Detailed equilibrium studies for various concentrations and for alternative
solvents that allow minimal ion pairing (e.g., nitrobenzene) could potentially elucidate
these issues in the future.
As in the case of the carboxam ide derivatives, negative-ion mode electrospray
ionization mass spectrometry fully supports the distribution and NMR results, indicating
the presence of I : 1 anion-receptor RoN0 3- complexes. Addition of excess
tetrabutylammonium nitrate to 10 JlM solutions of R in 1,2-dichloroethane (R = 10-13
and 18) gave peaks corresponding to RoN0 3- for receptors 10-13 and 18, suggesting a
primarily 1:1 complexation stoichiometry (e.g., for 2 mtz = 586.4). Peaks corresponding
to 2:1 R 2oN03- complexes were barely observable. On the contrary, none of the controls
14-17 gave any observable peaks for RoN0 3- complexes, consistent with very weak
The X-ray structure of monoprotonated 10 exhibits lipid bilayer-type networks ,"
suggesting a possible tendency of these amides toward self-association. Such association,
if present, could potentially play an important role, and therefore the discrepancies
between single-phase data and distribution data could potentially be explained . By use of
vapor phase osmometry. i" an investigation of the triamides 10-13 and 18 in 1,2-
dichloroethane at 25°C gave aggregation factors of 1.28 for 10 at 23.9 mM, 1.14 for 11
at 20.2 mM, 1.02 for 12 at 4.2 mM, 0.94 for 13 at 2 mM, and 1.22 for 18 at 20.8 mM.
These results indicate that self-association is rather weak and does not appear to be a
major cause for the differences observed in the extraction enhancements among the
various receptors. Likewise , vapor phase osmometry was also used to study possible
crown-amide association. It was found that the particle concentration of a solution

containing 5 mM crown ether plus 5 mM of amide 11 was 10 mM (±IO%), suggesting no

significant crown-amide association.
To summarize, our investigation of different tren derivatives as co-extractants for the
extraction of CsN0 3 in 1,2-dichloroethane by tetrabenzo-24-crown-8 gave significant
extraction enhancements, particularly for the tris-(4-t-butyl)benzoyl tren-derived triamide
18. The observed synergistic effects are in reasonable agreement with the ones predicted
theoretically from the IH-NMR Kcpx values for receptor-nitrate binding. It is currently
uncertain why the extraction enhancement is smaller with increasing chain length for the
trialkylamides. In this direction, the effect of ion pairing is an issue worth investigating.
Distribution experiments with branched or shorter chain analogs could potentially
elucidate this question.


1,3,5-Tris(aminomethyl)benzene is a versatile tripodal framework that has

been widely used in supramolecular and coordination chemistry, in particular by
the Anslyrri" and Raymond" groups . The neutral bicyclic analog of 2,4,6-triethyl-
1,3,5-tris(aminomethyl)benzene recently reported by Anslyn et al. binds nitrate
with an binding constant of 300 M- 1 (CDCI 3IDMSO) .21a Although the triamine
building block for synthesizing tripodal amides is not commercially available, the
synthetic routes for making those analogs are straightforward, and all reactions
proceed with reasonably good yields.t" As part of our efforts to design a
potentially improved dual-host system , we carried out a distribution study using the
triamide 20, which was synthesized'f from the corresponding triammonium
trihydrochloride salt. 43


The distribution experiments, carried out in exactly the same fashion as for the tren 29
and tricarboxamide'" derivatives gave significantly increased enhancement compared
with the trimesic acid analog 7 and the tren derivative 18 for the same receptor
concentration (Figure 10). These results thus show that 20 is a stronger nitrate receptor .
However, solubility limitations did not allow further study of this receptor in a sufficient
concentration range that would make its use practical in dual-host applications. In the
future, more lipophilic triethyl (rather than trimethyl) derivatives as well as more soluble
sulfonamide tripodal analogs will be synthesized and tested.

7.0 to·
,-------"T-----.....- - - - - - ,

I 7

2.0 10.3 ' - - - - - - - - ' - - - - - - - ' - - - - - - - - '


Figure 10. Cs extraction enhan cements for different ratios of an ion receptor 7 or 20 to CE (tetrabenzo-
24-crown·8) (I).


A promising approach using synergistic combinations of cation and anion receptors

for alkali metal salt extraction has been elucidated for the specific case of cesium nitrate.
Significant synergistic effects are observed when tripodal amide receptors for nitrate are
added to the organic phase. The synergism can be definitely attributed to anion binding,
as demonstrated by NMR spectrometry and electrospray ionization mass spectrometry.
An equilibrium model was suggested to explain the synergism in terms of 1:1 anion
binding in the organic phase. Generally, the synergistic enhancements predicted from
this model based on single-phase NMR binding constants were found to be of the
correct magnitude. Detailed equilibrium analyses remain to be performed,
including determination of the ion-pairing constants and intermolecular association
between the two recepto rs. Ion-pairing may be weak , as conditions were chosen to
favor dissociation, and the complexation of the anion is likely to decrease the
stability of the ion pair by increas ing charge separation. Likewise, the question of
possible intermolecular association needs to be addressed.
These studies have offered a new perspective on ligand design for ion-pair
extraction. It is now possible to take advantage of the versatility of such systems
for designing selective extractants and sensors for environmentally important ion
pairs. Comparisons with analogous "ditopic" ion-pair receptors capable of binding
both the cation and the an ion by a single host14 ,44 would be very instructive in
elucidating potential adva ntages or disadvantages of the dual-host versus the

single-host strategy. Further extraction experiments using different tripodal

receptors that show selectivity for nitrate, as well as systematic investigations of
other alkali metal salts in a variety of crown ether concentrations are anticipated to
elucidate aspects of speciation and selectivity . This would be further useful in
designing, synthesizing, and testing more efficient and practical dual-host systems
for applied processes.


This research was co-sponsored by the Division of Chemical Sciences, Geosciences

and Biosciences, Office of Basic Energy Sciences, under contract DE-AC05-000R22725
with Oak Ridge National Laboratory and the Environmental Management Science
Program, Offices of Science and Environmental Management, U. S. Department of
Energy, under Grant No. DE-FG-96ER62307 to the Department of Chemistry,
University of Kansas. Oak Ridge National Laboratory is managed and operated by UT-
Battelle, LLC. The participation of K.K. was made possible by an appointment to the
ORNL Postgraduate Program and subsequently the HERE program for guest faculty
administered by the Oak Ridge Associated Universities. K.K. would like to thank the FlU
foundation for funding.


1. (a) L. F. Lindoy, The Chemistry 0/ Macrocyc/ic Ligand Complexes (Cambridge Univ. Press, Cambridge,
1990). (b) Calion Binding by Macrocyc/es: Complexation ofCationic Species by Crown Ethers. edited by
Y. Inoue and G. W. Gokel (Marcel Dekker, New York, 1990). (c) Crown Compounds: Towards Future
Applications, editedby S. R. Cooper(VCH,NewYork, 1992).
2. (a) C. D. Gutsche, Cal ixarenes Rev isited (The Royal Society of Chemistry, Cambridge, 1998). (b)
P. V. Bonnesen, T. J. Haverlock, N. L. Engle,R. A. Sachleben, and B. A. Moyer, Development of process
chemistry for the removal of cesium from acidic nuclear waste by calix[4]arene-crown-6 ethers, in:
Calixarene Molecules/or Separations, edited by G. J. Lumetta, R. D. Rogers, and A. S. Gopalan, ACS
Symposium Series, Vol. 757 (American Chemical Society, Washington, DC, 2000), Chap. 3, pp. 26-44.
(c) R. Ungaro, A. Casnati, F. Ugozzoli, A. Pochini, J.-F. Dozol, C. Hill, H. Rouquette, 1,3-
Dialkoxycalix[4]arenecrown-6 in 1,3-altemate conformaticn-e-cesium-selective ligands that exploit
cation-arene interactions, Angew. Chem.• Int. Ed Engl. 33, 1506-1509(1994). (d) Z. Asfari, C. Bressot, J.
Vicens, C. Hill, I.-F. Dozol, H. Rouquette, S. Eymard, V. Lamare, and B. Toumois, Doubly crowned
calix[4]arenes in the 1,3-altemate conformation as cesium-selective carriers in supported liquid
membranes, Anal. Chem. 67, 3133-3139 (1995). (e) T. J. Haverlock, P. V. Bonnesen, R. A. Sachleben,
and B. A. Moyer,Applicability of a calixarene-crown compound for the removal of cesium from alkaline
tank waste, Radiochim. Acta 76, 103-108(1997). (f) P. V. Bonnesen, L. H. Delmau, T. 1. Haverlock, and
B. A. Moyer, Alkaline-Side Extraction of Cesium from Savannah River Tank Waste Using a
Calixarene-Crown Ether Extractant, Report ORNLrrM·13704 (Oak Ridge National Laboratory, Oak
Ridge, TN, U.S.A., December, 1998). (g) L. H. Delmau, G. J. Van Berkel, P. V. Bonnesen, and B. A.
Moyer, Improved Performance of the Alkaline-Side CSEX Process for Cesium Extraction from Alkaline
High-Level WasteObtained by Characterization of the Effectof Surfactant Impurities, ReportORNLrrM-
19991209 (Oak RidgeNational Laboratory, Oak Ridge, TN, U.S.A., October, 1999). (h) P. V. Bonnesen,
L. H. Delmau, B. A. Moyer, and R. A. Leonard, A robust alkaline-side CSEX solvent suitable for
removing cesium from Savannah River high-level waste, Solvent Extr. Ion Exch. 18, 1079-1108 (2000).
(i) T. Kakoi, T. Toh, F. Kubota, M. Goto, S. Shinkai, and F. Nakashio, Liquid-liquid extractionof metal
ions with a cyclic ligandcalixarenecarboxyl derivative, Anal. Sci. 14, 501-506 (1998). (j) P. Vanuraand
I. Stibor, Extraction of alkalineand alkaline-earth metal cations at synergistic action ofbis[undecahydro-
7,8-dicarbaundecaborato(2")]cobaltate(r) and substituted calix[n]arenes in nitrobenzene, Call. Czech.

Chem. Commun. 63, 2009-2014 (1998). (k) V. Lamare, J.-F. Dozol, S. Fuangswasdi, F. Arnaud-Neu, P.
Thuery, M. Nierlich, Z. Asfari, and J. Vicens, A newcalix4 arene-bis(crown ether)derivative displaying
an improved caesium over sodium selectivity: molecular dynamics and experimental investigation of
alkali-metal ion complexation, J. Chem. Soc.. Perkin Trans. 2271-284 (1999). (I) H. Otsuka, Y. Suzuki,
A. Ikeda, K. Araki, and S. Shinkai, Guestinclusion properties of calix[6]arene-based unimolecular cage
compounds. On their high Cs+ and Ag+ selectivity and very slow metal exchange rates, Tetrahedron 54,
3. (a) M. Dozol, in: New Separation Chemistry Techniques for Radioactive Waste and Other Specific
Applications, editedby L. Cecille, M. Casarci, and L. Pietrelli (Elsevier, Amsterdam, 1991), pp. 163-172.
(b) J.-F. Dozol, J. Casas i Garcia, and A. M. Sastre, ibid., pp. 173-185. (c) R. E. Gephart and R. E.
Lundgren, Hanford Tank Clean-up: A Guideto Understanding the Technical Issues, ReportPNL-I0773
(Pacific Northwest National Laboratory, Richland, WA, U.S.A, 1995).
4. R. A. Leonard, C. Conner, M. W. Liberatore, P. V. Bonnesen, D. J. Presley, B. A. Moyer, and G. J. Lumetta,
Developing and testing an alkaline-side solvent extraction process for technetium separation from tank
waste, Sep. Sci. Technol. 34,1043-1068 (1999).
5. (a) Y. Marcus, Ion Solvation (Wiley, Chichester, 1985). (b) Y. Marcus, Ion Properties (Marcel Dekker, New
York, 1997). (c) B. A. Moyer, Complexation and transport, in: Comprehensive Supramolecular
Chemistry. edited by G. W. Gokel (Pergamon, Elsevier, Oxford, 1996), Chap. 10,Vol. I, pp. 401-405. (d)
B. A. Moyerand P. V. Bonnesen, Physical factors in anionseparations, in: Supramolecular Chemistry of
Anions, editedby A. Bianchi, K. Bowman-James, and E. Garcia-Espana (Wiley·VCH, New York, 1997),
Chap. I, pp. 1-44.
6. F. Hofmeister, On regularities in the albumin precipitation reactions with salts and their relationship to
physiological behavior, Arch. Exp. Pathol. Pharmakol. 24,247-260 (1888).
7. (a) W. J. McDowell, Crown ethers as solvent extraction reagents: Where do we stand?, Sep. Sci. Technol. 23,
1251-1268 (1988). (b) E. Blasius, and K.-H. Nilles, The removal of cesiumfrom medium-active waste
solutions. I. Evaluation of crown ethers and special crown-ether adducts in the solvent extraction of
cesium, Radiochim. Acta 35, 173-182 (1984).
8. (a) W. J. McDowell, B. A Moyer, G. N. Case,and F. I. Case,Selectivity in solventextraction of metal Ions
by organic cation exchangers synergized by macrocycles: factors relating to macrocycle sizeand structure,
Solvent Extr. Ion Exch. 4, 217-236 (1986). (b) W. J. McDowell, G. N. Case, and D. W. Aldrup,
Investigations of ion-size-selective synergism in solvent extraction, Sep. Sci. Technol.18, 1483-1507
(1983). (c) D. D. Ensor, G. R. McDonald, and C. G. Pippin, Extraction of trivalent lanthanides by a
mixture of didodecylnaphthalenesulfonic acid and a crownether,Anal. Chem.58, 1814-1816 (1986). (d)
W. J. McDowell, G. N. Case,J. A. McDonough, and R. A. Bartsch, Selective extraction of cesiumfrom
acidic nitrate solutions with didodecylnaphthalenesulfonic acid synergized with bis(tert-butylbenzo)-21-
crown-7, Anal. Chem. 64, 3013-3017 (1992).
9. (a) K. Strzelbicki and R. A. Bartsch, Extraction of alkalimetal cationsfrom aqueous solutions by a crown
ether carboxylic acid, Anal. Chem . 53, 1894-1899 (1981). (b) T. M. Fyles and D. M. Whitfield,
Membrane transport systems. VI. Complexation of cations by lipophilic crownethercarboxylic acids, Can
J. Chem. 62, 507-514 (1984). (c) J. Tangand C. M. Wai, Solvent extraction of lanthanides with a crown
ether carboxylic acid, Anal. Chem. 58, 3233-3235 (1986). (d) R. A. Sachleben, B. A. Moyer, F. I. Case,
and S. A. Garmon, Alkylated lariat ethers as solvent extraction reagents: surveying the extraction ofalkali
metals by bis-(tert-octylbenzo)-14-crown-4-acetic acid by potentiometric two-phase titrimetry, Sep. Sci.
Technol.28, 1-23 (1993). (e) H. Sakamoto, H. Goto, M. Yokoshima, M. Dobashi, J. Ishikawa, K. Doi,
and M. Otomo, Benzocrown etherhydrazones as extractants for alkalimetal ions, Bull. Chem. Soc. Japan
66,2907-2914 (1993). (f) Y. Sakai, K. Nabeki, E. Uehara, M. Hiraishi, and M. Takagi, Effect of the
poly(oxyethylene) chainlength of ionic surfactants on the extraction of alkalimetal ions, Bull. Chem. Soc.
Japan 66, 3107-3109 (1993). (g) R. Lyazghi, Y. Pointud, G. Dauphin, and J. Juilliard, Interactions
between metalcationsand the ionophore lasalocid. Part 12. Alkaline-earth-lasalocid 1-2 complex salts in
chloroform. Formation andstructure, 1. Chem. Soc. Perkin Trans. 2 1681-1686 (1993).
10. (a) J. Honnadaly and Y. Marcus, Anion solvation properties of proticsolvents. 1. Spectroscopic study,1.
Phys. Chem. 83, 2843-2848 (1979). (b) Y. Marcus, E. Pross, and J. Hormadaly, Anion solvation
properties of protic solvents. 2. Salt distribution study, J. Phys. Chem. 84, 2708-2715 (1980). (c) Y.
Marcus, L. E. Asher, J. Honnadaly, and E. Pross, Selective extraction of potassium chloride by crown
ethers in substituted phenol solvents, Hydrometallurgy 7,27-39 (1981). (d) E. P. Horwitz, M. L. Dietz,
and D. E. Fisher, Correlation of the extraction of strontium nitrate by a crownetherwith the watercontent
of the organic phase, Solvent Extr. Ion Exch. 8, 199-208 (1990). (e) M. L. Dietz, E. P. Horwitz, S.
Rhoads, R. A. Bartsch, and J. Krzykawski, Extraction of cesium from acidic nitrate media using
macrocyclic polyethers: the roleof organic phase water, Solvent Extr.Ion Exch. 14, 1-12 (1996). (f) R. A.
Sachleben, Y. Deng, and B. A. Moyer, Ring-size and substituent effects in the solventextraction of alkali

metal nitrates by crown ethers in 1,2-dichloroethane and l-octanol, Solvent Extr. Ion Exch. 14, 995-1015
(1996). (g) R. A. Sachleben and B. A. Moyer, Ligand design for smallcations: the Li+/14-crown-4 system,
in: MetalIon Separation and Preconcentration. Progress and Opportunities, editedby A. H. Bond, M. L.
Dietz, and R. D. Rogers, ACS Symposium Series 716 (American Chemical Society, Washington, DC,
1999), pp. 114-144.
II. C. R. Duchemin, N. L. Engle, P. V. Bonnesen, T. 1. Haverlock, L. H. Delmau, and B. A. Moyer,
Solvatochromic solventpolarity measurements of alcohol solventmodifiers and correlation with cesium
extraction strength, Solvent Extr.Ion Exch. 19, 1037-1058 (2001).
12. (a) U. Olsher, F. Frolow, N. K. Dalley, 1. Weiming, Z. Y. Yu, 1. M. Knobeloch, and R. A. Bartsch, Crown-
ether alcohols as bifunctional ligands for simultaneous cation complexation and anion solvation, 1. Am.
Chem. Soc. 113, 6570-6574 (1991). (b) M. L. Dietz, E. P. Horwitz, M. P. Jensen, S. Rhoads, R. A.
Bartsch, A. Palka, J. Krzykawski, and 1. Nam, Substituent effects in the extraction of cesiumfrom acidic
nitrate mediawithrnacrocyclic polyethers, Solvent Extr. IonExch. 14,357-384 (1996).
13. For recent reviews on anion recognition see: (a) The Supramolecular Chemistry ofAnions, edited by A.
Bianchi, K. Bowman-James, and E. Garcia-Espana (Wiley-VCH,New York, 1997). (b) C. Seel and J. de
Mendoza, From chloride katapinates to trinucleotide complexes: developments in molecular recognition
of anions species, in: Comprehensive Supramolecular Chemistry, editedby F. Vogtle (Pergamon, London,
1996), Chap. 17, Vol.2, pp. 519-552. (c) B. Dietrich, Design of anion receptors: applications, Pure Appl.
Chem. 65, 1457-1464 (1993). (d) 1. L. Atwood, K. T. Holman, and J. W. Steed,Laying traps for elusive
prey: recent advances in the non-covalent binding of anions, Chem. Commun. 1401-1407 (1996). (e) F. P.
Schmidtchen and M. Berger, Artificial organic host molecules for anions, Chem. Rev. 97, 1609-1646
(1997). (f) M. M. G. Antonisse, and D. N. Reinhoudt, Neutral anion receptors: design and application,
Chem. Commun. 443-448(1998). (g) P. D. Beerand P. A. Gale,Anionrecognition and sensing: the state
of the art andfuture perspectives, Angew. Chem.. Int. Ed. Engl. 40,487-516 (2001).
14. M. T. Reetz,Simultaneous binding of cationsand anions, in: Comprehensive Supramolecular Chemistry,
editedby F.VOgtle (Pergamon, Oxford, 1996), Chap. 18,Vol.2, pp 553-562,andreferences therein.
15. L. A. 1. Chrisstoffels, F. De Jong, D. N. Reinhoudt, S. Sivelli, L. Gazzola, A. Casnati, and R. Ungaro,
Facilitated transport of hydrophilic saltsby mixtures of anion and cation carriers and by ditopic carriers, 1.
Am. Chem. Soc. 121, 10142-10151 (1999). (b) M. M. Murad, T. Hayashita, K. Shigemori, S. Nishizawa,
and N. Terarnae, Efficient ion-pair extraction of potassium chloride from aqueous phase into nitrobenzene
containing dicyclohexyt-18-crown-6 as cation binder and diphenylthiourea as anion binder, Anal. Sci. 15,
1185 (1999).
16.(a) For a review see: F. De Jong,and H. C. Visser, Carrier-mediated transport through liquid membranes, in
Comprehensive Supramolecular Chemistry; edited by D. N. Reinhoudt (Pergamon, Oxford, 1996), Vol. 10
pp 13-51. (b) H. C. Visser, D. M. Rudkevich, W. Verboom, F. De Jong, and D. N. Reinhoudt, Anion
carrier mediated membrane transport of phosphate, J. Am. Chem. Soc. 116, 11554-11555 (1994). (c) T.
Shioya, S. Nishizawa, and N. Teramae, Anion recognition at the liquid-liquid interface. Sulfate transfer
acrossthe 1,2-dichloroethane-water interface facilitated by hydrogen-bonding ionophores, 1. Am. Chem.
Soc. 120, 11534-11535 (1998).
17. D. M. Rudkevich, J. D. Mercer-Chalmers, W. Verboom, R. Ungaro, F. De Jong, and D. N. Reinhoudt,
Bifunctional recognition: simultaneous transport of cations and anions through a supported liquid
membrane. 1. Am.Chem. Soc. 117,6124-6125 (1995).
18. (a) T. G. Levitskaia, J. C. Bryan, R. A. Sachleben, J. D. Lamb, and B. A. Moyer, A surprising host-guest
relationship between 1,2-dichloroethane and the cesium complex oftetrabenzo-24-crown-8, J. Am. Chem.
Soc. 122, 554-562 (2000). (b) T. G. Levitskaia, R. A. Sachleben, R. C. Bryan, and B. A. Moyer,
Synthesis, structure, and extraction behavior of 4,5',4",5"'-tetra-tert-butyltetrabenzo-24-crown-8, 1. Chem.
Soc-Perkin Trans. 2,808-814 (2001). (c) R. A. Sachleben, Y. Deng,and B. A. Moyer, Sep. Sci. Technol.
32,275-284 (1997).
19. J. C. Bryan, K. Kavallieratos, and R. A. Sachleben, Unusual ligand coordination for cesium, Inorg.
Chem. 39, 1568-1572 (2000).
20. J. C. Bryan, R. A. Sachleben, and B. P. Hay, Structural aspects ofCs selectivity by tetrabenzo-24-
crown-8,Inorg. Chim. Acta 290,86-94, (1999).
21. (a) A. P. Bisson, V. Lynch, M. K. C. Monahan, and E. V. Anslyn, Recognition of anions through N-H. .,lt
hydrogen bonds in a bicyclic cyclophane-Selectivity for nitrate. Angew. Chem., Int. Ed. Engl. 36,
2340-2342 (1997) . (b) R. Jagessar and D. H. Burns, (c i s)-5,10,15,20-Tetrakis[2-
(arylurea)phenyl]porphyrins: novel neutral ligands for remarkably selective and exceptionally strong
chloride anioncomplexation in (CDJ)2S0.Chem. Commun. 1685-1686 (1997). (c) M.Staffilani, K. S. B.
Hancock, 1. W. Steed, K. T. Holman, J. L. Atwood, R. K. Juneja, and R. S. Burkhalter, Anion binding
within the cavity of p-metalated calixarenes, J. Am. Chem. Soc. 119, 6324-6335 (1997). (d) R. D.

Schnebeck, E. Freis inger, and B. Lippert, A novel highly charged (+12) receptor that encapsulates
simulatenously nitrate and hexafluorophosphate anions, Angew. Chem.• Int. Ed. Engl. 38,168-171 (1997).
(e) P. D. Beer, M. G. B. Drew, and R. Jagessar, Selective anion recognition by novel 5,10,15,20-
tetrakis(o-ferrocenylcarbonylaminophenyl-substituted) zinc metalloporphyrin receptors , J. Chem. Soc.
Dalton Trans. 881-886 (1997). (f) P. D. Beer, M. G. B. Drew, D. Hesek, and R. Jagessar , Spectral and
electrochemical anion sensing by a novel 5,1O,15,20-tetrakis(R-subst ituted) porphyrin receptor, 1. Chem.
Soc. Chem Commun. 1187-1188 (1995). (g) S. Mason, T. Clifford, L. Seib, K. Kuczera, and K. Bowman-
James, Unusual encapsulation of two nitrates in a single bicyclic cage, J. Am. Chem. Soc. 120, 8899-8900
(1998). (h) R. J. Motekaitis, A. E. Martell, J.-M. Lehn, and E.-I. Watanabe, Bis(2,2',2"-
triaminotriethylamine) cryptates of cobalt(II), nickel(II), copper(II), and zinc(II) . Protonation constants,
formation constants, and hydroxo bridging, Inorg. Chem. 21,4253-4257 (1982).
22. (a) G. A. Jeffrey and W. Saenger, Hydrogen Bondingin BiologicalStructures (Springer-Verlag, Berlin,
1991). (b) P. Gilli, V. Bertolasi, V. Feretti, and G. Gilli, Evidence for resonance-assisted hydrogen
bonding: Covalent nature of the strong homonuclear hydrogen bond. Structure of the O-H... O bond by
crystal structure correlation methods, J. Am. Chem. Soc. 1994, 116, 909-915. (c) B. P. Hay, D. A. Dixon,
J. C. Bryan, and B. A. Moyer, Crystallographic evidence for oxygen acceptor directionality in oxyanion
hydrogen bonds,J. Am. Chem. Soc. 124, 182-183 (2002).
23. B. A. Moyer, P. V. Bonnesen , L. H. Delmau, T. J. Haverlock , K. Kavallieratos , and T. G. Levitskaia,
Toward recognition of the anion in the extraction of alkali metal salts by crown ethers and calixarenes, in:
Proceedings of the International Solvent Extraction Conference (ISEC 2002), edited by K. C. Sole P. M.
Cole, J. S. Preston, D. J. Robinson (Chris van Rensburg Publications, Melville, South Africa, 2002), Vol.
I, pp. 299-306.
24. T. J. Haverlock, P. V. Bonnesen, R. A. Sachleben, and B. A. Moyer, Analysis of equilibria in the extraction
of cesium nitrate by calix(4)arene-bis(t-octylbenzo-crown-6) in 1,2-dichloroethane, J. Inclusion Phenom.
Macrocyclic Chem. 36, 21-37 (2000).
25 . K. Kavallieratos, J. C. Bryan, G. 1. Van Berkel, and B. A. Moyer, Complexation and extraction of
cesium salts by a dual-host strategy : structural and thermodynamic aspects of synergism and
selectivity, Inorg. Chem. (in preparation).
26. K. Kavallieratos, R. A. Sachleben, G. J. Van Berkel, and B. A. Moyer, Novel dual-host approach in ion pair
extraction: A simple tripodal nitrate host facilitates CsN03 transfer to 1,2-dichloroethane by a large crown
ether, Chem. Commun. 187-188 (2000).
27. (a) P. D. Beer, F. Szemes, V. Balzani, C. M. Sala, M. G. B. Drew, S. W. Dent, and M. Maestri, Anion
selective recognition and sens ing by novel macrocyclic transition metal receptor systems . IH-NMR ,
electrochemical, and photophysical investigations, J. Am. Chem. Soc. 119, 11864-11875 (1997). (b) A. P.
Davis, J. F. Gilmer, and 1. 1. Perry, A steroid-based cryptand for halide anions, Angew. Chem., Int. Ed.
Engl. 35 , 1312-1315 (1996). (c) Y. Morzherin, D. M. Rudkevich W. Verboom , and D. N. Reinhoudt,
Chlorosulfonylated calix(4)arenes: precursors for neutral anion receptors with a selectivity for hydrogen
sulfate, 1. Org. Chem. 58, 7602-7605 (1993). (d) C. Seel and F. Vogtle, Molecules with large cavities in
supramolecular chemistry, Angew. Chem.. Int. Ed. Engl. 104,542-63 (1992).
28. (a) A. P. Davis, J. 1. Perry, and R. P. Williams, Anion recognition by tripodal receptors derived from cholic
acid,J. Am. Chem. Soc. Il9. 1793-1794 (1997). (b) P. D. Beer, Z. Chen, A. J. Goulden, A. Graydon, S. E.
Stokes, and T. Wear, Selective electrochemical recognition of the dihydrogen phosphate anion in the
presence of hydrogen sulfate and chloride ions by new neutral ferrocene anion receptors, J. Chem. Soc.•
Chem. Commun. 1384-1386 (1993) . (c) C. Raposo, N. Perez, M. Almaraz , M. L. Mussons, M. C.
Caballero, and J. R. Moran, A cyclohexane spacer for phosphate receptors , Tetrahedron Lett. 36 ,
3255-3258 (1995).
29. (a) K. Kavallieratos, S. R. de Gala, D. 1. Austin, and R. H. Crabtree, A readily available non-preorganized
neutral acyclic halide receptor with an unusual nonplanar binding conformation, 1. Am.. Chem. Soc. Il9,
2325 ( 1997). (b) K. Kavallieratos, C. M. Bertao , and R. H. Crabtree, Hydrogen bonding in anion
recogn ition: a family of versatile, nonpreorganized neutral and acyclic receptors, J. Org. Chem. 64 ,
1675-1683 (1999) . (c) F. Garcia-Tellado, S. Goswami, S. K. Chang, S. J. Geib, and A. D. Hamilton,
Molecular recognition:.a remarkably simple receptor for the selective complexation of dicarboxylic acids,
J. Am. Chem. Soc. Il2, 7393-7394 (1990). (d) E. Fan, S. A. van Arman, and A. D. Hamilton, Molecular
recognition : hydrogen-bonding receptors that function in highly competitive solvents, J. Am. Chem. Soc.
115. 369-370 (1993).
30. K. A. Connors, BindingConstants- The Measurement of Molecular Complex Stability (John Wiley, New
York, 1987), pp. 189-215.
3 I. K. Kavallieratos, J. C . Bryan, G. J. Van Berkel , M. A. Kelly , R. A. Sachleben, and B. A. Moyer,
lon-Pair recognition by a dual-host strategy: binding and extraction of cesium nitrate by a
combination of tetrabenzo-24-crown-8 and a simple tricarboxamide anion host, 1. Org. Chem. (in

32. K. Kavallieratos, A. Danby, G. 1. Van Berkel, M. A. Kelly, R. A. Sachleben, B. A. Moyer, and K.
Bowman-James, Enhancement of CSNOl extraction in 1,2-dichloroethane by tris(2-aminoethyl)amine
triamide derivatives via a dual-host strategy, Anal. Chem. 72, 5258-5264 (2000).
33. S. Valliyaveettil, 1. F. J. Engbersen, W. Verboom, and D. N. Reinhoudt, Synthesis and complexation
behavior of uncharged anion-receptors, Angew. Chem.. Int. Ed. Engl. 32, 942-944(1993).
34. P. D. Beer, P. K. Hopkins, andJ. D. McKinney, Cooperative halide, perrhenate anion-sodium cationbinding
and pertechnetate extractionand transport by a novel tripodal tris(amido benzo-Ic-crown-S) ligand, Chem.
Commun. 1253-1254 (1999).
35. (a) L. Prodi, F. Bolletta, M. Montalti, andN. Zaccheroni, Searching for newluminescent sensors. Synthesis
and photophysical properties of a tripodal ligand incorporating the dansyl chromophore and of its metal
complexes, Eur. 1. Inorg. Chem. 455-460(1999). (b) L. Prodi, M. Montalti, N. Zaccheroni, F. Dallavale,
G. Folesani, M. Laufranchi, R. Corradini, S. Pagliari, and R. Marchelli, Dansylated polyamines as
fluorescent sensors for metal ions: photophysical properties and stability of copper(II) complexes in
solution, Helv. Chim. Acta 84, 690-706(2001). (c) X. Wang, andH,-J. Schneider,Binding ofdansylamide
derivatives to nucleotides andnucleic acids, 1. Chem. Soc.. Perkin Trans 2,1323-1328 (1998).
36. For a detailed experimental description for compounds 16-12 and 18, see ref. 32. The NMR titration
experiment and determination of the binding constant for compound 19 (ref. 37) was carried using the
well-established method reported elsewhere forotheranion receptors (ref. 29b).
37. O. D. Espetia, and K. Kavallieratos, Anion biding properties of tripodal N-dansylamide derivatives,
Presented at the 223dNational Meeting of theAmerican Chemical Society, Orlando FL, April2002.
38. A. Danby, L. Seib, N. W. Alcock, and K. Bowman-James, Novel structural determination of a bilayer
network formed by a tripodal lipophilic amide in the presence of anions, Chem. Commun. 973-974
39. A. V. Hill, Thermal method of measuring the vapor pressure of aqueous solutions, Proc. Roy. Soc.
(London), A127, 9 (1930).
40. The aggregation factor for a given concentration is defined as the ratio of the theoretical (total)
concentration of the receptor to the totalexperimental concentration of amidespecies as measured by the
vapor-phase osmometry experiment.
41. (a) T. D. P. Stack, Z. Hou, and K. N. Raymond, Rational reduction of the conformational space of a
siderophore analog through nonbonded interactions: the roleof entropy in enterobactin, 1. Am. Chem. Soc.
115, 6466-6467 (1993). (b) T. B. Kaprishin, T. D. P. Stack, and K. N. Raymond, Octahedral versus
trigonal prismatic geometry in a seriesof catechol macrobicyclic ligand-metal complexes, 1. Am. Chem.
Soc. 115, 182-192 (1993). (c) Z. Hou, T. D. P. Stack,C. J. Sutherland, and K. N. Raymond, Enhanced
iron(III) chelation through ligand predisposition: syntheses, structures and stability of tris-catecholate
enterobactin analogs, Inorg. Chim. Acta 263, 341-355(1997).
42. Prepared from 1,3,5-tris(aminomethyl)-2,4,6-trimethylbenzene trihydrochloride (ref. 40) using the
following procedure: To a suspension of 0.200 g (0.632 mmol) of 1,3,5-tris(aminomethyl)-2,4,6-
trimethylbenzene trihydrochloride (ref. 36) in methylene chloride was added excess of
triethylamine (1.28 g). After 20 min., a stoichiometric amount of 4(tert-butyl)benzoylchloride (3
equiv, 0.373 g, 1.89 mmol) was added, and the reaction mixture was stirred for 10 h. The
reaction mixture was then transferred to a separatory funnel using additional methylene chloride,
washed with I N HCI (three times each with 20 ml.), and subsequently with saturated NaHC03
(two times each with 40 mL) and water (two times each with 40 mL). During the washing it was
necessary to add methylene chloride in order to avoid precipitation. The organic layer was dried
through a Na2S04 column and reduced in volume by rotary evaporation until the product 20
started precipitating. The crystalline product was filtered and washed with cold methylene
chloridelhexanes (0.234g, 54.0% yield). lH_NMR (CDCI3, 400MHz): 7.67, (d., 2H, Ar-H), 7.42
(d., 2H, Ar-H), 5.89 (br., 3H-CONH), 4.72 (d., 6H, -CH-r), 2.47 (s., 9H, -CH3), 1.31 (s., 27H,
CH3-tert-Bu) .
43. T. Grawe, T. Schrader, M. Gurrath, A. Kraft, and F. Osterod, Self-organization of spheroidal
molecular assemblies in polar solvents, Org. Lett. 2, 29-32 (2000).
44. See chapter J. M. Mahoney, J. P. Davis, and B. D. Smith, Ditopic salt-binding receptors for
potential use in anion separation processes, This Volume.

Holger Stephan, I Karsten Gloe,2 Werner Kraus,' Hartmut Spies,I Bernd

Johannsen,' Kathrin Wichmann.' GUnter Reck,3 Dillip K. Chand," Parimal
K. Bharadwa]," Ute MUller,s Walter M. MUller,s and Fritz Vogtle s*


The design and synthesis of anion receptors of technical and biochemical

significance is receiving more and more attention .J·8 Currently, effective binding and
selective phase transfer of the oxoanions pertechnetate and perrhenate is of considerable
interest from different point of view . Due to its long half-life and environmental mobility,
the radioactive pertechnetate is one of the most hazardous contaminants. In this context ,
effective and selective separation processes are of utmost importance.v':' On the other
hand, there are some emerging possibilities for the application of the radiochemically
active oxoanions pertechnetate and perrhenate in nuclear medicine.P'" The most
commonly used isotope in diagnostic nuclear medicine 99mTc is readily available from a
99 Mol'9m Tc generator system . .
15 J9
Likewise, the p-emitting J88Re - discussed as one of
the most interesting radionuclides for specific therapeutic applications - is conveniently
produced by a 188W/188Re generator.20 •22 In both cases the radionuclides are ava ilable as
oxoanions in isotonic solution, and it appears highly desirable to directly complex
99mTc04' and J88Re0 4' as they exist in the generator eluate itself. But, the binding of such
large, lowly charged anions is a difficult venture. The enthalpic contribution for
complexation is rather small. Hence, host compounds being capable to encapsulate these
oxoanions are of great interest.
For both directions discussed above, some different requirements have to be fulfilled
for the design of ligands . The essential properties of ligands serving as extractants and
imaging/therapeutic agents are summarized in Table 1.

I Forschungszentrum Rossendorf, Institut fur Bioanorganische und Radiopharmazeutische Chemie, PF

510119,01314 Dresden, Germany. ' Technische Universitat Dresden, Institut fur Anorganische Chemie,
01062 Dresden, Germany. ' Bundesanstalt fur Materialforschung und -prufung, Richard-Willstatter-Strasse
II , 12489 Berlin, Germany. 'Indian Institute of Technology, Department of Chemistry, Kanpur 208016,
India. ' Universitat Bonn, Kekule-Institut fur Organische Chemie und Biochemie, Gerhard-Domagk-Strasse
I, 53121 Bonn, Germany

Fundamentals andApplications of AnionSeparations, edited by

Moyer and Singh, Kluwer Academic/Plenum Publishers, New York, 2004 ISl

Table 1. Requirements for the design of ligands capable of serving as extractant and
imaging/therapeutic agent.


The receptor has to interact reversibly with The receptor should form an inclusion
the anion in a specific way. This should compound with the oxoanion in isotonic
overcome the solvation energy. solution.

The host and the complex formed should The receptor/anion complex should have
display high Iipophilicity to avoid leakage a well-balanced Iipophilicity (LogP: 1-
of the extractant into the aqueous phase. 2.5).

The receptor should allow a rapid No exchange reaction with endogenous

complexation/decomplexation across phase species should happen in physiological
boundary. environment.

TeO,' ReO,

r = 252 pm r = 260 pm
pK, (IITeO, ) = 0 033 pK. (I/ ReO,) = -0.2
E' ': = 0 74 V £" J:= 0 51 V
toG.,.. = - 251 kl -rnol' toG = - 330 k.l-rnol'

Figure I. Properties ofperteehnetate and perrhenate .i' :"

Figure 1 presents the primary properties of the tetrahedral anions pertechnetate and
perrhenate. Cage compounds having the size and binding centers required for anions may
be very promising for the molecular encapsulation of these oxoanions. Azacryptands in
the protonated state represent such a kind of receptor and are well-known to bind
different oxoanions in the interior." Some solid anion cryptates were isolated, and the
encapsulation of nitrate." perchlorate,29.3o chromate, selenate, and thiosulfate" was
demonstrated. The cage structures of the above mentioned anion cryptates have
similarities. In all cases, the azacryptands are fully protonated, and as a result of
electrostatic repulsion , the cavity is expanded. In particular, at low pH these cryptands
serve as good anion receptors because of the geometric complementarity of H-bonding
and electrostatic interaction between host and guest. " In contrast to this, the environment
is completely different if the binding of pertechnetate and perrhenate in the generator
eluate is considered. At physiological pH, the azacryptands are only partially protonated,
and consequently the binding strength and the shape of the cages should differ
remarkably in comparison to the strongly acidic media.
We are especially interested to characterize the binding and distribution behavior of
selected cage compounds towards pertechnetate and perrhenate at neutral pH.
Azacryptands of different size possessing nitrogen atoms as binding centers have been
chosen (cf. Fig. 2). Aromatic spacer elements and amide groups were introduced in order
to improve the lipophil icity, to lower the flexibility, and to vary the binding mode of the

1 2 3

4 5 6

Figure 2. Structures of azacages investigated.


2.1. Liquid-Liquid Extraction Experiments

Liquid-liquid extraction studies were selected to characterize the binding and

distribution behav ior of the complexes formed by different azacages with pertechnetate
and perrhenate. Using 99m TC0 4' and 188Re0 4' radiotracer, it is easy to get reliable and

precise information about the extraction efficiency and the complex composition in the
organic phase. Figure 3 shows the structural influence of the azacryptands 1-6 on the
extractability of pertechnetate and perrhenate in neutral medium . The most effective
extraction of pertechnetate and perrhenate is obtained by cage compound 1 having both
triethanolamine and tris(2-aminoethyl)amine (tren) caps bridged with a tolyl spacer in the
ortho-position. Amazingly, the structure related compound 2 connected in the meta-
position gives significant lower extraction. Also the double tren-capped azacryptands
3 and 4 show a drastically reduced extraction efficiency . In all these cases, a rapid
attainment (within some minutes) of both the extraction and back extraction (by lowering
the pH) equilibrium was observed . Using the amidocryptands 5 and 6, no transfer of
pertechnetate and perrhenate into the organic phase was found under the experimentally
chosen conditions.

Extractability 1%1


/ '"
47 .2

- -
60 13.9

1 0.1 0.1

- -
40 ~
< <
20 0 .5 0.1 0.1
< <

2 J 5 6

Figure 3. Extractability of pertechnetate and perrhenate by azacryptands 1-6. [NaTcO,] or [NaReO,) =

1·10-1 M; pH = 7.4 (HEPESlNaOH buffer) ; [ligand) = I ·) 0" M in CHCh.

It is worth mentioning here that in all cases pertechnetate is slightly better extracted
in comparison to Eerrhenate . This is a general trend not only for these azacages but also
for guanidinium 3 and ammonium" compounds, dendrimersr':" and bimetallic
cyclotriveratrylene hosts." Differences in hydration state are likely responsible for this
behavior . Regrettably, there are only estimated thermodynamic data available for
pertechnetate, revealing a lower hydration energy (Mt\Ydr = -251 kJ/mol) than for
perrhenate (Mt\ydr = -330 kJ/mol) .24 Charge-density calculations" corroborate this
finding. The partial negative charge of oxygen atoms of perrhenate (-0.755) is clearly
increased compared to pertechnetate (-0.739). This fact should lead to a significant
stronger hydration, and consequently to a more difficult perrhenate transfer into an
organic solvent.
In order to obtain a deeper insight into the extraction equilibrium, the influence of
pH on the extraction of pertechnetate was studied. The results determined for azacages I
and 2 are summarized in Figure 4. The extractability increases with rising pH, reaching a
maximum at the pH range between 7 and 8 and decreases in the more basic medium . It
can be clearly concluded that this extraction behavior corresponds to the different
protonation state of the azacages in aqueous solution." Under acidic conditions, the cage
compounds are highly protonated, and as a result become really hydrophilic . This
explains their rather poor extraction ability from aqueous into organic solution. The
extraction efficiency is only slightly different for I and 2 in the pH range between 2 and
7. But, at higher pH azacage I is superior to 2. This fact corresponds to the dominating
formation of the LHr species of I between pH 7 and 8. In this range, 2 forms mainly the
LH/ +species.

