Laboratory
`Techniques in
`Electroa na lyt ica I
`Chemistry
`
`Second Edition, Revised and Expanded
`
`edited by
`
`Peter T, Kissinger
`Purdue University and
`Bioanalyticai Systems, Inc.
`West Lafayette, Indiana
`
`William R, Heineman
`University of Cincinnati
`Cincinnati, Ohio
`
`Marcel Dekker, Inc.
`
`New York* Basel ¯ Hong Kong
`
`Lupin Ex, 1054 (Page 1 of 19)
`
`

`

`Library of Congress Cataloging-in-Publication Data
`
`Laboratory techniques in electroanalytical chemistry / edited by Peter
`T. Kissinger, William R. Heineman. -- 2nd ed., rev. and expanded.
`p. cm.
`Includes bibliographical references and index.
`ISBN 0-8247-9445-1 (hardcover : alk. paper)
`1. Electrochemical analysis--Laboratory manuals. I. Kissinger,
`Peter T. II. I-Ieineman, William R.
`QD115.L23 1996
`543’.087 de20
`
`95-46373
`CIP
`
`The publisher offers discounts on this book when ordered in bulk quantities.
`For more information, write to Special Sales/Professional Marketing at the
`address below.
`
`This book is printed on acid-free paper.
`
`Copyright © 1996 by MARCEL DEKKER, INC. All Rights Reserved.
`
`Neither this book nor any part may be reproduced or transmitted in any form
`or by any means, electronic or mechanical, including photocopying, micro-
`filming, and recording, or by any information storage and retrieval system,
`without permission in writing from the publisher.
`
`MARCEL DEKKER, INC.
`270 Madison Avenue, .New York, New York 10016
`
`Current printing (last digit)’
`10 9 8 7 6 5 4 3 ’2 1
`
`PRINTED IN THE UNITED STATES OF AMERICA
`
`Lupin Ex. 1054 (Page 2 of 19)
`
`

`

`This materiai may be protected by Copyright faw (Titie 17 U.S. Code)
`
`15
`Solvents and Supporting Electrolytes
`
`Albert J. Fr~ Wesleyan University, Middletown, Connecticut
`
`!.
`
`INTRODUCTION
`
`Almost every electrochemical experiment is carried out in a medium consist-
`ing of a solvent and supporting electrolyte. Many such combinations have been
`used by electrochemical experimenters; as a glance at the Appendix will show,
`selection of a solvent and electrolyte has been the subject of much discussion
`in the literature. Yet the problem, is not intrinsically complicated; a handful of
`solvent-electrolyte combinations would probably suffice for the majority of
`electrochemical applications. For this reason, a few particularly good solvent
`systems (the term "solvent system" is used here to describe the medium con-
`sisting of both solvent and supporting electrolyte) are recommended here, in lieu
`of an extended but uncritical listing of all such systems. The systems to be
`recommended have been chosen from among many that have been reported.
`However, because the novice, will certainly want to develop a feel for what
`makes a "good" solvent system, the first part of the chapter discusses desirable
`criteria, as well as the ways in which electrochemical results may depend on
`the nature of the particular solvent-electrolyte combination employed. Because
`special situations frequently.do require special solutions, Section III describes
`some less common solvents and electrolytes that have been found useful in
`special situations. Finally, the Appendix offers a guide to further readings of a
`more comprehensive nature.
`
`A. Solvent Selection Criteria
`
`There is no universal solvent, and even for a given application one rarely finds
`an ideal system. One must factor some informed guesswork into one’s choice
`of solvent and electrolyte. In order to optimize conditions for an electrode re-
`action, one must consider how its chemical and electrochemical features, for
`
`469
`
`Lupin Ex. 1054 (Page 3 of 19)
`
`

