CA2194127A1 - Delocalized anions for use as electrolytic solutes - Google Patents

Abstract

The delocalized anions are substitution derivatives of perfluoromethanesulfonamides or polyalkoxysulfonamides, cyclic anions deriving from barbituric acids, or heterocyclic anions of general formula (see fig. I) in which Z1, Z2, Z3, Z4 and Z5 represent either a nitrogen atom ~ N, or a radical ~ CY, a group ~ S (= O) Q or ~ PQ'Q "where Q, Q 'and Q" identical or different representing - an organic group alkyl or alkenyl, aryl , alkyl-aryl or alkenyl-aryl, containing or not containing oxa or aza substituents, simple or perhalogenated, in particular perfluorinated of the RF type. Y representing: - a group Q, - a halogen F, Cl, Br, - a pseudo-halogen, that is to say a monovalent electron-withdrawing radical such as CN, SCN, N3, CF3, OCF3, OCnF2n + 1, OC2F4H, SCF3, SCnF2n + 1, SC2F4H ..., - an XCO- or XSO2- group in which X represents either a fluorine atom, a Q group, an alkoxy, vinyloxy or dialkylamino, dialkenyl amino group. Use as a source of ions in electrolytic solutions, as polymerization catalysts or in the catalysis of chemical reactions.

Description

INTRODUCTION

CHAPTER I: ELECTROLYTES

II-1 INTRODUCTION .............................................. . 1 II-2 CONDUCTORS ~ PROTONI ~ UES ............................... "

II-2-1 Introduction ............................................ 2 II 2-2 Phase diagram of the binary im ~ le / acid ............... 10 II-2-3 Conductivity of proton electrolytes ................ 14 II-2-4 Mechanism of proton conductivity in molten salt ... 21 II-2-5 Range of electrochemical stability of molten salt ........ 22 II-2-6 Conclusion on proton electrolytes .............. 23 II-3 ALKALINE CONDUCTORS ................................... 25 II-3-1 Introduction ............................................ . 25 II-3-1-1 Salt-solvent interactions ......................... 25Il-3-1-2 Acidity measurements. ................................. 32 Il-3-1-3 Salt-polymer interactions ........................ 37 II-3-2 Open anions ...................................... 43 II-3-2-1 General presentation .................................. 43 II-3-2-2 Synthesis of anions ................................... 5 7 II-3-2-3 Condrrctivity of open anions ....................... 61 II-3-2-4 Conclusion on open anions ...................... 7 1 II-3-3 Cyclic anions ..................................... 72 II-3-3-1 General presentation .................................. 72 II-3-3-2 General presentation of the synthesis of heterocycles. 81 II-3-3-3 Synthesis of anions .................................... 87 II-3-3-4 Conductivity of cyclic anions ...................... 93 II-3-3-5 Outlook .......................................... . 1 07 II-3-3-6 Conclusion on cyclic salts ...................... 108 II-3-4 Conclusion on alkaline salts ......................... 109 II-4 CONCLUSION ON THE ELECTROLYTES ............................ 110 Bibliography................................................. .. 111 ;
CHAPTER II: Mixed conduction polymers 114 II-l INTRODUCTION .............................................. ...... 114 II-2 CARBON GRAFTED ............................................ 115 II-2-2 Grafting of polymers ......................................... 116 11-2-2-1 Changes in the carbon surface ...................... 116 II-2-2-2 Grafting of polymers ....................................... 119 II 2-3 Grafting of salts ........................................... ... 130 II-2-4 Conclusion ............................................ ........ 131 II-3 REDOX SOLVENT POLYMERS .................................. 132 II-3-1 Introduction ............................................ ...... 132 II-3-2 Cationic redox polymers ............................... 133 II-3-3 Anionic redox polymers ................................ 137 Il-3-2-1 Summary .......................................... .......... 137 II-3-2-2 Electrochemical test ...................................... 141 ll-3-2-3 Molecular stacks ................................ 147 II-3-4 Application as me ~ te ~ r of electrons in batteries .... 153 II-3-5 Application as a power reserve in batteries ..... 155 II-3-6 Application to electrochromic glazing ....................... 167 II-3-7 Conclusions and perspectives ................................... 175 Bibliography................................................. ....... 176 GENERAL CONCLUSION

Annex A
Annex B

Introduction SENSOR ~ LECTROCHROME BATTERY

2 Platinum ~ HylrO2 Lithium - e ~

t ~ t ~ ~ ~ ~ a ~

Ho 5V2 ~ 5 HXWO3 TiS2 , ...............................................
. He ~

Chapter I:
ELECTROLYTES

11-1 lntroductlon:

Electrolytes are at the heart of all electrochemical systems, both generators electrochemical, electrochromic glazing, sensors, and others, in which they allow the transport of ions between the two electrodes where the electrochemical reactions take place.

Over the past fifteen years, the concept of thin film electrochemistry has emerged, mainly in the context of the development of high energy density batteries. The way more promising for the development of these systems is to try to reconcile a more complex than that used in current systems, with the possibility of manufacturing these products using proven technologies for producing thin films.

Such an approach requires in particular to have plastic electrolytes allowing work with relatively thin thicknesses (~ 50 ~ Lm), and able to withstand mechanical constraints inherent in these processes.

In terms of chemistry, we can distinguish two main families of polymer electrolytes:
proton polymer electrolytes, and polymer electrolytes based on alkali metal salt.

Proton electrolytes are of particular interest for the production of sources energetic of puics: ln ~ e, their conductivities being in general rather high, which motivated the part concerning them in this work.

Regarding polymer electrolytes based on alkali metal, lithium conductors are particularly studied for the development of lithium batteries with electrolytes polymers. They are obtained by dissolving a lithium salt in a solvating polymer matrix, such as poly (ethylene oxide). The cost of this salt generally represents a fraction not negligible of the total price of these systems we have sought to explore new avenues of synthesis of these anions. To do this, we analyzed according to literature data the physico-chemical criteria pennett ~ nt to consider obtaining dissociated salts in a polymer medium, and thereafter, explored the different voices of syntheses making it possible to obtain salts corresponding to these criteria.

~ 1 q4 '27 II-2 The proton conductors:

II-2-1 Introduction:

The proton differs markedly from other alkaline ions mainly by its ionic radius 105 times lower (r- 1 fm). Due to the polarizing power which ~ n rusulle, we pe ~ uppojur ~ u'ii ne exist ~ e not in an unsolvated state, but always associated with a donor atom of doublets (O, N, S, ...) in the case of organic proton conductors, or at crict ~ llins sites of the network for conductors minerals. We can distinguish two mechanisms: proton displacement lymmas [Colomban, A ~ and1,2, Kreuerl le vehicle transport and transport according to the me-c ~ nicme of Grotthus (also called translocation mechanism). In the case of vehicle mechanics, the proton is associated with a 7 7 7 basic molecule (H2O, NH3, ..) which diffuses and ~~ \ ~~ \ ~~ \
releases the proton at the reaction site (reactive site, H \ + / ~ "~ H ~ (~" J ~ "~ H
electrode, ..). The mechanism of Grotthus, typical lll I
protic solvents, involves the displacement of HHHH
Figure 2.1: Vehicle transport mechanism proton by the cooperative exchange between protonenmilieuaqueux, according to Armand2].
chemicals and hydrogen bonds ~ More specifically, taking the example, illustrated in figure 2.1, of the proton conduction in water, the process takes place in two stages: the proton transfers from an oxonium ion to a water molecule by effect tunnel along a hydrogen bond, then the water molecule thus reformed is redirected in order to fix a new proton. The real mechanisms of proton conduction are certainly more complex, we can however note that the displacement of the proton following the Grotthus mechanism corresponds to conductivities about ten times higher than those governed by the mechanism vehicular.

Proton conductors are the subject of intense research, both fundamental to understand the particular mechanisms of proton and applied conduction, essentially as a mineral or polymeric electrolyte for the development of fuels, electrochemical capacities, electroehrome glazing, sensors, and batteries [Colombar ~ n ~ f ~ r ~ "il. For the past ten years, various studies mainly conducted at LIES-G in Grenoble, and at the molecular spectroscopy and crict ~ lline laboratory (LSMC) of the University of Bordeaux have highlighted the possibility of making proton conductors anhydrous polymers [Trinquet.Defendini.Charbouillot.Prusse ~ ~ Ds7e ~ Remmudez1, Ravetl At the present time these electrolytes are used in the form of thin films of a few di7 ~ ines of microns for setting at the point of electrochromic glazing [Ravet ~ Ripoche]. Thereafter, no description will be made exh ~ ustive of all the materials studied, inviting the reader more particularly interested to refer to the summary articles published to date [1 ~ 5egu ~ - Armand1,2l, but rather to define the generalities common to these.

The first polymer electrolytes studied were obtained by dissolving a strong acid, the most often mineral, in a solvating polymer matrix comprising oxygenated heteroatoms or nitrogen, and therefore basic in the sense of Lewis. Thereafter, interest will go more particularly on protic electrolytes utilic ~ n ~ polyethylene oxide (POE), pattern monomer (CH2-CH2-O), as solvating matrix [De ~ endini]. POE chains being cut r: ll idcmcnt cn milicu; l ~ icl ~ ~ q-lcu: Y, il ~ sl irnrer ~ r dc lrav ~ r cn milicu ~: lrf ~ i ~ emcnl; lnlly (lr ~.
CU ~ CilU.lili, m ~ e Cll ~,; 'u ~ iioll U'L ~ (.; IUt; U ~ fUl ~ p ~ UL ~; ùnuuir ~ a une i ~ nLe ~ ie ~ ra ~ iaLion ~ ill polymer. In addition, the acid used must be low volatile to ensure the m ~ intiPn of its concentration to over time. The choice then fell on orthophosphoric acid, which is not very volatile, and whose first acidity of a pKa ~ 2 in aqueous medium is suffic ~ mment strong to induce many charge carriers, while limiting: lnt degradation of the polymer. Anhydrous orthophosphoric acid has a significant proton conductivity ~ 2.10-4 Ql.cm-1 at 20 ~ C, due to its autoprotonation and autodehydration, the balance of autoprotonation being described by the equation:
2 H3PO4 H4PO4 + + H2POi When the orthophosphoric acid dissolves in the POE, the free doublets present on the oxygen atoms of the polymer, which we can also assume here to keep its helical structure, solvate the cation H4PO4 + thus allowing the dissociation of the acid. The conductivity of these electrolytes is controlled by a phase diagram with a compound defined for a ratio of number of POE solvating units per O / P acid molecules - 4/3, and a first eutectic of Te ~ 37 ~ C for a composition O / P = 2 associated with the maximum conductivity of the electrolytes, as illustrated in figure 2.2 IA ~ and1l. It is important to note that unlike the works cited, the O / P ratio is defined relative to the number of moles of phosphorus (and therefore of acid phosphoric), not relative to the number of moles of protons (each acid molecule orthophosphoric containing three protons). This amounts to considering only the acid proton in H3PO4, and makes it possible to use a homogeneous definition with respect to the other electrolytes cited by the continuation, s ~ ch: lnt in the end that O / P = 3-0 / H.

Figure 2.2: Phase diagram of the mixture POE / H3P04, and isothermal conductivity for ce éla, b ~, lyte between 0 and 70 ~ C, according to ~ Armand2 ~.

21941 ~ 7 -It has been proposed that the eutectic composition adhesives ~ ond to a H4PO4 I cation solvated by helix, with an H2PO4- anion outside of it, as illustrated in figure 2.3. For this composition, the electrolytes are in the form of amorphous films in an infinitely metastable state, optiquemen ~ c transparent, and a conductivity of about 3.10-5 Ql.cm-l at 25 ~ C.

Figure 2.3: Mechanis ", e of solvation for coi" ~, eutectic position according to [Defendini].
The m ~ c ~ ni.cmes of displacement of the proton in these electrolytes are not yet completely elucidated because of their complexities, and this, already in the case of pure orthophosphoric acid [Dippel].
Pulsed field gradient and neutron scattering NMR studies have shown, however, that the mobility of the proton is also coupled there to the movement ~ s of the chain segments IDefendini polymer ~ u; ~ l ~ Yllulll. We can assume that the conductivity is ensured, according to a mec ~ ni.cme of Grotthus, by the transfer of the proton from an H4PO4 + cation to an H2PO4- anion associated with a méC ~ nicme of solvation / desolvation and reorientation [BerUlierl of the following ionic species:
H2PO4 + H4PO4 + solvated ~ H4PO4 'solvated + H2PO4-Considering that POE acts as a solvent equivalent to water in aqueous solutions of orthophosphoric acid, the electrolytes based on POE / H3PO4 mixtures are intrinsically acid proton conductors of pKa ~ 2, assuming that POE does little to modify the activity of proton relative to water. These electrolytes are one example of proton conductors obtained by dissolution / solvation of a CF3SO3 arnphoteric acid by a weakly ba ~ sic polymer matrix. NH2 NH2 NH3 + NH2, NH2 The chemistry of sol-gel processes has made it possible to develop another <) interesting class of glassy proton conductors obtained by polyconden.s ~ tion of organosilicon compounds such as aminopropyl trialkoxysilane and therefore called aminosils ~ _ Charbouillot, P ~ c ~ ~ and2l We thus obtain a composite material organic / inorganic whose structure, described on the ~ lgure opposite, is assumed to be that of a non-crosslinked comb polymer. These NH2 NH ~ 3 ~ NH2 NH2 polymers, stable in acidic medium, form solid CF3SO3 transparent solutions for protonated amine levels less than 50% with AMINOSILS

strong monoacids such as perchloric acid (HCl04), or trifluoromethane sulfonic acid (HCF3SO3). The rigidity of the material is ensured by the silicon frame, the tr ~ nsition temperature glassy is not observable before the start of its thermal decomposition, from a temperature of 280 ~ C. The proton conductivity, for its part, obeys a mechanism of Grotthus, by the jump of the proton from a ~ h ~ nn ammonium to an amine by tunneling along a bond hydr ~ c '~, S ~ "t" ~ .cr ~ .F ~ r. ~ i ~ rs ~ .ot ~ icr. ~, tou; ~ c ~ vr.
carbon / nitrogen. In addition, the different movements of free rotations possible around the connections carbon / carbon of the propyl segments induce a movement of b ~ l ~ ncçment of the side chains of the polymer which promote the transfer of the proton by bringing together the ~ min ~ es functions [Ch '~. Diffusion coefficient measurements by pulsed field gradient NMR made it possible to show that the conduction is ensured essentially by the proton (t + ~ 0.85 to 140 ~ C for a doping at H / N = 10% by CF3SO3H, this ratio expressing the molar fraction of protonated amine).
Electrolytes doped at H / N = 7.5% by CF3SO3H have a conductivity of approximately

3.10-5 Ql.cm-l at 25Oc [Rousseaul Unlike electrolytes based on POE / H3PO4 mixtures for which the proton activity is fixed by the force of the acid used, the proton moving by jumping from an ammonium cation to an amine in aminosils (just like liquid ammonia), the activity of the proton is determined by the force of the base serving as a relay for the protons. These electrolytes are therefore proton conductors basics of pKa ~ 11. Aminosils belong to the family of electrolytes obtained by doping by excess of protons of a basic polymer matrix.

More recently, other basic proton conductors have been studied in LIES-G [D ~ Z ~ J ~ 2]. The simplest of them is obtained by dissolving sulfonamide in a POE matrix, the sulfonamides being, in general, weakly acidic molecules of pKa = 10.4 in the water [GalwaYl. Sulfonamide is dissolved in POE via intermolecular hydrogen bonds between the oxygen atoms of the polymer, and the hydrogen atoms of the amino groups of the sulfonamide molecule. The ethers of POE are less basic than groupings sulfonamide amines, they are less likely to tear a proton from sulf ~ mide than the sl-lf ~ mide to itself. The hydrogen bonds interslllf ~ mi-the predominant in these electrolytes, it follows that the sl ~ lf ~ mides plésen ~ e a balance of autoproton ~ tion in the following POE:
2 H2NSO2NH2 ~ H2NSO2NH- + H2NSO2NH3 +
As for the POE / H3PO4 mixtures, it can be assumed that the solvation of the species cati - ni-read inside the POE propellers favor the displ ~ cment of the balance of autoprotonation to the right. The transfer of the proton is ensured in these electrolytes by the simultaneous presence of proton vacancies (molecules of s ~ llf ~ deprotonated mides) and protons (molecules of slllf ~ mides protonates) following a m ~ c: ~ ni.cme of Grotthus.

21, 941 27 The phase diagram of the POE / sulfonamide binary presents four eutectic steps for ratios of the number of solvating units from POE to s ~ llf ~ mide (O / S) of 2, 4, 9 and 30 respectively.
La ~ l --- t; of the eutectic level dimim ~ e when the concentration of sl-lf ~ mi ~ e auU ~ n ~ nte ~ the point of eutectic fusion for a composition with O / S = 2 being 25 ~ C. Because of the crist ~ linlinity of POE, the conductivities obtained at ambient are poor, of the order of 10-8-10-9 Ql.cm-l.
However, as he scra pr ~ cis ~ thereafter, with non-crystalline comb polymers unc horn, ductivity m ~ xim ~ le of about 10-5 Ql.cm-l is obtained with av ~ '' pc.i. la ~ e ~ ll.pûsLLicrl CIS =,.

The concentration O / S = 2 corresponds, as for orthophosphoric acid, to a molecule of snlf ~ mide for two solvating units of the polymer. This militates for a structure of eutectics at O / S = 2 identical to that obtained with H3PO4, with a molecule of sulfonamide protonated by helix of POE, associated with a deprotonated sulf ~ mide molecule outside of it. By ~ llelement ~ it is interesting to mention the synthesis of copolymers (PSJ600), integrating the function slllf ~ mide in their sllu ~; lures ~ by polycondensation of sulf ~ mide with POE segments di ~ min ~ s mass 600 (Jerr ~ e ~ 600-ED) depending on the reaction:

H2N ~ H2 - CH2 - O ~ NH2 + H2NSO2NH2 3 / NH ~ ~: H2 - CH2 - O ~ H - SO2 ~ H--The electrolytes based on POE / sulfarnide mixtures are the counterpart of the POE / H3PO4 electrolytes, thus, by the same principle of dissolution / solvation of an amphoteric molecule in a matrix weakly basic polymer, it is possible to lead to proton conductors for which the proton activity varies from a pKa ~ 2 for the POE / H3PO4 electrolytes to a pKa ~ 10 for the POE / sulfonamide electrolytes.

The conductivity of these electrolytes can be significantly improved by doping with a base more stronger than sulfonamide, such as gu: lnidine, playing the role of deprotonating agent. Doping allows the controlled creation of proton vacancies in the electrolyte, and the increase in the number of carriers of charges. The ion transport is then provided only by the proton vacancies. The conductivity has a maximum, depending on the doping rate, for a given concentration of sulfonamide given. Thus, the best conductivities obtained are between 10-4 and 10-5 Ql.cm-l for a temperature of 20 ~ C, i.e. a ~ lgm ~ nt ~ ion of conductivity which can reach five orders of gr ~ nde ~ lrs per Mpport to an undoped electrolyte at this temperature. These electrolytes are basic proton conductors obtained by the creation of proton vacancies.

As mentioned above ~ demm ~ nt, in order to overcome the problems related to the crict ~ llinit ~ of POE, other solvating networks were envisaged, including in the first place, comb polymers containing ~ nt a skeleton of meth ~ crylate with latéMles oligooxyethylene chains, and the formula of which is indicated below. The study focused on two polymers containing eight respectively (PMPEGM 400 Dalton) and twenty-two units (PMPEGM 1000 Dalton) of ethylene oxide on the side chain.

-The PMPEGM 1000 being semi ~ ict ~ llin, the conductivity studies were ~ ffectllé ~ s Essentially with the PMPEGM 400, it does not have any trace of tCH - c ~ -cri ~ t ~ llinité, for electrolytes of O / S compositions of 2, 4, and 30, doped with c - o gu ~ nirline for N / H concentrations of 0, 1, 5, 10, and 20% respectively (the o ratio N / H exrrim ~ nt the ratio in moles of the gu ~ ni ~ ine to the set of protons cH2 sulfonamide). For the three sulf ~ mide envic ~ ages ~ optimum cH2 concentrations dc CCI .. ùC.iiit ~, cs. obtained for a 5% N / H dopag ux, by conh-e with ~ in ~ m higher doping, the conductivity is CH3 - - - PMPEGM
lower than that without doping agent. The optimum 2 - - ~ - 2 - of conductivity results from a competition between the increase ~ tion of number of charge carriers via guanidine doping, and the concomitant increase in the transition temperature glassy polymer. Maximum conductivity of \ NH \ Ne N \, 2.10-5 S2-l.cm-l is obtained at 20 ~ C for the electrolyte of 2 <~ NH composition O / S = 2 doped with N / H = 5%. On the other hand, chemistry RR of sol-gel processes has helped develop networks (1-x) (x) (x) ~ (y) org ~ nosiliriés integrating in their structures the methyl function R = CH3 or ~> or benzenesulfonamidosil (respectively symbolized here by POE ~ MSAS and BSAS), the deprotonating base, and a chain of POE
playing the role of agent pl ~ ctifi: lnt, these materials have been called ~ sulfonamidosils (see figure opposite). The deprotonating agent SULFONAMIDOSILS chosen here is an alkoxysilane-alkylimidazoline of pKa ~ 13.
After optimization of the respective report of the different components, a conductivity of 7.10-8 Ql.cm-l is obtained at 25 ~ C, for an electrolyte integrating a chain of POE of mass 2000, and of composition x = 15% (in moles) in be ~ 7 ~ anesulfon: lmidosil and y = 30% (molar ratio relating to all of the siloxanes). Dissolving free sulfonamide in a grafted imidazoline and POE chain copolymer, part of the family of ormolytes (short for organically modified silicate electrolytes), has improved conductivities, thus, one obtains a conductivity of 10-6 Ql.cm-l at 20 ~ C with a chain of POE
mass 2000 for an electrolyte with O / S = 2 doped with imid ~ 7Oline grafted with N / H = 10%.

There is therefore a wide variety of proton conductors present ~ nt a conductivity of about 10-5 Ql.cm-l at ~ mp ~, dture ~ mbi ~ n ~, this being recognized as the minimum required for re ~ tion of electrochemical systems in thin film technology ~ The proton activity in these electrolytes can vary over a wide range of pKa ~ typically between 2 and 12. The cl ~ ccification of the electrolytes according to the activity of the proton makes it possible to define the dom ~ in ~ s of app ~ atiQn ~
possible for these m ~ téri ~ llx. Thus, the hydroxides present ~ nt ~ nt a dom ~ ine of stability in a pH window ranging from 4 to 12 associated, for example, with aminosils or mixtures POE / slllf ~ doped mide allow to consider the re ~ tion of proton batteries in thin films.

2 I q4t 27 The POEIH3PO4 proton conductors, for their part, have found an application within the framework of the development of electrochromic glazing for which the nature of the electrodes makes it more interesting the utility ~ tion of "acid" electrolytes [Ripoche]. However, rather than classifying electrolytes according to the activity of the proton, it seems more judicious to compare these following the nature of the doping allowing initiation of the proton conductivity in the solvating matrix considered. It is air ~ if possible to define three families of proton conductors ~ es, s. ~ It clc ~ ..ulj- ~ s ob ~ nus p; u:

~ dissolution / solvation of an amphoteric molecule in a solvating matrix ~ doping by excess of protons (doping "p") ~ doping by proton vacancies (doping "n") A parallel can be drawn between doping with holes or excess protons of the electrolytes with the n or p doping used to initiate electronic conductivity in the ~ semiconductors. Indeed, in both cases, the material acquires electrical conduction properties (respectively ionic or electronic) by creating an excess or a defect of charges (respectively of protons or electrons). Cepensl ~ nt the proton con ~ ctors can not be described simply, as semiconductors, by an energy band diagram.

Tables 2.1 to 2.3 thus summarize the different conductivities obtained at 20 ~ C, respectively for the family of protonic electrolytes obtained by dissolution / solvation of a molecule amphoteric, by doping by excess of protons, and finally by doping by vacancies of protons. He was added results for electrolytes using polyethylene imine (PEI) [RaVe ~ l, this polymer being the subject of numerous studies and being NNH
poter til ~ llement a candidate for the industrial development of windows POLYETHYLENE IMINE
electrochromic. PEI can be doped with acid over a wide range of concentration, typically from 0 to 2 molecules of acid per amine (defined by the mol ~ ire HIN ratio of acid by arnine function), one thus passes from an electrolyte for which the proton is transferred by jump of a cation ~ mmonium to an amine, to a mec ~ ni ~ me of transfer via the acid molecules ['aSce ~ 51.
Thus, for the electrolytes obtained by doping PEI with H3PO4, the activity of the proton depends on the acid concentration and goes from pKa ~ 11 to pKa ~ 2 [Tnnquetl.

Table 2.1 l ~ le ~, lrulytes obtained by ~ i ~ ssl ~ lisl- / solvation of a .I. ~ Lé ~ 'e amphotère.
Polymer dopant Composition Conductivity pKa amphoteric solvating agent (O / dopant) (S2-l.cm-l) relating to water H3PO4 POE O / P = 2 2.105 ~ 2 H2NSO2NH2 POE O / S = 2 8.10-1 ~ ~ 10 H2NSO2NH2 PMPEGM400 OIS = 2 z 10-5 ~ 1 0 Autodoped PSJ600 O / S ~ 14 3.10 ~ ~ 12 21 q41 27 Table 22 ~ IE ~ tr. ~ Tes obtained by doping by excess of protons.
Dopant Matrix Composition Conductivity pKa polymeric acid (Ql.cm-l) relative to water CF3SO3H Arninosils H / N = 7.5% 2.10 ~
H3PO4 PEI H / N = 1.5 2.10 ~ ~ 2 Table 2.3 Basic electrolytes obtained by doping by proton vacancies.
Base Base Matrix Composition Conductivity deprotonating polymer (Ql.cm-l) Sulfonamide Guanidine POE N / H = 0.1 O / S = 2 z 10-4 Sulfonamide Guanidine PMPEGM400N / H = 5% O / S = 2 2.10-5 BSAS Imidazoline Sulfonamidosilsx = 15% y = 30% 5.10-8 Ormolytes grafted sulfonamide O / S = 2 N / H = 0.1 10-6 Living chemistry offers many examples of the richness of the different possible mechanics.
of proton transfer, the latter playing a large role in biology, proton transfers being involved in practically all biological processes [HadZi]. Indeed, the properties functional proteins depend on their spatial organization, and therefore on the state loading amino acids components. The transfer of the proton between the Cil2 - CH - COOH
basic and acidic sites of proteins, as well as between these and the medium F ~ NH2 aqueous, therefore influences their functionality by modifying their N ~ NH
conformations. Among all amino acids, histidine plays a role HISTIDINE
important in the proton transfer processes due to the presence of the gro ~ -pem.ont imidazole, protonating at pH ~ 7, in particular as a proton transfer mediator in many enzymatic reactions. This last molecule gives strong bonds hydrogen, with forces at least comparable to those of water.

Also, has it been considered the specific study of imid ~ 7ole for the synthesis of conductors protonics. Preliminary tests ~ ires showing that the mixture of imid ~ ole and triflic acid presetlt ~ it has a melting point which may be near ambient temperature depending on its composition, and a conductivity of at least 10-5 Ql.cm ~ l. Before considering mixing them with a polymer matrix, it was therefore established the phase diagram of the binary mixture imi ~ 7ole / acid, this being essential for the understanding of the conductivity results, as has been specified before.

~ 2! 94'127 II-2-2 Phase diagram of the imidazole / acid binary:

The phase diagram of the im ~ 7olelacid mixtures makes it possible to determine, for a composition given, the temperature range where the system is in the molten phase, and the composition eutectic with the lowest melting point, this one because ~ ct ~ risking the most structural d serdonr. ~ du m. ~ t ~ r ~ u. , ~ t, e ~ a '' ~ f, r '' .- "'.' ~ e, llc; i scanning (Cf Annex) is a tool of choice. To assess the influence of the nature of the acid on the properties of the binary imidazole / acid, the study focused on the mixture of imidazole with respectively, the acid trifluoroacetic (CF3COOH), triflic acid (CF3SO3H), and bis (trifluorom ~ th ~ nesulfonimide) (CF3SO2NHSO2CF3) commonly symbolized by HTFSI.

Preparation of DSC samples ~ To avoid repeated handling of acid, salts corresponding to each acid are p, ~ p ~ s by adding l ~ Pntpm ~ p ~ nt to an aqueous solution of imi ~ 7ole a sta ~ chiometric quantity of a titrated acid solution. After being agitated for one hour, the solution is evaporated under vacuum, and the salt dried for several days under vacuum at room temperature, then stored in a glove box. The samples for DSC analysis are then prepared in a glove box, by grinding in an agate mortar one of the salts and the quantity imidazole n ~ 6cP ~ ss ~ ire to obtain the desired composition of the imidazole / acid binary, the latter being expressed as the ratio of the number of moles of imidazole per mole of salts (I / S). The mixture obtained is then sealed in et capsules: anodized aluminum sheets (Dupont brand capsule) For each sample, two successive passages in DSC are carried out, and this for a speed of scanning temperature of 10 ~ C / min.

First, the 12 0 Figure 2.4 shows the dia- - imidazole / triflic acid-gram of phase proposed with 1 0 ot; ~
triflic acid. Like the ~ ~
shows this one, the binary ~ 80-., 4, / ~
has a eutectic level, ~
for a composition VS = 3 ~ /
(i.e. 48% by weight of im ~ 7ole in the mixture imida- ~, O c "
zolelacide) d'unpoint defusion 20 c ~~~ 'n /
Te ~ 12 ~ C. It is important to OOOO ~ / o OOOOO -note that this is a 0. I, I, I, I

phase diagram "dyna-% by weight of i"'~ - ~ relabvement with salt mique ". Indeed, the point of FigurQ 2.4: Phase diagram of the mixture i".: '~;.' e 'aclde trifllgue.
fusion of the eutectic observed depends on the heating speed, and goes from 12 ~ C for a speed of 10 ~ C / min, to 7 ~ C for a ~ 10 speed of 2 ~ C / min, this difference of 90 temperature characterizing the 80 imidazole / trifluoroacetic acid kinetics of fusion of the part ~ Q ' cri.ct ~ llin ~ du m ~ - ~ ri ~ 70 _ By comparison, 1 ~ figure 2.5 _ 60 pri; ~ nL ~ i ~ phase diagram ~ "50 _ obtained by substituting the acid E ~ L ~ ll p triflic with trifluoro- acid ~ 40 - \ 'n /
acetic, in the binary imida- 30 ~ ~ ~ OO
zole / acid. Following the results ~
obtained with triflic acid, one 20 lllllll 40 50 60 70 80 n / a 1 oO
can be satisfied with the study of the% by weight of i "~ iq ~ ele relative to salt phase diagram for a Figure 2.5: Phase diagram of the mixture i ". '~ o.' ~ 'CF3COOH.
composition between I / S = 1 and the imi ~ 7.ole pure, this restricted domain being sllffl.c ~ nt to determine the composition and the melting point of the eutecti-read. As with the binary imi ~ 7.ole / triflic acid, the composition eutectic corresponds to a binary composition at VS = 3 and melts at a temperature Te ~ 34 ~ C.
The phase diagram is more complex with HTFSI, due to the appearance of phenomena thermal supplements ~ ires, the partial phase diagram established within the framework of this study is shown in Figure 2.6. He ... not - 100 III ~ This is only a draft of imidazole / HTFSI. o phase diagram, which allows OO. however to establish the area of .o temperature for which the binary is in the molten state.
_ 40 ~ Unlike the two ç ~ çmples precedents, eutectics seems 20 -. - correspond to a composition - ~ ..... ~. I / S = 4. Given the o- 'O ~ - apparent complexity of the diagr-- phase ami imi ~ 7olelHTFsI ~
- 200 '2 ~ 0' 4 ~ 0 '6 ~ 0' 8 ~ 0 '1 00 it would be necessary to make some % by weight of imi ~ 'e relative to salt a more detailed study, but this Figure 2.6: Didyldril ~ lle phase mixing; "~ JHTFSI. Is not the main goal of present work, the results partial obtained being s ~ lffi.c ~ nt to judge the influence of the nature of the acid on the properties physics of m ~ lqnges imi ~ q7.ol ~ 'acid.

All the results concerning the - 60 glass transition temperatures obtained, o HTFSi ~ "- 65 - o CF3COOH
for the three acids studied, and this for .--, ~ CF SO H
different compositions, are given on the O ~ 70 ~ ~ 3 3 figure 2.7. In the case of triQic acid and a ~ 75 _ ~~
HTFSI, the eutectic composition seems ~
correspond to a minimum read ~; ll dc la _ - 80 ~ A
glass transition temperature, this is ~ 85 ~ 'f / V
consistent with the fact that the eutectic corresponds ~ ll generally at the most disordered structure of 0 5 l / S 16 mat ~ .rian Figure 2.7: Influence of coi "~ osilion of surleurteo mixtures, f ~ érdllJre glass transition.
For its part, Table 2.4 summarizes the whole results for the three binaries studied, and indicate eg ~ l ~ men ~, for comparison, the point of fusion of the stoichiometric m ~ nEe imi ~ la7Ole / acid (Tf salt).
Table 2.4: Main physical characteristics of the binaries i ", '~' acid. Acid I / S eutectic T ~ eutectic Te Tf sel CF3COOH l / S = 3 -72 ~ C 34 ~ C 136 ~ C
CF3SO3H l / S = 3 -76 ~ C 1 2 ~ C 1 03 ~ C
(CF3SO2) 2NH l / Sz4 -88 ~ C z-10 ~ C 65 ~ C
From it, it is possible to conclude that the nature of the acid plays a role çcsenti ~ l on the properties physics of imid ~ ole / acid mixtures. In particular, on the eutectic melting temperature, which decreases by about 45 ~ C by snbstih ~ ant triQic acid by HTFSI, the melting temperature of stoechoimetric salt decreasing by 70 ~ C. By comparison, the transition temperature vitreous is little modified by the nature of the acid (z 15 ~ C), this seems to indicate that this parameter essentially characterizes the imidazole molecule, for the acids considered. It is probable that here intervenes both the interactions between the imidazole molecules, in the form of bonds hydrogen between the nitrogen atoms, and the crystal structure of imidazole, the molecule presenting a plane geometry ~ a ~ l, This point will be detailed later to try to explain the mechanisms of con-1uctivities envisaged for these electrolytes.

It would be, t ~ .essallt to be able to correlate these results with the respective strength of the acids used, to this Finally, Table 2.5 presents the pKa values available in the literature for these acids, ~ es ~ ecd ~, e.llent in water, and in pure acetic acid. These latter values are obtained from pardr of the measurement of dec ~ laEe in 1 H NMR of the acid proton signal between pure acetic acid, and a acid soludon studied in acetic acid, assuming there is a correlation between the pKaet this one.

219a127 Table 2.5 pKa of the acids used in the binary i ... ~ 1 'ac.i ~ e in water and in acid 2 ~ ketigue.
Acid pKa relative reference pKa relative reference with acetic acid water CF3COOH 0.2 [Streitwieser] 11.4 [Foropoulos]
CF3SO3H -14 [Bordwell6] 4.2 [Foropoulos]
(CF3SO2) 2NH 1.7 [F ~ ~ ooulos] 7.8 ~ Foroooulos ~

It appears that the strength of the acids strongly depends on the solvent, in particular, a inversion between trifluoroacetic acid and HTFSI, depending on whether the pKa is determined in the medium aqueous or non-aqueous protic. Indeed, in an aqueous medium the value of pKa of HTFSI is astonishing ~ mmPnt high, and corresponds to that of a moderately strong acid. We assume that this acid deriving from a soft delocated base, the water dispersing the anionic charges via bonds hard hydrogen, in the Lewis sense, is a bad solvent for it. However, the measurement of acidity in phase g ~ 7P, uce, that is to say without any solvation phenomenon occurring, shows that the HTFSI is one of the strongest acids known to date [Koppel]. At this point, it will not be pushed anymore far this reasoning, this particular point will indeed be studied more precisely later (Cf II-3-1). In the present case, the imidazole / acid binaries being anhydrous protic media, it is more relevant to use the scale established in acetic acid, according to this, the HTFSI is a acid with a strength comparable to sulfuric acid (pKa = 7 in CH3COOH). At the sight of these r ~ snl ~ ts, it seems difficult to establish a direct correlation between the physical properties of binary imi ~ 7ole / acid and the strength of the acid, but a ten ~ nce, CF3SO3H and HTFSI being more acids stronger than CF3COOH. We can argue for explain the results obtained with HTFSI, ~ ~ 50 that the eYictPnce in the molecule of connections ~ "55_ flexible SNS [Armand2l, is favorable to a. ~ /
decrease in eutectic temperature. o c - 65 -c ~ ~
The substitution of all or part of the imidazole for ~ 70 ~ /
by 4-methylimidazole (MeIm), should be ~ 5 /
allow to assess the influence of the base E - 75 ~ ~
heterocyclic, on the transition temperature - 8 0 l III
villous. For the imidazole / CF3SO3H mixture, ~ 2 0 4 0 6 0 8 o 1 0 o % by weight of 4-méU ~ yli, l ,. ~ '~ 7O'e whatever the composition (I / S = 2, 3, 4) ~ Figure 2.8 Influence of the suhstituti ~ n of the i ". ~ le this temperature increases with the quantity of Melm for the ~ mixture Im / CF3SO3H at US = 3.
MeIm, figure 2.8 presents the evolution for the composition VS = 3. Thus, the glass transition temperature increases by 23 ~ C when substitutes imidazole for MeIm. The strength of the two bases being substantially equivalent, ~ u ~ mPnt ~ tion of the glass transition temperature is probably related to the structure of the molecule, the presence of ~ ulx ~ ent methyl in ~; ",; ~ the number of degrees of freedom.