Log D T c04

.• .
' • I •

. .. . :-
•• ••

I •
• • ••
pK. , = 10.4S ••
pK., = 10.4.- • 2.
PJ<d -= S.S6 •
pK•• = 3.46· • pK. , = 9.7'-
-2 pK. , = 8.6
pK., - .4
• pK•• = S.6
pK•• = 2.4"

0 2 .. 6 8 10

Figure 4. Extraction ofpertechnetate with azacryptands I and 2 as a function of pH. [NaTcO.) = 1·10-4 M; pH
= 2.0-5.2 (NaOAc/HCI buffer ); pH = 5.4-61(MES/NaOH); pH = 7.1-8.0 (HEPES/NaOH); pH = 80-90
(TAPS/NaOH); [ligand) = 1·\0"' M in CHC\,; • pK, (HlO, 0.1 M [(CH,).N)NO,) from Ref 39.

Extraction of pertechnetate and perrhenate was also examined at different

concentrations of cage compounds. The results, shown in Fig. 5, reveal an essentially
linear relationship between the distribution ratio and the azacryptand concentration. The
slopes of the lines in the log Drc04/ReOJlog Cligand diagram were unity, indicating a clean
I: I composition of the extracted complexes . Loading experiments of the organic phase at

the same experimental conditions give a maximum ratio of ligand to anion of 1:2. It
follows that the cage compounds can be transferred into the organic phase in the mono-
and diprotonated form with one or two bound pertechnetate/perrhenate anions. It seems
to be plausible because the extraction of higher charged species should be energetically

Log D u ion




-4 -3 -2

Figure 5. Extraction of pertechnetate and perrhenate with azacryptands I and 2 as function of ligand
concentration . [NaTcO. _ , NaReO. .] = ).104 M; pH = 7.4 (HEPESlNaOH buffer); [ligand] = .10" M in CHC!,

Generally, the penetration of anions into the cavity of azacages 1 and 2 may be
hindered by encapsulated water molecules.t' '" Furthermore, derived from the size
relation and molecular modeling calculations, the cage cavity of these azacryptands
seems to be too small in particular for the large pertechnetate and perrhenate oxoanions.
On the other hand, the cryptands 3-6 have a more suitable size for the encapsulation
of these large anions. This was also shown by molecular modeling .P Unfortunately, these
compounds show a poor extraction for pertechnetate and perrhenate in the neutral media
(cf. Fig. 3).
Regarding the extraction as function of pH, the double tren-capped cryptands 3 and 4
(cf. Fig.6) follow the same trend as obtained for 1 and 2 (cf. Fig.4). That can be also
explained on the basis of the different protonation state in dependence on pH.43,44 The
lower extraction efficiency of 3 and 4 in comparison to 1 and 2 should lie in the higher
overall basicity connected with a greater number of secondary nitrogen atoms, and
consequently the complexes formed should have a lower lipophilicity in the acidic and
the neutral med ia. This can be illustrated by a high proportion of the higher charged
species LH/+, L~4+and LH/ +of 3 and 4 present at the pH range between 7 and 8.
As shown in Fig. 6, the extraction power of 6 is significantly stronger in comparison
to 1--4 in the acidic pH range . Because in the case of 6, only the two bridgehead nitrogen
atoms can be protonated and favor the anion transfer. Therefore, between pH 2 and 3 the
pertechnetate extractabilities are in the same order of magnitude as for 1 and 2 at neutral
conditions . Only slight differences of pertechnetate and perrhenate extraction were
obta ined for 5 and 6, apparently caused by the change in the lipophilicity in going from
the acetylated compound 5 to the benzoylated der ivative 6.

1 Log DTe~
pK. , = 9.4
" pK" = 8.8
pK. , = 7.8
" pK., = 7.1
pK. , = 6.2

pK•• = 5.4" ••• • •
-1 6 • " ... ••
"" " "

." •
"". ,.•• • • 3•
•• •
• pK. , = 9.6
I pK. , =9.0
pK.. = 8.6
-3 pK., = 7.4
• pK., = 6.7
pK•• = 6.5"
0 2 4 6 8 10
Figure 6. Extraction of pertechnetate with cryptands 3. 4 and 6 as function of pH. (experimental conditions are
the same as in Fig. 4) [3,4) = 1·10-) M in CHCI,; (6) = 5-10·) M in CHCI); • pK. (H,O, 0 I M [(C,H ,),N)CIO,
from Ref 43.

In order to obta in information about the Iipoph ilicity of the cages, we have
determ ined the partit ion of 1--4, and 6 between water and I-octanol. The concentrations
of azacryptands in the organ ic and aqueous phase were determ ined by UV measurements.
At pH = 7.4 more than 96% of 1,3 and 4 rema ins in the aqueous phase . On the other
hand, 70% of 2 and 95% of the amidocryptand 6 are preferentially located in the organic
phase. The part ition behav ior of 6 can be explained on the basis of low basicity and
increasing Iipophil icity caused by the nonprotonable amidofunction compared to
aminocages, and it is in good agreement with the results obtained by mass
spectrometry." Also, the aminocages 1,3 and 4 are preferentially located in the organ ic

phase after complete deprotonation. Accordingly, 98% of 1 is transferred into I-octanol

at pH > 10. It is true that amidocages have the appropriate Iipophilicity to act as an
extractant, but owing to the loss of the protonable secondary amine groups as binding
centers, the ability to transfer the oxoanions pertechnetate and perrhenate disappears in
particular in the neutral media. Knowing that pyridine nitrogen 43 should also not
protonate at pH > 2, we assume that the extraction behaviour of amidocages 5 and 6 is
caused only by the protonation of the bridging tertiary amine nitrogens at low pH.
Information about the overal1 stoichiometry of complexes extracted was also
obtained by measuring the distribution ratios of anions between the organic and aqueous
phase as a function of the ligand concentration (cf. Fig. 7). The data determined with
slopes of unity were consistent with preferential 1: I complex formation of pertechnetate
and perrhenate with the cryptands 3--6. In case of 3 and 4 the maximum loading of the
organic phase by pertechnetate or perrhenate is characterized by a ligand to anion ratio of

Log D an ion

o •



-4 -1-- - - - - - - - - - - - .....- - - - - - - - - - - --,
-4 -3 -2

Figure 7. Extraction of pertechnetate and perrhenate with cryptands 3 - 6 as function of ligand concentrat ion.
(experimental conditions are the same as in Fig. 5).

2.2. Structural Considerations

Unfortunately , solvent extraction studies reveal no information about the complex

structure . The anion can be arranged inside and outside the cavity. The formation of
anion cryptates is mainly influenced by the host/guest complementarity of size and
binding mode, the protonation and solvation state of the cage compounds, but also by
charge and solvation of the anion. Taking into account the large size of pertechnetate and
perrhenate , there is only the possibility to bind one anion inside the cavity of the
cryptands investigated .
As can be seen from Fig. 8, generally some different coordination patterns are
possible. Considering the monoprotonated form of a host with an adequate size, the
mononegative oxoanion may be bound inside of the cavity. But this desired binding
mode is often prevented by unfavorable ligand conformation . Furthermore, the cavity can
be blocked by bound solvent molecules, in particular water. Consequently, the
penetration of the oxoanion should be complicated. Nevertheless, the first perrhenate
inclusion complex with a hexaprotonated aminocage could be isolated."

[ H4]lcOi

Figure 8. Possible coordination patterns for binding pertechnetate and perrhenate by azacage 4 in mono and
diprotonated form.

In order to find more detailed information about the structural arrangement of the
anion with regard to the above questions , we have generated some solid crystals of
perrhenate complexes with azacages . In contrast to the usual procedure of applying
strongly acidic solutions to azacryptands 27•3o, we started from cage monohydrochlorides

dissolved in methanol. Sucha solution was passed through a column loaded with a strong
basic anion exchanger in the perrhenate form. This procedure was chosen because we are
especially interested to get more information about the binding mode of azacages with
perrhenate in the neutral media. In the case of azacryptand I, single crystals of sufficient
X-ray diffraction quality were isolated. The structure of the complex formed with
perrhenate is shown in Fig. 9.


Figure 9. X-ray structure ofazacrypland 1 with perrhenate [perspective views: (a) lube (b) space filling).
In the same way as found by solvent extraction, I: I comple x formation in the solid
state was observed. The perrhenate anion is arranged outside the cavity , partially
embedded between two benzene rings. This structure leads to a shortening of the distance
between the bridgehead nitrogens [N(\}-N(2): 5.812 A] compared to the free ligand
(6.249 A).38 Weak hydrogen bonding is observed between one perrhenate oxygen atom
and the two methylene hydrogen atoms of the tolyl spacer unit [O(Re04'}-H(CH 2) : 2.575
and 2.658 A]. Such weak hydrogen bonding of the perrhenate anion with methylene
hydrogen atoms is also known for crown compounds.V:" We found that one water
molecule is encapsulated by the cryptand. That is very similar to a rhodanide complex of
1.41 In this case, the cavity is also blocked by water, and the anions are arranged outside
of a diprotonated host. On the other hand, a crystal structure of perchlorate with 1 shows
that even if the cage is fully protonated, the anions are bound at the periphery ."
Molecular modeling calculations" confirm the preferential arrangement of anions outside
the cavity of 1.
Also for azacage 2 a complex structure was found where the perrhenate anion is
arranged at the periphery of the ligand molecule (cf. Fig. 10). In contrast to I, the azacage
2 is diprotonated, and it results in a 2: 1 complex (perrhenate:2) in the solid state, which is
in agreement with the maximum loading in solvent extraction experiments. The structure
of the perrhenate complex is very similar to the corresponding perchlorate complex"
isolated. Accordingly, the distances between the bridgehead nitrogens [N(I), N(2)] of
9.003 A and 9.052 A are almost the same for the perrhenate und perchlorate complex .
Likewise, two water molecules are bound inside the cavity . In the case of the perrhenate
complex, one perrhenate anion forms a strong hydrogen bond to the protonated
secondary nitrogen N(3). The distance from one oxygen atom of Re04' (\) to N(3) is
2.77 A. Moreover, the two perrhenate molecules are bridged by water, forming two
additional hydrogen bonds (2.72 A and 2.96 A).
In the case of azacryptand 3, the X-ray crystal structure analys is led to a further
different structure arrangement (cf. Fig. 11). The asymmetric unit contains two
independent, but very similar molecules . There is no indication of any interactions
between these molecules. That is comparable to a structure of a bistren capped azacage
bridged with xylyl spacers in the meta-position recently described.V In the case of the
perrhenate complex, the cage structure is stabilized in a rather flat arrangement. The
distances between the bridgehead nitrogen atoms are 10.224 and 10.072 A, respectively .
The formation of two intramolecular hydrogen bonds between secondary amine nitrogen
atoms of each cage molecule should be responsible for that (N-H-N: 2.86,2.96 A; 2.87,
2.96 A). The resulting long and narrow structure of the azacage leads once again to an
arrangement where the perrhenate anions are bound outside the cavity. Each cage
molecule is surrounded with two perrhenate and three water molecules.
To sum up, the occupation of cryptands by water molecules as found for azacages 1
and 2 as well as the formation of intramolecular hydrogen bonds evidenced for 3 hamper
the penetration of anions into the cavity . Furthermore, a low state of protonation seems to
favor complex structures where the anions are bound outside the cavity. Nevertheless, in
particular bistren capped azacages are very promising in view of the encapsulation of
large anions. As evidenced in the solid state, the pyridine-containing bistren cryptand 4 is
able to accommodate the hexafluorosilicate anion .29 The SiF/ anion has a radius of 2.59
A that is very similar to Re04' (2.60 A). Likewise , the bistren capped azacage bridged
with xylyl spacers in the meta-position encapsulates the large dinegative oxoanions

chromate (2.40 A), selenate (2.43 A), and thiosulfate (2.50 A).3\ Further studies using this
compound as anion receptor have shown, that oxalate can be included with distinctive
stability." In case of the fluoride ion, three different complex structures with similar
cages were found. Among them the first anion-based cascade complex, in which a water
molecule bridges two fluoride ions inside the cage 3.50




Figure 10. X-ray structure of azacryptand 2 with perrhenate [perspective views: (a) tube (b) space filling).

ReO , (2')

ReO ,- (2)


(b )

Figure 11. X-ray structure of azacryptand 3 with perrhenate [perspective views: (a) tube (b) space filling].


Aminocages 1--4 and amidocages 5 and 6 are capable of extraction of the large
oxoanions pertechnetate and perrhenate. Extraction behavior was consistent with clean
1: I complex formation at an excess of the ligand over the anion concentration. A
maximum loading of two anions per ligand molecule was observed at higher anion
concentration. In all cases, pertechnetate is slightly better extracted than perrhenate.
The efficiency of oxoanion extraction correlates especially with the acid-base
behavior of the cage compounds and the Iipophilicity of the anions. In view of the
application as extractants, the Iipophilicity of aminocages has to be increased.
Preliminary results show that both the Iipophilicity and extraction efficiency are
remarkably enhanced after methylation of secondary amine nitrogen atoms of the cage."
The amidocages have the appropriated Iipophilicity. But, due to the preferred protonation
at low pH, these cryptands only exhibit high extraction of oxoanions in acidic solution.
The replacement of tertiary by secondary amide groups seems to be an interesting way to
improve the extraction efficiency in the neutral media. In this case the oxoanions can be
additionally stabilized by hydrogen bonds.
Another promising approach is based on the use of open-chain counterparts of the
cage compounds having the tren unit modified by lipophilic moieties."
The molecular encapsulation of pertechnetate and perrhenate for imaging and
therapeutic purposes is a challenging task. To achieve a high stability in vivo the
oxoanions have to be mechanically locked into cage compounds.


4.1. Synthesis

Reagent-grade chemicals were used as provided. The cryptands 1--4 were prepared
as described earlier (1,40 2,51 3,27 and 427). The amidocryptands 5 and 6 were obtained by
acetylation and benzoylation, respectively, of pyridine-containing cryptand 4. In a typical
experiment, 3 mmol acetylchloride (benzoylchloride) dissolved in 30 mL dry
dichloromethane was added over a period of 1.5 h to a solution of 0.5 mmol 4 and
4 mmol triethylamine in dry dichloromethane (50 mL). After the addition was complete,
the solution was heated to reflux for 2 h. The solution was cooled up to room
temperature, and was washed with KHC03 (10% in water) and finally with water. After
drying the organic layer with anhydrous Na2S04, the solvent was removed in vacuo. The
residue was purified by MPLC (Lichroprep Si0 2, 15-25 11m, CH2ChIMeOH = 100:25).
Amidocryptand 5: glassy solid in 55.9 % yield;
FAB MS (matrix: 3-nitrobenzyl alcohol): m/z (%): 854.5 (100) [M+];
C4sH63NII06: 854.06.
Amidocryptand 6: glassy solid in 61.5 % yield;
FAB MS (matrix: 3-nitrobenzyl alcohol): m/z (%): 1226.5 (100) [~] ;
C7sH7SNII06: 1226.49.
Perrhenate complexes of the azacryptands 1-3 were prepared on the following way:
10 mg of cryptand was dissolved in 10 mL methanol/acetonitrile (9/1) and the equimolar
amount of 0.1 M HCI was added. This solution was passed through a column filled with
2 mL strongly basic anion exchange resin (DOWEX®IXA-200) in the perrhenate form.
After washing the column with 50 mL methanol and finally with 5 mL acetonitrile, the
eluate was completely evaporated. The residue obtained was dissolved in CH 2CI 2, and the
solution was dried using anhydrous Na2S0 4. After complete removal of the solvent,
colorless solidified oils were obta ined. X-ray-quality crystals were grown by slow
evaporation of an acetonitrile solution at room temperature.

4.2. Liquid-Liquid Extraction Procedure

Extraction studies were performed at 25 ± I °C in 2 em' microcentrifuge tubes by

mechanical shaking. The phase ratio V(org):V(w) was I :I (0.5 crrr' each) ; the shaking
period was 30 min. The extraction equilibrium was achieved during this period. All
samples were centrifuged after extraction. The pertechnetate and perrhenate
concentration in both phases was determined radiometrically using p-emission (99Tc0 4',
188Re0 4' ; liquid scintillation counter LS 6000 LL'Beckman). The aqueous solution was

adjusted using 0.05 mol-drrr' NaOAc/HCI (pH 2.0-5.2), 2-[N-morpholino]ethanesulfonic

acid (MES)/NaOH, 5.4--6. I), N-[2-hydroxyethyl]piperazine-N '-[2-ethanesulfonic acid]
(HEPES)/NaOH, 7.1-8 .0), and 3-([2-hydroxy-I,I-bis(hydroxymethyl)ethyl]-amino)-1-
propanesulfonic acid , TAPS)/NaOH, 8.0-9.0).
In order to determine the partition coefficients of azacages in the waterlI-octanol
system , 0.00 I M stock solution s of aminocages in buffer, saturated with l-octanol, were
prepared . In the case of amidocage 6, 0.00 I M stock solution in l-octanol, saturated with
buffer , was prepared. Partition experiments were performed with 0.0001 M solution of
azacage in aqueous solution (HEPES/NaO H, pH = 7.4; 2-amino-2-methylpropanol/HCI
buffer , pH = 10.6), and 0.0001 M amidocage in l-octanol, respectively. The phase ratio
V(org):V(W) was I: I (0.8 ern' each ); the shaking period was 2 h. After separation of both
phases by centrifugation, the concentration of azacages in the aqueous and organic phase
were analyzed by UV-spectroscopy (Lambda 2, Perkin-Elmer).

4.3. X-ray Crystallography

The X-ray data were collected at room temperature (293 K) on a SMART-CCD

diffractometer (SIEMENS), using graphite-monochromatized Mo-K, radiation (A. =
0.71073 A). The structures were solved by direct methods using SHELXS-90 and refined
with SHELXL-97. 52 An empirical absorption correction (IJI-scan) was applied. The
anisotropic refinement of all non-hydrogen atoms was only possible for the perrhenate
complex of azacage 2. Because of the relatively poor quality obtained for the perrhenate
complexes of azacages 1 and 3, only the heavy atoms could refined anisotropically. This
explains the relative high R-values of these complexes. Therefore also some restraints for
bondlength and angles are applied . The positions of hydrogen atoms were calculated
corresponding to their geometrical conditions and refined using the riding model. Atomic
positional and thermal parameters, full lists of bond lengths and angles , and FJF c values
have been deposited as supporting information at the CCDC. Crystallographic data
(excluding structure factors ) for the structure reported in this paper have been deposited
with the Cambridge Crystallographic Data Centre as supplementary publication no.
CCDC 177808, 177809 and 177810. Copies of the data can be obtained free of charge on
application to CCD C, 12 Union Road, Cambridge CB2 I EZ, UK (Fax: +44- 1223/336-
033; E-mail : depos it@ccdc.

[Hie H20)] . Re04: C33~gNs08Re, FW = 828.98, a = 10.535(10), b = 12.661(11),

c = 13.870(12) A, a = 74.19(2)°, ~ = 75.19(2)°, Y= 76.11(3)°, V = 1692(3) A3, triclinic
P-I, Z = 2, ~ = 3.649 mm'', 1381 reflections collected, 1369 unique reflections, Rmt =
0.0688, RI = 0.1478, wR2 = 0.3050 [I> 20(1)], RI = 0.1873, wR2 = 0.3228 (all data),
GOF = 1.132.
[H22 c (H20hJ . (Re04h . H20: C33Hs3Ns014Re2, FW = 1116.22, a = 11.583(4), b =
28.744(9), c = 12.337(4) A, ~ = 92.003(6)°, V= 4105(2) A3, monoclinic P21/n, Z= 4, ~ =
5.960 mm", 12553 reflections collected, 3815 unique reflections, Rint = 0.1008, RI =
0.0427, wR2 = 0.0981 [/> 20(1)], RI = 0.0726, wR2= 0.1076(all data), GOF = 0.789.
{[H23] . (Re04h . (H20hh: C36~2N801IRe2, FW = 1155.35, a = 15.05(3), b =
17.42(3), C = 16.97(3) A, ~ = 93.77(4)°, V = 4441(14) A3, monoclinic P2/n, Z = 4, ~ =
5.509 mm", 13586 reflections collected, 7941 unique reflections, Rm l = 0.1139, RI
0.0979, wR2 = 0.2220 [I> 20(1)], RI = 0.1398,wR2 = 0.2388 (all data), GOF = 1.102.


We thank the Sachsisches Ministerium fUr Wissenschaft und Kunst (project no.
7531-50-03-0370-01/4), the Deutsche Forschungsgemeinschaft, Germany, and the
Department of Science and Technology, India, for support of this work. The authors are
indebted to Ms. 1. Schubert and Ms. U. Stockgen for their careful experimental
assistance, Dr. T. Rambusch for molecularmodeling calculations, and Prof. 1. Nelson for
stimulating discussions.


1. Comprehensive Supramolecular Chemistry, edited by 1. L. Atwood , 1. E. D. Davies, D. D. MacNicol, F.

Vogtle, and 1.-M. Lehn (Pergamon , Oxford , 1996).
2. Supramolecular Chemistry of Anions, edited by A. Bianchi , K. Bowman-James, and E. Garcia-Espana
(WILEY-VCH, Weinheim, 1997)
3. B. Dietrich , Design of Anion Receptors : Applications, PureAppl. Chern. 65, 1457-64 (1993) .
4. C Seel, A. Galan , and 1. DeMendoza , Molecular Recognition of Organic Acids and Anions - Receptor
Models for Carboxylates, Amino Acids, and Nucleotides, Top. CurroChem. 175, 101-132 (1995) .
5. F. P. Schmidtchen and M. Berger, Artificial Organic Host Molecules for Anions, Chem. Rev. 97, 1609-
46 (1997) .
6. M.M.G. Antonisse and D. N. Reinhoudt , Neutral Anion Receptors : Design and Application, Chem.
Commun. 443-448 (1998) .
7. P. D. Beer and P. A. Gale, Anion Recognition and Sensing : The State of the Art and Future Perspectives,
Angew. Chem. Int. Ed. Engl. 40,487-516 (2001) .
8. K. Gloe, H. Stephan , and M. Grotjahn, Quo Vadis Anion Extraction , Chem-Ing-Tech. 74, 767-777
(2002) .
9. P. V. Bonnesen , B. A. Moyer, D. 1. Presley, V. S. Armstrong, T 1. Haverlock , R. M. Counce , and R. A.
Sachleben , Alkaline-side Extraction of Technetium from Tank Waste using Crown Ethers and other
Extractants , Oak RidgeNationalLaboratory (ORNL)-Report TM 13241 (1996) .
10. R. A. Leonard , C Conner , M .W. Liberatore , P. V. Bonnesen , D. J. Presley, B. A. Moyer, and G. 1.
Lumetta , Developing and Testing an Alkaline-side Solvent Extraction Process for Technetium
Separat ion from Tank Waste, Sep. Sci. Technol. 34, 1043-1068 (1999) .
11. K. L. Nash, R. E. Barrans, R. Chiarizia, M. L. Dietz, M. P. Jensen , P. G. Rickert, B. A. Moyer, P. V.
Bonnesen , J. C Bryan, and R. A. Sachleben, Fundamental Investigations of Separations Science for
Radioactive Materials, Solvent Extr. Ion Exch. 18, 605-{)31 (2000) .
12. J A. Gawenis, J F. Kauffman, and S. S Jurisson, Ion Pairing as a Strategy for Extraction by Modified
Supercritical Carbon Dioxide: Extraction of Radioactive Metal Ions, Anal. Chem. 73,2022-2026 (2001)
13. H. Stephan, R. Berger, H. Spies, B. Johannsen, and F. P. Schmidtchen, Efficient Phase Transfer of
Pertechnetate with Bicyclic Guanidinium Compounds, J. Radioanal. Nucl. Chem. 242, 399-403 (1999).
14. H. Stephan, H. Spies, B. Johannsen, E. Nicoletti, and F. P. Schmidtchen, ITC Analysis of Binding
Perrhenateusing Monopyridinium-a-Cyclodextrin, Annual Report Forschungszentrum Rossendorf FZR-
312,54-56 (2000)
IS. K. Schwochau, Technetium: Chemistry and Radiopharmaceutical Applications (W1LEY-VCH,
Weinheirn. 1997)
16. B. Johannsen. and H. Spies, Technetium(V) Chemistry as Relevant to Nuclear Medicine, Top. Curro
Chem. 176,77-121 (1996)
17. D. Parker. Imaging and Targeting, in Comprehensive Supramolecular Chemistry. vol. 10
"Supramolecular Technology. " edited by J. L. Atwood,J E. D. Davies, D. D. MacNicol, F. Vogtle, and
J.-M. Lehn(Pergamon, Oxford, 1996), pp. 487-536.
18. D. E. Reichert, J S Lewis. and C J Anderson, Metal Complexes as Diagnostic Tools, Coord. Chem.
19. S. S. Jurisson and J D. Lydon, Potential Technetium Small Molecule Radiopharrnaceuticals, Chem. Rev.
20. M. J Heeg and S S. Jurisson, The Role of Inorganic Chemistry in the Development of Radiometal
Agentsfor CancerTherapy, Acc. Chem. Res. 32, 1053-1060 (1999)
21. P. J Blower and S Prakash, The Chemistry of Rhenium in Nuclear Medicine, in Perspectives on
Bioinorganic Chemistry. vol. 4 (JAI Press, London, 1999), pp. 91-143.
22. W. A Volkertand T J. Hoffman, Therapeutic Radiopharrnaceuticals, Chem. Rev. 99,2269-2292 (1999).
23. Y. Marcus, Ion Properties (Marcel Dekker, New York, 1997).
24. B. A. Moyer and P V. Bonnesen, Physical Factors in Anion Separations. in Supramolecular Chemistry
of Anions, edited by A. Bianchi, K. Bowman-James, E. Garcia-Espana (W1LEY-VCH, Weinheim,
1997), pp. 1-44.
25. T Nakashima and K. H. Lieser, ProtonAssociation of Pertechnetate, Perrhenate and PerchlorateAnions,
Radiochim. Acta 38, 203- 206 (1985)
26. Handbook of Chemistry and Physics, edited by D. R. Lide and H. P. R. Frederikse (CRC Press, Boca
Raton, 1997)
27. J Nelson, V. McKee, and G. Morgan, Coordination Chemistry of Azacryptands, in: Progress in
Inorganic Chemistry. vol. 47. edited by K. D. Karlin(Wiley, New York, 1998), pp. 167-316.
28. S Mason, T Clifford, L. Seib, K. Kuczera, and K. Bowman-James, Unusual Encapsulation of Two
Nitratesin a Single Bicyclic Cage. J. Am. Chem. Soc. 120,8899-8900 (1998).
29. G. Morgan. V McKee, and J Nelson, Caged Anions: Perchlorate and Perfluoroanion Cryptates, Chem.
Commun. 1649-52 (1995).
30. M. J. Hynes, B. Maubert, V. McKee, R. M. Town, and J Nelson, Protonated Azacryptate Hosts for
Nitrateand Perchlorate. J. Chem. Soc.. Dalton Trans. 2853-2859 (2000).
31. B. M. Maubert, J. Nelson, V. McKee, R. M. Town, and 1. Pal, Selectivity for Dinegative versus
Mononegative Oxoanionic Guests within a Cryptand Host, J. Chem. Soc.. Dalton Trans. 1395-97
32. M. Arthurs, V McKee. J Nelson, and R. M. Town, Chemistry in Cages: Dinucleating Azacryptand
Hostsand their Cation and AnionCryptates, 1. Chern. Ed. 78, 1269-1272 (200I)
33. H. Stephan. H. Spies. and F. P. Schmidtchen, unpublished results.
34 K. M. Rohal. D. M. Van Seggen, J. F. Clark, M. K. McClure, C. K. Chambliss, S H. Strauss, and N. C
Schroeder, Solvent Extraction of Pertechnetate and Perrhenate Ions from Nitrate-rich Acidic and
AlkalineAqueousSolution. Solvent Extr. Ion Exch. 14,401-416 (1996)
35. H. Stephan. H Spies. B. Johannsen, L. Klein, and F. Vogtle, Lipophil ic Urea-funclionalized Dendrimers
as EfficientCarriers for Oxoanions,Chern. Commun. 1875-1876,(1999)
36 H. Stephan, H. Spies, B. Johannsen. K. Gloe, M. Gorka. F. Vogtle, Synthesis and Host-Guest Properties
of Multi-Crown Dendrimers towards Sodium Pertechnetate and Mecury(lI) Chloride, Eur. J. Inorg.
Chern 2957-2963 (2001)
37. 1. L. Atwood. K. T Holman. and J W. Steed, Laying Traps for Elusive Prey: Recent Advances in the
Non-covalent Binding of Anions, Chem. Commun. 1401-{)7 (1996).
38. DFT calculation using ADF version 23 (Scientific Computing & Modeling, Chemistry Department,
Vrije University. De Boelelaan 10, 1081 HV Amsterdam, The Netherlands) on SGI CRAY Origin2000.
39. C Bazzicalupi, P Bandyopadhyay, A. Bencini, C Giorgi, B. Valtancoli, D. Bharadwaj, P. K.
Bharadwaj. and R. J Butcher, Complexation Properties of Heteroditopic Cryptandstowards Cul +, Znl +,

Cd2+. and Pb2+ in Aqueous Solution: Crystal Structures of (H sL')(CIO,)s]-4H 20 and

(NiL 2Cl)CI)·5.5H20·CH)OH, Eur. J. Inorg. Chem. 2111- 16 (2000).
40. P. Ghosh, S. S. Gurta, and P. K. Bharadwa], Complexation Properties of a Heteroditopic Cryptand
towardsCu" and Nil, Crystal Structures of the Cryptand and its Nickel(H) CascadeComplex, J. Chem.
Soc.. Dalton Trans. 935-938 (1997).
41. D. K. Chand, K. G. Ragunathan, T. C. W. Mak, and P. K. Bharadwaj, Tetrahedral Recognition of a
Water Molecule by Heteroditopic Cryptands: X-ray Structural Studies, J. Org. Chem. 61, 1169-71
42. D. K. Chand, and P. K. Bharadwaj, Heteroditopic Cryptands of Tunable Cavity Size: Imposition of
Distorted Geometry onto Copper(H) and Nickel (H) and Molecular Recognition of Water Molecules,
lnorg. Chem. 37,5050-55 (1998).
43. F. Amaud-Neu, S. Fuangswasdi, B. Maubert, J. Nelson, and V. McKee, Binding Properties of
Octaaminocryptands,lnarg. Chem. 39, 573-579 (2000).
44. A. Bencini, A. Bianchi, E. Garcia-Espana, M. Micheloni, 1. A. Ramirez, Proton Coordination by
Polyamine Compounds in Aqueous Solution, Coord Chem. Rev. 188,97-156 (1999).
45. M. Moder, K. Wichmann, K. Gloe, and F. Vogtle, Study on Formation and Stabilityof Azacage Metal
Complexes usingElectrospray MassSpectrometry, Int. J. MassSpectr. 210/211,327-339 (2001).
46. The hexamethylated bis-tren cage with m-xylyl spacers has a significant higher Iipophilicity as the
structure related m-pyridine bridged compound 4 (72% towards 0% in octanol); the resulting
extractabilities for bothcagesusingthe experimental conditions ofFigJ are 48% and 14%,respectively:
D. Farrell,K. Gloe, K. G1oe, G. Goretzki, V. McKee, 1. Nelson, I. Pal, H. Stephan, R. M. Town, and K.
Wichmann, Towards Promising Oxoanions extractants: Azacages and Open-chain Counterparts, J.
Chem. Soc.. Dalton Trans. to be published in 2003.
47. 1. C. Bryan, cis-syn-cis-Dicyclohexano-I8-crown-6 Sodium Perrhenate, Acla CrySI. C54, 1569-157\
48. 1. C. Bryan, and R. Sachleben, Synthesisof a New Dibenzo-14-Crown-4 LariatEtherand Structureof its
SodiumPerrhenate Complex, J. Chem. CrySI. 29, 1255-1259(1999).
49. 1. Nelson, M. Nieuwenhuyzen, I. Pal, and R. M. Town, Dual-Mode Recognition of Oxalate by
Protonated Azacryptate Hosts; Conformational Response of the Guest Maximizes p-Stacking
Interactions, Chem. Cammun. in press.
50. M. A. Hossain, 1. M.L1inares, S. Mason, P. Morehouse, D. Powell,and K. Bowman-James, Parallels in
Cation and Anion Coordination: A New Class of Cascade Complexes, Angew. Chem. 114,2441 -2444
51. D. K. Chand and P. K. Bharadwaj, Synthesis of a Heteroditopic Cryptand Capable of Imposing a
Distorted Coordination Geometry onto Cu(H): Crystal Structures of the Cryptand (L),
[Cu(L)(CN»)(picrate), and (Cu(L)(NCS»)(picrate) and Spectroscopic Studies of the Cu(H) Complexes,
lnorg. Chem. 35, 3380-3387(1996).
52. G. M. Sheldrick , SHELXL-97, Universitat Gcttingen(1997).

Spiro D. Alexandratos


The design and development of methods for the removal of toxic ions from water in the
environment, industrial process streams , and waste storage facilities cont inues to be a high-
priority problem, Such methods must be cost-effective and environmentally compatible.
Polymer-supported reagents prepared by the immobilization of ion-selective ligands onto
crosslinked polymer beads offer an important avenue for achiev ing these objectives.'
Additionally, such polymers have a broader applicability into sensor technology and
chromatographic separations.
Situations centering upon the removal of ions from aqueous solutions, including cases of
environmental remediation, are documented in the literature . The example of the Hanford
reservation in Richland, which stores fifty-four million gallons of radioactive waste, has been
often analyzed.' Bioassays suggest that the groundwater around Hanford is being
contaminated by tanks that are leaking waste.' Additional example s include findings that:
leachate from areas at electric utilities plants where coal is stored can contaminate
groundwater with heavy metal ions, including lead:" acid mine drainage can release
significant levels of heavy metal ions into surrounding streams, adversely affecting life within
those streams;' mining wastes can contaminate surface water with large amounts of zinc,
copper, lead, nickel, and cadmium :" and groundwater contam ination with zinc, copper, and
arsenic by the wood preservation industry is possible.' The perchlorate anion has been found
in the groundwater of the southwestern United States as well as the Colorado River ,S Its
toxicity is associated with its adverse effect on thyroidal iodide uptake ." A similar
environmental problem is found with the presence of the pertechnetate anion in
groundwater. 10

• HunterCollege of the City University of New York, Department of Chemistry. 695 Park Avenue. New York.
NY 10021.

Fundamer&tals andApplications of Anum Separations, edited by

Moyer and Singh, KluwerAcademic/Plenum Publishers, New York, 2004 169


In separations accomplished by solvent extraction, II the aqueous phase is contacted with

an organic solvent in which is dissolved a complexant known to be selective for a targeted
metal ion. Complexation occurs at the aqueous/organic interface thus requiring vigorous
mixing of the two layers to provide sufficient contact and allow the rapid attainment of
equilibrium. The advantages of solvent extraction include high throughput and the numerous
complex ants available for many different separations.
Research continues in the preparation of highly selective complex ants for specific
applications; the development of a crown-calixarene for cesium separations from nuclear
wastes is one important recent example.f Phosphorus-based complexants are useful in
solvent-extraction chemistry because of their selectivity for transition metal, lanthanide, and
actinide ions under a range of conditions. That selectivity can be tuned depending upon the
precise structure of the organophosphorus ester, acid, or oxide. Interestingly, certain
solutions containing both acidic and neutral coordinating complexants show synergism,
wherein metal ion uptake is greater than expected based on the performance of each
complex ant alone. One example of such a solution is that comprised of di(2-
ethylhexyl)phosphoric acid and tributylphosphate. This combination is especially useful in
the extraction of lanthanides and actinides .
Soluble complex ants can bind anions through acid-base and/or coulombic interact ions.
Ligands in such complexants that are not pre-organized, as with linear polyamines and
guanidines, may have high ionic affinities but minimal selectivity. Complex ants with pre-
organized ligands (i.e., having two or more ligands in fixed positions that allow for
cooperation in binding of an ion) may have high affinities with significant selectivities.P
Azamacrocycles and bicyclic cryptates are two important examples of macrocyclic hosts
wherein protonation of their amine sites allows them to encapsulate anions by coulombic
interactions, giving high affinities and selectivities." Quaternization of amines, either by
protonation or alkylation, allows for the facile preparation of hosts with high ionic affinities.
Coulombic interactions are also enhanced when a metal cation is ligated to the amine, and the
metal then becomes the site to which the anion is bound, forming a cascade complex.f
A significant disadvantage to the use of soluble complex ants for separations by solvent
extraction is complexant and solvent loss through dissolution and entrainment (i.e.,
incomplete phase separation of finely dispersed droplets of solvent in the aqueous phase) .
This disadvantage, in fact, represents the limiting feature determining the economic
competitiveness of the technique, especially as feed solutions become dilute « 100 ppm). In
addition , the environmental consequences due to these solvent losses and diluent volatility
may be unacceptable.


A significant amount of research has been directed at the preparation of crosslinked

polymer beads with ligands expected to be selective towards targeted cations and anions.
The desired selectivity should approach that found with complexants such as crown ethers for
the various alkali metal ions at rates found under homogeneous conditions. The polymers
most often used are polystyrene, poly(vinylbenzyl chloride) and poly(glycidyl methacrylate).
Polymer-Supported Reagents for Anionic Recognition 171

They may be crosslinked with divinylbenzene (DVB) and prepared by suspension

polymerization to give beads appropriate for use in columns." The beads can be microporous
(i.e., gel) or macroporous particles. I? The polymers have good physical stability and can
undergo a wide range of reactions to yield different covalently bound ligands (Figure 1).

6 --~


(O-c~CI - L

--~ -~

Figure 1. Polymerization of styrene. vinylbenzyl chloride. and glycidyl methacrylate and their subsequent
modification withligand L.

Ion-exchange resins with sulfonic and carboxylic acid ligands are widely available and
can be used to remove metal cations from aqueous solutions. While often displaying rapid
rates of complexation, they both are insufficiently selective, the former becoming quickly
saturated with benign ions such as sodium, potassium, and calcium always present in
environmental water ." More selective ligands have thus been immobilized. Pyrazole,
imidazole, and triazole moieties, immobilized on poly(glycidyl methacrylate), selectively
complex Cu(II) from a solution also containing Cd(II), Co(lI) , Ni(II), and Zn(II) at pH > 2.5
(Figure 2).19 Poly(methacrylohydroxamic acid) has an affinity sequence of Cu(II) > Zn(II) >
Ni(II) > Co(lI) > Pb(II) > Fe(II) within an optimal pH range of 3.5-5 (Figure 3).20 A
macrocyclic Schiff base (Figure 4) was found to have a high selectivity for copper ions."

Figure 2. Pyrazole, imidazole. and triazole substituted on poly(glycidyl methacrylate).

Figure 3. Poly(methacrylohydroxamic acid).

Q HN a
C-N Ar
H2 H

Figure 4. Polystyrene-immobilizedmacrocyclic Schiffbase.
Polymer-Supported Reagents for Anion Recognition 173

Additional examples include reaction of glycidyl methacrylate with iminodiacetic acid

and polymerization of the resulting monomer gave beads that had high capacities for both
Cr(Ill) and Cu(ll) in acidic solutions ;22 calix[4]arene carboxylate was immobilized onto
polyallylamine and found to have a selectivity sequence of Pb » Cu » Zn;23 and an
azathiacrown ether was bonded to crosslinked poly(glycidyl methacrylate)[polyOMA] and
found to be selective for silver ions in the presence of copper, zinc, and cadmium.i'
Positively charged ligands are most useful for anion complexation reactions .
Quaternized amines and immobilized metal ions have been widely studied for their selectivity
towards anions. Piperazine resins (Figure 5) have good selectivity for precious metals in
their anionic form, being able to separate Au(III) and Pd(ll) from 0.1 M HCl solution also
containing copper, nickel, and iron.25 PolyOMA has been found to be an effective support
for polyamines, including ethylenediamine (EDA, Figure 6), diethylenetriamine (DETA),
triethylenetetramine (TETA), 3,3'-iminobispropylamine (IBPA), and N,N'-bis(3-
aminopropyl)ethylenediamine (BAPED) . All were contacted with an aqueous cyanide
solution containing 50 ppm KAu(CNh.26 At pH 9.2-9.8, the percent Au(CNh- complexed
from the resins as listed above was 100%,96%,93%,97%, and 80%, respectively. All have
a higher affinity for the gold complex than Cu(CNh", Co(CN)l, and Ni(CN)/".
Poly(vinylpyridine) has been quaternized with substituted alkenes to give resins capable of
anion exchange (Figure 7).27 Contact with 0.01 M Fe2(S04h in a solution at pH 2 gave
quantitative sorption of Fe3(S04h(OHk after a seven day contact time.

Figure 5. Polystyrene-immobilized piperazine.

Figure 6. Ethylenediamine immobilized on poly(glycidyl methacrylate).

+ /'R

R =CN, C(O)NH 2 , C(O)CH 3

Figure 7. Poly(vinylpyridine) quaternized withalkenes.


Ligands with both sulfur and nitrogen atoms have significant affinities for anions.
Thiazoles immobilized on polystyrene(Figure8) havebeen contactedwithAuC1 4' , ptCll,
PdCll, IrCll, and base metals in aqueous HC1.28 Distribution coefficients show a high
selectivityfor gold over the base metals(Fe(III), Cu(II), Co(ll), and Ni(II)). The resin also
showsselectivityfor Pt, Pd, and Ir over the base metals,but the selectivityfor theseplatinum
group metals is not as high as that for gold.

Figure 8. Polystyrene-supported thiazole.