`

`470
`
`Fry
`
`example, the chemical properties of expected intermediates, might be affected
`by the solvent or electrolyte. A solvent is chosen such that its merits outweigh
`its disadvantages in a particular situation. A good solvent system for one type
`of experiment may be wholly unsuitable for other applications. The most im-
`portant properties that the ideal solvent system ought to possess are electrochemi-
`cal inertness (stability over a wide range of potentials), high electrical conduc-
`tivity, good soIvent power, chemical inertness, availability in pure form or ease
`of purification, and low cost, although others must sometimes be considered.
`We now examine each of these in detail.
`
`Electrochemical Inertness
`The solvent system should not undergo any electrochemical reaction over a wide
`range of potentials from very positive (strongly oxidizing) to very negative
`(strongly reducing). The potential at which electrochemical reaction of the sol-
`vent system (either the supporting electrolyte or the solvent itself) commences
`is known variously as the solvent breakdown potential, solvei~t decomposition
`potential, or solvent background limit. Certain systems can be used over a very
`wide potential range; one of the best for this purpose is a solution of tetrabutyl-
`ammonium hexafluorophosphate in acetonitrile, which exhibits anodic and ca-
`thodic breakdown limits of +3.4 and -2.9 V (vs. SCE), respectively [1]. In
`practice, however, one rarely needs both extended positive and negative ranges
`in a single set of experiments; solvents are frequently used that exhibit either a
`very negative or very positive decomposition potential, but not both. Thus, both
`nitrobenzene [2] and methylene chloride [3] have been used for studies of oxi-
`dation processes because they are hard to oxidize, notwithstanding the fact that
`they are quite easily reduced. Conversely, liquid ammonia and methylamine are
`good solvents for electrochemical reductions, although poor for oxidations [4].
`It is reported that potentials as positive as +4.5 V (vs. SCE) can be reached in
`liquid SO2 [5].
`The potential limit is frequently set by the electrochemical behavior of the
`supporting electrolyte, not the solvent. For example, dipolar aprotic solvents
`such as acetonitrile (AN), dimethylformamide (DMF), and dimethyl sulfoxide
`(DMSO) are very difficult to reduce; the reduction of alkali metal ions to the
`corresponding metals (ca. -2 V vs. SCE) is the potential-limiting process in such
`soIvents, and by use of the more difficult-to-reduce tetraalkylammonium salts
`one can reach considerably more negative potentials. The decomposition poten-
`tials of tetraalkylammoniumions at mercury themselves become more negative
`as the alkyl groups become larger (Me4N+: -2.65 V; Bu4N+: -2.88 V) [6]. One
`can frequently extend the accessible cathodic range by using a large tetraalkyl-
`ammonium ion, but double-layer and adsorption effects can negate this gener-
`alization [6,7]. DMF and DMSO are not suitable as solvents for anodic stud-
`ies because they are fairly easily oxidized, but as noted earlier, AN is very
`difficult to oxidize, and in this solvent the decomposition potential of the sup-
`
`Lupin Ex. 1054 (Page 4 of 19)
`
`

`

`Solvents and Supporting Electrolytes
`
`471
`
`¯ porting electrolyte is usually the limiting process. At very positive potentials,
`oxidation of the anion of the supporting electrolyte is often the limiting process.
`Not only should easily oxidized ions such as bromide and iodide not be used
`for anodic studies, but even nitrate and perchlorate salts should be avoided when
`one wishe~ to operate at very positive potentials; at ihese potentials tetrafluoro-
`borate and hexafluoroborate salts are preferred because of their stability to
`oxidizing conditions [1].
`¯ Electrochemical reaction of the solvent or supporting electrolyte need not
`be undesirable. Sometimes one can take advantage of such behavior by using
`the product of such a reaction as a chemical reagent. For example, one can
`generate bromine by anodic oxidation of bromide ion or lithium by cathodic
`reduction of lithium ion. There can be real advantages to electrochemical gen-
`eration of reagents this way instead of using the bulk reagent, for example,
`avoiding the need to handle dangerous substances and being able to add the
`reagent in regulated amounts by careful control of the electrolysis current [8].
`
`Electrical Conductivity
`In order to support passage of an electric current, the solvent system should have
`low electrical resistance. This implies that the solvent should have a moderately
`high dielectric constant (_>10); the prevalence of ion pairing and even multiple
`ionic association in less polar ffolvents results in relatively low ionic mobility
`and conductance. The fact that nonpolar solvents tend not to be very good sol-
`vents for salts in the first place makes it even harder to contemplate using theml
`A number of the most common solvents for organic electrochemistry have sat-
`isfactory dielectric constants (DMF: 36.7; AN: 37.5; DMSO: 46.7), and their
`electrolytic solutions have acceptably high conductances [9-11]. Solvents of
`substantially lower polarity than these, however, have been used successfully
`despite their high electrical resistance (e.g., tetrahydrofuran, THF [12], and
`methylene chloride [3], whose dielectric constants are 7.3 and 8.9, respectively
`[13]). Lower conductivity can be tolerated more readily in low-current appli-
`cations such as voltammetry than in preparative electrolyses. Although volt-
`ammograms can be badly distorted by uncompensated resistance under condi-
`tions of high electrical resistance, the distortion can be largely corrected by using
`high electrolyte concentrations (not always possible in resistive solvents, which
`tend to be poor solvents for salts) and by computational correction of the volt-
`ammogram [11]. A new element has entered the scene in recent years with the
`development of small-diameter (diameter _<10 I.tm) microelectrodes [14]. These
`can be used in highly resistive media [11,15-19], including the absence of
`supporting¯ electrolyte [20], and in fact even in the gas phase [21].
`
`Good Solvent Power.
`An electrochemical solvent must be able to dissolve a wide range of substances
`at acceptable concentrations. In general, this means that electrolytes must be
`
`Lupin Ex. 1054 (Page 5 of 19)
`
`