21941 ~ 7 -II-2-3 Conductivity of protonic electrolytes:

The dirré ~ mPl ~ ngP ~ c imi (l ~ 7ole / acid present ~ nt a eutP ~ ti-lue melting at a le- ~ lp ~ ralule inf ~ riPllre or close to that of ambient, these thermal properties make it possible to envisage the use of such mixtures for the realization of electrochemical systems function ~ nt al ~ temp ~ rature ambient ~ nte.
Cepen.l ~ n ~ of a poirll of applied sight, ~ the nombl ~ ux ~ rav ~ ux re; liisej these ten last years ~ onL
emphasized the interest of coupling the development of new materials for electrochemistry with technology sector in thin films [F ~. However, it is known that it is possible to confer properties of elastomers to a molten salt by mixing it with a small proportion (~ 10% in weight) of a high mass polymer [Angell] ~ provided that there are interactions between it and the liquid. In fact, in a high mass polymer, there are many tangles between the chains which allow to obtain a m: ~ t ~ ri ~ ll elastomer, and this for small amounts added.
Thus, it is possible to impart a mechanical outfit to the materials described above.
incorporating POE with a mass of five million (POE 5.106), in order to obtain electrolytes solid polymers (EPS).
Study of conductivity of conductors protonics is limited to mixtures ~
imid ~ ole and methylimidazole with acid - ~ NQ ~
triflic, and mixtures of imidazole with N ~ ~ NHCH3 ~?
5-phenyl-1H-tetrazole respectively. and 5-phenyl-1H-t ~ r Acid ~ luè ~ e ~ llfonique p-toluenesulfonic acid, an acid with a flat aromatic molecule. The quantity of polymer respectively in binary is defined by the O / H ratio, expressing the ratio of the number of moles of poly (oxide) solvating units ethylene) to the number of moles of staichiometric salt.

Preparation of samples ~ 7 ~ 7ti ~ 10ns: A solution is prepared in anhydrous acetonitrile from binary to desired composition, and the amount of POE 5.106 n ~ cess ~ ire (this is placed before a few days in a vacuum desiccator at room temperature then stored in a glove box). After homogeneous ~ ic ~ tion of the solution, it is poured onto a steel strip, the film obtained is then stored for 48 hours before use, in order to evaporate the solvent. We then hot press a of ~ .me steel strip, in order to obtain the conductivity cell, which has a surface of 3.9 cm2. All these manipl ~ l ~ tions are carried out under argon in a glove box. The three samples liquids are prepared by grinding in an agate mortar. Once the mixture is liquefied, permeates a s ~ pa, ~ their in Celgard (e = 50 ~ 1m), and it is enclosed in a button cell between two steel p ~ till ~ 5. All eçh ~ ntillon ~ thus obtained are sealed in sachets and ~ nrhes so to be able to carry out their manipulations in the air. Table 2.6 summarizes all of the éçh ~ ntillol-s p, ép ~ s, and specifies their physical and visual aspects after a period of 48 hours of , ~ h ~ in a glove box.

.

Table 2.6 Cor ~ ition and appearance of the different samples, ~ ltpales for conductivity.
BINARY I / SO / H% weight ASPECT ASPECT
- PHYSICAL VISUAL POE
11 very sticky 2: 1 2 20 3 27 tran ~ n ~ rent Imid ~ ole / 2.5: 1 18 triflate 3: 1 2 17 sticky imidazole 4: 1 15 1: 2 15 1: 3 11 trouble 1: 4 9 Imidazole / 1 13 very sticky triflatede 2: 1 2 19 transparent sticky methylimidazole 3 26 sticky Melm / Melm, tr 2: 1 2 18 transparent sticky 2: 1 24 2.5: 1 2 22 opaque non-sticking Imidazole / 3: 1 20 phenyllel, ~ ole 2: 1 liquid liquid 2.5: 1 no translucent POE little 3: 1 after viscous mixing Imidazole / 2: 1 2 22 crystalline not very sticky p-tol.sulfon acid. 3: 1 19 a little sticky Conductivity measurements: The conductivity measurements were carried out by us at the Hydro-Québec Research Institute (IREQ), on an automatic measurement bench developed internally, in series of ten cells using an impé-l ~ ncçm ~ tre Hewlett-Packard HP 4192A coupled to a 10-channel multiplexer The cells are placed in an oven having a cooling (licsPmPn ~ with liquid nitrogen, and regulating better than +/- 1 ~ C All these devices are controlled by a micro-gold ~ in ~ e - r HP 9000 Measurements are always taken from 60 ~ C in steps of 20 ~ C, and t ~ rr ~ -l .. ~ s after 4 hours of stabilization from 60 ~ C to 20 ~ C, then after 2 hours from 20 ~ C to -40 ~ C.

- 21 ~ 4127 -First, Figure 1 o- 2 ~ 1 1 1 1 1 1 2.9 has conductivity 3_ ~ -obtained at decreasing temperature ~ -healthy, for an electrolyte of E 10 4 ~ 3 composition I / S = 2 in mixture ~ E 10-5 imida70l ~ / ~ clde trifllqu ~ pourd ~ so 6 -u / ~ r ~ p ~ v ~ m ~ nL ~
(i.e. for around 10 to 30% in ~ 10-7 r ~ O ~ H = 1 ~ ~ -POE weight 5.106). Figure ~ -8- ~ O / H = 2 2.10, meanwhile, allows ~ o O / H = 3 co - lpalel the obtained condllctivity 10 at increasing temperature, relative- 10- 1 0 temperature temperature 3 3.2 3.4 3.6 3.8 4 4.24.4 decreasing. Electrolytes Figure 2 9: Influence of the quantity of) OEs on the conductlvit for O / H = 2 and 3 do not have the mixture i ". '~ - ~' - 'ac" I triflique for a cGn "~ os; tion al / S = 2.
no hysteresis phenomena, between the measurements made at decreasing and increasing temperature, the kinetics of recri.ct ~ llic ~ tion for these two compositions is therefore slow over the measurement time scale (a few hours), unlike the composition O / H = 1 which is also the least conductive. Even if the better conductivities are obtained at 1 o-2 ~ ll O / H = 2 (20% by weight of POE), for these 10 3 ~ three compositions the electrolytes can o 10-4r ~ be used for electro- systems E 10 5- ~ _ chemicals operating at temperature ~ 10-6- ~ ambient, the conductivities being ~ 10 7 o particularly high, and around g 10-3 Ql.cm-l at 20 ~ C, i.e. at least 10-8- OO / H = l T down c minus an order of magnitude at this ~ 10 9 - ~ O / H = 1 T up - 1 0 ~ II ~ temperature, with anhydrous electrolytes 10 3 3.2 3.4 3.6 3.8 4 4.2 4.4 studied so far.
1000 / T (K) Fi9l.nuarl.ere2i1 ~~ HJ ~ l ~ el ~ earc ~ Jé ~ trJiflel.queào ~ dsuct2vite ~ cplonugrée The composition O / H = 2 presenting the with POE 5.106 for a composition with O / H = 1. better conductivity, the study focused on following on the influence of the composition of the binary for this poly (ethylene oxide) composition, the results are shown in the figure 2.11, for I / S = 2, 2.5, 3, and 4, and in figure 2.12, for I / S = 1/2, 1/3, 1/4. The electrolyte of composition O / H = 2 and I / S = 2 still has the best conductivity, except for the composition I / S = 2.5, and this for a temp ~ l ~ tulc of -20 ~ C, for which we probably observe a ph ~ name ~ does not delay the cry ~ t ~ tion of the electrolyte. It is interesting to note the appearance of the ., curve for the composition VS = 3, the conductivity is certainly lower than that for the composition at VS = 2 by a factor of three at a temperature of 60 ~ C (10-3 vs. 3.10-3 Q-1.cm-l), but joins it prog ~ essivelllent at low temperature, this composition corresponding to that of the eutectic of binary imi ~ l ~ 7Ole / triflic acid. Ultimately, the electrolytes of composition VS <1 pr ~ sçnt ~ nt a conductivity at least an order of magnitude lower than the electrolyte at VS = 2.
_ 10, ~ 0 3 ~
E10 ~ ~ O 8 E 10 a O
~ E 10 4 - OO ~ E 10 5; ~
s 10-5 ~ ~ ~ c 10-6 _ ~ O
1o 0 211 ~ - ~ 10 ~ 7 10 r O 3; 5/1 0 ~ 10 ~ 1/3 o 10-8 ~ OO 10 1/4 ~.) 9. IIII lI O ~ - 1 0 10 3 3.2 3.4 3.6 3.8 4 4.2 4.4 10 3 3.2 3.4 3.6 3.8 4 4.2-4.4 1000 / T (K) 1000 / T (K) Figure 2.11: Influence of the composition of Figure 2.12: Influence of the composition of binary i., '~ -.' e 'Cf3SO3H containing POE binary i "' ~ o. '~' CF3SO3H containing POE
5.106at 0 / H = 2.on the conductivity forl / S> 1 5.106at O ~ H = 2.on the conductivity forl / S <1 In order to assess the influence of POE on the physical properties of the binary imidazole / CF3SO3H, it is n ~ cess ~ ire to establish the partial phase diagram of the binary in its presence. Instead of establishing this for a composition in POE at O / H = 2, it is preferable to study different compositions of the binary mixed with 20% by weight of POE, to be able to compare it with the established phase diagram previously without POE. To avoid the use of a solvent when preparing samples, this n ~ ceased ~ nt careful drying, POE 5.106 is substituted by polyethylene glycol mass 3400 (PEG-3400), corresponding to the fastest kinetics of crict ~ llic ~ tion binary mix in an agate mortar in a glove box. The alcohol functions the end of the chains being present in small quantities, and of a pKa ~ 15, we will suppose that they do not influence the physical properties of the m ~ t ~ ri ~ u The proposed phase diagram is shown in Figure 2.13. Figure 2.14 and Table 2.7, meanwhile, allows ~ tçn ~ to compare the phase diagram and the physical properties obtained in abs ~ nce and ples ~ llce of POE. Poly (ethylene glycol) does not modify the eutectic composition, obtained for VS = 3 in both cases, on the other hand the tempe, ~ lure of the eutectic level ~ u ~ mente slightly from 5 ~ C, and thus goes from 12 to 17 ~ C. The eutectic composition still corresponds to a minima marked with temp ~ c; of tr ~ n ~ ition vill ~ use at -57 ~ C, or a ~ mPn ~ tion of 19 ~ C.
Table 2.7 Influence of POE on the physical properties of the binary i ", ~ - Triflic acid.
Electrolyte I / S eutectic Tv eutectic Te Tf salt Without POE l / S = 3 -76 ~ C 12 ~ C 103 ~ C
20% by weight of POE l / S = 3 -57 ~ C 17 ~ C 33 ~ C
17.

2 1 q4 1 27 -1 1 1 1 70 11 ~ II
40 _ ~ _ 60 - 20% by weight of POE "e Without POE, 2 0 - A ~ _ 0 5 o -_ ~ 30-- ~
E-20 - E 20 -o ~ ~, ~
~ a ~ ~ ~ ~ 10 ~ --e-- ~ -o- ~ n -60 1 1 1 o o 20 40 60 80 100 20 3040 50 60 70 80 % by weight ~ 'e relative to salt% by weight ~ 70'Q
Figure 2.13: Partial phase dkyldlllllle of Figure 2.14: Cou ~ ison of the diagram binary i "~ '- ~' - 'd - ,; trfflique mixed with 20% phase of binary i". '~,' trffligue acid alone, and by weight of poly (ethylene gtycol) of mass 3400. phase diagram in presence: presence of PEG-3400.

The results of figure 2.15 allow to evaluate the influence on the conductivity of the substitution, all or part of the imi (1 ~ 7.ole with 4-methylimid ~ 7.ole, in the particular case of the electrolyte of composition I / S = 2 and O / H = 2.
The conductivity of electrolytes 10-rl decreases with increase of 10-3 ~
the amount of methylimidazole. ~ ~ 3 This is consistent with study '~ 10-4 ~
previous showing the increase- - 1 0- 5 ~~ g ~.
temperature of ~ a glass transition of binaries ~ "10 o imid ~ 7.ole / CF3SO3HWhen we ~ 10-7 ~ o Imidazole o ~
substitutes imidazole with ~ -8 ~ ~ Im / Melm (2/1) ~ -methylimid ~ 7.ole. Figure 2.16 ~ O Im / Melm (1/2) 0 shows the influence of quantity 10 9 ~ o Methylimidazole o of POE on conductivity for 10 10 ~
an electrolyte at I / S = 2 per 1000 / T (K) 4.2 4, 4 which one imid ~ 7Ole out of three is Figure 2.15 Influence of 4-methylimidazole on conductivity substituted by a methylimid: .7.ole from the binary i "~ F 'ac; Je triflique, in the presence of POE at O / H = 2.
(~ n / MeIm (2/1)), for O / H of 1, 2, and 3. The best conductivity at low temperature is obtained for O / H = 1, the conductivities from 20 ~ C being substantially equivalent. The pr ~ sellce of methylimi ~ l ~ 7Ol ~
modifies the quantity of POE for which the optimum conductivity is obtained, however for this optimum (O / H = 1), the conducdvities are equivalent to those obtained without methylimi ~ l ~ 7.ol ~., as illustrated in Figure 2.17.

lB

21941 ~ 7 _ -2 ~ 1 1 1 1 1 1 _10 2 ~ 1 0 - 3 ~ ~ 'E 1 0 - 3 - 10 4 - o E 10 4 ~ ~

~ 10- 6 o ~ ", 1 0 5 ~, ~ - 1 0 ~ 7 r ~ O / H 2 o ~ 10 6 Ir ~ 'e 0 / H = 2 10-8 ro 0 / H = 3 0 ~ o 10 7 ~ Melm (1/2) 0 / H = 1 9 1 1 1 ~ - 8 3 3.2 3.4 3.6 3.8 4 4.2 4.4 10 3 3.2 3.4 3.6 3.8 4 4.2 4.4 1000 / T (K) 1000 / T (K) Fi ~ ure 2.16: Influence of the quantity of POE Figure 2.17: Comparison of the conductivity of 5. 10b- on the conductivity of i., Eldnge I m / Melm (2 / lJmixtures Im / Cf3SO3H at l / S = 2 and O / H = 2 and and triflic acid for a co "", osition at l / S = 2.Im / Melm (1 / 2.1 / CF3SO3H at l / S = 2 and O / H = 1.
The nature of the acid playing a key role on the physical properties of binaries imi (l ~ 7ole / acid, it is interesting to study the effect on the conductivity of acid substitution triflic with p-tolu ~ nesulfonic acid and phenyltetrazole. Results for phenyltetrazole for VS of 2, 2.5, and 3 at O / H = 2 are given in figure 2.18 ~ The best conductivity is obtained for VS = 3, but this is about an order of magnitude lower at 20 ~ C compared to the same electrolyte 10-2 ~ II ~ II
sition u ~ ilic ~ nt triflic acid ~ ~ -The conductivity of the mixture ~ 10 ~ ~
imidazole / phenyltetrazole alone 'E 1 0 ~ 4 ~ ~
is shown in Figure 2.19. 5 _ Even if it is less, the 'E 1 o ~
comparison with that obtained ~ 10 0 in the presence of POE gives' 10-7 ~ ~ -information on the inflllence of ~ 8 - ~ 2: 1 o the presence of POE on the ~ 10 ~ 2,5 / 1 conductivity. We can already 10-9 ~ o 3: 1 o -observe an inversion in -10 ~ I o ~. 10 3 3.2 3.4 3.6 3.8 4 4.2 4.4 1 evolution of conductlvlté in 1000 / T (K) function of the composition of Figure 2.18: Influence on the conductivity of the co ", f ~ osilion du binary which proves the interaction ".el..nge in .. ~ J ', ~, hénJ ~ ltétrazole contenanl from POE 5.106 to OA ~ = 2.
between POE and binary imidazole / triflic acid. The most interesting result is presented in Figure 2.20, by the comparison of the results for the respective optimums of conductivity in the presence (O / H = 2) or no POE. n appal ~ l that the addition of POE significantly improves the co ~ .h ~ cl; lives ~ (one to two orders of quantities depending on the temperature) of the binary imidazole / phenyltetrazole for the temperatures inf ~ ..; ~ .... ~ à60 ~ C.

2194 ~ 27 .

_ The conductivity of the binary im ~ 7okolphenyltétra7-ole without POE allgm ~ nte abruptly above 40 ~ C; however, the binary with composition I / S = 2 has a melting point at a temp ~ u ~ of about 50 ~ C and the jump in conductivity is prob ~ blP.mPn ~ to be linked to its fusion.

10-2 ~ ~ 10 10-3 ~ 2/1 ~ ~ 10 3 ~~ OO 2/1 Without POE ~
~ ~ 2.5 / 1 - "1 ~~ 4 ~ C ~ O 3/1 O / H = 2 E lo S ro ~ 0 311 ~ ESO

10 9 - b- ~ '10 9 o -10 1 1 1 1 ~ -10 1 1 1 1 ll 3 3.2 3.4 3.6 3.8 4 10 3 3.2 3.4 3.6 3.8 4 4.2 4.4 1000 / T (K) 1000 / T (K) Figure 2.19: Influence on the conductivity of Figure 2.20: Influence of the presence of POE
co ", f, osition of", mixture i "~ 'phenyltetrazole. on the conductivity of the mixture Im / phenyltetrazole.

In the same way, figure 2.21 presents the results of conductivity with the binary imi ~ 7el ~ acid p-toluènesl-lfonique for I / S of 2 and 3, mixed with POE at O / H = 2. As for binary imi ~ 7ole / phenyltétra_ole, the best conductivity is obtained for the composition I / S = 3, moreover, this is roughly equivalent to that obtained with the binary imida_ole / phenyltetrazole at same composition, as shown in Figure 2.21.
-2 ~ ~ 10 2 ~ E 10 3 ~ ~ '' E 10 3 ~ OQ
_. 10-4 _ O ~ _ 10 O
E 10-5 r ~ ~ 'E 1 0 5 - ~ ~
10 6 _ o ~ ~ ~ 10 6 _ b : ~ 10 ~ 7 - o Acidetriflique 10-8 _ O 2: 1 ~ 10-8 - ~ Toluenesulfonic acid O 10-9 _ ~ 3: 1 ~ ~ 10 - O phenyltetrazole o 3 3.2 3.4 3.6 3.8 4 4.2 4.4 3 3.2 3.4 3.6 3.8 4 4.2 4.4 1000 / T (K) 1000 / T (K) Figure 2.21: Influence of b co "~ os; tion duFigure 2.21: Comparison of the conductivity for , ~ clange 1 ~ If / p-toluene sulfoniguel' triflic acid p-toluene sulfonic, and the with POE at O / H = 2, on the conductivity. phenyltetrazole al / S = 3 with POE at OKH = ~

II-2-4 Mé ~ ~ e of conductivity of the proton in the molten salt:

In the introduction, let us point out that given the melting point of the eutectic of the mixture imi ~ 7ole / triflic acid at a temperature of 12 ~ C, and the conductivity obtained at 20 ~ C of 10-3 Ql.cm-l, when this mixture is dissolved in poly (ethylene oxide), this binary can be considered to be a proton conducting molten salt. Before trying to describe the meC ~ assumed conductivity nism of the proton in these electrolytes, it is first of all n ~ cess ~
pl ~ specify the properties and nature of the imidazole molecule.

The imi ~ 7ole belongs to the f ~ thousand of azoles which include the cycles with aromatic characters to five atoms which conl; p ~ n ~ two nitrogen, one nitrogen and one oxygen, or one nitrogen and one sulfur. Its structure electronics is special because the pair of free electrons in the nitrogen atom of the NH bond is delocalized and is part of an aromatic system with 6 electrons 7 ~. But the free pair of the other atom nitrogen is available and can accept a proton. The imid ~ 7ole is therefore an amphoteric molecule have two acid-base equilibria, as illustrated in Figure 2.22, which can both accept ~
a proton of pKa = 7, and lose one of pKa = 14 [Ca ~ nl.

H - N / ~ \ N - H pK. - 7 N ~ N - H pK. = 14 N / ~ \ N
Figure 2.22: Amphoteric character of imidazole.
On the other hand, whether in protonated form (imidazole is then in cationic form) or deprotonated (imidazole is then in anionic form), the charge present on the aromatic cycle is delocalized by resonance on the two nitrogen.

The proton of an imidazole molecule linked by a covalent bond to a nitrogen atom, can initiate a hydrogen bond with the nitrogen atom free of an adjacent molecule, as illustrated in figure 2.23. The ~ ~
presence of such hydrogen bonds explaining the boiling point Figure 2P23: TIiNaiHNon hydro high of imi ~ 7ole (Teb = 263 ~ C). The geometry of the NHN bridges thus gene between two i "~~ s ~ s formed was studied by X-ray diffraction, and diffraction of neutrons, the latter measurements being the most reliable [Ma ~ ars ~]. So in the particular case of imidazole, neutron diffraction studies show that the position of the proton is not symmetrical to the two nitrogen atoms, the length of the NH bond being 1.038 ~ compared to a length of 1.83 ~ for the N --- H hydrogen bond (i.e. 2.86 ~ for the N --- HN bond). The imi ~ 7ole molecules are therefore linked to each other via hydrogen bonds intermolecular. Therefore, we can consider this en ~ h ~ înem ~ nt as a pseudo-polymer, based on li ~ hydrogen iconconc.

21 ~ 41 27 ._ Like slllf ~ mide or orthophosphoric acid, it is possible to write a balance autoprotonation for imid ~ 7.ole, as illustrated in Figure 2.24, however, in order to induce a cor ~ uction protonic subst ~ ntiPllP, it is nPcess ~ ire to create charge faults in the m ~ tP.ri ~ -2 N ~ N - H _ N / ~ \ N + H - N / ~ \ N - H
Figure 2.24: Balance of self-association of the 7ole.
To do this, the imidazole molecules are partially doped with an acid, in order to introduce an excess of protons. Figure 2.25 shows the mec ~ ni.cme of conductivity envisaged, in the particular case of the binary imidazole / triflic acid, for the eutPctic composition at VS = 3.

H - N / ~ \ NHN ~ NHN ~ rJ HN ~ N - HTran ~ proton fert along l = \ I = \ / = \ I = \ ~ Rotation of imidazoles and transfer H - N ~ NHN ~ NHN ~ NHN ~ VN - Hdu proton at the end of the chain.

r ~ r = \ / = \ ~ ~
H - N ~ NHN ~ N l; N ~ N ~ N ~ N - HReturn to the initial state.
Figure 2.25: Possible proton conduction mechanism in molten salts of; ll..c '- ~ le.
It is of course obvious that this mechanism is relatively simple and very schematic, however, it highlights the need to rotate molecules imidazole, to allow the proton to move over long distances. This mechanism of deployment: -cement is ultimately quite similar to Grotthus mechanic used to describe transport of the proton in an aqueous medium. Since protons move from a protonated imidazole molecule, to a free imidazole molecule, we can consider these electrolytes as conductors neutral protonics, the imid ~ 7ole deprotonation pKa being 7 in an aqueous medium.

In a way, the proton conductors based on an imidazole / acid mixture, and based on sulfonamide, for example, can be viewed as a generalization of electrolytes aqueous protonics. Indeed, one can eg ~ lPment considered an aqueous solution of acid electrolyte, as an excess proton doping of an amphoteric solvent, and an aqueous solution of a basic electrolyte, like doping by creation of proton vacancies of an amphoteric solvent.

II-2-S Field of electrochemical stability:

The stability domain of the imidazole / triflic acid electrolyte at I / S = 3 with POE SM at O / H = 2 is de ~~ ed in ~ ltili ~ nt a platinum microelectrode of 125, um, and a pseudo-electrode of reference, which is made up of a nickel sheet. Indeed, this metal contains ~ n ~ a quantity important of gn) upelllents hydroxides on its surface, it plays the role of a comparison electrode.
Thus, from two ec ~ ntillo ~ c of the same electrolyte poured onto a nickel sheet, one performs an oxidation sweep, and a reduction sweep, the domq-inP of stqhility éleeh ~ cl, it ~ ique of this electrolyte given by the gap between the two waves of currents.

We thus establish a domqine of electrochemical stability of about 1.6 volts, which is a priori compatible with an application to electrochromic systems for example, and close to the values for the other anhydrous proton electrolytes, as illustrated by the resllltqtc reg. ~ upés in the table 2.8. This result thus confirms the interest in molten salts of imi ~ lq.7ole in electrochemistry.
Table 2.8 High stability domain: truch "~ le des élé ~, t .., l ~ tes pruton, es anhvdres.
Electrolytes POE / H3Po4 Im / CF3SO3HUPOE POE / Sulfamide Aminosils Domain 2a 1.6 1.7b l ~ 2c from a: [Defendini], b: [De7P ~ Rermudez1], c: lRou-cse ~ u].

II-2-6 Conclusion on proton electrolytes:

Anhydrous polymeric protonic electrolytes formed by the mixture of a poly (oxide ethylene) of high mass, and a protonic molten salt obtained by acid doping of a base heterocyclic, in this case imidazole, allow to significantly improve the conductivity comparison of the electrolytes described so far, as illustrated in Figure 2.26 ~ ' ~ 1 û-6 ~ 'CF3S03H US = 2 O / H = 2 _ O / ~ p ~ / El POE / H3PO4 O / H = 2 ~ 10 7 ~ (O BPEl / H3PO4 N / H = 1.5 - ~ POE / Sulfamide O / S = 4 N / H = 5%
1 0 '~''~ ~ I,,,, I

Temperature (~ C) Figure 2.26: Comparison of the conductivity results for the p, i "-; ~, with the ele -I, olytes presented POE / H3PO4 from ~ Defendinil BPEI / H3PO4 from [Ravetl POE / Sulfamide from [DeZeaBermudez ~.
The con ~ uclivile obtained with the m ~ lqnge imi ~ lq-7olelacide ll; rli (read is probably not the oplilllulll to bass ~ IIlp ~ lalul ~, if we consider that we can qbqisser the ~ Illp ~ lalult; of fusion of its eutectic from 12 to # -11 ~ C as a substitute for triflic acid with bis (trifluoromethesulfonyl) imide.

Such electrolytes would therefore make it possible to operate electrochemical systems in films thinS, such as electrochromic glazing, with fast kinetics even at low temperature, ignoring the kinetics specific to electrochemical reactions producing at the electrode, which is not meaningless in view of the very thin thickness of the electrodes l ~ tili ~ es in this case. On the other hand, the conductivity coll ~ crgeallt at - 2.10-3 Q ~ l.cm ~ l at 60 ~ C
in almost all cases, this value is probably a limit linked to the transfer of proton of an imid ~ 7ole molecule protonated to a neutral imi ~ 7olç molecule, but this matters little, since a conductivity of 10-5 Q ~ l.cm ~ l is snffis ~ nte most of the time.

These electrolytes are, to our knowledge ~ ics ~ nce ~ the first example of the demonstration of the possibility of obtaining neutral anhydrous proton electrolytes; previous work referent indeed to acidic or basic electrolytes and the use of a heterocyclic base as imid ~ 7O1e for the synthesis of protonic molten salts at room temperature.

In addition, even if this study focused on molten salts based on imid ~ 7ole, tests pr ~ limin ~ ires show that the stoichiometric salt of the triazole protonated by HTFSI is liquid at room temperature, even after one year of storage in a glove box. So we can consider obtaining anhydrous protonic polymer electrolytes using triazole in su / as ~ itl1tion to imi ~ 7ole, and thus to obtain relatively acid electrolytes, the pKa of triazole for the protonation balance being ~ 3. This is an advantage for their possible application to uti1ic electrochromic systems: mt t ~ -n ~ ctèn ~ WO3 oxide whose films dissolve for p H> 5 In addition, it is also possible to exploit the many possibilities of organic chemistry and macromolecular for directly grafting heterocyclic groups and acid functions on the macromolecular chain. So we can consider obtaining polymer electrolytes self-doped anhydrous protonics no longer containing volatile species, with all the advantages inherent in such an approach in terms of f ~ brir ~ tion technology.

Let us mention eg ~ 1rm ~ nt the potential interest of these electrolytes for the realization of energy sources strong pnics ~ nr ~ es ~ for which the conductivity is an essential parameter, and the solid state a advantage for the f ~ brir ~ tion of these systems by technologies derived from thin film processes.

It is of course obvious that these materials are present ~ 1 ~ ment great interest fon ~ m ~ ntal, a share to understand the m ~ c ~ nicmes of conduction of the proton in these systems, and by the link of between chemistry and biological processes, and more particularly proteins, and their studies relative to the dom ~ ine of electrochemistry of protein electrolytes anhydrous.

Bibliography:

[Angell] PubO Nature [Armand1] Slngapour [Armand2] GFP
[Bertier ~
[Bie ~ gen] Biel ~~ gen, G. Electrochim. Acta 1992, 37, 1471-1478.
[Cob "ll, an] Chemistry of Solid State Materials 2, Proton Conductors, (P. Ca! A., ~ An, Ed.), Can, l, ridge University Press, Cambridge, 1992.
[Charbouillot] Thesis of the University of y, erLb '?, INPG, 1988.
[Dt: fen ', i] Defendini, F. Thèse de rUniversité de Grenoble, INPG, 1987.
[De Zea Bermudez1] De Zea Bermudez, V. Thesis of the University of Grenoble, 1992.
[De Zea Bermudez2] De Zea Bermudez, V .; Sanchez, JY Solid State lonics 1993, 61, 203-212.
[Dippel] Dippel, T. Solid State lonics 1993, 61, 41-46.
[Hadzi] Hadzi, DJ Mol. Struct 1988, 177.1.
[Kreuerl Kreuer, KDJ Mol. Struct. 1988, 177, 265.
[LasseguesJI ~ ssègues, JC in Cr,! C ~ Iban [Malarski] Al ~ ki, Z .; Sobczyk, LJ Mol. Struct. 1988, 177, 339.
[Po,;, s, ~ "on] P ~ i., Siyllon ~ C. Publi ILL
[Ravet] Ravet, N. Thèse de l'Université de Grenoble, INPG, 1994.
Ripochel Ripoche, X. ECS El ~ h ~ h, - ", ~ Materials ll [Rousseau] Rousseall F. Thesis of the University of Grenoble, INPG, 1990.
[St, ~ it ~ ias_,] Sb.iit-.ieser, Jr .; AT.;
[Trinquet] Trinquet, O. Thesis from the University of Bordeaux, 1990.

2! 941 27 _ II-3 CONDU ~ l ~; UKS ALKALINS:

II-3-1 ~ production:

As part of the development of high energy detec batteries, the electrolytes protonics with a limited electrochemical stability range (~ 1.5 Volts), elccLIochimictes se ni ni; ournes vers l'uiilisaiioil SU; V3 ~ orgLLIli4ues r ~ a ~ 3r ~ 1, and ~ lu ~ r ~ t; n polymlArmand ~ l, aprotic, which have an extended domain of stability which can exceed

4 Volts if the ionic vector is an alkaline salt. In polymer electrolytes, the polymer can party ~ m ~ nt be ~ ccimile to a stationary aprotic solvent, the solute being constituted by a salt of alkali metal, the m ~ c ~ ni.cmes depl ~ cemçnt ions on the microscopic scale are partly described according to the theories from the physico-chemistry of solutions. Also, it seems important to specify the mechanisms of solvation of ions in liquid medium, before tackling the problem of polymeric solvents, these notions will be useful to us in the rest of this presentation.

II-3-1-1 Salt-solvent interactions:

In a simplified way, an ideal electrolytic medium of an MX salt is obtained in a given solvent.
if it is able to dissolve the salt in the form of ion pairs, then to solvate the charges ionic in order to separate them, as described in figure 3.1.

MX lonisati ~ n _ [M ~ + X ~] + S solvation _M + S + X-, S
solvent: S
Ionic crystal _ Pair of ions _ Electrolyte Figure 3.1: Obtaining an elE ~ lyt medium; that ideal of an MX salt in a solvent S.
An ionic medium is thus obtained, in which the salt is dissociated, capable of transporting the ions within any electrochemical system (battery, electrolyser, fuel cell, ...). We then understand ~ icement that the properties of the electrolyte depend on both salt and nature solvent. According to Lewis theory, we define a base by the capacity of a chemical species to yield an electronic doublet (ie the anion X ~ in this case), and an acid like a species having empty orbitals capable of accepting a pair of electrons (ie here the cation M +); a acid-base reaction collGspond then to an ec ~ nge of doublets. Separation of the ion pair MX can only be observed if the salt is placed in the presence of a stronger base than X ~ (solvation of M +), or even in pr ~ sellce of an acid stronger than M + (solvation of X). n is therefore crucial to define the acid-base properties of solvents. To this end, the chemistry of solutions establishes a marked difference between protic and aprotic solvents, this is expressed by the concepts of donor number DN [GUtmann1 ~ 2l (linked to the tendency of a solvent to partially transfer doublets with positive charges) and of acceptor number AN IMaYer ~ G ~ l'r ~ 121 (expressing the tend ~ nce of a solvent to be accepted by ~ i ~ llPmPnt an electronic doublet of negative charges).

219 ~ 127 These parameters because ~ ct ~ ri ~ nt q ~ l ~ ntit ~ tively acid-basicity in the sense of Lewis of solvents, we we have indicated in table 3.1 the donor and acceptor numbers of some solvents usual [Unertl. The most common dipolar aprotic solvents such as acetonitrile (MeCN), ~ dimethylformamide (DMF), or Table 3.1 l ~. ~, L- ~ s donors and ~ -c ~ t, ~ urs for ql ~ PI ~ ues sol ~ & nt: j usual, according to [Linert]. dimethylsulfoxide (DMS0) may be more ANDN solvates as water to c ~ lions D ~ ~ 'rvd ~ l, dl ~ eCl - CH2 - CH2 - Cl1tj, 7 o (~~ so = ~ .o ~ .e ~ U = i ~
however, the lack of hydrogen bonds Water H ~ O ~ H 54,819 ~ 5 allowing load stabilization Methanol CH3 - O - H 41 5 191 negative (ANdmSo = 19.3 to compare with ANeaU = 54.8) implies that these. solvents Ethanol C2H5 - OH 37 ~ 9 19 interact only weakly with anions.
1 ~ "u ~ lllaneCH3 - NO2 20.5 27 It should be noted that given the role preponderant of the hydrogen bond in the Acetonitrile CH3 - C - N 18 9 141 anion-solvent interactions, it has been D ~ tl, flu ~ r CH3 - O - CH3 12.5 17 suggested [Bordwell11 to use the CH3 ~ terminology of hydroxylic solvents Dimethylfor ".a", ~ 11 H 16 26 6 CH3 O '(protlques), et non-hydroxyllques (aprotlques) CHy S - CH3 to differentiate them. This remark being DimethylsuHoxide ll 19.3 29.8 ~ done, we will use either of these terminologies. Subsequently, the introduction of the concept of hardness-softness ~ Pearson1,21 completed the notion of donor and acceptor number.
This criterion which reads makes it possible to differentiate the acids and bases Table 3.2 Typical examples of of Lewis, in fact, quantum mechanics shows that hard bases and soft bases.
reaction involving an acid and a base is favored when HARD BASES SOFT BASES
the orbits involved have spatial extensions CF3S03 1-comparable. Also, derivatives of light elements (lère and Cl04- SCN
2nd line of the periodic table) being characterized N03- BPh4-by low-volume orbitals, Cl- acids and bases drifters are not very polarizable, and are described as hard. Br-Heavier elements, on the other hand, have orbitals de exten.cion. ~ important spatial ~ i. ~ mel-t deformable, we speak of acids and soft bases.
Some bases ~ epr ~, sel t ~ tives of these two f ~ miles are given in table 3.2 ~ Chaba9no], It it is important to note that the soft bases are generally said to be hard bases by their less r ~ if ~ t ~ nces to oxidation.

If unlike protic solvents, aprotic solvents do not stabilize anions whose charge is loc ~ ée, they are on the other hand better solvents of bulky anions whose charge is very dis ~. ~ [E-- ~ '"11, the difference having its origin in the nature of the interactions put 21 ~ 4127 in play, as illustrated in Figure 3.2. Indeed, the etqblissoment of a hydrogen bond implies a partial overlap between the hydrogen orbital (for example of a hydroxyl group) and that of an electronegative atom; however, water being a hard solvent, the stabilization of an anion by bond hydrogen is favored when the charge is localized, i.e. in cases where the density of charge on the anion is maximum (for example F- for the halogen family). However, aprotic solvents stabilize anions through dipole interactions, this type ~ inL.; ï; lction is ~ energetically weaker than a hydrogen bond, but the number of dipoles likely to stabilize the load is inversely proportional to the surface density of the load, this mechanism is therefore in favor of anions whose charge is dispersed in a large volume.
Solvent solvent non-hydroxylic hydroxylic , R ~ ~
H, _ ~ _ "O ~ ~
~ electricality ~ charge charge on the anion Figure 3.2: Influence of the nature of the solvent on the solvation of anions.
All this explains why to obtain dissociated ionic media and of high concentration, electrochemistry of miliel) x non-aqueous uses anions whose charge is dispersed. Different phenomena can contribute to dispersing the charge of an anion, we will limit ourselves thereafter only for the stabilization of the anionic charges by attracting effect and by resonance. In in particular, if we are primarily interested in carboxylates, sulfonates, or perchlorates, we can consider that they R - C + R - C ~~
derive from the addition of an electrophilic radical on the oxanion o2- 0 ~
lC ~ ~ i ;, as illustrated in Figure 3.3. In each case, it is ~ G
possible to write mesomeric forms allowing to place the R + ~ ~ RO
charge sucks-ccively on all oxygen, thus contributing to OO
scatter it. The carboxylates only delocqliserlt the charge on o = a + o = c ~ o ~
two ol ~ ygenes, in the sulphonates in addition to the delocqlicq ~ tion oo on the three oxygen atoms, the interaction p ~ -d7 ~ between the Fdi ~ 9, cure, 3 ~ s35EdX, ea'p-rèss [cdhoexrakanoonl. ~ s doublet free of anionic charge and a vacuum orbital of the sulfur atom makes it possible to relocate the charge eg, ql ~ ment thereon, this is consistent with the lower bqsicity observed for sulphon ~ your than carboxylates. This effect alone is not s ~ fficq-nt to obtain soluble derivatives of sulfonates and carboxylates, and only the adjoncdon of an electron attracting group (R = CF3) rlr ~ inqnt the electrons, is likely to decrease the charge density on oxygen is sufficient to obtain lithium salts which are soluble in the medium aprotic. The low basicity of perchlorates is linked to the combination of deloc ~ lis-qtion of charge on the four oxygen atoms and with the attractor effect of chlorine, due to its electronegativity. The basicity of XO4- anions decreases with the electronegativity of the following halogen: Cl04 ~ <BrO4 ~ <IO4 ~. The main mesomeric forms allowing deloc ~ licçr Ia charge on these anions are shown in Figure 3.4.
~ 0 _ ~ R - C = O
O O ' ~ O, - ~ +
--O _ ~ R - = O ~ ~ R-- --O
O
O 0 ~ 0 1 ~ ao ~ _ ~ o = ll ~ - - ~ = b ~ - - - ~ - ~ = ~
Figure 3.4: Examples of mesomeric forms of delQc ~ ian oxanions In the case of bases derived from fluorinated coordination anions such as BF4-, AsF6-, or PF6-, the effect attractor of fluorine atoms combined with the fact that the free doublets on fluorine atoms are can donors, fluorine being a very hard base, makes it possible to obtain particularly soluble salts in an aprotic environment. The effect is further accentuated for BPh4- and B (CH3) 4-, by the absence of doublets free in the molecule, so a solution of the lithium salt of the latter in dioxolane allows to obtain an ionic conductivity of 1.3 10-2 Q ~ l.cm ~ l [Klemann].