Chitin (poly(N-acetyl-D-glucosamine), Figure 9) can be deacetylated to give chitosan

(poly(D-glucosamine)), and the latter has been studiedfor itsmetal ionaffinity. Chitosan can
be crosslinked with glutaraldehyde, and it is found to have a high uptake capacity for
molybdateand vanadate ions,"
A class of polymers, termed Polymer Ligand Exchangers, has an ion such as copper
chelated to an immobilized ligandand which,in tum, has the abilityto bind anions.l" Such
polymershave a high affinityfor phosphateand chromate ions.


~~~I_-"'~:~O O:~
( v HO NH (V
Figure 9. Chitin.

In theseand mostother examplesin the literature,a singletypeof ligandis immobilized

on a polymersupport. In somecases, a reagentwithhighselectivityis produced but with low
rates of complexation; in other cases, the reagent has good selectivity but not sufficiently
high for environmental or industrial applications. Bifunctional polymers offer the
opportunityof combininghighselectivitywithrapid complexation kinetics. The importance
of bifunctionality was first studied with phosphorus acid resins, then extended to ligands
capable of anion exchange (vide infra).
Polymer-Supported Reagents for Anionic Recognition 175

Phosphorus acid ligands (phosphoric," phosphonic.f and phosphinic'" have been

immobilized on polymers. Each displays selectivities much greater than sulfonic acid
resins." Most importantly, research has found that bifunctional polymers bearing ion
exchange (phosphonic acid) and neutralcoordinating (phosphonate diester) ligands(Figure
10)complexed greater levelsof metal ions than the analogous monofunctional resings.35 For
example. as quantified by the distribution coefficient (D) at equilibrium (mmol Mn+ per
g,esi"lM n+ per mLsolulion), the value for Ag(l) from 2 M HN0 3 is 2590 for the bifunctional
resin, 680 for the monofunctional phosphonic acid resin, and 370 for the monofunctional
diester resin.

2 I


Figure 10. Bifunctional ion-exchange/coordination polymer.

The concept of bifunctional polymers has received continued emphasis inour laboratory.
In all cases, one ligand nonselectively allowsall ions rapid access into the matrix, while the
second ligand complexes or otherwise retains the targeted ion within the matrix. This
approach has resulted in the commercialization of two new ion-selectivepolymer-supported
reagents: Diphonix® (prepared in collaboration with the Separations Group at Argonne
National Laboratory) is especially selective for actinides and Fe(III) from highly acidic
solutions, while BiQuat™ (prepared in collaboration with the Chemical Separations Groupat
the Oak Ridge National Laboratory) is selective for pertechnetate and perchlorate anions
from groundwater(Figure II). Diphonixbears the nonselectivesulfonic acid group and the
selective diphosphonic acid ligand. Its performance is indicated by comparing the
distribution coefficients for the complexationof Am(III) by three resins from a solution of
0.10 M HNOi4 M NaN0 3 : D(Diphonix)= 5600; D(phosphonicacid) = 24; D(sulfonic acid)
= 49. BiQuat extends the concept of bifunctional polymers for selective anion separations
and is comprised of a nonselective triethylammonium group and a selective
trihexylammonium ligand. Resins were contacted for 24 hours with a groundwater test
solution that is 0.06 M each in NaCI, NaN0 3, and Na2S04, and 6 JLM in Tc04'; distribution
coefficients for the monofunctional triethylammonium resin, the trihexylammonium resin,

and BiQuat were: 16200.1540. and 37300. respectively. In comparison, the commercially
available resins Amberlite IRA-900 and Purolite A-520E have distribution coefficients of
2460 and 12800. respectively . The combination of ligands clearly results in far higher
distribution coefficients than either ligand gives alone. Reasons for this have been proposed
and center on the balance between the matrix hydrophilicity and the electrostatic properties
of the binding site.
The monofunctional and bifunctional polymers published to date have the ligands
randomly arranged within the matrix. The ligands can be expected to interact with each other
and with the complexed metal ions based on studies of site-site interactions in crosslinked
polyrners'" and evidence provided by solid-state 31p nmr of bifunctional polymers." Given
the importance of bifunctionality in enhancing ionic affinities, it is the objective of current
research to achieve greater control of where the ligands are placed in relation to each other.

Figure 11. Diphonix'" and BiQuat™ resins.


The immobilization of anion-selective ligands onto polymer support will continue to be

studied for applications in environmental remediation, chromatographic separations. and
sensor technology. A wide variety of amines, both cyclic and acyclic , are of great interest.
Thiazoles and phosphorus-based ligands also give promising results . Bifunctional polymers
offer important complementary results to those from monofuntional polymers .


We are grateful to the Department of Energy , Office of Basic Energy Sciences, for
funding of our research in polymer-supported reagents through grant DE-FG02-02ERI5287.
Polymer-Supported Reagents for Anionic Recognition 177


I. S. D. Smith and S. D. Alexandratos, Solvent Extr. Ion Exch. 18,779 (2000).

2. Proceedings ofthe First Hanford Separation Science Workshop, PNL-SA-21775, Pacific Northwest National
Laboratory, Richland, WA (1993); Science and Technologyfor Disposal ofRadioact ive Tank Wastes, W. W.
Schultz and N. J. Lombardo, eds.; Plenum Press, NY (1998).
3. D. A. Delistraty and J. Yokel, Environ. Toxicology 14, 473 (1999).
4. A. M. Cook and S. J. Fritz, Water, Air, Soil Pollut. 135,371 (2002).
5. D. M. DeNicola and M.G. Stapleton, Environmental Pol/ution (Oxford, UK), 119, 303 (2002).
6. J.-E. Yang, H.-J. Kim, and S.-H. Jun. Korean J. Soil set; Fert. ,34,33 (2001).
7. M. F. Azizian, P. O. Nelson, and P. Thayumanavan, Abstracts of Papers, 223rd ACS National Meeting,
Orlando, FL, ApriI7-1!, 2002, ENVR-157.
8. W. E. Motzer, Environmental Forensics, 2,301 (2001).
9. M. A. Greer, G. Goodman, R. C. Pleus, andS. E. Greer, Environmental Health Perspectives, 110,927 (2002).
10. G. D. DelCui, W. D. Bostick, D. R. Trotter, and P. E. Osborne, Sep. Sci. Technol., 28, 551 (1993).
II. T. Sekine, and Y. Hasegawa, Solvent Extraction Chemistry: Fundamentals and Applications, Dekker: New
York, 1977.
12. P V. Bonnesen, T.J. Haverlock, N.L. Engle, R.A. Sachleben, and B. A. Moyer, Am. Chem. Soc. Symp. Ser.,
757(Calixarene.I'.Iilr Separations), 26 (2000).
13. F. P. Schmidtchen and M. Berger, Chem. Rev., 97, 1609,(1997).
14. M. Mitewa and P.R. Bontchev, Coord. Chem. Rev. 135, 129 (1994).
15. V. Amendola, L. Fabbrizzi, C. Mangano, P. Pallavicini, A. Poggi, and A. Taglietti, Coord. Chem. Rev. 219-
221,821 (2001).
16. W. L. Sedereland G. L. De Jong, J. Appl. Polym. Sci. , 17, 2835 (1973).
17. H. Jacobelli, M. Bartholin, and A. Guyot, J. Appl. Polym. Sci. , 23, 927 (1979).
18. G. E. Boyd, F. Vaslow, and S. J. Lindenbaum, J. Phys. Chem., 71, 2214 (1967).
19. P. M. Van Berkel, W. L. Driessen, J. Reedijk, D. C. Sherrington, and A. Zitsmanis, React. Funct. Polym. 27, 15
20. Y. K. Agrawal and K. V. Rao, React. Funct. Polym., 31, 225 (1996).
21. T. Matsushita, N. Kubota, M. Fujiwara, and T. Shono, Chem. Lett. 657 (1984).
22. C-C , Wang, c.- Y. Chen, and c.- Y. Chang, J. Appl. Polym. sa.. 84, 1353 (2002).
23. K. Ohto, Y. Tanaka, M. Yano,T. Shinohara, E. Murakami, and K. Inoue, Solvent Extr. Ion Exch., 19,725
24. L. G. A. vande Water, F. ten Hoonte, W. L. Driessen, J. Reedijk, and D. C. Sherrington, lnorg. Chim. Acta.
303, 77 (2000).
25. C. Yi-Yongand Y. Xing-Zhong, React. Polym. 23, 165(1994).
26. P. Haineyand D. C. Sherrington, React. Funct. Polym., 43, 195(2000).
27. V. Gutanu, C. Luca, C. Turta, V. Neagu, V. Sofranschi, M. Cherdivarenco, and B. C. Simionescu, J. Appl.
Poly. Sci., 59,1371 (1996).
28. B. K. Leungand M. J. Hudson, Solvent Extr. Ion Exch., 10, 173(1992).
29. E. Guibal, C. Milot. and J. M. Tobin, Ind. Eng. Chern. Res., 37,1454 (1998).
30. A. Senguptaand D. Zhao, U.S. Patent6.136,199.
31. J. Kennedy, E. S. Lane. and B. K. Robinson, J. Appl. Chelll.. 8,459 (1958).
32. H. Egawa, K. Yamabe, and A. Jyo, J. Appl. Polym. Sci., 52,1153 (1994).
33. S. D. Alexandratos, M. A. Stand. D. R. Quillen, and A. J. Walder, Macromolecules, 18, 829 (1985).
34. S. D. Alexandratos, D. W. Crick, and D. R. Quillen, Ind. EIlR. Chem. Res., 29, 772 (1991).
35. S. D. Alexandratos, D. R. Quillen, and M. E. Bates. Macromolecules, 20,1191 (1987).
36. M. Stem, M. Fridkin, and A. Warshawsky, J. Polym. Sci. Polym. Chem. Ed. 1982,20, 1469; L. T. Scott, J.
Rebek, L. Ovsyanko, and C. L. Sims, J. Alii. Chem. SOl'., 99, 625 (1977).
37. D. W. Crick and S. D. Alexandratos. Magn. Reson. Chern., 32 (Spec. Issue),S40 (1994).

Grant C. Lukey and Jannie S. 1. van Deventer


For the greater part of the 20 th century. the hydrometallurgical pathwa y used for the
treatment of gold ores co nsisted of leach ing the ore in cyanide solut ion followed by a
co mplex solid-liquid separatio n stage. which separated the solid residues from the leach
liquor. The leach liquor was then treated by zinc cementation to recover the gold.
However, this process resulted in a poor gold recovery if the ore treated conta ined a high
co ntent of clay or sulphide mineral s. Consequently. new hydro metallurgi cal processes
have been de veloped in the last 20 years for the recovery of gold from these complex ore
types. The implementation of the carbon-in-pulp (CIP) and carbon-in-Ieach (CIL)
processes in the gold industry has effectively replaced the filtration and countercurrent
dec antation stages in most plants, leading to a significant economic gain. Publications by
Fleming (1992 ) and La Brooy et al. (1994 ) provide extens ive general information on the
hydro metallurg y of gold processing.
In recent years there has been an increasing need to process highly refractory and re-
adsorbing ores of a very low gold grade (less than 1 g Ault) compared to that of the base
metals such as copper (ca. 200 g Cult) . The poor performance of the CIP and CIL
proce sses for the treatment of these ore types has led to the development of the resin-in-
pulp process (RIP ). Th is process uses ion-exchange resins in direct contact with the pulp
to recover valuable metal cya nides.

Gram C Lukey and Janme S. J. van Deventer, Department of Chemical Engineering. The University of
Melbourne. VIctoria 301O. Australia. Tel.: +61-3-8344-6620. Fax: +61-3-8344-4153. E-mail:
l anll1 ~ lc, lI " , mclh .cll~.i!.~ .

Fundamentals andApplications of AnionSeparations, edited by

Moyerand Singh,K1uwer AcademicIPleoum Publishers, NewYork,2004 179

The literature on the recovery of gold using ion-exchange resins may be classified
into several distinct types of work. First the sorption properties of existing commercial
resins for gold and other metal-c yanide complexes have been determined and described
In many publications. ThIS work established that anion-exchange resins were capable of
recovering metal-cyanide complexes. Consequently, the development of methods for the
elution and regeneration of anion-exchange resins was undertaken . The majority of this
work has been reviewed previously by Fleming (1991), Rivero s and Cooper (1987) and
Hosking (1984) .
Although the use of resins for the recovery of gold is commonplace in the former
Soviet Union (Bolinski and Shirley, 1996), the first commercial application in the
Western world of the use of ion-exchange resins for the extraction of gold from leachate
occurred in the late 1980s at the Golden Jubilee mine in South Africa . This experience
established that commercially available ion-exchange resins were nonselective for the
gold-cyanide species, and consequently a high proportion of the resin was loaded with
undesired metals with a high affinity for the resin, namely copper cyanide, zinc cyanide
and iron cyanide. It was also found that cobalt cyanide had the potential to poison the
resin under certain conditions.
The use of ion-exchange resins for the selective recovery of gold from leach streams
is becoming increasingly important, not only because of the need to treat ores of a lower
grade and of a more complex mineralogy, but also because of the more stringent
environmental requirements being placed on mining operations. A series of investigations
has been conducted to study the effect of various resin characteristics on the sorption of
metal-cyanide species, in an attempt to understand which characteristics predominantly
determined the selectivity of the resin . In addition, various institutions have synthesized
new ion-exchange resins claimed to be selective for gold cyanide. These resins have been
synthesized especially for the recovery of gold and include those developed by Mintek in
South Africa as well as the Henkel Corporation. Despite this work, little has been written
about the successful determination of the exact structure of the resin or the effect that the
functional group structure may have on the speciation of the sorbed metal complex. Such
studies are needed before the particular sorption properties or characteristics of a resin
can be properly understood .
The speciation of copper cyanide in aqueous solution has been the subject of many
investigations. However, the effect of ionic strength on the distribution of copper-cyanide
species in solution has not been established. Furthermore, the effect of highly saline
solutions on the capacity and selectivity of ion-exchange resins for gold cyanide has not
been studied in detail. Moreover, few studies have investigated the use of highly saline
eluants for the recovery of metal-cyanide species.
This chapter will review the literature on the use of ion-exchange resins for the
recovery of gold from cyanide solution, and show that the research to date has produced a
wealth of information on the sorptive and elution properties of various commercially
available ion-exchange resins for gold-cyanide and other metal-cyanide complexes.
Recent spectroscopic studie s on the speciation of metal-cyanide complexes in aqueous
cyanide solution have provided new insights into copper-cyanide speciation as the
salinity of the leachate is increased. This particular result has significant implications for
the selectivity of ion-exchange resins for gold cyanide and has led to an increased
fundamental understanding of the ion-exchange behavior of metal-cyanide complexes.


2.1. Cyanidation

The dissolution of gold by cyanide was originally described by Elsner in 1849

(Marsden and House , 1992). However, it was not until 1887 that the process by which
gold was leached from ores using a cyanide Iixiviant was patented (Eisele, 1988). The
cyanidation process was first used for the recovery of gold on a commercial scale in New
Zealand in 1889 (Tran , 1992). Since that time , the cyanide leaching of gold has been a
well-established and important hydrometallurgical process.
Kudryk and Kellogg (1954) demonstrated that the dissolution of gold in cyanide
solution is due to an electrochemical reaction. The overall reaction is:

2Au (s) + 4CN- (aq) + 1/202 (g) + H20 (L) ~ 2[Au(CN)2r (aq) + 20H- (aq) (1)

An additional reaction important for the leaching of metals from an ore using cyanide is
represented by:

HCN (aq) ~ H+(aq) + CN- (aq) (2a)

CN- (aq) + H20 (L) ~ HCN (aq) + OH- (aq) (2b)

The cyanidation reaction (Eq. I) takes place in an alkaline environment. Since the pK. of
the reaction presented by Eq. 2a is 9.2 (Hogfeldt, 1982), a high pH (>10) is necessary to
ensure that most of the cyanide is present in solution as the "free" form.
Gold is found in nature mainly as native gold . A number of gold alloys such as
electrum (a gold-silver alloy) and tellurides constitute the main gold minerals. To these
three types of gold-bearing minerals, a fourth class, referred to as "invisible" gold must
be added . When this invisible gold is dispersed in the form of solid solution or submicron
inclusions in sulphide minerals (Gasparrini, 1983) , this form of gold is effectively
"locked up" and, as a consequence, cyanide is unable to come into direct contact with the
metal to leach it. The recovery of gold from the leached solution is made difficult because
these sulphide matrices that contain copper must be broken down initially so that the gold
can become leachable. The leaching of such minerals results in the formation of a series
of copper-cyanide complexes, thiocyanate, and cyanide degradation products such as

2.2. The Resin-In-Pulp (RIP) Process

The RIP process for gold recovery is an alternative process to carbon-in-pulp (CIP)
and was first investigated when the first anion-exchange resins were synthesized
(Sussman et al., 1945). A number of features of the RIP process make it more attractive
than CIP for certain applications. These features have been well documented by Fleming
(1991) and include: (1) improved kinetics and equilibrium loading of gold cyanide, (2)
elution at ambient temperatures, and (3) minimal fouling by organics. It should be noted ,
however, that resins are not as selective as activated carbon, are less dense (which may
cause dispers ion problems), and their physical strength and resistance to attrition

(Fleming and Seymore, 1990) and abrasion in the sorption tanks are largely unknown.
These factors provide the motivation for the current research on the recovery of gold
cyanide using resin technology.
The RIP process involves the direct contact of ion-exchange resins with the cyanide-
leached pulp in countercurrent flow. The gold loaded onto the resin is later recovered by
elution once the resin has been separated from the pulp by screening . If necessary, the
resin is regenerated after gold elution is complete. The RIP process is analogous to that of
the CIP process . However, the type of elution and regeneration processes used will
depend upon the type of ion-exchange resin used. The wide industrial application of RIP
technology as a means of processing gold-containing ores has not occurred in the
Western world, presumably because no resin has yet been identified that is sufficiently
gold-selective over base metals, such as copper.
Resins are more versatile substrates than activated carbon because specific functional
groups can be introduced into the resin matrix during synthesis . As a consequence,
features such as selectivity , resistance to chemical attack, durability and loading capacity,
can be considered in tailoring a resin to a particular application .
It has been demonstrated that resins have superior rates of adsorption and greater
equilibrium loading capacity of gold cyanide as compared to carbon . They also can be
effectively eluted at ambient pressures and temperatures, thereby eliminating the need for
the expensive high-temperature elution and acid-wash regeneration processes discussed
previously for activated carbon. Moreover , they do not appear to be poisoned by organic
foulants that severely inhibit carbon adsorption .
The gold-sorption properties of several commercially available resins have been
investigated previously. These resins generally exhibit nonselective sorption properties .
It has been proposed that the hydrophilic nature of the resin, the chemical structure of the
functional group and the ionic density (i.e., the number of "active" sites per unit volume
of resin) significantly affect the ability of the resin to selectively extract the gold-cyanide
complex (Riveros and Cooper , 1987; Riveros, 1993).

2.3. Chemistry

Our current understanding of the ion-exchange behavior of metal-cyanide complexes

and the factors affecting a resin's selectivity and capacity to load gold cyanide, as well as
effective eluants for the recovery of gold, can be attributed to a series of pioneering
studies .
Investigations by Burstall et aI. (1953) and Burstall and Wells (1955) established that
gold could readily be sorbed from cyanide solution onto a strong-base ion-exchange
resin, and that it was possible to selectively remove the metals from the resin using
different eluants , including hydrochloric acid, sodium cyanide, and organics solvents
such as acetone. However, the expense and technical difficulties of this type of selective
elution procedure were not considered.
A study by Aveston et aI. (1958a) established for the first time that the number of
weak-base groups within a strong-base group environment affected the selectivity of the
ion-exchange resins for gold cyanide. Work on the effect of weak-base group content on
the selectivity of ion-exchange resins for gold cyanide was advanced by Aveston et aI.
(19581;» . The increase in selectivity of the resin for gold cyanide was aptly described by
the fact that with multivalent anions, sorption is limited by the increased distance
between the ionized resin sites and the equilibrium position of the adsorbed anion at the

"centre of gravity" of these groups, which reduce the polarization of these anions by the
functional groups on the resin.
More recently, the sorption and elution properties of commercial and experimental
ion-exchange resins for metal-cyanide species have been studied extensively, and have
been the subjects of several reviews , includ ing those by Hosking (1984), Riveros and
Cooper (1987), Fleming (1991), and Jennings (1991).
Resins may be categorized as either strong- or weak-base exchangers. By definition,
weak-base anion-exchangers contain predominantly primary, secondary, or tertiary amine
functional groups, while strong-base resins contain quaternary ammonium groups.
Importantly, weak-base resins may contain a small proportion of strong-base groups with
a fixed positive charge .
Ion-exchange of the gold-cyanide complex on a strong-base resin can be described
by the following equation (Fleming and Cromberge, I984a):

---- (3)

where the symbol I- denotes the inert matrix of the resin (usually a copolymer of
polystyrene and divinylbenzene).
In a highly alkaline environment such as a leach stream, it has been proposed that
gold is reversibly sorbed onto the resin via the mechanism described by Eq. 3; the
equilibrium lies strongly to the right because the affinity of the strong-base group for the
highly polarized gold-cyanide complex is relatively high. However, this affinity is also a
function of the nature of the alkyl groups attached to the nitrogen atom . Experimental
studies have shown that the selectivity of the resin and, consequently, the sorption
equilibrium depend upon the length of these alkyl chains (Schwellnus and Green, 1988;
Riveros, 1993).
Chemical or physical processes can be used for the elution of metal-cyanide species
from strong-base resins. The three methods developed involve treating the resin with zinc
cyanide, thiocyanate or thiourea . The basis for the zinc-cyanide elution method is that the
zinc-cyanide anion is more strongly sorbed onto the resin than the gold-cyanide complex
(Hazan, 1957). The equilibrium shown in Eq. 3 is shifted to the left when the system
contains an anion having a stronger affinity for the resin than the gold-cyanide complex.
The equation for the physical displacement reaction is:

2(1- N+R3[Au(CN)2n + [Zn(CN)4f- ~

(I- N+R3MZn(CN)4]2- + 2[Au(CNhr (4)

The kinetics are slower than the chemical elutions detailed below . However, the elut ion
has been shown by Fleming and Cromberge (I984b) and Seymore and Fleming (1989) to
be very efficient, with excellent gold and base-metal recovery. After elution the resin
needs to be regenerated to remove the strongly sorbed zinc-cyanide species. The most
efficient method known is treating the resin with a dilute mineral acid such as sulphuric
acid (Eq. 5).
184 G. C. LUKEY ET AL.

Thiocyanate has more affinity for the resin than does the gold-cyanide complex.
Therefore, under the correct conditions the thiocyanate anion can also displace the gold-
cyanide complex in an ion-exchange process represented by the following equation :

As was the case with the zinc-cyanide elution method, the thiocyanate-eluted resin must
be regenerated before returning the resin to the sorption circuit. Mineral acids can be used
for this regeneration step, but Davison et al. (1961) has reported that, in strong acids, the
thiocyanate anion can decompose to elemental sulphur, thus poisoning the resin. An
alternative approach proposed by Fleming (1985) involves complexing the thiocyanate
anion with ferric sulphate to form a cationic complex.
Treating the loaded resin with thiourea is the elution method most widely practiced
in former Soviet RIP plants (Bolinski and Shirley, 1996) . As shown by Eq. 7, the
reaction between the gold-cyanide complex and thiourea also results in the formation of a
cationic complex that is no longer bound to the resin.

~ WR.3[Au(CN)2r + 2CS(NH2)2 + 2H2S04 ~

~ N+R3HS04- + 2HCN + [Au(CS(NH2)2)2]~S04- (7)

Ion-exchange of the gold-cyanide complex on a weak-base resin differs from that on

a strong-base resin because the former requires initial protonation of the primary,
secondary, and tertiary amine groups (Eq. 8).


From Eq. 8, it is evident that the pH of solution has an important influence on the ion-
exchange behavior of weak-base resins. In the free-base form, weak-base resins are
uncharged and so need to be protonated prior to the extraction of gold cyanide.
Protonation of the resin occurs at pH values below the pK. of the tertiary amine
functional groups. Functional groups present on weak-base resins have a pK. value
between 8 and 9 and are, therefore, theoretically unable to sorb strongly species in the
leach stream with typical pH between 10 and 12 (Fleming, 1998). Therefore the pH of the
pulp must be lower than the pK. of the resin for sorption of the gold via ion-exchange to
be possible. This dependence on pK. has led to considerable efforts to develop weak-base
resins containing specific functional groups that will increase the pK. of the resin. This
permits the efficiency of the weak-base resin at the natural pH ofthe leached pulp.

Once protonated, the ion-exchange of the gold-cyanide complex is analogous to that

for strong-base resins. The elution of the gold-cyanide complex from the resin is
achieved by increasing the solution pH (Eq. 9):


The concentration of the hydroxide ions during elution is important; otherwise, the base
metals will not be completely eluted from the resin. The elution of metal-cyanide species
is very economical compared to the expensive reagents required for the successful elution
and regeneration of strong-base resins (Harris et aI., 1992).


The aqueous chemistry of metal-cyanide complexes in solution has been studied

extensively using UV, IR, Raman, and 13C_NMR spectroscopic techniques (Fagan, 1998).
However, the speciation of copper in cyanide systems is of most interest in the processing
of gold because of the ability of copper to form three copper(I)-cyanide complexes of
different size and geometry. For this reason , the discussion below is limited to the
speciation of copper(l)-cyanide and complexes in nonsaline and saline solutions. The
effect of salinity on the distribution of copper cyanides in solution is of practical
significance because, as shown in the following section, it is possible to selectively
extract gold from saline solutions containing other metal-cyanide complexes.
The UV spectrum of copper cyanide in nonsaline and saline solution (3 M KCl) is
presented in Fig. I . The absorption spectra of copper cyanide only (with no addition of
KCl) show two distinct absorbance peaks at 237 nm and 208 nm, respectively,
confirming the existence of copper-cyanide species (Pierrard et aI., 1971). There is a
gradual change in the spectra as the CN:Cu molar ratio is altered. The absorbance peaks
in Fig. I cannot be separately assigned to the respective individual copper-cyanide
complexes [Cu(CNh]2- and [CU(CN)4]3-. The significant change in the spectra when
copper cyanide exists in solution with 3 M KCI suggests that a change in copper-cyanide
speciation may occur in saline solutions. Work reported by Kappenstein and Hugel
(1969) and Pierrard et al. (1971) has also shown that the UV spectra of copper-cyanide
complexes are quite distinctive, with two maxima occurring at approximately 210 nm and
235 nm, and three isosbestic points observed at 205-210 nm, 234-236 nm, and 243-246
nm, respectively. Furthermore, Kappenstein and Hugel (1969) reported a gradual change
in the UV spectrum as the CN/Cu molar ratio was altered, indicating a change in
speciation in solution.
The Raman spectra obtained for copper-cyanide solutions in nonsaline and saline
solution for different CN/Cu molar ratios are presented in Figs. 2 and 3. A summary of
the peak assignments is presented in Table 1. Figure 2 shows that as the CN/Cu molar
ratio is increased from 2.2 to 3.5, the broad peak at 2137 cm' gradually disappears and a


" : ... .. . [KGI) = 0 M!

--[KGI) = 3 Mi
.c 1.5


oL----L_......._ ....... _.&...--.l_--'-_...J:::::::::.._........-l...........J

190 200 210 220 230 240 250 260 270 280 290
Wavelength [nm]
Figure 1. UV-visible spectra of copper cyanide. [Cu] = 10 mglL, [CN-].. cess = 200 mglL. Reprinted from
Lukey et al. (I 999b), Copyright (1999), with permission from Elsevier Science.

Table 1. Raman and IR peak assignments for aqueous copper(I)-cyanide complexes.

Raman data are taken from Kappenstein et at. (1978) and Lukey et at. (1999a). The IR
data are from Penneman and Jones (1956).


v , Polarized (ern") v, Depolarized (em") v, sym v, asym
(ern") (cm'")

CuCN 2175 2172

CN"' 2077 2080
[Cu(CNhr 2137 2125
[Cu(CNhl2- 2108 2094 2094
[Cu(CNM~'" 2094 2078 2076

new peak at 2077 cm- I gradually forms, The intensities of the two peaks seen in each of
the spectra at 2108 ern" and 2094 ern", respectively, gradually increase to produce well-
identified peaks as the CN/Cu molar ratio is increased to 3,5. Neither of the peaks in Fig.
2 can be assigned to the precipitation of CuCN from solution at low CN/Cu molar ratios
because the Raman peak for solid CuCN occurs at a wavelength of2174 em-I, Lukeyet
al. (1999a) therefore assigned the line at 2137 em" to the symmetric vibrational stretch of

[Cu] = 0.01 M

t [NaCI] = 0.0 M

CN I Cu =3.5



2050 2065 2080 2095 2110 2125 2140 2155 2170 2185
Raman Shift [cm·1]
Figure 2. Raman spectra of aqueo us copper(l) cyanide as a function of the CN/Cu molar ratio. [NaCI) = 0.0 M.
Reprinted from Lukey et al. ( 1999a), Copyright ( 1999), with permission from Elsevier Science .

[Cu] = 0.01 M
[NaCI] = 4.0 M



0:: 3.0


2050 2065 2080 2095 2110 2125 2140 2155 2170 2185
Raman Shift [cm ' ]

Figure 3. Raman spectra of aqueous copper(l) cyanide as a function of the CN/Cu molar ratio . (NaCI) =4.0 M.
Reprinted from Lukey et al. (I 999a ), Copyright (1999 ), with permission from Elsevier Science .

[Cu(CN hr in solution, because this peak gradually disaPf-ears as the CN/Cu molar
ratiois increased, making the formation of [Cu(CN h l - and [Cu(CN)413- more

predominant. Similarly , the peak at 2108 cm- is assigned to the symmetric vibrational
stretch of [CU(CN)3f-. The Raman peak at 2094 em" is assigned predominantly to the
symmetric vibrational stretch of [Cu(CN)41 3- and also to the weak asymmetric stretch of
[Cu(CN hl z-. Similarly , the weak asymmetric stretch of [Cu(CN)41 3- reported to occur at
2077 cm- 1 is coincident with the vibrational stretch of free cyanide also occurs at the same
These peak assignments are in good agreement with previous Raman studies
(Chantry and Plane, 1960; Kappenstein et al., 1978) and with the respecti ve structures in
aqueous solution. However, the assignments contradict other work on peak assignments
(Reisfeld and Jones, 1965). Th is latter study, however , assigned the Raman peaks
observed in the solution spectra on the basis of the poorly-resolved peaks observed in
powdered samples of KCu(CN)z and K3Cu(CN)4 '
A comparison of Fig. 3 with Fig. 2 shows that the Raman spectra of aqueous copper
cyanide changes significantly in highly saline solution compared with nonsaline solutions
(4.0 M NaC!, which is equivalent to 233 .7 gIL). For a CN/Cu molar ratio of 2.2, the
broad peak at 2137 ern" (Fig. 2) assigned to [Cu(CNhr is not as distinct for the copper-
cyanide solution spectrum in 4.0 M NaCI (Fig. 3). The peak appears to have shifted to be
coincidental with the peak assigned to [Cu(CNhl z-. For the formation of [Cu(CNhf-
from [Cu(CNhr in highly saline solutions where the free cyanide concentration is
effectively zero, the formation of CuCN must also occur to liberate additional cyanide
required for complexation. A rise in the background occurs at 2174 cm- I in the spectrum
for a CN/Cu molar ratio of 2.2 (Fig. 3). Lukey et al. (1999a ) proposed that this is
indicative of the format ion of CuCN.
The effect of free cyanide concentration on the proposed phenomena is further
illustrated in Fig. 4. The copper concentration of each solution in Fig. 4 is 0.01 M. The
exces s free cyanide concentration was 200 ppm. With such a large excess of cyanide ,
both [Cu(CN hl - and [Cu(CN)41 3- should form predom inantly in solution. Without the
addition of KCI, three peaks are observed in the Raman spectra for copper cyanide (Fig.
4). As the concentration of KCI is increased, the peak at 2108 em" gradually disappears
so that for the spectra obtained for KCI concentrations of 2 M to 4 M, only two peaks are
observed; therefore , the distribution of copper-cyanide complexes changes in highly
saline solut ion, preferentially forming [Cu(CN)41 3- when the free cyanide concentration is
not limiting.


Most work on the use of ion-exchange resins for gold recovery has assumed ideal
ion-exchange between the metal-cyanide complex and the functional group attached to
the ion-exchange resin surface. As a consequence, the possible precipitation of AuCN,
CuCN. or polymers thereof, as has been shown to occur on activated carbon (Adams and
Fleming, 1989), has never been established . Apart from a speciation study using an
iterative slope technique (Riveros and Cooper , 1988), few published papers have

QI [KCI] =0 M
~ 2M


2040 2060 2080 2100 2120 2140 2160 2180

Raman Shift [em-1]
Figure 4. The effect of KCI concentrat ion on the speciation of copper(l)-cyanide complexes (0.01 M) in
aqueous media. [C~l",,,,,, = 200 mgIL. Reprinted from Lukey et al. (1999a), Copyright (1999), with permission
from Elsevier Science.

considered the nature of the sorbed metal-cyanide species on the resin. The form in which
the copper-cyanide complex predominantly exists on the resin surface is of major
importance because it will influence the ability of ion-exchange resins to selectively
recover gold cyanide from solution.
As discussed in Section 3, the aqueous chemistry of metal cyanides has been studied
extensively, and there is considerable knowledge about the species present under a given
set of experimental conditions. However, there is some uncertainty as to the exact
composition of the speciation of complexes that are sorbed by ion-exchange resins from
dilute cyanide solutions. To study the effect of resin structure on the speciation of sorbed
copper-cyanide comr,lexes, one must first characterize the resin substrate.
The CPIMAS I C-NMR spectra obtained for a variety of ion-exchange resins (Lukey
et aI., 2000a) are presented in Fig. 5, and the chemical shift values are shown in Table 2.
The peaks appearing in each spectrum at 40.7 ppm, 128 ppm, and 144 ppm are due to the
polymeric carbons (~H2CH-)n that are a part of the polystyrene - divinylbenzene
matrix, the aromatic carbons and the quaternary aromatic carbon, respectively. These
peaks are observed in all the resin spectra presented and are in agreement with results
obtained in a previous study (Bassedas et aI., 1989).
Spectrum (a) and spectrum (b) in Fig. 5 exhibit peaks that are typical of the presence
of trimethylamine and tripropylamine functional groups on the resin. Of more interest,
however , are spectra (c) and (d), as these resins have been synthesized using secondary
amine reagents, dimethylamine and dipropylamine, respectively. These resins should





200 180 160 140 120 100 80 60 40 20 o

Figure s. CPIMAS 13C-NMR spectra of (a) trimethylamine, (b) tripropylamine, (c) dimethylamine, and (d)
dipropylamine resin respectively. Peaks labelled with an asterisk (*) are spinning side bands. Reprinted from
Lukeyet al., (2000a), Copyright (2000), withpermission of Elsevier Science.

Table 2. Chemical shift assignments for each resin: (a) polymeric carbon atoms in
polystyrene-divinylbenzene matrix , (-~Hz-CH-)n ; (b) aliphatic carbons attached to the
nitrogen atom in the functional group, (-N~); (c) benzylic carbon, (-kHz-WR3) or
(-kHz-Cl) ; (d) aromatic carbon; and (e) quaternary aromatic carbon.


Peak 1 Peak 2 Peak 3 Peak 4 Peak 5 Peak 6 Peak 7
Trimethylamine 40 .7, a 52.9, b 68.9, C 127.9, d 146.2, e
Tripropylamine 11.6, b 16.8, b 40 .9, a 63.1, C 128.1, d 144.6, e
Dimethylamine 41.0 , a 65.2, C 69.7, C 128.3, d 146.1, e
Dipropylamine 12.5, b 21.5, b 40.7, a 59.0, C 63.9, C 128.3, d 146.1, e

contain only weak-base functional groups . However, the CPIMAS 13C-NMR spectra (c)
and (d) of the aminated resin show two different types of amino functional groups on the
resin. Th is is evident from two benzylic carbon signals at 69 .7 ppm and 65.2 ppm for the
dimethylamine resin and at 63.9 ppm and 59.0 ppm for the dipropylamine resin (Lukey et
al., 2oo0a) . The peaks at 69.7 ppm for dimethylamine resin and 63.9 for dipropylamine
resin are at approximately the same chemical shift as the benzylic carbon peak identified
for the quaternary ammonium functional groups [spectra (a) and (bj] . The second peak
correlates well with the model compounds dimethyl-benzylamine and
dipropylbenzylamine, respectively, confirming the presence of weak-base functional
groups on the resin .
It has been proposed by Lukey et al. (2000a) that the formation of strong-base groups
on ion-exchange resins synthesized with secondary amine reagents arises due to the
internal attack of the neighbouring unreacted chloromethylated sites by the tertiary amine
functional groups that are formed dur ing synthesis (Fig . 6). The mechanism by which
strong-base groups are formed during synthesis using a secondary amine reagent , such as
dimethylamine or dipropylamine, has been proposed previously (Conradie et al., 1995);
however work by Lukey et al. (20ooa) using CPIMAS 13C_NMR has clearly identified the
existence and the chemical structure of these groups.
It is expected the predominant copper-cyanide species that sorbs onto ion-exchange
resins would be the copper tricyano complex because it is thermodynamically the most
stable species in solution (Fagan, 1998). Raman spectra presented in Fig. 7 show three
peaks at 2074, 2093 , and 2109 em" , respectively, for both the commercial Amberjet
4400 resin and the trimethylamine resin. The peak shifts for each complex sorbed onto
the anion-exchange resins were predicted to slightly shift, relati ve to the peak shifts
observed in aqueous solution, because this had been observed in previous work by Jones
and Penneman (1954) . These three peaks (Fig . 7) are assigned to the symmetric
vibrational stretch of [Cu(CNhl z- and the symmetric and asymmetric vibrational stretch
of [Cu(CN)41 3-, respectively. Notably, the weak asymmetric vibrational stretch of
[Cu(CNhl z- is coincidental with the intense symmetric vibrational stretch of [Cu(CN)413- .
The peak at 2074 cm- is not due to the sorption of free cyanide. The assignment of the
peaks observed in the Raman spectra are supported by the peaks observed in the FfIR
spectra that were obtained for the same resins (Lukey et al., 2000a) . The observed Raman
shifts and FfIR peak assignments reported are in good agreement with prev iously
reported results (Table 3).
192 G. C. LUKEY £T AL

Me ow


Partial structure of D2780

/ Intramolecular
displacement of chloride

Presence of weak-base along with strong -base sites

Figure 6. The mechanism of formation of strong-base groupson resins using weak-base amine reagentsduring
synthesis. Reprinted from Lukeyet al. (2000a). Copyright (2000). with permission of ElsevierScience.

f [Cu(CN).J3-

Amberjet 4400

Trlmethylam ine

2000 2025 2050 2075 2100 2125 2150 2175 2200

Raman Shift [em·1]
Figure 7. The distribution of copper-cyanide complexes on commercial resin Amberjet 4400 and synthesized
trimethylamine resin. The difference between the two resins is the ionic density of the resin. Reprinted from
Lukeyet al. (2000a), Copyright (2000). with permission of Elsevier Science.

Table 3. Observed Raman shifts for gold-cyanide and copper-cyanide loaded resin.


v, sym (cm'") v, asym (em") v, sym (em") v, asym (em")
CN- 2077 R 2077 R
20751R 2080lR
[Au(CNhr 295R&2164R 2164 & 2162 R
2140lR 2140 & 2141 IR
[Cu(CNhr 2137 R 2137 R
21251R 21251R
[Cu(CNhl2- 2108R 2094 R 2108R 2094 R
20901R 20941R
[Cu(CN)413- 2094 R 2078 R 2094 R 2078 R
20711R 2076IR
a Raman shifts and IR wavenumbers are those observed in Kappenstein et al. (1978), Penneman and Jones
(1956),Jones (1957), and Chadwickand Frankiss (1976), respectively.

No conclusive evidence of the formation of [Cu(CNhr at 2137 em" exists for resins
containing trimethylamine functional groups (Fig. 7). However, the FfIR spectra
obtained for dimethylamine and dipropylamine resins do contain a peak at 2125 em"
(Fig. 8), confirming the sorption of [Cu(CNhr on each resin, and establishing that the
length of the alkyl chain and the stereochemistry of the functional groups affect the
speciation of the sorbed copper-cyanide complex .
Raman and FfIR spectroscopy can also be used to determine the speciation of gold
cyanide sorbed onto ion-exchange resins. The gold-cyanide complex exhibits two peaks
in Raman spectroscopy. The peak in the lower wave number region of the spectra (200-
400 em") is attributed to the vibrational stretch of the Au-CN bond (Nakamoto, 1962).
The peak observed in the higher region of the spectra (the lower energy region) is that of
the asymmetric carbon - nitrogen stretching vibration (Jones, 1953, 1957). The peak
observed at 2140 ern" in the FfIR spectra of each resin loaded with gold cyanide (Fig. 9)
is evidence of the loading of [Au(CNhr (Jones and Penneman, 1954), which is expected
due to the thermodynamic stability of the [Au(CNhr complex (log 13 = 47.0) (Hogfeldt,



The detrimental effects of highly saline water on the CIP process have been
documented in the literature (Ariti and La Brooy, 1988). These include a decrease in the
rate of gold adsorption onto activated carbon due to the increased viscosity of the pulp,
and increased carbon loss, as the salts loaded on the carbon can catalyze its
194 G. C. LUKEY ET AI.

of [Cu(CN)2l'


2000 2025 2050 2075 2100 2125 2150 2175 2200

Wavenumber [em"]
Figure 8. FTIRspectraof copperloaded ion-exchange resin. Reprinted from Lukey et al. (2000a); Copyright
(2000), withpermission of Elsevier Science.


I Amberjet 4400
of Trimethylamine


2050 2075 2100 2125 2150 2175 2200

Wavenumber [em'1]

Figure 9. FTIRspectraof gold loaded ion-exchange resins. Reprinted from Lukey et al. (2000a), Copyright
(2000), withpermission of Elsevier Science.

decomposition during regeneration . The effect of salinity on the capacity and selectivity
of ion-exchange resins containing different functional groups is, therefore, of interest
because RIP technology may have advantages compared with CIP/CIL in locations with
highly saline ground water.
Reports have indicated that the decrease in gold loading on ion-exchange resins as
ionic strength increases (Fleming and Cromberge, 1984a) is attributed to the increase in
competition of anions, such as chloride, for active sites on the resin that occupy gold
cyanide . In contrast, further studies have reported no change in gold adsorption with
strong-base resins and only a slight decrease in gold adsorption by weak-base resins with
increasing salinity (Hla et aI., 1991). In addition, the selectivity of strong- and weak-base
resins has been shown to improve slightly in highly saline solutions (RIa et aI., 1992),
although a plausible explanation was not offered for the observed sorption behavior.
Recent work by Lukey et aI. (1999b) has shown that the use of synthetic solutions
that increase salinity has a beneficial impact on the selectivity of ion-exchange resins.
The results obtained by Lukey et aI. (l999b) for the sorption of a mixture of metal
cyanides in saline solutions onto resins containing different functional groups are
presented in Figs. 10 to 12 ([Aulo = 5 mgIL, [Agl o = 5 mgIL, [Culo = 15 mgIL, [Znlo = 10
mglL, [Fe]o = 10 mg/L), Increasing the total ionic strength of solution by the addition of
KCI has the effect of significantly increasing the selectivity of each resin . It is shown that
as the chloride concentration is gradually increased to 3 M, the loading of gold onto each
resin remains unchanged, while the loading of copper cyanide is significantly decreased.

~ 7.0
,g, 6.0


3.0 I\-~~-+------+-----J

0.0 0.5 1.0 1.5 2.0 2.5 3.0
Chloride Concentration [M]
Figure 10. The effect of chloride concentration on the selectivity of a trimethylamine resin. [CN1excess = 150
ppm. Reprinted from Lukeyet al. (1999b) , Copyright (1999) , withpermission from ElsevierScience.


.§. 3.0


.9sGl 2.0 •
== 1.0

0.0 ~--....L--- ....--.....------.. . . --__iI:

0.0 0.5 1.0 1.5 2.0 2.5 3.0
Chloride Concentration [M]

Figure 11. The effect of chloride concentration on the selectivity of a tripropylamine resin. leWl" "" = 150
ppm. Reprinted from Lukeyet al. (1999b). Copyright (1999). withpermission from Elsevier Science.

At a chloride concentration of 3M copper cyanide did not load onto any of the resins.
Similar trends were observed for the iron-cyanide complex.
As discussed previously. there is an increasing need to process ores of a very low
gold grade «1 glt) compared to that of the base metals such as copper (ca. 200 glt). For
this reason. the effect of salinity on the selectivity of the trimethylamine resin was
investigated when the initial copper concentration in solution was high (2 15 ppm) relative
to that of gold (5 ppm). The concentration of the other metals remained the same as in the
previous experiments. Figure 12 shows that the trimethylamine resin is very selective as
the chloride concentration in solution is increased . even though the copper concentration
in solution is significantly higher than to the other metals. As expected. when the total
ionic strength of the solution is not adjusted. the loading of copper onto the resin is
significantly higher than on the same resin at a much lower copper concent ration (Fig.
10). The additional sorption of copper cyanide impairs the loading of zinc and iron onto
the resin. The latter result further shows that the iron-cyanide complex interacts only
weakly with the functional group on the resin. possibly due to its octahedral structure.
Although the loading of silver cyanide decreases slightly when there is a high level of
copper in solution. the loading of gold cyanide is seemingly unaffected. This result
further confirms the stability and strong sorption of gold cyanide onto strong-base anion-
exchange resins.

I·AU .Ag .Cu xZn )KFe I
~ 14.0
.§. 12.0
:a1'0 10.0
..J 8.0
Sell 6.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Chloride Concentration [M]
Figure 12. The effect of a high (215 ppm) initial copper concentration on the selectivity of trimethylamine
resin. Chloride concentration adjusted by NaCI. ICW)"ce,," 150 ppm. Reprinted from Lukey et al. (I999b).
Copyright (1999), withpermission from Elsevier Science.

It was shown in Section 4 that the equilibrium distribution of copper-cyanide

complexes changes in highly saline solutions, preferentially forming [Cu(CN)41 3- when
the free cyanide concentration is not limiting. The tetracyano copper complex is
tetrahedral in geometry and has a negative 3 charge . Consequently, three functional
groups of the correct stereochemistry must be in close proximity for sorption via an ion-
exchange mechanism to take place. This has the effect of significantly decreasing the
ability of [Cu(CN)41 3- to load onto the ion-exchange resin, thereby increasing the
selectivity for univalent ions such as [Au(CNhr. Lukey et al. (1999a) proposed that this
is the reason for the increase in selectivity of ion-exchange resins for gold cyanide over
copper cyanide is observed in highly saline solution.


The recovery of gold cyanide from anion-exchange resins has been the focus of
many investigations in the Western world (Hosking, 1984). From this work, several
suitable eluants have been proposed for the efficient recovery of gold and other metal-
cyanide complexes, including ammonium thiocyanate, thiourea, and zinc cyanide
(Fleming and Cromberge, 1984b). Other reagents have also been investigated, including a
combination of mineral acids and polar solvents, such as acetone (Aveston et al., 1958;
Burstall et al., 1953) or organic/water mixtures (Stamboliadis et al., 1978). However, a

simple and cost-effective elution procedure has not been developed that is able to
selectively strip metal cyanides from resins. Furthermore, the majority of studies on the
sorption and subsequent elution of gold and other metal cyanides have been performed on
commercially available ion-exchange resins. As a consequence, the types of functional
groups and resin matrices investigated have been limited.
As discussed previously, the thiocyanate anion and the zinc-cyanide complex have
more affinity for the resin than gold cyanide, and so, under the correct conditions , are
able to displace the gold complex (Fleming and Cromberge, 1984b). It has also been
shown that efficient elution of iron cyanide from strong-base resins is accomplished with
a 2 M NaND) solution. However, gold cyanide was also partially eluted in this process,
which made the eluant unfeasible (Riveros et al., 1993). Sodium chloride solutions have
been shown to efficiently elute both copper and iron cyanides from resins (Leao et al.,
1998). The focus of these investigations was on the recovery and recycling of cyanide
from tailings streams ; therefore, the strong-base resins investigated were loaded with only
copper, iron and nickel (high consumers of cyanide).
Lukey (2000) recently investigated the effect of the ionic density and hydrophilicity
of the resin and also the stereochemistry of the functional group on the elution of metal
cyanides. Saline solutions, the thiocyanate anion, and zinc cyanide were used as eluants
for ion-exchange resins. Figures 13 and 14 show that copper cyanide and iron cyanide are
selectively eluted from trimethylamine and dimethylamine resins using 2 M KCl,

90 -+-Au Ag
--.-Cu ~Zn
80 -.-Fe


*' 40
0 2 4 6 8 10 12 14 16 18
Bed Volume

Figure 13. Elution of metals from trimethylamine resin. Ionic strength adjusted to 2 Musing KCI. [CWIrm: ::::
200 ppm. Reprinted from Lukeyet al. (2000b), Copyright (2000).withpermission of Elsevier Science.

e 60
::I 50
m -+-Au Ag
~ 40
---K- Zn
- . - Cu
_____ Fe
0 2 4 6 8 10 12 14 16 18 20
Bed Volume
Figure 14. Elution of metals from dimethylamine resin. Ionic strength adjusted to 2 Musing KCI. [CN-1rr",, ::
200 ppm. Reprinted from Lukey et al. (2000b), Copyright (2000), with permission of Elsevier Science.

containing 200 ppm cyanide as KeN. Figure 13 shows after the passage of 8 bed volumes
of eluant that 100% of the iron and 45% of copper are eluted from the trimethylamine
resin. A similar result was reported for a resin containing tripropylamine groups.
As discussed in Section 4, the dimethylamine resin contains predominantly weak-
base functional groups with some strong-base functional groups attached to the resin
matrix. Consequently, the dimethylamine resin has a lower strong-base capacity than
other resins. such as the trimethylamine resin. Figure 14 shows that the elution of copper
cyanide from the dimethylamine resin exceeds 80% within the first 8 bed volumes of
eluant. The elution of iron cyanide is initially rapid. with 50% of the iron being eluted
within the first bed-volume. After the passage of 2 bed volumes of eluant, 87.5% of the
iron is eluted from the resin. The gold-cyanide complex is not eluted from either resin.
The rapid elution of iron cyanide from the dimethylamine resin occurs due to the
resin's low strong-base capacity and a sterically bulky functional group. Consequently,
the resin cannot easily accommodate iron cyanide because it requires at least four
functional groups in close proximity for sorption to occur. Furthermore, the structure of
the functional group on the dimethylamine resin can vary significantly (Fig. 6), unlike
that of simple alkyl amine-type functional groups.
In contrast to the elution curves obtained for the trimethylamine and dimethylamine
resins, only a limited amount (25%) of copper cyanide is eluted from the dipropylamine
resin using 2 M KCl containing 200 ppm cyanide (Fig. 15). The poor elution of copper

80 --+- Au Ag
70 -.-Cu ~Zn
--lIE- Fe
c 60
0 40
0 2 4 6 8 10 12 14 16 18 20
Bed Volume
Figure 15. Elution of metals from dipropylamine resin. Ionic strength adjusted to 2 Musing KCI. [CWlr"", ::
200 ppm. Reprinted from Lukey et al. (2000b), Copyright (2000), with permission of Elsevier Science .

cyanide from the dipropylamine resin is attributed to the loading of the linear [Cu(CNhr
on the resin as well as [Cu(CNhf- and [Cu(CN)413-, as discussed in Section 4 (Lukey et
al., 2oooa). The [Cu(CNhr is able to strongly associate itself with the dipropylamine
functional group because of its smaller size and linear geometry, thereby being less
accessible to chloride ions in the eluant.
The results presented in this section establish that the speciation of the sorbed
copper-cyanide complex affects the effectiveness of saline water as an eluant. However,
the elution of copper from resins containing trimethylamine functional groups is possible
using KCl because the chloride anion is readily able to displace both the iron cyanide and
[Cu(CNh1 2- and [Cu(CN)413- complexes. The [CU(CN)4]3- complex is rapidly eluted from
the resin because it cannot associate itself closely with the functional group due to its
tetrahedral geometry. Lukey (2000) has proposed that the selective elution of cop-per by
saline solutions occurs because [Cu(CNh]2- preferentially converts to [Cu(CN)41 - in the
resin phase. Because the [Cu(CN)41 3- complex is more hydrated than [Cu(CNh12-, the
former complex is less readily accepted into the hydrophobic resin framework; hence, the
elution of this complex is more easily achieved than that for [Cu(CNh1 2-. The same
elution mechanism can be used to explain the selective elution of copper from resins
containing tripropylamine, dimethylamine , and dipropylamine functional groups.
Fleming and Cromberge (l984b) performed a comprehensive investigation of the
elution properties of resins using nonsaline eluants, showing that thiocyanate anions,
zinc-cyanide anions, and thiourea in acid solution are effective eluants of strong-base
resins, and that weak-base resins are eluted efficiently by sodium hydroxide solution . The
elution of only gold from commerically available resins was examined by Fleming and

Cromberge (l984b), and as a result , the possible prec ipitat ion of double-metal complexes
of the type Zn2[Fe(CN)6) or Fe4[Fe(CN)6h within the resin bead pores was not
considered. The formation of these types of species had been previously postulated using
a NH 4SCN eluant (Fridman et al., 1971a).
Under strongly alkaline conditions (pH > 12), zinc-ferrocyanide complexes are
soluble (Loos-Neskovic et al., 1989), and so their formation will not be observed for a 2
M KSCN eluan t in which the pH is maintained higher than 12. More than 80% of iron,
copper, and silver are eluted from a resin containing trimeth ylamine functional groups
within the first 4 bed volumes of a 2 M KSCN eluant containing approximately 200 ppm
cya nide (Fig. 16). The strongly sorbed gold- and zinc-c yanide complexes have
substantially slower elution rates. In addition, the large, octahedrally shaped iron-cyanide
com plex and the co pper cyanides are more easi ly eluted from the resin than the highly
polarized, linear gold complex (Fig. 16). Additional work repor ted by Lukey (2000) has
shown that the length of the alkyl chain of the functio nal group has no substantial
influence on the elution of metal cya nides, when using a KSC N eluant.
A com parison of Figs . 16 and 17, reveals that copper, silver, and iron are eluted more
rapid ly from the trimethylamine resin using the zinc-cyanide eluant than for the KSCN
eluant. The elution of gold from the resin is similar whether zinc cyani de or KSCN is

c 60
'5 50
0~ 40
-.-Au _Ag
20 -la-CU -*- Zn
10 -.- Fe

0 2 4 6 8 10 12 14 16 18 20
Bed Volume
Figure 16. Elution of metals from trimethylamine resin using 2 M KSCN. [CWlr,~ ::: 200 ppm. Elution
temperature of 22 °C. Reprinted from Lukey et al. (2001). Copyright (2001). courtesy of Marcel DekkerInc.

c 60
~ 50
0~ 40
20 --.-Cu
0 2 4 6 8 10 12 14 16 18 20
Bed Volume
Figure 17. Elutionof metals from trimethylamine resin using 0.5 M zinc cyanide. [CWlr,,,,:::: 200 ppm. Elution
temperature of 50 °C. Reprinted from Lukeyet aI. (2oo\), Copyright(2001), courtesyof Marcel Dekker Inc.

used. This has been reported previously (Fleming and Cromberge, 1984b) and is possibly
due to the higher temperature of the zinc-cyanide eluant (50 "C), as well as the divalent
charge of the zinc-cyanide complex, versus the univalent thiocyanate ion. Poor elution of
copper from a dipropylamine resin was observed using the zinc-cyanide eluant (Lukey,
2000) .
Whereas the elution of other metals seems to be unaffected by changing the structure
of the functional group on the resin when a zinc-cyanide eluant is used, the elution of the
valuable gold-cyanide complex appears to be significantly inhibited (Fig. 18). It is
believed that the poor elution of gold from resins of a low ionic density is predominantly
due to the stereochemistry and charge of the tetrahedral [Zn(CN)4]2- complex . For the
zinc-cyanide complex to load onto the dimethylamine resin and, therefore , elute the
loaded gold complex, at least two functional groups of the correct size and spatial
geometry are required to satisfy electroneutrality. This requirement is less easily satisfied
for resins of low ionic density because the functional groups are not close together and,
consequently, gold is poorly eluted .


A study of the interaction between metal-cyanide complexes and functional

polymers, such as ion-exchange resins, is relevant to the gold processing industry, as
resins have the potential to replace activated carbon for the in-pulp processing of gold.
Ion-exchange resins can be tailor made to selectively recover the gold-cyanide complex,
enabling the processing of ores of low grade and complex mineralogy.

c 60
~ 50
iii ~Au
0 40 ___ Ag

30 --.-Cu
0 2 4 6 8 10 12 14 16 18 20
Bed Volume
Figure 18. Elution of metals from dimethylamine resin using 0.5 M zinc cyanide. [CWlr,,,, ::: 200 ppm. Elution
temperature of SO "C. Reprinted from Lukeyet al. (2001), Copyright (2001), courtesyof Marcel Dekker Inc.

To fully characterize the interaction between metal cyanides and ion-exchange

resins. it was necessary to determine the effect of salinity on the speciation of copper
cyanide in solution . The results presented in this chapter have shown that the equilibrium
distribution of copper-cyanide complexes changes significantly in saline solution . It was
established that the copper-cyanide speciation depends upon the salinity of the solution as
well as the CN/Cu molar ratio. whereby the predominant formation of [CU(CN)4]3- is
observed when the free cyanide concentration is not limiting.
Recent work has also shown that increasing the total ionic strength of solution during
sorption increases the selectivity of resins without affecting the loading capacity for gold
cyanide . It was proposed that the observed decrease of copper-cyanide loading in saline
solutions is due to the formation of the tetrahedral [CU(CN)4]3- complex, which requires
at least three active sites of the correct size and spatial geometry for sorption to occur.
The use of Raman and FfIR vibrational spectroscopy to determine the nature of the
sorbed copper and gold-cyanide complex onto resins has also been described. It was
shown that for a resin conta ining trimethylamine functional groups, [CU(CN)4]3- and
[Cu(CN hf- are sorbed at significant levels. However. for resins characterized by more
sterically bulky functional groups and a low ionic denisty, such as a resin containing
dipropylamine functional groups , [Cu(CNhf-, as well as appreciable amounts of the
linear [Cu(CNhr complex, sorbed. Thus. the ionic density and hydrophilicity of the resin
and the chemical structure and stereochemistry of the functional group was shown to
directly affect the type of copper-cyanide species that loads onto the ion-exchange resin.
These same factors had a signficant effect on the elution performance of saline solutions,
potassium thiocyanate and zinc-c yanide eluant s for ion-exchange resins.

The main findings detailed in this chapter, which are the effect of functional group
structure and salinity on the selectivity of ion-exchange resins, will have a significant
impact on the current understanding of the sorption behavior of functional polymers in a
wide range of applications . The spectroscopic methods developed in these works also
provide a valuable characterization technique for other sorption systems . In particular,
the work described in this chapter paves the way for future investigations into the
sorption behavior of ion-exchange resins in gold lixiviant systems of more commercial

Adams , M. D., and Fleming, C. A., 1989, The mechanism of adsorption of aurocyanide onto activated carbon ,
Metall. Trans. B 20: 315-324.
Ariti, J., and La Brooy, S. R., 1988, Effects of saline water in gold extraction, in: RandolGold Forum '88, H.
von Michaelis, ed., Randol Intemational Ltd., Perth, Australia, pp. 212-215.
Aveston, J., Everest, D. A., and Wells, R. A., 1958b, Adsorption of gold from cyanide solutions by anionic
resins, J. Chem. Soc., 231-239.
Aveston, J., Everest, D. A., Kember, N. F., and Wells, R. A., 1958a, Recovery of gold, silver and nickel from
alkaline cyanide solutions by means of weak-base ion exchange resins, 1. App. Chem. 8:77-86.
Bassedas, M., Lopez-Calahorra , F., Castells, J., and Sanz, J., 1989, Solid-phase 1JC_NMR study of bis(3-
polystyrylrhiazolin-2-ylidene), React. Poly. 10:275-278.
Bolinski, L., and Shirley, J., 1996, Russian resin-in-pulp technology, current status and recent developments, in:
Randol Gold Forum '96, H. von Michaelis, ed., Randol International Ltd., Squaw Creek , California, USA,
Burstall, F. H., and Wells, R. A., 1955, Studies on the recovery of gold from cyanide solution by ion exchange,
lon-Exch. App., 83.
Burstall, F. H., Forrest, P. 1., Kember, N. F., and Wells, R. A., 1953, Ion exchange process for recovery of gold
from cyanide solution, Ind. Eng. Chem. 45(8):1648-1658.
Chadwick, B. M., and Frankiss, S. G., 1976, Vibrational spectra and structures of some dicyanoaurate(l)
complexes , J. Mol. Struc. 31:1-9 .
Chantry, G. W., and Plane, R. A., 1960, CN stretching bands in the Raman spectra of some group Ib and group
lIb complex cyanides, J. Chem. Phys. 33(3):736-740.
Conradie , P. J., Johns, M. W., and Fowles, R. J., 1995, Elution and electrowinning of gold from gold-selective
strong-base resin, Hydrometallurgy 37:349-366.
Davison, J., Read, F. 0 ., Noakes, F. D. L., and Arden, T. V., 1961, Ion exchange for gold recovery, Trans. lnst.
Min. Metall. 70:247-290.
Eisele, J. A., 1988, Gold metallurgy-A historical perspective, Can. Metall. Quart. 27(4):287-291.
Fagan, P. A., 1998, Development of an ion chromatographic system suitable for monitoring the gold
cyanidation process. Ph.D. Thesis, The University of Tasmania , Australia .
Fleming, C. A., 1985, Novel process for recovery of gold cyanide from strong-base resins, in: Proceedings of
Extraction Metallurgy '85, Institution of Mining & Metallurgy, London, UK, pp. 757-787.
Fleming, C. A., 1991, Resin-in-pulp as an alternative process for gold recovery from cyanide leach slurries, in:
Proceedings of the 23'dCanadian MineralProcessors Conference, Ottawa, Canada, pp. 1-17 .
Fleming, C. A., 1992, Hydrometallurgy of precious metals recovery, Hydrometallurgy30:127-162.
Fleming, C. A., 1998, The potential role of anion exchange resins in the gold industry, in: EPD Congress, B.
Mishra , ed., The Minerals, Metals & Materials Society, San Antonio , TX, USA, pp. 95-117.
Fleming, C. A., and Cromberge , G., 1984a, The extraction of gold from cyanide solutions by strong and weak-
base anion-exchange resins, J. S. Afr. Inst. Min. Metall. 84(5):125-137.
Fleming, C. A., and Cromberge , G., 1984b, The elution of aurocyanide from strong- and weak-base resins, J. S.
Afr. Inst. Min. Metall. 84(9):269-280.
Fleming, C. A., and Cromberge , G., 1984c, Small-scale pilot-plant tests on the resin-in-pulp extraction of gold
from cyanide media, J. S. Afr. Inst. Min. Metall. 84(1 1):369-378.
Fleming, C. A., and Seymore, D., 1990, Golden Jubilee RIP plant - process economics and recent changes and
improvements in plant performance, in: Proceedings of Randol Gold Forum '90, H. von Michaelis, ed.,
Randol International Ltd, United States, pp. 237-241.

Fridman , I. D., Klyueva, N. D., Pochkina, L. E., and Zdorova , E. P., 1971a, An examination of the process of
regenerating anionite AP-2 when using the ion exchange technology of gold extraction, Sov. J. Non-Fer.
Met. 12(12) : 70-74.
Gasparrini , C , 1983, The minerology of gold and its significance in metal extraction, CIM. Bull. 76(851):144-
Harris, W. I., Stahlbush, 1. R., Pike, W. C , and Stevens , R. R., 1992, The extraction of gold from cyanide
solutions using moderate base polyamine ion exchange resins, React.Poly. 17:21-27.
Hazan, W. C., 1957, Method of eluting adsorbed complex cyanides of gold and silver , United States , Patent
number: 2,8/0,638.
Hla, E., Kyle, 1., and Browner , R., 1991, The effect of highly saline ground water on the gold adsorption
capacity of ion-exchange resins, in: Proceedings of World Gold '91, The Australasian Institute of Mining
and Metallurgy, Cairns, Australia , pp, 177-180.
Hla, E., Kyle, J. H., and Browner, R. E., 1992, The selectivity of ion-exchange resins for Au, Ag, Cu, Zn, and
Fe in highly saline ground water, in: Extractive Metallurgy of Gold and Base Metals, V. N. Misra, D.
Halbe, and D. J. Spottiswood, eds., The Australasian Institute of Mining and Metallurgy , Kalgoorlie,
Australia, pp. 381-385.
Hogfeldt , E., 1982, Stabilityconstants ofmetal-ioncomplexes. Part A: Inorganic ligands, Pergamon Press.
Hosking, 1. W., 1984, The recovery of gold from ores by ion exchange resins, in: Gold-Mining, Metallurgy,
and Geology, AusIMM , pp. 127-139.
Jennings, M. A., 1991, Advantages of ion exchange for gold and silver recovery , in: EPD Congress '91, D. R.
Gaskell, ed., The Minerals, Metals and Materials Society, New Orleans , Louisiana, USA, pp. 719-724.
Jones, L. H., 1953, Polarised infrared spectrum of KAu(CNh,J. Chem. Phys.21:1891.
Jones , L. H., 1957, Vibrational spectrum and structure of metal cyanide complexes in the solid state . II.
KAu(CNh, J. Chem. Phys. 27(2):468-472.
Jones, L. H., and Penneman, R. A., 1954, Infrared absorpt ion studies of aqueous complex ions: I. Cyanide
complexes of Ag(l) and Au(I) in aqueous solution and adsorbed on anion resin, J. Chem. Phys. 22:965-
970 .
Kappenstein , C; and Hugel, R., 1969, Etude des solutions aqueuses de tetracyanocuprate(l) de potassium, Rev.
Chilli. Min. 6(6):1 107-1121.
Kappenstein , C., Hugel , R., Alix, A. J. P., and Beaudoin , 1. L., 1978, Spectres raman , structures et analysis
vibrationnelles en coordonnees normales des cyanures complexes de cuivre(l) en solution aqueuse, J.
Chilli. Phys.75(4) :427-443.
Kudryk, V., and H. H. Kellogg, 1954, Mechanism and rate-controlling factors in the dissolution of gold in
cyanide solution , Journal of Metals 6(5):541-548.
La Brooy, S. R., and Muir, D. M., 1994, Gold processing with saline water, in: Proceedings of the Australasian
Institute of Mining and Metallurgy, AusIMM , 299(2) pp. 81-88.
La Brooy, S. R., Linge, H. G., and Walker , G. S., 1994, Review of gold extraction from ores , Min. Eng.
7(10) :1213-1241.
Leao, V. A., Ciminelli, V. S. T., and De Souza Costa, R., 1998, Cyanide recycling using strong-base ion-
exchange resins", Journal of Metals,71-74.
Loos-Neskovic, C., Fedoroff , M., and Gamier, E., 1989, Preparation, composition and structure of some nickel
and zinc ferrocyanides : experimental results , Talanta. 36(7):749-759.
Lukey, G. C., Van Deventer, J. S. 1., Chowdhury, R. L., and Shallcross, D. C, I999b, The effect of salinity on
the capacity and selectivity of ion exchange resins for gold cyanide, Min. Eng. 12(7) :769-785.
Lukey, G. C , Van Deventer , J. S. J., Chowdhury , R. L., Shallcross, D. C., Huntington , S. T., and Monon, C 1.,
2000a , The speciation of gold cyanide and copper cyanide complexes on ion exchange resins containing
different functional groups, React. Func. Polym. 44(2):121-143.
Lukey, G. C, Van Deventer, 1. S. J., Huntington, S. T., Chowdhury, R. L., and Shallcross , D. C, 1999a, Raman
study on the speciation of copper cyanide complexe s in highly saline solutions, Hydrometallurgy
53(3) :233-244.
Lukey, G. C; Van Deventer , J. S. J., and Shallcross, D. C; 2000b, Selective elution of copper and iron cyanide
complexes from ion exchange resins using highly saline solutions , Hydrometallurgy 56(2):217-236.
Lukey, G. C., 2000, The interaction between metal cyanides and functional polymers in saline solutions . Ph.D.
Thesis, The University of Melbourne , Australia.
Lukey, G. C., Van Deventer , 1. S. J., and Shallcross , D. C., 2001, The effect of functional group structure on the
elution of metal cyanide complexes from ion exchange resins, Sep. Sci. Technol. 35(15):2393-2413 .
Marsden, J., and House, I., 1992, The ChemistryofGold Extraction. New York, Ellis Horwood .
Nakamoto, K., 1962, Infrared Spectra of Inorganic and Coordination Compounds. New York, John Wiley and
Sons .
206 G. C. LUKEY £T AL

Penneman, R. A., and Jones, L. H., 1956, Infrared absorption studies of aqueous complex ions. 11 Cyanide
complexes of Cu(1) in aqueoussolution, J. Chem. Phys.24(2):293-296.
Pierrard, J. C., Kappenstein, C., and Hugel, R., 1971, Etude des solutions aqueses de dicyanocuprate(l) de
potassium en presence d'ionscyanure, Rev. Chim. Min. 8(1):11-19.
Reisfeld, M. J., and Jones, L. H., 1965.Raman spectraof somecyanidecomplexes of copper.J. Mol. Spectrosc.
Riveros, P. A.• 1993, Selectivity aspects of the extraction of gold from cyanide solutions with ion exchange
resins, Hydrometallurgy 33:43-58.
Riveros, P. A.• and Cooper. W. C., 1987, Ion exchange recovery of gold and silver from cyanidesolutions, in:
Proceedings of the International Symposium 011 Gold Metallurgy, R. S. Salter, D. M. Wyslouzil, and G.
W. McDonald, eds., Pergamon Press,Winnipeg, Canada, pp. 379-393.
Riveros, P. A., and Cooper, W. C., 1988. Kinetic aspects of the ion exchange of gold. silver, and base metal
cyanocomplexes, SolventExtr. lon Exch. 6(3):479-503.
Riveros, P. A., Molnar, R. E., and Mc Namara, V. M., 1993, Alternative technology to decrease the
environmental impact of gold milling-a progress report on CANMET research activities in this field.
CIM. Bull. 86(968):167-171.
Schwellnus, A. H., and Green. B. R., 1988, Structural factors influencing the extraction of gold cyanide by
weak-baseresins,in: IonExchange for Industry, M. Streat, ed., Ellis Horwood Ltd, pp. 207-218.
Seymore. D., and Fleming, C. A.• 1989, Golden Jubilee resin-in-pulp plant for gold recovery. in: Randol Gold
Forum '89, H. von Michaelis, ed., Sacremento, California, pp. 297-307.
Stamboliadis, E., McHardy, 1., and Salman, T., 1978, Ion exchange techniques for the recovery of gold from
cyanidesolutions, CIM. Bull.• 124-127.
Sussman, S.• Nachod, F. C., and Wood, W., 1945, Metal recovery by anion exchange. Ind. Eng. Chem.
Tran, T.• 1992, The hydrometallurgy of goldprocessing, Interdisciplinary Science Reviews 17(4):356-365.

Minoru Isshiki, Kouji Mimura, and Tamas Kekesi*


Recent advances in modem technology have created an increasing demand for ultra-
high purity metals. For example, ultra-high purity copper could be used as a bonding wire
and interconnector in integrated circuits or as a superconductor stabilizing material , high-
field magnet, besides its use in cryogenics and ultra-high vacuum technology. I Iron,
together with chromium and manganese, is used in the production of special magneto-
optical materials for digital recording. Development of new opto-electronic devices based
on the application of p-FeSiz 2 also demands extreme purification of the constituent
elements. The use of ultra-high purity manganese and chromium in semiconductor
technology is evolving. Ultra-pure zinc has a major significance in the properties of ZnSe
compound used as semicond uctor' materials .
Ultra -high purification of metals is also important in determining intrinsic properties
of compounds. which provides guidelines for material s development. The elimination of
trace impurities usually brings near extreme changes in the physical and chemical
properties of compounds. therefore ultra-high purity metals are new functional materials.
whose importance is strategic in future technology.
Conventional methods of metal purification via high temperature or aqueous
processing are only capable of eliminating a limited number of impurities. In order to
reach ultra-high purity , often expressed with higher than 99 .9999% concentration or
residual resistivity ratios" of the order of 10 , several purification steps are required.
Hydrometallurgical separation is a potentially efficient way of eliminating large amounts
of various impurities. Impurity metals may be eliminated to a degree from the solutions
by simple precipitation as their hydroxide or sulfide compounds. The procedure,
however, would consist of many steps , involving addition of various reagents and
requiring elaborate solid-liquid separations. Chemical similarity of many transition
metals would make the conventional purification methods even more difficult. Therefore,

Minoru lsshiki and Kouji Mimura, Institute of MUltidisciplinary Research of Advanced Materials, Tohoku
University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577. Japan. Tamas Kekesi, University of Miskolc,
Miskolc-Egyetemvaros, 3515. Hungary.

Fundamentals andApplicationsof AnionSeparations, edited by

Moyer and Singh, Kluwer AcademicIPlenumPublishers, New York, 2004 207

the efficiencyof purification and the removal of contaminants can be little guaranteed by
the conventional aqueous separation methods.
Anion exchange separation methods are superior to conventional aqueous separation
methods, as they are simple to operate and are highlyefficient.The preferred medium for
ultra-high purification is hydrochloric acid, since any traces of residual acid (chlorine)
can be efficiently eliminated during the melting of extracted metal, which is an
indispensable step for the production of bulk specimens. Extraction of the metal from
their purified chloride solution is practically possible either by direct electrowinning or
via hydrogenreduction of the crystallized chloridesalt.'
Hydrochloric acid is a favorable medium for anion exchange separations. Due to the
large differences in the formation and sorption of chloro-complex species, this technique
has the potential of eliminating all the major impuritiesfrom the solutionsof a numberof
With increasing HCI concentrations, chloride ions actively displace water molecules
in the coordinationspheresof many metal (Me n+) ions:


These chloro complex species may be denoted as

Me CIP (n- x)
P px

Eq. 2 also includesa parameterp for the polymerization of primary mononuclear species,
possibly taking place at high chloride ion concentrations. The absolute average charge,"
associatedwith one metalatom in the complex,z = In- x I.
Chlorocornplex-anions are subject to exchange with chloride counter ions of an
anion exchange resin:
MepClC~- + pzcr ~ MepCIC~- + pzcr

In Eq. 3 the superscriptbar indicates the species sorbed in the organic (resin) phase. The
concentration ratio of the metal sorbed in the resin to that in the solution is denoted as the
distribution coefficient (D), whichcan be expressedas (Eq. 4):


Transformation of the equilibrium constant of exchange reaction presented in Eq. (3)

results in the following relationship:

log D = log pK
CMe c/ pz- ) ~r~-PZ
+ log--""Cl:..-__ pzlog Ccr
p p.

'LMe rM e c/p.pz-
p "a:

Eq. 5 describes the anion exchange distribution coefficient as a function of the chloride
ion (HCI) concentration. In this chapter, we have reported the anion exchange behaviour
of a number of elements in hydrochloric acid solutions.


In order to design an anion exchange procedure for reliable elimination of virtually

all significant impurities from a solution, the characteristics of equilibrium distribution
have to be investigated . If the tendencies in chloro-complex formation and sorption are
sufficiently different for iron and the impurities, proper control of the hydrochloric acid
concentration in the eluent can achieve the desired effect of purification . Optimization of
rinsing and elution requires separation experiments to be carried out in ion exchange
columns .

2.1. Equilibrium Tests

In order to expand, correct or confirm available data, the fundamental equilibrium

distribution functions were determined by the method of batch equilibration, as shown in
Figure I. This experimental technique is also suitable in the instances of extremely high
distribution coefficients and offers higher accuracy and better control of the oxidation
states than the chromatographic band displacement method.' A strong base polystyrene-
divinylbenzene quaternary amine type of anion exchange resin (DIAION SAIOA) was
used for the separation of impurities. Ion exchange experiments were carried out at room
temperature (20 ±2 QC).
A chloride salt (in the air-stable oxidation state) of the desired element was dissolved
to prepare stock solutions of 2000 mg dm" concentration. Hydrochloric acid was also
added (to a safe level) to avoid hydrolytic precipitation. Reducing conditions, if needed,
were obtained by stirring the solution with the fine powder of a suitable metallic element
(such as Zn, Co, Fe, or Cu) in a reactor purged with nitrogen gas. Alternately, addition of
hydrogen peroxide stabilized the highest practical oxidation states of elements . Solution
samples of various HCI concentrations were prepared by mixing equal volumes of the
stock solution with varied volumes of concentrated HCI (analytical grade) and de-ionized
water (of less than 0.5 IlS em" conductivity). The solutions were, de-aerated when so
required. The concentration of the desired element (- 100 mg dm") was kept lower than
approximately I % of resin's total ion exchange capacitr to avoid saturation of the resin.
Metal solutions (30 ern') and resin samples (10 ern . in the centrifuged state) were
contacted in 100 em' plastic bottles for 1 hour in a vertically shaking machine. In order to
avoid inadvertent re-oxidation during the operation, in some cases, a small amount (0.02 -
0.03 g) of reducing agent was also added to the resin sample. Amount and type of
reducing agent which would not precipitate metal from solution, would not cause a
considerable extra load on the resin sample and would not interfere with the analysis of
the desired elements was determined by hit and trial blank experiments . The distribution
coefficients were expressed according to Eq. 6:

V:wl(Ci,.I'ol - Ce, .wl)

D=' =----'-""----
ce,.wl ce,.I'ol
110 M. ISSillKI ET AL

In Eq. 6, Ct.",. is the equilibrium concentration of the desired element in the resin phase,
C~sol is the analyzed concentrations in the initial solution, C.,sol is the analyzed
concentration in the equilibratedand separated solution, V.ol is the volume of the solution
sample and Vr.. is the wet volumeof the resin sample."
Solutions were analyzed by Inductively Coupled Plasma Atomic Emission
Spectrometry (ICP-AES), applying one standard series with 6M HCl matrix and
correcting for the effect of acid concentration by mathematical functions.! The generally
applied linear least mean squares type calibration was replaced by a forced real-zero type
model to determine low concentrations of the elements (less than - I % of the highest
standard concentration), allowing reasonably accurate trace analysis and the
determination of correspondingly high distribution coefficients. Trace concentrations
were analyzed under carefully controlled plasma and sample introduction conditions,
adjusted by a washing cycle with multiple HCl concentrations in the solution, and
verified by checking blank intensities before and after sample introduction. Computations
were assisted by a computer program (ANDIS), specifically developed for the purpose of
analyticaldata correctionand for the evaluationof the distribution coefficients.

Wet resin
1 1
a,0 Dissolution Centrifugation I
I Reduction/oxidation I
5to ck solution 1(2000 ppm) Mass measuring
(observing continuous
changes due to dehydration)
HC 1_ (controlled
a,0 redox cond.)

In itial solution
(100 ppm)
L Equilibration
(vertical shaking for 1h)
Resin sample
(Vresin =10 ern!
(V sol. = 30 crn-) ~

Correcting for:
Equilibrated solution
1 - machine s tability
I ICP analysisand data processing 1-- calibration
&'J:CI .
Figure I. Aowsheetof the anion exchange equilibrium experiments.

2.2. Separation Experiments

The general procedure of the anion exchangeseparations is summarizedin Figure 2.

According to the designed scheme of separation, the solution was prepared by dissolving
the impurities (to be tested) and the main (desired) metal as its chloride salt in an HCl
solution of adequateconcentration. If reducedconditions were required for the separation
of an impurity element, the air-free solution was reduced by stirring the powder of the
main element in a reducing reactor, flushed with nitrogen gas. Subsequently the ion
exchange column was loaded directly from the reactor, driving the solution through the
connections by increased pressure of nitrogen. In this case, loading and the first rinsing
step of ion exchange were carried out with the exclusion of air. The void space in the
resin bed had been previously filled up with the same concentration of de-aerated HCl
solution that was used for loading. In practice, the first effluent should be used for a
subsequent ion exchange separation step as loading solution after oxidation. Separations
under oxidized (or air-stable) conditions however were tested independently, by applying
freshly prepared solutions for loading, aerated and/or oxidized by the addition of
calculatedamountsof hydrogen peroxide.
When the main metal can only be fixed in the column under reduced conditions, the
rinsing steps have to be carried out under nitrogen atmosphere. In this case, subsequent
elution may be enhanced and separation sharpened by converting the main element to its
higheroxidationstate. It has been confirmed that in situ oxidation of the sorbed species in
the resin phase can be affectedby applyingcalculatedamountsof hydrogenperoxide.

(N 2l
CMe ~ 25%saturation of the bed
Starting solution CMi - 100 ppm each.
Rinsing procedure }
(with varied CHCll

HCl+Hp Samples/
(N/H 202l Collected Me-

Correction: - machine stability

- 6.cHCI
- spectral interference
- calibration method
Figure 2. Flow sheet ofthe anion exchange separation experiments.
6. , i , i , I N

C NO: no sorption (HCI cone .: 0.1 - 12M)
RTlrLJl 41-r-t
ITtJtiiJ g' Oxidation
oJ 21- S!ate(~) -I SL: slight sorption: 0.3 < 0 < 1 at 12 M HCI I>J S, p S Cl

U.l.J STR: st~ongl sorptlo7' 0 ~> 1i. I I I ,I I t

00 4 8 12 ~ ... . ~_ 0 , ~ IJ - . M
He l cone., mol-ern? ( f\ -9 . .4 f./7 \. Q "\. , ( / I~
ffiJEm (a l -
VI I((D-V) 0II1 J, IoIt I)o.t '
, I..,.,. .... ICc flU
-- -
'" -
~ ~h. , :I II G&II
r ' l[ -; IGt
- -
: ,Sob
" IAI '' ,.
-., ~~.-
J ~,
• (I
00 - (SLL(\
L - (II ', I .;'l.. \. (t . - J( II "'
, SL SL I I I Sl. 1g', . i" ";J.' II"
- (M (VV" - lU ) . :J. -
~ At.: - ~ .
( l ~ ItO :~: i~' f\ ) /1/1(. \ri:: A I 1...
J oooo~ (~l: :. ll l'-K ',~...::o:" '. . ~(j~ ~\ 1:\.~(~Th;' ..-,
,~~v, ".,,!!,!