`

`1
`
`soluble at least to the extent of 0.1 M, while the electroactive substance must
`be soluble to the extent of 0.1-1 mM for application of the various electroana-
`lytical techniques, and higher concentration for preparative electrolyses. Salts
`in which either the cation or anion, and preferably both, are large, such as
`tetraalkylammonium perchlorates, hexafluorophosphates, and tetrafluoroborates,
`generally exhibit the greatest solubility in organic solvents. The latter two are
`becoming more popular than perchlorates because of safety hazards with per-
`chlorates, especially in acidic media. Ethyltributylammonium tetrafluoroborate
`is quite soluble in nonpolar solvents [12], as is tetrabutylammonium hexafluoro-
`phosphate.
`
`Chemical Inertness
`The solvent should not react with the electroactive material nor with interme-
`diates or products of the electrode reaction under investigation. No one solvent
`can reasonably be expected to be unreactive toward all of the many kinds of
`reactive species that can be generated electrochemically. When selecting a sol-
`vent, one does have to make some informed guesses about likely intermediates
`and products, based on chemicaI experience. For example, many organic and
`anodicprocesses generate cationic, that is, electrophilic species [22]; one would
`therefore not use a nucleophilic solvent if one wants such intermediates to be
`long-lived. This is equally true for organic and inorganic species. Acetonitrile
`is frequently used in anodic studies because of its high conductivity and wide
`accessible potential range, but it is moderately nucleophilic; it can react with
`organic intermediates and replace ligands on inorganic intermediates. Attempts
`to characterize anodically gefierated intermediates are therefore frequently car-
`ried out in a less nucleophilic solvent such as methylene chloride or one of the
`"superacid" solvents to be discussed in Section II.A. Conversely, reduction of
`organic substrates frequently forms anionic intermediates, and one must there-
`fore avoid the use of electrophilic solvents, in particular protic substances, to
`avoid acid-base reactions between such intermediates and the solvent. Finally,
`one would like the solvent to be reasonably stable, so that purification, prepa-
`ration, and storage of standard solutions will present no major problems.
`
`Liquid Range
`The solvent should exhibit a convenient liquid range; that is, it should be nei-
`ther too low-boiling nor too high-boiling. The former makes for potential safety
`and fire hazards and.difficulty in storing, solutions, particularly when degassing
`them, and the latter makes purification by distillation more difficult. On the other
`hand, the definition of convenient liquid range depends on the particular experi-
`ment involved, in particular the temperature at which it is to be carried out.
`High-boiling solvents are preferred, for experiments at higher temperatures,
`whereas low-boiling solvents are better for low-temperature electrochemistry
`
`Lupin Ex. 1054 (Page 6 of 19)
`
`