All these criteria, even if they make it possible to define the parameters giving alkali metal salts, in particular of lithium, soluble in aprotic medium, are unfortunately only qualitative.
A first attempt to quantify the interactions between solvents and anions led to try to establish, in a similar way to the solvent (DNsolv) ~ a scale of own donor number to the anions l ~ 1; m ~ lhellrell $ emçn ~ the reliability of these values was questioned later.
Reliable determination of donor numbers of different anions in different solvents (DNX-, Solv) has only recently been established ~ ~ Unert]. It should be noted that the measurement of DNX- ~ solv being linked with the solvatochromic properties of the complex [Cu (acac) (tmen)] + (with acac = acetylacetone and tmen = tetrarnethyl ~ nedi ~ mine), and more precisely to the modification of the energy level of the orbital d of Cu2 +, the values obtained refer to interactions with a hard acid. For to overcome solubility problems, the measurements are made from the salts of anion tetrabutylammoniums. Correlation of whole e ~ p ~ data from spectroscopic, chemical equilibrium, kinetic data, or data measurements thermodynamics using anion donor numbers prove their reliability and value intrinsic to this concept. The donor numbers of different anions relative to the dichloroeth ~ are only given in Figure 3.5, as well as that of the triflate anion in various solvents.
It is eg ~ l ~ m ~ nt reported in this figure the numbers do ~ u .. ~. and acceptors of common solvents.

21 q41 27 DNx- ~ DcE DNsolv ~ DcE DNCF3SO3-, solv ANsolv --BPh4- - 0DCE_ _ B F4 - --EtOH
CIO4 ~ _10 _ Me2CO
MeCN _ -_ MeCN ~ DMS0 DMF
CF3SO3- _ Me2CO --_ DMF ~ _ DCE

NO3 --- _ 20EtOH Me2CO DMSO_MeCN

--CN- DMF_ CH 3C ~ 2 ~ --30DMSO _ --SCN-Br ~ -Cl ~
I Basesl IBases I EtOH_ ¦ hard ¦ ¦ soft ¦ --40 "
Figure 3.5: Non.l ~ res donors of various years; relative to dichloroethane.
Let’s look first at DNX- ~ dce ~ we can consider that it measures the trend of an anion to yield an electronic doublet, in other words its basicity in dichloroethane, solvent guidelines for donor number measurements. In accordance with what we have exposed previously, we can assume that the DNX- ~ dce depends both on the center carrying the load anion, the structure of the anion, and for a given family of anions, the charge density anionic. The DNX- ~ dce therefore gives a measure of basicity integrating all of these settings. In particular, the least donor anions are obtained with the derivatives boron tetrasubstituted (BF4- and BPh4-), which confirms the advantage of the ~ absence of free doublets, or the presence of low donor doublets in the molecule on its basicity. The effect is accentuated in BPh4- by the combined effect of the steric hindrance of the four phenyl groups and of the deloc ~ lic ~ ion of the charge they in ~ ent ~ this justifies that this anion is taken as a reference of the DNX- ~ dCe scale. It is particularly interesting to note the difference between the perchlorate and triflate, these two anions being gener ~ lemPnt considered as bases of a equivalent force. The DNX-, dCe for triflate and perchlorate therefore seem to infirm this hypothesis, it is to our knowledge ~ i ~ s ~ nce the only results indicating that the triflate is a base more strong as suggested by pKa measurements in an aqueous medium, the anion is even more donor that ~ star ~ itrile and just as much as dimethyl ether in dichloroethane. We can assume 2 1,941 27 that the symmetry of the molecule plays a role, the perchlorate anion being p ~ perfectly symmetrical, we ~ u ~ Jne-nte the volume in which the load can be located, in other words the configuration entropy of the anionic charge in the molecule. The value for the acetate may seem low, the measurement being made from ammonium salt, not tetrabutylammonium, the lesser solubility of this salt in dichloroethane or the existence of aggregates could explain this result.

D ~ s lGrs, if 1 in s;;; ~ ~ 1 inter ~ Lh ~ nir ~ 1 ~ s ~ l M, ~ cL 1 ~ solvent (S), if L ~ l ~ X ~, solv <D ~ solv the acid (M +) is solvated, and the salt fully dissociated, a favorable condition for obtaining a conductive electrolytic medium. On the other hand, if DNX- ~ solv> DNsolv the acid interacts prefer ~ ntiell ~ mPnt with anion, MX salt remains in the form of pairs of ions, the medium obtained ~ ~
being a solution of MX little or not x MX + S _ [M + X '] + S pay ~ dions conductr this for ions (we do not consider v not the case where the salt is insoluble in the o solvent). It is obvious that when ~ - MX + S ~ c ~; onsolvanVanion DNX ~, solv ~ DNsolv ~ the anion and the solvent are in x competition to bind with acid. These v concepts are summarized in Figure 3.6. These c, MX + S ~ ~ '+, S + X' I: ssoci ~ n considerations are confirmed by measurements of conductivity in DMSO [Linenl, solvent X, solv Figure 3.6: Influence of the donor number of the anion X ~
much appreciated due to the large number of su ~ / a fl; ~ soc; ~ i "n MX salt for a given solvent.
bodies that can be studied in this environment.
Thus, the anions which are less donor than DMSO are then completely dissociated, moreover, it is possible to correlate molar conductivity with DNX- ~ solv when it is greater than DNdm trad-lis ~ nt thus the gradual appearance of pairs of ions in the medium.
The donor number of an anion strongly depends on the solvent, so for the trinate the number donor goes from 19.2 in dimethyl ether to -4 in water, Figure 3.5 shows the DNX- ~ solv of triflate opposite DNSolv. We can consider that an anion whose donor number in a solvent is less than DNSo ~ v is not a donor with respect to a hard acid in this solvent, thus for triflate only DMF and DMSO are likely to meet this criterion for aprotic solvents. In the same way, a comparative scale is obtained for all the anions in dec ~ l ~ nt the DNX scale, Solv for the triflate of the difference between the donor numbers eyist ~ nt in dichloroeth ~ ne, thus Cl04-, BF4-, or BPh4- are not given with respect to a hard acid in the solvents indicated. One of the most interesting conclusions from the work cited on anions is to have shown that the DNX-, Solv is linearly dependent on the following A ~ olv:
DNX-, solv = 129.6 - 0.548 ANsol v - a The parameter a is a con ~ t ~ nte for a given anion linked to the measurement of the absorption in the visible from the dd band of the complex [Cu (acac) (tmen)] + in dichloroeth ~ ne The influence of ANSolv on DNX- ~ solv can easily get ~ l ~ lé ~ r, indeed, an anion will have as much less likely to interact, with a hard acid in this case, when its anionic charge is stabilized by the solvent, it is a partial dirc ~ nt Lransférée sur cclui ci. In particular, DNX- ~ solv in water is obtained in tliminll ~ nt the value determined in dichloroethane of 21 units. So, all the anions mentioned in figure 3.5 are not donors compared to a hard acid in water, the strong solvating power of water partially or completely leveling ~ the basicity of the anion.

The energy of transfer from one anion from one solvent to another, and more specifically from water to solvent aprotic, is a characteristic parameter of the solvation of ions. This value is all the more weak that the anion is easily solvated by the aprotic solvent, we can then consider that the anion is relatively indifferent to its surroundings ~ mPnt DNX- ~ solv have made it possible to correlate the e ~ nent ~ values of transfer of water to different solvents (in kcal / mole) according to the equation:
~ Gtrans = 0.35 + exp (0.092 ANsolv) The non-linear evolution of transfer energy is linked to the increasing interaction with DN
the donor and acceptor effect of the solvent induced by the presence of the anion, implying a addict ~ nt solvent-solvent interactions, and by the same a modification of the structure local solvent, this notion refers to that of activity coefficients in thermodynamics. It is interesting to mention that we usually associate the donor number with a measure enthalpy, the correlation of free enthalpy with DNX- ~ solv therefore implies that the term entropy, linked to the transfer of the anion, is negligible or evolves linearly. This assertion is criticizable ~ a modification of the structure of the solvent introduces the notion of order, and by the same an entropy term in the value of DNx- ~ solv.

The interdependence between the different parameters makes it possible to define the limit of validity of the concept of ANSolv and DNSolv ~ you can only use one of these parameters to perform correlations or forecasts, and this only in cases where the other may be corl ~ considered as a constant or negligible. These precautions being taken, the DNX- ~ solv is a tool of great value in the understanding of salt-solvent interactions, the measurements being reliable and easily e ~ t ~ -ncibles many anions.

21941 ~ 7 -II-3-1-2 Acidity measurements:

Unfortunately, the reliable determination of given numbers, since anions are recent, we do not at present have only DNX- ~ solv for a qllin ~ ine of anions. However, the determination of the pKa of an HX acid makes it possible to measure the basicity of the anion X ~, in other words the ten-lan ~ e of the anion to yield an electronic doublet to a hard acid, we can then reasonably assume that pKa is somehow related to DNX-, solV. The determination of the pKa of an acid There are many ways to do this, the first one usually referring to pH measurements in water using a glass electrode; indeed, according to the dissociation equilibrium of HA acid in water (valid only for dilute solutions), when [AH] = [A-] the pH is equal to PKa of the acid considered, this result is a direct application of the definitive equation ~ nt the pH in the water:

HA ~ Ka) A- + H + hence ~ K = [] [] and) pH = PKa -log ([A ~ l / [H +]) The pH measurements in water are however very limited ~ its relative acidity and its low power solvent for neutral organic compounds allows acidity determination only for -1% of known organic compounds. Many methods developed by the chemists organics to determine the pKa of various organic compounds exist; it is indeed possible to ~ qualitatively optimize the electrophilic power of a cation to its acidity in the Lewis sense, and the nucleophilic power of an anion at its basicity, this partially explaining the interest for these measurements on the part of the chemists All of these methods will not be detailed, only will be exposed those showing interest thereafter.

In order to overcome the problem posed by the use of water as a solvent for the determination of acid-base properties of organic molecules, many other solvents were considered for these measures such as ben7 ~ ne, or cyclohexylamine; even if these solvents have allowed to provide information on the pKa of certain families of organic compounds, the impossibility establish a reliable comparison according to the solvent and the nature of the substrate involved gr ~ ndemPnt the interest of these resultsc. The use of the acidity function H_ relative to the anions, valid for concentrated aqueous miliPux or mixtures thereof, theoretically allows establishing pKa measurements which can be referred to equivalent values in water. In this case, is more interested in the pH measurement obtained by a glass electrode, but in the evolution of the degree of protonation of the base induced by the presence of a stronger acid than its conjugated acid (in general sulfuric acid or perchloric acid), i.e. at equilibrium:

A- + H + ~ AH

"32 2 1 94 ~ 27 ~.
The acidity function H_ for the solvent considered is then defined by the equation:
H_ = pKa + log ~ A ~ l / [AH]
The pKa in this equation is relative to water. The degree of protonation is obtained by measuring the optical absorption specific to the anion A- as a function of the amount of acid added to a solution of HA acid. The value of H_ is determined ~ using the IH indicator of which 1 ~ pK ~ is known in water, by the detcrmhla; ion; iu degl ~ of prolonaLioll c ~ rlcspulld ~ uilible:
I- + H + ~ IH
We can thus dete-rmin ~ r, in theory, the pKa of any acid, but this supposes that it have the same behavior in the solvent as in ~ lic ~ te ~ lr; it was therefore necessary to establish the functions acidity for different f ~ mill ~ s of organic molecules (for example we know the functions acidity for ketones and aldehydes, for primary amines, for cyano acids, ...). he there is therefore no acidity function, for a given solvent, independent of the base considered, the structure of the anion being ~ etermin ~ nte in this. We can again relate this phenomenon to degree of delocalization of the anion charge, anions with a high charge density inter ~ gi.ssent strongly with water via li ~ icons hydrogen, local destructuring ~ mPnt the solvent, when the anion is protonated these interactions disappearing the water molecules reor ~ nicent; it results in an addiction to the entropy of the medium. The gain of entropy during protonation is obviously all the weaker as the anion is delocalized, this explains why the acidity function is not necessarily valid even for anions of equivalent structure, the charge density depending on the nature of the substances present on the molecule. A special example striking is provided by acetic acid and its chlorine derivatives, as illustrated in Table 3.3, by the part of the entropy in the acidity resulting from the reorg ~ nic ~ tion of the water molecules around the anion during its dissociation lMarch]
Table 3.3 Thermody data "d" ,, les for the ni ~ ni3ation of carboxylic acids in water at 25 ~ C.
Acid pKa ~ G (kcal / mole) ~ HT ~ S
CH3COOH 4.76 +6.5 -0.1 -6.6 CICH2COOH 2.86 +3.9 -1.1 -5.0 Cl3CCOOH 0.65 +0.9 +1.5 +0.6 What is at issue is therefore not only the structure of the anion, but also the degree of significant order of water induced by the self-association of molecules with each other, via hydrogen bonds, hence the advantage of using aprotic solvents. In this, the DMSO presents numerous advantages over other aprotic dipolar solvents, and in particular of being stable very strong base, very weak acid (pKa = 35.1), and a good solvent for many molecules organic (it is estimated that it solubilizes at least 50% of organic mol ~ s). DMSO is therefore a solvent of choice for developing an acidity scale.

The choice of an aprodic solvent calls for some details, in particular, we can write the balance ionic ~ tion of an HA acid in a solvent S by equilibrium:
HA + S ~ HS ++ A-This assumes that HA in this solvent is more acidic than the solvent itself. We then set the absolute con c ~ n ~ e of Kabs acidity by:

Kabs = [HS] [A l / tS] [HA] ~ pH = -log [HS] = PKabs - log (~ A] ~ [HA]) ~ PKabs = PKa + log [S]
We could thus build an absolute scale of pKabS, by correcting the pKa values of log ~ rithm of the solvent concentration. If we are only interested in the difference between pKa in DMSO and in water, we get ~ pKa = log (56/14) - 1.4. On the other hand, the balance of acid-basicity between the solvent and the HA acid depends on the acidity of the solvent, so we were able to determine a difference in acidity of water and DMSO in the DMSO of pKd ~ so-pKeau = 3s ~ l-3l ~ 4 = 3 ~ 7 According to these two hypotheses, we could predict a ~ pKa between water and DMSO of 5.1. The problem is m ~ lhe ... cuse ~ llent far from being as simple, we do not observe a difference conct ~ nte between pKa measured in water and in DMSO. More specifically, the acids whose charge ionic is localized are more acid in water than in DMSO, for example, for phenol on obdent a ~ pKa = 18.03 - 9.98 ~ 8, the very delocalized anions are on the other hand more acid in the DMSO than in water, so for picric acid the ~ pKa is ~ -1.5. Again, the gap is connected to the charge density of the anion, which facilitates or makes more difficult its solvation depending on the nature of the solvent.

The determination of pKa in non-hydroxylic solvents poses other problems pardculier. It is carried out by measuring the acidity of the molecule considered by a known pKa acid, The evolution of the equilibrium being determined by spectroscopy. To obtain precise measurements, it is preferable that the difference between the pKa of the acid studied and that of in ~ ic ~ te ~ lr is less than 2 units pKa, for this we currently have in-lic ~ t ~ urs pcrrnett ~ nt to determine values of PKa between 0 and 32, the pKa scale being anchored by the agreement for some acids of spectroscopic measurements with potentiometric measurements. In addition, in solvents aprodques, li ~ i.con.c intermolecular hydrogen being reinforced (for example between phenol and phenate anion), the acid dissociation balance and the balance of self-association of molecules for the determination of pKa ~ in order to eliminate the influence of self-association. This phenomenon has enabled the development of a methodology for th! , ~ er conct ~ ntes of associ ~ tiorls as well by li ~ i.conc hydrogen, as by the formation of ch ~ l ~ tes with cations 7 ~ lr ~ lin.c. DMSO being stable for only a few minutes in an acid medium strong, it does not allow m ~ lheu ~ eucG ... erlt not to establish an acidity scale integrating these.

21 9 ~ 1 27 For the moment we are interested in the influence of the solvent on the establishment of a scale of pKa ~ it is also possible to overcome the inflllence of solvation in deterrnin ~ nt balance of dissociation in phase ~ 7ellse We will not detail here the principle of the measurement, the setting work is more complex than in a liquid medium, we do not access the pKa of the acid but its free dissociadon enthalpy, using various direct methods: spectroscopy mass at high pressure, ion resonance spectroscopy, using a cyclotron ..., or indi ~ S ~ A, photoelectron spectroscopyCatàlanl. More specifically, both formal ga ~ use dissociation equilibria following:
HAj (g) + Aj- (g) ~ Aj- (g) + HAj (g) (a) HAj + (g) + Aj (g) ~ Aj (g) + AjH + (g) (b) The free enthalpy of equilibrium (a) measures the acidity of HAj (g) relative to that of HAj (g), that of equilibrium (b) the basicity of Ai (g) ~ G, ~ (g) +
relative to that of Aj (g). The value ArOH ~ H + ArO
free enthalpy standard collGsl) ond to ~ G ~ (ArOH) '~ Gt ~~ H +) ~ ~ Gt ~ (ArO ~) the hypothesis of a perfect gas with a fugacity +
of 1 atmosphere. We can refer the acidity ArOH ~ G ~ (S) H + ArO
in phase ~ 7P-uce, for example for Figure 3.7: Thermodynamic cycle of ionisa ~ i ~ n of phenol.
phenol, to pKa in solvent S from of the thermodynamic cycle given in figure 3.7.
The conn ~ iss ~ nce of free transfer enthalpies (~ Gt ~) between the solvent and the gas phase would allow the pKa of the acid in solvent S to be determined from its acidity in the ga7ruSe phase by the relation: ~ Gi ~ (S) = 2,302 RT pKa ~ the enthalpy of ionization deriving from the equation:
~ Gj ~ (g) - / ~ Gj ~ (S) = ~ Gt ~ (ArOH) - ~ Gt ~ (H +) - ~ Gt ~ (ArO ~) The transfer energy of the g phase is known for the proton; ~ 7P, nse to water (- 260.5 kcaVmole) and to DMSO (-265 kcaVmole). But the calculation of pKa also implies knowing the transfer energy of the anion and the acid in the g ~ 7ous phase, it again rests the problem of the effect of solvation on acidity, i.e. the results are always as dependent on the structure of the anion. However, many acidity or basicity values in ~ 7Puce phase are given by the dirrélGnce with respect to an in ~ tsnr, also even the conn ~ i ~ s ~ n ~ e of the acidity of this in ~ t ~ ollr in a solvent does not make it possible to obtain a reliable value of the pKa of the anion in this solvent. So, the free ionization enthalpy of meta or para-phenols in water, DMSO, or in phase g ~ l7J ~ U ~; e, shows that even if we can linearly connect the different values of a medium to the other, the effect of s ~ lbstitu ~ nt increases as the anion is less solvated. So for the phenol substi ~ Gi ~ decreases with a slope of -2.2 in water, -5.3 in DMSO, and -12.3 in phase g; ~ 7 ~ uce ~ according to the coefficient of E ~ mmett ~ of the substitute ~ nt (Cf II-3-2-1), as this is illustrated in Figure 3.8. The effect of the substituent on the acidity is thus glob ~ lPment 2.4 21 94 ~

times greater in DMSO relative to water, and ~ 6 times greater in gas phase than in water. Let’s look at m ~ int ~ n ~ nt the relative importance of enthalpy and entropy in the value of the free enthalpy, and this from the information available in figure 3.9.
As we can conct ~ entropy control - 80% of the dissociation energy of phenol in aqueous solution, this result again being linked to the destructuring of water during the ionic ~ tion of phenol; this fact is confirmed by the negligible part of the entropy in the energy of dissociation of phenol in D ~ SO, this s ~ lv ~ nL ~ Lanl of a less structured p ~ r ~ and on the other hand a bad solvent for hard bases like phenol.

5,,,, 2 ¦ Phend ¦ ¦Phenol ¦
~ _ O ~ ~ PeaU = 2 ~ 2 ~ 1 ~
"5 ~ _ ~ () e ~ u '3 ~ J j ~

aGE ~ u ~ Pd 5.3 I, ~ / = DMS0 '~ dmso Z -gas p --12.3 ~ G) dmo 0 ~ 96 -1 5 '''' -3 -0.2 0 0.2 0.4 0.6 0.8 ~ 4 -3 -2 -1 0 1 2 ~ 5 HQ ~ (kcal / mole) Figure 3.8: Influence of sllhct; tl ~ A ~ 7t on ~ G Figure 3.9: Fraction of enthalpy in energy ion; sdti ~ n phenol and this in various media. of ionisdti ~ n of phenol in water and in DMSO.
All this shows us the difficulty of comparing the anion acidity from one medium to another, and even in the same medium, both the structure of the anion and the nature of the solvent play a role in the balance ionization. However, there is a qualitative rule to remember: an anion whose charge density is important is more acidic in water than in an aprotic solvent, an anion whose charge is dispersed is more acidic in an aprotic solvent than in water. Information we have on the referent anions m ~ lh ~ urellsçm ~ nt just as well to spectroscopic measurements or potentiometric, in aqueous or partially aqueous medium, in DMSO, or in phase ga ~ use. As we will see later, this explains the difference obtained between the different values of the literature, and the caution necessary in their uses. The problems we have just exposed for the determination of the acidity of an anion only reinforce the interest that we are concerned with the concept of anion donor numbers and the methodology for accessing to this parameter. Since the measurement is made from the tetrabutylammonium salt, it is possible to measure the DNX ~ V for both very strong acids and for strong bases, ie on at least 30 pKa units. This concept, if it is extended to other anions, will be cert ~ inem ~ n ~ a tool of great value for the compr ~ h ~ nsion of salt-solvent interactions.

-II-3-1-3 Interac ~ ons salts-polymers:

If we are now interested in polymers [A ~ and2-51, the presence in the chain hetero ~ tomes (Y, N, P, S ...) having free electronic doublets, allows interactions with cations, their conferring thus solvating properties. It was known that the oligomers of the glymes family (n <20) solvate ~ i ~ ment the cations, CH3O (CH2 ~ H2 ~) nCH3 1 ~ flexibility of the molecule allowing the atoms ~ O, ~ O ~
GLYMES of oxygen to orient themselves so as to complex them by ~) the interm ~ re of the chelate thus formed. Indeed, like the CH ~ LATE
shows the figure opposite, we thus obtain a five-atom ring known for promote the formation of stable complexes. This mechanism extends to polymers of high m ~ cses, in particular poly (ethylene oxide), or POE, of monomeric unit (CH2-CH2-O); of similar to the glymes, the flexibility of the chain allows it to adopt a configuration in propeller favorable to the solvation of cations. According to this hypothesis, the electronic doublets of oxygen atoms facing the inside of the propeller are placed in a structure close to that of crown ethers, the solvation of cations by POE is thus favored by the entropy of resulting configuration. This parameter clearly differentiates POE from aprotic solvents dipolar, and explains its exceptional solvating power with respect to cations, and lithium in by ~ iculier. We can estimate, from the values obtained on the glymes, that the donor number of the POE is at least 22, this value being probably higher due to the proximity imposed solvating patterns; the ANpoE can be estimated at ~ 10. We could also define an apparent donor number (DNapp) of the POE integrating the entropy factor, knowing that solubilizes lithium chloride to give very poorly conductive electrolytes, we can assume that the DNapp is close to the donor number of Cl- in a solvent of AN ~ 10 ILiner ~], we We therefore arrive at an estimate of the DNapp close to 40. It is possible to ~ csimile, from a point of formal view, poly (ethylene oxide) to a very viscous non-hydroxylic solvent having a degree of order important at the microscopic level (where in this sense it would be closer to water).
Figure 3.10 illustrates sclléma ~ uement the difference between the modes of conduction of ions respectively in liquid medium, and in polymeric medium [Almal ~ d31.

S ~ lu ~ ions lig ~ la ~. Poly. "~, ~ S ~ is ... es lr nS, ~ G. l of solvate ions ~ Ss of solvatdtlon / desolv '- ~ ion Figure 3.10: Comparison of ", éoan; _." Es of conduction between cls ~ b ~ your liquids and él ~ t ~ ul ~ tcs po / ~ ", er ~ s according to [Arrnand3 ~.

21 9 ~ 1 27 Thus, con ~ e-llent to liquids, for which the mechanism of ion transport ~ involves the displ ~ cem ~ nt solvated ions by solvent molecules, cations propagate in passing from one propeller to the other following a m ~ C ~ nism of solvation / desolvation in the POE, the anions located outside the chain, hence the qu ~ lifi ~ tif of stationary solvent.

Before det ~ iller further characteristics specific to polymer electrolytes, POE being a aprotic solvent, during prelni ~ res studies on baL ~ erics ii eleclloly ~ es pulymeles, it was considered using hard localized anions such as CF3SO3-, Cl04-, or coordination anions fluorinated as BF4-, for the synthesis of lithium salts. Unfortunately, these anions exhibit major drawbacks for this application:

~ The fluorinated coordination anions result from the action of the base F- on the Lewis acid AFn, thus BF4- corresponds to the acid-base balance: BF3 + F- ~ BF4-. Because of the pol ~ ris ~ nt power of Li + and the significant reticular energy of LiF, the equilibrium may have been shifted to the left, results in degradation of the electrolyte, linked to the cutting of the POE chains by BF3 lchabasno ~ Robitailbl ~ PF6- and AsF6- have similar behavior. In addition, the presence of arsenic in the AsF6 anion poses problems as to its tOXiCitf ~.

~ The chemical instability of perchlorate involves risks of unpredictable explosions during its handling. It is difficult to im ~ iner the handling of large quantities in a industrial process, recent unsuccessful trials on this have only confirmed what was so known for a long time [Picotl.

~ In batteries with liquid electrolytes, the reactivity of the fluors of the CF3SO3 anion with respect to metallic lithium limits its use to primary systems. In a polymer environment, the various reaction products on lithium create a thin layer of a solid electrolyte [Bélangerl stabilized mechanically by the polymer, limiting the reduction of anions by avoiding their contact with the lithium metal. However, as we will specify later, the phase diagram unfavorable lithium triflate involves working ~ mlont batteries at a temperature high [~~ '~ 21 (~ 130 ~ C) to obtain a pui ~ s ~ nce slJffls ~ nte, leading in particular to a lower chemical stability of the POE with respect to lithium.

The anions originally used for lithium batteries with electrolyte F ~ E3 ~ F
liquid is not conducive to the development of batteries with electrolyte FF, C ~ F
polymer, it quickly appeared the need to find anions specific to these AN1OON OF TFSI
latest systems. This led to the use of bis (trifluoro-méth ~ neslllfo ~ imi ~ le) of lithium (LiTFSI) whose formula we give above, and whose interest has been confined for more than ten years as part of the development of ACEP bats.

. .
Compared to the anions previously considered, the TFSI has different advantages, and between others: to be non-explosive, to contain only covalent bonds, gar ~ ntiCs ~ n ~ thus its stability in the presence of POE, and as we will specify, to improve the properties of electrolytes.

Independent of the problems posed in terms of application, if one is primarily interested in physical properties of POE, lc polymi ~ re, derived from polym ~ ris ~ lion of oxide of ~ 1ylene, present ~
a regular and symmetrical structure which gives it a semi-crict character ~ llin Properties POE thermals stabilize from a molar mass> 104 g / mole (i.e. -230 units monomers), it is of course obvious that this value is greater than the entanglement threshold of POE chains (~ 3.103 g / mole). Under these conditions, the POE melting temperature stabilizes at ~ 65 ~ C (for T> Tf the POE is amorphous), below this temperature the polymer is composed of a mixture of a crystalline phase (in the form of spherulites) and an amorphous phase, the high rate of entanglement preventing all chains of crict ~ llicer. At a temperature of 25 ~ C, the crystalline phase occupies ~ 60% of the volume of the polymer, which leaves a part not negligible amorphous phase. In the presence of a solute, the POE tends to form complexes cri.ct ~ lli.c ~, c (OR solvates) and defined stoichiometry, different stoichiometry are known for complexes with light alkins (Li, Na), essentially: 6: 1, 3: 1, 2: 1, 1: 1, complexes 4: 1 and 8: 1 are more common with heavy alkalis (K, Cs). Just like a mixture solvent-solute, the physical state of the POE with its different solvates is described by a diagram of phase, the liquidus-solidus distinction specific to a liquid medium is made here between crist: -llines and the amorphous phase, considered as a liquid from the thermodynarnic point of view. We observe a once again a similarity with liquid media, the phase diagram of the POE / NaI mixture being similar to that obtained with the H20 / NaI mixture, the two diagrams being simply offset by approximately 70 ~ C, which corresponds approximately to the melting difference / ~ Tf = 65 ~ C
between the two solvents. We present on ~ lgure 3.11, the POE-salt phase diagrams for the following lithium salts: LiCF3S03, LiC104, LiTFSI [Valléel, Annand3] and LiASF6 [Robitaille]
250 64 32 16 10 6 3 2 1 EO, 'Li 300 32 16 ~ 4 4 3 2 1 EOIU

200 PEO-LiN (SO2CF3) 2 i 250 PEOLiCI04 f 150-:
~, 100 -, ~ _ 200 t_150- "
f ~
- ~, ~ Tg 50 ~
.00.2 0.4 0.6 0.8 0.0 0. 2 0. ~ 0.6 0. 8 Ponderale in salt F ~ ponderale in salt 2 1,941 27 64 ~ 16 8 4 3 2 1 EO / Li 350 ~ 200 300. PEO LiCF3S03 PE0-LiAsF6 250 ~ 150 -": ~
t 200 : ~ 100-150 ~ E

"""
0. 0 0.2 0.4 0. 6 C. 8 0 0.2 0.4 0.6 0.8 rr ~ 1ion porderaJe in salt Weight fraction in salt Figure 3.11: Didylailulle phase of mixtures of LiCF3SO3, LiTFSI, LiC104 with PEG3400, after ~ Armand3] and Li4sF6 with POE 5M, after [J ~ ~. 'lel.
In all cases, between the pure solvent (POE) and the stoichiometric salt concentration solvate the weakest, systematically exists a eutectic whose melting occurs at a temperature lower than that of the two cot-ctit ~ l ~ nts The phase diagram with the triflate differs from three others by the absence of the 6: 1 composition lower melting point complex, this shifts the eutectic composition towards low salt concentrations, it is generally accepted that the non-existence of this solvate is linked to the asymmetry of the anion, it would then be only about the influence of a geometric parameter on the phase diagram. The reverse rule is not always verified, despite the symmetry of BF4- the phase diagram of the POE-LiBF4 mixture does not has no solvate of composition 6: 1 but 3: 1 ~ When this exists the temperature of fusion of the solvate of composition 6: 1 strongly depends on the following anion:
Tf (CF3SO2kN ~ <Tf Cl04- <Tf AsF6-The melting temperature thus goes from 136 ~ C for AsF6- to z 50 ~ C with the TFSI anion, i.e. a gain of about 85 ~ C The decrease in the melting point between perchlorate and hexafluoroar.sé ~ t ~.
can be explained by the structure in which cri.st ~ 11i.ce the anion, the hexafluoroarsenate by its octahedral geometry, cri.ct ~ llise in a more compact structure, and therefore of higher energy reticular than perchlorate. The large volume of the TFSI anion and the existence of interactions orbital "7 ~" dp in the SNS bond (similar to the Si- ~ Si structure in siloxanes), indlli.c ~ nt significant flexibility of the molecule, may explain the low melting point of complex 6: 1. In addition, the difference between the temp. ~ 6: 1 melting lule of the complex for perchlorate and for TFSI ~ Tf ~ 15 ~ C lV ~ Iée11 ~ is similar to the difference in the melting temperature of the two salts of lithium ~ Tf = 251 - 234 = 17 ~ ClVallée ~ l, which tends ~ col-fi. ". ~ .r the inflllçnce of the structure crist ~ llinP.
lithium salt on the melting point of the complex.

~ - 21 941 27 Knowledge of the phase diagram is essential for understanding the results of conductivity, in fact, only the amorphous phase allows appreciable mobility of the cations.
~; ssçnti ~ m ~ nt two equations represent the evolution of conductivity with temperature:
Arrhenius law and the so-called free volume law. Arrhenius' law usually makes it possible to account of the evolution of con ~ uctivity ~ as a function of t ~ m ~ lalulc ~ in a liquid or in a ionic solid (glass, ~ AgI ...), where it is assumed that the m ~ C ~ nism of displ ~ cem ~ nt causes the sau ~ active of a barbar of polcnLiel by the ions, it is expressed by the equaLion:
~ = ~, exp (-Ea / RT) In this equation Ea represent the activation energy. It also applies to the conductivity of salt-POE mixtures below their melting point, and reflects the evolution of the conductivity in the amorphous phase in which the crystalline phase gradually dissolves. Energy of high activation is then connected to the free enthalpy of fusion of the crystallites in the phase amorphous. Above the melting point determined by the phase diagram, even if the electrolytes are composed of a homogeneous arnorphic phase, two cases occur according to the anion:

~ Lithium salts derived from a symmetrical anion and give ~ 6: 1 low point complex fusion, like LiTFSI or LiCl04 follow a free volume type conductivity law, one of which expressions is given by the Vogel-Tamman-Fulcher equation (VTF):
c ~ = A / ~ IT exp (-Ea / (T-TO)) This equation assumes that the diffusion of the species depends on the disordered local movements in the middle, it involves three parameters (while Arrhenius' law only depends on two): the term To represents the ideal glass transition temperature, the term pre-exponential A is proportional to the concentration of free charge carriers, the term Ea is a term of activation pseudo-energy. This equation generally describes the behavior of electrolytes in an amorphous state, like the POE-salt mixture above Tf, OR the mixture of a salt with a electrolyte non-crict ~ llic ~ ble, such as poly (propylene oxide) or POP. It should be noted that the kinetics of recrystallization ~ llic ~ tion of the POE-LiTFSI mixture, for the eutectic composition (several days), being slow on the time scale of the conductivity measurements, the evolution of the conductivity follows a law of the free volume type even at low temperature.

~ Dorln lithium salts ~ nt a complex 6: 1 or 3: 1 p ~ se ~ nt a high melting point, as LiAsF6, LiBF4 or LiCF3SO3 (Tf comple% e> l3ooc) ~ or in other words with high reticular energy, still obeys a law of Arrhenius but with less activation energy. We have summarized these notions in figure 3.12.

~ 1,941 27 Tf ~ Tf ~~
~ _ ~ _ ~ _ ooo ex: POE / LiAsF6 ex: POE / LiCI04 ex: PPO / LiTFSI
1 / T 1 / T l / T
Arrhenius T ~ Tf A "l enins Free volume Figure 3.12: Typical law of evolution of conductivity as a function of temperature.
The ideal glass transition temperature of the VTF law defines that where the excess of entropy of configuration of the polymer is canceled, this value can be obtained by extrapolation from the conductivity measurements. In practice, one rather reaches the glass transition temperature Tv by calorimetric or spectroscopic measurements like NMR. For electrolytes polymers, we can estimate as a first approximation that TV-To - 50 ~ C, the difference is explained by the fact that To is a thermodynamic parameter specific to the polymer, whereas Tv is a parameter kin ~ tique depends ~ nt on the system observation time scale (~ 100 s in DSC to 10-6 s in NMR). When a salt dissolves in the polymer, the concomitant introduction of strong electrostatic interactions induces conformation restrictions on the chains which result by an increase in the glass transition temperature. Table 3.4 shows the influence of the anion on the variation of Tv with the concentration of lithium salt LiX [A ~ and3].
Table 3.4 Variation in the glass transition of the POE with the lithium salt content.
Litl salt, iul .. LiBPh4 * LiSCN LiCl04 CF3SO3Li LiTFSI
aTV / ~ Csel (~ C / mo%) 6.7 3.8 4.2 2.8 2.8 ~ from [Vallée2].
The nature of the anion notably influences the variation of the glass transition temperature, according to these res ~ llt ~ tS, it is minim ~ for anions incorporating the triflyl group (CF3SO2).
It is likely that the effect comes mainly from the presence of the CF3 group, its very hydrophobic character reducing the interactions between the anion and the hydrophilic polymer. These measurements are effe ~ killed by rapidly soaking (~ -300 ~ C / min) a molten sample of electrolyte, this is possible up to 40% by mole of LiTFSI in the electrolyte, on the other hand with the triflate of lithium from ~ 15% by mole, the electrolyte recri.ct ~ llise immediately ~ t The flexibility of the anion TFSI gives it pro ~ t ~ s pl ~ stifi ~ ntes on the chains of POE which reduce net ~. ~ slow rewriting speed ~ t ~ qtion of electrolytes.

2 ~ 941Z7 Even if the POE is the polyether pr ~ sçnt ~ nt ~ tllçll ~ m ~ nt the best solvating power with respect to cations, it has the disadvantage of being semi-crict ~ llin To overcome this problem, derivatives POE non-crist ~ llic ~ bles have been developed. One of the ways to limit, or eliminate the crict ~ llinit ~ of the POE is to break the regularity of the chain by the introduction of point defects, e ~ çh ~ nt thus the chains to fold back to form cri.ct ~ llit ~ s We will limit ourselves to two examples of such polymers. One of the possibilities is to copolymerize in a reactor, by catalysis coordinated type Vandenberg [Alloin1] OR by polymerization anioni ~ read [Allcin1], the oxid ~ of Eillylella with epoxies carrying a lateral go-lpe ~ nt. We also introduce allyl-glycidyl ether units allow ~ nt to improve the mechanical strength of the material by crosslinking. More recently the extension of the reaction of Willi ~ mcon of formation of ethers to the synthesis of polyethers [800th, Alloin1,2l, a made it possible to avoid the manipulation of toxic gaseous monomers and thus to simplify their synthesis.
The most interesting synthesis process uses the conduction of poly (ethylene glycol) and a dichlorinated derivative carrying a crosslinkable function [Alloin1 ~ 2] ~ the reaction is carried out in the absence of solvent by mixing the two components at room temperature in the presence of a strong base like KOH, followed by steps of purifi ~ ation of the polymer.