~~~M~~ - u ----.\"i\ ~ ~/ ~ ~ p ~I V r i ~ i
U1JLtJiJIJiftJ ;H"'f, :~ ~~3 j - "I ru , .. J,,'.,,,
Figure 3. Sorption of elements from hydrochloric acid solutions on strong base an ion exchan ge resins . 1:1
(Solid lines and underlined notes refer to equilibrating tests;' broken lines and plain notes represent column elution result s,' ) 23

The resin pro ved to be stable against degradation and could be used for many cycles ,
without losing any appreciable portion of its orig inal exchange capacity or physical
integr ity. Addition of H202 to a resin bed saturated with a solution of more than 4M HCI
concentration should however be avoided because of the chlorine gas evolution causing
gas-bubble formation in the resin bed. Expansion of the resin on abruptly releasing the
main metallic species after its con version into non-adsorbing oxidation state may become
violent and could potentially result into breaking the column when the sorption band is
close to the bottom. Th is can be avoided by relax ing the bed by a short backwash before
oxidation. Similar dan ger is posed by changing the chloride form of the resin into OH-
form during conditioning.


3.1. Anion Exchange Equilibrium

Batch equilibrium experiments generated sorption (distribution) functions, which in

general confirmed the tendencies found in earlier work, using different methods,
summarized by Kraus and Nelson ." The new equilibrium results are arranged according
to the layout of the periodic table in Figure 3, showing a comparison to the data compiled
earlier. Minor discrepancies may be attributed to the slightly different resin materials,
temperatures, humidity levels, resin volume measuring and analytical methods.
Agreement in man y cases, such as Co(ll), Cu(II), Fe(II) , 2n(II), Cd(II), Pb(II) , Au(III)
and Hg(lI ) is remarkable. However, our results significantly differ from earlier data on
Fe (I II), Mo(Y ) W(YI), Y(lY ), Y(Y), Cr(lll), Mn(II) and Te(lY). In the cases of Cu(l ),
Ag(l ), Pd(II ), Pt(lY ), Sn(II), Sb(III), Sb(Y) and Bi(lII ), available data have been extended
or corrected in the lowest HCI range . There are also some new functions revea ling the
behavior of Mo(III), Mo(YI ), W(III ), W(Y), Cr(lI) and Y(IY) , which was missing in
previous report s. Distribution coefficients of Mn(II), Cr(lll), Y(lY ) and Y(Y), referr ing to
the I hour equilibration method , did not significantly exceed unity in the examined 0-10
M HCl range, therefore this data have not been plotted. Figure 4 presents the new results
(at a larger scale ) of those examined cases where significant sorption has been found . in
view of the required density difference between the resin and the solution Higher HCl
concentrations are not usable in applied methods due to preferential absorption by the
resin sample. 6 In view of the required density difference between the resin and the
solution, high HCl concentration range is also not very practical.
The feasibility of reduct ion, oxidation and complexation was estimated on the basis
of available thermodynamic data lO•1I and with the application of computer software
(ROCC) developed specifically for this purpose.f To interpret the results of anion
exchange distribution function s obtained in this work, associated changes occurring in the
structure of ionic species were also determined using spectrophotometric analysis.
Characteristic peaks of the different oxidation states of metallic species, established by
preliminary work, were used to observe the changes in the oxidation states of ions. As an
example , spectral evidence of the formation of the trivalent iron during the air-oxidation
of a divalent iron solution is illustrated in Figure 5. The initial Fe(II) spectrum has only
one peak at -253 nm, while after oxidation formation of a triplet (at -24113111363 nm)
characteristic of Fe(III ) in 8-10 M HCl solut ions is obtained. These spectra also showed
reducing efficien cy of Fe powder as well as relative abundance of oxidized species .

3-<- - - - - - - - - - - - - - - -,..

"'"G a(II I) ':"

~ 0


0 2
" 6 8 10



o 2 " 6 8 0 0.5
Hel concentration (mol dm -3 )

Figure 4. Anion exchangedistribution of the elements in differentoxidation states.