`

`Solvents and Supporting Electrolytes
`
`473
`
`(they retain favorable viscosity properties at low temperatures). Other factors
`to be considered are (a) whether (and how) the solvent must be removed and
`solutes recovered at the end of the experiment, and (b) the particular purifica-
`tion and transfer procedures to be used in the experiment..
`
`Miscellaneous Considerations
`Solvents and electrolytes should also be inexpensive, nontoxic, and nonflam-
`mable. The latter two characteristics are not well satisfied by most organic
`solvents, but with reasonable safety precautions and reasonable ventilation they
`can be used routinely without incident. Another solvent property, viscosity, may
`be of importance on occasion. High viscosities are useful when one wishes to
`extend the time interval over which mass transport occurs purely by diffusion,
`such as for potential-step experiments, but a low-viscosity solvent is preferred
`when efficient mass transport is required, as in preparative electrolyses.
`
`!1. RECOMMENDED SOLVENTS AND ELECTROLYTES
`
`A. Anodic Electrochemistry
`
`Solutions of inorganic salts in water meet many of the preceding desirable sol-
`vent criteria, and have been used for innumerable electrochemical studies. Much
`inorganic electrochemistry has been carried out in water with little difficulty.
`Organic and organometallic compounds either are not very soluble in water or
`are reactive toward it; this has stimulated interest in the use of nonaqueous
`solvents in electrochemistry. Ethanol and methanol are good organic solvents,
`inexpensive, readily available in high purity, dissolve salts readily, and in gen-
`eral have much to commend them for electrochemical purposes. On the other
`hand, alcohols are relatively easily oxidized and are highly nucleophilic. For
`this reason, aprotic organic solvents are more commonly used, and of these, the
`most commonly used are acetonitrile and N,N-dimethylformamide. Acetonitrile
`(AN) is probably the most widely used solvent for organic anodic electrochem-
`istry, and we believe that it is the preferred solvent for most such purposes. It
`is a good solvent for organic compounds and many inorganic and organic salts,
`has a high dielectric constant and a convenient liquid range (mp -45.7°C, bp
`81.6°C), and is transparent in the ultraviolet region. (For purification of AN,
`see Sec. VI.) Recommended electrolytes include tetrabutylammonium hexa-
`fluorophosphate (TBAHFP) and tetrafluoroborate (TBATFB), both of which are
`readily obtained in pure form (See. VI), have high solubilities in AN, and are
`stable to very positive potentials. These salts, particularly the hexafluoro-
`phosphate, are rather insoluble in water and can often be separated from reac-
`tion products by evaporation of the AN followed by extraction with a nonpolar
`solvent, chromatography, or volatilization of the desired product. Sometimes
`
`Lupin Ex. 1054 (Page 7 of 19)
`
`

`

`474
`
`Fry
`
`separation of the electrolyte from an organic product presents difficulties, how-
`ever. An inorganic electrolyte (e.g., NaBF,~, LiC104, or NaC104) may be nec-
`essary in such cases.
`Acetonitrile is rather nueleophilic; this is a disadvantage when the electro-
`chemical reaction produces a reactive electrophilic intermediate, or when the
`substrate is an inorganic or organometallic species containing a replaceable
`ligand. In this case, solutions of TBATFB or TBAHFP in methylene chloride
`have frequently been used as alternatives, although the latter solutions are re-
`ported to decompose slowly [23]. Meyer, however, has recently found that 1,2-
`difluorobenzene is a better solvent than methylene chloride [24]. Other choices,
`although chemically more reactive than the preceding halogenated solvents,
`include trifluoroacetic acid (no electrolyte needed) [25], liquid sulfur dioxide
`[26], an aluminum chloride-alkali halide melt [27], or trifluoromethanesulfonic
`acid [28]. Anodically generated reactive electrophilic inter.mediates are also rela-
`tively long-lived at low temperatures [29] or in very dry AN [30,31].
`
`B. Cathodic Electrochemistry
`
`Water and alcohols are often useful in inorganic electrochemistry. Dimethyl-
`formamide (DMF), acetonitrile, and dimethyl sulfoxide (DMSO) are all excel-
`lent solvents for reductive electrochemistry of organic compounds. TBAHFP and
`TBATFB are good electrolytes for use in these solvents. Despite their rather high
`boiling points (156°C and 189°C, respectively), DMF and DMSO are particu-
`larly good solvents for cathodic electrochemistry. DMF is fairly easily purified
`(See. VI.B) and DMSO is in fact fairly pure as,. received. Both are excellent
`solvents for both organic and inorganic solvents. Water (which is always present
`in polar aprotic solvents) is a poor proton donor in DMSO and DMF and can
`be reduced by appropriate procedures (Sec. VI.A). Probably the most widely
`used organic solvents for cathodic electrochemistry are DMF and acetonitrile.
`DMF has a number of desirable properties, but when wet is prone to slow de-
`composition, liberating dimethylamine. Decomposition is much less of a prob-
`lem in dry DMF, which can be prepared by vacuum distillation from calcium
`hydride, followed by treatment with 3-/~ molecular sieve [32]. Acetonitrile would
`be our primary recommended solvent for cathodic electrochemistry except for
`problems related to its purification. Very pure, dry AN can be obtained by
`proper treatment, but this can be tedious [33]. The efficacy of a number of
`simpler procedures in the literature for. drying AN has been disPUted [34].
`
`I!!. SOME OTHER SOLVENTS
`
`Many solvents and electrolytes have been used in electrochemical applications.
`In the preceding sections we suggested a few solvents for general use. It may
`
`Lupin Ex. 1054 (Page 8 of 19)
`
`