To conclude, let’s point out that there is a potential interest for electrolytes with grafted anions, which we expect significant improvements on the cyclability and power performance of INeWrnan ~ Annand5l polymer electrolyte batteries. Indeed, such an electrolyte eliminates the gradients of anion and cation concentrations in the polymer electrolyte, and should result, in particular, a reduction in the degradation of the various interfaces present in the b ~ heads with polymer electrolytes.

II-3-2 Ar ~ ions ~. L:

Il-3-2-1 General presentation:

Fan ~ ille of TFSVTFSM: At present, LiTFSI allows to obtain polymer electrolytes having the best performance both from the point of view of conductivity, stability electrochemical, or the crict ~ lineage of the polymer (or its amorphous nature for polymers non-crict ~ llic ~ bles). Also, it is important to specify the physicochemical properties of the salts of the TFSI family, the data specific to it being m ~ lheureuseluent reduced we we rather refer to the information available on the salts of the family of bis (trifluorom ~~ h ~ n ~ sll1fonylméth ~ ne) common ~ m ~ -n ~ symbolized by the initi ~ TFSM, both salts present significant similarities of sl ~ u. ~. Indeed, we can consider from a point of formal view that the anions of the family of imides or m ~ th ~ nes ~ icubstitués r ~ sulten ~ de the ~ edition of an elecllophilic radical respectively on the nucleophilic radical N3- or HC3-, such as shows figure 3.14, and in a similar way to the family of anions deriving from the 02- oxanion.

e N - C - C - C9N N_C ~ --C_N
N_C + H
C3- ee CF3 - HERE + + CF3 - C - C - C - CF3 CF3 - C - N - C - CF3 o NOHO ~ ~
+ e G ~ 3 G

- ~ H ~ ~ ~ 1 Tamil Anion ~ Family Anions TFSM of TFSI
Figure 3.14: Anions of the family of imides and mé ~ hanes ~ / is ~ hstit ~ Jé.s.
Unlike this last f ~ thousand, the anionic center carrying the charge does not derive from a non-polarizable atom "hard" like oxygen but polarizable atoms "soft" (carbon or nitrogen), it follows that the anions are better solvated by the dipolar aprotic solvents since the charge is on the one hand easier to disperse, and on the other hand delocalized by two radicals éle ~ tl ~ philes instead of one. As proof of this concept, it is possible to cite the example of anions derived from the meth ~ nes-llfonyl radical (R = CH3SO2), the bis (meth ~ ne ~ ulfonimide) of lithium is soluble in POE unlike the lithium salt of methanesulfonic acid [Cherkaoui]. Of plus, the lithium salt of TFSI is very soluble in most organic solvents, we observe a non-negligible solubility of 63% by weight of dry extract, even in 13M ether], despite the low acceptor number of this solvent (AN = 12.5).
It can be noted that for the anions of the TFSM family, we thus find the common anions in organic chemistry that is the anion of malononitrile (of formula (CN) 2CH2), and the anion of he ~ fl ~ oroacetylacetone (of formula (CF3CO) 2CH2). Therefore, we can be interested in the study of these anions following a "horizontal" reading of this figure, ie to the influence of the central atom on the properties of anions, or on a "vertical" reading allowing to specify the influence of groupem ~ nt electrophilic on the anions. Let’s first look at this last possibility, in the particular case of the anions of the TFSM family, a complete study of the pKa of their acid conjugate and their electrochemical stability being available in the literature.

We have indicated in Table 3.5 the pKa of different acids of the methan family .lisukstitl. ~ s (of generic formula R2CH2) and the oxidation potential of anions deriving from it, in DMSO [~ l, these latter values given with respect to the normal hydrogen electrode in the cited work are ~ shell ~ es here of 3 Volts in order to reference them to the lithium electrode (the oxidation potential of the TFSM anion being greater than the oxidation potential of DMSO, this is d ~ termin ~ in acétQnitrile).

~ 21q4127 Table 3.5 pKa and p-, uAy lerltiel: l ~ tion of years from the TFSM family in DMSO (except ~ CH3CN).
R SO2CF3 CN SO2Ph COCH3 COPh SO2CH3 CO2Et pKa 2.1 11 12.2 13.3 13.4 15 16.4 Eox (A ~) ** 5.35 * 3.81 4.09 3.76 3.78 4.08 3.73 ~~ po ~ oxidation enliel vs. Li + / Li ~.
The inucucibility of a substance: on the physico-chemical properties of a molecule, and, by ~ i ~ ulier ... . ..... . . ..
on the ~ cldl ~ peuL ~ measure p; lr ul ~ l ~ ren ~ s coelllelenls a ~ ~ unm ~ tt c ~ OUS ~ lons ~ n ~ vement recall the meaning of these coefficients from the example of benzoic acid and its derivatives, as illustrated in the figure opposite, which is at the origin of the first works concerning them [March].
The study of the ionization of benzoic acid in water, substituted in meta or para, allowed to formulate the equation of ~ mmett that we XP ~ -8 OH
we have eyrrim ~ here in terms of pKa such that ~ pKa = p ~ compared to ~ /
non-substib benzoic acid ~ Constant p measures susceptibility xm benzoic acid with electrical effects induced by the substituents ~ nts pKa (Xj) = pKa (Xj = H) - P ~
by resonance or by inductive effect, in order to fix a reference on HAMMETT EQUATION
takes p = 1 in this specific case. The constant ~ characterizes the groupemP-nt Xj, there is in fact the co ~ .ct ~ nte <5m for groups substituted in meta and ~ p for those sl ~ bstitués in para. As a first approximation, we can consider that <~ m measures the effect induced by induction (attractor if ~ m> 0 and donor if ~ m <0), and that ~ p measures the cumulative effect of the resor ~ nce and the inductive effect on the ionization of the acid. The v31eurs of ~ being thus defined for different groupings, the value of p is determined for a given reaction using the reaction constants for at least two substituents (usually four), thus allowing to extrapolate to other substituents. Harnmett's equation has proven to be of great value in organic chemistry to establish correlations for many reactions and groups fon ~ tiorln ~ lc In the case of the anions which interest us more particularly, the dispersing groups the anionic charge by attracting effect and by resonance, it is more judicious to use the values of ~ p ~ developed to take into account the possible interaction of a substituent with a charge anionic appearing in a transient state during a chemical reaction [Marchl They do not tell us not seem inconsistent to use the ~ p ~ rather than the Taft parameters determined for the non-aromatic molecules [Mask 11], the anions of the TFSM family described being very delocalized and poss ~, of ~ t 4n + 2 electrons ~ T.

Table 3.6 C ~ ff :: - ent of I lal "" ~ tl ~ sp ~ of different attracting radicals.
R SO2CF3 CN COCH3 CO2And ~ p- 1.36 1 0.87 0.68 Reference [Fisher] [March] [March] [March]

2 ~ 941 27 -The results ~ tc indicated on figure 3.15 allow to correlate the pKa with the ~ p ~, for the groupçlnf! nt ~ whose value is given in Table 3.6, and thus to determine a value of p = 21.3 according to the equation:
pKa = 31.5-21.3csp ~
This v ~ is particularly high, firstly compared to that obtained for the acid benzoic in DMSO [~ ordwell3] (p - 2.4), but more generally with respect to phenols or derived from phenylmalononitrile in DMSO [E ~ 1 for which 4 <p <6, it reflects the ease with which the anions of the TFSM family disperse the anionic charge. Since it there is no general rule allowing to add the contributions received ~ COOE ~ I ~ II 1 of two substituents on a substrate 15 - \ IR CH ¦ ~
given IMa ~ hl, we can criticize the value of ~ 1 2 ~ h3 ~ 1 2 2 p obtained referring to a double ~ \ ~ CN
substitlltio ~ Cepen ~ n ~ the symmetry of c 9 ~
molecule and the identity of the two ~ 6-substituents, suggest that the c ~ \
value of this parameter characterizes 3 CF3 well charge dispersion by olllllll substituents within the meaning of the equation of o, 6o, 7 o, 8 0, 9 1 1, 11, 2 1, 3 1, 4 Hammett. However, it is not possible P
Figure 3.15: Influence of the attractor group on lePKa to establish a linear correlation between the delocalized anions ~ derived from disubstituted methanes.
oxidation potential of the anion and values of <7p- or pKa, in this case, due to the antagonistic effects of the groupings attractors. These can disperse the anionic charge, helping to decrease the basicity of the anion, but eg ~ lPment stabilize the radical formed following the oxidation of the anion, and thus decrease the stability of the carbon-hydrogen bond, as illustrated in Figure 3.16 with the example of ~ -pyridinium acetonitrile.

~ \ N + E3 ~ N Oxidation> Q ~ C-- ~ N ~ S ~ ~ '~' icn ~ ~ N + C ~ N
~ / H \ ~ H by resonance ~ H
- Figure 3.16: Effect of stabilization of radicals on the oxidation of o ~ -pyridinium acetonitrile.
Indeed, the conn ~ ics ~ nce of the pKa of the acid and the oxidation potential of the anion, in DMSO, allows to calculate the dissociation energy of the link (EDL) CH l30rdwel ~], according to the relation:
EDL (kcal / mole) = 1.37 pKa + 23.06 Eox (A-) + 56 This relation is in fact resulting from the coupling of two equilibria corresponding to the rupture homolytic of the CH bond of HA acid, i.e. the dissociation of the acid followed by the oxidation of the anion. If we co ~ si-lère the anions of the TFSM family as r ~ s ~ nt of the substitution - 2 1,941 27 s ~ lccessives of two of the hydrogen atoms of m ~ th ~ ne by an acceptor group, we can m ~ inten ~ nt if ~, sser to the infll ~ence of these s ~ lbstitutions on dissociation energy. We have lCgl'UU ~) e for this purpose in Table 3.7, the energies of ~ soci ~ tion ~ for the corresponding acids one and two substitutions respectively by the same attractor group X. Methane p ~ u ~ t an energy of ~ issoci ~ tion of 105 kcaVmole, we can obse. ~ er a ~ ul; on of this energy under the effect of the s ~ lbstitution of one of the hydrogen atoms. This difference actually corresponds to the energy of s ~ mobilization of the radic ~ (~ SR) dc XCH2 ~ rcla ~ only to CH3-.
Table 3.7 ~ l ~ energy, so ~ on of mono and dis derivatives ~ ~ hsti ~ llés methane (in kcaUmole).

X EDL ESR ¦ BDE ~ ESR ~ (ESR) CN 93 +12 90 +3 +15 SO2CH3 99 + 6 100 - 1 +5 SO2Ph 99 + 6 98 + 1 +7 SO2CF3 103 +2 113 -10 -8 CO2And 95 + 10 95 0 + 10 COCH3 94 + 11 92 + 2 + 13 COPh 93 + 12 92 -1 + 11 NO2 98 +7 The results for the first subst ~ iQn show that the groups capable of delocalizing the free electrons (CN, CH3CO, PhCO) weaken the CH bond by 10-12 kcaVmole, on the other hand groupemerlt ~ sulphonyles attractors essentially by indllctive effect little stabilize the radicals, from 2-6 kcaVmole. However, it should be noted that the value of this stabilization energy is not a constant specific to each group, but depends on the molecule 0, 6 ~~ ~ F
on which the substitution takes place. 0.5- \ 3 IPhCH2so2G¦
So, if the group PhSO2 in I o 4_ the species PhSO2CH3 stabilizes the radical of g 0 3

6 kcal / mole, the same grouping ~ ~ OPh substituting an acidic proton in the ~~. 2 -toluene, xylene, B-methylnaph ~ ne, O 0.1 - ~ H2P ~ -NMePh 9-methylanthracene, or fluorine O _ -Bu (t) ~ -Me dést ~ kili ~ e the corresponding radical from 2 to - 0.1 lllllll

7 kcaVmole. However, with the exception of 12 14 16 18 2 0 2 2 2 4 2 6 2 8 glou ~ e .. ~ nt triflate, the effect of a second pK Db ~ acid subs ~ itn ~ ion is leveled by the presence of Figure 3.17: Influence of pKa DMSO on the pot: ntiel o ~ dation, dé ~ erminé in the DMSO relative to the ENH, first groupc-llent acceptor, and this in acids of the family PhCH2SO2G, according to [Bor ~. '; 21.
especially for sulphonates. The study PhCH2SO2G type acids confirm this result, the dissociation energy of the CH bond .
being practically identical, between 89 and 91 kcaVmole, for the groups G of the series: Me, t-Bu, Ph, CH2Ph, NMePh, OPh, and CF3, while the acidity varies ~ 11 pKa units (i.e.
z 15 kcaVmole). In this case, there is a linear correlation between the oxidation potential of the anion and pKa of the acid with a slope close to unity when the two parameters are expressed in kcal / mole, as shown in Figure 3.17. In fact, this linear relation must be pr ~ llle when ~ dissociation energy is not const ~ nt ~ but it is then disturbed p ~ r the effect stabilizing or destabilizing acceptor groups with respect to the radical.
On the other hand, the gloup-llent triflate being essentially attractor by inductive effect, the second substitution on methane results in an increase in the stability of the CH bond of 10 kcaVmole, this bond being fin ~ l ~ ment more stable in TFSM than for methane. The trifluoromethanesulfonyl group makes it possible to synthesize slightly basic anions and particularly stable in oxidation. Let us specify to fix the orders of magnitude, that the substitution of the methyl group of the sulfonyl radical with its fluorine equivalent, makes it possible to gain z 13 pKa ~ units and increase the oxidation potential of the anion by z 1.2 Volts. We can thus reconcile the advantages of a delocalized soft base favorable to solvation in medium aprotic, while gar ~ ntics ~ nt anion oxidation potential consistent with the application envisaged in the field of b ~ tteries with polymer electrolytes.

Now let's take a look at the anions of the imide family, for this purpose Table 3.8 brings together various values of pKa making it possible to specify the influence of the anion center on the strength of their conjugated acid. Thus, pKa decreases on average z 9 units by substituting carbon with a nitrogen atom, the same result is obtained by restricting the acid CF3SO2XHPh which is the most close to our concerns.
Table 3.8 Influence of the center an- ~ n, _e (-N-- ou = C ~ -) on pKa ~ relative to DMSO, their cor. acids, ugués (except for ~ whose values are relative to the acidity scale in water).
Acid pKa (X = CH) reference pKa (X = N) reference / \ pKa CF3SO2XH2 18.8 [B1] 9.7 [B3] 9.1 CF3SO2XHPh 14.6 [B1] 5'7 [B3] 8.9 XH3 48 ~ [March] 38 ~ [March] 10 Ph2XH 32.1 [B2] 25 [B3] 7.1 (PhSO2) 2XH 12.2 [B1] 1.4 ~ [F&D] 10.8 On the other hand, Table 3.9 presents the effect of substitution on the acidity of ammonia and method, as we can see the effect of the substituents on their acidity is try ~ m ~ nt independent-i ~ nte of the anion center. All of this suggests that the acidity of TFSI family anions evolve identically to their TFSM family counterparts with the nature of the subssibl ~ nt ~ but with a ~ éc ~ lage z 9 pKa units.

...

Table 3.9 Influence of the acid center XH on the pKa of the RR'XH acids by derivative, with respect to XH3.
~ PKa with reference to XH3 Acid X = CH X = N ~ (~ PKa) / (~ average pKa) Ph2XH -16 -13 20%
CF3SO2XH2 -29.2 -28.3 3% CF3SO2XHPh -33.4 -32.3 3% (PhSO2) 2XH -35.8 -36.6 2%
Thus, knowing the pFS of the TFSM of 2.1 in the DMSO, we can therefore assume that the TFSI should present a pKa ~ -7 in this same solvent, even if it is insnffis ~ mment stable to strong acids to consider confirming this result. Identiq ~ lemçnt the pKa of the form bis (cyanoimide) acid, of formula HN (CN) 2, would be ~ 2, ie a substantially identical value to that of the TFSM. At this point, we think it is appropriate to come back to the acidity measurements of TFSI present during the study of protonic molten salts (Cf II-2), in the light of these results and collcep ~ that we introduced previously on the salt-solvent interactions. PKa TFSI in water of 1.7 Ir ~ 5], therefore differs ~ 8.7 units from the value estimated in the DMSO, such a difference expressed in terms of free enthalpy, according to the relation ~ Gj ~ (drnso) ~ Gi = 2,3RTapKa, corresponds to a TFSI ~ H + + TFSI energy difference ~ 12 kcaVmole. The anion ~ Gt ~ (TFSI), ~ Gt ~ (H +),, ~ Gt ~ (TFSI-) TFSI being a very TFSI ~ H + + TFSI soft base reloc ~ lisée, non-hydroxylic solvents ~ Gj ~ (water) Figure 3.18: Thermodynamic ionization cycle of the TFSI.
some are better solvents than hydroxylic solvents; if we assume that the difference in ionization energy (~ G;) corresponds to the gain in terms of solvation energy in passing from water to DMSO, the thermodynamic cycle of figure 3.18 makes it possible to write:
o ~ G; (TFSI) = ~ Gt ~ (TFSI) - / ~ Gt ~ (H +) ~ ~ Gt ~ (TFSI-) We can consider that the transfer energy from the non-dissociated TFSI from water to DMSO is negligible, this neutral species involving few electrostatic interactions with the solvent, from then knowing the energy of the proton transfer from water to DMSO [E ~
(~ Gt ~ (H ~ = -4.5 kcaVmole), we deduce a transfer energy from the TFSI anion from water to DMSO aGt ~ ('IFSI-) ~ -7.5 kcal / mole. This value is consistent with the transfer energy of BPh4- water with DMSO ~ -10 kcal / mole, classic example of a soft base delocalized little basic, and confirms us in the hypothesis that the pKa of TFSI in water is not abnormally high lemP-nt, but translates only ~ m ~ nt the least solvation of water with respect to anions vol.lminellx very deloc ~ lized n is also ~ l ~ ment possible to estimate from these results the number donor of the TFSI anion in dichloroeth ~ ne, the energy of transfer of water to another solvent ~ ugJnent ~ nt lin ~, ent with DNX- ~ dce lune ~ l, the assumed value of ~ Gt ~ ('IFSI-) ~ -7,5 kcaVmole would then correspond to a DNX- ~ dce ~ 4. If this calculation proves to be correct, the anion of ~ Lj +
TFSI would be less donor than BF4- in dichloroethane (DNX, dce = 6.03),: ~: ~ ~
while the molecule contains free doublets, carried by the atoms CF3 ~ _ CF3 of oxygen, donors in the sense of Lewis, with respect to the cation Li + in particular, EFFECT DONN ~ UR
as illustrated in the figure opposite. This result could probably be DE LOXYGENE
confirms by spectroscopic measurements, starting from the salt of letra-~ utyiammoniums of TFSi, and thus clearing the hypo ~ heses allows ~ ant to es ~ imer ie Dl ~ TFSI ~, dce.

In order to complete this study on the acidity of TFSI, Figure 3.19 illustrates the evolution of energy ionization determined in the anion gas phase containing at least one group CF3SO2 [Koppell ~ according to their pKa determined in DMSO ~ Bordwell6], except for the TFSI for which value is the one previously estimated The apparent linear relationship, between energy (~ G / pK) = 1,86 ~
ionization phase g ~ 7euse and the pKa of these 330 ~ a / ~ JNanions, is an el ~ -mynt which militates in favor of O / ~; 3 assumptions used to estimate the acidity of ~ - ~, 320 CF3 ~) 2N ~
TFSI at ~ -7 in DMSO. It is of course ~ 310 _ ~) 2NHPh obvious that these should be confirmed by ~ ~; 3N
further studies, such as the 300 ~
determination of the donor number of the anion of 290 TFSI in the DMSO. On the other hand, from - 1 o - 5 o 5 1 o 1 5 2 o vibration frequencies characteristic of TFSI, PKa determined by spectroscopic measurements Figure 3-19- ~ ionization energy in the gas phase one could probably estimate the energy of CF3SO2 as a function of their pKa in DMSO.
dispersion by dipole-dipole interactions between the solvent molecules and the anionic charge carried by the anion. Such an approach allows effect of justifying the low basicity of the picrate anion in DMSO [Gunwaldl. The TFSI anion is an excellent example of chemistry specific to dipolar aprotic media, such as electrolytes polymers utilic ~ n ~ poly (ethylene oxide) as a solvent.

The TFSI and the TFSM present ~ nt a similar structure, the g ~ vupel ~ ent CH being substituted by a nitrogen atom more electronegative, justifi ~ nt thus the least basicity of the anion, we can assume that the oxidation potential of TFSI should be greater ~ 530 mV than that of TFSM, i.e.
value ~ 5.88 V relative to a lithillm electrode This value is close to that recently detPrmin ~ e in imid medium ~ 701 fade ~ 6 Volts relative to a lithium ID electrode ~ mineY].
. .
Therefore, subject to the validity of the assumptions previously made ~ nt the conjugated acid of a delocalized soft base does presel ~ nt a moderately acidic pKa in water ~ 1, perhaps a c ~ ndid ~ t as a soluble and dissociated lithium salt in an aprotic medium, such as poly (ethylene oxide).

._ This comment partly justifies the orientation of our work on anions, which are not focused on looking for very strong acids in water, as a criterion for choosing anions, both the nature of the solvent modifies the dissociation as a function of the ~ clu.e of the anions envisaged.

Thus, the introduction of TFSI in electrochemistry widens the field of investigation for synthesis salts, for an application to lithium batteries, to the family of soft bases delo ~ illi ~ e ~. Cepe ~ L, me ~ lle ~ i u ~ a ~ ivn of TFSl d ~ l ~; e ~ b ~ ries e ~ l ellVi ~ and ~ pUi ~ piU ~
ten years, ec.ce.ntiel of the publications on the development of salts still refer to derivatives perfluorosulfonates, or in the development of solvent mixtures making it possible to use the salts derived from coordination anions. This only highlights the few interactions between the approach from electrochemistry, and that from organic chemistry; however, it is clear that concepts that we introduced previously justifying the choice of the TFSI in an aprotic environment have their origins in organic chemistry, and this for an application to electrochemistry The p ~ sellce of a labile proton on the TFSM limits its long-term stability with respect to an electrode of lithium ~ however, the possibility of making nucleophilic substitutions on its carbanion extends the possibilities of grafting an acid function on different IKoshar ~ Benrabahl substrates. In in particular, the addition of CF3SO2F on the dim ~ gnesian of the TFSM makes it possible to synthesize the triTFSM, of formula (CF3SO2) 3CH, carbanionic equivalent of TFSI. The TFSI and the TFSM are ~ ct ~ l ~ llçm ~ nt among the strongest acids known in the ga phase ~ use IKoPpell ~ with respectively an ionization enthalpy of 291.8 and 289 kcaVmole. POE being a dipolar aprotic solvent, the conductivity of a salt is essentially an image of the basicity of the anion in this medium when the electrolyte is in a fully molten state. The conductivity of the lithium salt of TFSI and of triTFSM being roughly equivalent in IBenrabahl POE, we can assume that the two anions have an equivalent basicity also in aprotic medium.

One of our main goals being to find substitutes for TFSI based on more chemistry simple to implement, we have envisaged the use of acids derived from cyanated substrates, such as malononitrile, cyanamide, and tetracyanoethylene. We will present in a first time a bibliographic study allowing to specify the synthesis and the properties physico-chemical of these salts, and thereafter, the res..lt ~ tc obtained in conducdvity for some of them. It may seem strange to be interested in cont ~ n ~ n ~ acids of grol.pe- ~ ntc cyanées conn ~ ic ~ nt their instabilities with respect to lithillm, in particular for acetonitrile. However, if we s ~ 30se that the slow cyan ~~ gloups placed on the considered anion participates in the relocation of the charge, these are st ~ bili.céc. Therefore, it is possible to use g, oupe ... çntc cyanés for salt synthesis, this justifies our interest in derivatives of malononitrile and c ~ e ~

!
Fan ~ ille of anions deviating from malononitrile: Malononitrile, cyan equivalent of TFSM, being more basic than this one - 9 pKa units ~ its carbanion is more nucleophilic than that of TFSM. Therefore, it is easy to synthesize many derivatives by s ~ lbstit ~ ltion nucleophile on c ~ rbanion of m ~ lononitrile by an electrophile radic ~ l (E), as illustrated in Figure 3.20. With perrnett ~ nt radicals to relocate the anionic charge (such as M (1ic ~ aryls, carbonyls, sulfonyles, or cyanos), soft deloc ~ li.cées bases are obtained, the conjugated acids of which are relatively strong. If we take the example of tricyanomethane, l., ~ Igu ~ C 3.21 nGUS shows that the anionic charge is delocalized on the central carbon and by the three cyan functions, this justifying its pKa ~ -5 [March] relative to water, 16 units lower than that of malononitrile.
c ~ N c ~ N c ~ N C4 ~ N

~ N E - C ~ l ~ O ~ C ~ 3N = CC ~
C ~ N
Figure 3.20: Synthesis of delQ ~ AIi ~ és Figure 321: D, _, ~ r:, 3 ~ n of the charge an ~ e of from the anion of malononite. :. 'e tricy ~ noi "t: ~ l, dne by resonance.
We could also have considered hexafluoroacetylacetone as a substitute for TFSM, cepen ~ nt, M (~ ir ~ n ~ c electrophiles reactiCs ~ n ~ preferell ~ iPllPment on the oxygen of its carbanion to form a mixed anhydride IV ~ I, it is not possible to synthesize derivatives thereof starting from simple chemical reactions, which greatly limits their interest. In addition, the methane tris (trifluoroacetyl) hydrolyzes in water (or in alcohol) at room temperature giving he ~ fluoroacetylacetone and trifluoroacetic acid [V ~], this is incompatible with basic chemistry such as that specific to lithium batteries. On the sidelines of the problems posed by the synthesis and chemical stability of derivatives of hexafluoroacetylacetone, the compounds B-dicarbonylated chelating cations, and in particular the lithium cation, it & ut expect a less dissociation of their lithium salts in the polymer. Note, however, that this property could be taken advantage of for the transport of lithium by a mechanism vehicular II P ~~ "'~ 3], the principle of which is to compensate for the low number of transport of lithium (t + ~ 0.3) in the POE by masking it with the polymer using complex ~ n ~ c, thus allowing its transport in anionic form.

Let’s come back to the synthesis of malononitrile derivatives, the carbonyl derivatives of malononitrile s'obtipnnp-nt essentip ~ llement by acdon of a carbonyl possé ~ nt a leaving group in the form of an acyl chloride [Kemll, an acyl cyanide [~ l ~ l, an acyl azide [7], a anhydride U, of a carboxylic ester ltl on a salt, usually sodium, of malononitrile.
Unlike the previous methods, diethyl phosphorocyanidate [JaP1], of formula (EtO) 2P (O) CN, in the presence of a base, allows the directPment malononitrile to be coupled with a carboxylic acid. In the same way, we obtain the tricy ~ named ~ h ~ ne by acdon of bromide or cyanogen chloride on the anion of malononitrile 0. Strangely, the only example to our conn ~ isC ~ rce of sulfonyl malononitrile described in lillél ~ tu ~, in this case, e-lce tolu ~ neslllfonyl malononitrile, is not synthesized by the action of a sulfonyl halide on the anion of m ~ lononitnlp ~ but by treating the anion of tolu ~ nesulfonyl acetonitrilP by BrCN [a'lc ~ ar ~ dl Les aryls malonor i ~ rile5 are obtained in the same way from anions of aryl acetoni ~ riles treated by BrCN or ClCN a.

We have grouped in Table 3.10 various pKa values of malononitrile derivatives by relative to the acidity scale in water or in Dl ~ SO. It appears that the pKa measures of anj l ~
water / dioxane mixture are more basic than the values obtained in sulfuric acid solution or in DMSO, there is thus a difference of 3 pKa units for malononitrile. In addition, if we compare the acidity of acetyl malononitrile in the water / dioxane mixture and that of the ester of malononitnl.o. in aqueous sulfuric acid, the two values differ from 8 pKa units. Gold, based on the pKa determined for TFSM in DMSO and taking into account that the effect of s ~ lbs ~ ih ~ n ~ is lessened in water [Bo ~ well3], we can expect a maximum difference - 2 pKa units ~
or a difference of ~ 6 pKa units considering only the effect of the solvent on the acidity.
Table 3.10 Acidity of malononitrile and its derivatives in water and in DMSO.
Acid pKa F ~ h ~ lle Solvent reference CH2 (CN) 2 11.2 H2o H2O [Pearson3]
CH2 (CN) 2 14.1 H2OH2O / Dioxane 1/9 (Vol.) [Kem2]
CH2 (CN) 2 11.1 DMSO DMSO [Bordwell1]
PhCH (CN) 2 5.8 H2O? [Josey]
PhcH (cN) 2 4.2 DMSO DMSO [Bordwell1]
p-NO2PhCH (CN) 2 1.9 H20? [Hartzler]
MeO2CCH (CN) 2 -2.8 H2O H2O / H2SO4 [Boyd1]
MeCOCH (CN) 2 5.2 H20H20 / Dioxane 1/9 (Vol.) [Kem2]
PhCOCH (CN) 2 4.3 H2OH2O / Dioxane 1/9 (Vol.) [Kem2]
p-NO2PhCOCH (CN) 2 3.6 H2OH2O / Dioxane 1/9 (Vol.) [Kem2]
HC (CN) 3 -5.1 H2O H2O / H2SO4 [Boyd1]
This significant difference can be explained by the fact that if dioxane makes it possible to solubilize acids organic insoluble in water, this solvent is not very acceptor in the sense of Lewis, moreover the symmetry of the molecule makes it unpolar. On the other hand, the pH measurement is done using a glass electrode whose measurement is referenced to that obtained for an oAo solution of 10-5 M perchloric acid, it assumes that the behavior of anions derived from malononitrile is identical to that of perchlorate in the mixture Dl water / dioxane. It is therefore more interesting, in this case, to refer to the values obtained in aqueous sulfuric acid solution, concentrated solutions (~ 50% in volume) of sulfuric acid are more donor than water alone. In addition, the soft bases deloc ~ li.sé ~ s derived from malononitrile must be better solvated in DMSO, as illustrated the example of phenyl malononitril ~, this reason ~ ment may extend to poly (ethylene oxide).

!
Another parameter to consider is the electrochemical stability of the anions deriving from the malononitrile, also we have grouped in table 3.11, the oxidation potentials, d ~! ~, ... in ~ s in the DMSO, of some anions [Kem21, referenced here to a lithium electrode.
Table 3.11 '~ t ~ l oxidation vs. (Li + / Li ~) of malononitrile and denv ~ s in DMSO.
Acid CH2 (CN) 2 CH3COCH (CN) 2 PhCOCH (CN) 2 p-NO2PhCOCH (CN) 2 Eox (A-) ~ 3.75 z 4.48 ~ 4.52 ~ 4.62 These values are determined by cyclic voltammetry using lithium perchlorate as bottom salt, gold, for anions çhél ~ trnts the lithillm cation, the anionic charge being thus partially stabilized, the measured oxidation potential is apparently higher. So for acetylacetone which has a high association constant with Li + (pKaSs = 4.77), the potential of apparent oxidation is 520 mV higher with LiCl04 as the bottom salt compared to the value obtained in llt ~ rnt Et4N + BF4-, the tetraethylammonium cation not being complexed by acetyl ~ ketone IB "~ 1. However, the anions that we consider are not very complex with respect to Li + screws, for example pKaSS <1 for malononitrile [~ o ~ well1], we do not have to causes the values of the oxidation potentials for these. On the sidelines of this problem of comple ~ rtion of Li +, the potentials being determined with respect to ferrocene, it is necessary ~ ire to calibrate its potential with respect to the normal hydrogen electrode. The differences observed between different authors are explained by a problem in the reference chosen for this couple, at the time current, we admit that the ferrocene couple is ~ 719 mV in DMSO compared to ENH [E '~' ~ "4l We therefore directly obtain potentials referenced to a lithium electrode in adding ~ 3.719 V to the values referenced to ferrocene. We can then connect the potential oxidation of the anion to pKa of its conjugate acid [Kem21, as illustrated in figure 3.22, for lead to the following linear relation:
Eox (A ~) = 4.89 ~ - 0.0812 pKa If we express this equation in trr ~ -icrnt 4.8 the potential and the pKa in kcaVmole, one 4,6- ~ p-NO2PhCO¦RCH (CN) ¦
leads to a slope # 1.37. This result is PhCO ~ CH CO 2 not restricted to the derivative of malononitrile, J 4.4- \ ~ -but this is also true for the anions t ~ 4 2- \
derivatives of hydrocarbon acids (cyclo- \
pentr ~ i ~ nyle, indenyl, fluorenyl, ..) [KelT ~ 2]. 0 4 -More recent work [E ~ ~ 11 on these 3, 8 - ~ H ~ rniPrs anions for which the potential 3,6 oxidation and acidity were determined 2 4 6 8 10 12 14 16 in the DMSO show a slope close to pKaH20 of the unit between the oxidation potential and Figure 3.22: the evolution of the anion oxidation policy ~ ~ Derivatives of malononitrile according to their acidity.
PKa expressed in kcal / mole ~ ela assumes in 2i94127 ..
fact that we can link the pKa in water and in the water / dioxane mixture by a linear relation, which is conceivable by referring to r ~, slllt ~ ts obtained for phenol and its derivatives in water and in the DMSO, as we mentioned before ~ nt We will retain this hypothesis to estimate the oxidation potentials of certain anions in DMSO. It should be noted that this re ~ isonnpment can only apply for anions of illU ~; lult; simil ~ ire, and i ~ lement on anions where we substitute) e one of the gronpçm ~ ntc on a given substrate, limit ~ nt thus the effect of securing radicals, as we have set it out for the anions of the TFSM family. In particular, conn ~ iss ~ nt the effect of replacing an acetyl group with a group trifuorométh ~ neslllfonyle in ~ ketone (~ pKa ~ 17), and in acet ~ mide (~ pKa - 16) IBordWell6] ~ on can expect increased oxidation potential from acetylmalononitrile to trifluororn ~ th ~ neslllfonylmalononitrile - 1 V, i.e. an anion oxidation potential ~ 5.5 V close that of the TFSM (Eox = 5.35 V).

We have also considered cy ~ n ~ mide (of formula H2NCN), product of great chemistry organic, in the cyan carbanion family, for the synthesis of salt. Indeed, one can im ~ gjner to synth ~ ticer ~ for example, a sl1bstitut at TFSI in which one of the fluorinated groups (-SO2CF3) is replaced by a non-fluorinated gronp.oment (-CN) reducing the cost and the weight accordingly of the molecule. This salt for which we can expect less acidity ~ 4 pKa units per compared to TFSI, an estimated value of pKa ~~ 3 in DMSO, should however remain so sllffic ~ mm ~ nt to be dissociated in the POE.

All this fully justifies our interest in cyan anions, despite the doubt that may remain on the stability of cyan groups with respect to lithium, this particular point cannot be decided, in our opinion, only by long-term battery tests, this going beyond the framework of our works. Also, we will present below the possibilities offered by the cyanated acids deriving tetracyanoethylene.
Family of anions derived from tetracyanoethylene: Work on tetracyanoethylene (TCNE) and its derivatives come essentially from a research program pursued by the Dupont de Nemours company in the 60s and 70s aiming to develop a cyanide substitute for Teflon ~ 'developed by the same company in the 1950s.
the impossibility of syn ~ h ~ ticking such a polymer, this research resulted in an N ~ ", N
extensive knowledge about the chemistry of cyan compounds and their characteristics> ~ <
physico-chemical (pKa ~ optical absorption, oxidation potential, ...), and at N ", cc ~ N
writing of articles of complete reviews on this subject ~ Websbr1, y. We will not specify TCNE
not in detail the possibilities offered by this chemistry, the reader more particularly interested who can refer to the previously cited review articles. We will have the opportunity to return to this work during the study of heterocyclic anions, and molecules with redox properties.

21 94 ~ 27 , ..
We have grouped in Table 3.12 the acidities obtained for some of these anions in sulfuric acid solution relative to the acidity scale in water. From the TCNE, it is therefore envi ~ a ~ P ~ IP to obtain deloc soft bases ~ lily ~, es whose conjugated acids have '~ ci ~ lit6s typically between 0 and - ~; the lithium salts thereof should therefore be potentially dissociated in aprotic medium, and more particularly in POE.
Table 3.12 pKa of cyan acids derived from tetracyanoethylene relative to water.

NC ~ ~ C ~ H + 0.6 ~ C: = C; -N = C ~ -C ~ 3 H + -6.07 p ~ li ~. .;., ~ I) -phenyl- e ~ al ~ o ~ mine di ~

NC CN, CN NC Hl CN
OC ~ -C = C ~ -C ~ H + -3.55, C = C ~ -C ~ H + -8 D ~ i nc T ~ - ~ pJ ~ pène NC H
NC O H + NC ~ ~ CN
~ C: = C -5.3 ~ C = C ~ -8.5 NC CN NC C; -CN
NC ~ H +
Alcooll ~;. "Li ~ e ~ c 'Db ~ tylène It should be noted that for acids incorporating the tricyanovinyl radical, it is not possible to scour a form of resonance for the cyano group placed in a of the malononitrile motif, as this is illustrated in Figure 3.23. Theoretical calculations of quantum mechanics [Boyd2l thus show that the ~ R = 13, ~ R _ C ~ O - C = R =

Figure 3.23: Resonance form of the tricyanovinyl radical.
effective load is lower on the nitrogen of this groupemPnt We can therefore consider a lower st ~ bilit ~ de ce ~, pe ~ "Pnt cyano with respect to a lithium electrode. Even if it is only of a hypothesis, which should be confirmed by a battery test, we did not in ~ éless ~ s in a plc; first time with these anions.