Changes in absorption spectra with the variation in HCI concentration could be

interpreted in terms of distribution functions. When comparing distribution curves with
the corresponding series of spectra, it is necessary to take the regular shape of the
sorption functions, as predicted by Eq. (5) into consideration. The rising section in the
distribution curves is indicative of the formation of a sorbable complex species. This is
also reflected by the changes in corresponding spectra. As the HCI concentration is
further increased, at certain HCI concentration the complex-ion peaks are fully
developed. Their peak-position becomes stable in the spectrum, indicating no more
change in the structure of ionic species with the increase in HCI concentration. At the

same time the anion exchange distribution coefficient starts declining due to the
competing effect of excess cr ions. Trivalent iron, divalent copper and divalent cobalt
are important examples, which clearly demonstrated the relationship between their
observed light absorption spectra and distribution coefficients during incremental
changes in HCI concentration. The sorption functions of these three metal ions are
presented in Figure 4 while their corresponding light absorption spectra, which
demonstrated the formation of the sorbable chloro-complex species, are depicted in
Figures 6-8. The apparent average charges of the predominantly sorbed species of Cu(l),
Zn(II), Pd(II) , Pt(lV) and Cd(II) were estimated from Eq. 5 by applying the distribution
functions and the experimentally determined chloride ion activity coefficients."

o- Fe(lI) (stabilized with Fe powder as reductant),
1· Oxidized with air (2 min),
2 . Oxidized with air (4 min),
3 . Oxidized with air (6 min) .


200 300 400 500

Wavelength, nm
Figure 5. The conversion of Fe(H) into Fe(I1I) in - 8 M HCI solutions detected in the photometric cell.

A comparison of sorption functions of practically feasible different oxidation states

of Cu, Fe, Mo, W, and Sb suggests the vital importance of controlling oxidation states
during their anion exchange separation. The differences in the anion exchange behavior
of different oxidation states of these element s can be applied in developing new processes
for the ultra-high purification of transition metals.

3.2. Anion Exchange Separation Examples

In general . the HCI concentration and the oxidation states of elements are the two
essential parameters to be optimized in a separation procedure.

Fe(III) 10 mg dm
o - 0.005M HCI,
1 - 1.00 M HCI,
2 - 1.95 M HCI,
3 - 2.90 M HCI,
4 - 3.85 M HCI,
~ S - 4.80 M HCI,
6 - 5.75 M HCI,
7 - 6.70 M HCI,
II) S - 7.65 M HCI,
9 - 8.60 M HCI,
~ 10 - 9.55 M HCI,
11 - 11.0 M HCI.

o -+---.----,----r===--=;r===~ .......--1
200 300 400 500
Wavelength, nm
Figure 6. The effectof HCI concentration on the Fe(lII)absorption spectrum.

2--n__- - - - - - - - - - - - - - - - - - - ,

CU(II) 25 mg dm-3

o • 0.01 M HCI,
1 . 1.00 M HCI,
2 - 1.99 M HCI,
3 - 2.98 M HCI,
4 - 3.97 M HCI.
S • 4.96 M HCI,
6 • 5.95 M HCI,
7 - 6.94 M HCI,
S - 7.93 M HCI,
9 - 8.92 M HCI,
10 - 9.91 M HCI,
11 • 11.5 M HCI.

200 300 400 500

Wavelength, nm
Figure 7. The effectof HCI concentration on the Cu(lI)absorption spectrum.

0.06 --r-----------------,,-....,----,
Co(lI) 200 mg dm,3

o - 0.02 M HCI,
1 - 1.00 M HCI,
2 - 1.98 M HCI,
3 - 2.96 M HCI,
0.04 4 - 3.94 M HCI,
~ s - 4.92 M HCI,
c: 6 - 5.90 M HCI,

Cll 7 - 6.88 M HCI,
a - 7.86 M HCI,
til 9 - 8.84 M HCI,
« 10 - 9.82 M HCI,
11 - 11.4 M HCI.

450 500 550 600

Wavelength, nm
Figure 8a. The effect of He) concentration on the Co(ll) absorpt ion spectrum (Range, I )

Co(lI) 200 mgdm,3

S - 4.92 M HCI,
6 - 5.90 M HCI,
7 - 6.88 M HCI,
a - 7.86 M HCI,
9 - 8.84 M HCI,
~ 10 - 9.82 M HCI,
c: 11 - 11.4 M HCI.


550 600 650 700 750

Wavelength, nm
Figure 8b . The effect of HC) concentration on the Co(ll) absorption spectrum (Range-ll)

Based on the sorption functions presented in Figures 3 and 4, procedures for the
purification of Cu, Zn, Co and Fe have been designed. Their elution curves have always
been found in agreement with their anion exchange distribution curves presented in
Figure 4.
A large number of elements do not form sorbable anionic species in chloride
solutions. Therefore, strong sorption of the main element is indispensable at one point of
the separation procedure if overall purification is the aim. The essential steps of the tested
processes are described in Table I. Efficiency of separation is dictated by the elution
curves of critical impurity elements of practical importance or the impurity elements
separated together as a large group. Most of the impurities are separated by the rinsing
steps (while the main element is in the resin). However, some strongly sorbed elements
are retained during the main elution step (elution of product). These impurities may
accumulate unless they are removed by the conditioning steps executed before each re-
use of the resin.

Table 1. Established Anion Exchange Separations.

Purification steps Groups of separated or

(HCI concentration in mol drn" (partly separated)
- redox conditions) impurities
Passing (-1.5 - Red.) Class-Il + Cu+Ag+Zn+Cd+(In)+Pb+Sn+Sb+(Te)
Loading+ Rinsing (-5 - Ox.) Class-I + Zr+Hf+Nb+Ta+(Pa)+W+Mn+Co+(Cu)+(In)+
Elution (-I - Ox.) Class-Il + Mo+U+Ag+Zn+Cd+Sn+Pb
Remainingon the resin (Pa)+Ga+(In)
Passing(-1.5 - Red.) Class-Il +(Mo)+Cu+Ag+Zn+Cd +(In)+Pb+Sn+Sb
Loading+ Rinsing (-9 - Ox.) Class-I + (Zr)+(Hf)+Mn+(Cu)+Ag+(In)+Pb+As
Elution (-5 - Ox.) Class-Il + Mo+U+Fe+Zn+Cd+Ga+(In)+Sn+Sb
Remainingon the resin Nb+ Ta+Pa+W+Ge+Se
Loading + Rinsing (-4 - Red.) Class-I + Zr+Hf+Nb+Ta+Pa+Mn+Co+(Ag)+Ge+As+Se
Rinsing (0.001-0.05 -Red.) Group-A + W+U+Fe+Zn+Cd + Ga+ln+Pb+(Sb)+Te
Elution (-4 - Ox.) Class-Il + Mo+Sn+Sb
Passing (-1.5 - Ox.) Class-II + Mo+Ag+Zn+Cd+Pb+Sn
Loading + Rinsing (-5.5 - Ox.) Class-I + Nb+Ta+Pa+Mn+Co+(Cu)+Ge+Pb+As+Se
Rinsing (-I - Ox.) Group-B + W+U+Fe+Co+Cu+Ga+ln+TI+Sb+Te
Elution (-0.2 - Ox.) Class-Il + Nb+Ta+Mo+Ag+Cd
Red. Reducedwiththe pure powderof the main metal.
Ox. Oxidizedwith H,O".
Class-I Not sorbed (alkali, alkaline earth, rare earth, Sc,Y,La,Ac,Ti"Th,V,Cr,Ni, first two rows of the P:
Class-II Stronglysorbed-underoxidizedconditions- rre. Re, Ru, Rh, Pd, os. lr, Pt, Au. Hg, T1. Bi, Po).

Separation schemes devised for the purification of copper and iron are presented in
Figures 9 and 10, respectively; the most important elution curves for the two stages of the
processes are depicted. In copper purification procedure.i!' reduction of the dissolved
copper chloride with Cu powder assured strong sorption of copper in the monovalent
state when the solution was fed into the anion exchange column under reducing
conditions. Conversely, in situ oxidation by the controlled addition of H202 was
necessary to remove c(e1ute) opper from the resin bed because of the CuCI precipitation
at low HCl concentrations and the high distribution coefficients of Cu(1) in high HCl
In case of iron purification, initial reduction of the impurity Cu(II) in the starting
solution with Fe powder is accompanied by precipitation of cemented copper. The
residual Cu content of the solution remained in the monovalent state and was separated
by anion exchange sorption. The amount of metallic Cu deposited on the Fe particles is
typically very low and is helpful in stabilizing the Cu(l) state." Application of H202
(-0,3-0,5 %) in the Iron-2 step was indispensable to assure the hexavalent state of
molybdenum, thereby avoiding leakage in the slightly sorbable Mo(III) state previously
stabilized during the treatment of the solution with Fe powder.
During optimization of the separation procedure, the resulted elution curves were
numerically integrated to express the purification ratio of an impurity element (Eq. 7) and
the yield of the main element (Eq, 8):


y = 100% rec v, m

In these Eqs., V, and Ve are the effluent volumes at the starting and ending points of the
purified solution collection, respectively. The loaded amount of the impurity is Em, and
that of the main element is• The amounts collected in the purified solution are m, and
mm, as determined from the corresponding concentrations (c, and cm)' Computed ratios of
purification may refer to different intervals of effluent collection, i.e., different widths (Ve
- Vs ) of the principal elution peak:

cm ~ c m•max Ik

Standard values (PRj •Sld , Y,td) corresponding to the volume range defined by the half-
height concentrations of the main element (k = 2) are presented in Table 2 for Copper-I,
2 and Iron-I, 2 processes. The characteristic values were computed with a computer
program (ANEL) specifically developed for the analysis of elution test results.

(S n) Zn Cd Max. cone . ,ppm Cu Sn

Cu: 10000
Cd : 60
Sn: 18 0
c: Zn : 150
~ d
'E I
'C Sb"
~ r~
'E :;,
~ I Bi
o I ¢

o 1000 2000 3000 4000 5000 6000 7000

Volume of effluent, em 3

e --- Cu
- .... -,- - Pb A9 Max . conc. pp

c: Cu : 411 2
Aq: 31
Pb : 83
X 0.5
c: 2M NaOH
0 10 20 30 40 50
Volume of effluent, dm 3

Figure9b. Elution curves of the"Copper-2"'separation process


c: Ma x . conc .ppm
X 0.5

o 500 1000
Volume of effluent, cm3
Figure lOa. Elution curves of the "Iron-I " separation process .

Max . conc . ,ppm

Fe : 2 668 1
Mn : 110
Mo : 17.2
P : 99 . 2
Si : 1. 05
c: Ti : 15 6
.2 Fe Mo
0 0
X Max. conc .,ppm

~ Fe : 26681
Ai : 171
c: Co : 86 .2
0 cr : 112
lU Zn : 11 6

o 500 1000
Volume of effluent, ern"
Figure lOb. Elution curves of the "Iron-Z" separation process .
222 M.ISSffiKI ET AL

Table 2. Efficiency Characteristics of Anion Exchange Separations.

Process YSld (% ) Standardpurification ratios of elements, PR'Sld

Ag Cd In Pb Sb Bi.Fe,Hg,Zn,Sn
Copper-I 70 2 840 >100 46 >600 >1000
Copper-Z 92 >1000

Cu Co Si Mo Al,Cr,Mn,P,Ti,Zn
Iron-I 87 >1000
lron-Z 55 50 -1000 >1000 >1000

Some of the first row transmon metals do not lend themselves to efficient
purification by anion exchange in HCl solutions. Nickel is not sorbed, allowing the
separation of only those impuritiesthat show considerable distribution coefficients at any
HCl concentration. Manganese, chromium and vanadium have been found to form higher
than neutral chloro complexes only in the highest possible HCl concentrations and the
complex formation process may take several hours, much longer than typically
experienced in most other cases. A combination of cation and anion exchange steps for
the ultra-high purification of these metals is underconsideration.


Batch equilibrium experiments under controlled reducing/oxidizing conditions have

revealedsome new results for the anion exchange distribution of a numberof elements in
hydrochloric acid solutions. A systematic investigation of equilibrium behavior has led to
the development of several anion exchange separation procedures. The diversity in the
formation and sorption characteristics of chloro-complex species has enabled the
separationof chemicallysimilar elements and devising the separation scheme of virtually
all the significant impurities from several metals. Reduction of the starting solutions with
the pure powder of the metal to be purified proved to be an indispensable technique to
separate certain impurities. A subsequent anion exchange step under oxidized conditions
may result in the separation of the rest of the elements. In situ oxidation by the addition
of controlled amounts of hydrogen peroxide has been found technically feasible at
moderate concentrations of HCI in the eluent. The variation in the HCI concentration in
influent solution was used as a major tool for affecting the desired separations. Special
indexesderived from the numerical analysisof the experimental elution curves have been
proposed and applied for the optimization of the separation procedures. The results have
demonstrated the feasibility of anion exchange separation methods for the preparation of
ultra-high purity Fe, Co, Cu, and Zn. Expansion of these investigations to the purification
of other elements is currently under investigation. Extraction of the metals from purified
solution is possible by electrowinnig or hydrogen reduction after evaporation to dryness.


I. Y. Miyake. Recentaspectsof manufacturing methods and applications of superpure copperfor ndustrial use,
Bulletin Japan ITw. Metals 31, 267-276 (1992).
2. M. Powalla and K. Hen , Co-evaporated thin films of semiconducting ~-FeSi" Appl. Surf Sci, 65166, 482-48
3. M. Isshiki, Photoluminescence and cyclotron resonance studies on highly purified ZnSe single crystals, J.
Crystal Growth 86, 615-624 (1988).
4. K. Mimura et al, Precisepurity evaluation of high purity copper by residual resistivity ratio, Mater. Trans.
JIM, 38, 714-718 (1997).
5. T. Kekesi et aI, Preparation of ultra high purity copper by anion exchange, Met. Mater. Trans. B, 28B, 987-
993 (1997).
6. T. Kekesi and M. Isshiki, Anion exchange behaviorof copperand some metallic impurities in HCI solutions,
Mater. Trans. JIM, 35, 406-41 3 (1994).
7. H. Small: PlenumPress, NewYork, London,1989. pp.I-39.
8. T. Kekesi, K. Mimura, M. Uchikoshi and M. Isshiki: Bulletin of the Institute for Advanced Materials
Processing, Tohoku University SOZAIKEN IHO, 56, 9-16 (2000).
9. K.A. Kraus and F. Nelson, Anion exchange studiesof the fission products, Proc. Int. Cant Peaceful Uses of
Atomic Energy, (Geneva) 7,113-1251956).
10. L.G. Sillen and A.E. Martell, Stability Constants ofMetal-Ion Complexes (The Chemical Soc., London, No.
17,1964) pp. 2-301.
II . J.F. Zemaitis et al., Handbook of Aqueous Electrolyte Thermodynamics (American Inst.Chem. Eng. Inc.,
NewYork, NY, 1983) p. 101.
12. T. Kekesi , M. Uchikoshi, K. Mimura, Y. Zhuand M.lsshiki, Proc.Lnternational Conference on the
Processing Materialsfor Properties, PMP2000, 4-8 Nov. 2000, San Francisco, Ca, USA, pp. 979-984.
13. T. Kekesi and M. Isshiki, Ultra high purification of copper chloride solutions by anion exchange,"
Hydrometallurgy 45, 345-361 (1997).

Eric Guibal, Javier Guzman, Ricardo Navarro, Montserrat Ruiz, and

Ana Sastre'


The increasing enforcement of environmental regulations, as well as the need to

recycle metal from wastes and industrial processes, has led to the development of new
hydrometallurgical processes. Derived from conventional processes, the current ones use
new conditioning procedures (e.g., solvent-impregnated resins, membrane processes for
liquid/liquid extraction) or new materials (including materials of biological origin)
(Brooks, 1991). In the past two decades, biohydrometallurgy has focused on leaching
processes (Ciminelli and Garcia, 2001); however, more recent studies have been directed
toward biosorption. A wide range of materials has been investigated for the recovery of
various metals ions, especially from dilute solutions . Preliminary investigations have
focused on the use of fungal, algal or bacterial biomass (Guibal et al., 1992; Volesky and
Holan, 1995; Veglio et al., 1997); however, more recently intensive research has been
carried out on biopolymers for the uptake of toxic or valuable metals. Alginates
(extracted from algal biomass) (Coughlin et al., 1990; Chen et al., 1997; Konishi et al.,
1998) and chitosan, a partially deacetylated form of chitin, the most abundant biopolymer
after cellulose , have received a great deal of attention (Findon et al., 1993; Kawamura et
al., 1993; Guibal et aI., 1995a,b; Inoue, 1998; Piron et al., 1998a; Bassi et al., 2000).
These materials are of particular interest because they are widely available in nature and
are characterized by a large number of reactive sites. Chitosan's greater reactivity enables
this amino-polysaccharide to sorb metal ions through several mechanisms, including
metal chelation of cations on the free electronic doublet of nitrogen (a weak electron
donor, able to react with soft-acid metal cations), and ion exchange for metal anions
(Roberts, 1992). This electrostatic attraction capability is also responsible for the sorption

Eric Guibal, Ecole des Mines d' Ales, L.G.E.I., 6 avenue de Clavieres, F-30319 Ales cedex, France, Javier
Guzman, Ricardo Navarro, Universidad de Guanajuato, I.I.C., Cerro de la Venada, Guanajuato, 36000, Gto.,
Mexico. Montserrat Ruiz, Ana Sastre, Universitat Politecnica de Catalunya, E.T.S.E.I.B., Diagonal 647,
E-08028 Barcelona, Spain. Corresponding author : E.G., Phone : +33 (0)466782734, Fax . +33
(0)466782701, E-mail Eric.Guibal@emafr

FIUIIlmnoItDIs andApplications of AnionSeparations, edited by

Moyer and Singh, Kluwer Academic/Plenum Publishers, NewYork, 2004 215

of mineral and organic acids (Yoshida et aI., 1994; 1995), and organic compounds such
as dyes (Yoshida et aI., 1993; Guibal et al., 2001). In acidic solutions, chitosan is
protonated, a unique property among biopolymers. The pK. of chitosan is dependent
upon its degree of deacetylation and neutralization, and varies between 6.1 and 6.8
(Sorlier et al., 2001). Consequently, metal anions can be electrostatically attracted to
protonated amine groups and exchanged with the counter ions on the biopolymer.
Chitosan is soluble in organicand mineral acids, with the exception of sulfuricacid, and
it may be necessary to modify the polymer usinga cross-linking treatment to reinforce its
stability in acidic solutions. For example, glutaraldehyde has been used for chitosan
cross-linking: a Schiffs base reaction between aldehyde groups on the cross-linking
agent and aminegroupson the biopolymer leads to the formation of iminefunctions and
new linkages between the chains of the polymer. However, this treatment results in a
decrease in the number of free reactive aminegroupswhich is especially important in the
case of chelating mechanisms, since a reduction of amine-group density reduces sorption
efficiency. In some cases metal sorption is completely inhibited (Dzul Erosaet aI., 2001).
Although, the steric hindrance aroundthe reactive protonated amine groups may cause a
decrease in sorption efficiency. This inhibitory effect is less significant in the case of ion-
exchange mechanisms (Milotet al., 1998, Ruizet al., 2000).
Chitosan and someof its derivatives have been widely used for the sorption of metal
ions. Indeed, an interesting property of chitosan is its chemical and physical versatility.
The chemical versatility of chitosan has led to the preparation of many chitosan
derivatives with improved sorption properties. The grafting of new functional groups
increases the sorption capacities or the selectivity in the sorption process (Inoue, 1998;
Inoueet al., 1999). For example, the graftingof thiourea on the chitosan backbone brings
to the polymer new chelating functions and decreases the sensitivity of the sorption
process to environmental parameters, especially to the presence of competitor or co-ions.
Chitosan membranes and fibers have also been used for the sorption of metal ions and
dyes (Modrzejewska and Kaminski, 1999; Yoshida et aI., 1993). Recently, hollow
chitosan fiber has also been used for the extraction of chromate ions has been
accomplished by the use of hollow chitosan fiber using the fiber for the sorption of the
metal anions in acidic solutions, with quaternary ammonium salt (Aliquat 336) flow ing
into the lumen of the fiber, resulting in the simultaneous desorption/stripping of the metal
(Vincent and Guibal, 200I).
Due to chitosan's low porosity, intraparticle diffusion control often causes kinetic
restrictions in sorption. The polymer can be physically modified by the preparation of
chitosan gel beads to improve diffusion properties (Guibal et al., 1998). This procedure
consists of dissolving the chitosan in an acetic acid solution and coagulating small drops
of the viscous solution in an alkaline coagulation solution. The dissolving procedure
reduces the residual crystallinity of the polymer and the coagulation step maintains the
expanded structure. This treatment enhances the accessibility of metal ions to internal
sorption sites, especially for large metal ions (e.g., polynuclear forms) in cross-linked
materials (Milotet aI., 1998).
Despite the large number of studies focusing on the sorption of metal ions on
biosorbents and chitosan derivatives, the influence of metal chemistry and, more
specifically of metal speciation on sorption properties has been frequently ignored. Only
a few studies have recently addressed this issue (Guibal et al., 1994; 1995a,b; Juang and
Ju, 1997; Saba et aI., 1998; Lasko and Hurst, 1999; Wu et aI., 1999; Remoudaki et aI.,
1999). The objective of the present work is to illustrate the influence of metal-ion

speciation on the ion-exchange process for the uptake of metal anions, such as molybdate
(Guibal et al., 2000a), vanadate (Guzman et al., 2001). platinum-group metal anions
(PGMs) (Guibal et al., 2000b), copper (Juang and Ju, 1997), and silver-chelated anions
(Lasko and Hurst, 1999). The type of sorption mechanism is frequently controlled by
metal speciation and pH. Changing the pH of the solution may results in a change in the
mechanism involved in metal uptake: copper sorption and silver uptake on chitosan
derivativeswill be discussed (Juang and Ju, 1999; Lasko and Hurst, 1999,respectively).


2.1. Materials

Chitosan was supplied by Aber Technologies (France). The biopolymer was

characterized by its molecular weight MWn : 125,000 g mol" (using size exclusion
chromatography coupled with light-scattering and refractometer equipment), and its
deacetylation percentage DO: 87% (using FTIR facilities). As a consequence, the
molecularweight of the equivalent monomer unit was approximately 166 g mol". The N
content was then 6 mmol N g", and the free amine content was 5.2 mol -NH 2 g'l. The
chitosan flakes were ground and sieved and the smallest-size fraction was collected for
experiments (e.g.. GI , d, < 125 urn, dp: particle diameter). The pK. of the chitosan used
in these experiments was 6.2. Figure I demonstrates the effect of pH on the chitosan's
Ammonium heptamolybdate, ammonium vanadate, glutaraldehyde (50% w/w), and
thiourea were supplied by Fluka (Switzerland) as analytical grade products;
hexachloroplatinic acid and sodium tetrachloropalladate salts were supplied by Acros
(U.S.A.). Other common reagents (sulfuric acid, sodium chloride, sodium sulfate,
hydrochloric acid. and sodium hydroxide) were supplied by Carlo Erba (Italy).


R-NH, '

a ~-------=--------'

0.5 25 4.5 6.5 8.5


Figure I. Protonation of chitosanas a function of the pH.

2.2. Chitosan Modification

When required (e.g., sorption experiments carried out in hydrochloric acid

solutions). cross-linking was performed by mixing known amounts of chitosan and

glutaraldehyde solutions for 16 hours (Rorrer et al., 1993; Hsien and Rorrer, 1997). The
slurry was then filtered and rinsed thoroughly until the typical smell of excess
glutaraldehyde disappeared . Finally, the product was dried at room temperature. This
common procedure involved the reaction of glutaraldehyde and chitosan on the basis of
equimolar concentrations of aldehyde (from the di-aldehyde cross-l inking agent) and
amine (from the biopolymer).
Chitosan gel beads were prepared by dissolving the polymer in an acetic acid
solution, mainta ining an excess of acid compared to the concentrat ion of amine groups.
The chitosan concentration was 4% (w/w) . The solution was filtered to remove insoluble
particles and then left to stand at room temperature for 3 days. The degassed viscous
solution was then pumped and distributed as small drops through a thin nozzle into a
neutralizing and coagulat ing alkaline bath (NaOH, 1-2 M). After 24 hours of contact, the
beads were collected and thoroughly rinsed until a neutral pH was obtained .
Thiourea grafting was performed through a preliminary reaction with glutaraldehyde
(Schiff's base reaction between aldehyde functions and amine groups on the thiourea) for
8 hours, and then a reaction with chitosan for another 16 hours. The free aldehyde groups
of the cross-l inking agent acted then as links between the intermediary product and the
biopolymer; the intermediary product was linked onto the polymer backbone through a
second Schiff's base reaction. The polymer was then filtered, thoroughly rinsed, and
dried at room temperature. The reagents consisted of 3 g of thiourea and 2 mL of
glutaraldehyde for I g of chitosan.

2.3. Methods

Sorption experiments were performed with controlled pH. Micro volumes of molar
solutions of sulfuric acid (or hydrochloric acid, depending on the experiment) and sodium
hydroxide were used for the continuous control of the pH during the sorption process.
For known sorption isotherms, varying amounts of sorbent (m.) were mixed with
fixed volumes (Vi) of metal-containing solutions at different fixed concentrat ions (Co,,)
for 3 to 5 days of contact , depend ing on the time required to reach equilibrium. Samples
were then filtered using a 1.2 urn pore-size membrane and the filtrate was analyzed for
metal content (Ceq,,) using ICP-AES facilities (Jobin Yvon, JY-36 or JY 2000,
Longjurneau , France) or spectrophotometric measurements (platinum colorimetric
method with the stannous chlor idelHCI method) (Charlot, 1978), depending on the metal
and the concentration range. Sorption capacities (q, mg metal g", or mmol metal s'')
were then calculated using the mass balance equation : q = (Co,;- Ceq,,)V/m,.

2.4. Distribution of Metal Ion Species

The distr ibution of metal species was calculated using Hydraql software (Papelis et
al., 1988) and the equilibrium constants reported by Baes and Mesmer (1976) , in order to
obtain the distribution of the metal species as a function of the pH and the total metal
concentrat ion (and in some cases as a function of the presence of chloride ions). It was
particularly important to determine the predominance of selected species, and to correlate
the sorption performance with the presence of adsorbable species. This correlation was
used to confirm the affinity of the sorbent for selected species.


Molybdate ions are strongly sorbed onto chitosan (Guibal et al., 1998). The optimum
pH was found to be approximately 3. In agreement with the hard and soft acids and bases
theory, the functional groups containing N donor atoms strongly interact with soft acids;
chitosan .can sorb metal ions (especially cations) through a chelation mechanism .
However, the main interaction mechanism in the present case was due to the electrostatic
attraction/anion exchange between the metal anions and the protonated amine functions.
Below pH 5.5, more than 90% of amine functions are protonated and thus available for
metal anion sorption (Fig. I). Figure 2 shows the influence of pH on the shape of the
sorption isotherms.

3.1. Effect of pH

The optimum pH range is close to pH 3 because the sorption isotherms overlap

almost completely at pH 3 and pH 4, and these curves are characterized by a high
maximum sorption capacity (on the plateau) and also a steep gradient for very low
solution concentration values. The initial slope characterizes the affinity of the sorbent
for the solute. The maximum sorption capacity tends to be 750-800 mg Mo s" (7.8-8.3
mmol Mo g"). Taking into consideration the sorbent's degree of deacetylation (ca, 87%),
the nitrogen content of the sorbent is 6 mmol N s:'. which corresponds to an amine
content of 5.24 mmol -NH 2 g", The molar ratio between the metal and the protonated
amine groups is thus greater than I for the most favorable pH conditions (ca. 1.5 mmol
Mo mrnol' -NH 2) .


• •
a • ee • pH I
• 0
0 •

.s ,
o pH 2
D 0 o
.a.o 0 • pH 3
:; D pH 4

0 • pH 5
.... 300 o pH 6
00 • pH 7
6 pH 8

80 120 160 200

C ... (mglL)

Figure 2. Influence of pH on the sorption of molybdate ions on chitosan gel beads (pH controlled with either
sodium hydroxide or sulfuric acid). (Reprinted from Guibal et at. 2000a. p. 1027. by courtesy of Marcel
Dekker, Inc.)

Taking into account that all the sites cannot be saturated due to steric hindrance and
electrostatic repulsion, the molar ratio between the metal ions and the sorption sites is

certainly greater than 1.5. This suggests that the molybdate sorption on chitosan involves
the participation of other functional groups (hydroxyl groups for example), or the
sorption of polynuclear metal species [Mo702l(OHf" for example].
It can also be seen from Figure 2 that at pH 2 and pH 5, the maximum sorption
capacities are comparable to the levels reached at pH 3 and pH 4. However, the initial
slope of the isotherm is rather flat at very low equilibrium solution concentrat ions. This
is an unfavorable trend because sorption capacities remain low (below 50 mg Mo g") up
to a residual concentration for which the sorption capacity increases strongly. Above this
limit concentration, the sorption isotherms become favorable. A similar trend was
observed at pH 6, though the maximum sorption capacity (in the experimental
concentration range) was significantly lower than the levels reached at more acidic pH
values. The residual concentration at which the sorption becomes favorable changes with
the pH; increasing the pH results in an increase in the residual concentration of the metal
in the liquid phase.
At pH 1, pH 7 and pH 8, the sorption capacities remain very low, below 50-80 mg
Mo g" . At pH I, the decrease in sorption properties is due to the excess of counter ions
brought by the acid used for pH control. This explains the strong competition of these
anions for the sorption on protonated amine groups. In the case of pH 7 and pH 8, though
some amine groups remain protonated (especially at pH 7; about 40% of amine groups
are protonated), the polymer is significantly less protonated and the sorption capacity
decreases strongly .
The protonation of the polymer is thus a key parameter; more specifically, the
electrostatic balance appears to be an important criteria However, it also appears
necessary to take into account other parameters such as the chemistry of metal ions. The
molar ratio between molybdate and protonated amine groups exceeds I by a significant
amount, as stated above. Therefore, molybdate uptake might result from the sorption of
polynuclear species. The shape of the sorption isotherms is another confirmation of the
necessity to take into account metal chemistry. Indeed, at pH 2 and pH 4, amine groups
remain protonated; thus the sorbent is expected to remove molybdate ions with a degree
of efficiency similar to that obtained at pH 3 and pH 4. However, their sorption isotherms
were comparable only to those at pH 3 and pH 4 at high residual concentration . At low
residual concentration, the sorption capacity remained low despite the protonation of the

3.2. Molybdenum Species

Figure 3 illustrates the distribution of molybdate species as a function of pH and total

metal concentration. A number of molybdate species may co-exist in the solution, and the
actual distribution of these species depends strongly on these experimental parameters.
Molybdenum can be found under the form of non-ionic species or anionic species.
However, among these anionic species, there is a great diversity; both mononuclear and
polynuclear species (hydrolyzed or non-hydrolyzed) have been identified.
Previous observations have shown that the high level of maximum sorption
capacities requires the presence of polynuclear molybdate species (Guibal et al., 2000a).
The low sorption capacities, observed at low residual concentration at pH 2, pH 5 and pH
6, may be related to a change in the distribut ion of metal species under these
experimental conditions. By increasing the pH, the predominance of molybdate species is
displaced toward the formation of species that have a greater affinity for the sorbent due

to ionic charge and molecular size. The sorption capacities increase with the appearance
of species more favorable for sorption on chitosan. The competition of counter-ions
(brought by pH control) may be suggested as an explanation for the decrease in sorption
properties, but Figure 3 also shows that at pH I, especially at low residualconcentrations,
the only significant metal species are non-ionic species (H2M04 ) that are less adsorbable
on protonated amine groups.

100 100
IMol=I mM (a)
75 75

c~ ~
.~ 50

ell 25 ell 25

0 0
2 3 4 5 6 1 2 3 4 5 6
pH pH

Figure 3. Distribution of molybdate species as a function of pH - Influence of total metal concentration [(a):
MoO/; (b): HMoOi ; (c): H,MoO.; (d): M07O,:-;. (e): Mo,o,,(OH)s-; (f) : Mo,o,,(OHj,4-; (g) : M0702I(OH),'-;
(h): Mo,o,oOH)/j .

Several attempts have been made to correlate the shape of the sorption isotherms,
relative to the increase of the sorption capacity at a fixed residual metal concentration,
with the presenceof differentmolybdate species. The best correlationwas obtained when
the shape of the sorption isotherm curves was correlated to the concentration of
polynuclear molybdate species. Figure 4 shows the sorption isotherm curves compared to
the distribution curves of mono- and polynuclear molybdate species, at pH 3, pH 4 and
pH 5. It appears that the increase in the sorption capacity corresponds to the appearance
of polynuclear molybdate species. It is expected that polynuclear species would have a
greater affinity for the sorbent than mononuclear species. Nekovar and Schrotterova
(1999, 2000) also observed that tungsten, vanadium, and molybdenum polynuclear
species are preferentially extracted by the primary amine Primene JMT. The interactions
between protonated amine groups and polynuclear species involve multiple linkages
between several units of the polymer and a single metal molecule, which contributes to
the cross-linking phenomena observed in the coagulation of chitosan in molybdate
solutions(Dambieset aI., 2002).
In order to verify the importance of these polynuclear species (together with the
protonation of the polymer), the sorption isotherms have been plotted as a function of the
actual concentration of the polynuclear species, instead of the total metal concentration,
using the Langmuirequation. For the concentrations below the limit concentration for the
appearance of polynuclear species, the sorption capacities have been set at O. Figure 5
shows the simulation of sorption isotherms according to this model. A quite good
approximation of the experimental data was obtained. It appears that under the conditions
of chitosan protonation, molybdate sorption occurs through anion exchange/electrostatic
attraction when the metal is present under a suitable polynuclear anionic form. This is
important since it makes it possible to predict the experimental conditions favorable for

the sorption of molybdate. Giventhe total concentration, the optimum pH for the sorption
of molybdate can be obtained by calculating the distribution of metal species and
determining the pH for the predominance of polynuclear species.
4 r -,----- - - -----r 10 10

8 ... ...
.. 0.75

.. 0.75
8. 0.5 2- 0.5 2-
co 4 3: co 4 3:
pH3 co pH4
2 e ~0.25
2 e
0 0 0
0 0.25 0.5 0.75 0.3 0.6 0.9 1.2 1.5
C.. (mmollL) C.. (mmollL)

\ ...
.. 0.75
• 6
0.5 •
4 2-
~0 .25
It co
• • It
0 • 0
0 0.5 I 1.5 2
C.. (mmollL)

Figure 4. Correlation between the sorption isotherms (e) and the distribution of polynuclear (0) and
mononuclear (6) molybdate species. (Reprinted fromGuibal et al., 2000a,p. 1033-1034, by courtesy of Marcel
Dekker, Inc.)

8 "-:~=:::;:::;;;::::;;;;;=1;;;;iijii"""====Z:;
c; 4
.§. 3
... 2
o 0.25 0.5 0.75 I 1.25 1.5
C.. (mmol MoIL)

Figure S. Modeling of sorption isotherms with the Langmuir equation (using the actual concentration of
polynuclear molybdate speciesinstead of the totalresidual molybdate concentration). (Reprinted fromGuibal et
al., 2000a,p. 1035, by courtesy of Marcel Dekker, Inc)


For vanadate sorption on chitosan, the sorption isotherms were found to be similar to
those obtained with molybdate. At the optimum pH (i.e., in acidic solutions) the sorption
isotherms were almost rectangular. while for the other pH values, the curves were
characterized by a sigmoidal shape, i.e., sorption was unfavorable at low metal
equilibrium concentrations and increased as metal concentration increased . The limit
metal concentration was also pH-dependent.

4.1. Sorption Isotherms

In order to verify again the influence of metal speciation, the sorption isotherms were
plotted against the total metal concentration. the total concentration of all the polynuclear
species , and the actual total concentration of deca-vanadate. Indeed, vanadate chemistry
has similarities with molybdate chemistry, though it is more complex due to a greater
number of polynuclear species , which include cationic species , non-ion ic species, and
among anion ic species, a number of different compounds including rnono-, di-, trio, tetra-
and deca-vanadate species . Figure 6 shows the plots of sorption isotherms as a function
of the different variables of concentration, and the distribution of metal species at
selected pH values.

4.2. Vanadium Species

The protonation of chitosan is required for the sorption of anion ic vanadate species ,
but the sorbent has a greater affinity for the deca-vanadate species compared to other
anionic species. Figure 6 clearly shows that the sorption isotherms are more favorable in
the case of curves plotted as a function of deca-vanadate concentrations. At pH 3, the
curves overlapped due to the predominance of deca-vanadate species over the whole
range of concentrations. But at pH 5.2 and pH 6, the only curves characterized by a
rectangular shape (very favorable sorption - almost irrevers ible) were those plotted as a
function of deca-vanadate concentrations.
The maximum sorption capacities (around 450 mg V gol) correspond to a molar
sorption capacity of 808 mmol V s". which is about 1.7 mmol V rnmol' -NH z. Again in
this case. the molar ratio between vanadium and protonated amine groups is greater than
I. The sorpt ion of vanadium occurs through the uptake of polynuclear species. As for
molybdate sorption , polynuclear species may be involved in multiple linkages with
different monomer units. The sorpt ion of I deca-vanadate unit induces the sorption of 10
vanadium units. which may explain the high sorpt ion capacity of this metal compared to
other mononuclear metal ions. A similar behavior was observed with molybdate sorption .
Most adsorbable species are formed by heptamolybdate ions, and the sorption of I mole
of heptamolybdate results in the sorption of7 molybdenum units (Gu ibal et al., 2000a) .
The preference for deca-vanadate rather than other polynuclear species (di-, trio,
tetra- species) may be related to the relative ionic charge (RIC) of these anionic species
(RIC : ionic charge divided by the number of vanadium units in the species). Indeed , for
deca-vanadate species. the RIC varies between -0.4 and -0.6. while for other polynuclear
species, the RIC is equal to - I. The electrostatic repulsion between the metal anion
species in solution and the species fixed on the sorbent is thus weaker with the deca-
vanadate species than with other polynuclear species .

500 0.8

• 400

•••• •• • ••• -
: 0.6 VIOOziOH)s-
",. t
300 VO(OH)3 VIOOz6(OH)z4-
!.,. 200
100 pHJ <Il VO z+
0 0
0 50 100 150 0 0.001 0.002 0.003
COOl (mg VIL) V(V) cODceDtntioD (M)

500 0.6

6 66 66
" a. a. a .a~.
6 aa:a'"• a. a
,~ 0.4
100 zi OH)s,

ae ae

pH 5.2
.. 0.2
100 l

0 0
0 50 100 150 0 0.001 0.002 0.003
Ceq (mg V/L) V(V) coac:eDtra tioD (M)

0 2 3 4 0.8

a a• -: 0.6
':II: " V3O /'

.- 300 6 I
66 6 aD• • •
~ ~ 0.4
r 200 66 a. a •
V40 124-
"-' a.
D.D •
K. 0.2
c:r 100
0 0
0 50 100 150 0 0.001 0.002 0.003
Ceq(mg V/L) V(V) CODceDtrac:ioD (M)
Figure 6. Sorption isotherms at different pH values and distribution of vanadate species as a function of total
metal concentrations. Sorption isotherms are plottedas a function of the total metal concentration (. ), the total
concentration of polynuclear species (0), and the total concentration of deca-vanadatespecies (~) (at pH 6, the
top scale gives the concentration of the deca-vanadatespecies). (Reprinted from Guzmanet al., 2002. p. 1570-
1571 . by courtesy of American Chemical Society.)


The recovery and separation of platinum-group metals (PGMs) from dilute or

complex solutions is of great importance due to the increasing demand for PGMs for

catalyst applications and due to the limited sources of these metals. Platinum-group
metals form coordinated complexes with a wide variety of nucleophile ligands (Chanda
and Rempel , 1990; Iglesias et aI., 1999). These ligands are generally more strongly bound
to these metals than to base metals because the 4d and 5d orbitals (spatially larger than 3d
orbitals) give rise to better orbital overlap in bond formation with ligands. Noble metals
tend to form their most stable complexes with ligands containing polarizable centers
(heavier donor atoms). Resins with donor N in their functional groups are thus efficient
exchangers for the pre-concentration and separation of noble metals. Likewise , due to the
large number of amine groups on chitosan , PGMs are expected to be efficiently sorbed on
the biopolymer . However, the strong effect of co-ions (chloride and nitrate for example)
on the sorption efficiency for platinum and palladium indicates that ion-exchange
mechanisms are involved in the uptake mechanism (Guibal et aI., 1999, Ruiz et aI.,

5.1. Sorption Isotherms in Hel and H 2S0 4 Media

The protonation of the amine groups is responsible for the electrostatic attraction of
PGM anions. The influence of the acid used for pH control was tested on platinum
sorption using glutaraldehyde cross-linked chitosan (GC). Figure 7 shows the sorption
isotherms obtained with sulfuric acid and hydrochlori c acid. Though the maximum
sorption capacity tends to a similar value (180 mg Pd g" , ca. 1.7 mmol Pd g") with both
sulfuric acid and hydrochloric acid solutions , Figure 7 clearly shows that the affinity of
the sorbent for palladium (given by the initial slope of the sorption isotherm curve)
depends on the acid used for pH control. In the case of hydrochloric acid solutions , the
isotherm curve is almost rectangular ; i.e., the initial slope is very steep and the plateau is
reached for a residual concentration as low as 10 mg Pd L" . In the case of sulfur ic acid
solutions. the sorption capacity progressivel y increases up to a plateau that is only
reached for a residual concentration of 40 mg Pd L". These differences may be correlated
to the change in the distribution of metal species as a function of the pH, the total metal
concentration , and the chloride concentration. Figure 8 shows the speciation diagrams for
different experimental condition s.

200 .------ - - - - - ---,

Hel a a

i fa
~ 150 aa 6 6

100 6 6666 6

... 50 6

O +-- --t-- ---;- - --r-- --t-- ---l

o 10 20 30 40 50
Ceq (mg PdIL)
Figure 7. Influence of the acid used for pH control on palladium sorption using glutaraldehyde cross-linked

In hydrochloric acid solutions, the dissociation of both tetrachloropalladate salt and

HCI leads to a high concentration of chloride compared to palladium. The predominant
species at pH 2 with a 0.01 M concentration of chloride are PdCh' and PdCl/"

(independent of the total metal concentrat ion). These species are anionic forms, which
are suitable for sorption on protonated amine groups through anion exchange/electrostatic
attraction . On the other hand, when the pH is controlled by sulfuric acid, the presence of
chloride ions is only due to the dissolving of palladium salt. Consequentl y, the total
chloride concentration does not exceed 2 mM. Under these conditions (lack of chlor ide
ions), the distribution of palladium is significantly influenced by the pH and the total
metal concentration, so the formation of cationic or non-ionic species is expected . When
the concentration of palladium increases, the concentrat ion of anion ic species also
increases as the affinity of the sorbent for the metal species increases.