`

`Solvents and Supporting Electr.olytes
`
`475
`
`. happen, however, that none of these is acceptable for a particular application.
`There are many exhaustive discussions of solvent systems available in the lit-
`erature, which we shall not duplicate here (see the Appendix). A few solvents
`deserve specia! mention here, because each has a property likely to be useful
`in certain special situations.
`
`A. Low-Temperature Electrochemistry
`
`Low-temperature applications (Chap. 16) present special problems. With most
`solvents, solvating power decreases (this is a particular problem with support-
`ing electrolytes, which have to be soluble in concentrations of 0.1 M or higher
`for conventional electrochemistry) and viscosity increases as the temperature is
`lowered. Both effects result in higher solution resistance at low temperatures.
`Reilley and Van Duyne [29] and, more recently, Evans [11] and Murray [15]
`and their coworkers have examined a variety of solvent combinations and have
`found several that suffer the least degradation of solvent properties with decreas-
`ing temperature. Reilley and Van Duyne recommended butyronitrile for use with
`cortventional electrodes at temperatures as low as -105°C. The development of
`microelectrodes in recent years has been a boon to those interested in low-tem-
`perature electrochemistry, since these electrodes can function well in highly
`resistive media (Chap. 12). Evans recommended acetonitrile for operations down
`to -40°C and butyronitrile to -80°C for high-speed low-temperature voit-
`ammetry at microelectrodes. Murray found a number of solvent mixtures that
`are suitable for voltammetry at ultra-low temperatures: 1:1 butyronitrile
`(BN):ethyl bromide (with 0.2 M TBAP as electrolyte) is usable to -148°C;
`1:1:1:1 isopentane: methylcyclopentane:BN:ethyl bromide to-158°C; and re-
`markably, 1:1 BN:ethyl chloride, in which voltammetric data could be obtained
`at temperatures as low as -185°C! The use of microelectrodes also permits
`voltammetric measurements in frozen solvent [16] and in solids [17]. Gosser has
`carried out voltammetry in frozen DMSO with conventional electrodes, but has
`suggested that the electrochemistry actually occurs in pools of liquid solvent
`within the frozen medium [18].
`
`B. Solvated Electrons
`
`Solvated electrons are readily prepared in hexamethylphosphoramide (HMPA),
`ammonia, and methylamine [19]. Lithium chloride is the preferred electrolyte
`for this application.
`
`C. Nonpolar Media
`
`Nonpolar media are of interest as electrochemical solvents because they permit
`the investigator to observe the properties of polar intermediates undistorted by
`
`Lupin Ex. 1054 (Page 9 of 19)
`
`