~ - 56 -II-3-2-2 Synthesis of anions:

TFSI family anion:
, ~, ~ CF3SO2NLiSO2N (CH3) 2: 26.85 mmol (4.004 g) of CF3SO2NH2 (Fluorochem), 26.85 mmol (3.856 g) of medium dimethylsulf chloride (Aldrich, 99%), and 53.71 mmol (5.435 g) of tncthyl umir.c scnt acts ~ s penu ~ nt .lu ..; re days in TH ~, u, h ~ dre. L., pr ~, hydrochloride response of triethylammonium is then removed by filtration, the THF evaporated, then the product taken up in the water. Then added ~ 30 mmol of KHCO3 (3 g), after release of CO2, the water is evaporated, the product dried under vacuum at 60 ~ C overnight, then taken up in methyl formate, filtered, and dried again. The product is obtained with 85% yield NMR analysis of fluorine in CD3CN, gives a majority peak at -74.019 ppm (CF3SO2N (TEA) SO2NMe2), and a peak at -75.741 ppm (CF3SO2NH2), which represents 2.6% of the fluors present. An aqueous solution of product is then buffered to pH ~ 4, washed with ether to remove the trifluorométh ~ nPs.llfAmi ~ e unreacted, acj ~ ifiPR at pH ~ 1 with a 4M solution of hydrochloric acid, then extracted with diethoxyethane. Lithium salt is obtained by adding lithium carbonate (10% excess), once dry the product is taken up in methyl formate and then dried under vacuum at 60 ~ C overnight. NMR analysis of fluorine and proton shows that the salt of the dimethylsulfamoyl trifluorométh ~ nP-sulfonyl imide in the form of a pure white product.

Malononitrile & Thousand Anions:

~ CHLi (CN) 2: the lithium salt of malononitrile (Aldrich, 99%) is obtained by treating it with a equivalent of lithium hydride in dimethoxyethane or cold tetrahydrofuran (~ 0 ~ C); the hydrogen evolution stops quickly and the salt is then used in situ after a period of time 15-30 mins. If this salt is to be treated with a halide, two equivalents of LiH, the second later serving as a base.

~ CH3COCLi (CN) 2: acetylmalononitrile is a commercial product (Aldrich, 98%). Salt lithium is prepared by treating the acid with an excess of lithium carbonate (~ 20 mol%). After evaporation of water, and ~ h ~ ge of the salt overnight at 50 ~ C under vacuum, the product is taken up in the bike ~. ; the, filtered, and vacuum dried at SO ~ C for ~ 24 hours. The infrared spectrum of the product shows an intense peak of absorption of the nitrile function (~ 2200 cm-l), and the absence of a peak co ~ l ~,., polldant with a carboxamide function (~ 1650 cm-l) which could have resulted from the hydrolysis of cyan functions of the product.

~ CF3COCLi (CN) 2: the trifluoroacetylmalononitrile is synth ~ ti ~ ed by treating the lithium salt of malononitrile by trifluoroethyl trifluoroacetate (Aldrich, 99%), this reagent being a substitute stable and manipulable with trifluoroacetyl chloride which has the disadvantage of being ._ gaseous (Teb = -30 ~ C) and easily hydrolysable, in cold dimethoxyethane (z 0 ~ C) for ~ 2 hours, then at room temperature ~ .alulG for ~ 24 hours. After evaporation of the solvent, the product is taken up in methyl formate, filtered, then dried for z 24 hours at 60 ~ C under empty. We obtain at this stage a rçn ~ çment z 80% lithium salt of trifluoroacetylmalononitrile, the produces p ~ se ~ nt a slight fluorescent green tint. Fluorine NMR analysis performed in CD3CN has a major peak at -68.325 ppm (CF3COC (CN) 2Li) and a minor peak at -71.4239 ppm. Le recol-i des;, urraces ues deuic pics inuiqUeill a pUI ~ C / ~ 0 pOUI '; e pl ~ Jd ~ vi ~ ~ L Vi ~ i du fluorine.

~ PMAMLi: first, the polymerization of trifluoroethyl methacrylate (Lancaster, 98%), initiated by AIBN (1% by weight), is carried out in anhydrous THF at 80 ~ C for z 24 hours. After bringing this solution to 0 ~ C, a solution is added to the EMR of m ~ lononitrile (20% in excess) treated with lithium hydride (10% in excess). The polyelectrolyte formed pl ~ cipite in THF during the reaction. After 24 hours, the solution is centrifuged, and the polymer dried at 60 ~ C under vacuum. The synthesis is illustrated in the figure below:

CH2 = C THF 80 ~ C ~, ~ n CH2cN2 ~ 2LiH, ~ 3'n AIEN DME
C = OC = OC = O
OO, C
CH2 ICH2 N Lj + N

~ RSO2CLiC (CN) 2: the synthesis of sulfonated derivatives of malononitrile from chlorides of sulfonyles show a rapid coloration of the solution concomitant with the oxidation of the malononitrile, especially for trifluoromethanesulfonyl known to be a dimethyl malonate IT'Jli ~ nl chlorinating agent. So we have used mainly suLfonyles fluorides, fluorine being more solvated than chlorine, it is thus made less oxidizing than this despite its superior oxidation potential.

Meth ~ nesl-lfonyl and paratoluenesulfonyl malononitrile are obtained by treating the salt of malononitrile lithium by respectively, an equivalent of methanesulfonyl fluoride (Lancaster), or an equivalent of paratolu ~ nesulfonyle (Aldrich, 99%) in pl ~ sellce of an equivalent of LiH, in cold tetrahydrofuran (z 0 ~ C) for z 2 hours, then at room temperature ~ .d ~ l ~
pçn ~ nt ~ 24 hours. The solution is then centrifuged to remove the LiF formed and then evaporated.
After drying ~ ge at 50 ~ C under vacuum for z 24 hours, the lithium salts are taken up in the formate methyl to remove the LiF which may remain in the product, then dried for # 24 hours at 60 ~ C under vacuum. The slightly colored lithium salts are thus obtained, with a yield of 90%.
Microanalysis of these products and 1 H NMR (80 MHz) confirm ~ the purity of these products:

2 ~ 94127 CH3SOzCLiC (CN) 2:
RQUII ~ lS% C% H% N% S
e 32.37 2.06 18.63 20.58 C4H3S02N2 32.01 2.01 18.67 21.36 1 H NMR (80 MHz, CD3CN):

CH3PhSO2CLiC (CN) 2:
Results% C% H% N% S
~ analysis 52.67 3.11 12.32 13.3 CloH7SO2N2 53.1 3.12 12.39 14.37 1 H NMR (80 MHz, CD3CN):

Since trifluorometh ~ nPsulfonyl fluoride is not commercially available and gaseous, we used trifluorometh ~ nesulfonyl imid ~ 7ole (Fluka, - 97%), imidazole being a fairly good leaving group which also has the advantage of being relatively stable to hydrolysis and liquid, while by being less oxidizing than the chlorinated derivative. Thus, the lithium salt of malononitrile is treated with an equivalent of this product in dimethoxyethane when cold (~ 0 ~ C) for ~ 2 hours, then at room temperature for ~ 24 hours. After evaporation of the solvent, the product is dried for ~ 72 hours at 60 ~ C under vacuum in order to eliminate the imidazole produced during the reaction, taken up in methyl formate, filtered, and finally dried for ~ 24 hours at 50 ~ C under vacuum. The product is obtained colored with a yield of ~ 70%.

~ LiC (CN) 3: The lithium salt is obtained by ion exchange, by treating the potassium salt (Alfa) with an equivalent of anhydrous lithium chloride in THF. After ~ 30 min, the solution is evaporated, the product dried in a vacuum desiccator at 60 ~ C overnight, taken up in etQnitrile anhydrous, the solution filtered, evaporated, and lithuim salt dried overnight at 60 ~ C under vacuum.

C ~ -ide:

~ LiN (CN) 2: sodium bis (cyanoimide) is a commercial product (Aldrich, 96%). We obtain its silver salt by treating it with an equivalent of silver nitrate in water, this insoluble salt is then treated with an equivalent of lithium chloride After ~ .limin ~ tion by filtration of the chloride of silver formed, the solvent is evaporated, then the product dried for 24 hours at 50 ° C. under vacuum. Salt is then taken up in methyl formate, filtered, and again dried ~ 24 hours at 60 ~ C under vacuum.

~ CF3SO2NLiCN: the trifluorom derivative ~ th ~ neslllfoné of cy ~ n ~ mide is obtained by treating the di-salt sodium (~ 20% molar excess) of cy ~ n ~ mide with trifluorometh chloride ~ nesulfonyl in tlim ~ cold thoxyethane (# 0 ~ C) pçn ~ l ~ nt - 2 hours, then at room temperature for 24 hours. By filtration and evaporation of the solvent, the excess cy ~ n ~ mide and the chloride are eliminated no reacts. After evaporation of the solvent and drying, the sodium salt is taken up in THF and treated by an excess of sodium chloride (z 10%). After evaporation and drying, the product is taken up in methyl formate, filtered, and finally, the evaporated solvent and the product dried z 24 hours at 50 ~ C
under vacuum. Fluorine NMR analysis in CD3CN gives a major peak at -74.1891 ppm (CF3SO2NLiCN) and two minor peaks at -74.3893 ppm and -75.0165 ppm, the area ratio indicating purity ~ 80% with respect to lluor.

!
II-3-2-3 Conduct:

TFSI family anion:

The conductivity of the lithium salt, for a composition at O / Li = 12/1, of dimethylsulfamoyl trifluorométhanesulfonyl imide, which we will call more simplem ~ nt diméthylaminoTFSI est rlncc on figuic 3. ~, cn comp, ù.iis., il ~ iv ~; C h CùiluuC; i vii ~ du Li l r. ~, / Li = b / 'i.

1 o-2,,,, I, --- 1 0 3 ~ ~ ~ ¦ Me2NSO2NLiSO2CF3 ¦
E ~ _ E 10'4 ~
O
.10 5 -- o O / Li = l 2/1 0 6 TFSI O / LI = 8/1 - -2, 5 3 3, 5 1000 / T (K) Figure 3.24: Conductivity curve for dimethylaminoTFSI.
Despite a pKa that can be estimated at least five units greater than that of the TFSI, this anion gives a conductivity which is only a factor two lower than that obtained for the anion of the TFSI. This example proves that it is thus possible to synthesize derivatives of TFSI not containing than a perfluorinated group, while retaining appreciable conductivity in the POE. This the result is interesting for several reasons. In fact, most of the cost of the TFSI comes from the n ~ ce-csit ~ to go through an expensive electrochemical fluorination step, therefore, if we reduce half the amount of fluors present in the molecule, it is possible to reduce its cost. On the other hand, these compounds open the way to a whole family of salts with properties surfactants, for which we can hope to control the physicochemical properties, by nature and length of the alkyl chains present on the nitrogen; well, it's likely that his homologous with two grwp ~ e ~ ts propyles give even better conductivity of this salt in the POE.
This synthesis opens eg ~ llomçnt the pe ~ e to obtain salts carrying crosslinkable functions pe puts ~ nt to graft them. This relatively simple reaction to transform a halide from sulfonyl in a dissociated salt, should even ~ lemçnt allow to solubilize in the POE various molecules, such as redox species, and thus introduce different functionalities within polymer electrolytes, such as overload protectors I

Anions d ~ of the ~; ~, ~ ide:

~ LiN (CN) 2: The conductivity obtained, at decreasing temperature ~ nte, for the lithium salt of the bis (cyanoimide) for different concentrations is given in Figure 3.25, with reference to the conduc ~ é of LiTFSI of composition O / Li = 8/1 [SYllal which will serve as a comparison ~ l ~ mPnt in most c ~ s. The three conductivity curves (O / Li = 8, 12, and 16) initiate a break li ~ e à la recr ~ LroiyL ~ ~ i ~ SO ~ c. Above this temperature, the lithium salt bis (cyanoimide) makes it possible to obtain conductivities lower by about an order of magnitude to that of LiTFSI at O / Li = 8/1, which nevertheless represents a completely interesting result for views of the simplicity of the anion in terms of chemistry and its low mass (LiN (CN) 2 = 71 compared to 287 for LiTFSI).

1 o 2 1 ~
¦ (Nc) 2NLi ¦

E ~ o - O ~

'~ D \ Q \
- 6- ~ O / Li = 8/1 ~
10 - OO / Li = 12/1 o 7- 0 0 / Li = 16/1 10 ~ TFSI O / Li = 8/1 1 o-8 2.5 3 3.5 1000 / T (K) Figure 3.25: Conductivity of the lithium salt of bis (cyanoimide).
The potential interest of this salt is obviously ~ nt conditioned to the electrochemical stability of the anion of bis (cyanoimide), especially with respect to a lithium electrode, the oxidadon potential of the anion which can be esdmated, as for LiTFSI to a value greater than 530 mV than that of malononitrile or ~ 4.3 V.
We could be criticized for not having sought to confirm this result by a study electrochemical of the redox stability domain of the bis (cyanoimide) anion in POE, however, at first, it seemed to us more important to define the f ~ miles of anions may give rise to lithium salts suffi.c ~ mment dissociated in the POE. Furthermore, as we have already begged for it, the experience acquired for more than ten years as part of the project ACEP shows that it is difficult to escape a real test, in standard con ~ itions, of cycling in battery of new materials to validate or invalidate their uses. These tests allow compare the performances obtained for a new salt both in terms of cyclability, p - iss ~ nce, and thus define the potential field of application of a new material. He is in ~ 21 9 41 2 7 effect n ~ stop ~ ire to keep in mind that the applications of batteries to electronics are gener ~ l ~ ment less contr ~ i ~ n ~ n ~ es in terms of pui ~ s ~ nce and cyclability, unlike applications for electric traction. This first example shows us that it is completely possible to find lithium salts based on non-fluorinated chemistry present ~ nt a c ~ nductivity certainly inf ~ -; e ~ to that of LiTFSI, but allows ~ nt at least to consider their Uses in primary or secondary battery systems of low cost using non-crict electrolytes ~ is: ~ bles e ~ poLenLiellement plasti ~ les.

On the sidelines of the problems of possible applications of this m ~ t ~ n we have gathered on the figure 3.26, the con ~ uctivity of different anions of the TFSI family for O / Li = 8 / l. It is interesting to note that the order of the conductivity values respect the order of the attractiveness of the substit1l ~ ntc, allowing to link the increase in conductivity to a better dissociation:
, attractor effect of the substituent: CH3S02 <CN <CF3CO <CF3S02 If it is once again assumed that the difference in acidity obtained in DMSO, for the anions of the family of the TFSM, is kept for the anions of the family of the TFSI, the gain -4 pKa units, resnlt ~ nt of the substitution ~ ltion of the methylsulfonyl group with a nitrile, corresponds approximately to a increase of an order of magnitude by conductivity. On the other hand, the substitution 1 o-2 of the nitrile group by a group ~ E 1 ~ -3 ~ CF3SO
trifluorom ~ th ~ nesulfonyl causes ~ 1 0 4 - C ~
decrease in the basicity of the anion ~ 9 E ~ C ~ F CO ~ E}
pKa units, resulting in an increase ~ 1 o - 3 conductivity of only one order _ 1 o 6 of greatness. It is likely that everything ~ R2NL ~
as for liquid media, the ~ 10 - CH3SO2 conductivity ~ increases when the salt is 1 o-8 ~
more dissociated, in other words when 2, 4 2, 6 2, 8 3 3, 2 3, 4 the anion becomes less basic, until 1000 / T (K) as the solvation energy of the anion polymer dis ~.; ~ s belonging to the family of imides.
with respect to salt becomes greater than the energy of solvation and dissociation of salt. This remark actually refers to the concept of given number of anions previously introduced (Cf II-3-ll).

Identically to the TFSI, it would be important to be able to have the values of the DNX- ~ solv of these different anions, and in particular in media such as low mass liquid glymes (M ~ 100-300), which would correlate the thermodynamic data obtained by spe .; ~ tri ~, to conductivity measurements cl ~ termin ~ es in poly (ethylene oxide) medium, and so to make a link between the physical chemistry of liquid solutions, and that specific to the polymer medium.

~ t 941 27 _ CF3SO2NCNLi: The conductivity of the lithium salt of trifluorom ~ sh ~ nesl-lfonyl cy ~ n ~ mide is given in Figure 3.27 for dirré ~ ntes compositions (O / Li = 8, 12, and 16). Conductivity obtained is only two to three times lower than that of TFSI, and therefore allows to gain almost an order of magnitude on the conductivity of bis (cyanoimide), as illustrated in figure 3.28.
1 o-2 CF3SO21'1 L ~ 5 _ ~ 0 ~ 4 - ~

: ~ o O / Li = 8/1 ~ \
1 o-6 _ o 0 / Li = 12/1 \ -oo O / Li = 16/1 1 o'7 - TFSI O / Li = 8 1 o.8 2.5 3 3.5 1 oOO / T (K) Figure 3.27: Conductivity of the lithium salt of trifluoromethanesulfonyl cyanamide.
In addition, the electrolyte of composition O / Li = 8/1 does not recreate ~ t ~ lli.ce that from ~ 40 ~ C, or - 10 ~ C in below the electrolyte of the same composition with bis (cyanoimide), it is likely that this lowering the temperature electrolyte melting, i.e. to be connected to the 1 o-2 presence of the triflyle grouping in -E 10'3 ~ ~ _ ¦ 0 / Li = 8/1 ¦ -the molecule, in accordance with the _ "4 _ ~ --3 ~ - ~ -previous results ~, on the influence of E
salt concentration on the --1 0 ~ 5 glass transition temperature of '1 o-6 _ e N (CN) 2 \ \ E3 _ polymer electrolytes (Cf II-3-1-3, ~, ~ CF3SO2NCN \ ~
table 3.4). o 10 7 _ TFSI \ ~, 1 0 ~ 8 To conclude with the anions 2, 5 3 3, 5 1 000 / T (K) cyanimide derivatives, Figure Figure 328 Com ~ iSOn of conduc ~ ivity of Ijthium salts ~ .28 allows to compare the of the bis (cydn ~, 'e) and the trfflu ~ ru ,,, ~ lhanesulfûnyl cyanimide conductivity of the lithium salt, bis (cyanoimide) and trifluoromethanesulfonyl cyanamide, respectively, for a composition at O / Li = 8/1.

A ~ ions of;. ~ Ls of malononitrile:

The chemistry of malononitrile being relatively easy to access, we wanted to study the infl-lP.nce on the conductivity of the nature of the substituent, namely acetyl, methanesulfonyl, para-tol ~ l ~ nesl ~ lfonyle, cyano, trifluolu ~ célyl, and finally trLfluorû ~ h ~ neslllfonyl m ~ onitrilp ~ In order to to simplify the presentation of the results, we give on figures 3.29, 3.30, and 3.31, the iso ~ hermes of conduclivities respec ~ ively for the compositions at 0 / Li = 8, 12, and 16/1, and this for lel ~ peldtures between 40 and 100 ~ C for these different anions.
1 o. 2 I 0 / Li = 8/1 1 ~
- ~ 10 3- _.
_o - 100 ~ C ~
0 1 0 - = ~ = = ~ ~

, 5- ~ /

0 1 o- 6 _ e /

1 0 7 1 1 ~ oN <~, O
s ~ s ~
Figure 329: Isotherm of conductivity for anions derived from malononitrile at O / Li = 8/1.

1 o 2 1 ~, ¦ O / Li = 12/1 ¦
~ 10-3-_ "- 1 00 ~ C ~
F 1 o 4 - 80 ~ C ~ ~~ a ~) 5 - 50 ~ C ~ - ~ / -~ a ~ 10 ~ 6 - 7,

8 oo ~ ~, o Figuro 3.30: Is ~ ll.er ", e of conductivity for the derhan ~ anions of malononitrile at O / Li = 12/1.

102 1 1 1 1 ll ¦ O / Li = 16/1 ¦ ~
--1 o 3 - ~ _ - - 80 ~ C ~ L ~
s 10 4 -60 ~ COO Ef ~ -- 50 ~ C [1 ~ -10 5 ~
f 3--1 o-6-40Oc o 0 '= O ~ ~ O

Figure 3.31 ~ l, e "" e of conductivity for the anions deriving from malononitrile with O / Li = 1 6/1.
As we can see, the conductivity of lithium salts for O / Li = 8/1, and when the electrolyte is in a fully molten state, augmen ~ P, with the nature of the following substituent:
MePhSO2 <MeCOz MeSO2 ~ CN <CF3CO <CF3SO2 The lower conductivity observed under these conditions for toluenesulfonyl malononitrile vis-à-vis vis acetyl malononitrile, despite the lower attractiveness of this substituent, can be explained by the presence of the phenyl group which must stiffen the polymer matrix, and thus cause a higher increase of the glass transition temperature of the POE with the concentration of salt, as illustrated by the results above ~ emment exposed on tetraphenylborate (Cf II-3-1-3, table 3.4).

It is interesting to conct ~ er, that for a given substrate, the conductivity is essentially related to electron-withdrawing power of the substituent, thus helping to facilitate the dissociation of the envisaged salt in poly (ethylene oxide). Apart from problems with the structure of anions, such as the presence of a rigid phenyl group ~ the electrolyte, the laws governing ~ n ~ c dissociation of salts in the medium polymer, therefore seem to be after all coherent with the results obtained in liquid medium.

The two figures given below make it possible to assess the influence of the substitution of a carbonyl by a sulfonyl in acetyl malononitrile (figure 3.32) and its fluorine equivalent (figure 3.33). In the first case, the conductivity is close for the two salts, but the presence of sulfonyl dimimulates the melting temperature of the electrolyte for a composition O / Li = 8/1. For fluorinated products, the presence of sulfonyl allows eg ~ l ~ m ~ nt to extend the area where the electrolyte is amorphous, and also leads to an increase ~ tior ~ of the conductivity by a factor of about two.

1 o-2,, I,,,, 10 3 O ~
E 1 ~ 3 ~ O / Li = 12/1 1 - ~ E 1 0 4 r ~ g E 1 0'4 o _ o '~ D 10 5 o ~
o CF3COCLi (CN) 2 10 7 - o LiNCN) 2 ~
c 10 6 - o CF3SO2NLiCN la O 1 0 ~ ~ o CH3COCLi (CN) 2 10 7 ~ 10 9 2.5 3 3.5 2.5 3 3.5 1000 / T (K) 1000 / T (K) Figure 3.32: Comparison between CH3CO and CH3SO2 Figure 3.33: Comparison between CF3CO and CF3SO2 malononitrile for a c ~ "" ~ sitien at O / Li = 8/1. malononitrile for a co "", osition at O / Li = 8/1.
We then wanted to compare the conductivity of the anions derived from cyanamide and malo ~ ollitrile As illustrated in Figures 3.34 and 3.35 given ~ Rs below for electrolytes of compositiQn.CO/Li = 12/1, the cyan derivative of cy ~ nimi-le has a conductivity equivalent to acetylmalononitrile, and the trifluorometh derivative ~ nesulfonated to that of trifluoroacetylmalononitrile We can conclude that the g- ~ up ~ lllent malononitrile allows better dissociation of lithium salts in the POE for a sl) bstib) ~ nt given compared to the cyanimide group.
1 o-2 ,,, ~ 10 2 E103 ~ O / Li = 12/1 ¦ _ E10'3 ~ ~ ¦0 / Li = 12 / 1¦ -o 10 4 - o - o 10 4 o 0 5 o - ~ 1 o'5 -o CF3COCLi (CN) 2 "~ C ~ 3CCcLi (cN) 2 o 1 ~ ~ O CF3SO2NLiCN _c 1 0 - OC ~ 39t) 2NLiCN b ~
107- '''' I '''' ~ 107 '''' I
2.5 3 3.5 2.5 3 3.5 1,000 / T (K) 1 OOOIT (K) Figure 3.34: conl ~ ardison between CH3COCLI (CN) 2 Figure 3.35: Comparison between CF3COCLi (CN) 2 and andLiN (CN) 2 for a cen "~ o ~; tion 0 / Li = 12/1. CF3SO2NLiCN for a con" ~ os; tien O / Li = 12/1.
Figure 3.36 gives the conductivity respectively of the lithium and potassium salt of tricyanorn ~ th ~ ne for the same composition O / M = 14/1. The salt of po ~ csium of tricyanométh ~ ne is certainly more conductive than lithium salt at ~ ll-pel ~ tures below z 60 ~ C, and has thus a conductivity - 10-5 nl.cm-l at 25 ~ C. On the other hand, it is interesting to note that the salts lithium and potassium, when the electrolyte is in a molten state, have the same conductivity which seems indicated that in these con ~ itions the lithium salt of the tricy ~ name ~ th ~ is whole ~, .- slow ~ absorbed in the POE; this assumption is consistent with the pKa ~ -5 determined for the tricy ~ nom ~ th ~ ne relative to water.

10 3 1 o-2 E10 ~ ~ C) 3CLi¦ _ E10 , 1o-6 ~ 10 5 - (CF3so2) 2NL
o O / Li = 14/1 ~ ~ ~ CF3So2cLi (cN) 2 10 ~~ O / K = 14/1 - O 10 0 LiC (CN) 3 o-8- ~ ~ lll - 107-2.6 2.8 3 3.2 3.4 3.6 3.8 2.5 3 3.5 1000 / T (K) 1000 / T (K) Figure 3.36: Comparison between the conductivity Figure 3.37: C ~ i "f ~ on between the conductivity of lithium salt and pQ's-e ~ rn of tricyano-LiTFSI at O / Li = 8/1 and LiC (CN) 3 and i "~ V, donkey for a composition O / M = 14 / 1.CF3SO2CLiC (CN) 2 for c ~" ~ ositi ~ n at O / Li = 12/1.
It is eg ~ lpm ~ nt possible to refer to the acidity measured in phase g ~ 7 ~ use to correlate it with the conductivity obtained for the different anions, also we have grouped in table 3.13 the dissociation enthalpies, experiment ~ le for TFSI, and estimated for tricyanométh ~ ne and le trifluorométh ~ nesulfonyl malononitrile [Koppel], as well as the respective conductivity optimums at each of them in Figure 3.37.
Table 3.13 Acidity in gas phase of TFSI and derivatives of malononitrile, according to [Koppel].
Acid HC (CN) 3 (CF3so2) 2NH CF3SO2CH (CN) 2 ~ Gd, SS (kcaVmole) 293 291.8 282 In view of its results, the low enthalpy of gas phase dissociation of these acids indicate that they are probably fully dissociated in the molten electrolyte, however, this parameter is not snffi ~ nt to predict the conductivity of lithium salts. Indeed, even if the tricyanométh ~ does not have an enthalpy of dissociation in the ga7ous phase equivalent to that of TFSI, its conductivity is two to three times lower. This difference can be justified by the flat and symmetrical structure of the tricyanom ~ th ~ ne which rigi (probably unifies the electrolyte. On the other hand, trifluorométh ~ neslllfonyl malononitrile, despite a lower dissociation enthalpy 10 kcaVmole to that of TFSI, has a conductivity equivalent to or less than this.
From these r ~ slllt ~ tc, it seems that the TFSI, among all the anions considered so far, provides the electrolytes with the largest amorphous domain, and the best conductivity for a composition at O / Li = 8/1. This result is linked to the conjunction of an acidity suffl ~ mme ~ t high to obtain lithium salts fully dissociated in the POE, and flexible loop of the anion which allows to plasticize the electrolyte.

. ~ _ 1 o'3e,. ~ ~
b1 o - ~ 0 / Li = 12/1 ¦ -0 ~ 5 ~

0 6 - o UN (CN) 2 ~ \
0 7 ~ o UC (CN) 3 ~ -. 1 0 2.5 3 3.5 1 OOO / T (K) Figure 3.38: Influence of the anion center between Figure 3.39: Influence of the center al,. - r.-, e between , ~ s, ~ ti "~ ement N-X2 and CX3 for X = CN, lespe ~ ement N X2 and C-) ~ 3 for X = CF3SO2, for a co. "f ~ ositi ~ na O / Li = 12/1. for O / Li = 8/1 (LiTFSI) and O / Li = 10/1 (LitriTFSM) Figure 3.38 compares the conductivity of the li ~ hium salt, for a composition with O / Li = 12/1, bis (cyanoimide) and tricyanomethane. This ultimately amounts to assessing the influence of the anion center on the dissociation of salt, for a given s ~ lbstitll ~ nt, in this case a nitrile function. It appears that unlike the anions substituted by CF3SO2, that is to say respectively the anion of TFSI and the anion of triTFSM, whose conductivity for their optimum respective composition is given on figure 3.39, one observes a clear increase in conductivity passing from imide to trisubstituted methane. This difference is explained by the larger steric hindrance of the trifluorometh ~ nes ~ -lfonyl group, relative to group ~ l, ent cyano. Indeed, in the case of triTFSM, it is difficult to place the three groupings CF3SO2 symmetrically in the same plane, geometry favorable to the stabilization of the anion. The distortion relative to flatness results in less acidity than that at which one could expect by simply adding the cumulative effect of three substitutions, we generally speak of a saturadon effect to qualify this phenomenon [Koppel]. However, the small volume of the group cyano allows you to place the three substitll ~ nts in the same plane, and this symmetrically. It results in a higher acidity of tricyanomethane relative to bis (cyanoimide) which justifies the results observed in conductivity.

To finish with the anions derived from malononitrile, the mixture of the lithium salt poly (acetyl malononitrile methacrylate), which we will symbolize to simplify a ~ 3 by PMAMLi, with a poly (ethylene oxide) matrix, is an example of ~ n the application of the chemistry of malononitrile to the particular field of electrolytes at c = o grafted anions. Even if this anion is not grafted directly onto the POE chain,, ~
this electrolyte can be considered as an electrolyte with a unipolar collection, of ANION
made of the low diffusion coefficient of the polyanion, relative to that of the cation of PMAM
lithium within the solvating m ~ tnce. Figure 3.40 gives the conductivity of the salt lithium poly (m ~ th ~ acetyl malononitrile crylate). As we can note, the grafted lithium salt of acetyl malononitrile has a conductivity of 10-5 Q ~ l.cm ~ l at 60 ~ C; which is after all a very honorable value for a grafted anion, with regard to the ease of synthesis of this polyelectrolyte. In addition, the anion being immobilized, one can expect less reactivity of the nitrile functions with respect to the lithillm electrode, and even consider the use of such a polyelectrolyte only in the cathode of batteries to limit the harmful effect of gr ~ dients of concentrations ~ pp ~ r ~ iss ~ nt cssentiellement in this area.

10-4 ~ ~ ~
PMAMLi ¦ -E 10 5 ~ ~.

O ~ _ \
~ 10 6 ~

~ o O / Li = 8/1 ~ -~ 10 7 - OO / Li = 12/1 ~
~ O / Li = 16/1 ~

1 o-8 2.6 2.8 3 3.2 3.4 3.6 1000 / T (K) Figure 3.40: Conductivity curve of the tithium salt of the grafted acetyl malonitoneum.
Finally, Figure 3.41 compares the conductivity of the lithillm salt, for a composition with O / Li = IV, acetyl malononitrile, relative to its grafted counterpart. To clarify the things, the conductivity values for the grafted anion are multiplied by a factor of five, which roughly fits away from conductivity between the free anion and the anion 10'3 '''' I
grafted when the electrolyte is fully --- 1 0-4 - ~ ~ O / Li = 12/1 ¦ -molten. However, below the point of _y ~ x 5 electrolyte fusion which is E 10 \\
approximately at 50 ~ C, the conductivity ~ 1 o-6 4 of the grafted anion electrolyte decreases more slowly than that with free anion, this c ~ 10 oPMAMLi phenomenon probably related to 8 1 o-8 ~ o MeCOCLi (CN) 2 \ ~ -kinetics of recrystallization ~ ation plus 10 9 slow. So at room temperature 2.5 3 3.5 the grafted anion electrolyte is approximately a 1000 / T (K) Figure 3.41: Comparison of salt conduction order of magnitude plus conductive which of acetyl malononitrile and PMMA.
the free anion electrolyte.

_ II-3-2-3 Conclusion on open anions:

An analysis of data from the literature concerning the basicity of anions of the family of meth ~ nes disubstituted and imides, allows to correlate the properties of the lithium salt of the bis (trifluorom ~ -h ~ neslllfonimide) in polymer electrolytes, on thermodynamic criteria rel ~ tively simple from the physico-chemistry of ~ protic solutions.

These criteria confirm p ~. r ~; ~ ... ent the hypotheses having led to the development of this salt, a ten years ago, and in particular, the importance of using base-derived anions soft deloc ~ licées which are best solvated in aprotic dipolar solvents, among which poly (ethylene oxide) which interests us more particularly.

It is thus possible to find substitutes for the anion of bis (trifluoromethanesulfonimide) based on a chemistry relatively easy to implement, and derived from precursors of the great organic chemistry such as trifluoromethanesulfonamide, cyanamide, and malononitrile. For each of these precul ~ iuls, it is possible to obtain anions whose salts of lithium has a conductivity close to that obtained for the anion of the bis (trifluoromé-h ~ nesulfonimide), the lithium salt of which is ~ ctuellemer-t used in the context of various industrial research programs aimed at developing batteries with lithium with polymer electrolytes.

From a more fundamental point of view, the study of the influence of the electron-withdrawing power of the substituting on the conductivity of lithium salts derived from malononitrile, shows without ambiguity that this, for a given family of anions, is directly linked to the dissociation of salt, in other words to its basicity, and what such conn ~ ics: lnces from the .
chlmle orgamque.

Finally, we must unfortunately deplore, at the moment, the lack of data thermodynamics available in the literature, concerning the acid-base properties in medium aprotic of these anions. These data will probably allow, once their knowledge has been extended, to refine the concepts re ~ iss ~ nts m ~ c ~ nicmes of salt conduction in polymer medium.

II-3-3 Cydic anions:

II-3-3-1 General presentation:

Anions derived from heterocyclic compounds, and more generally cyclic, have no practically not been studied with a view to developing new anions for polytrophytes polj; n ~ r ~. Ur.e sculc e ~ iuipc s'., S. in; c.cs ~ c ~ the conductivity of imida; zole e ~ de ccr.ai..s de sc ~, derived in POE IWnght1,2]. The chemistry of heterocyclic compounds being particularly varied, it seemed interesting to us to evaluate the potential interest of these anions in electrolytes polymers in ETU (li ~ nt the col ~ dllctivity of different compounds of this family in the POE.

Among all these species, the anions deriving from cyclic compounds with five or six atoms content ~ nt 4n + 2 electrons 7 ~ are assumed to be stabilized by resonance [Bo ~ wel ~ 7]. This energy of aromatic stabilization (ESA) of the anion, of the order of 10 to 20 kcaVmole, is characterized among other things by a diminlltion of its b ~ cicity, relative to its non-cyclical equivalent. We have illustrated in Figure 3.42 this concept in the particular case of cyclope.nt ~ di ~ ne:

Cy ~ n R ~
CHz = C ~ CHz - C ~ CHz ~ ~ ESA = 2,3 RT ~ pKa - 23 kcal / mole pKa ~ 35 pKa = 1 8 Figure 3.42: Example of ~ eneryeti4ue by cyclization of a molecule.
As we can co ~ ct ~ ter, the resonance energy of the cyclopçnt ~ does not allow to decrease its basicity -17 pKa units in DMSO relative to 1,4-pentadiene. R
However, this anion is still too basic to consider obtaining ~ Rz lithium salts dissociated in POE medium. In order to make it less basic, it F ~ ~ R3 is necessary to substitute all or part of the hydrogen atoms with ~ r grouping.s allowing dl ~ eccentuer the deloc ~ li.c ~ tion of the anionic charge cYcLOpERN4TADIENE
(Ri = CN, CF3, CO2Me, F "..), as illustrated by the pKa values of such SUBSTITUTED anions given in table 3.14.

These acidity values call for some comments. First, it is surprising to note that the pKa of pent ~ fluorocyclopent ~ di ~ only decreases ~ 2 units reladvement at cyclopent, ~ iene; this only illustrates the ambivalent character of the atoms of auors (and halogen in general) which are electron attractors ~ ar inductive effect ~ via the bonds ~, but which perhaps potentially ~ electron donors by resonance, via ~ bonds, as the confirms the values of ~ 1 and C ~ R given ~ Rs in Table 3.15. These two parameters measure in effect the relative share of the inducdf effect (~ 1) and the resonance (~ R) for a given substibl ~ nt. Of more, ~ SR is generally close to the value of ~ m, and the sum of ~ 1 and ~ R actually corresponds to the value of ~ p IMa ~ hl, and this in agreement with the results relating to the equation of ~ mmett (Cf II-3-2-l).

- 2 1,941 27 Table 3.14 pKa of the acids con ~ -gués of anions deriving from cy ~ p ~ "~ adiene.
R pKaH2 ~ Medium Rér ~. ~ .. this R1 ~ = H 1 ~, 5 / [Paprott]
R1 s = F. ~ 14 Deuterated solvent [Paprott]
R4 = CN 9.78 H20 [Webster3]
R3 = R5 = CN 2.52 H20 [Webster3]
R1 = R2 = CN 1.11 H2O [YVebster3]
R1 = ~ 2 = Fl4 = CN R3 = CH3 -5.7 CH3CN [Webster3]
R1 = R2 = R4 = CN -6.1 CH3CN [Webster3]
R3 = R4 = R5 = CN - -7.5 CH3CN [Webster3]
R1 = R2 = R3 = Rs = CN -9 CH3CN [Webster3]
R ~ = CN ~ -11 CH3CN [Webster3]
R1.s = CF3 <-2 H2O / H2SO4 [Laganis]
R1.5 = CO2Me ~ -7 H20 [Cookson1]

Table 3.15 Inductive and lesional attracting effects of certain groups, according to [March].
C; ~ 'JI ~ R ~ I ~ R ~ p (C5'l + C ~ R) F 0.5 -0.31 0.19 Cl 0.46 -0.18 0.28 Br 0.44 -0.16 0.28 CF3 0.42 0.08 0.5 CN 0.56 0.08 0.64 CO2 and 0.2 0.16 0.36 The structure of pentafluorocyclopentadiene is unfavorable in this sense due to the overlap possible between the 7 ~ orbitals of fluorine atoms, on the other hand their substitutions by groupings CF3 essPnti ~ llPment attractors by inductive effect allows to obtain a cyclopçnt molecule ~ dièn ~
which is the strongest known conjugate-free carbon acid to date [Laganis]. Acids even stronger are obtained by substituting the protons with groups. ~ carbomethoxy or cyano which allow to delocalize the load by inductive effect and by resonance, the pentacyanoplopene being the strongest carbon acid known to date. The estimated pKa value for pentacarbomethoxycyclopentadiene may appear weak in view of the attractor effect of groupem.o.nt CO2Me relative to that of group CF3 or CN, however, the relocation ~ li.s ~ tion of the oxygen charge of the carbonyl functions probably acts in favor of a effioa ~ e solvation of the anion in an aqueous medium. In contrast, the pKa of pentacyanocyclopen relative to water, is d ~ nnin ~ in con ige ~ nt the value obtained in acetonitrile ~ 9 pKa units ~
either the difference measured for the acidity of dicyanocyclopentadiene between water and acetonitrile; the pentacyanocyclopent ~ di ~ .ne being a very delocated soft base ~ ée ~ it is therefore likely that the pKa value relative to water is under ~ stimulated 2 1 9 ~ 1 27 -Tetracyanocyclopentadienide is a potentially more interesting molecule than pentacyanocyclopent ~ enide, due to the low basicity of the anion which allows to consider obtaining very dissociated lithium salts in POE medium, and by the possibility of making substitutions electrophiles on the molecule. It is thus possible to obtain from this molecule, the derivatives nitro, amino, chloro, acetyl, trifluoromethanesulfenyl tetracyanocyclopentadiene, and especially synth ~ tize from the derivative .; a ~ etylé, vinyltétracyano ~ yclopentadienid ~ ~ Webster4]. We obtain easily a polyelectrolyte soiubie in acetonitrile in ch ~ lff ~ n ~ the vinyl derivative at ~ 140 ~ C under argon. It is then possible to mix the lithium salt of this polyelectrolyte with POE, which should thus behave like a very dissociated fixed anion, but above all to graft this molecule on a modified POE containing allyles polym ~ -ri.c ~ bles functions. So even if there is a doubt about the stability of the cyano functions with respect to a lithium electrode, the immobility of the anion must guarantee the chemical inertness of the grafted salt, while combining the advantages inherent in the use of purely cationic conduction polymer electrolytes on the performance of catteries. On the other hand, the oxidation stability of the anion must be potentially largely su ~ -ieur at 4 volts, the anionic charge being very delocalized and therefore not very oxidizable.