60 80

~ CI : 100 mM
. ~
:l 40
r- (d)
2.5 3 3.5 4 2 2.5 3 3.5 4
pC (Pd) pC (Pd)

80 90
Pd: 0.1 mM
(b) (e)
! 60 Pd: 0.1 mM
:l 40 .~
'2 CI : 100 mM
l30 (d)
:- 20
(a) '" 1-------- ------
0 0
I 2 3 4 1 2 3 4
pH pH


rl 40


2 3 4
pC (CI)

Figure 8. Influence of pH. chloride concentration. and palladium concentration on the distribution of metal ion
species (pH 2, unless specified) « a): Pd'+; (b): PdCr; (c): PdCJ,; (d) PdCJ,"; .(e): PdCI/ ; (I) : Pd(OH( ; (g):
Pd(OH h; (h) Pd(OH)/). Whenthe concentration was less than 5%, the specieswasnot included in the
speciation diagrams.

Similar observationson the influenceof the acid used for pH control have been made
on the sorption of mercury on polyaminated highly porous chitosan chelating resins
(Kawamura et al., 1993). Chitosan gel beads were prepared by the usual coagulation
procedure and then polyethyleneimine was grafted onto the beads using a cross-linking
procedure. Kawamura et al, (1993) observed that mercury was adsorbed on the sorbents
under experimental conditions that correspond to the predominance of non-dissociated
species (ca. HgCI2) , and concluded that mercury uptake occurred through a chelation
mechanism on the nitrogen groups. However. in the presence of a large excess of
chloride in the solution, the sorption capacity decreased due to the ionization of HgCI 2 to
form HgCI 2' . Comparing the sorption capacities for solutions prepared in the presenceof
sodium chloride and hydrochloric acid. Kawamura et al. (1993) observed that for a
similar chloride concentration, the saturation capacity was greater in HCI solutions than
in NaCI solutions. This change in the effect of solution composition was explained by a
change in the sorption mechanism .

Chelation mechanism:
R1R2R)N + HgCh ~ RI R2R)N-HgCI 2 (I )

lon-exchange mechanism:
R1R2R )N + HCI ~ R,R2R)NH +Cr (2)
RjR 2R )NH+Cr + HgCI)' ~ RJR2R)NH +HgCh' + cr (3)

Kawamura et al. (1993) also observed that mercury was not adsorbed from solutions
in which pH was controlled with sulfuric acid. The decrease in sorption capacity is
explained by the protonation of amine groups and the formation of ion pairs with sulfate
ions in the vicinity of protonated amine groups [in equation (2), sulfate ions are
substituted by chloride ions). Since mercury does not form complexes (and. more
specifically. anionic complexes) with sulfate, HgCh is not ionized and mercury species
cannot be exchanged. as seen in equation (3). Therefore. mercury speciation reveals a key
parameter and, more specifically, the complexation of mercury strongly influences
sorption efficiency.
Saba et al. (1998) prepared a chitosan derivative (N-(2-pyridylmethyl)chitosan] in
order to improve the sorption selectivity of copper(lI) over iron(III). The sorption
properties for copper were compared in different media (including sodium nitrate and
ammonium nitrate) at several pH values. In the presence of sodium nitrate. the
distribution coefficient [D = (metal concentration in the solid) / (metal concentration in
the aqueous phase), L kg'J] for copper sorption increased continuously with pH up to pH
5-6 (below the precipitation domain), while in the presence of ammonium nitrate. the
distribution coefficient increased up to pH 3-4 (with values comparable to those obtained
with sodium nitrate solutions). remained constant up to pH 6-6.5. and then decreased
again. In the presence of ammonia, at near neutral or alkaline pH, copper ions were
complexed as Cu(NH) / + (i=I-4) and did not precipitate. These complexes cannot be
adsorbed onto chitosan derivatives. In the low pH region, copper complexes (with
ammonia and hydroxide) are not formed and copper can be directly adsorbed in the form
of copper nitrate. as indicated in the following equations:

Polymerprotonation in the low pH region:

RNN1s) + 2 H+ + 2 NO)' ~ (4)

RNN(s) + Cu(NH l ),2+ + 2 NO l ' ¢:> (5)

In the pH range in which copper is strongly complexed by ammonia, the addition of

ammonium chloride strongly reduces the sorption efficiency, since the metal is less
available for sorption on amine groups. On the other hand, by decreasing the pH, the
formation of ammonia-copper complexes is reversed, and ammonia does not change
sorption performance .
Compared to the sulfate anions, the presence of chloride ions is more favorable for
the sorption of palladium on cross-linked chitosan due to the formation of favorable
ionized species that may interact with protonated amine groups in acidic solutions. Figure
9 shows the sorption isotherms for palladium in the presence of increasing concentrat ions
of chloride ions (the pH was controlled with HCI). The addition of chloride ions (0.05-
0.10 M) in the salt solution results in a decrease (by 20-25%) of the sorption capacity for
palladium, due to the competitor effect of chloride ions for sorption on protonated amine
groups. The initial slope of the isotherms is also significantly changed with increasing
chloride concentration. It can be seen that while at low chloride concentration the
sorption is very favorable (almost rectangular) with a very steep initial slope, when the
concentration of chloride ions exceeds 0.05 M, the slope of the curve (i.e., the affinity)
decreases significantly. By increasing chloride concentration above 0.1 M (up to 0.5 and
1.0 M), the sorption capacity strongly decreases below 50 mg Pd s" (not shown, Ruiz et
al., 2000). The distribution coefficient is plotted on the figure as a function of the
equilibrium coefficient for an addition of chloride ions of 0.0 I M, 0.05 M and 0.10 M.
All the curves overlap throughout the entire range of metal concentrations. In HCI
solution, metal speciation is not influenced by the excess of chloride ions, which act only
as competitor ions.
I .E-+{)7 ,.---- - -
- ..._--_. ...._-_. - - - -----,
- - - _. _ -

I .E-+{)6

~ I .E-+{)5

I.E-+{)3 +----t---+- -+-- +----t

o 10 20 30 40 50
C .. (mg PdIL)

Figure 9. Influence of increasing concentrations of chloride ions on palladium sorption at pH 2 using

glutaraldehyde cross-linked chitosan (distribution concentration and sorption isotherms) .

5.2. Effect of Chitosan Modification

The acid used for pH control also influences the sorption of platinum. Figure 10
shows the influence of the acid on the sorption of platinum on glutaraldehyde cross-
linked chitosan. For similar reasons, the larger sorption obtained at a low metal
concentration is related to the speciation of platinum and to the predominance of

adsorbable compounds (anionic platinum species) in HCI solutions (compared to sulfuric

acid). However, in the presence of sulfuricacid, by increasing the residual concentration
of platinum the sorption increases. Yet, this increase is not sufficient to reach the level
attained in HCl solutions. Figure 10 also shows the influence of the type of acid on
platinum sorption on a sulfur derivative of chitosan. Sulfur is known to be a strong
electron donor that is very reactive with PGMs. Sulfur compounds are very efficient at
chelating PGMs. Pesavento and Biesuz (1998) and Pesavento et al. (2000) have pointed
out that chelating resins are assumed to be characterized by their intrinsic complexation
constants which, according to Gibbs-Donnan model, are related to the complexation
constants in solutions with ligands havinga similar structure to that of the active groups
in the resin. Therefore, taking into account the strong tendency of sulfur compounds to
react with PGMs in solutions, a similar increase in the sorption efficiency is expected
after chemical modification of chitosan.
A number of resins have been modified in order to bestow chelating properties to
more traditional ion-exchange resins (Chanda and Rempel, 1990; Iglesias et al., 1999).
Binman et al. (1997) immobilized sulfur derivatives of jojoba wax on polystyrene beads
in order to sorb mercury and chromate. Many sulfur derivatives have been prepared in
that improve sorption efficiency for metal ions (Muzzarelli and Tanfani, 1982; Arguelles-
Monal and Peniche-Covas, 1993; Guibal et al., 2000; Cardenas et al., 2001). One of the
most significant limitations in the use of these derivatives is the weak stability of sulfur
derivatives on chitosan. Thiourea was selected as the chelating group to be grafted on
chitosan and to prepare TGC (thiourea-grafted chitosan). The modification was made by
a pre-reaction of the sulfur compound with glutaraldehyde and a subsequent reaction of
the reactive product with chitosan. Glutaraldehyde acts as a link between the chitosan
backbone and the substitutent.

_- -0-- - -~

. --
a 0
a 0

300 0
a a _~a .--';;=::--'IJ" 0
"~ . - ' - _.---~ a
~ •
200 •
! •
...- •
o ,
100 . ,' •
:,. -~~
..... .3-- ------- • •
0 5 10 15 20 25 30
Ceq (mg PtIL)

Figure to. Influence of the acid used for pH control (sulfuric acid: filled symbols; hydrochloric acid: open
symbols)on platinum sorption isotherms using glutaraldehyde cross-linked chitosan (GC) (. /0) and a thiourea
derivativeof chitosan(TGC) (.10) (the lines represent the modelingof the isothermcurves with the Langmuir
equation). (Reprinted fromGuibaJ et al., 2000b,p. 132, by courtesyof John Wiley & Sons,lnc .)

The grafting of thiourea results in an increase in the sorption capacity, due to an

increase in the number of chelating and sorption sites: amine groupsand sulfur groups in

HCI solutions. The actual sorption mechanism has not yet been identified, but both ion
exchange and chelation are suspected to occur. More interesting is the form of the
sorption isotherm in the presence of sulfuric acid solutions. Similar to the observations
made on palladium recovery, at low residual concentration the sorption remained low but
increased as the residual concentration of platinum increased. The affinity of the sorbent
for platinum (initial slope of the isotherm curve) was significantly lower than those
observed in HCI solutions; however, the affinity and the saturation capacity increased
strongly with the grafting of sulfur compounds in sulfuric acid solutions. At saturation of
the sorbent, platinum sorption capacities are comparable for hydrochloric acid and
sulfuric acid solutions in the case of sulfur derivatives of chitosan. The speciation of
platinum does not influence its maximum sorption on TGC, but at low metal
concentration, the sorption capacity is significantly lower than in hydrochloric acid
solutions. The sorption isotherms can be modeled using the Langmuir equation, and the
lines on Figure 10 show the modeling of the experimental data according to this model
with the parameters summarized in Table I. The model fits the experimental data well.

Table 1. Modeling of sorption isotherms for platinum at pH 2 with the Langmuir

equation(qm: mg Pt g'l, b: L mg") .
Sorbent Acid used qm b MSR
TGC HCI 346 3.1 24
GC HCI 304 1.1 24
TGC H2S04 387 0.14 32
GC H2S0 4 288 0.21 28
MSR: [1:,." .,(F".(C,)·F". (C,))' /n)"' , whereF - the Langmuir equation.

In order to verify the influence of the presence of co-ions, the sorption isotherms
were compared for glutaraldehyde cross-linked chitosan and for a thiourea derivative of
chitosan in hydrochloric acid and sulfuric acid solutions. The solutions were completed
by addition of complementary salts (NaCi in sulfuric acid solutions, and Na2S04 in
hydrochloric acid solutions). Figure II shows the influence of increasing the
concentrations of added salts in these acidic solutionson platinum sorption capacity.
In HCI solutions, sorption capacities were comparable for the two sorbents in the
absence of sulfate (130 mg Pt g" , under selected experimental conditions: pH 2, initial
metal concentration 10 mg Pt L" ; sorbent dosage 75 mg L"). However, increasing the
concentration of the co-ionsabove 100 mM, drastically decreasedthe sorptioncapacity to
30 mg Pt g.1 in the case of glutaraldehyde cross-linked chitosan.On the other hand, in the
case of the thiourea derivativeof chitosan, a small decrease in the sorption capacity was
observed, but it remained higher than 120 mg Pt g.l, even when 500 mM sulfate was
added. In the case of the sulfur derivative, sulfate ions have a weak inhibitory effect on
sorption capacities. The introduction of strong chelating groups made it possible to
decreasethe influenceof platinum speciation.
In sulfuric acid solutions, the difference was significant between glutaraldehyde
cross-linked chitosan and the thiourea derivative of chitosan: The sorption capacity was
2.5 times greater for the sulfur derivativethan for the cross-linked material in the absence
of chloride anions. Increasing the amount of chloride increased the sorption capacity of
the two sorbents up to a chloride concentration of 50 mM. Above this value, the sorption
sn cl" TlO~ 0" M nAl IO~ S 241

capacity decreased sl ighlly with the sulfur deriv ative, and mo re drastically with the cross-
linked sorbent. In the laner case, the sorpt ion capac ity tended to 30 mg PI g" (comparable
10the sorption capacity reached in He l med ia with a high sulfate co ncentra tion).
The presence o f chloride at low concentrat ion favored the formaticn o f chlcro-
an ionic species that are readily sorbed on peotonated am ine gro ups . By add ing chloride
anjons, the specialion of platinum was disp laced to the formation of these anionic
species . and the sorptio n increased. In the case of gl utara ldehyde cross-linked chirosan,
the sorption capac ity was IWK:e tha t obtained in sul furic acid solution (without chloride )
when chlori de concentration reached SO mM. The introduction of ch loride an ions also
enhanced plat inum sorption on the su lfur derivative. Though the chelation o f platinum
Io\I.S not expected to be controlled by the speciation of the metal. the pos itive effect of
add ing chloride may have been due to the enhancement of ion-cxchange properties on
prot:Oflaled amine groups . Indeed. in the: sulfur derivative, both ion Cl:change sites
(prolonal ed amine groups) and chelaring groups (sulfur and amine sites) were active in
the uptake of platinum. The introd uction of chloride an ions enhanced ion exchange
performance in add ilion to the chelation process. AI high chlori de concentralion (in large
excess). chlori de tons interfered in the ion exchange mechan ism, but the chelation was
r'lOl affected and the sorptio n capacity remained high (compared to cross-li nked material ).
The com pet itive effect of sulfate and the influenc e of the f)-pc of ro-ions on plat inum
sorption Io\ ere cort$istenl with the conclusions reached by Kawamura et al. ( 1993) on the
sorpt ion of mercury . The ability of chloride ions to form complexes with mercury and
plat inum contrasts with the inabi lity of sulfate ions 10 ion ize these metal s. which in tum
affC'cts the abili ty of the soetenr 10 adsorb plaIinu m through ion-cxchange reaction. The
gra fting of a sulfur compound on chitosan brought to the ioo-<xchange resin new
chelaling funct iona l groups thai were less sensitive to the presence o f suIfare ions and to
an exces s o f chloride ions.

ISO, -- - - - - - -
I SO 1 H,SO,
• • •
_ 100.
, • • •

~ I • •
~ so;' •
• • •
o ~~~~i>C, ,J'-----+~~_---'
o 100 200 roo 400 0 100 200 roo 400 '00
S.. lf l t.. eee e. (mM} C. lor;".. fOU . (m M)

..Ie..... II . InnUofnc.. of tIM: ICld USfd for pH control, of IIIoIl lntT(I,~lna fOrK:fntrlUon of Co- IOll~ (s..lfl l.. lOllS ,n
He l sol..lIons and cbloride Ions In H,SO. solutlOllsl on pl.." num sorpllr>n lsolhfr-ms "" na a1utll al<k1r) <kcross-
lonked cbllosan (e l and I tb'Ollr... dmvat,ve of cllltosan (0) \ Rl'plln.ed from GUlbai ~ aI. 2000b. 1l 12&. by
co..nny of Jobn WIIl'y &: Sons. IrK: 1


The preceding section has shown that the limiting effect of the composition of the
solution may be overcome by a simple modification of the sorbent which causes a change
in the sorption mechanism. However, several studies have recently shown that the
composition of the solution and the speciation of the metals may also change the sorption
mechanism .
Juang and Ju (1997; 1998) used polyaminated chitosan beads, prepared by
polyethyleneimine grafting on porous chitosan beads, for the sorption of copper in the
presence of a chelating agent (EDTA). Copper sorption on chitosan and derivatives has
been studied extensively (Muzzarelli and Tanfani, 1982; Coughlin et aI., 1990; Dixon and
Deans, 1992, Findon et al., 1993; Lee et al., 200 I). Copper sorption on chitosan proceeds
by chelation on amine groups, though some interactions may also occur with hydroxyl
groups on the polymer chain (Domard, 1987). The optimum pH for sorption is
approximately pH 5. Depending on the total concentration of copper, it is impossible to
investigate sorption at higher pH values, due to precipitation phenomena. However, when
chelating agents (such as tartarate, citrate, EDTA) are present in the solution, the limit pH
for precipitation is increased; this change in the pH for precipitation can be a problem for
wastewater treatment since conventional precipitation processes cannot be applied.
Juang and Ju (1997) have shown that in the presence of chelating agents copper can
be sorbed on chitosan , but at lower pH. The protonation of amine functions on the
sorbent and on the PEl grafted on the polymer makes the material efficient at sorbing
anionic copper species by electrostatic attraction. Indeed, in equimolar diluted solutions
containing cupric ions and EDTA, within a wide range of pH values (between pH 3 and
pH 12) the predominant species is Cu(EDTAt, while at a lower pH, copper is present in
solution in the form of CuH(EDTA). Juang and Ju (1997) observed that the sorption
capacity was very low below pH 3 due to the weaker electrostatic attraction of
monocharged species by the ammonium groups of PEl immobilized on the sorbent. This
decrease in sorption may also be due to the strong chelation of hydroxo complexes of
Cu(EDTA) chelates, which makes this water-soluble compound less reactive for amine
and ammonium groups on the sorbent in the very acidic region. Finally, Juang and Ju
(1997) observed that by increasing the pH, the sorption capacity increased again. They
suggested that it may be caused by competitive reactions of protonation and complex
binding. Under selected experimental conditions corresponding to the formation of
anionic chelated copper species, the interaction between chitosan and the metal is an
electrostatic attraction. When the experimental conditions change (e.g., a change in the
protonation, change in the speciation of copper to form less anionic species), the sorption
mechanism can be partially replaced by a chelation mechanism.
Wu et al. (1999) investigated the influence of pH on copper(II) sorption on chitosan
in the presence of chelating agents. The sorption isotherms are strongly influenced by the
nature of the chelating agent, the pH, and the concentration of the metal (equimolarity of
metal and ligand). An increase in the pH of the solution was also observed after metal
sorption. Under non-controlled acidic pH conditions, the sorption isotherms were very
favorable at a low metal concentration in the absence of chelating agent and in the
presence of gluconate, tartarate, and citrate. Under similar conditions, in the presence of
EDTA, the affinity of chitosan for copper was strongly decreased. In acidic solutions,
copper was present in the form ofCuH(EDTA)' and to a lesser extent Cu(EDTA)2.(Juang

and Shiau, 1998); however, H2(EDTA)2. anions were also present and competed with
copper chelates for the sorption on protonated amine groups. This competition effect did
not occur with the other ligands (not ionized under selected pH conditions). By
increasing the residual concentration (above 0.4-0.5 mM), the sorption capacity
decreased again in the presence of EDTA, tartarate, and citrate, while with gluconate as
the chelating agent, sorption capacities continued to increase up to a residual
concentration corresponding to a plateau at a value slightly lower than the maximum
sorption capacity obtained in the absence of a chelating agent.
The weak effect of gluconate may be explained by the weak formation constant of
this ligand compared to those of the other ligands. Depending on the pH of the solution,
the mechanism may change: electrostatic attraction of anionic chelates, and/or
coordination of Cu(Il) with unprotonated amine groups. Changing the pH changes the
protonation of the polymer but also the distribution of metal species. Consequently, the
optimum pH for copper sorption in the presence of ligands depends on the chelating
agent: pH 3.1-4.2, pH 5.0-5.8, pH 5.0-6.0, and pH 5.2-5.8 for EDTA, citrate, tartarate,
and gluconate, respectively. The optimum pH presents a slightly acidic front compared to
the equivalent points on titration curves: the aqueous-phase complexation of Cu(Il) and
chelating agents start to proceed after optimum pH, and coordination of Cu(II) with
unprotonated amine groups dominates in less acidic solutions.
Lasko and Hurst (1999) studied silver sorption on chitosan under different
experimental conditions, changing the pH (2, 4, 6, 8 and 10) in the presence of several
complexing agents. The sorption performances of chitosan were compared to those of 4
commercial resins: Amberlite 1RA-67 (weakly basic, polyamine functional groups), IRA-
458 (strongly basic, quaternary ammonium functional groups), IRC-718 (iminodiacetic
acid functional groups) and Duolite GT-73 (thiol functional groups).
In the presence of chelating agents, silver ions (Ag+) appeared in the solution in
cationic forms, Ag(NH 3h\ with ammonia, and anionic forms, Ag(CN)2-, Ag(SCN)/,
Ag(S203)/ with cyanide, thiocyanate, thiosulfate, respectively. As expected, cationic
forms were adsorbed at a near neutral pH or weakly acidic pH (4-8), when the chitosan
amine groups were unprotonated and the electron pair on the amine nitrogen was
available for donation to silver. On the other hand, at low pH, the protonation of amine
groups allowed metal anions to be sorbed: Ag(SCN)/and Ag(S203)/ were efficiently
sorbed at pH 2, while at pH 4, sorption was drastically reduced and null at neutral or
alkaline pH. Silver cyanide ions were not sorbed on chitosan regardless of pH: because of
precipitation of silver cyanide at pH 2, it was impossible to study behavior at this pH,
.whereas this pH should be the optimum pH for the sorption of this anionic form. In the
case of Ag" sorption, sorption performances (at each pH) were comparable for chitosan
and Amberlite IRC-718, Duolite GT-73 was better at pH 2: the chelating resins exhibited
a behavior similar to that of chitosan.
On the other hand, the anion-exchange resins (Amberlite lRA-67 and lRA-458) were
not efficient at removing silver, independently of the pH. For silver sorption in the
presence of chelating thiosulfate, the chelating resins (Duolite GT-73 and Amberlite IRC-
718) only adsorbed silver in alkaline solutions, while Amberlite lRA-67 removed silver
thiocyanate with comparable efficiency between pH 2 and pH 6. Amberlite lRA-458
strongly adsorbed silver thiocyanate independently of the pH. Chitosan was able to sorb
silver species through different mechanisms (ion exchange and chelation), but compared
to reference commercial resins bearing similar functionalities this biopolymer was more
sensitive to the pH.

Chitosan is inefficient at removing alkaline and alkaline-earth metals, though some

derivatives of chitosan , for example whose bearing phosphorylated functions, have
shown magnesium or calcium sorption properties. Hsien and Rorrer (\995) explain this
inefficiency of raw chitosan by the configuration of the electronic orbitals of these ions,
which have only s- and p-electrons in their outer shells. Heavy metals have s- and p-
electrons, but they also possess d and f unsaturated orbitals that make them more reactive
with chitosan.
Piron and Domard (1998b) have shown that strontium cannot be adsorbed onto
chitosan under standard conditions . However, in alkaline solutions, in the presence of
carbonate, strontium forms an ion-pair (s.-l+ C0 32-) that can be efficiently sorbed onto
chitosan . The interaction between chitosan and these ion pairs is weaker than that
between chitosan and transition metals. Chitosan interactions interactions with strontium
and carbonate were proven not to proceed by electrostatic attraction: the ion pairs were
first formed and then bound to amine groups. Due to the dehydrating power of carbonate
(direct interaction with chitosan or with the water molecules of the outer shell), the
accessibility of metal ions to amine groups decreased and the sorption efficiency
dropped: the polymer can be mixed with carbonate ions only after the formation of the
ion pairs. They obtained selectivity in the sorption of alkaline earth-metals in relation to
the ionic radius of the metals: sorption efficiency was about 5 times greater for strontium
than for barium due to the shorter ionic radius of barium.


Metal-ion sorption on chitosan may proceed by several mechanisms, including

chelation on unprotonated amine functions (for cations) and ion exchange or electrostatic
attraction on protonated amine groups (for anions). Metal speciation may change due to
the presence of chelating agents in the solution. This change in the distribution of metal
species influences the charge of metal ions and the mechanisms involved in their uptake.
In some cases, the presence of co-ions enhances the sorption of metal cations (e.g.,
strontium sorption in the presence of carbonate ions by formation of ion pairs adsorbable
on amine functions). The pH also strongly influences metal speciation and thus both the
sorption performances and the uptake mechanism (e.g., copper sorption in the presence of
chelating agents).
The case of oxo-rnetal ions was considered with special attention given to molybdate
and vanadate . These metal anions were adsorbed on protonated amine groups through
electrostatic interactions. The correlation between the metal species present in the
solution (under given experimental conditions: pH and total metal concentration) with the
shape of the sorption isotherms made it possible to identify the absorbable species.
Indeed, at low residual concentration, sorption isotherms were unfavorable for given pH
(e.g., very mild acidic pH, sorption capacities were very low or negligible) , but by
increasing the concentration, the sorption increased sharply . This limit concentration was
dependent on the pH and the metal but can also be correlated to the appearance of
selected polynuclear species such as heptamolybdate forms (free or hydroxylated) and
deca-vanadate species (among other polynuclear species).
In the case of PGM sorption on glutaraldehyde cross-linked chitosan, the sorption
also proved to be controlled by the distribution of metal species. Since sorption also
occurs through anion exchange , the presence of anionic species is required . The optimum

pH was found to be approximately pH 2. The acid used for pH control revealed a very
important parameter: in sulfuric acid solutions. the lack of chloride disp laced the PGM
salt dissolution to the formation of non ionic or cationic species that were less adsorbable.
On the other hand. in hydrochloric acid solutions. PGM ions were present under the form
of anionic chloro-complexes. which are strongly adsorbable (very high affinity of the
sorbent for these specie s, materiali zed by the sharp initial slope of the isotherm curves) .
Similar trends were observed with mercury sorption on polyaminated chitosan gel beads .
In sulfuric acid solutions, the addition of chloride ions (up to a concentration that induced
an ion-exchange competition mechanism. ca. 0.) M) improved sorption performance. A
simple modification of the biopolymer, obtained by grafting sulfur compounds (thiourea)
onto a chitosan backbone , allows the sorbent to remove metal ions by a dual mechanism,
including both ion exchange and chelation mechanisms. The introduction of sulfur
moieties brought a dual sorption structure to the sorbent (anion exchange and chelation),
making the resin more versatile (less sensitive to the presenc e of co-ions) and more
efficient (higher sorption capacities).
Understanding the sorption mechanism and the influence of metal speciation on
sorpt ion performance is very important because it makes it possible to predict the
optimum pH for a given metal concentration. It is also important information needed to
predict the possibility of separating metals in mixed solutions by adequately controlling
the pH.


The Spanish Ministry of Education and Culture CICYT (QUI99-0749) and DURSI
(SGR2000-00075) have supported part of this work. J.G. is indebted to CONACYT and
CONCYTEG (W 99-16-203-201 Annex 2) for a fellowsh ip. M.R. also acknowledges
CIRIT for the fellowship received during her sabbatical in France. J.G., R.N., and E.G.
acknowledge the Franco-Mexican program PCP (99/4) for financial support .


Argu elles-Monal , W., and Peruche-Covas , C , 1993, Preparation and characterization ofa mercaptan derivative
of chitosan for the remova l of mercury from brines, Angew. Mu cromol. Chem ., 207 : 1-8.
Baba, Y.. Masaaki , K , and Kawano , Y., 1998, Synthesis of a chitosan derivative recongnizing planar metal ion
and its selective adsorption equilibria of copper(lI) over irondll), React . Funct. Polym. , 36(2) : 167-172.
Baes, CF ., Jr., and Mesmer , R E., 1976, Hydrolysis ofCutions, Wiley, NY, pp. 1-489 .
Bassi, R., Prasher, S.O., and Simpson , BK , 2000, Removal of selected metal ions from aqueous solutions
using chitosan flakes, Sep . Sci Technol., 35(4) 547·560
Binman , S.. Belfer , S , and Shan i, A , 1997, Metal sorpti on propertie s of sulfur-chlorinated jojoba wax bound to
polystyrene beads 1. Appl . Polym . Sci., 63(5) : 625-633.
Brooks , C S., 1991, Metal Reco very/rom Industrial Wustes , Lewis Publish ers, Chelsea, pp. 1-267 .
Cardenas , G , Orland o, P., and Edelio , T., 2001 , Synthe sis and applicat ions of chitosan rnercaptanes as heavy
metal retention agents , Int. 1. Bioi Mucromol., 28(2) : 167-174
Chanda, M , and Rempel, GL , 1990, Polybenzimidazole resin based new chelat ing agents . Palladium(lI) and
platinum(l V) sorption on res ins with immobili zed dithiooxarn ide, Reactive Polym ., 12: 83·94.
Charlot , G , 1978, Dosages Absorptiomet riques des Elements Mineraux , Masson , Paris, pp.I-444.
Chen , J., Tende yong , F , and Yiacoum i, S , 1997, Equilibrium and kinet ic studies of copper ion uptake by
calcium alginate, Environ Sci. Techn ol., 31(5) : 1433-1439

Ciminelli, V.ST, and Garcia. 0 ., Jr., 200I, Biohydrometallurgy: Fundamentals, Technology and Sustainable
Development. Part A: Bioleaching, Microbiology and Molecular Biology - Part B: Biosorption and
Bioremedianon., ProcessMetallurgy II A and II ", Elsevier, Amsterdam, pp. 1-683.
Coughlin, R.W., Deshaies, MR., and Davis, EM ., 1990, Chitosan in crab shell wastes purifieselectroplating
wastewater, Environ. Prog., 9( I): 35-39.
Dambies, L., Vincent, T., Domard, A., and Guibal, E , 2002, Preparation of chitosan gel beads by ionotropic
molybdategelation, Biomacromolecules , 2(4): 1198-1205.
Dixon, B.G., and Deans, J.R., 1992, Uptake of Pb" and Cu" by novel biopolymers, Water Res., 26(4): 469-
Domard, A., 1987, pH and CD. measurements on a fully deacetylated chitosan: application to Cu"-polymer
interactions, Int. J. BioI. Macromol., 9: 98-104.
Dzul Erosa, M.S., Saucedo Medina. T/. , Navarro Mendoza, R., Avila Rodriguez, M., and Guibal, E , 2001,
Cadmium sorption on chitosanderivatives, Hydrometallurgy; 61(3): 157·167.
Findon, A., McKay, G., and Blair, H.S., 1993, Transportstudies for the sorptionof copper ions by chitosan, J.
Environ. Sci. Health. Part A, 28(1): 173-185.
Guibal, E., Roulph, C , and Le Cloirec, P., 1992, Uranium biosorption by a filamentous fungus Mucor miehei:
pH effecton mechanisms and performances of uptake, Water Res., 26(8): 1139-1145.
Guibal, E., Saucedo, /., Roussy, J., and Le Cloirec, P., 1994,Uptakeof uranyl ions by new sorbing polymers:
discussion of adsorption isotherms and pH effect, React. Polym., 23: 147·156.
Guibal, E, Roulph, C and Le Cloirec, P., 1995a. Infrared spectroscopic study of uranyl biosorption by fungal
biomassand materials of biological origin, Environ. Sci. Technol., 29(10): 2496-2503.
Guibal, E., Jansson-Charrier, M., Saucedo, /., and Le Cloirec, P., 1995b, Enhancement of metal ion sorption
performances of chitosan: effectof the structureon the diffusionproperties, Langmuir, 11(2): 591-598.
Guibal, E , Milot, C , and Tobin, J.M., 1998, Metal anion sorption by chitosan beads: equilibriumand kinetic
studies, Ind. Eng. Chem. Res., 37(4): 1454-1463.
Guibal, E., Vincent, T., Larkin, A., and Tobin, J.M., 1999, Chitosansorbents for platinum recovery, Ind. Eng.
Chem. Res., 38(10): 4011 ·4022.
Guibal, E., Milot, C , and Roussy, J., 2000a. Influence of hydrolysis mechanisms on molybdate sorption
isothermsusingchitosan, Sep. Sci. Technol., 35(7): 1021 -1038.
Guibal, E., Vincent, T, and Navarro Mendoza, R., 2000b, Synthesis and characterization of a thiourea-
derivativeof chitosan for platinumrecovery, J. Appl. Polym. Sci., 7S(I): 119-134.
Guibal, E., McCarrick, P., Tobin, J.M., 2001, Comparative study of the sorption of anionic dyes on activated
carbonand chitosanderivatives, submitted for publication in Sep. Sci. Technol..
Guzman, J., Saucedo,/. , Revilla. J., Navarro, R , and Guibal, E., 2002, Vanadium(V) interactions with chitosan:
Influence of polymer protonation and metalspeciation, Langmuir, 18(5): 1567-1573.
Hsien, T- Y., and Rorrer, G.L., 1995, Effect of acylation and crosslinking on the material properties and
cadmiumion adsorption capacity of porouschitosan beads, Sep. Sci. Technol., 30(12): 2455·24
Hsien, T- Y., and Rorrer, G.L., 1997, Heterogeneous cross-linking of chitosan gel beads: kinetics, modeling,
and influenceon cadmium ion adsorption capacity, Ind. Eng. Chem. Res., 36(9): 3631-3638.
Iglesias, M., Antico, E , and Salvad6, V., 1999, Recovery of paJladium(1I)and gold(lII) from diluted liquors
using the resin DuoliteGT-73, Anal. Chim. A('ta, 381( I): 61-67.
Inoue, K., 1998, Application of chitosan in separation and purification of metals,in: Recent Advances in Morine
Biotechnology, Volume 2, Environmental Marine Biotechnology, M. Fingerman, R Nagabhushanam and
M.-F. Thompson. eds., Oxford & ISH Publishing PVT. Ltd. New Delhi, pp. 63-97.
Inoue, K., Yoshizuka, K., and Ohto, K., 1999, Adsorptive separation of some metal ions by complexing agent
typesof chemically modified chitosan,Anal. Chim. Acto, 388(1-2): 209-218.
Juang, R-S ., and Ju, C-Y ., 1997, Equilibrium sorption ofcopper(ll)-ethylenediaminetetraacetic acid chelates
onto cross-linked, polyaminated chitosanbeads, Ind. Eng. Chem. Res., 36(12): 5403·5409.
Juang, R.-S , and Ju, C-Y ., 1998, Kinetics of sorption of Cu(ll)- ethylenediaminetetraacetic acid chelated
anionson cross-linked, polyaminated chitosan beads,lnd. Eng. Chem. Res., 37(8): 3463-3469.
Juang, R-S ., and Shiau, L.-D., 1998, Ion exchangeequilibriaof metal chelates of ethylenediaminetetraacetic
acid (EDTA)with Amberlite IRA-68.lnd. Eng. Chem. Res., 37(2): 555-560.
Kawamura. Y.; Mitsushashi, M., Tanibe, H., and Yoshida. H., 1993, Adsorption of metal ions on polyaminated
highlyporouschitosanchelatingresin, Ind. Eng. Chem. Res., 32(2): 386-391.
Konishi, Y., Shimaoka, J.-/., and Asai, S., 1998, Sorptionofrare-earth ions on biopolymer gel beadsof alginic
acid, React. Funct. Polym., 36(2): 197-206.
Lasko, CL. , and Hurst, M.P., 1999, An investigation into the use of chitosan for the removal of soluble silver
from industrial wastewater, Environ. Sci. Technol., 33(20): 3622-3626.
Lee, S.-T , Mi, F.-L., Shen, Y.-J., and Shyu, S.-S., 2001 , Equilibrium and kinetic studies of copper(ll) ion
uptake by chitosan-iripolyphosphate chelatingresin, Polymer, 42(5): 1879-1892.

Milot, C, McBrien, J., Allen, S , and Guibal , E, 1998, Influence of physico-chemical and structural
characteristics of chitosan flakeson molybdate removal, J Appl. Polym Sci ., 68(4): 57I-580.
Modrzejewska, Z , Kaminski, w., 1999, Separation of Cr(VI) on chitosan membranes, lnd. Eng. Chem . Res.,
38(12): 4946-4950.
Muzzarelli, RAA. , and Tanfani, F , 1982, N-(O-carboxybenzyl)chitosan, N-carboxymethyl chitosan and
dithiocarbamate chitosan: new chelating derivatives of chitosan, Pure Appl. Chem., S4(II) : 2141-2150.
Nekovar, P., and Schrorterova, D., 1999, Extraction of molybdenum(VI) by Primene lMT, Solvent Extr. Ion
Exch., t7(I) 163-175.
Nekovar, P., and Schronerova, D., 2000, Extraction of V(V), Mo(VI) and W(V1) polynuclear species by
PrimeneJMT, Chem . Eng. J , 79(3): 229-233.
Papelis, C , Hayes, K.F., and Leckie, 1.0., 1988, HYDRAQL . A programfor the Computation of Chemical
Equilibrium, Technical report 306, Environmental Engineering & Science, Department of Civil
Engineering, StanfordUniversity, Stanford,CA, September 1988.
Pesavento, M., and Biesuz, R., 1998, Characterizationand applications of chelatingresins as chemicalreagents
for metalions, basedon the Gibbs-Donnanmodel , React Funct Polym .,36(2 ): 135-147.
Pesavento, M., Biesuz, R., Alberti, G., and Dalla Riva, F., 2001 , Evaluation of the sorption of metal ions on a
complexing resin from different solutions basedon the Gibbs-Donnan model, React . Funct. Polym ., 46(3) :
Piron, E., and Domard, A., I998a, Interaction between chitosan and uranyl ions Pan 2. Mechanism of
interaction,lnt. J Bioi. Macromol., 22(I) 33-40.
Piron, E., and Domard, A., I998b, Formation of a ternarycomplex betweenchitosan and ion pairs of strontium
carbonate, Int. J Bioi Macromol , 23(2): 113-120
Remoudaki, E., Tsezos, M., Hatzikoyian, A., and Karakoussis, V 1999, Mechanism of palladium biosorption
by microbial biomass. The effects of metal ionic speciation and solution co-ions, in: Biohydrometallurgy
and the Environment toward the Mining of the 21" Century. Process Metallurgy 9B. R. Arnils, and A.
Ballester, eds., Elsevier,Amsterdam, pp. 449-462.
Roberts, GAF., 1992, Chitin Chemistry , MacMillan, London, U.K, pp 206-229.
Rorrer, G.L., Hsien, T.-Y., and Way, J.D., 1993, Synthesis of porous-magnetic chitosan beads for removal of
cadmiumions from wastewater, Ind. Eng Chem. Res., 32(9): 2170-2178
Ruiz, M., Sastre, A., and Guibal,E , 2000, Palladium recovery fromdilute effluents usingglutaraldehyde cross-
linkedchitosan, React . Funct . Polym ., 45(3) 155-173.
Shiau, L.D., and luang, R.-S., 1998, Ion exchange equilibria of metal chelates of ethylenediaminetetraacetic
acid (EDTA)with Amberlite IRA-68, Ind. Eng. Chem Res , 37(2): 555-560.
Sorlier, P , Denuziere, A., Viton, C , and Domard, A., 2001 , Relation betweenthe degree of acetylation and the
electrostatic propertiesof chitin and chitosan, Biomacromolecules, 2 (3): 765-772.
Vegho, F., Beolchini, F., and Gasbarro, A., 1997,Biosorption of toxic metals: an equilibrium study using free
cells of Arthrobacter sp., Process. Biochem ., 32(2): 99-105.
Vincent, T., and Guibal, E, 200I, Cr(VI) extraction usingAliquat336 in hollowfiber module madeof chitosan,
Ind. Eng. Chem Res.,40(5) 1406-1411
Volesky , 8. , and Holan, ZR. , 1995. Biosorptionof heavy metals, Biotechnol. Prog., 11(3): 235-250.
Wu, F.-C , Tseng, R.-L., and luang, R.-S, 1999, Role of pH in metal adsorption from aqueous solutions
containingchelatingagentson chitosan, Ind. Eng. Chem. Res., 38(I) 270-275
Yoshida, H , Okamoto, A., and Kataoka, T., 1993, Adsorption' of acid dye on cross-linked chitosan fibers,
equilibria,Chem Eng. Sci ., 48(12) 2267-2272.
Yoshida, H., Kishimoto, N., and Kataoka, T , 1994, Adsorption of strong acid on polyaminated highly porous
chitosan: equilibria, Ind. Eng. Chern. Res., 33(4) 854-859
Yoshida, H., Kishimoto, N, and Kataoka, T ' 1995,Adsorptionof glutamicacid on polyaminated highly porous
chitosan: equilibria, Ind. Eng. Chem Res., 34(I) 347-355

Zdenek Matejka," Helena Parschova,' Pavla Ruszova, I Ludek Jelinek, I

Pavla Houserova.' Eva Mistova,' Milan Benes,2 and Martin Hruby2


The contamination of public drinking-water resources by oxoan ions of metals such

as Mo, V, W, Ge, and As is caused either by industrial activity (i.e., industrial waste
waters penetrating without sufficient and proper treatment into surface and/or ground
waters) or by the contact of water with metal-bearing ores' or solid waste materials such
as deposits of fly ash. These contaminated waters usually contain trace concentrations
(low mgIL level) of metal oxoanions and a large concentration (200-1000 mg/L) of
accomp anying nonhazardous anions such as chloride, sulfate , and bicarbonate.
To meet the environmental and WHO-limits on drinking water, these oxoanions should
be removed from water down to low J1, gIL levels. Th is requirement calls for a very
efficient treatment process for the separation of oxoanions. The whole metal-rem oval
process should be highly selective toward contaminant oxoanions, for the following
reasons : (i) no substant ial change in ionic composition of drinking water is allowed
during the treatment; that is, the accompanying anions such as chloride, sulfate, and
bicarbonate should remain in water to keep mineral content of the drinking water intact
and ; (ii) the separated metals such as Ge, Mo, W, and V should be recovered as by-
products . However, the currently used conventional anion exchangers lack sufficient
selectivity and separation ability toward above-mentioned oxoanions.

• To whom correspondance should be addressed . Email: zdenek .matejka@vscht.c z.

I Inst. Chern . Technol ., 16628 Prague 6, Czech Republic .
, Inst. Macromolec . Chemistry. 16000 Prague 6, Czech Republic .

Fundamentals andApplications ofAnionSeparations, edited by

Moyer and Singh, Kluwer Academic/Plenum Publishers, New York, 2004 249

In this chapter, we have investigated several sorbents in the bead form, having
different polyol moieties , for selective uptake of several metal-oxoanions from water by a
diol-cornplex formation mechanism. It was also investigated: (i) whether steric
configurations of diol groups present in polymeric matrices, such as styrenic and
polysaccharides, are appropriate enough to form stable coordination complexes with
oxoanions; and (ii) whether these polymeric materials can be considered as selective
sorbents for the removal/recovery of oxoanions.


It is well known'? that low-molecular-weight substances containing diol-groups

such as hexitols and saccharides can form cyclic complexes with oxoanions in an
appropriate steric conformation. These complexes can be categorized into two types : (a)
mono-nuclear complexes formed by Ge, B, and, As and (b) bi-nuclear or poly-nuclear
complexes formed by V, Mo, and W.

Table 1. Optimum pH-range for diol-complexes.

Oxoanion pH Oxoanion pH
V(V) 2-7 B(III) 7-11
Mo(VI) 1-6 Ge(IV) 6-12
W(VI) 1-6 As(I1I) 9-12
As(V) 1-6

The complexation reactions between oxoanions and diol-cornpounds are rather pH-
dependent. Table 1 presents" the optimum pH range for the formation of various
oxoanion-diol complexes. Typically, reactions are more pH-sensitive when a proton (H+)
is required for complexation.'
For example, protons are required and spontaneously consumed during the Mo-
oxoanion and diol-cornplex formation (Eq. 1). Consequently, this reaction results in an
increase in pH of the treated solution to neutral values, which slows down the kinetics of
molybdate complexation. Therefore, to keep the complexation reaction in progress, pH-
value of the solution needs to be maintained within the acidic range.