`

`476
`
`Fry
`
`interaction with the solvent medium, but they present major problems because
`of their high electrical resistance. This is compounded by the poor solubility of
`most electrolytes in such solvents. The high electrical resistance causes severe
`signal distortion in low-current experiments (voltammetry) and resistive heat-
`ing (which can be so severe as to cause solvent boiling) in high-current experi-
`ments (preparative-scale electrolyses). As mentioned in the preceding section,
`the use of microelectrodes can largely mitigate the resistance problem in volt-
`ammetry in such media.
`There are hydrocarbon-like media that can permit voltammetric studies. It
`turns out that molten naphthalene and its 1-chloro and 1-methyl derivatives are
`highly electrically conductive and good solvents for organic salts at elevated
`temperatures (1.0 M TBAP solutions can be prepared) [35]. Well-formed ac and
`cyclic voltammograms can be obtained at 150°C. These media were found useful
`for substrates such as phthalocyanines, which are completely insoluble in all
`common solvents near room temperature. A barely explored idea in nonaqueous
`electrochemistry is that of incorporating the functions of solvent and support-
`ing electrolyte into a single substance, that is, a liquid or low-melting organic
`salt. A solution of tetrabutylammonium tetrafluoroborate (TBATFB) in toluene
`separates into two liquid phases; the upper is pure toluene, and the lower is a
`novel "liquid electrolyte" whose composition corresponds to three molecules of
`toluene and one of TBATFB [36]. Fry and Touster examined the properties of
`this substance as a voltammetric solvent and found them to be remarkably similar
`to those of DMF, AN, or DMSO [37]. Tetrahexylammonium benzoate, a liq-
`uid at room temperature, has been found suitable for electrochemical use [38].
`Better candidates might be found among the many low-melting organic salts now
`available [39].
`
`IV. SOLVENT- AND ELECTROLYTE-DEPENDENT PHENOMENA
`
`Electrochemical reactions are unavoidably influenced by the nature of the sol-
`vent and supporting electrolyte. We have already alluded several times to the
`possibility of chemical reaction of electrochemically generated intermediates with
`the solvent. One must always be on guard against this possibility. The nature
`of the problem is exemplified by a study of the electrochemical reduction of
`iodobenzene in DMF containing tetraethylammonium bromide [40]. A variety
`of experiments demonstrated that the h.ighly basic and nucleophilic phenyl an-
`ion is generated in this reaction. Benzene is the major product; presumably it
`is formed by attack by the phenyl anion on traces of moisture in the solvent.
`When the water content in the solvent was lowered, benzene was still found as
`the major product, but now it was found to be formed by proton attack upon
`the cation of the supporting electrolyte (Hoffmann elimination). When the elec-
`trolyte was changed to lithium bromide to prevent Hoffmann elimination, phe-
`
`Lupin Ex. 1054 (Page 10 of 19)
`
`

`

`Solvents nnd Supporting Elect~rol~tes
`
`4 ~7 .
`
`nyl anion proceeded to react with the solvent by a combination of proton ab-
`straction and nucleophilic addition! This is a reminder that carbanions gener-
`ated in dipolar aprotic solvents such as DMF are especially reactive; they can-
`not bind covalently to tetraalkylammonium counte.rions, .and even ion pairing
`to the coUnterion is minimal in such solvents because of their high dielectric con-
`stants. Electrochemically generated phenyl anions might perhaps be longer lived
`in a less reactive solvent (e.g., THF) with a metal salt as electrolyte. The an-
`odic counterpart of this problem has already been mentioned: the fact that an-
`odically generated cations readily attack nucleophilic solvents such as
`acetonitrile, mandating the use of electrophilic solvents when it is desired to
`extend the lifetime of such cations. All of the special solvent systems mentioned
`at the end of Section ii meet this criterion.
`The solvent and electrolyte can affect an electrode process in a number of
`ways other than by reacting chemically with the electrolysis intermediates or
`products. The rate of heterogeneous electron transfer from the electrode to the
`electroactive substance can be affected by. the structure of the electrical double
`layer at the electrode surface, which in turn is dependent on the nature of the
`solvent and supporting electrolyte [41]. This isn’t the only role the electrolyte
`can play when incorporated into the double layer. Acidic cations, such as tri-
`ethylammonium and guanidinium ions, can protonate short-lived intermediates
`generated at the electrode surface before such intermediates can escape into bulk
`solution [42]. Similarly, preferential solvation of small ions by one constituent
`of a mixture can raise the local concentration of this substituent at the electrode
`surface if the ion is in the double layer (anions at potentials positive of the
`potential of zero charge [PZC], cations at potentials negative of PZC). Examples
`of both cathodic [42] and anodic [43] electrochemistry thought to be affected
`this way have been reported. A third type-of surface electrolyte effect has been
`reported, in which optically active products have been obtained from electrode
`reactions in which an optically .active supporting electrolyte has been used [44].
`The degree of asymmetric induction obtained in this fashion has generally been
`small. Reasons for this have been suggested [45].
`The most eornmon electrochemical effects exerted in bulk solution are
`related to association (solvation, ion-pairing, complex formation, etc.) with the
`electroactive substance or electrochemically generated intermediates [4,19]. The
`importance of solvation can be gauged by comparing calculated and measured
`values of the parameter !XEI~2 (defined as the difference, in volts between the
`half-wave potentials of the first and second polarographic Waves) exhibited by
`polycyclic aromatic hydrocarbons (PAH) in dipolar aprotic solvents [46,47]. It
`can be shown that AEIt2 is related to the equilibrium constant for disproportion-
`ation of the aromatic radical anion into neutral species and dianion, that is,
`
`2Ar:~ArH + ArI-I2-
`
`Lupin Ex. 1054 (Page 11 of 19)
`
`