Data from the literature on anions of the cyclopentadiene family confirm a priori our interest in anions derived from cyclic molecules with aromatic characters. As well, will we m ~ inten ~ nt present the information available on the azoles family with characters aromatic, i.e. essentially to five rings containing only carbon atoms and nitrogen. The generic elements of this family are obtained by gradually replacing the carbon atoms of cyclopen ~ dièn ~ - by nitrogen atoms, either succesively: the pyrrole tlN), imidazole, (2N), pyrazole (2N), triazole (3N), and tetrazole (4N), the structures of which are illustrated in Figure 3.43.
R ~ AA N - N ~ ~ ~
~ N ~ ~ NH ~ N ~ NH 6 ~ N ~ N'N ~ NN ~ N, N
HHHH
PYRROLE IMIDAZOLE PYRAZOLE 1,3, ~ TRIA ~ OLE 1,2,3-TRIAZOLE TETRAZOLE
Figure 3.43: Family of azoles composed of five-atom rings.
For all these azoles, it is possible to write mesomeric forms making it possible to delocalize the anionic charge on all nitrogen atoms, as shown in Figure 3.44 for 1, 3, 4-triazole.

N - N N - N ~ N - N

Figure 3.44: Resonance form of 1,3,4-tri ~ oiQ
As shown by the pKa values given in Table 3.16, the ~ acicity of azoles ~ iminl ~ e ~ 4 units every time we replace a carbon in pyrazole with a nitrogen atom, this di ... in ~ l ~; we r ~, sult ~ nt in part of the more electronegative character of nitrogen relative to carbon. The 2 ~ 94127 less basicity of 1, 2, 3-triazole, with respect to 1, 2, 4-triazole, results from the elimination ~ tion in the more basic pyrrol nitrogen cycle. In addition, the acidity measured in the DMSO (remove it ~ 4 units (l ~ pKa) relative to that obtained in water, tr ~ dui ~ nt lower solvation energy azoles in DMSO from 4 to 8 kcaVmole.
Note that in the azole family, pentazole is a molecule too unstable to be isolated, but we can however estimate its pKa ~ O ~ z 0, g lCata ~
Table 3.16 pKadans H20 and in DMSO pi ~. To azoles to c ~ ld ~; teres ~ loll.,. ~ Es Hél ~ keys Pyrrole Imidazole ~ l ~ E 1,2,4-Triazole 1,2,3-Triazole Telr__nl ~
pKaH2O 17.51 14.4 14.21 10.04 9.26 4.9 pKaDMSO 23 18.61 19.84 14.75 13.93 8.23 ~ PKa 5.49 4.21 5.63 4.71 4.67 3.33 Let’s take a closer look at the effect of substitution on the acidity of azoles, we’ll start in this by tetrazole which has the advantage of being able to carry only one substituent (R) on the cycle. We have shown in Figure 3.45, the evolution of pKa relative to water [Ca ~ al ~ l in function of the ~ p of the substitution ~ n ~ lM ~]. In agreement with what we exposed above ~ lemm ~ nt on the anions of the TPSM family, it might be better to use the values of ~ p ~, however the use of ~ p perrnet already to obtain a good correlation according to the equation:
pKa = 4147-5.3 ~ p We can deduce from this equation a value Me- '' of 5.3 for the coefficient p of the equation of H \ ~ Ph Hammett which is to compare with the value l 2.2 for the substituted phenols in the 2 ~ CF3 Br same conditions [Bordwell31. Tetrazole is N o _ therefore more influenced than phenols by Y NO2 ~ \
substituting, this probably being to - 2 -correlate with a better deloc ~ li.cation of 4 _ \ Nz + _ load. In addition, the value of the coefficient p depends on the nature of the solvent and thus passes 60, 5 0 0, 5 1 1, 5 2 2.2 to 5.3 of an aqueous solvent with DMSO ~ P
Figure 3.45: Influence of s ~ hst ~ f on acidity for phenols. If we extend this result to t ~ relative to water.tetrazole, it is likely that we can obtain relatively strong acids in DMSO, in s ~ lbssitu ~ nt tetrazole by groups very au. ~ u ~ s of electrons such as: CF3, CN, CF3CO, or CF3SO2.
n is also possible to colupafcr the acidity of azoles fully s ~ lbstihlés by a gloup ~ luent R
given, i.e. compare the effect of a monosubstitl ~ tion on tetrazole relative to a double substitution on the triazole, and so on. For this purpose, we have grouped in the table 2 1,941 27 3.17 the pKaH2O available for tetrazole, sorting ~ 7.ole, and imid ~ 7ole. We also have indicated, in italics, the difference in acidity (~ pKa) reduced to an attracting group, and this relative to the pKa of non-substitute azole ~ - ~. It should be noted that unlike the pKa d ~ tPrmin ~ for malononitrile derivatives in m ~ l ~ nge water / dioxane 1/9 vol. (65.5 mol% dioxane), the PKa of bis (trifluoromethane) triazole determined in a water / dioxane equivolumic mixture is ~ s.simil: -klc has its valcur d ~ s water, dioxane does not represent [while only 17.4% of the molar fraction.
This r ~ sultaL is comïrmé by tait that the pH of a solution of perchloric acid in a mixture water / dioxane 1/1 vol. only varies at m ~ ximllm by 0.2 pH units, and this at infinite dilution [Shukla].
Table 3.17 pKa ~~ of different azoles Sl ~ hStitlléS ~ from [Catàlan].
Br CF3 NC NO2 CF3S02 ~ N ~ ~ N ~ ~ N ~ ~ N ~ ~ N ~
N ~ N, NN, NH ~ N N'H ~ N
PKa 2.13 1.7 0.7a -0.83 -1l15a Ka 2.77 3.2 4.2 5.73 6.05 N - N N - N NC CN N - N
Br ~ Br F3C ~ CF3 ~ 6 o2N ~ N9 N ~ 2 PKa 5.17 3b 1.47 -0.63 Ka 2.44 3.52 3.9 5.34 NC ~ ~ CN

~ 9 d _5 ~ ~ 3 ~ -4 PKa 2.1c ~ PKa 4, 1 a: estimated from the relationship: pKa = 4.36 - 5.23 ap.
b: determined in ", mixture 50/50 vol% water / dioxane, according to [Brown].
c: from [Apen].
d: M. Armand, Personal communication As we can cost .ct ~ ter, according to the available values, the decrease in pKa induced by a g ~ or ~ ... ~ -nt given is substantially equivalent whatever the azole considered. This result allowed to estimate the value of pKa ~ ~ ssPntie1lPmPnt for imidazole derivatives, ad-litionn ~ nt the effect observed for a given grouping in tetrazole. From all of this data, we give on figures 3.46 and 3.47 the evolution of pKa as a function of ~ p of the substitute ~ n from which we can deduce the equations of ~ mm. ~ tt for imi ~ 7.ole and triazole:
Imidazole: PKa = 14.1 - 18.2 ~ p Triazole: pKa = 9.3 - 11.4 ~ sp 21 94i Z7 Estimating pKa values assumes that it is possible to place two substituents on the tn ~ 7: olP, and especially three on the imid ~ ol ~ o without inducing a steric constraint between the substitutes, which çntra ~ me ~ it a ~ n ~ l; one of the stability of the anion, and this in particular for CF3SO2.

12 e '''''' 3 p.m.
10 - ~ H
8 - \ - lo -Br ~ ~ Br ~~
Q 4 C ~ ~ Y 5 - CF3 2 - CN O \ CN
~ ~ N0z \ ~ \ ~ n N0z ~ \
-4 '''''5 1 1 1 1 1 ~
-0.2 0 0.2 0.4 0.6 0.8 1 1.2 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 ~ p ~ P
Figure 3.46: Influence of suhstitllAnt on Figure 3.47: Influence of 5l ~ hstit ~ Ant on the acidity of triazole with water. The acidity of i ".. '- ~' e relative to water.
Figure 3.48 allows you to compare the evolution of acidity as a function of the parameter ~ p of the equation of H ~ mmett respectively for imidazole, triazole, and tetrazole snbstit ~ lé As shown this figure, the slope p of the Hammett equation gradually increasing from 5.3 to 18.1 in passing from tetrazole to imidazole, the order of acidity for these azoles depends on the attractiveness of the substinl ~ nt So for the substituents with a ~ p <0.75 such as halogens, CF3, or CN, the acidity order is such that pKa (imidazole) <pKa (triazole) <pKa (tetrazole), this order being reversed ~
for the very attractive substituents present ~ nt a ~ p> 0.75 such as NO2, or CF3SO2. According to these résl.lt ~ ts, we can assume that the acidity of these azoles substituted by CH3SO2 (~ p = 0.73) should be practically identical and such as pKaZl.

Imidazole Triazole \ \ Tetrazole Y ~
o- ~

-0.5 0 0.5 ~ 1 1.5 2 Figure 3.48: Influence of s ~ hsti ~ A ~ It on the acidity of i "'- ~ -.' E, triazole, t zl, In order to complete these results ~ ts, Figures 3.49 and 3.50 show the influence of a double substitution on the acidity and basicity of pyrazole respectively in the gas phase and in water lEI9u ~. As we can see, the p ~ r ~ meter p of the equation of ~ mmett for the acidity of pyrazole is 5 2.7 times greater in g ~ 7ellse phase than in water, this result is compare with the factor 5.6 determined previously for the substituted phenol between the values in ph ~ se g ~ zcuse and d ~ ns e ~ u.
400,,,,,,,, 20, I
350 _ 3 H ¦Pyrazole ¦ 15 - ~ ¦Pyrazole ¦
~ 53 ~~ - PA - 31.7 ~ 10 - p = 11.6 ~ 3 ~, 250 -CH H p = 25.2 cY ~ 5 ~ ~ = 18.9 200 - ~ B
<I ¦ + ~ G basic I CF ¦ + pKa basic \ _ 1 50 - I ~ G aade ¦ -5 - I pKa acid 3 100 '''''''-10 -0.2-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6-0.2-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 ~ p ~ P
Figure 3.49: Influence of 51'hS ~ itll ~ rlt on Figure 3.50: Influence of s ~ hstitu ~ t7t on the acidity of triazole relative to water. The acidity of i "; L '- ~' 8 relative to water.
It is interesting to note that the parameter p of the Hammett equation describing the acidity of the pyrazole in water (p = 11.6) is roughly equivalent to the parameter determined for triazole under the same conditions (p = 11.4); on 5 can therefore assume that the effect of the, ~, Pyrazole / H 0 J ' solvation on these two acids is identical 2 - O PhénoUDMSO
in an aqueous medium. In addition, the information available on pyrazole, carrying different - ~
substitute ~. ~ nts, allow to show, as ~ H / ~ G) d, T, so = 0.96 illustrated in Figure 3.51, that the entropy I - 5 ~ ~ '-represents a small part of its energy '1,' ionization (~ 15%), unlike ~ (~ H / ~ G) = 0.86 phenol in water. These data are -10 (~
-15 -10 -5 o 5 10 referenced to the AGo energy values (kcal / mole) ionization and acid enthalpy, Figure3.51: By ~ del'e "lhr / ~ edanslenergy ~. ionization of pyrazole in water, comparatively respectlvement for pyrazole and phenola the share for phenol in DMSO.
unsubstituted This result is not surprising, pyrazole being a soft base, the charge of the anion does not imply a structuring important water molecules; in the same way as the part of the entropy of phenol in the DMSO is negligible due to the lesser structuring of this solvent compared to water.

To en ~ e .... i - er with pyrazole, the values grouped in table 3.18 allow to evaluate the inflllçnce of the sllbstitlltion of grou ~ e ~ ts CH3 by CF3 groups on its acidity in ph_se ~ 7ellce and in the water. It appears that the effect of substitution is practically additive in phase g ~ 7ellce ~ acidity (liminl1 ~ nt ~ lS kcaVmole by CF3; even if the effect is less additive in water, acidity decreases ~ 3.5 pKa units per CF3 to +/- one unit.
Table 3.18 Influence of the substituents on the ~ G ~ j (in kcaUmole) and the pKa of pyrazole, according to [Elguero].
Substitute R ~ G ~ G PKa ~ PKa Rl = Rt = CH3 348.1 0 15 0 Rl = CH3 R2 = CF3 333.6 -14.5 12.3 -2.7 Rl = R2 = CF3 317.4 -16.2 7.5 -4.8 The app ~; nte additivity of the effect of an attracting group on the acidity in the gas phase allows to estimate the acidity in phase g ~ 7 Chip of 3,5-bis (CF3SO2) -pyrazole [El9Uer ~ l, this group present ~ nt gener ~ lemert an effect twice as important as CF3 on acidity. So we end up with a estim ~ tion of ~ GaC ~ 295 kcaVmole, a value between the acidity in ga phase ~ use of TFSM (aG = 301.5 kcaVmole) and TFSI (~ G = 291.8 kcaVmole), pyrazole being an intermediate, the load being carried by a carbanion placed in a resonant structure. We we also wanted to estimate the acidity in the gas phase of 3,5-bis (CF3) -triazole, conn liss ~ nt its lower basicity in the gas phase relative to ammonia (~ Gbas = l9s ~ 6kcavmolet ~ ippin9]) from -50.2 kcaVmole [EI9Uer ~], we end up with a / ~ GbaS = l4s ~ 4 kcaVmole, i.e. 41.5 kcaVmole less than the ~ GbaS of bis (CF3) -pyrazole. If we assume that this difference is maintained on ~ GaC, we can estimate the ~ GaC of 3,5-bis (CF3) -triazole at 283.3 kcaVmole, i.e. a gain in terms of acidity ~ l0 pKaH2O units, hence a pKaH20 (bis (CF3) -triazole) ~ -2. This value differing from S pKa units from that determined experimentally ~ ~ ~ it is likely that the basic behavior of bis (CF3) -triazole differs from that of bis (CF3) -pyrazole in phase g ~ 7Pllce- It is therefore preferable to start from the difference in acidity observed in water ~ 4.5 pKa units ~ hence we can deduce a lower acidity in the gas phase ~ 17 kcal / mole, i.e.
~ Gac (bis (CF3) -triazole) ~ 300 kcal / mole.

Examples of weak basic anions derived from heterocycles are not limited to five-cycle atoms, as illustrated by the example of the cyano derivative (C7N7H) [Wiley] and trifluoromethane (C7FgN4H) [~ nl du tétr ~ 7 ~ pçnt ~ lène, of which we give the respective pKa in figure 3.52.

C7N7H N ~ CN pK ~ H ~~ = 3 C7FsN4H N ~ pK --2.75 Figure 3.52: pKa of the cyan derivative and peffluon "" éV, ~ le of tetrazapentalene.

219 ~ 127 , It should be noted that it is not possible to delocalize the load on all the nitrogen of the t ~ tr ~ 7 ~ pent ~ lan ~
substituted by any g ~ upe ~ slow R, nor on the sub ~,!; lu ~ nt R placed on the triazolo cycle for groul) e ... entc capable of dispersing the charge by reson ~ nce, such as nitriles, as illustrated ~ /
Figure 3.53. However, the trisubstituted derivatives ~ N ~
t ~ r ~ zap ~ n ~ alene are more acidic than lcurs ~ N ~ 3 ~ 3Nb ~ 2 ~ R
counterparts lli7U'0 ~ 7iiiUt ~ i U ~; liV ~ 7 U ~; illliU ~ lG (~ \ / R
this tr ~ dnis ~ nt an extension of the deloc ~ lic: ltion Figur ~ 3.53: Resonance form i "" ~ ass' ~.
of the anionic charge passing from the nucleus imid ~ 7Ole to the tetrazapent nucleus ~ lane In addition, chlorine re ~ gics ~ nt on C7N7- to give a chloro-C7N7 which gives back C7N7H on contact with water, we can deduce an oxidation potential of the anion ~ 4.5 Volts vs Li + / Li ~.

We also focused on the cyclic anions of the acid family barbiturate. Indeed, it is possible to obtain strong acids by substituting the labile proton present on the carbon of barbituric acid by a ~
g ~ upelllent R attractor, for example the nitro derivative (R = NO2) has a Me ~ N ~ N'Me pKadmr, o = 0.8 [Bordwell6]. In addition, it is possible to play on the length of ~ O
alkyl chains to induce potel ti ~ llPment of surfactant or R properties pl ~ ctifi ~ ntes to these anions with respect to POE chains, this concept having Acid ri ................................................. .................................., ~ tl "~ lGarbiturique already highlighted for the anions of the family of tetraalkyls suhs ~ itu ~
borcher lcherkaoui]
The interest in cyclic anions is not limited to the possibility of obtaining dissociated salts in dipolar aprotic medium, but also chemistry allowing ~ to yield to these materials, this one offering diverse and varied ways of synthesis, as we will show later. he it is possible to apprehend this problem in two ways, either by treating the main possibilities synthesis by anion family (cyclopentadiene, imidazole, triazole, tetrazole, ...), or focusing on anions derived from different pr ~ c ~ ll ~ ul ~ of great organic chemistry. These two present approaches: lnt an interest in our eyes, we commlonces by describing the anions by f ~ thousand, then in a synoptic way the possibilities of syntheses offered from a given precursor. We we are not limited to do this to fully substituted anions ~ which do not therefore more labile protons on the cycle, by glvl ~ pe ~ ntc attractors potentially ~ stable at electrochemical point of view, but also to the cycles on which it is possible to envisage modifi ~ tion.c simple chemical p ~ rm ~ tt ~ nt to obtain such deloc ~ anions licensed.

2194t ~ 7 . ..
II-3-3-2 General presentation of the heterocycles synthesis:

Cyclopentadiene family: The mono, di, and tricyane derivatives of cyclopent ~, are not synth ~ ticked by the action of cyanogen chloride on the cyclopent ~ di-ne in the presence of hydride sodium capable of extracting labile protons from the aromatic cycle. Cycle responsiveness to su ~ electro stitutions ~ hiles dim.inu2nt ~ with Ic num de grou ,, cmcn.t ~ n ~ s ce c ~ nticnt, Ics derivej ieli-, l et pen; ~ cy, ul ~; s du ~ y ~: lupe ~ die ~ 'o'O ~ mellt p, lr l'd ~ L; ul ~ d ~; Chii '~; ii ~; ulijullciiùll with aluminil-m chloride [Websbr3l. The tetracyanocyclope, nt ~ liè, rlP, is obtained é ~ l ~, m ~ nt from cyclopentane- 1, 2, 3, 4-tetracarboxylic by conversion of the carboxylic groups into nitriles, then dehydrogenation of the cycle by PCls [Cookson2l ~ as illustrated in Figure 3.54:

? ~ PaS ~ ~ POCI3> ~ PCIs ~ ~
HO2C-- co H NH3 H No ~ CONHz NC CN ~ CN
Figure 3.54: Synthesis of tetracyanocyclopentadiene.
As we have already mentioned, the possibility of making electrophilic substitutions on the carbanion of tetracyanocyclopent ~ e allows to synth ~ tick several derivatives [~ J ~ L 11, including we have grouped the essentials in Figure 3.55, vinyl tetracyanocyclopent ~ not present ~ nt our eyes most of interest for the synthesis of electrolytes with purely cationic conduction.
NC CN NC CN

N ~ CN ~ N ~ CN

NC ~ CN
Cl2 ~ ~ 6 NC ~ y - CN
NC ~ CN Cl N ~ CN CF3scl ~ ~ N

N CN

NC ~ CN NC ~ CN NC ~ CN
Ac20 ~ D ~ NaBH4 ~ ~ 6 (1) SOCI2 / Pyridine ~ ~
CF3CCkH N ~ CN N (~ CN (2) Pyridinel ~ N (~ CN
C = O H - O - OH

Figure 3.55: S ~ hs'ihnion éle., ~ R ~, ph 'e surl'anion du pentacy, l ~ pe "~.': Èr, e.

2i9 ~ 127 ._ Amine-tetracyanocylopent ~ can not ~ ~ lçmellt be obtained from tetracyanobutè.ne ~ liide, derived from tetracyanoethylene, which cyclizes after protonation in the form of the amino-pentacyanocyclopent ~ di ~ .ne, eliminating ~ nt thereafter one of its cyano groups by hydrolysis in concealed HCl followed by decarboxylation [WebsteR] ~ as illustrated in Figure 3.56.

NC CN ~ NC ~ CN
N ~ CN 6N ~ N ~ 12N N ~ CN
NC CN CN
NH7 N ~
Figure 3.56: Synthesis of am; no tetracyanocyclopentadiene.
The usual derivatives of the reaction are accessed by diazotization of this compound by nitrous acid.
S ~ n ~ meyer [~ revet WebSterl, and in particular to another synthesis of tetracyanocyclopentadiene as shown in Figure 3.57.
NC ~ CN
CuCN ~ ~
NC ~ CN NC ~ CN N ~ CN
~ HN02 ~ ~ _ CN
N ~ CN NC ~ ~ ~ CN
I + NC ~ CN

NH2 ~ N ~ CN

Figure 3.57: Synthesis of pentacyano and iodo cyclopentadiene.
The synthesis of perfluoromethylcyclopent ~ diene is also described in the literature [La9anis], however this is carried out with poor performance ~ (~ 10%) by the action of trifluorodiazoethane on perfluorotetramethylthiophene, desulfurization with triphenylphosphine, and pyrolysis.

To finish with the anions of the & mille of cyclopentadiene, let us mention the synthesis of penta-carbomethoxycyclope.n ~ n ~ ILe Gofq from dimethylmalonate and the acetyl diester ~ .ne-lic ~ rboxylate as illustrated in Figure 3.58.

002Me C02Me R ~ R MeCO2 CO2Me Cl Pyridine ~ MeC02 ~ H20 ~

CD2Me 2 MeCO2H R ~ ~ R ~ C0 M
Figure 3.58: Syn ~ hese du perlt2 ~ a ~ xy ~ yel ~ pentddWnQ

21941 ~ 7 -Azole family:

Pyrrole: Among the pyrrole compounds, two pr ~ se.nt ~ rt of interest for the synthesis of anions, it it is respectively tetracyanopyrrole (pKaH20 = -2.7) synthesized from tetracyano-1,4-dithiin [] (Figure 3.59), and perfluorotetramethylpyrrole obtained from ??? [Silvesterl from the reaction given in figure 3-60-NC ~ S ~ CN NC ~ CN F3C CF3 F3 C CF3 N3 ~ ~ CF3CF = C - C = CFCF3 Nucleophile ~
NC ~ S ~ CN NCf ~ N ~ CN R2NH F3C-- ~ N ~ --CF3 - HH
Figure 3.59: Synthesis of tetracyanopyrrole. Figure 3.60: Synthesis of perffuoromethyl pyrrole.
Imidazole: Many delocalized anions of the imidazole family are NC CN
synth ~ ti.cés from ~ minomaléonitrile (DAMN) IVJ ~ L l-RI, this molecule ~
being the HCN tetramer obtained by basic catalysis; this leads to H2N DAMN NH2 different 2-substituted derivatives of 4, 5-dicyanoimi ~ 7.ole, as illustrated in Figure 3.61.

RC (OEt) 3 H2N ~ R = Me PhOMe NC CN And o DDt ~ H2N ~ G ~ N = CH - R R - CH H2N ~ NH2 Alkyls CH3CN NC CN C ~ 3CN NC CN CF3 NC) [~ NCS / ~ le / DMF NC ~ CN ~ R ' HCOOH H2N ~ R- H
Diglyme NC CN
RO
Butanol H2N NHc = cHco2Et RC ~ CH2C02Et H2N NH2 ¦ Alkyls NC CN ~ P20s / EtOH NC CN R = ¦Phenyl Figure 3.61: Synthesis of derivatives 2-s ~ h ~ és of 4,5-dicy ~ noi "'~
The first molecules of this family were thus obtained by reaction of DAMN with the methyl orthoformate (HC (OEt) 3) or different ortho esters lU ~ - 'i Thereafter, a route of synthesis from aldehyde [Be9 ~ ed allowed to extend the possibilities of synthesis, the basis of Schiff resulting from the addition of the aldehyde to the DAMN being cyclized by oxidation by the dichlorodicyanobenzoquinone (DDQ). The use of N-chlorosllcinimide, more oxidizing than DDQ, coupled with nicotin ~ mitle allows these compounds to be obtained in a single step l0ht ~ ka1l. It is eg ~ lP.m ~ t possible to condense the B-keto esters (RCOCH2CO2R ') with the P205 / ethanol mixture, ~ 1,941 ~ 7 then cyclize the ~ n ~ mine obtained at reflux of butanol lOhtsuka21. The reaction of DAMN with acid formic reflux of the diglyme (Teb ~ 160 ~ C) results in an optimization of the synthesis of dicyanoimitl ~ 7.ole [Ohtsuka3]

Another important derivative of DAMN, 2-amino-4, 5-dicyanoimid ~ 7.ole, is obtained by reaction of DAMN with ClCN. By dia_otation of this compound, it is possible by the reaction of Sar ~ m ~, r to obtain among others, the iodo [Webs Patent]], cyano lApen, Webs Patent], and phenyl O dicyanoimidazole, t ~ i as described in Figure 3.62.

NC

NC H

NH2 ~ 2 ~) ~ ~ N + CuCN

~ ~ [~ IR
Figure 3.62: Synthesis of 2-sll-ctitués derivative of 4 5-dicyanoi ".i ~ 'e from the diazol derivative.
As shown in Figure 3 ~ 63, the reaction of orthophenylenediamine with acid trifluoroacetic in an acid medium allows syn ~ h ~ tick 2-trifluorométhylben7.imid ~ 7.01e which is oxidized by hydrogen peroxide to 2-trifluoromethyl-imid ~ 7.ole-4, ~ dicarboxylic lSmi ~ BelCher]. Per action of SF4, the carboxylic groups are converted to trifluoromethyl to result in perfluoromethylimid ~ 701e 0.

~ NH2 ~ HO2C F, ~ H
Figure 3.63: Synthesis of perHu ~ rJ ", ~ ll" / l jlnj; the The bromination of imidazole (Figure 3.64) or dicyanoimid ~ 7ole (Figure 3.65) allows synthesize respectively tribromoimidazole, and 2-bromo-4, 5-dicyanoimidazole ~ Apen], on which we can hope to make the nucleophilic substitutions specific to the molecules with characters aromatic.
H 13 H ~? ~ ~
Figure 3.64: Synthesis of Iril ~ u ~. . "idazole. Figure 3.65: Synthesis of bromodicyan ~ le.

Pyrazole: As we can see in figure 3.66, we can synthesize tricyanopyrazole by con ~ enc ~ ion of an acetylated derivative with a diazo compound, transformation of the esters present on the cycle in ~ mides ~ followed by their dehydration by P2O5 [Weis].

NC ~ CCN NC ~ CN
+ ~ ~
N2CHCOOAnd H NC2 ~ CN

MeOOC COOMe ~ N ~ CN
MeOO ~ COOMe ~ ~ H
+
N2CHCN N ~ CN
Figure 3.66: Summary of tricyanopyrazole.
Triazole: The addition of nitrous acid to the DAMN, as described in Figure 3.67, allows the synthesis of 4,5-dicyanotriazole [~ T ~ U ~ R ~ ere ~ kl. This acid is the first known example of a very delocalized organic anion of aromatic character, as attested by the electrical conductivity of its sodium salt in water IFialkof ~ proving that this compound is dissociated more than 90% in dilute solution.
NC CN
H2 ~ e ¢ HNO2 ~ h ~

Figure 3.67: Synthesis of dky ~ n ~ tr The cyclization of bis (tri ~ luoroacetyl) hydrazinelBr ~ Wn1ChamberSAbdul ~ Ghanilpar P205 allows to obtain 2,5-bis (trifluoromethyl) -ox ~ 7ole [Charnbers ~ Brown1] whose cycle is sufficient.
reactive to nucleophiles to be opened by anhydrous ammonia. By dehydration of compound obtained, we end up with 2, S-bis (trifluoromethyl) triazolel ~ wn2 ~ Tippins ~ Abdul-Ghani] such that described in Figure 3.68. This acid is particularly thermally stable, in fact the product sealed under vacuum in a test tube is not altered after several hours at 350 ~ C. Plus the acid is not hydrolyzed in concentrated sulfuric acid, nor oxidized by potassium perm ~ ng ~ n ~ e acid medium, which suggests that the oxidation potential of the anion is greater than 1.3 V / ENH (i.e. = 4.3 Volts vs Li + / Li ~).

N Nl lZ ~ ,, ~ CF3 - ONN C - CF3 2 5 ~ F30 ~ CF3 OO

N - N P20sH H NH3 anhydrous F30 ~ CF3 CF3 - ONN O - CF
H NH O
Figure 3.68: Synthesis of bis (triflu ~, u ", ~ tl, ~ l) h 3z ~

Another interesting compound of the triazole family is obtained (Figure 3.69) by the action of a azide on acetyl ~ ne ~ ic ~ rboxylate, or its diester, followed by the hydrolysis of esters in the latter case [Ch ~
Me4C ~ CO2Me HOeC ~ CO2H
HO2C - OC - CO2H NaN3 N ~ ~ N Hydrolysis ~ N ~ ~ N

Figure 3.69 Synthesis of dicarboxy triazole.
Finally let us mention the synthesis of 2-amino-5-trifluoromethyltriazole by azeotropic distillation of a m ~ l ~ nge of acetic acid and aminog ~ l ~ ni ~ inP in toluene ILoPYreVI, as illustrated in the figure 3.70. This synthesis has the advantage of being relatively simple, and of arriving at a product inexpensive potential. In addition, it is possible to envisage obtaining different derivatives thereof by diazotization, like aminoimid ~ 701e, and in particular 2-cyano-5-trifluoromethyltriazole:

I k ~ r ~ l ~ c ~ NH ~ h upe ~ N9--Figure 3.70: Synthesis of the aminotriflu ~ run, ~ / l triazole.
Tetrazole: Tetrazole and most of its derivatives are synth ~ tic ~ s by the action of an azide on a nitrile group. Nitriles substituted by an attracting group presents a special case, these groupings making attack of the azide easier, and make it possible to carry out the reactions with moderate temperatures, thus obtaining tetrazoles substituted by groups aryles [r ~ l, trifluorom ~ th ~ ne [Norrisl, and cyano [Oliveri ~, as shown in Figure 3.71. The cyanotetrazole is also obtained by a complex reaction based on simple and inexpensive products expensive lr ~ ~. schPm ~ ticked in Figure 3.72.
R ~ SO4 NC
P ~ N ~ ~ ~ NR = Ph ~ ls + MnO2 + CuSO4 ~ N ~ ~ N
H HCOOH H
Figure 3.71: Synthesis of tetrazole s ~ hstitué Figure 3.72: Synthesis of cyano tetrazole.

Tetrazapentalene: The reaction of potassium cyanide - CN H CF3 H
ssium with cyanogen leads to the formation of a ~ _N ~ ~ _N ~
cyané bicycle, le lH-imi ~ 1 ~ 7.0 [1, 5-b] s-triazole- ~, N_N ~ Nq ~ N ~ CF3 -2, 5, 7-tricarbonitrile [wileyl ~ commonly symbolized CN CF3 by HC7N7. Identically the reaction with trifluoro-C7N7H C7F9N4H
acetonitrile leads to the synthesis of 2, 4, 6 Figure 3.73: synthesis of estelrdzapenldlene ~
tris (trifluoromethyl) -l, 6a-dihydro-1, 3, S, 6a-letf ~ aL) P ~ ne; ~ "~ 1, which we will symbolize by thereafter by HC7FgN4. Figure 3.73 shows the two derivatives of léll ~ apçnt ~ lene obtained.

2 ~ 941 21 . ..
II-3-3-2 Synthesis of anions:

Famile du., ~ Clo ~, ... ladiene:

~ Pentamethyl cyclopentadiene-1,2,3,4,5-pentacarboxylate (PMCPH) is a product commercial (Aldrich, 97%). The lithium salt is obtained by treating this product with an excess of lithium carbonate (20% excess) in water, the solution immediately takes on a coloring intense red. After evaporation and drying, the product is taken up in acetonitrile, filtered, and evaporated (yield - 50%). The weak rendem.qnt is probably explained by a reaction of Knoevenagel between a carbonyl and the carbon carrying a labile proton in basic medium, also did we check the purity of the lithium salt by NMR in D20 / CD3CN. 1 H NMR: a peak (CO2CH3) at 3.6 ppm. 13C NMR: a peak (CC) at 169.353 ppm, a peak (CO2Me) at 117.855 ppm, a peak (CO2CH3) at 52.661 ppm.

Imidazole family:

~ 2-bromo-4, ~ -dicyqnc; ~ 7Ole: 0.127 mole (15 g) of dicyanoimidazole (TCI America, 95%) and 0.14 mole (19.4 g) of potassium carbonate is mixed in a beaker containing 250 ml of water, we then observe a release of carbon dioxide concomitant with the formation of the salt of dicyanoimidazole potassium. After placing the beaker in an ice bath, add by portion (for ~ 15 min) 0.127 mole of bromine (- 6.6 ml), there follows a release of CO2 in the following minutes. After one hour, 2 cc of bromine and 7 g of potassium carbonate are added.
After one night, isopropanol is added which makes it possible to eliminate the potassium hydrobromide present in the solution, then the water is evaporated on a rotary evaporator, and the synthesis product dried at 50 ° C.
under vacuum. Once dry, the product is taken up in methyl formate, filtered to remove the KBr formed during the reacdon and the potassium salt of dicyanoimidazole which may not have reacted.
We have indeed verified that this salt is insoluble in methyl formate. After evaporation solvent, and drying at 50 ~ C under vacuum, the product is recreated ~ t ~ llicé in water (~ 1 g / cc) We obtain thus ~ 28 g of the potassium salt of bromodicyanoimi (l ~ 7ole (yield ~ 94%). The lithium salt is obtained by ether extraction of an aqueous solution of potassium salt, acidified with 4M HCl in stoechiomé.trie After evaporation of the ether, the acid is taken up in water, treated with an excess of Li2CO3 (20% excess), evaporated, dried, taken up in methyl formate, filtered, and finally dried.

~ 2-chloro-4,5-dicyanoin ~ idazole: In order to avoid the manipulation of ga_eux chlorine, we used as a substitute for the sodium salt of dichloroisocyanuric acid, which is a chlorinating agent electrophile fa ~ ilement manipulable. The reaction takes place in water at low temperature (~ 0 ~ C) in p-ésellce of sodium carbonate (the choice of the base being important because of the pH which it imposes solution), as described in Figure 3.74.

o. o C 1I NC ~ CN ~ NC ~ CN
~ N ~ NNa) ~ 1 ~ 0 ~ N ~ NNa ~ 0 ~ J Igl ~ N ~ O ~ N
Figure 3.74: Summary of the sodium salt of chloro dicyan ~, ~ 7ole.
After one night, the solution is centrifuged, the water evaporated, the product dried, taken up in the formate of methyl, filtered, and dried again. This gives the sodium salt of chlorodicyanoimi ~ 7Ole (makes ~ 75%). In order to purify the product, we repa ~ sse with pot salt ~ ssinm, after extraction of acid with ether, then it is recrict ~ llized in water. The lithium salt is then obtained, after a new ether extraction, following the same operating mode as for the brominated derivative.