Mo0 2(OH)/ + 2Hidiol-ligand) + 2H+ - Mo0 2(diol -ligandh+ 4H 20 [1]

Tungstate and vanadate, on the other hand, are not so sensitive to low pH-value, as
they do not require protons for diol-cornplex formation . Consequently, the solution pH
remains constant during complex formation . These complexes are stable and can be
formed even in the neutral medium.
The optimum pH for the complexation of oxoanions of Bill, Ge IV, and As lII has been
found in the alkaline range. In these reactions, a proton (Hj is released during the
complexation of oxoanion with diol-ligand, and consequently the pH of the solution
decreases (Eq. 2).

As 20h + H 2(diol-ligand) - AslII(diol-ligand)(OHh" + W + H 20 [2]

The spontaneous changes in solution pH are the reason, for slow complexation
reaction between oxoanion and hydroxy ligand compound (i.e., diol ligand) during the
formation of a low-stability diol-complex of Mo and As. The formation of these
coordination diol or polyol complexes can be utilized for selective uptake of oxoanions
from the water stream. provided that a solid matrix containing diol-ligands with
appropriate sterical configuration of OH-groups is used as a selective sorbent. However.
the complex formation between oxoanion and the low- molecular weight diol-cornpound,
alone does not guarantee the efficient removal of oxoanion from contaminated stream
down to /lgIL level by this selective solid sorbent.



The removal of oxoanions from contaminated water, containing a high excess of

accompanying anions. by a solid sorbent with immobilized diol-ligand is usually carried
out under dynamic column operation. This is the most suitable working condition to
achieve a very low concentration of oxoanion (ug/Ldevel) in the treated water. There are
two crucial factors necessary to attain high sorption efficiency: (i) a highly stable
complex of diol-ligand with oxoanion inside the sorbent matrix and (ii) fast overall
kinetics of oxoanion-uptake. These factors will enable efficient uptake of oxoanions
even at pH values outside the optimum pH range . Stable complex formation would also
overcome the competition from accompanying anions . Fast kinetics will allow a
decrease in the oxoanion concentration down to /lg/L-level within the short period of
time available when the solution is passed through the sorbent bed in the column.
To achieve highly stable diol-complex formation and the fast kinetics of oxoanion
uptake, an appropriate steric configuration of immobilized -OH groups inside the sorbent
bead are required. The location of -OH groups at the adjacent C-atoms in a cis- position
is a key requirement for the formation of a stable mono-nuclear complex [Structure 1].

-C-C\./""' C - C -
-CH -C-C-C-C-CH20H
2 I I I I I I I 0I
OH H 0\ /0
o 0 0
\ / \ /
B ( Ge. AS) O-Me-O-Me-o
/\ / \ OH / \ OH
o 0 HO HO
[Structure 1]
[Structure 2]

The location of -OH groups at more distant C atoms or in a trails position will result
in a less stable complex. To form a stable binuclear complex [structure 2], which is the
case for Mo, V, and W oxoanions, the hydroxy compound should have at least 3 or more

-OH groups available . The more suitable steric configuration of -OH groups in the diol-
ligand will lead to the faster complexation reaction with oxoanions. A highly porous and
hydrophilic sorbent matrix will provide an adequate space, necessary for the formation of
a stable diol- complex, and, at the same time, it will provide a high diffusion rate for the
oxoanion inside the sorbent 's particle.


4.1. Synthetic Sorbents

These sorbents are based on a poly(styrene divinylbenzene) copolymer matrix,

functionalized by methylamine-glucitol and diethanolamine moieties . To produce such
sorbents, the polymeric matrix has to be functionalized by a suitable diolligand.

4.2. Polysaccharide-Based Biopolymer Sorbents

The main sorbents in this category are: (i) chitosan (poly-D-glucosamine) beads, (ii)
bead cellulose and (iii) sea-weed Ascophyllum nodosum, a marine brown algea
biopolymer having alginic acid as a main active compound. The whole structure of the
polymeric matrices of these biosorbents consists entirely of functional oligo-hydroxy-
compound units (i.e., diol ligands) ; no inert matrix is present in these sorbents.
Consequently, there is no need for any matrix, functionalization to create a selective
sorbent. These sorbents contain a highly hydrophilic matrix enabling the fast internal
diffusion-rate of oxoanions. These sorbents are available in abundant quantity, often as
natural waste material from crabs, shrimp, insects, fruit-cores, etc.


Dynamic sorption runs were carried out in glass columns (Ld.12 mm) packed with 30
mL of sorbent with solutions containing 0.5-20 mgIL metals (Mo, V, W, Ge, and As) and
100-1000 mgIL accompanying anions (Cl' and SOl). The sorpt ion run was terminated
at a metal-breakthrough concentration of 0.05 mgIL in the effluent. Solution flow-rate
through the sorbent bed was kept at 4-6 BVh·' . The effect of pH and concentrations of
metals and anions in the inlet solution on sorption capacity was determined. The sorption
ability of particular diolligands with different structures was compared .


6.1. Uptake on Sorbent with L-decxy-methyl-amlno-glucltol Moiety Immobilized in

StyreneIDVB Matrix

The metal uptake ability of this sorbent in the protonated hydrogen chloride form,
where primarily a standard anion-exchange process takes place, was investigated
previously.f'" In the present work, experiments were carried out with the commercial
sorbents Purol ite D-4123 (Purolite Int.) and Diaion CRB-02 (Mitsubishi Chern. Corp.) in

the free-base form. where a standard anion-exchange process is completely excluded.

Only complexation reactions took place under these experimental conditions.
This sorbent exhibits a suitable steric configuration of -OH groups to form very
stable mononuclear complexes with germanium [Structure 3] and boron oxoanions. The
uptake of Ge from alkaline solution is very efficient. Ge concentration is decreased from
the initial value of 20 mg/L down to less than 0.01 mg/L in the effluent (Figure 1). The
optimum pH range for Ge-uptake is 6-12; no sorption is observed at pH < 6.


f~ r yH r r
OH H 0 0
o 0
[Structure 3]


pH 6
• pH 4
E. 10 I
pH 10
(!J T

• )J pH 9

0 500 1000 1500 2000

Figure 1. Breakthrough curves of Ge-oxoanion from waterof differentpH on glucitol sorbentcolumns. Initial
solutions (feed)contained Ge at 20 mg/Land sulfateat 1000mglL(exceptfor solutionof pH 9 whichcontained
5 mglLGe and 500 mglLsulfate).

Arsenic forms a low-stability mononuclear complex, [As-glucitol] , and the kinetics

of this complexation reaction are slow.' That is why As-oxoanions (Asv and As ) are
not taken up efficiently via the mechanism of diol-complex formation.
The structure of methylamine-glucitol ligand (where 3 -0H groups are close
together in an appropriate conformation) is able to form a stable binuclear complex with
V, W, and Mo. These oxoanions prefer acidic pH-range during complexation . However,
V and Ware taken up efficiently from solution even at neutral and slightly alkaline pH
(Figure 2), as these complexes are very stable and exhibit a low requirement of H+ions.
The efficient sorption of Mo is restricted to solution of pH < 4 only, as protons are
required in complexation reaction (Eq. 1).

2 pH 10.5


1000 2000 3000


Figure 2. Breakthrough curves of V-oxoanion from water of different pH on glucitol sorbent columns. Feed
= =
solution: V 5 mglL, sulfate 100 mglL.

6.2. Uptake on Sorbent with DiethanoIamine (DEA) Moiety Immobilized ines

StyreneIDVB Matrix

An experimental sample ofPUROLITE D-3441 in the free base form was used. The
synthesis of sorbent with diethanolamine (DEA) ligand [Structure 4] is rather simple and
inexpensive as compared to glucitol sorbent. But the steric configuration of this ligand
obviously does not fully meet the requirements for a stable mononuclear complex, as the
mutual distance between -OH groups in the ligand are too large. Therefore, Ge-oxoanion
is not taken up by this sorbent. However, sorption of V, W, and Mo-oxoanions was quite
efficient (Figure 3) at neutral and acidic pH, as they form binuclear-complexes. The
stable binuclear complexes are formed even though this ligand has only two -OH groups
available for complexation. This happens due to matrix-crosslinking effect, which puts
two diethanolamine units from adjacent crosslinked polymer chains close together to
create suitable steric position of -0H groups for complexation.

_ CH

[Structure 4)


oS 0.2

a 500 1000 1500 2000 2500 3000 3500
Figure 3. Breakthrough curves of V and W oxoanions from water of pH 6.5 on (diethanolamine) sorbent
columns . Loading (feed) solution: V. W) = I mglL; sulfate =100 mglL.

6.3. Uptake on Crosslinked Chitosan (poly-D-glucosamine) Beads

Experiments were carried out with the commercial product Chitopearl CR-CH (Fuji
Spinning Co., Ltd., Japan) in the free- base form, without any functionalization . To
produce water-insoluble chitosan beads or fibers, the water-soluble chitosan unit
[Structure 5] is crosslinked by ethylenediglycidylether." Inside this crosslinked chitosan
structure , positions of four -OH groups, from adjacent polymeric chains, are fixed close
together in a cis-position, which enable oxoanions of V, W, and Mo to form binuclear
diol complexes [Structure 6]. Experimental results have shown (Figure 4) that W-
oxoanion is quantitatively removed from water by crosslinked chitosan in the free-base



[Structure 5]


[Structure 6]

0.30 .----....,...--.---....,...--.---.......----,
pH 7.2

§. 0.15


O .OO~~ .........--"""---.........--"""---........- - - I
o 100 200 300 400 500 600
Figure 4. Breakthrough curves of W oxoanion from water of pH 5.7 and 7.2 on cross-linked chitosan (in free
base form) columns. Loading solution: W =5 mglL; sulfate =500 mglL.

Ge-oxoan ion is not quantitativel y removed from water by a crosslinked chitosan .

The steric requirement s (the position of -OH groups) for a stable mononuclear comple x
of Ge-oxoan ion-chitosan are not fulfilled inside the crosslinked-chitosan matrix .
Guibal et al.8 investigated the uptake of molybdate by crosslinked chitosan,
precond itioned by mineral acid. Under these conditions, the -NH z group of the
chitosan's structure gets protonated as -NH 3+Cr and the overall sorpt ion process consists
of two simultaneous reactions: (a) diol-complex formation (i.e., rnetal-oxoanion -chitosan )
and (b) anion exchange of metal-oxoanion for cr ions.
In the present work, the quantitative uptake of W-oxoanion by protonated chitosan
was stud ied, and a very high sorpt ion capacity for W has been found (Figure 5). W-
oxoanion prefers the acidic pH for complex formation (Table 1), and protonated chitosan
is very efficient in keeping the loading-solution pH constantly at optimum low pH value
of abo ut 4 , due to amine-gro up hydrolys is (Eq. 3).


6.4. Uptake on Crosslinked Bead Cellulose (without Functionalization)

The steric structu re of cellulose is very similar to that of chitosan . The crosslinking
process will bring the -OH groups close together and binuclear complexes with the
oxoanions of V, W, and Mo can be formed . Figure 6 presents the uptake of oxoanions of
V, W, and Mo. But the -OH groups in a cis-posit ion are not located at the adjacent
C-atoms , and there fore they are not close enough even after crossl inking to form a
mononuclear diol complex such as Ge-cellulose type.



-. 0 .15

3: 0.10



0 1000 2000 3000
Figure S. Breakthrough curves of W oxoanion from waterof different pH on crosslinked chitosan(in NH)+Cr
form)columns. Loadingsolutionfor a) and b): W =5 mglL; sulfate = 100mglL; Loading solutionfor c): W =
5 mgIL; sulfate =500 mgIL.

Cellulose is known as a very porous material with a low total volume sorption
capacity. For this reason, cellulose sorbents are more suitable for rather low
concentrations of metals (around 0.5 mgIL) in a loading solution.

6.5. Uptake on Brown AigeaSeaweed (Ascophyllum Nodosum)

Seaweed contains a heterogeneous mixture of different polysaccharides (20% alginic

acid, 8% mannitol, 20% other sugars); all of them are able to react with oxoanions.
Typically, the ratio of different hydroxy compounds in the structure changes with season,
geographical locality, and the maturity level of seaweed plants. Seaweed is mostly
available as a waste material from production (recovery) of alginate (called as "partially
de-alginated seaweed ") and as a raw harvested seaweed material prior to the de-
algination process.



[Structure 7]

0, 0.4
0 0.3
->. 0.2
Mo w

0 100 200 300 400 500 600
Figure 6. Breakthrough curves of Mo, V, W -oxoanions on crosslinked bead cellulose sorbent columns.
Loading solution: Mo.V, W =0.5 mglL; sulfate =100 mglL; ... =pH 3.5; • =pH 5.5.

Alginic acid exhibits an optimal steric conformation of -OH groups in the molecule
[Structure 8] for diol-complex formation. However, the dissociation of -COOH groups
to -COO-Na+ in the sorbent's free-base form would create electrostatic repulsion of
oxoanions. Consequently, better sorption ability of this sorbent is achieved with
protonated seaweed, where the carboxylic group remains in the nondissociated free-acid
form, -COOH (Figure 7). Chemical modification of -COOH group (for instance by
esterification) to eliminate the negatively charged repulsive forces may also make this
sorbent more suitable for the uptake of oxoanions.

[Structure 8]

1.0 rr---""'---~-~~------'-----'

• •
200 400 600 800 1000
Figure 7. Breakthrough curves of W-uptake on dealginated seaweed columns. Loading solution: W = 0.9
mglL; sulfate = 100mglL; pH =5.5; • =protonated form; ... = free-base form.


The formation and a stability of coordination diol-complexes (oxoanion-

hydroxycompound) is pH sensitive. The desorption of V-oxoanion, which prefers acidic
solutions for sorption, can be easily achieved by NaOH solution (Figure 8).

.!: 200
0 2 3 4 5 6
Figure 8. Desorptionof oxoanion of V loaded on glucitolsorbentby 1 M-NaOH.


The separation of oxoanions is based on the different requirements of particular

oxoanions for H+ concentration to form stable diol-complexes . Mutual separation of the
oxoanions of Mo and V for instance was achieved using glucitol sorbent in the free-base
form at a solution pH of 7.0 (Figure 9). Under these conditions, Mo-oxoanion is not

taken up from solution at all, as it requires pH < 4 for efficient sorption. Sorption of V-
oxoanion is quant itative at pH 7. NaOH as eluent , as shown in Figure 8, can desorb the
loaded V-oxoanion from the column.


5 • •

> It
~ 2

0 500 1000 1500 2000 2500
Figure 9. Mutual separation of Mo and V using glucitol sorbent in the free-base form. Loading solution: Mo.
V = 5 mg/L: sulfate =500 rng/L; pH = 7.


I. B. N. Kabir, S. A. Ahmed, and I. Z .Zaman, Arsenic contamination of groundwater: quest for solutions. in:
Proceedings of 25" WEDC Conference on Integrated development for water supply and sanitation (Addis
Abeba, Ethiopia, 1999). pp. 192- 193.
2. A. Mikan and M. Bartusek, The reactions of oligoalcohols with arsenic. arsenous. boric and germanic acids.
Coli. Czech. Chem. Commun. 45. 2645-2655 (1980).
3. M. Bartusek and V. Sustacek, Chelates of Vanadium (V) with organic hydroxy compounds in aqueous
solutions, Coli. Czech. Chem. Commun. 48, 2785-2797 (1983).
4. M .Mikesova and M. Bartusek, Complexes of germanic acid with mannitol, sorbitol, maltol, and with
chromotropic and kojic acids, Coli. Czech. Chem. Commull. 44, 3256-3 263 (1979).
5. M. Mikesova and M. Bartusek, Complexes of molybdenum and tungstate with organic hydroxy compounds,
Coli. Czech. cso« Commun. 43,1 867-1 879 (1978).
6. S. Yasudaand K. Kawazu, Bunseki Kagaku 37,17-21 (1988).
7. Y. Kawamura, M. Mitsuhasi, H. Tanibe, and H. Yoshida, Adsorption of metal ions on polyaminated highly
porous chitosan chelating resin, Ind. Eng. Chem. Res. 32, 386-39 1 (1993).
8. E. Guibal, C. Milot. and 1. Roussy, Water Environment Research 71,10-17 (1999).
9. U. Schilde and E. Uhlemann, Reactive Polymers 20,1 81-1 88 (1993).
10. U. Schilde, H. Kraudelt and E. Uhlernann, Reactive Polymers 22.1 01-106 (1994).

Katsutoshi Inoue," Kedar Nath Ghimire,' Rabindra Prasad Dhakal,'

Kenjiro Makino,' and Tohru Miyajima '


Arsenic, selenium, and phosphorus are among the serious water pollutants
specifically generated in the effluents of mineral and chemical industries. In addition,
arsenic pollution has also been serious in some ground water or hot spring water over a
large area in Bangladesh, West Bengal in India, Inner Mongolia in China, and Japan as
well. To date, arsenic and selenium have been removed by means of the following
methods : precipitation with lime, co-precipitation with ferric sulfate, alum precipitation,
and precipitation as sulfide using either sodium sulfide or hydrogen sulfide . Although
among these methods the iron co-precipitation method has been reported to be the most
successful in lowering arsenic content to the drinking-water standard level, it still suffers
from a post-treatment problem due to the excess use of Fe(lII) salt and generation of
alkaline sludge of high water content.' Since it is difficult to remove As(lII) directly by
traditional methods, it should be oxidized into the pentavalent state prior to their
treatment by using suitable oxidizing agents like hydrogen peroxide.i
Although lime precipitation is in frequent use, it suffers from gypsum formation and
such low effectiveness in reducing arsenic to the optimum level that additional ferric
treatment is necessary. Furthermore, calcium arsenate formed in such a process and
disposed in a landfill site may interact with carbon dioxide in air to form calcium
carbonate, releasing arsenic back into the environment. Although alum precipitation is
also in practice , its performance has been unsatisfactory.
Arsenic and selenium precipitation using sodium sulfide is useful in reducing arsenic
and selenium contents in water; however, the precipitates are very colloidal in nature,
causing a filtration problem and are readily oxidized, thereby making it difficult to
dispose in a landfill site. Moreover, it is not possible to apply the traditional methods for

I Faculty of Scienceand Engineering, Saga University, Honjo I, Sage 840-8502 (·Correspondingauthor:

Katsutoshi Inoue).
2 Yamasoh Micron Inc., 1-21-12, Uenoshibamachi, Sakai 593,Japan.

Fundamentals and Applications ofAnion Separations, edited by

Moyer and Singh, KIuwerAcademic/Pleoum Publishers,New York, 2004 263

the treatment of arsenic and selenium in an acidic medium. On the other hand, phosphate
removal has been an important issue for sewage disposal plants, as the excess amount of
phosphorusgenerated in the plants induces the growth of microorganisms, thereby posing
in itself an environmental problem.
In an attempt to overcome the drawbacks of conventional treatment processes, many
works have been made regarding the removal of arsenic and selenium by using iron(III)-
loaded chelating ion exchange resins having an either acidic or basic moiety as the
functional group.2.4--7 But this treatment is expensive owing to the cost of the materials
associatedwith the removal process.
To address this goal in the present work, an alternative adsorption gel has been
prepared from cheap biomass. Effective use of biomass wastes is highly economical due
to availability of raw materials almost at no cost and is suitable for water treatment
because of their environmental friendly nature.
In this paper, we will discuss the efficiency of the adsorption gel prepared from
orange-juice residue by means of simple chemical modification by phosphorylation
followed by loading with iron(lII) for the removal of oxo anions like arsenic, selenium,
and phosphorusfrom aqueous media.


2.1. Materials

2.1.1. Phosphorylated Crosslinked Orange Juice Residue (POJR)

Adsorption gel was prepared from orange-juice residue, abbreviated as OJR hereafter,
which was kindly provided from JA Beverage Saga. The OJR principally consists of
cellulose, pectin substances, hemi-cellulose, chlorophyll pigments, and other low-
molecular-weight compounds like limonene and so on," The active binding sites for
metals are carboxylic acid groups in pectin substances. However, since the content of
carboxylic groups in the original OJR was unsatisfactory to strongly load enough iron(II1)
to adsorb oxo anions, the cellulose part of OJR was phosphorylated after crosslinking in
order to partially convert its abundanthydroxyl groups into phosphoric groups with high
binding affinity to ferric iron. A detailed chemical modification procedure that has been
already discussed in our previous paper' is representedby the following reaction equation.

+ Ho--P-OH "'ipIi~

The phosphorus content in phosphonylated OJR (POJR) was measured as the P20 S
mass% with the aid of Rigaku ZSX lODE fluorescence X-ray crystallography as 20%,
which is equivalentto 1.15 mol of phosphorus per glucose unit of cellulose.

2.1.2. Chemicals

AlI chemicals used in this study were of pure analytical grade purchased from
Katayama Chemical Co. Ltd. Japan, unless otherwise mentioned. Aqueous solutions of
iron, arsenic, selenium, and phosphorus were prepared by dissolving corresponding
analytical grade ferric chloride, arsenic trioxide, sodium arsenate, selenic acid, selenious
acid, and trisodium phosphate, as appropriately, in deionized distilled water and dilute
hydrochloric acid or sodium hydroxide solution.

2.2. Methods

Batch and column tests were replicated (three times) in order to assess the
reproducibility of the results. The variation was below 4% ofthe mean value .

2.2.1. Batch Test

In the batchwise tests, 25 mg of dried adsorbent was taken into a conical flask
together with a 15 mL of corresponding aqueous solution . The pH of the aqueous
solutions was adjusted by adding smaII amount of HCI or NaOH, while just in the case of
iron(1I1) adsorption, 0.1 M of 2-[4-(2-hydroxyethyl)-I-piperazinyl] ethanesulphonic acid
was used as the buffer reagent. The flask was shaken vigorously in a thermostated shaker
at 30°C for about 24 h to attain equilibrium . The initial and equilibrium concentrations of
the aqueous solution of iron, arsenic, and selenium after adsorption were measured by
using Shimadzu Model ICPS-5000 ICP/AES spectrometer or Shimadzu AA-6650 atomic
absorption spectrophotometer, while phosphate was monitored with the aid of DIONEX
DX-120 Ion Chromatograph.

2.2.2. Column Test

In the tests using a column , arsenic removal was carried out in a glass column of
8 mm diameter packed with 0.1 g of iron-loaded POJR. A schematic diagram of the
column experimental set up is shown in Figure I. The column was conditioned by
passing water of pH = 3.1 overnight. The sample solution containing 15.8 ppm of
arsenate whose pH was maintained at 3.1 was percolated into the column at a constant
flow rate of 6 cm'zh using a peristaltic pump (IWAKI PST-lOON, Japan) . Effluent
samples were collected at each one-hour interval of time by using a fraction colIector
(BIORAD Model 2110 Fraction ColIector). Effluent samples colIected at each time were
used for pH and arsenic concentration measurements.
For the elution tests, the column was pre-washed with deionized distilled water so as
to expel any residual arsenic by physical adsorption . A I M hydrochloric acid solution
used as the eluent was percolated into the column at the same constant flow rate of
6 cm3/h using the peristaltic pump. The concentration of the eluted arsenic effluents
colIected at each interval in the tube of fraction colIector was measured by using a
Shimadzu Model ICPS-5000 ICP/AES spectrometer or Shimadzu AA-6650 atomic
absorption spectrophotometer.

E'fl.-..t ()

A dsofboo n ' - - - - I. .

Cott er )

1 . Adsorotion colu rn n
3 2 . Micr 0 lulJ<l cum",
3 . Feed solution
4 . Fra c tion oolleclor 1 . A dsorp tio n c o lumn

Figure I. Schematic diagramof columnexperiment.


3.1. Batch Experiment

3././. Adsorption ofIron(lll)

ITOn(lII) was nearly quantitatively adsorbed from aqueous solution of an initial

concentration of 56 ppm iron as ferric chloride over the pH region of 1_3.9. 10 The
adsorption isotherm ofiron(lII) on P01R at various equilibrium concentration is shown in
Figure 2. The maximum iron uptake capacity was 3.7 mol/kg dry gel.

• qmax= 3.7mol/kg

t +-----~---~----~----~
0.000 0.004 0.008 0.012 0 .016
Equilibrium concentration [mol/l]

Figure 2. Adsorption isotherm of iron(lII) on POJR.

The high value of Fe(lII) uptake capacity by the POJR gel may be due to a
synergistic effect of additional carboxyl functionalities present within the polymer matrix,

which could accommodate more metal ions. After determination of the maximum loading
capacity, those corresponding gels were loaded with Fe(III) at their optimum
concentration and pH range in order to further investigate their adsorption characteristics
for oxo anions of arsenic , selenium, and phosphorus.

3.1.2. EffectofpH on Arsenic Distribution and Its Removal

Before going through the discussion about the adsorption characteristics for
arsenic(III and V), a simple picture of their chemical speciation calculated based on their
stability constants is outlined in Figure 3 11• The monovalent anionic species of arsenate
are dominant at a pH between 2 and 6, while in the case of arsenite they are significant at
alkaline region of pH = 9-12 .

1.0 1.0

0.8 0.8
I: I:
.~ .~
U 0.6 Uco
0.6 H3AsOJ
~ <l::
0.4 Q)
"0 "0 0.4
0.2 0.2

0.0 0.0
0 2 4 6 8 10 12 \' 0 2 4 6 8 \0 12 \4
pH pH

Figure 3. Distribution of arsenic(I11 and V) as a function of the pH.

Figure 4 shows the relationship between percentage removal and equilibrium pH in

the adsorption of arsenic(III and V) as AS20 3 and Na2HAs04 on iron(III)-loaded POJR
gel at the initial concentration of 15 ppm. It is seen from this figure that the pH of the
aqueous solution is an important variable for adsorption . It is considered likely that
iron(III) is adsorbed by releasing protons from the phosphorylated unit of cellulose
according to a cation-exchange mechanism. The adsorbed iron will be coordinated
octahedrally by hydroxyl ions and neutral water molecules available in the aqueous
medium . The adsorption of an As(lII or V) ion will take place by releasing a hydroxyl
anion from the above-mentioned coordination sphere . For this reason, the adsorption of
anionic species on metal-loaded gel is termed as "ligand-exchange adsorption"." From
Figure 4, the optimum pH condition for arsenate is mildly acidic (pH = 2-6), while that
for arsenite is weakly alkaline. The maximum removal was around 90% in the case of
arsenic(V) and 85% in the case of arsenic(III) .
Since the pH range where monovalent arsenic(V) species is predominant in Figure 3
is almost identical to that where high removal of arsenic(V) has been achieved as shown
in Figure 4, it can be concluded that an effective removal of monovalent arsenic(V) is
accompanied by using iron(III) loaded POJR gel. On the other hand, since monovalent
arsenic(III) ion is predominant only at pH > 8 in Figure 3, arsenic(III) adsorption on
POJR gel taking place at pH < 8 in Figure 4 is attributable to the adsorption of neutral

arsenic(lII) species, and at pH > 8 the adsorption of monovalent arsenic(lII) anions is

assumed to be taking place.
100 - , - - - - - - - - -- - - - -- - - - - - ----,
• • • ••
80 • 0

60 •

• 0

40 %0


20 0

0 0 • •
• As(V)
As (III)
I 0

0 2 4 6 8 10 12 14
pH e

Figure 4. Removal of arsenic(lll and V) on iron- loaded POJR.

• • ..
alRu - O.9 __mol /kg
qmax = 0 .91 mol /kg

• •
~ 0.6
• •

.§.. 0.4
0 .2 ..
• • As(V)
I • A s( III)

0 .004 0.006 0.008 0.0 I0 0.012 0.014

Equil ibrium concentration [m o l/L]

Figure 5. Adsorption isotherm of arsenic (Ill and V) on iron-loaded POJR.

Figure 5 shows the adsorption isotherms of As(lII) as AS20 3 and As(V) as

Na2HAs04 at pH 10 and 3.0 with respect to their respective equilibrium concentration.
The adsorption appears as a Langmuir-type adsorption. The maximum amounts of
adsorption of As(lII) and As(V) were evaluated as 0.91 and 0.94 mol/kg, respectively,
which is higher than 0.84 and 0.74 mol/kg, respectively . The latter two values have been
reported for the iron-loaded synthetic chelating resins.'

3.1.3. Effect ofpll on Selenium Distributionand Its Removal

Figure 6 shows the distribution profile of selenium(lV and VI) as functions of pH

based on their respective stability constants . 10-1 I Figure 7 shows the percent removal of
selenium(lV and VI) at the initial concentration of 15 ppm as H2Se03 and H2Se04,
respectively . The optimum removal condition for selenium(IV) is acidic within the pH
range 2-6 and a highest adsorption of 85% is achieved at pH = 3. On the contrary,
selenium(VI) is not effectively adsorbed, only up to the extent of 36% at pH = 3, as seen
in Figure 7. It means that in the case of selenium(lV) (as suggested by the distribution
diagram) only monovalent selenite oxo anions are adsorbed, while divalent selenate oxo
anions are adsorbed in the case of selenium(VI) . The maximum amount of selenium(lV)
adsorbed on iron-loaded POJR was evaluated as 0.51 mo/kg as shown in Figure 8.

10 -,--;:----?"'"~_::_--___;?'--.........., 10

SoO,' s.o'

08 .~ 06

o. ] o.


o 10 12 14 10 12 14

Figure 6. Distribution of seleniteand selenateas function of pH.


•0 Se(VI)
80 0 0 % I
60 0
E 40

;f 0
• •

0 • 0
• 0 •

o 4 6 8 10 12 14

Figure 7. Removal of selenium(IV and VI) on iron-loaded POJR.


qmax= 0.51 mollkg
0.5 • • •

'"iiii 0.4


0.000 0.002 0.004 0.006 0.008
Equilibrium concentration [molll]

Figure 8. Adsorption isotherm of selenium(lV) on iron-loaded gels.

3.104. Effect o/pH on Phosphate Distribution and Its Removal

Figure 9 shows the speciation of phosphate with respect to pH based on their

stability constants!' . Figure 10 shows the removal behavior of phosphate on iron(I1I)-
loaded POJR at the initial concentration of 46 ppm as Na3P04.12H20. The maximum of
97% removal has been achieved at pH -3, and the removal is significant up to pH = 6. In
accordance with the distribution diagram, monovalent phosphate oxo anions are adsorbed
on iron-loaded POJR. Figure II represents the adsorption isotherm of phosphate at pH =
3, from which the maximum loading capacity was evaluated as 0.83 mol/kg.
1.0 r-;::----,--::;:p--.:::-r--.--::=o-r-"'L"--r----,


'D 0.6
H2 P04 HPO/'

'0 0.4

0 .0
0 2 4 6 8 10 12 14

Figure 9. Distribution of phosphate as a function of pH.


• •
• •,

> 60




0 2 4 6 8 10 12 14
pH e
Figure 10. Removal of Phosphate on iron loaded POlR.

qma x = 0.8 3 mol /kg
• • •
0. 8

"0 0 .7 •

0 .6

0.000 0.002 0.004 0 .006 0 .008 0 .010
Figure I I . Adsorpt ion isotherm of phosphate on iron-loaded POlR.
272 K. INOUE £T AL.

3.1.5. Column Experiment

As a model experiment, adsorptive removal of As(V) using a column packed with

the iron(111)-loaded POJR has been tested . Figure 12 shows the breakthrough curve of
As(V) from the column at the initial concentration of 15.8 ppm. The breakthrough began
at 81 bed volumes , and the bed was completely saturated at 342 bed volumes.

•• ••• •
0.8 • ••
0.6 •
v<::> •


0 100 200 300 400
B. V.
Figure 12. Breakthrough profile ofarsenic(V)on iron loaded POlR.

Arsenate adsorbed onto the column bed can be recovered by eluting either with
alkaline or acidic eluent. Since stripping using alkaline eluent still leaves the As(V) to a
considerable extent in the column, it has been recovered using I M Hel as the eluent. The
elution profile is shown in Figure 13. As seen in Figure 13, the elution was completed
within 38 bed volumes and was concentrated 14 times that of the feed solution . Since
Fe(111) elutes together with arsenate, it is necessary to load iron(111) once again onto the
gel for the next operation.

3.1.6. PlausibleAdsorption Reaction Mechanism

As mentioned in the previous section , arsenic is postulated to be adsorbed on the

iron(111) immobilized onto POJR gel by the mechanism of ligand exchange . It is obvious
that the chemical species of iron(111) involved in the loading at acidic condition is the
hydrolyzed form of iron(111).13 Iron(111) is adsorbed on the POJR gel through a cation-
exchange mechanism, supported by the fact that the initial pH of the aqueous solution
had decreased after adsorption thereby leaving the proton in the aqueous phase. Both
dependences of the chemical species on pH in Figure 3 and their subsequent removal in
Figure 4 suggest that arsenic(111) removal at pH > 8 is attributable to monovalent arsenite
anions (ligand exchange), which can be represented by mechanism I below , while the






o 10 20 30 40
Figure 13. Elution profile ofarsenic(V) from iron loaded POJR.

adsorption at pH < 8 is attributable to their neutral species accompanied by the release of

a water molecule according to mechanism 2. In the case of arsenic(V), monovalent
arsenate anions are considered to be adsorbed according to the mechanism I as evidenced
by the distribution diagram in Figure 3. In the same manner , monovalent selenite is
prevalent at acidic condition in accordance to its distribution diagram, and its maximum
removal (pH = 2-6) is in harmony with mechanism 1. Subsequently, the distribution
diagram of phosphate and its maximum removal pH also suggest that monovalent
phosphate oxo anions are adsorbed by the same mechanism.
The lower loading capacities for oxo anions on iron(III)-loaded POJR may be due to
the coulombic interactions between the hydrolyzed iron surface and the oxo anions . On
the other hand, since selenate species at around pH = 2-6, where some adsorption has
taken place, is the highly hydrophobic divalent selenate oxo anion, the iron-loaded POJR
is insignificant for its removal. The same argument applies for the decrease in oxo anion
removal where their corresponding divalent anionic species are prevalent (Figures 4, 7,
and 10).


A novel adsorption gel for the removal of oxo anions has been explored by making a
simple chemical modification of orange-juice residue . In the treatment of arsenic in water,
there is no need to oxidize trivalent arsenic into the pentavalent form, which is inevitable
in the conventional methods . Direct removal behavior and attractive economics of the
POJR made it suitable for various types of water treatment-processes; in addition , it is
more efficient than the commercially available synthetic chelating resins . The most
advantageous aspect of this technology is that a free biomass waste can be converted into

Loading mechanism.

+ OH


valuable resources for the remediation of polluted environment. Hence, the technology
offered in this paper could be potentially applied for oxo anions removal from aqueous


The present work was conducted with the aid of the financial support from Japan
Science Technology. Corporation as the Modeling Project for New Concept of
Technology. We would also like to acknowledge Saga Ceramics Research Center
(SCRC), Arita, Japan for their kind assistance in providing X-ray crystallography

1. T. Shuzo and U Akira, A Reviewon Waste Treatment and Disposal (Japanese), Busshitsu Kogaku Kogyo
Gijutsu Kenkyusho Hokoku S(I) , 2 I-38 (1997).
2. A. H. Mirza and V. Ramachandran, In Removalof Arsenic and Seleniumfrom Waste Waters- A Review,
Second International Symposium on Extraction and Processing for the Treatment and Minimization of
Wastes, Salt Lake City, Utah, Edited by V. Ramachandran and C. C. Nesbitt, (The Minerals, Metals &
Materials Society, 1996), pp. 563-582.
3. R. W. Lawrence and T. W. Higgs, JOM, 27-29 (1999).
4. I. Yoshida, K. Ueno, and H. Kobayashi, Sep. Sci. Technol. 13(2), 173-184 (1978).
5. H. Egawa, T. Nonaka, and H. Maeda, Sep. Sci. Technol. 20(10), 653-664 (1985).
6. M. Chanda, K. F. Odriscoll, and G. L. Rampel, React. Polym. 8,85-95 (1988).
7. H. Matsunaga, T. Yokoyama, R. J. Eldridge, and B. A. Bolto, React. & Func. Polym . 29,167-174 (1996).
8. S. Nagy, P. Shaw and M. K. Veldhuis, Citrus Science and Technology, Nutrition. Anatomy. Chemical
Composition and Bioregulation, (The AVI Publishing Company, lnc., Westport, Connecticut 1977), Vol. I,
9. K. N. Ghimire, K. Inoue, K. Makino and T. Miyajima, Adsorptive Removal of Arsenic Using Orange
Juice Residue, the IJh Internationalconference SIS '01, (Bratislava, June 5-10, 200 I).
10. K. N. Ghimire, K. Inoue, K. Makino and T. Miyajima, Adsorptive Removal of Arsenic Using Orange
Juice Residue, Sep. Sci. Tech. , 37( 12), 2785-2799 (2002).
11. A. J. Bard, R. Parsons, and J. Jordan, Standard Potentials in Aqueous Solutions (Marcel Dekker, New
York, 1985).
12. R. C. West, ed. Handbook of Chemistry and Physics. 67'h ed. (Chemical Rubber Publishing Co.,
Cleveland, Ohio, 1986).
13. F. A. Cotton, and G. Wilkinson, AdvancedInorganic Chemistry, 5'" ed. (John Wiley & Sons: New York,
1988), pp. 717.

Jan D. Miller, Gustavo A. Munoz, and Saskia Duyvesteyn


Gold is dissolved in alkalin e cyanide solut ions in the presence of oxygen to form the
aurodicyanide anion comple x, Au(CNh', accord ing to the follow ing stoich iometric
electrochemical reactions (Marsden and House, 1993):



The aurodic yanide anion is the only cyanide complex with Au+ and has a stability
constant ~2 of 10 . The high stability of the Au(CNh ' complex is the basis for the
industrial gold recovery process currentl y used worldwide , as originally patented by
MacArthur and the Forrest brothers (Sharpe , 1976).
Conventional gold-recovery methods involve crushing and grinding of the gold ores
followed by dissolution of the gold in an oxygenated alkaline cyanide solution . Gold,
dissolved as the Au(CN h ' comple x, is recovered from solution by zinc precipitation,
electrowinning, or by one of several methods using activated carbon adsorption . After
separation of the activated carbon from the pulp or solution , the adsorbed Au(CN h '
complex is stripped from the activated carbon for subsequent precipitation and refining .
Figure I is a schematic representation of a conventional carbon-in-pulp (CIP) or carbon-
in-leach (CIL) countercurrent circuit for gold adsorption by activated carbon. In the case
of the CIP process, gold leaching by alkaline cyanide solutions is performed in a series of
reactors (not shown in Figure 1). After the gold has been dissol ved, the slurry is

Jan D. Miller, Gustavo A. Munoz, and Saskia Duyvesteyn, Department of Metallurgical Engineering,
University of Utah, 135S 1460 E Rm 412, Salt Lake City UT 84112-0114, USA.

Fundamentals and Applications ofAnion Separations, edited by

Moyer and Singh, KluwerAcademic/PlenumPublishers,New York, 2004 277

Slurry flow after/du ring cyanide

leaching of gold (CIP/CIL)

Activated carbo n flow


Gold-loaded activated carbon Gold-free slurry

Figun 1. Schematic representation of a conventional carbon in pulp (CIP) or carbon in leach (CIl)
countercurrent circuit for goldadsorption by activated carbon.

transferred to the CIP tanks shown in Figure I. In each tank, the slurry containing the
dissolved gold is contacted for a specific time in countercurrent direction with activated
carbon. By way of contrast, in the case of the CIL process, activated carbon is in contact
with the slurry during the cyanide leaching stages; thus, dissolution of gold and
adsorption of the aurodicyanide complex by activated carbon occurs in the same leaching
During 1999, the amount of activated carbon purchased for gold recovery in the
United States and the rest of the world was estimated to be 4 thousand and 7 thousand
metric tons respectively, which represents sales on the order of $8 million and $14
million respectively (Faulkner, 2000). The United States produced 340 metric tons of
gold from mining operations during 1998, of which 78% was recovered by the use of
activated carbon, as shown in Figure 2 (Marsden and House, 1993; US Geological
Survey, 2000). The remaining 22% was recovered by other methods.
From a fundamental point of view, activated carbon has a disordered
microcrystalline structure, similar to graphite. Crystals of pure graphite consist of
parallel layers of condensed regular hexagonal rings spaced at 0.335 nm. The c-e bond
distance in individual layers is 0.142 nm (Jankowska et al., 1991). This random
organization of graphite microcrystals in activated carbon creates a porous structure,
which accounts for the high the surface area of activated carbon and its notable
adsorptive capacity. A significant amount of research has been carried out to explain the
mechanism of gold adsorption by activated carbon, the nature of the adsorption sites, and
the role of specific surface functional groups in gold adsorption (McDougall et aI., 1980;
McDougall and Hancock, 1981; Jones et aI., 1989a, 1989b, 1989c, Ibrado and
Fuerstenau, 1992; Sibrell and Miller, 1991, 1992; Lagerge et aI., 1997, 1999). Currently,
the most accepted theories regarding the mechanism of gold adsorption include the
adsorption of ion pairs Mn+-[Au(CNh'] onto less active sites of activated carbon. This
ion-pair adsorption process is somewhat selective due to the structure of the
aurodicyanide anion, which is considered to be less hydrated than other cyanoanions.

Car bon in Pulp: 36 %

Carbon in
Column: 27 %

Othe rs : 9 %

Zn prec ipitation: 13 %
Carbon in Lea ch: 15 %

Figure 2. The United States produced 340 metric tons of gold from mining operations during 1998, from
which 78% were produced by the use of activated carbon. Zinc precipitation and other methods produced the

On this basis, the selective solvent extraction of gold from alkaline cyanide solutions has
been previously explained (Mooiman and Miller, 1984; 1986; Miller et aI., 1987). In
addition, adsorption of the unpaired anion Au(CNh - occurs through electrostatic
interactions at highly active sites of activated carbons that have appropriate polarity
(Lagerge et aI., 1997, 1999). In terms of the adsorption sites, the most accepted theory
suggests that gold is adsorbed by the edge sites rather than at sites on the basal planes
(Sibrell and Miller, 1991, 1992; Lagerge et aI., 1999). Regarding the role of surface
functional groups on the adsorption of activated carbon, researchers suggest that the
presence of basic surface functional groups, specifically the pyrone type, favors the
adsorption of gold onto activated carbon (Papirer et aI., 1987, 1991, 1995; Polania-Leon
et aI., 1993).