`

`478 F~
`
`The computed value of AE~/2 by a reliable computational method [48] is -5 V
`[49]. The experimental values are generally only about one-tenth of this (0.4-0.6
`V) in DMF, AN, or DMSO, showing that the disproportionation equilibrium
`is significantly perturbed by solvation [46,47,49]; the computed value is essen-
`tially a gas-phase, solvent-free value. The disproportionation equilibrium depends
`on the nature of the supporting electrolyte, demonstrating that ion pairing is
`operative even in solvents as polar as these, although it appears to involve only
`the dianion, not the radical anion. In low-dielectric-constant solvents, where ion
`pairs and even higher aggregates are prevalent, or with electrolytes (such as
`lithium salts) that form strong covalent bonds [50], association between the elec-
`trolyte and charged intermediates plays an even greater role in affecting mea-
`sured potentials. In fact, in the specific case of radical anion disproportionation
`under discussion here, AEIt2 can even become negative; that is, the two one-
`electron waves in polar media coalesce into a single two-electron wave [49,51].
`
`V. EXPERIMENTAL PROCEDURES
`
`A. Obtaining Dry Solvent
`
`Water is the most ubiquitous impurity in organic solvents. Considering how
`sensitive many reactions are to the presence of water, the literature on drying
`solvents is surprisingly full of contradictions and misinformation. Anyone
`doubting this statement is referred to a remarkable series of papers by Burfield
`[32,34,52,53]. The Introduction sections to these papers will provide a healthy
`dose of reality to those who feel they can rely solely on their chemical intuition
`when deciding how to purify their organic solvents. Burfield developed a highly
`accurate method for determining the water content of organic solvents and then
`used it to assess the dryness of a number of common organic solvents, each of
`which was dried by a variety of methods. (The analytical method is based upon
`treatment of the anhydrous solvent with tritiated water, followed by drying and
`analysis of the solvent for residual tritium by scintillation counting [54].) Burfield
`made a number of useful discoveries in the course of this work, including the
`following: (a) a number of common laboratory drying procedures are almost
`totally ineffective with some solvents; (b) a reagent that is good for drying one
`solvent may be rather poor with another; and (c) probably the best general
`method for drying most solvents is by the use of molecular sieves, assuming
`they are used properly, that is, (1) activating 3-,~ molecular sieves for 15 h at
`300°C, (2) allowing them to cool in a P205 dessicator, (3) treating the solvent
`with activated sieve (5-10% w/v) for 6 h to 7 days, and (4) allowing the sol-
`vent to stand over a second batch of fresh molecular sieve. This procedure was
`shown to reduce water levels in the.following solvents to the indicated levels:
`DMF: <1.5 ppm; AN: _<0.6 ppm; DMSO: _<10 ppm. A series of experiments
`
`Lupin Ex. 1054 (Page 12 of 19)
`
`

`

`Solvents and Supporting Elect.rolytes
`
`479
`
`in which the water content of DMF standing over molecular sieve was moni-
`tored by gas chromatography confirmed the fact that several days are necessary
`to achieve the lowest water levels [55]. An important qualification must be added
`at this point, since it was not considered by Burfield: His analysis assumes that
`the only ~olvent impurity is water. This is frequently not the case. DMF con-
`tains readily detectable quantities of dimethylamine, DMSO contains traces of
`dimethyl sulfide, AN can contain acrylonitrile, THF and other ethers contain
`peroxides, etc. Thus it is frequently necessary to subject the solvent to a pre-
`liminary treatment to remove such impurities and lower its water content be-
`fore drying. (Commercial DMF is frequently >0.1 M in water, for example.)
`Peroxides can usuaIly be removed by passing the solvent through a column of
`activated alumina.
`Dipolar aprotic solvents are generally hygroscopic, and contact with the
`atmosphere should be avoided after drying. For critical applications, the solvent
`should be distilled into the electrochemical cell [56] or in a dry box, or passed
`through a column containing the dry agent repeatedly in a sealed system [57].
`Simpler and almost as satisfactory alternatives include drying the solvent over
`a molecular sieve either in a dropping funnel attached to the cell, followed by
`direct addition of the solvent to the cell [55], or in a separate flask, followed
`by solvent transfer using cannula techniques [58]. Supporting electrolytes are
`frequently hygroscopic and should be dried before use; this is readily done in
`an Abderhalden appar