~ 2-trifluoromethyl-4,5-dicyanoimidazole: The reaction described in the literature ~ B ~ g ~
being carried out in two stages with a yield ~ 25%, we have considered other ways of synthesis inspired by the operating procedures developed for other derivatives of dicyanoimidazole.
First, we tried to get the product at reflux of an acid mixture trifluoroacetic and DAMN in the diglyme: we mix in - 40 cc of diglyme, 0.034 mole of DAMN (3.718 g) and 0.069 mole of trifluoroacetic acid (4 cc), then this solution is brought to the reflux. n appears after ~ 1 hour a black precipitate in the solution, after ~ 24 hours the solvent is evaporated, the product obtained is then stirred with ether, the solution filtered, this operation being repeated three times. The different ether fractions are mixed, the ether evaporated, and the product obtained taken up in water and treated with 0.02 mole of lithium carbonate (1.48 g). We thus obtain the lithium salt of trifluoromethyldicyanoimid ~ 7Ole with a yield ~ 60%, however, it is necessary note that the lithium salt probably forming a complex with the diglyme, the salt is pasty and therefore still contains a fraction of solvent.
We also tried to cyclize DAMN and trifluoroacetic acid by P2Os in acetonitrile: 0.03 mole (3.243 g) of DAMN and 0.03 mole (3.42 g) of trifluoroacetic acid and 0.06 mole (8.7 g) of P2Os, added per serving, are mixed with ~ 50 cc of anhydrous acetonitrile, and this solution brought to reflux under argon overnight. The acetonitrile is then filtered to to eliminate the black precipitate present in the solution, evaporated, then the product is taken up in water in order to hydrolyze the remaining P2O5, and the aqueous solution extracted with ether. We recover ether 1 g of trifluoromethyldicyanoimi ~ 7.ole (yield ~ 18%), then transformed into its salt pot ~ Ccillm- The same reaction carried out under reflux of ethanol results in a ren-lem ~ nt ~ 25%.
In any case, an insoluble carbon appears during the reaction, so we have eg ~ ltoment tried to start from trifluoroacetylamide in order to no longer carry out the synthesis in medium acid: l ~ ntemerlt (~ 20 min) is added 0.0216 mol (3 ml) of trifluoroacetic anhydride to 0.0216 mole of DAMN in solution in anhydrous acetonitrile at 0 ~ C. After ~ = 2 hours, the solvent is evaporated ~, and the 4.4 g of the product obtained (yield qu ~ ntit ~ tif) are dried under vacuum at 50 ~ C. This compound is then brought to reflux of a solution ~ 50 cc of diglyme under argon for ~ 24 hours, a black precipitate appears again in the solution. After evaporation of the solvent, the product is taken up in ether, then the ether solution filtered, and evaporated. We move on to salt jar ~ csillm by adding 0.012 mole (1.7 g) of potassium carbonate to the water, and excess carbonate is then removed by taking up the salt in acetone. The pasty product obtained is then washed several times with ether, armed with eliminating traces of diglyme, then we go back to lithium after extraction with ether. We end up with the lithium salt of trifluoromethyldicy ~ noimi ~ 7ole with a yield ~ 30%.
We also tried to perform an aeotropic lictillion of a mixture of DAMN, trifluoroacetic acid, and pyridine (3 times in excess of the acid), there is a again the appearance of an insoluble charcoal.
The salts obtained being all blackish, we have combined all the fractions, extracted the acid with ether, and sublimed the acid obtained at 100 ~ C under vacuum. We then obtain a white crystallized product which is transformed into its lithium salt The purity of the lithium salt is confirmed by carbon NMR and fluoride in CD3CN. 19F NMR: a peak (CF3) at -59.4914 ppm. 13C NMR: a quadruplet (C-CF3) to 150.357; 149.595; 148,847; 148.085 ppm, a quadruplet (C-CF3) at 128.532; 123.182; 117,841;
112.491 ppm, a peak at 120.037 ppm (C = C), and finally a peak at 115.160 ppm (C-CN). The microanalysis of the lithium salt gives the following proportions: C 35.15% H 1.56% N 26.09% F
24.51% (C6N4F3Li: C 37.53% N 29.18% F 29.68% Li 3.61%), the deviation from stoichiometry probably corresponds to the monohydrate of the lithium salt (C6N4F3LiH2O: C 34.31% H
0.96% N 26.67% F 27.13% O 7.62% Li 3.3%).

~ 2 ~ 4 ~ 5-lr; ~ o-imidazole: Aminodicyanoimid ~ 7ole acidified with an equivalent of HCl (4M) is diazotized with an equivalent of potassium nitrite at 0 ~ C. A precipitate of diazodicyanoimidazole, after 15 min an equivalent of NaCN and two equivalents of CuCN, it is then released from the nitrogen in the solution. After one night, add half an equivalent of carbonate of po ~ csillm allowing to precipitate a copper carbonate, then the solution is evaporated, the dried product, taken up in methyl formate, and dried. The acid is extracted with ether, then transformed into its lithium salt (yield ~ 30%).

Fan ~ Ille du Triazole:

~ Dicyanotr; ~ ole: The synthesis is carried out in accordance with the procedure described in literature 1 ~ T ~ - ~. 0.185 moles (20 g) of diaminomaleonitrile (Aldrich, 98%) are put suspended in 200 ml of water placed at the temperature of an ice bath. We then add 0.18 moles of potassium nitrite (Janssen), then the solution is brought to a pH of 2 by adding slowly 5 cc of 4M hydrochloric acid. After one hour, diaminomaleonitrile has passed solution, and the solution brought to a neutral pH by adding bicarbonate of pot ~ sillm ~ filtered on paper, acidified with 46 cc of 4M HCl, and extracted several times with ether. The dirrélelltes 21 9412 ~ ~

ether fraction are then evaporated using a rotary evaporator, after being filtered through paper. After secll ~ ge overnight at room temp ~ ure under vacuum, 18.77 g of the acid dicyanotriazole are obtained (rtondem ~ nt 86%). The present product has a slight yellow coloring, it is then sublimated under vacuum at 100 ~ C, the acid is finally obtained in the form of crystals whites with a yield of 75%. First, a spectroscopic analysis infrared acid shows the ~ bs ~ ~ -this of a band relative to the carbo ~ amide which could have resulted hydrolysis of cyano functions. The result of the microanalysis of the sublimed acid, given below, confirms the purity of the product obtained:
dicyanotriazole C4NsH analysis C: 40.44% H: 0.75% N: 58.78% C: 40.34% H: 0.85% N: 58.81%
1 H NMR in CD3CN gives a peak at 9.5 ppm corresponding to the acidic proton of the molecule.

~ 2 years ~ ino-5-trifluoromethyl-triazole: 0.22 mol (z 17 ml) of acid is added dropwise 0.2 mol trifluoroacetic acid (27.2 g) of aminoguanidine bicarbonate in 50 ml of toluene. A
once the deg ~ g ~ m ~ nt of CO2 ceased, the solution is brought to reflux in a dictation device ~ tion azeotropic. After one night, the toluene is evaporated and the product dried for 24 hours at 50 ~ C
under vacuum. It is then washed with several ether fractions, thus obtaining 28.6 g of amino trifluoromethyltriazole (yield - 94%). A fraction of the product is puri ~ ed by sublimation at 100 ~ C under vacuum, its NMR spectrum in CD3CN proves its purity. 1 H NMR: a peak (C-NH2) at 5.4 ppm, and a peak (-NH-) at 11.1 ppm, the peak ratio being from 2 to 1. 19F NMR: a peak (CF3) at -61.914 ppm. 13C NMR: a quadruplet (C-CF3) at 153.082; 152.318; 151.555; 150.801 ppm, a quadruplet (C-CF3) at 128.738; 123.404; 118.063; 112.734 ppm, and a peak (C-NH2) at 158.851 ppm.
Microanalysis gives the following result: C 23.36% H 1.85% N 36.24% F 36.82% (C3N4F3H3: C
23.69% H 1.99% N 36.84% F 37.48%).

~ 2-cyano-S-tritluoromethyl-triazole: The amino trifluoromethyltriazole acidified with an equivalent of HCl (4M) is diazotized with an equivalent of potassium nitrite at 0 ~ C. It quickly appears precipitate of diazotrifluoromethyltriazole, after 15 min an equivalent of NaCN and two CuCN equivalents, it then releases nitrogen from the solution. After one night, add a half equivalent of potassium carbonate allowing to precipitate a copper carbonate, the solution is evaporated, the product dried, taken up in methyl formate, and dried. The acid is then extracted with ether, then sublimed, and finally transformed into its lithium salt (yield - 30%).

~ 1,941 27 Tetrazole family:

~ P-nitrophenyltetrazole (Lancaster, 97%) and 3, S-bis (trifluoromethyl) tetrazole (Maybridge) are commercial products, the latter has been sublimated under vacuum at 100 ~ C before use. The lithium salts are prepared in a conventional manner, except that the lithium salt of paranitro phenyltétrl ~ ole e ~ t taken up in an ethanoVacetonitrile mixture, being insoluble in acetonitrile.

~ S-trifluoromethyltetrazole is a commercial product (Aldrich, catalog of rare products):
1.81 mmol (250 mg) of trifluoromethyltetrazole are treated with 1 mmol of carbonate lithium in water. After drying ~ ge and overnight at 50 ~ C under vacuum, the salt is taken up in acetonitrile, filtered, and evaporated. This gives ~ 250 mg of the lithium salt with a yield of ~ 96%.

~ 5-trifluo ~ ac8 ~ 1tetrazole: This compound is synthesized by the action of a stoichiometric amount of trifluoroacetic anhydride on sodium cyanide in cold THF, to form cyanide trifluoroacetyl. After ~ 30 min of reaction, a st ~ chiometric quantity of azide is added.
of sodium, and the solution is then stirred at room temperature for ~ 48 hours. After evaporation of the solvent, the product is taken up in water, acidified by an excess of 4M HCl (50% in excess) and extracted with ether. We then return to the lithium salt in the manner already explained above.

Barbituric acid family:

~ 1,3-Dimethylbarbituric acid is a commercial product (Aldrich, 98%), its salt is obtained pot ~ c.sium by shaking it with an excess of potassium carbonate (20% in excess) in ethanol for ~ 48 hours, filtration of the solution and evaporation of the ethanol. Lithium salt is obtained with lithium carbonate (20% excess) in water, then the excess carbonate is removed in taking up the lithium salt in an equimolar mixture of ethanol and acetonitrile.

~ 5 trifluo ~ etyl ~ dimethylbarbituric: the potassium salt of dimethylbarbituric acid is treated with 1.2 equivalents of trifluoroethyl trifluoroacetate in the DME. Potassium salt insoluble in EMR gradually dissolves during the reaction, once all dissolved product (~ 24 hours) the solution is evaporated, thus obtaining the potassium salt of the acid trifluoroacétyldiméthylbarbiturique with a yield qu ~ ntit ~ tif. The lithium salt is then obtained by ether extraction, tr ~ item ~ nt with Li2CO3, and the product taken up in acetol-itrile.

dibutyl-2-barl;; lul; que: We m ~ l ~ nge an equimolar amount of dibulylul ~ e and dichloride of malonyl in dichlorom ~ th ~ n ~ "we observe a release of concomitant HCl ~ nt in the reaction.
After 48 hours, the solution is evaporated, and dibutylbarbituric acid is obtained with a r ~ ndemtont qu ~ n ~ if ~ The lithium and potassium salts are obtained from their carbonates ~ spec!; rs, whose excesses are ~ limin ~ s by taking up the salts in a m ~ l ~ nge eth ~ nollacetonitrile ~

~ 1,941 27 dibubl-2-sulfonylbarbiturique: We add the ~ t ~ ment an equivalent of sulfuryl chloride in a solution of dichlorom ~ th ~ ne at 0 ~ C, conttqn ~ nt 4 equivalents of butyl ~ mine After a night, the solution is evaporated, then the product taken up in heptane, and filtered in order to ~ liminsr butylamine hydrochloride formed during the r ~ action This gives dibutylsulfonylurea with a return of 80%. Thereafter, the procedure is the same as for the dibutylbarbiturate.

._ Il-3-3-3 Conductivity of cyc anions:

Cyclopentadiene family anions:

PMCPLi: The conductivity of the lithium salt of pentamethylcyclo- MeO2C CO2Me pe.n ~ .n ~. 1, 2, 3, 4, 5-pentacarboxylate (PMCPLi) is given on the MO c ~ CO2Mefigure 3.75, for dcs compositions at C, ~, 12, e; 1 C ~ 1. romme nous + '~
can the const ~ t ~ r, the conductivity obtained is poor and does not reach a 'CO2Me value of 10-5 Ql.cm-l only for a temperature above 120 ~ C. PMCPLi ~ o-2 ~ ¦ PMCPLi ¦

E 10 4 - ~ _ O 1 0 ~ = = = ~

".7: 0 O / Li = 8/1 \\\ ~
10 - OO / Li = 12/1 8- ~ O / Li = 16/1 ~ 3 - E ~
TFSI O / Li = 8/1 - ~ 3 10 9 '''' I
2.6 3 3.5 1000 / T (K) Figure 3.75: Conductivity curve of the lithium salt of pentacarbomethoxy cyclopentadiene.
At first glance, this low conductivity may seem strange with regard to the pKa of the acid of the pentacarbomethoxy cyclopentadiene 1 o'4 in an aqueous medium, which is estimated <-7 (Cf II-3-3-1, table 3.14). In 'E 10-5 - ~ ¦ O / Li = 8/1 ¦ ~
done, this example illustrates perfect- - "-the concepts introduced E 1 o-6 ~
prev ~ emm ~ .n ~, on the influe.nce of the ~ - 1 0'7 - \ _ nature of the solvent on the solvation. ' ~ 5 ~
anions. PMCP acid is ~ '1 o-8 0 (CH3SO2) NLi \ E ~ E -facil ~ mP.nt solvated in aqueous medium, ~ ~ PMCPLi which suggests that the 1 o 2 4 2, 9 load is essentially delo- 1000 / T (K) localized on the oxygen atoms Figure 3.76: Com ~ ison of the conductivity of lithium salts PMCPLi and (CH3SO2) NLi for a c ~ i "~, os; ti., na O / Li = 8/1.
present on the cycle. However, the ., poly (ethylene oxide) being a bad solvent for hard bases, such as oxygen atoms, PMCP lithium salt is visibly very little dissociated in such a medium. Figure 3.76 shows as well as the conductivity of the lithium salt of PMCP is close to that obtained for the salt of bis (methanesulfonylimide) lithium, for the same composition at O / Li = 8/1. In assuming that these two anions pre ~ ntçnt an equivalent basicity, we can estimate, from the hypotheses formulated above mmP.nt (Cf II-3-2-1), that the pKadmSo of PMCP acid is close that of bis (meth ~ nesulfonylimide) of pKadmS ~ ~ 15 - 9 = 6, or ~ 13 pKa units more than aqueous medium.

'' _ Fan anions ~ Ille de l'j ~

BrD (: Tm ~ The condu-; livi ~ of the lithium salt of bromo-dicyanoimid ~ 7ole is given in the figure 3.77, for compositions with O / Li = 8/1, 12/1, and 16/1. For a composition with O / Li = 8/1, the con curve; vile has an identical appearance to that corresponding to a law of free volume.
This is consistent with the visual observation of the electrolyte film, which is for this composition perfectly transparent after two weeks in glove box, characterizing the amorphous state of polymer. This composition must be close to the eutecti ~ read composition, as attested moreover the a - gm ~ n ~ ion of the melting temperature of the electrolyte with the salt dilution.
1o2 10 5 - o 0 / Li = 8/1 ~
- o O / Li = 12/1 ~ -1 o-6 - e O / Li = 16/1 - ~ TFSI O / Li = 8/1 \ ~

2.5 3 3.5 1000 / T (K) Figure 3.77: Conductivity curve of the lithium salt of bromo-dicyanoi "'~ e.
This result is consistent with the hypothesis previously expressed (Cf II-3-2-1), according to which the deloc ~ licée soft bases, of which the conjugated acids have a pKaH2O ~ 1, can be envisaged to obtain appreciably dissociated lithium salts in the poly (oxide ethylene). Indeed, bromo-dicyanoimidazole, which we can estimate pKaH2O 4 (Cf II-3-3-1, table 3.17), has a conductivity which is only a factor of four lower than five, temperature dependent, relative to that of the lithium salt of TFSI at O / Li - 8/1.

Finally, Table 3.19 gives the parameters corresponding to a free volume law for the BrDCImLi and LiTFSI at O / Li = 8/1. Note that the ideal glass transition temperature for BrDCImLi is thus = 14 ~ C greater than that obtained for LiTFSI under the same conditions.
Table 3.19 raldlll ~ 1st ~~ e the law of free volume pou LI ~ FSl and BrDClmLi ~ O / Li = 8/1.
Anion A (Ql.cm-l.Km) Ea (eV) T (~ C) BrDCIm 3.2 +/- 1.6 0.08 +/- 0.005 -56 +/- 6 TFSI 15 +/- 12 0.08 +/- 0.02 -70 +/- 1 1 ~. .
ClDC ~ The conductivity of the lithium salt of chloro-dicyanoimidazole (ClDCImLi) is given in Figure 3.78, for a composition at 0 / Li = 12/1, compared to that of the brominated derivative.
1 o 2 _103- o ~ ~~ O ~ O ~ XDCImLj¦ -- ~ 10-4 - ~
Q 1 0 5 ~ ~

, 10 ~ - OX = CI O / L; -12/1 \\ E3 ~
OOX = Br O / L; = 12/1 ~ 10 7 - TFSI O / L; = 8 1 OOO / T (K) Figure 3.78: Curve of conductivity of the thium salt of chloro-dicydr ~ ,,, '7th. *.
It is remarkable to note that the substitution of bromine by chlorine improves the conductivity by a factor of about two. This salt ultimately has a lower conductivity of only ~ a factor two relative to that obtained for the lithium salt of TFSI at 0 / Li = 8/1.
This increase in conductivity of the chlorinated product, relative to the brominated product, is probably related to the respective electron-withdrawing effect of these two halides. If we refer to the data relating to this subject in table 3.15 (Cf II-3-3-1), it appears that the coefficients of H ~ mmett of these two halogens are substantially equivalent. 1 o-2,,,, ~,. .
However, it should not be forgotten that - 1 o-3 - ~ ~ L ~
the resonance part depends on the recovery of the orbital 1 ~ of E 1 0-4 halogen with the aromatic cycle. ~ 10-5 Therefore, it can be assumed that this. ~ \ \
~ _ 1 o-6 ~ e ClDClm O / Li = 12/1 \ \
recovery is influenced by g \ \
~ ~ DMATFSI O / Li = 1211 1 nature of halogen, and may encourage ~ 10 7 1- TFSI O / Li = 8/1 ~ 3 synth ~ tick, thereafter, the derivative 1 oH-fluorinated dicyanoimi ~ 701e. 2, 5 3 3, 5 1 oOO / T (K) Figure 3.79: Co "" ~ a, ai ~ n of the conductivity of the lithium salt of As illustrated in fililage 3.79, the choro-dicyan ~, '.,.''-,' eetdudiméthylaminoTFSlàO / Li = 12/1.
conductivity of this salt is practi-cely identical to that obtained, for the same composition, with the lithium salt dudiméthylaminoTFSI, while p ~~ se ~ t the advantage of not using fluorine chemistry.

CF3DCImLi: The conductivity curve of the lithium salt of trifluoromethyl-dicyanoimi ~ 7ole (CF3DCImLi) is given in figure 3.80, for compositions with O / Li = 8/1, 12/1, and 16/1.
102,,, I

1 0 ~ 3 - ~ a ~ CF3DClmLi ¦

,. o - ~

~ O / Li-12/1 ~
o - o O / Li = 16/1 ~ ~
10 ~ TFSI O / Li = 8/1 ~ ~

.8,,,, I,,,, ~
2.5 3 3.5 1000 / T (K) Figure 3.80: Conductivity curve of the lithium salt of the trfflwr "" ~ dicy ~ noi ".: '- .'e.
It is interesting to note that the conductivity is identical for these three compositions, when the electrolyte is in a molten state. On the other hand, it is surprising to note that for the same composition at O / Li = 12/1, the conductivity of the lithium salt of dicyanoimidazole substituted by a trifluoromethyl group, is less than that for the molecule substituted by a chlorine, and identical to that of the molecule subs-bromine-like, as illustrated in 1 o-2 Figure 3.81 for the latter. In --1 o-3_ ~ ¦ O / Li = 12/1 ¦ _ assuming that this expresses the influence E ~
of the substituent on the dissociation of 'E 10 ~ p ~
salt, the information given in the ~ ~ o-5- ~
table 3.15 (Cf II-3-3-1), would tend to ~ ~ \
suggest that in this case the effect ~, 1 ~ o CF3DClm \\ ~ 1 of the substituent is described essen- O 10-7- 1 ~ 1BrDClm \ 9 -partly by the inductive part which ~ 8- ~ -is ~ specifcally equal to 0.42 for the 1 2, 5 3 3, 5 trifluoromethyl group, and 0.44 1 oo 0 / T (K) Figure 3.81: Com, comparison of the conductivity of the brominated derivative for bromine. This reasoning etdudérivétrifluorom ~ Jrlédudicyan ~. ',.''7. ~, aO / Li = 12/1 simple, whose validity, if it turns out exact, po ~ likely to be confirmed ~ by quantum mechanical calculations, does not explain however not the conductivity obtained with chloro dicyanoimid ~ ole. This family of salts deserve a complete study of their properties by developing their phase diagrams in polymer medium, and the determination of their pKa and DNX- ~ solv in aprotic medium.

2i 941 27 An ~ ions of the family of Ir. ~ ~ E-TFCTrLi: The conductivity curve of the lithium salt of trifluoromethyl-cyanotriazole (TFCTrLi) is given in Figure 3.82, for compositions with O / Li = 12/1, 14/1, and 16/1.
-2,,,, I

10-3 ~ Tr ~, T ~ Li I _ ~ 1 0-4 10 5 ~
~ OO / Li = 12/1 ~ OO / Li = 14/1 o ~ o O / Li = 16/1 ~
10 ~ TFSI O / Li = 8 ~ ~
-a,,,, I,,,, ~
2.5 3 3.5 1 OOO / T (K) Figure 3.82: Conductivity curve of the lithium salt of trifluoromethyl-cyanotriazole.

N - N Figure 3.83 compares the conductivity of the trifluoromethylated derivative of cF3 ~ e, ~ cN cyanotriazole, relatively to that of dicyanotriazole, for the same u + composition at O / Li = 12/1, this comparison drawing its interest from the fact that these TFCTrLi two salts contain each at least one 1 o 2 cyano grouping, and a ~ 3 ~ grouping trifluoromethyl, although in positions not E ~) ~ _ ~ _ identical. As this figure illustrates, ~ 10 4 the conductivity for these two salts is o .5-substantially equivalent. This result is in agreement with the proximity of the values of ~, 1 0 oTFCTr O / Li = 12/1 ~ -pKaH2O of the two acids as derivatives, that O 1 ~ ~ ~ ~ 3TFSI O / Li = 8/1 ~ 1 ~
we can estimate between 2 and 3 (Cf II-3-3-1, C) ~ 8. I
Table 3.17), and suggests, in the 1 ~ 2.5 3 3.5 present case, that the conductivity is 1000 / T (K) - Figure 3.83: C ~ ", f ~ r .i ~ n of the conductivity of the derivative essentially linked ~ the basicity of triflu:, rui ,, t ~ lédudicyanoi ".e etducy ~ nc.lriazole anions. However, the nature of the cycle aromatic bearing the subs ~ itu ~ nts ~ in other words the imid? 7ole or the tria_ole, do not seem to play here a big role on the conductivity of the lithium salts in derivatives in the poly (ethylene oxide).

DCTrLi: The conductivity curve of the lithium salt of dicyanotriazole (DCTrLi) is given on Figure 3.84, for compositions with O / Li = 6/1, 8/1, IV, 14/1, and 16/1.
1 o-2 E 1 ~ 4 ~ 10 5 - OO / Li = 6/1 '~ D e O / Li = 8/1 \' ~
o-6 0 O / Li = 12/1 ~
O / Li = 14/1 ~ \ ~
1 0 7 ~ ~ O / Li = 16/1 ~
TFSI O / Li = 8 ~ 3 0 ~
2.5 3 3.5 1 OOO / T (K) Figure 3.84: Conductivity curve of the lithium salt of dicydnolr _; 'e.
Unlike the previous examples, the lithium salt of dicyanotriazole has an optimum of particularly marked conductivity, and this for a composition with O / Li = 14/1. It is probable let this fact be linked to the structure of dicyanotriazole, which is a flat and symmetrical molecule.
For this optimum conductivity, this salt has a conductivity 1.5 to 2 times lower, in the temperature range included between 40 and 60 ~ C, at the optimum of 6 o A
conductivity of LiTFSI. This result is ¦ ~ Lixpt T = 80 ~ C v = 20mV / s particularly encouraging for the 2 0 ~ ~ _development of non-fluorinated substitute for ~ ~ ¦
TFSI, due to the simplicity of synthesis of '- 2 ~ ~ ~ ---- H2O
this molecule and its moderate cost. - 4 0 ~
- 6 0 - ¦ DCTr O / Li = 20/1 ¦ -Therefore, the study of the stability domain - 8 ~ ~ Li ~ + e ~ Li ~
electrochemical of dicyanotriazole present 1 0 0 0 1 2 3 4 a very special interest. The result of E (Volts) vs. Li '/ L i ~
cyclic voltammetry, obtained on a Figure3.85: Voltai "", ~ l, iecycliquedudicydn ~ zule.
platinum microelectrode with a diameter of 125 ~ 1, for an electrolyte with a composition at O! Li = 20/1, is given in Figure 3.85 [~ anl3], As we can coll-ct ~ ter ~ this anion has an oxidation potential of at least 4 Volts relatively to lithium, and seems stable in reduction with respect to a lithium electrode. This dom ~ ine of stability of at least 4 Volts is a priori sl ~ ffis ~ nt for the intended application, and justifies a cycling test of a battery using this anion as a salt, replacing LiTFSI.

Let's take a look at the fan:

Figures 3.86 and 3.87 pr ~ se.ntent ~ eelivell, ent, the conductivity of the lithium salt for a composition at O / Li = 8/1, lV1, and 16/1 of nitrophenyl-tetrazole (NPTétLi), and for a composition at O / Li = 12/1 for bis (trifluoromethyl) phenyl-tetrazole (bT ~ letLi). In both cases, the con. ~ u. ~, deliveries obtained are low, and reach a value of 10-5 Q ~ l.cm ~ l only for z 60 ~ C.
10'4 '''' I ''''10'4'''' I
1 oS ~ ~ -E 1 0 ~ bTFPTétLi ¦ -E 10 6 - \ - '1 o 6 - ~ _ ~ 10 7 ~ -, 10 ~: O 8/1 - ~
C ~ 1 o'8 - o 12 ~ 1 o-8 0 16/1 ~ ¦ o 0 / Li = 12/1 ¦

1 OOO / T (K) 1 OOO / T (K) Figure 3.86: Salt conductivity curve Figure 3.87: Salt conductivity curve of para-n ~ lithium. uf ~ hey, lyl-t ~ lr. ~ 'e. of lithium from bis (trifluoromethyl) phenyl-telr - ~' e Figures 3.88 and 3.89 show the conductivity of the lithium salt of the triQuoromethyl (CF3TétLi) and trifluoroacetyl-tetrazole (CF3COTétLi), for the same composition at O / Li = 12/1.
1 o 2,,,,,,,,,, 1 o-2 1 o 3 - ~ _ ¦ CF3TétLi ¦ _ 1 o 3 - ~ - ¦ CF3COTétLi ¦ -u 10 - ~ 10 . ~ 10 ~10 6_ \ -~ 1 0 7- \ ~ - ~ 10 7 - ~ -- o O / Li = 12/1 - ~ - o 0 / Li = 12/1 ~ -10 -TFSI O / Li = 8/1 1 0 3 - _ ~ TFSI O / Li = 8/1 \ ~

~ 2.5 3 3.5 102.5 3 3.5 1 OOO / T (K) 1 OOO / T (K) Figure 3.88: Salt conductivity curve Figure 3.89: Salt conductivity curve lithium triflu ~ run. ~ t ~ 'e.Iithium triflwr.Jacetyl- ~ elr. - 'e . ~
It may seem sul ~, renant, to note the lower conductivity of tetrazole substituted by a trifluoroacetyl group, with respect to that carrying a group trifluoromethyl, conn ~ s ~ nt the least attracting power of the latter. It 3 ~ Cprobable, that the trifluoroacetyl-tetrazole has such a geometry that it can ~, C - N "
form a five-atom ring (indicated in bold in the diagram ~ m ~ above) which chelates (~) N ~ Ne ~ N
the lithillm cation The resulting association constant decreases the CHELATE accordingly apparent dissociation of salt, this result could be confirmed, by a study of the acidity of this salt in DMSO, thanks to the methods developed for this purpose (Cf II-3-1-2).
..
Figure 3.89 compares the conductivity of trifluoromethyl-tetrazole to that obtained for dicyanotriazole, for the same composition at O / Li = 12/1, even if it does not correspond not at the optimum conductivity for dicyanotriazole, it seems more judicious to compare the respective conductivities under these conditions. This comparison is of particular interest, since dicyanotriazole and trifluoromethyl-tetrazole have a substantially pKaH20 equivalent -1.5 (Cf II-3-3-1, table 3.17). As a result, under these conditions, the conductivity of trifluoromethyl-tetrazole is less than a factor of two less than that of dicyanotriazole. he would be n ~ ce.s ~ ire to extend the range of compositions stu ~ i ~ es for trifluoromethyl-tetrazole, in order to determine whether the conductivity obtained at O / Li = lVl corresponds to the optimum for this salt.
1 0-3 ~ '1 0 3 2 / ~ ¦ 'E 1 o ~

~, o DCTrLi \ ~ -10 7- 0CF3TétLi - cl CF3TétLi ~ o 1 o ~ 8 - TFCTrLi ~ -~ 1 o-8 ~ ~ 1 0 2.5 3 3.5 2.5 3 3.5 1 oOO / T (K) 1 oOO / T (K) Figure 3.89: Co ,,,, v ~ r ~ ison of the conductivity of the salt Figure 3.90: Comparison of the conductivity of the salt lithium from Cf3TétLi and DCTrLi at O / Li = 12/1. lithium from Cf3TétLi and TfCTrLi a O / Li = 12/1.
Figure 3.90 shows the conductivity of CF3TétLi, relative to that of trifluoromethyl-cyanotriazole, for the same composition at O / Li = 12/1. The latter, despite a pKaH2O also close to that of tetrazole (~ 2 to compare to 1.7 for tetrazole) is practically three times more costly. This result poull ~ l suggest that the nature of the substrate in this case, but it would be risky to try to draw too hasty conclusions from it, considering the little thermodynamic data available on azoles s ~ lkstitués in aprotic medium.
To finish with tetrazole, it would probably be int ~ l to seek to synthesize the derivative trifluorom ~ th ~ nesulfoné, which should improve nettPm ~ nt the condu ~; livi ~ relatively to CF3TétLi.

~. 101 ~ 1 9 ~ 1 27 Fan anions of barbiluric acid:

MeBb: Figures 3.91 and 3.92 give ~ -nt respecLi ~ el "ent the conductivity of lithium salts and potassium of dimelhylbarbiluric acid (MeBbM), for a composition with O / Li = 8 / l, 12/1, and 1/16. Even if the con ~] vcliviles obtained are relatively low, and n '~ ttei ~ n ~ nt conductivity close to 10-5 Ql.cm ~ l than around 100 ~ C for the best of this family of anions is not devoid of illt ~ lGt, from the many po.ccihilit ~ s to play s., i the chemistry of these acids.
1 0'4 10 ~ 4 o-5 ~ MeBbLi ¦ _ ~ 1 0'5 -El o ¦ MeBbK ¦
E; ~ 10b ~

~ - O 8/1 ~ ~ 0 8/1 8 1 o-8- o 12/1 \\ -8 1 o-8- 0 12/1 \ ~
- o 16/1 o 16/1 \ ~

10 9 ~ 10 9 '''' I
2.5 3 3.52.5 3 3.5 1 OOO / T (K) 1 OOO / T (K) Figure 3.91: Conductivity of the lithium salt of Figure 3.92: Conductivity of the pot salt ~ s ~ iurn of dimethyl acid ~, b.itl ~ rique.l'acid dimét ~ ylL, arb, ~ rique.
First, it is surprising to note that the conductivity in lithium and potassium presents a very marked optimum for the composition at O / Li = 12/1, and especially that for this one, the lithium salt of dimethylbarbituric acid is more conductive by a factor of two than the salt of potassium. The same goes for the composition at O / Li = 8/1.

This result is surprising to say the least, potassium salts being generally more dissociated and less crosslinked energy ~ ire than their lithium counterparts and gives electrolyte films to the times more amorphous and more conductive. It is possible to advance the hypothesis that the displacement of the lithium is carried out, in this case, by a m ~ vehicle transport mechanism, to explain these conductivity values. The principle of such m ~ C ~ nism is illustrated by the equation below:
Vehicle transport: 2 MeBb ~ + 2 Li + () 2MeBb ~, Li + Li +

This amounts to considering that a fraction of the lithium is transported in the electrolyte in the form anionic, and therefore, with a higher transport number (t- ~ 0.7) than that relating to the cation lithium (t + ~ 0.3) moving by a mechanism ~ nicm ~ of solvation / desolvation along the propellers of poly (ethylene oxide). This hypothesis implies that the anion of dimethylbarbituric acid is .
capable of complexing the lithium cation, as illustrated in Figure 3.93, and ~ o thus masking the charge of the cation with respect to poly (ethylene oxide). Me ~ NJ ~ N, Me I el This assertion is not meaningless, since the two carbonyls and the ~, co acid carbon form a five-ring capable of complexing the cation (~) H
lithillm, and similarly to acetyl acetone which has a constant o ~, c ~ "o high association for lithium of pKaSs = i, 77 [~ o ~ welq. On the contrary, the c ~ c association constant is significantly lower for potassium, Me '~' Me pKasS = 4.77 18ordWelll. All this could explain in the end, the slightest o conductivity of the potassium salt of dimethylbarbituric acid, Figure 3.93: Con.r '~; on relative to the lithium salt. Again, it would be required ~,; re deLilparllanionduMeBb.
to carry out a study of the complexing properties of dimethylbarbituric acid in DMSO, to using the methods developed for this purpose (Cf II-3-1-2), in order to confirm the hypotheses put forward previousmP.nt The carbanion of dimethylbarbituric acid being sllffis ~ mment reactive to perrnettre nucleophilic substitutions on this substrate, Figure 3.94 gives the conductivity of the lithium salt of trifluoroacetyl-dimethylbarbituric acid 'N, ~ Li +
(TFAMeBbLi), for a composition with O / Li = 8/1, 12/1, and 16/1. The presence o = ~ e) - Icl - CF3 of the trifluoroacetyl group makes it possible to gain more than an order of magnitude / N ~ o on the conductivity obtained at 120 ~ C (- 2.10-4 Q ~ l.cm ~ l), relative to that TFAMeBbLi for dimethylbarbituric acid (= 2.10-5 Q ~ l.cm ~ l), and this for the optimums respective conductivity for each of these salts, i.e. at O / Li = 12/1 for the lithium salt of the acid trifluoroacetyl-dimethylbarbituric, and at O / Li = 8/1 for dimethylbarbituric acid.
10-3 '''' I '''' 10-3 - '~ 4 ~, ¦TFAMeBbL ~ O / LI = 8 / 1¦
- 10 5 ~ - 10 - ~ -10 O 1 6/1 ~ CF3COMeBbL;
1 ~ 2 5 3 3.52 5 3 3 5 1000 / T (K) 1000 / T (K) F; 9Ure 3.94 Conductivity of the lithium salt of F; 9Ure 3.95 Conductivity of the potassium salt of trifwruacélyl-dimethylbarbituric acid. Trifu ~ r acid, ac ~ tyl-dimethylbarbituric.

The ~ increase in conductivity is even more spect ~ ul ~ ire for the same composition to O / Li = 8/1, between MeBbLi and TFAMeBbLi, as illustrated in Figure 3.95.

BuBb: It is also possible to play on the length of the alkyl chains to modify the physicochemical properties of barbituric acids. Figures 3.96 and 3.97 thus give the con-lu; ti ~ ilé lithium salts and pot ~ bcc ~ lm dibutylbarbituric acid (BuBbM), for a col ~ position at O / Li = 8/1, 12/1, and 16/1. Conl. ~ U ~ .uent with lithium salts and pot ~ csillm of dimethylbarbituric acid, a marked optimum of conductivity is no longer observed; Moreover, potassium salts are more conductive by at least an order of magnitude relative to their counterparts in li ~ hinm This is not in contradiction with the preceding hypothesis of a transport partial ~ vehicular lithium with dimethylbarbituric acid, steric hindrance r ~ sult ~ nt of the presence of groupem ~ ntc butyles limit ~ nt probably the possibilities of existence of a stable complex between a lithium cation and two anions of dibutylbarbituric acid.
1 0 ~ 3 1 0'3 0 ~ 4 - ¦ BUBbLi ¦ ~ 1 0 ~ 4 _ ~ ¦ BUBbK¦

0 ~ o 8/1 10- ~ - + 12/1 - ~ 1 o-8, = 12/1 10 9 ~ 10'9 2.5 3 3.52.5 3 3.5 1000 / T (K) 1000 / T (K) Figure 3.96: Conductivity of the lithium salt of Figure 3.97: Conductivity of the p ~ t ~ c ~ n salt of dibutyl acid ~~ ique.I'acid dibutylbarbiturigue.
The visual aspect of films is profoundly modified by the substitution of methyl groupings by butyls, in fact, very crystalline films for methyl take on the appearance because ~ ctéristique of electrolytes in an amorphous state, by their transparency, with butyl. It exists therefore an effect of internal plasticization of the poly (ethylene oxide) chains by this salt, resulting from the increase in the length of the alkyl groups. It would be necessary to qu ~ ntifier this parameter by the study of p-opl; thermal summers of m ~ l ~ nges polymers / salts.

On the other hand, the presence of gloupe --- butylentents makes it easier to dissolve the salts, made of the ~ ulion of their reductive energy which results from the d ~ so ~ t; introduced by these.

2194 ~ 27 BuSBb: Figures 3.98 and 3.99 data ~ .nt lc; spe ~; live ~ salt conductivity ~~ "~
lithium and potassium salt of dibutylsulfonylbarbituric acid 'N'S'N'BuLj +
(BSBbM), for a composition with O / Li = 8/1, 12/1, and 16/1. The conductivity of O ~ O
lithium salt is no more than a factor of three lower than the salt of su potassium from dibutylsulfonylbarbituric acid. Here too, the films obtained, by their ~ ~, lce, p. ~ é ~ PlPmçnt calaetl ~; cti ~ read vi. ~ u ~ lles of an amorphous electrolyte.
10'3 10 3 _.10 ~ __ ~ BuSBbLi¦ - _ 10 'sl ~ BuSBbK¦ ---1 0 6 ~ 1 0 6 _ ~ _ 10'7- ~ 8/1 ~ - ~ 10'7- ~ 8/1 ~ -1 o 8 - = 12/1 ~, 9 ~ 1 o 8 _ = 16/1 1 0 9 ~ 10 9 '''' I
2.5 3 3.5 2.5 3 3.5 1000 / T (K) 1000 / T (K) Figure 3.98: Conductivity of the lith salt; um of Figure 3.99: Conductivity of the potassium salt of dibutylsulfonylL ~, b. "~ rique. dibutylsulfonylbarbituric acid.
Figures 3.100 and 3.101 compare the conductivity obtained, for the same composition at O / Li = 12/1 respectively for the lithium and potassium salt of BSBbH acid.
10'3 10 3 ~ 10 4 - ¦ 0 / Li = 12/1 ¦ _ 10.4 ~ 0 / K = 12 / 1¦ -~ ~ 01 0 ~~
~ 10 7 - \ \ "10 -O o MeBbLi ~ ~ ~ ~ c oMeBbK ~
o 1 o-8; ~ BuBbLi 1 ~ 8 ~ 9 o BuSBbLi oBuSBbK
2, 5 3 3, 5 2, 5 3 3, 5 1000 / T (K) 1000 / T (K) Figure 3.100: Con, f, ar .. ison of the conductivity of Figure 3.101: Co "" ~ ar ~ i., On of the conductivity of lithium salts of various barbituric acids. salts of po '~ sc ~ 7 of various barbituric acids.