Conventional activated carbon circuits for gold recovery use granular activated
carbon particles that are significantly coarser than the ground ore in order to produce an
effective separation of the gold-loaded activated carbon from the slurry phase by
mechanical screening . Mechanical screening is the simplest method for separation of
particulate solids of this size, and it is based solely on the difference in particle sizes.
Separation of activated carbon from the slurry phase by screening has significant
problems. The screens may blind due to extended use, break due to excessive wear,
require considerable horsepower to operate, and need frequent maintenance and screen
replacement. Also, gold loaded onto the exposed carbon may be lost by abrasion. Losses
of activated carbon due to these factors can be as high as 50 grams of carbon per metric
ton of ore, (Marsden and House, 1993) and cumulative losses of fine activated carbon
within a year of operation can be high, with replacement of as much as 20% annually
(American Norit Company, J 989). This represents significant amounts of gold adsorbed

by activated carbon that is not recovered by screening. It is important to note that fine
powdered activated carbon shows a significant increase in gold adsorption kinetics and a
modest increase in gold adsorption capacity at pseudoequilibrium (24 hours of adsorption
time), as shown in Figures 3 and 4. However, powdered activated carbon currently
cannot be used in gold recovery circuits, since both the carbon and the solids in
suspension will have particle sizes of similar magnitudes; thus, separation of the gold-
loaded activated carbon from the slurry phase by screening, filtration, and sedimentation
are not viable options.


Activated carbon is a generic term for complex polymeric compounds where carbon
is the main component. It is used extensively for purification of aqueous solutions and
gases due to its high capacity for adsorption. The adsorptive characteristics of activated
carbon are due to its highly developed pore structure, extended surface area, and surface
functional groups that are generated during the production process (Bansal et al., 1988).
Activated carbon is often preferred over other more sophisticated purification
products, such as ion-exchange resins and reverse-osmosis membranes, mainly because it
is less costly and easier to regenerate (Silverberg, 1999). The United States was the
world's largest activated-carbon producer during 1999 at 190 thousand metric tons
(Silverberg, 1999). Market studies reveal that demand for activated carbon in the US,
Europe, and Japan has been increasing at a rate of about 6% per year during the last
thirteen years, and this rate is expected to increase (Roskill Consulting Group , 1998;
Freedonia Group, 1998).
Activated carbon can be produced from a variety of carbonaceous materials,
including bones, coals, wood dust, peat, nutshells, and wood charcoal (Jankowska et aI.,


~ 60
~ Particlesize, 11m

~ 0 116
20 0 325
Do 2375

0 4 8 12 16 20 24
Time (hr)

Figure 3. Particle size effect on gold adsorption kinetics by commercial activated carbon. Adsorption
cond itions: I gil activated carbon, 10 mgIL Au, 0.0 1 M NaCN, pH 11,200 rpm, 298 K.

~ 100
rl 0

OJ 10
Carbon size Adsorption capacity

o 1,400 - 3,350 urn o 40 Kg Au/ton C

o 53 - 106 ~m o 54 Kg AulTon C


0.01 0.1 10 100

Au in solution (mg/L)

Figure 4. Panicl e size effec t on gold adsorption capacity by comm ercial activated carbon. Adsorpti on
conditions 1-5 gIL activa ted carbon . 100 mglL Au, 0.01 M NaCN. pH II . 200 rpm. 298 K. 24 hours of
adsorption time.

There are several ways to produce activated carbon . For gold adsorption
appli cations, the two-step pyrolysis-gasification process is the preferred method of
Pyrolysis, or carbon ization, of the raw material is typically performed in an inert
atmosphere at a temperature between 700 and 1100 K. During pyro lysis, most of the
noncarbon elements are remo ved as gases, with the release of some of the carbon, mainly
as methane (CH4 ) , acetic acid (CH3COOH), carbon mono xide (CO), and carbon dioxide
( C O~ ). Carbon rearrange s into organized crystallographic formations known as
elementary graphitic crystallites or microcrystallites. The arrangement of these
microcrystall ites is irregular and free interstices rema in betwe en them . Tars that are
formed durin g the pyrolysis process tend to fill and block these interstices. The
carbonaceous produ ct obtained has a small specific surface area, and low adsorpti on
capacity (W igrnans, 1989; Sw iatowki, 1999). The produ ct formed during pyrolysis is
called carbon or char.
Carbonization is followed by activation, durin g which the extensive microporosity of
the activated carbon is produ ced . Activation is done by reacting the carbonaceous
product with an ox idiz ing gas such as water vapor, carbon dioxide, or oxygen at
temperatures between 1100 and 1300 K. In some cases , the carbonization and activation
can be carried out simultaneo usly. Typically, products have specific surface areas of 800
to 1500 m~/g with averag e pore sizes rang ing from 0.5 to 20 nm. The physical properties
of the acti vated carbon are highly dependent on the initial raw material.

The stoichiometr ic reaction of carbon with water vapor is


The overall reaction of carbon with carbon dioxide is

C +C02~2CO , [4]

and the overall reaction of carbon with oxygen is

C +02~C02 ' [5]

At pract ical activation conditions, which are above 1100 K, the reaction of oxygen
with carbon is about 100 times faster than the reactions of carbon with steam or carbon
dioxide. This leads to a diffusion-controlled reaction with extensive consumption of the
carbon that is located on the surface, and as a result, there is no development of porosity
in the carbon . For these reasons, activation with oxygen is avoided.


The concept of magnetic activated carbon (MAC) addresses several of the

technological limitations brought up in the previous discussion. It can be easily separated
from solution, even if the solution contains significant concentrations of solids , using a
wet magnetic separator, as opposed to the traditional screening technology. By
producing a smaller sized carbon that is magnetic , the kinetics and capacity of gold
adsorption can be improved.
The use of powdered magnetic activated carbon for gold recovery increases
significantly the loading kinetics of gold on MAC and is expected to provide important
savings in equ ipment and operating costs. The use of magnetic activated carbon for other
applications, such as wastewater treatment, will have a notable impact on current
operations. Since it is not necessary to remove any solids in the stream before contact
with the MAC, not only will the types of water and effluent streams that can be treated
increase significantly, but current treatment cost could also be expected to be reduced.
The feasibility of producing a magnetic activated carbon can be explained using
equilibrium diagrams for the Fe-O-C and Fe-O-H systems. These diagrams are well
known and are extensively used in the analysis of the gas-solid reactions involving the
reduction of iron ores to yield metallic iron.
Figure 5 shows the equilibrium between iron, oxygen, and carbon at 1 atmosphere
total pressure. The predominance regions of Fe, FeO, and Fe)04 are labeled accordingly.
The equilibrium line between CO and CO 2 that describes Eq. [4] is also included .
Regarding the gasification equation given in Eq. [4], at temperatures of about 1000 K
and above, equilibrium favors the formation of CO, and at temperatures above 1100 K,
conversion of C to CO is greater than 90% . An atmosphere rich in carbon dioxide will
favor the gasification of the carbonaceous material, due to Le Chatelier's principle. From
Figure 5 it can also be seen that the same atmosphere rich in CO 2 will favor the formation
of magnetite, Fe)04' The analysis suggests that if carbon is in contact with an iron oxide
at a suitable temperature and under an atmosphere rich in CO 2, these are optimum
conditions for the gasification of carbon , and the formation of magnetite as the iron
species. A similar analysis can be done for carbon gasification with water vapor and the
Fe-O-H system.

1.0 0.0

0.8 0.2

I I.e
0.6 0.4

0 ON
0.4 0.6 ~

0.2 08

EO C + cO 2 = 2CO
0.0 1.0
300 600 900 1200 1500

Temperature (K)

Figure 5. Equilibrium diagram for the Fe-O-C system and for the C + CO, = 2CO reaction .


A number of synthesis procedures for powdered MACs for gold recovery from
alkaline cyanide solutions were developed at the University of Utah (Miller et aI., 2001) .
The synthesis procedure described below is an example of one of these procedures.

5.1. Synthesis of MACs

5.1.1. Preparation of Material for Carbonization. A pinewood board, without the bark,
was processed in a sawmill until it was reduced to shavings and dust. The shavings were
further processed using a Gilson micro-hammer cutter-mill , and the size fraction 600-
1000 11m obtained from both the dust and the milled shavings was selected .

5.1.2. Impregnation. Two batches of 200 g of pinewood of particle size 600-1000 11m
were immersed in 1.5 L of solutions containing 2.5 and 5 g Fe3+/L as FeCb respectively,
and the systems were agitated inside a 2-L plastic bottle using a VWR orbital shaker at
200 RPM for 24hr. The solutions were removed from the solid wood by vacuum
filtration using Wathman paper No 1. The solid residues were then dried in an oven at
333 K for 24 hr. Products were labeled as "low Fe" and "high Fe."

5.1.3. Pyrolysis and Gasification. For pyrolysis and gasification, a Lindberg 3-chamber
tube furnace was used. In the case of pyrolysis, the tube furnace was purged with N2 for
30 minutes before heating was performed. The N2 flow was fixed in 1.5 I/min measured
at 273 K and I atm (STP) . The raw material was placed in the tube furnace at room
temperature, and the temperature was raised to 900 K at a heating rate of 12K/min. The
material was kept at the furnace for a total time of I hour. The pyrolyzed material was

transferred to another tube where it was cooled down to room temperature at a rate of 60
Klmin under a fixed N 2 flow of 1.5 Llmin.
Gasification was done using CO2 at a fixed flowrate of 1.5 Llmin measured at 273 K
and I atm (STP), until 21% and 45% of bumoff was accomplished . Bumoffis defined as
the weight percent that is lost during the gasification process.

5.1.4. Size Reduction and Classification . The activated chars obtained were wet-ground
and sieved, and the fraction 53-106 urn was used for characterization .
Figure 6 shows a schematic representation of the flowsheet for MAC production. A
nonmagnetic char and nonmagnetic activated carbon are included for comparison

5.2. Characterization

5.2.1. X-ray Diffractometry. XRD analyses were performed using a Rigaku DMAX
2200V diffractometer, using a copper anode. The diffractometer was operated at 40 kV
and 30 rnA, with a start angle (28) of 5° and a stop angle (28) of 80°. The step size was
set at 0.05°, scan speed at 2 deglmin, and step time in 1.5 seconds. The divergence slit
width was set at 1.0, scatter slit width at 1.0, and the receiving slit width was set at 0.15.
Powdered samples were mounted in a glass sample holder.

5.2.2. Magnetometry. A Lakeshore vibrating magnetometer was used for determination of

magnetic parameters such as magnetic susceptibility , saturation magnetization , remanent
magnetizat ion, and coercivity , among others. Analyses done included the complete
magnetization curve in a magnetic field range (H) from - 1.5.104 to 1.5.104 Oe.

5.2.3. Specific Surface Area. A Micromeritics Accelerated Surface Area and Porosimetry
(ASAP) Analyzer model 2010 was used to determine the specific surface area of chars
and activated carbons . Weighed solid samples were placed in an analysis tube and
degassed at 500 K and a vacuum of 5 urn for 4 hours to remove volatile contaminants and
moisture. The tube with the sample was then transferred to the analysis port where it was
analyzed using N 2 as the adsorbate . Specific surface area was calculated by application
of the BET equation in the relative pressure (PIPo) range ofO.05-{).3.

5.2.4. Gold Adsorption Kinetics. A O.I-gram sample of carbon was placed in plastic
bottles containing 100 mL of a 10 gIL of NaAu(CN)2 solution at pH II and at room
temperature (296 K). Bottles were sealed and stirred for 24 hours. After certain time, a
bottle was opened, filtered, and a 20-mL sample was transferred to a glass vial for gold
determination by ICp·ES.

5.2.5. Recovery ofMACs by Wet High Intensity Magnetic Separation (WHIMS). Selected
MACs were recovered from solution using a laboratory scale wet high intensity
separator, at a constant solution flowrate of I mLisec, I g MACIL, and a current intensity
of 3 amperes.


1 t 1

Figure 6. Schematic representation of the flow sheet for MAC production .


Figure 7 shows X-ray diffractograms of pinewood, the char formed during pyrolysis,
and the activated carbon formed during gasification to 21% bumoff. The broad peaks in
the 28 region of 10-30 degrees are an indication of a highly amorphous material,
characteristic of the short-range order species such as the lignocellulosic material and the
chars being analyzed. Figure 8 shows an X-ray diffractogram of pinewood impregnated
with ferric chloride in a solution containing 5 g Fe3+/L, the magnetic char formed during
pyrolysis, and the magnetic activated carbon formed during gasification to 21 % bumoff.
The broad peaks in the 28 region of 10-30 degrees found in the nonmagnetic materials
are also found in the magnetic materials, but the diffractogram of pinewood char shows
sharp peaks at the 2e region of 35.42°, 62.51°, and 56.94°, that are characteristic of
magnetite. Hematite, maghemite, or elemental iron or other iron species were not
detected in the analysis, at least within the detection limits of the X-ray diffractometer.
Figure 9 shows the magnetization curves for an activated carbon with no iron, a
MAC with low Fe. and a MAC with high Fe content, both at 21% bumoff. It can be seen
that the activated carbon shows diamagnetic behavior, and the two Fe-containing MACs
have ferromagnetic behavior. These magnetization curves are also useful to classify the
MACs as soft or no-permanent magnets due to the low remanent magnetization that they
exhibit. This property, together with the low remanent magnetization of the magnetic
matrix, shown in Figure 10, makes it possible the magnetic separation using a WHIMS,
as shown in Figure 11. Figure 12 illustrates a single-stage CIP/CIL circuit for gold
recovery using MAC.

::i Activated carbon



o 20 40 60 80

Figure7. X-ray diffractograms of pinewood, the char formed during pyrolysis, and the activated carbon
formedduring gasificationto 21% bumoff.

Magneticactivated carbon



o 20 40 60 80

Figure8. X-ray diffractogram of pinewood impregnated with ferric chloride iron in a solution containing 5 g FeJ+/L, the
magneticchar formedafter pyrolysis,and the magneticactivatedcarbon formedafter gasificationto 21% bumoff.

MAC(high Fe)

1~ 0.5 MAC(low Fe)

g- 0.0 I - - - - - - - - - - " " " "......~--------_i


·10xl0 3 ~~~ 0 5xl~ 10xl0 3

Magnetic Field, H (Oe)

Figure9. Magnetization curves for an activatedcarbon with no iron and two magnetic activated carbons with
differentlevelsof ironcontent All sampleshave been activatedto 21% burnoff.

x ; 3.63 X 10-2 (egs)

100 J. = 104emulg
J, 1.54 emulg
~E 50
-Hc= 26.4 Oe

§, -50

-10x103 -5x103 0 5x103 10x103

Magnetic Field, H (Oe)

Figure10. Magnetization curves for steel wool used as magneticmatrix.

288 J. D. MILLER £T AL.

WH IMS feed
(Gold-free slurry plus
gold-loaded MAC)
Magnet ic field (H)
. J.....£
, 't ~ I L ~

WHIM S discharge
(Gold-free slurry)
MAC particles captured
on the magnetic matrix

Figure 11. Schematic representation of magnetic separation usinga WHIMS.

Slurry (from cyanidation circuit)

Gold-free slurry and +-- MAC

gold-loaded MAC


Gold-loaded M
+-----' '--.........
Gold-free slurry

Figure 13 shows the effect of burnoff degree and iron content of the different chars
on gold adsorption kinetics for a particle size range of 53-1 06 urn. It can be seen that for
no activation, the nonmagnetic char and the two magnetic chars, the amount of

• 45% bumoff
(53 - 106 j1IT1)

21 % bumoff

- (53 · 106 j1IT1)

.... .....
.. ... . .
• • • • Conventional activated carbon
(1,400 -3,350 I'm)

No activation

(53 - 106 j1IT1)

4 6 8
Time (h)

Figure 13. Effect of bumoff degree and iron content of the different chars for a panicle size range of 53-106
I'm on gold adsorpt ion kinetics . Results for a conventional activated carbon with a panicle size of 1,400 -
3,350 I'm is included for comparison Adsorption conditions: 1 g/L activated carbon, 10 mgIL Au, 001 M
NaCN , pH 11,200 rpm, 298 K

aurodicyanide complex that is removed from solution is very small. For the case of21%
burnoff, the system does not reach equilibrium after 8 hours of adsorption , and the
aurodicyanide complex uptake from solut ion is similar for the activated carbon and the
two MACS with different iron content, reaching a level of about 60% after 8 hours. For
the activated carbon and the two MACs activated to 45% burnoff, adsorption of gold in
solution is almost 100% in less than I hour of adsorption. The results for a conventional
activated carbon with a particle size of I,400-3,350 IJm is included for comparison. It can
be seen the conventional activated carbon is less effective, when compared to the MACs
at 21 and 45% burnoff.
Table I shows the results for burnoff, BET surface area, gold adsorption , and
WHIMS recovery for samples after 2 hours of adsorption. It can be seen that for the char
and the two magnetic carbons, there is an increase in gold adsorption and in BET surface
area with an increase in percent burnoff. Carbon recovery with WHIMS for the
nonmagnetic carbon was not attempted . For the two magnetic activated carbons , with
different levels of magnetite, recovery from solution was greater than 98%.


Magnetic activated carbons have been obtained by mixing a magnetic precursor

(FeCI) with a carbon source (pinewood), and heat treating the mixture under controlled
conditions of temperature and gas composition. XRD analysis shows that magnetite

Table 1. Results for gold adsorption, WHIMS recovery and surface area using
carbons synthesized from pinewood.

Sample Fe Burnoff BET surface Au Adsorption WHIMS

(%) area (m2jg) % Rec.%
C None 0 250 4.6 0
AC None 21 480 50.1 0
AC None 45 890 89.5 0
MC Low 0 250 3.9 98.9
MAC Low 21 490 55 99.7
MAC Low 45 870 92.1 98.6
MC High 0 220 5.3 98.3
MAC High 21 470 54.1 99.4
MAC High 45 870 95.1 99.2
C: char; MC: magneticchar; MAC: magneticactivatedcarbon

(Fe304) is the main magnetic species in the MAC. Magnetization curves show that MAC
has paramagnetic characteristics, and behaves as a soft magnet. Gold adsorption curves
reveal that MAC has faster kinetics for the adsorption of the aurodicyanide anion
complex, due to the smaller particle size, when compared to a conventional activated


Bansal, R. C; Donner, J. B., and Soeckli, F., 1988, Active Carbon Marcel Dekker, Inc., New York,) p. 482.
Faulkner, w., 2000, PACCD International, personalcommunication.
FreedoniaGroup, 1998, ActivatedCarbonto 2002. Study, p. 1065.
lbrado. A. S., and Fuerstenau, D. W., 1992, Effect of the structure of carbon adsorbents on the adsorption of
gold(l) cyanide, Hydrometallurgy 30:243-256
Jankowska, J., Swiatkoski, A., and Choma, J., 1991 , ActiveCarbon Ellis Horwood, England,p. 280.
Jones, W. G., Klauber, C., and Linge, H. G., I 989a, Fundamental aspects of gold cyanide adsorption on
activated carbon, Gold Forum on Technology and Practices - WorldGold '89, Proceedingsof the First
Joint International Meeting Between SME and AusIMM, Reno, Nevada, November 5-8, SME, Littleton,
Colorado, pp. 278-281 .
Jones, W. G., Klauber. C; and Linge, H. G., 1989b, Loading of Au(CNh' solution onto activated carbon, The
AuslMMAnnualConference, Perth-Kalgoorlie, WesternAustralia, May I989b, pp. 225-228.
Jones, W. G., Klauber, C., and Linge, H, G., 1989c,The adsorptionchemistry ofgold(l) dicyanide on activated
carbon, Nineteenth BiennialConference on Carbon. Penn State, June 25-30, pp. 38-39.
Lagerge, S., Zajac, J., Partyka, S., and Groszek A. J., 1999, Comparativestudy on the adsorption of cyanide
gold complexesonto differentcarbonaceoussamples: Measurement of the reversibility of the process and
assessmentof the active surface inferredby flow microcalorimetry, Langmuir 15:4803-4811 .
Lagerge, S., Zajac, J., Partyka, S.. Groszek, A. J., and Chesneau M., 1997, A two stage adsorption of cyanide
gold complexes onto activated carbon inferred from various experimental studies, Langmuir 13:4683-
Marsden,J. and House, I., 1993, TheChemistry ofGold Extraction, Ellis Horwood, New York, p. 597.
McDougall, G. J. and Hancock, R. D., 1981 , Gold complexes and activated carbon, a literature review, Gold
Bulletin 14(4):138-153.

McDougall, G. J., Hancock, R. D., Nicol, M. i, Wellington, O. L., and Copperthwaite R. G., 1980, The
mechanism of the adsorption of gold cyanide on activated carbon, The Journal of the South African
Institute ofMiningand Metallurgy 80(9):344-356
Miller, J. D., Munoz, G. A, and Duyvesteyn, S., 2001 , Magnetic Activated Carbon Particles for Adsorption of
Solutes from Solution, February 25,200 I, Provisional PatentNo. U2975.1P.
Miller, J. D., Wan, R. Y , Mooiman, M. B., and Sibrell, P L., 1987, Selectivesolvation extraction of gold from
alkalinecyanide solution by alkil phosphorous esters, Separation Scienceand Technology 22:487-502.
Mooiman, M. B. and Miller, J. D., 1984,Selectivity considerations in the amine extractionof gold from alkaline
cyanidesolutions, Minerals and Metallurgical Processing 153-157.
Mooiman, M. B. and Miller, J. D., 1986, The chemistry of gold solvent extraction from cyanide solution using
modified amines, Hydrometallurgy 16:245-261 .
AmericanNoritCompany, 1989, Inc., Norit" , RO 3515 Bulletin. p 8.
Papirer, E., Dnetzer, L, Li, S., and Donnet, l-B., 1991 , Surface groups on nitric acid oxidized carbon black
samplesdetermined by chemicaland thermodesorption analysis, Carbon 29( I ):69-72.
Papirer, E., Li, S., and Donnet, l-B., 1987, Contribution to the study of basic surface groups on carbons,
Papirer, E., Polania-Leon, A., Donnet, J.-B., and Montagnon, P., 1995, Fixation of potassium aurocyanide on
active carbons, Carbon 33(9):1331-1337
Polania-Leon, A., Papirer, E., Donnet, l-B., and Dagois G., 1993, Modification et interaction des fonctions
oxygenees en surfacedes charbonsactifs, Carbon31(3):473-479.
Roskill ConsultingGroup, 1998. The Economics ofActivatedCarbon.
Sharpe, A. G., 1976, The Chemistry of Cyano Complexes of the Transition Metals. Academic Press, London,
p.302 .
Sibrell, P. L. and Miller, J. D., 1991, The search for adsorbedgold cyanideon carbon surfaces, WorldGold '91,
The AusIMM, pp. 21-25.
Sibrell, P. L. and Miller, J. D., 1992,Significance of graphitic structural features in gold adsorption by carbon,
Minerals and Metallurgical Processing November, pp. 189-195.
Silverberg, P., 1999, Activated carbon, for the "hole" cleanup job," Chemical Engineering December, pp. 33-
Swiatowki, W I., 1999, Industrial carbon absorbents, Adsorption and its Application in Industry and
Environment Protection. Studiesin SurfaceScienceand Catalysis. 120A:69-94.
U.S. Geological Survey, 2000, MineralCommodity Summaries. February.
Wigmans, T., 1989, Industrial aspectsof production and use of activatedcarbons, Carbon 27(I):13-22.

Masamichi Tsuji'


Ionic constituents present in environmental electrolyte solutions, such as groundwater,

seawater, and river water, are involved in ion exchange reactions with in-contact
environmental inorganic solids such as, clay minerals, oxides and hydroxides of multivalent
metals, and metal carbonates. These ionic substitution reactions could be effectively
described using thermodynamic principles, provided that the systems have reached
equilibrium. Unfortunately, chemical thermodynamic expressions are often not utilized for
the evaluation of ion-exchange processes involving environmental solids.
Typically, distribution coefficients (Kd) , ion exchange and/oradsorption isotherms, suchas
Langmuir equation and other empirical equations, are used for evaluating ionic reactions
between environmental solids and waters.!" However, these evaluation methods require
numerous time-consuming distribution experiments to understand biphasic distribution
equilibria of ionic species. Moreover, these techniques are not suitable for describing the
equilibria of non-ideal systems, particularly the equilibriums involving highly selective solid
Adsorbents exhibiting high selectivity are desired for effective separation of trace
concentration of ionic constituents from highly concentrated electrolyte solutions. Solid
extractants with highly selective anion-exchange properties may be referred to as anion sieves.
These anion sieves are requisite for the removal of toxic anionic species (such as the anions
of selenium, arsenic, chrom ium and fluorine) from industrial effluents, groundwater used for
drinking and nuclear waste streams. This is because considerably high concentration of
ubiquitous sulfate, chloride and other anions present in aqueous matrices strongly interfere in
the uptakeof minor and trace components by solid extractants. Recentstringent legislative
control on discharging toxic constituents into the aqueous environment has necesseciated the
development of anionsievesto obtain high level of separation.

'TokyoInstituteofTechnology, Research CenterforCarbonRecycling and Energy, 2·12-1, Ookayama, Meguro-ku,

Tokyo 152-8552, FAX: +81-3-5734-3337, E-mail:

Fundamentals andApplications of AnionSeparations, edited by

Moyer and Singh, KluwerAcademic/Plenum Publishers, NewYork, 2004 293
294 M. TSUJI

The distribution of ionic constituents strongly depends on the chemical environment of

the reactive sites in inorganic solid adsorbents. Structural properties of these compounds
(inorganic solids). which make them useful in selective separation of ionic species. can be
well characterized by analytical techniques. such as MAS-NMR, ESCA, FTIR, and XRD.
These probesare very useful in the interpretation of the chemical reactivity of these inorganic
ion exchangers.
This chapter provides an overview of the research recently undertaken to design novel
molecular anion sievesfor their use in a variety of separation applications.


2.1. Inorganic Solids and Modes of Anion Uptake

The following three types of inorganic solids showing ion-exchange properties can be
obtained either naturally or via synthetic methods. This new classification, which is based
on the charging mechanism, is recently proposed by us 9 and provides useful information in
designing solids for a particular application.

(1) Variable-charge type compounds: oxides, oxyhydroxides, and acidic salts of metals, such
as activatedalumina, titanicacid, aFeO(OH), and zirconium phosphate.
(2) Substitution type compounds: layered double hydroxides (LDHs), aluminosilicates (clays,
zeolites, and geopolymers), tobermorites, and nonstoichiometric mixed-valence
compounds, e.g., Mn02,.S and theirsalts.
(3) Intrinsically charged type compounds: hydroxyapatites and calcium silicates including
wollastonite and xonotlite.

Of these compounds, oxidesloxyhydroxides of multivalent metals and LDHs are

well-known anion-exchange compounds. The oxides and oxyhydroxides of multivalent
metals are weakly basic anion exchangers. Uptake of the anions by these compounds is
assumed to occur through a hydrogen-bonded reaction (denoted by : in the following Eqs.).
2(=M-OH) + AO)2, + 2H20 -> [(=M-OHh:(OHhAOj + 20H' (1)

Uptakeof halide ionsmay be described via the following two schemes.

At lowerpH =M-OH + HX -> =M-OH:HX (2)
At higherpH =M-O' + X' + H20 -> =M-OH:HX + OH' (3)

Uptakemechanism via formation of a chemical bond,=M-X, has also been claimedto be

involved. These mechanisms could be corroborated using MAS-NMR and ESCA
spectroscopies of X', as the binding energies of "X" in chemical species, =M·OH:HX and
=M-X will be different. At higher pH, simultaneous anion and cation uptake will occur,
and hence the extent of pH changewill depend upon the basicity of the M ions. The active
sites on LDHs are fully "dissociated" and hencethese compound show usual anion exchange,
as described by Eq. (4).
nOH+A n- ~An- + nOW (4)

The over bar denotes ions in the solid phase. This reaction can be described in same manner
cation exchange, which has been extensively studied.

2.2. Method of Anion Uptake Evaluation

Several methods for evaluating ion uptake have been used. Amongthese, the Langmuir
plot and the corrected selectivity coefficient (CSC) plots (Kielland's plot) are the most
commonly used methods. Other methods such as the power-exchange function and Gapon
equation have also been used, especially, in geochemistry, mineralogy, and soil-chemistry
applications. The following sections present a brief review of CSC plots and other useful

2.2.1. escApproach
The thermodynamic constant(K) can be utilized to describean ion-exchange system. The
K is expressed by the activity of each constituent a; as shown in Eq. 5 for anion exchange
reaction presented in Eq. 4:

K = ~AaOH n (5)
aOHna A

This quantity is the integrated quantity over the whole extent of exchange reaction. Its
existence meanspossible formation of An. form of the solids. The differential function to be
integrated is called the corrected selectivity coefficient as described below.
It has been well established that a plot of the corrected selectivity coefficient (CSC) as a
function of the chargeconcentration in an ion exchangerphase are the reduced expression of
thermodynamic generality.lo.'2 This is independent of the total charge concentration in
solution phase (as derived later) and can be used to calculate distribution coefficients (K.!) as
functionsof several parameters, e.g., the concentration of exchanging ion in solution phaseand
ratio of solid massto solution volume.":"
In the case of the anion exchange, the thermodynamic constant K of the OH·-An.
ion-exchange reaction (Eq, 4) can be written as:

K =XA[OWlnygHfA
where fA and fOHdenote the activity coefficient of the ion in solid phase; [An.] and [OH'l: the
molalconcentration Qfthe ion!!! liquidphaseandYOH andYA the activity coefficients of the ions
in solution phase. XA and XOH are the chargefraction of An and OH ions in solid phase,
defined by the ratio of content of ion-exchange to the ion-exchange capacity. Generally, the
ion-exchange experiments are carried out at a constant total ion concentration in the solution
phase. Although the ionic strength varies during the exchange reaction, the total charge
concentration is kept constant in solution phase as well as in solid phase. Hence,


Here, the totalchargeconcentration, c, of the exchanging ionsin solution isgivenby meq/cm'.

The total charge concentration, C" of the exchanging ions in the solid phase (ion exchange
capacity) is expressed in meq/g. When theconcentration of theelectrolyte solution is low, the
molar concentration can be used instead of the molal concentration. It will be necessary to
take the solutiondensity into consideration for concentrated electrolyte solution.
The relationship presented in Eq. (7) is often neglected for constructing an ion-exchange
isotherm, especially, in chemical engineering.
A different form of the CSCplotcan be derived, uniquely, forthe selectivityevaluation, as
it does not depend on the chargeconcentration in the solution.
The thermodynamic constant is rewritten in Eq. 8 usingthe activity coefficients of ions in
the solid phaseand the corrected selectivity coefficient ( K~H ).

OH fn
.1... (8)

When the total charge concentration in each phase is kept constant, K~H can be written as

A - XAXgH r
K OH (9)
- -n T]
T]= nc~-I and r = 'YOH (10)

The following equation (Eq. II) is also derived from Eq. (5).

logKOH = log = + n(pH-14) -loga A (II)

Assuming the application of Gibbs-Duhem equation to equilibrium system, the following

equation (Eq, 12) is derived by combining Eq. 9 and the thermodynamic equilibrium constant
(K)15.16 (CSC is the "differential form" of the thermodynamic equilibrium constant K which
is the "integral form.")

I _
InK=(1-n)+ J(1nK~H)dXA (12)

Eq. 12doesnot include the activitycoefficient of exchanging anions in thesolidphase. As the

thermodynamic equilibrium constant is notdependent on the totalcharge concentration ofions,
the CSC plot is also independent of the total charge concentration in the solution phase.
Hence, Eq. (II) can be usedwhenthe ionicstrength or charge concentration changes by anion

In case of inorganic solids (ion exchangers) often the maximum value of XA greatly
differs depending on the kind of exchanging anions. However, even though distribution
experimentsin a binaryexchangesystemare carriedout in differenttotal charge concentrations,
experimental values fall on the same CSC plot.
The CSC plot can also be expressed by Eq. 13 using common logarithms. Cm may be
referred to as a generalized Kielland's coefflclent.'" 17


Initial concentration of ion A, [An,]o (mol/drrr'), the volume of solution, V (crrr'), and the total
charge concentration, c, (meq/crrr'), can be used for the calculationof equilibrium composition
of the solution phase when solid mass w (g) in the OH' form was used. Equations(9) and (13)
are combinedto obtain the followingrelationship (Eq. 14).


The right hand side is determined by distribution experiment. Hence, values of XA and
!Qg K~H values from the CSC plot are substituted in Eq. (14) to calculate XA• A set of (XA,
XA ) obtained can be used for calculating the distribution coefficient (Kd). The K" value is
defined by the ratio of the equilibrium ion concentration in solid and solution phases,
respectively. Eq. 15 can therefore be written using the definition of the charge fractions at a
total charge concentration, ct.
K = [A - ] = C,XA (15)
d [An - ] c,X A
K, valuesat different C, and the same XA values can be calculated, becausethis equation is valid
for any concentration. Hence, ~ values at any total charge concentrations in solution phase
are calculatedby changing a c, value. The initial concentration of exchange anions, [An]; at
the experimental point that gave XA and 10gK~H values used for this calculation is given by
the material balance.


This process is very simple and convenient, and does not require a lot of time-consuming
experimentsto obtain parametersof the biphasic distribution. The constraintsof Eq. (17) are
often neglected in the chemical engineering.
Parametersrequired for the separationof chemicalspeciescan be evaluatedas functions of
the total charge concentration and ratio of solid mass to solution volume. Using the CSC plot,
the K" values at the infinitesimal exchange that are too large to determine directly can be
estimated using the followingequation derived from Eqs. (9), (14) and (15).
298 M. TSUJI

n n
(KA) _ - nC t • .1lli. .(K ) (18)
OH X A ' X A -+0 - Ct 'Y A d X A ' X A--+ 0

The relation of K.t value of An- ions at diluted solution and concentration of OH- can be
determined by takingthe logarithm of both sides.

10g(Kd)XA · -xA~.n = 10g~

nf (K~H )X A·-xA~. n -n logc,

The c, value equals the concentration of OH- at zero concentration of Ano. The slope of the
log-log plot of K, vs. c, will be -no- This relationship is also used as one of the methods to
corroborate the stoichiometry ofOH-IAn- exchange. The first term of the right hand side has
beenoftenassumed as a constant. But,the I' valuewillbe not the samein the rangeof thec, or
[OH"] used in experiments, and therefore the first term is not a constant. As a result,the slope
of the plot often does not equalto -n o
The integration of the CSC plot over the whole range of XA = 0-1 gives the
thermodynamic constantfor the specified ion-exchange reaction.

In K = (I - n) + 2.303{I Cm + 10g(K~H )XA.XA-+O I (20)


6G o = - RTInK = -RT [(I - n) + 2.303{ IC m + 10g(K~H)xA .xA-+O}] (21 )


van't Hoff's equation can be used to evaluate the standard enthalpy change6H o and then the
standardentropy change6S o. These are closelyrelatedwith the bonding strength of anions.

0(6GO) =-6H"
6G o = 6Ho - T6S o (23)

When 6G o is a positive value, pure An- form solid will not be formed within the specified
experimental conditions. It shouldbe notedthat this integral valueshouldbe usedonly for the
prediction of formation of a specified ionic form and not for the evaluation of selectivity at a
charge fraction in the solid phase. For the evaluation of selectivity, the differential form
presented in Eqs. (9) or (II) shouldbe used.

2.2.2. Langmuir andOther Equations

The Langmuir equation IS and the power-exchange function, K.x, have been applied in the
tabulation of the cation-exchange data obtained from certain aluminosilicates. The
power-exchange function was used by Walton as well as by Garrels et al." The
cation-exchange data from aluminosilicates were collected mostly in low electrolyte
concentrations of 10-6 to 10-4 M.

Some of the drawbacks associated with these equationsare: I) K.,. value is not a constant
although it is regardedas an exchangeconstant; 2) the n value is generally not an integer; 3) the
stoichiometry of the ion-exchange reactions is neglected; and 4) the power-exchange function
cannot be used to predict ion-exchange behavior at different concentrations. In addition,
hydrolysis of aluminosilicates may have interfered with the uptake of ions at low electrolyte
concentration. In such cases, the ternary exchange system has to be treated and would
therefore make the analysis complicated", as Kex value for exchange of trivalent ions is not
available." Both these equations are the derivative or abbreviated form of the CSC
equations. 10. 21

2.2.3. Activity Coefficient

It is important to correctlydetermine activity coefficients of ions involved in ion exchange

reactionsto be use for the evaluation of the CSC values. For mixedelectrolytesolutions, the
averagesalt method is often used.22-24


Here, the activity coefficient of mixed electrolyte solution should be used instead of the
activity coefficient of a single electrolyte." Several equations to evaluate the activity
coefficients are reported. The equation used by Kraus et al. includes the ionic strengthalone
and has neglected the nature of each electrolyte." Hence, a largedeviation cannotbe avoided
at concentrated electrolyte solutions. Appropriate activity coefficients of actinide elements
are not available," This requires numerous equilibrium experiments to obtain parameters for
the diphasicdistribution of actinides.
The mixed electrolytesolution method described above wouldthereforebe more efficient
for the determination of activity coefficients.


3.1. Crystal Structure and Ion Selectivity

Selectivity control for a targeted anion requires appropriate selection of molecular

structure and the space in the solids to accommodate anions. There is no straightforward
relationship between the exchanging anion selectivity and crystal structure of solids. In
other words, the ion-exchange selectivity has not been deduced on the basis of crystal
structureand other basic properties of solids.
However, from several cation selectivity studies on inorganic solids, geometry and
dimension as well as the electronic configuration of exchanging ions have been found to be
crucial to ion-selectivity. This can be exemplified by a cryptomelane-type manganic acid
(CMA) with the 2x2 type tunnel structure with a tunnel radius of ca. 0.14 nm. The
exchangeable proton in this compound is located in the tunnel. This compound shows an
extremely large K, value (ca. 1014 cml/g in O.IM HNOJ ) for K+and Rb+cations both having
nearly the same crystal ionic radius (0.138 and 0.152 nm, respectively) as the tunnel radius.
However, manganese oxide with the same crystal structure" showed low K, values(I Olcml/g
in 0.1 M HNOl ) for these cations.

This large difference in K.. values can be theorized due to some differences between the
exchange sites of the two ion exchangers at the molecular level, which cannot be inferred
using the crystal structure alone. CMA also showed unusually high K.. value for Pb2+
(_1020 cm'zg in 0.1M HN03) having the crystal ionic radius of 0.119 nm, while small K..
values (_102 cm3/g in O.IM HN03 ) for Ca2+ and Sr+ having nearly the same crystal ionic
radius as Pb2+ (0.100nm and 0.118nm, respectively)." This suggests that electron
configuration of cations plays a major role in their ion exchange selectivity on inorganic ion
exchangers. The binding energies of manganese and lead determined by ESCA
spectroscopy of the cation-exchanged CMAs indicated that these cations are exchanged with
protons in the tunnel. FT1R spectra of cation-exchanged CMAs revealed shifts in
absorption bands in 74G-400cm') range, the extent of shift depended upon the kind of cations
exchanged. 27• 28 This type of lattice flexibility has not been reported for other manganese
Though anion selectivity has not been so actively studied as cation selectivity, geometry
and dimension of exchange sites and space of solids to accommodate them should be
consideredas importantfactors. Therefore, size-fit strategy should reasonably be the primary
consideration to discover inorganic solids with high selectivity for target anions. This size-fit
strategy may not work on organic resins which have very flexible structure, but may be valid
for the inorganic solids due to their much more rigid structure than organicresins.
Hydrotalcites, which are found attractive for several industrial applications, have been
extensively studied for their anion-exchange properties, e.g., as adsorbents for CO2 and other
acidic gases, as host materials for catalysts and for intercalation of macromolecules including
blomolecules.""? However, the lack of atomic-level information relevant to their
anion-exchange selectivities, as well as poor XRD patterns resulting from small crystallite
sizes and structural disorder hinders a fundamental understanding of the relationship between
their ion-exchange behavior and chemical, structural, and dynamical properties of interlayer
and surface species.
Solid-state NMR spectroscopy of inorganic materials has been proved to be a powerful
tool to probe the atomic-level information of chemical species" as well as to study the local
structure and dynamics of surface and interlayer anions in inorganic ion exchangers such as
hydrotalcites and their analogues. Using static IH and 13C NMR technique for Mg/AI
hydrotalcites with 25% AI3+.substitution, Pol et al.32 found that the molecularsymmetry axes
of both carbonates (C3) and water (C2) between the interlayers are oriented parallel to the
crystal axis.
During FTIR spectroscopy of a hydrotalcite family of compounds, the present author
found that these compounds could be characterized by IR absorption band typically observed
at 1370em" . The V3 absorption band for Mg/AI hydrotalcite with 45% AI3+-substitution (high
level AI3+.substitution) split into two bands observed at 1397 and 1364 ern". In addition to
the split, another absorption band due to the c=o bond at 1570 em" also appeared. These
results indicated that carbonatecould exist in differentchemical environments in the low and
high level AI3+-substituted hydotalcites. Hence,the anion selectivity of hydrotalcites could be
different depending upon the AI3+-substitution as well as upon the nature of metal in the
brucite layer.

3.2. Design of SeOJ2· Sieves

Excellent anion sieves can be characterized by a very large (K~H)X .XA"'O and small
ICml values in a wide range of charge fraction in solid phase. In this case, the CSC value will
give a very large K, value irrespective of the extent of anion uptake. Suitable geometry and
dimension of ion-exchange sites present in an inorganic solid can be decided on the basis of
heir crystal structure.

Other important factors to be considered will be electronic configurations of atoms

surrounding or forming the ion-exchange/adsorption sites. The following subsections
present a brief review of potential anions exchangers, which could be used as SeOl2.sieves .

3.2./ . Variable-ChargeType Compounds

This type of anion exchangers includes oxides and hydroxides of divalent, trivalent and
tetravalent metals, and their mixed hydroxides. Some of these solids have been reported as
selective adsorbents of oxyanions. For example, amorphous titanic acid shows very high
select ivity for the anions of As(lII), Te(IV), and Se(IV) as well as for F' ions. 33, )5 Gothite
and aluminum oxides were extensively studied for oxyanion separations,36-31 and for the
coordination chemistry ofSeO/ ' and SeOlby EXAFS spectroscopy.P'"
Although hydro xides and mixed hydroxides of divalent and trivalent metals behave as
anion exchangers (adsorbents), the information is not enough to understand and predict their
relative selecti vities for different anions .":" A clear relationship has not been found
between the atomic arrangement of iron oxy hydroxides, FeO(OH)s, and their anion
selecti vities.P" More work using effective probes may be needed to understand the
mechan ism of ion exchange behavior of these compounds and to design and predict their
anion selectivity.

3.2.2. Substitution Type Compounds

This type of anion-e xchange compound includes layered double hydroxides (LDHs) and
PS+-substituted silicates. M/AI LDHs are represented by [MI,xAlx(OHhy+[(An')xJonH20j'",
where MI Al denotes All +-substituted hydrotalcites in the M(OH)2 brucite layers. Anions
present in the interlayers of LDHs can be exchanged for other anions; this makes them
suitable anion exchangers, Their ion exchange properties can be modified by substituting
metal cations in the brucite layers and the anions in the interlayers. Thennodynamic
treatment of anion exchange reactions may be helpful in the understanding of their exchange
behavior, but more work is needed in this area."
Mg 2+, Cu2+, C0 2+, Zn 2\ and some other cations can be substituted for metal , M, and AI)+
can be replaced by Fe)+, c-", and Ga)+. The interlayer distance ofLDHs may vary with the
type and amount of subst ituting anions. The typical interlayer