~ 1 94127 -The conductivity of li ~ hillm salts, when the electrolyte is in a molten state, changes as follows:
BuBbLi <MeBbLi <BuSBbLi The m ~ illellre conductivity of the sulfonated derivative can probably be explained by a lower basicity of the anion, the sulfonyl group dispersing part ~ ment the anionic charge by inductive effect according to the link network ~, more çfflc ~ cemellt than the carbonyl of dibutylbalbilu ~ ic acid, and thus makes it possible to gain approximately a factor two in conductivity. However, it is interesting to note that the lithium salt of dimethylbarbituric acid is more conductive than its dibutylated counterpart, but in this case, the possibility of vehicular transport of lithium makes difficult the colllpal; lison of values of con-lu-; livi ~ s.

To overcome this problem, it is more judicious to compare the conductivity of the salts of potassium. The strength of dimethyl and dibutylbarbituric acids must be relatively close, the dramatic increase in conductivity by more than an order of magnitude is essentially assign to the pl ~ ctifi ~ nt effect butyl segments.

A more detailed study of this family of salts presellt it has an interest both theoretical, and convenient. The ease of modifying the structure of the anions, and the pKa in the acid DMSO
barbiturate, which we can estimate close to 5, allows us to consider making a correlation between the data likely to be determined in the DMSO, both in terms of acidity and complexing properties, and thermal and conduction properties in poly (oxide) medium ethylene) of these anions. This approach requires extensive work at the experimental level, should make it possible to provide interesting information on the relationships between the structure of anions and the properties of the salts as derivatives in a solvating polymer medium, such as POE.

From a practical point of view, it would be interesting to seek to optimize the conductivity of the salts derivatives of these anions, first of all in the influence of the length of the segments aL ~ yle, and in particular propyl and ethyl, the butyl groups probably do not represent the optimum in this sense. It is also possible to introduce reactive functions such as allyles or vinyls which would allow grafting or homopolymerizing these salts.

On the other hand, for groul) elllent aL ~ cyle presents ~ nt the opdmum of conducdvité, it would be n ~ cess ~ ire to study the influence of the labile proton substitution by an electron-withdrawing group, such as a trifluoroacetyl or cyano group, and more particularly by a trifluoro-group meth ~ nesulfonyl. We can thus expect to obtain anions present ~ nt conductivities of interest, and whose surfactant properties could have a beneficial effect on the function ~ b ~ t ~, laughs, and pardculier on the mic ~ s ~ ulc; of lithillm, following its dissoludon and redeposidon during battery charge / discharge cycles.

lI 3-34 Perspectives:

The chemistry of heterocyclic compounds makes it possible to envisage the synthesis of anions whose conductivity is close to that obtained with the lithium salt of bis (trifluorom ~ th ~ ne-sulfonylimide). Given the richness of this chemistry, it is likely that there is at least one CG ~ .rOS ~ of this ~ e family of anions which ~ .... and ~. ~ it to obtain both appreciable conducliviles in a poly (o ~ ethylene cyde) medium, a stil ~ elelochromic s ~ lffic ~ ntP ~ for gal ~ ltir a duration life impol ~ te batteries incorporating them and a moderate cost, the latter point being not to be overlooked since the lithium salt of bis (trifluoromethanesulfonylimide) represents a significant fraction of the calculated cost of b ~ tteries with polymer electrolytes.
If we restrict ourselves to primary generators, cost is still a more determinative element, and a anion like dicyanotriazole, by the conductivity of its lithium salt in POE medium, and its ease of manufacture and purification, will be cert ~ inPmPnt of great interest, both for ~ n ~ r ~ primary tellrs utili.c ~ nt liquid electrolytes than polymer electrolytes. Moreover, the lithium salt of dicyanotriazole is not hygroscopic, unlike LiTFSI, this parameter is to be taken into consideration as part of an industrial process.

The use of dicyanotriazole in a rechargeable generator poses the problem of stability at long term nitrile functions with respect to a lithium electrode, which remains to be demonstrated.
However, it would be possible to use the perfluoromethylated equivalent of this anion, which can be hope to synthesize by the action of an azide on hexafluorobutyne, a by-product of PTFE manufacturing, as illustrated in the figure below:
F3C ~ CF3 CF3 - C ---- C - CF3 + NaN3 ~ N ~ ~ NN a In addition, the synthesis of the perfluoromethylated derivative in 1-3 is known, and one can even consider obtain it by a simpler route than that described in the literature, by adding the trifluoroacetyl hydrazide on trifluoroacetonitrile, then by dehydrating this product using P2Os. Trifluoroacetonitrile being a relatively simple precursor to manufacture by dehydration of trifluoroacét ~ mide ISolVaYI, which can be obtained simply by the ~ listill ~ tion ~ eotropic in toluene of m ~ l ~ n ~ e of an aqueous solution of ammonia and trifluoroacetic acid.

CF3CN + CF3 - HERE r '~ N ~ CF3 - C N - N C - CF3 2 5 ~ F3C ~ ~ CF3 OH ~ NH2 H
In addition, the pru ... ~ lle ~ results on halogenul ~ s of dicyanoimid ~ 7ole, l ~ issçnt glimpse a potential interest for the synthesis of bis (trifluoromethyl) pyrazole halogenules, and in particular chloro-bis (trifluoromethyl) pyrazole (see figure below), which has the advantage of not ~ 194127 cor ~ hold of ~, pe ~ -Nitrile ~ in the é ~ bed ~, of their ~ & ~ Delivered screw of a lithinm electrode The conductivity of the lithium salt of this anion should, a priori, be of the same order of magnitude as that obtained with the lithium salt of chloro dicy ~ noimi ~ 7ol ~

F3C ~ CF3 ~ F3C ~ CF3 K + a Finally, let us point out the synthesis opportunities offered by the trifluoromethanes ~ llfin ~ te which is an agent s-lffic ~ mment nucleophile 1 ~ 1 to consider its reaction on diazotized derivatives amino azoles, which would synthesize the trifluorométh ~ neslllfonés derivatives described on the figure below:
N ~, ~ CN
CF3SO2 ~ g ~
N CN

CF ~ Br + KZ5204 ~ CFzSCzK ~) ~ CF350 ~ rN \
N ~
The interest in this compound is motivated by its synthesis which does not pass through a fluorinatioh path electrochemical expensive, but by the chemical reduction of Freon 13Bl ~, product of the great organic chemistry, by sodium dithionite.

II-3-3-5 Conclusion on cyclic salts:

The present study proves the interest of heterocylic anions in the field of electrolytes polymers. This is motivated by both the results obtained in summer ~ nt the conductivity of the salts of lithium from these anions in poly (ethylene oxide), due to the great richness of the chemistry used for the synthesis of these anions, perrnett ~ nt thus to consider the development of other derivatives. This chemistry should find an important place in the field of batteries in the future polymer electrolytes, and néccssiL ~ to extend the study of the chemistry of these anions, but also to optimize the synthesis of those which will present the best performances in the batteries.

In the particular case of the family of barbituric acids, the ease of modifying the hydrophobicity and the strength of these acids suggests interesting applications in the particular field phase transfer catalysis. Such an app ~ ation ~ even if it exceeds the framework of the present works, ~.; Le. ~ l however that it is devoted to a complete study of the properties of these acids.

_ II-3-4 Conclusion on alkaline salts:

The chemistry of anions to obtain metal salts ~ lc ~ linc dissociated in the media liquid or polymer aprotics in the end have hardly changed in the last twenty years. Except the introduction of the bis (trifluoromethanesulfonimide) anion a decade earlier, which is fast becoming a reference in the dom ~ inP- of b ~ tter ~ es with lithinm, the majority of work focused on fluorinated coordination anions or on perfluorosulphonates.

In fact, most of this work has focused on the .echelche of anions derived from very acidic strong, but relatively to the acidity in aqueous medium, so triflic acid is considered as a superacid in water with a pKa of the order of -10. However, while it is preferable to obtain a very strong acid in water use anions whose charge is carried by a base hard, like the oxygen in sulphonates, on the contrary, obtaining strong acids in the environment aprotic requires the use of anions whose charge is carried by a delocalized soft center pennP, tt: ~ nt to decrease the charge density in the volume of this anion as much as possible.

Thus, we could estimate that the acid of bis (trifluorom ~ th ~ nesulfo ~ imide) has a lower pKa eight orders of magnitude when passing water to DMSO. Based on this observation, it is more it makes sense to focus only on the chemistry of delocalized bases. Concepts from the organic chemistry are in perfect agreement with the results obtained by studying the conductivity lithium salts derived from these anions in poly (ethylene oxide).

This study made it possible to widen the field of the chemistry of these anions, on the one hand to molecules relatively simple, such as the anions derived from malononitrile or cyanamide, and especially to the chemistry of heterocyles which offers many opportunities for system, and whose application to the field of lithium batteries with polymer electrolytes will be cert linemPllt fruitful in the years to come.

In addition to the application problems of these materials, it appears that the concepts from the organic chemistry are of great interest for polymeric media, and it is likely that joint work on the acidity in a liquid aprotic medium of the different anions studied, relatively to their polymer conduction properties, as well as the establishment of their phase, will enrich the knowledge specific to the particular field of physico-chemistry polymer electrolytes.

.
II-4 CONCLUSION ON THE ELECTROLYTES:

The study of polymer electrolytes is m ~ int ~ n ~ nt initiated since a qllin7 ~ ine years, and more and more works are devoted to them.

The aim of the present study was, as well for the electrolytes with proton cor ~ uction, that the electrolytes conduction by alkali metal salts, to conduct prospective research allow ~ nt to consider new ways of synthesis of these materials.

As a result, organic chemistry still offers many opp ~ llunities to improve properties of these m ~ téri ~ both from the applied point of view, whether in the field of b ~ tt ~ ries polymer electrolytes than electrochromic glazing for example, that from a more background ~ m-have ~ l, showing the importance of concepts from organic chemistry in the field materials for the electrochemistry of solids. We can also attribute the little renewal concepts sometimes found in this area to the lack of interactions between the world of electrochemistry and that specific to organic chemistry.

~ 21,941 ~ 7 Bibliography:

[Abdul-Ghani] Abdul-Ghani, M .; Tipping, AEJ Fluorite. Chem. 1990, 48, 149-152.
[Al11]
[Alloin1] Alloin, F. Thèse de l'Université de Genoble, INPG, (1993).
[Alloin2] Alloin, FJ C / ~, ~ t ~ hei, L Soc. 1994, 141, 1915-1920.
[Angell] Publi Nature [Apen] Apen, PG; Rasmussen, PG llJter ~ rcles 1989, 29, 1325-1329.
[Armand1] Armand, MB; Chabay "o, JM; Duclot, M. Fast lon Tidns ~ ll in Solids (P. VashRa, JN Mundy, GK Shenoy Eds.), Elsevier-North Holland New-York 131 (1979).
[Armand2] Singapore [Armand3] Armand, M. r ~ ly ,, ~ e ~ S at Condonique lonique, 8th GFP educational course, ~ o ~ ri ~ lss electric despolymères, September 7-9, 1992, Dijon, published by the GFP.
[Armand4] Armand, M. SolidState lonics, 1994, 69, 309-319.
[Armand5] Arrnand, M .; Sanchez, JY; Gauthier, M .; Choquette, Y. Electrochemistry of Novel Materials (P. Ross, J. Lipkowski Eds.), VCH pu '' he, ~ New York 1993.
[Armand61Arrnand, M. P ~ erElectrolyte ncvi ~ "~ 1 p1-22.
[Bertier]
[Biel ~ ~, n] Biel ~ ~, en, G. cie ~, tlu ~ rn. Acta 1992, 37,1471-1478.
[Bordwell1] Bordwell, FG; Branca, JC; Hughes, DL; O '..,:, ~ ead, WNJ Org. Chem. 1980, 45, 3305.
[Bo, d- ~ _112] Bordwell, FG; l lar. ~ lson, JA, Jr .; Zhang, XJ Org. Chem. 1991, 56, 4448.
[Bordwell3] Bordwell, FG; McCallum, RJ; Olrnstead WNJ Org. Chem. 1984, 49, 1424-1427.
[Bordwell31 Bordwell, FG; Branca, JC; Hughes, DL; Ol.,.:, ~ Ead, WNJ Org. Chem. 1980, 45, 3305.
[Bo- ~ w. "L1] Bordwell, FG; l lan ~ lson, JA, Jr .; Satish, AVJ Org. Chem. 1989, 54, 3101.
[Bordwell5] Bordwell, FG; Zhang, XM; Filler, RJ Org. Chem. 1993, 58, 6067.
[Bordwell6] Bordwell, FG Acc. Chem. Res. 1988, 21, 456-463.
[Bordwell7] BGrdwell, FG; Fried, HEJ Org. Chem. 1991, 56, 4218-4223.
~ Baril1] Baril, D. Thèse de l'IniversRé de Genoble, INPG, 1993.
[Barrel2] Barrel, DJ Cle ~, l, ~ he ".. Soc.
[Baril3] Internal report Hydro-nuPhec [Begland] Begland, RW; Hartter, DR; Jones, FN; Sam, DJ; Sheppard, WA; Webster, OW;
Weigert, FJJ Org. Chem. 1974, 39, 2341-2350.
[Beh) ger] Gbh.) Ger, A; Gauthier, M .; Robitaille, M .; ~ 8 "~ l ~ al'e, R. Second Inte", dtional Symposium on r ~ ", ~, a ~ -bulytes (B. Scrosati Ed.) Elsevier Applied Science London & New-York 347-381 (1990).

[Belcher] Belcher, R .; Sykes, A., Tatlow, JCJ Chem. Soc. 1954, 4159-4161.
[Benrabah] Benrabah, D .; Barrel, D .; Sanchez, JY; Armand, M .; Gard, GG
J. Chem. Soc. Faraday Trans. 1993, 89, 355-359.
[Booth] Booth, C .; N- ~ 'olas, CV; Wilson, DJ r ~ "~ er ~ u ~ tc n" i ..
(JR MacCallum, CA Vincent Eds.), Elsevier London 229-254 (1989).
[Boyd1] Boyd, RHJ Phys. Chem. 1963, 67, 737.

; .
lBoyd2] Boyd, RHJ Phys. Chem. 1961, 66.1834.
[C ~ cdeleclc] 6redel ~ .k, H .; Scl "~ ker, G .; Oehler, E. Ann. Chem. 1956, 600, 81-108.
lBrownl] Brown, HC; Cheng, MT; Parcell, LJ; Pilipovich, DJ Org. Chem. 1961, 26, 4407-4409.
[Brown2] Brown, HC; Ming, TCJ Org. Chem. 1962, 27, 3240-3243.
[Catalàn] Catalàn, J .; Abboud, JLM; Elguero, J. Adv. Heterocycl. Chem. 1987, 41, 187-274.
[ChdbdyllO] Chabdy-, o, JM Thèse de run; ~ r ~ t de G ~ n L ', INPG, 1980.
[Chdml ~ c,:,] Chall ~ tcr ~, WJ; Coffman, DDJ Org. Chem. 1961, 26, 4410-4412.
[Charbouillot] Thesis of the university of ylen ~? ~ INPG, 1988.
[Cherkaoui] Cherkaoui, F. Thèse de l'Université de Grenoble, INPG, 1982.
[C ~ ':., Lan] Chemistry of Solid State ~ idlj 2, Proton Conductors, (P. C ~ 'c ~ ILrdl), Ed.), Ca..lLIkJge U-, i ~ r ~ ity Press, CarlL.idge, 1992.
[Cookson1] Cookson, RC; Hudec, J .; Whitear, B. Proc. Chem. Soc. 1961, 117-118.
[C ~ J ~ k ~ n2] Cx! - ~ n, RC; EricJli ~ h, KRJ Chem. Soc. (C) 1966, 1641-1644.
[Defendinil Defendini, F. Thèse de l'Université de Gr ~ r, Ab '~, INPG, 1987.
[De Zea Bermudez1 l De Zea Bermudez, V. Thesis of the University of Grenoble, 1992.
[De Zea Bermudez2] De Zea Bermudez, V .; Sanchez, JY SolidState lonics 1993, 61, 203-212.
[Dippell Dippel, Th .; Kreuer, KD; I ~ ssèglles, JC; Rodriguez, D. Solid State lonics 1993, 61, 41-46.
[Dominey] Dominey, LA 7th l ~ r "~ ional Meeting on Lithium Batteries, May 15-20,1994.
[Elguero] Elguero, J. & ALII J. Org. Chem. 1991, 56, 3942-3947.
[Fialkoff] Fialkoff, JA Bull. Soc. Chim. 1927, 41,1209-1217.
[Fi. "Legan] Finnegan, WG; Henry, RA; I ~ ct, RJ Am. Soc. 1956, 80, 3908.
[Fisher] Fisher, A .; Leary, GJ; Topson, RD; Vanghan, JJ Chem. Soc., B. 1966, 786.
[Foropou '~, s' Foropoulos, J., Jr .; DesMarteau, DD Inorg. Chem. 1984, 23, 3720.
[Friede.i.; L,] rriedericl1, WUS Patent 2,710,297 (1955).
[Gauthier1] Gauthier, M .; B ~ langc ,, A .; Kapfer, B .; Vassort, G .; Armand, M.
F ~ ly. "Er Cl ~ tlulytes ncvie ~ ~ -11 (JR MacCallum, CA Vincent Eds.), Elsevier London 285 (1989).
[Gauthier2 ~ Gauthier, M .; Fauteux, D .; Vassort, G .; Béhnge ~, A .; Duval, M .; Ricoux, P .; Chabagno, JM;
Muller, D .; Rigaud, P .; Armand, MB; Deroo, DJ Ele-; tl ~ he "-. Soc. 1985, 132, 1333.
[Gunwald] Gunwald, E .; Price, EJ Am. Soc. 1964, 86, 4517-4525.
[Gutmann1] Gutmann, V. TheDonnor-Accep ~ orApproachtoA1 ~ 'e ~ lnteractions Ed. Plenum Publ. Corp. New York (1978).
[Gutmann2] Gutmann, V. Ele ~, t ~ himica Acta 1976, 21, 661-670.

Gutmann3] Gutrnann, V .; Mayer, U. A1 ~ n ~ t ~ h. Chem. 1968, 99.1383.
Gi, Lhi. ~ Trochimowski] Giszkiewicz-Trochimowski, E. Rocz. Chem. 1921, 1, 468-478;
Chem. Abstr. 1923, 17, 1424-1425.
[Hadzi] Hadzi, DJ Mol. Struct. 1988, 177.1.
[Hd ~ lcl] Hartzler, HDJ Am. Chem. Soc. 1964, 86, 2174.
[lernJ ~ i ~, l ~ on] I lend ~ i ~, l ~ on, JB; Giga, A .; Warting, JJ Am. Chem. Soc. 1974, 96, 2275-2276.
[Hinkel] Hinkel, LE; n; ~; hdlJ ~ GO; Thomas, OJ Chem. Soc. 1937, 1432-1437.
[Jap1]

[JOSeYI JOSQY, AD; D: _ki-, son, CL; Dewhirst, KC; McKusik, BCJ Org. Chem. 1967, 32, 1941.
[Keml] Kem, JM Thèse Université Louis Pasteur, Strasbourg, 1976, Order No CNRS Archives AO 12.508.
[Kern2] Kem, JM Te ~ rahedron 1978, 34, 661.
[Klemann] Klemann, LP; N_.l ... year, GH; Whitney, TA; Stogyn, EL; Farcasiu, D.
Meeting of the Cl. ~ Trv- hemical Society October 5-10 (1980), Cf. Extended ~ hstPrtc n ~ 47.
IKoppel] Koppel, IA; Taft, RW; Anvia, F .; Zhu, S. -Z .; Hu, L. ~ .; Sung, K. -S .; Desmarteau, DD;
y, A, 91 ~ ''L'i;,LM; yA ~ (IrOIS ~ Y. L; Ignat'ev, NV; Kondatrenl. ~, NV; V ~ '! c nskii, AY; Vlasov, VM;
Notario, R .; Maria, P. -CJ Am. Chem. Soc. 1994, 116 3047-3057.
[Koshar] Koshar, RJ; Mitsch, RAJ Or9. Chem. 1973, 3fl 3358-3363.
[Kreuerl Kreuer, KDJ Mol. Struct. 1988, 177, 265.
[Lagan] Laganis, ED; Lemal, DMJ Am. Chem. Soc. 1980, 102, 6634-6636.
[I Acse ~ es1] 1 A ~; segues. JC in C ': I, l year lLe Goff] Le Goff, E .; Lacount, RBJ Org. Chem. 1964, 29, 423-427.
[Linert] Linert, W .; Jameson, RF; Taha, AJ Chem. Dalton Trans. 1993, 3181-3186.
[Lopyrev] Lopyrev, VA; nacl. ", ~ laulina, TN Zh. Obsh. Khim. 1983, 53,1684.
[Malarski] Mabrski, Z .; Sobczyk, LJ Mol. Struct. 1988, 177, 339.
[March] March, J. ~ Advanced Organic Chemistry ~, 3rd ed .; Wiley-lntel ~ ~ nce, 1985.
[Maskill] Maskill, H. ~ The Physical Basis of Organic Chemistry ~; Oxford science publications. (To complete) [Mayer] Mayer, U .; Gutmann, V .; Gerger, W. ~ 1ohdt ~ ll. Chem. 1975, 106.1235.
[Middleton] Middleton, WJ; Metzger, DJ Org. Chem. 1970, 35, 3985-3987.
[NOT .. . ~ year]
[Norris] Norris, WPJ Or9. Chem. 1962, 27, 3248-3251.
[Ohtsuka1] Ohtsuka, Y. Synthesis 1984, 1057-1058.
[Ohtsuka2] Ohtsuka, YJ Org. Chem. 1976, 41 629 ~ 33.
[Ohtsuka3] Ohtsuka, YJ Org. Chem. 1976, 41, 713-714.
[Oliveri M-, ''] Oliveri-Mandalà E .; rAY ~ I ~ c ~ uA. T. Gazz. Chim. Ital., 41, Il, 430-435.
[Paprott] Paprott, G .; Seppelt, KJ Am. Chem. Soc. 1984, 106, 4060-4061.
[Pearson1 ~ Pearson, RG Advances in linear free-energy relat; onships Ed. Cha ~, "~ nn & Shorter Plenum Press New-York (1972).
[r ~ - ~ on2] Pearson, RGJ Chem. Edu. 1968, 45, 581-587.
[Pea, ~ r3] Pearson, RG; Dilbn, RLJ Am. Chem. Soc. 1953, 75, 2439.
[Picotl Picot, A .; Genouillet, P. L'actualité Chimique, April 1986, 41.
[P &-.:,;. J. on] Pc. ~; ~ - on, C. Publi ILL

[Ravet] Ravet, N. These of the University of Grenoble, INPG, 1994.
[Ripoche] Ripoche, X. ECSEle ~ tru tuv ,, 1 ~ 1 ~ 'll [Robitaille] Robitaille, CD; Fauteux, DJ Ele ~ t ~ u he ",. Soc. 1986, 133, 315-325.
[Rou ~ seAu] Rousse ~ Au, F. Thèse de l'Université de Grenoble, INPG, 1990.
[Shuklal Shukla, JP; Tandon, SGJ Cl ~ buanaL Chem. 1972, 35, 423-427.
[S ~. ~ T ~ ,.] S! Vestcr, MJ Aldrichimica Acta 1991, N ~ 2, 31-38.

21 941 2 ~

lSmith] Smith, WT, Jr; Steinle, EC, Jr. J. Am. Soc. 1953, 75, 1292-1294.
[Solvay] Doc ~ ", ~ nt ~ technical lion Solvay: Specialty Fluorine product.
[Sb ~.] Sb_ih ~ ie3er, A., Jr .; Heathcock, CH Introduction to O ~ y ~ n ~ e Chemistry (Translated by B. Coste, B. Daniel, Y. Dugenet), Edition Marketing, Paris, 1986.
[Sylla] Sylla, S .; Sanchez, JY; Armand, MJ Electrochim. Acta 1992, 37, 1699-1701.
[Tst.dhed ~ un] T ~ hdheJI ~, n Lett. 1979, 3643.
[Tipping] Tipping, AE & al J. Org. Chem. 1994, 59.1Qq ~ -1046.
[Trinquet] Trinquet, O. Thesis from the University of Bordeaux, 1990.
IVallée1] Vallée, A .; Besner, S .; Prud'homme, J. Electrochim. Acta 1992, 37.1579-1583.
[Vallee2] Vallée, A. Thesis of the University of Mcnl, ~ al, 1994.
[Volkonsky] Volkonsky, A. Yu .; Rokhlin, EMJ Fluorine Chem. 1984, 26.183.
[W ~ b ~ 11] Ciganek, W .; Linn, WJ; Webster, OW PATAT
ler2]
[WeL ~ er3] Webster, OWJ Am. Chem. Soc. 1966, 88, 3046-3050.
[\ ~, lel4] Webster, OWJ Org, Chem. 1967, 32, 39-42.
[Webster5] Webster, OWJ Am. Chem. Soc. 1966, 8fl 4055.
[Webster6] Webster, OW Adv. Heterocycl. Chem. 1987, 41, 1-40.
[Weis] Weis, CDJ Org. Chem. 1962, 27, 3695.
[Wiley] Wiley, DW; Webster, OW; Blan ~: l, ard, EPJ Org. Chem. 1976, 41, 1889-1895.
[Wao ~ d] Woo ~: ~ d, DWUS Patent 2, 534, 331 (1950), Chem. Abstr. 1951, 45, 5191.
[3M] Docl.,., ~ Nt ~ lion techn, .e 3M: FluoradLithium Trffluvr., I "~ ll, anesulfonimide Battery Ele ~ tlolytc HQ- 115.

2t94127 Chapter II:
Mixed conduction polymers II-1 INTRODUCTION:

In general, a mixed conduction polymer can be defined as a material macromolecular allows ~ nt the transport of both electrons and ions, in a system any electrochemical. We take this definition broadly as well at the level mac ~ ascopic than microscopic. This definition can bring together various approaches in terms of m ~ t ~ ri ~ l ~ x, and lead to dirÇé-ltes function ~ lit ~ s in electrochemistry, and more particularly in the field of polymer electrolyte batteries. We briefly describe, below, the dirréi ~ e ~ tes a ~ es ~ -n ~ which we summss i ~ ssés.

A first possibility for obtaining a polymer with mixed con ~ luction is to m ~ l ~ nger an electrolyte polymer with a charge of microparticles of an electronic conductor, such as grains of carbon. In this context, we are interested in the possibilities of grafting on the surface of carbonaceous material segments of solvating polymers, or dissociated anions, in order to locate at microscopic level the two conduction modes.

On the other hand, the synthesis of polymers comprising both solvating segments and functions electroactive leads to solvating polymers with redox properties, which we we will name more simply by redox solvating polymers. The ionic conducdon being ensured by the solvating segments, and the electronic conductivity via the, or the internal redox couples. The confinement of electroactive functions in a portion of space, induced by nature macromolecular mi ~ erii ~ -lx allows, in particular, to use them to achieve sources of energy of high intensity as we will see later (II-3-5).

Finally, the molecules with redox properties dissolved in a polymer electrolyte, if they are capable to di ~ ser quickly in such an environment, allow to transport electrons within a system electrochemical when the potential applied to the system exceeds a threshold potential (Es). In this sense, these materials behave like a polymer electrolyte when E <Es ~ and com does a polymer with mixed conduction when E> Es ~ and can potentially be used as a protector overcharging in batteries with polymer electrolytes.

114 f3-II-2 CARBON GRAFTED:

II-2-1 Introduction:

It is common practice in the field of batteries (alkaline, saline,.) And batteries, to introduce into the ca.bolle method, in order to favor electronic exchangP, between the electroactive material and the coll ~ Pcte ~ lr of current. As a general rule, a carbon black prepared by thermal decomposition of acetylene (in particular "Shawinigan black"). Cathode composite of lithium batteries with polymer electrolytes is thus made up of the mixture of a maté-ri ~ ll insertion (TiS2, V2Os, MnO2, ...), carbon black, and a polymer electrolyte, as illustrated in Figure 2.1, the latter permp ~ tuant both ion transport and serving é ~ alPmPnt of binder to the electrode.

~ O Insertion material e ~ ~ ~ Li ~ Graindecarbone _r ~ ~ _ ~! 511 ~; 5 ~. E ~ electrolyte polymer Figure 2.1: Desc, i ~, schematic lion of the composite cathode of a polymer electrolyte battery.
It is also possible to consider that the cathode is con ~ titlled by mixing a polymer with mixed condllction (polymer / carbon electrolyte) with the insertion material. The insertion reaction involves bringing a cation into the structure which is placed in the vacant sites of the material (between the sheets ~ in the particular case of those with a two-dimensional structure like the titanium disulfide), the charge of the cation being compensated for by the concomitant introduction of an electron which is delocalized on the available electronic sites (in general the "d" orbitals of the metal of transition of the host structure). The role of the mixed conduction polymer is therefore to allow the simultaneous transport of ions and electrons on the surface of the insertion material. Therefore, the ion transport taking place in homogeneous phase at the molecular level in the electrolyte, it seems reasonable to try to transport the electrons also in the most homogeneous way possible between the current collector and the surface of the insertion m ~ t ~ ri ~ n.

In this sense, the Shawinigan black presents the disadvantage of being organized in the form of chains (strings) which thus constitute a tri ~ limPnsionnel network of carbon grains allowing the transport of elects. This structuring of the carbon grains makes it possible to obtain composites presP-n ~ n ~ good overall electrical conductivity. However, this results in inhomogeneity in 21 94t27 , micloscopic level which t ~ imin ~ e all the more uniformity of the distribution of electrons on the surface of the m ~ teri ~ U of insertion. In addition, the structuring of carbon grains in a tri ~ im ~ nsional network induces, as a rule, a rheology difficult to control going as far as behaviors "~ lil ~ t ~ ntc" as well as the porosity in the c ~ composite method, ~ nt its energy density volume. In this case, the tortuosity of the ion path is also p ~ n ~ lic ~ nte in terms of kinetic.

Also, it seemed to us interesting to graft on the surface of carbon grains segments of a solvating polymer, grafting a polymer to the surface of a particle pçrmett ~ nt to modify its interaction with a polymer matrix. So, for silica particles grafted with chains hydrocarbons [Donnet21, the surface energy (Ys) of the grafted particle, characterizing its potential of physical interaction with the environment, gradually tends towards that of pure polyethylene at as the grafting rate increases. The grafting of 300 mg of polyethylene segments by gram of silica is sllffic ~ nt to obtain this result, illustrating by the same the effect produced by a weak mo ~ ific ~ tion of the surface of the particle. This approach should make it possible to make the surface of the carbon compatible with the polymer electrolyte, facilitating its dispersion in the matrix, in order to to obtain a microdispersion of carbon in the composite cathode. Such microdispersion should lead to a homogeneous distribution of carbon grains, which one can hope for an effect beneficial on the performance of the cathodes during an important pllisC ~ nce call; a lesser porosity of the composite can also be expected. Seen from this angle, the problem is close to that of m ~ t ~ ri ~ nx structural composites. Indeed, it is recognized ~ PaPirer1l that obtaining of a high performance material through the development of a homogeneous composite, promoting optimal dispersion of the mechanical stresses exerted between the polymer matrix and the filler serving as reinforcement When a call of ~ ~ important nce, it must also ~ ~ consider that it is created in the electrolyte polymer component of the method, gradients of salt concentrations due to mobility predc-min ~ nt ~
anions, resulting in the same a decrease in the performance of the cathode. If it turns out possible to graft a salt on the carbon surface, this would benefit from the advantages expected from the use of fixed anions to limit the penile ~ read ~ nt effect of these gradients concentrations.

II-2-2 Grerr ~, e of polymers:

II-2-2-1 modifications of the carbon surface:

As a general rule, the surface chemical composition of a solid differs from its composition mass, this res..lt ~ nt of an environment ~ ment different from the atoms of the surface which represents a diccontinuity of matter. In the particular case of carbon, its surface condition is a parameter important for many applications, in fact, it defines the int ~ this between the m ~ t ~ ri ~ l carbonaceous and the medium el-vi ~ n ~ By e ~ mrle ~ in carbon / polymer composites, the nature of the surface of the carbon considerably modifies the adhesion of the fiber with the polymer, and by the same token pro ~ les ~ ues of the composite IHushes]. Many works therefore relate to the poscihilit ~ s to modify this IE ~ 'l surface, for example, controlling its acid-basicity makes it possible to carbon whose surface is made acidic, compatible with a basic sorting [SChlUtZl. The poscibilit ~ s of such mor1ifir ~ til n ~ surface cl ~ hangs ~ nt largely from the sl.u.; lu ~ e du m ~ teri ~
carbonaceous envisaged.

Indeed, carbonaceous materials are obtained ~ nnent gener ~ le-m ~ nt by pyrolysis of organic precursors such as hydroc ~ l ~ ures, or aromatic compounds for example. In a simplified way, the elul ~; of the carbonaceous materials obtained depends in part on the temperature at which the pyrolysis, low temperatures (aS500-1000 ~ C) lead to very desiccated materials, while the high temperatures (> 2500 ~ C) allow synth ~ ti ~ pr very crict materials ~ llinc to ~ lluelule graphitique. Now, in graphite, the sp2 hybridized carbon atoms are chemically inert, and only incompletely bound atoms placed at the edge of the graphite planes, or located on structural defects, are likely to react chemiquçm ~ nt On the other hand, the m ~ téri ~ lls carbonés désorg ~ nic ~ s, because of their many surface defects, have a multitude of points attacks to chemical agents making it possible to modify their s ~ lrface ~ c, and this by various methods.

It is thus possible to introduce simple chemical functions to the carbon surface, these allowing subsequently to initiate the reactions of organic chemistry which are their own.
Carbon oxidation is the most used process to do this, it can be done by example, in the gas phase (air, oxygen, ozone, water vapor, ...), in the liquid phase (nitric acid, hypochlorite, sulfuric acid, ...), or electrolytically. Among all these processes, oxidation in phase g ~ 7Pu ~ e is most used for carbons obtained in powder form, oxidation electrochemically being of great interest for carbon fibers (in the form of fabrics, felts, or threads), due to the possibility of using continuous processes.

We give in table 2.1, by way of example, the physicochemical characteristics of two of these carbonslT5Ub0kawa ~ l ~ the FW200 (Degussa) and the NeospectraII (Columbian Carbon) ; tunes of this class of m ~ t ~
Table 2.1 Car ~ rK ~ physico-chemical ues of two ca.L, oxidized ones, according to [Tsl ~ 1.;. 'L 1].
Surface BET Surface groupings (meq / gr) Carbon Type (m2 / gr) pH COOH OH C = O
FW 200 Channela 4S0 2 0.61 0.1 1.42 Neospectra ll Fumaceb 906 3 0.4 0.24 0.92 a: obtained by ~,). uly;, e in a flame b: obtained by pyrolysis in an oven Oxidized carbons of this type are available from various manufacturers; an important outlet by being the -tili.c ~ Not as a pigment for printing ~ nt ~ q ~ s high speed inkjet [Degussa].

This oxidation results firstly in ~ ugmPnt ~ tion of the specific surface of carbon, concomitant with the introduction of oxygen functions on the surface, part of the carbon planes p ~ t a sllu ~ iluu ~ d ~ solgal ~ isée being oxidized to carbon dioxide (this is called carbon activated [Donnet3]) In these con ~ litiorlc ~ the p.ill ~ ipales fonctionc introduced on the carbon surface are carboxyls, hydroxyls, carbonyls. As shown in Table 2.2, the functions carbonyls can app ~ lenil to different grollpçmPnt.c present on carbon like quinones or l ~ onF ~ by P ~ emplr Table 22 F ~ i, .. ;; ~ aux gro ~ ", t:" t ~ chemical pr ~ s ~ rit ~ on the surface of carbonaceous materials.
R, OH
ACID ~ ~ ~ H ~ S03H
Acid ~ Acid Acid Ca.Lu> ~ jl;,) e Pl, os ~ .hor ,, ~ eS ~ n ',: e ~, OH ~, NH2 ~ 1 [~

Arnine Pyridine Alcohol Pyrrolic Nitrogen X
HALIDE ~ X = Cl, Br It '-3'ne CAR90NYLES ~}

LactonePyrone Quinone Anhydride This same table describes the main groupings that it is possible to create on the surface of carbonaceous materials. We will not specify the various conditions allowing to reach these, the more particularly interested reader being able to use the data of the lil ~ lulc to do this ; c "~ sl. As we can see, the possibilities of fon ~ til ~ rln ~ liQ ~ ti ~) n of the surface calbone are QllffiQ ~ mmPnt varied to consider initiating many chemical re ~ rtionc to From its s ~ l ~ ce Finally, note that the conception of these dirrefell ~ s gfou ~ e ~ PntQ on the surface carbon vade typi4ut; lllenl between Ql and S meq by ~ r ~ same carbon.

2t 9 ~ 1 ~ 7 . .
II-2-2-2 Grafting of polymers:

The grafting of macromolecular chains on the carbon surface can be carried out by coupling a prepolymer carrying a lclive function with respect to its surface, or by a polym ~ ris ~ tion initiated with from this surface whether anionically, c ~ tioniq ~ let or r ~ ir ~ l ~ ire.

a) CQ- ~ PI ~ ge of a prepo; yJ ~ I ~ re In ~ l ~ do we first of all study on the coupling of a prepolymer with the surface carbon, in most of these works this coupling is carried out via the hydroxyl functions and carboxyls p ~ present on oxidized carbons like FW 200 or Neospectra II. Also, by For the sake of simplification, we have limited ourselves exclusively to these two possibilities. This grafting can errecluer dilec ~ nt from pl ~ sen ~ es functions on carbon, for example, is ~ rifi ~ nt the carboxylic gro ~ ents with alkanols [Pa ~ l, or by reaction of poly (propylene glycol) terminated by an isocyanate (PPG-NCO) with the hydroxyl and carboxyl functions [TsuL ~ .2].
However, in most cases, the groups present on the surface of the carbon undergo a tr ~ chemical itemçnt modifying their functionalities and reactivities, Figures 2.2 and 2.3 summary ~ nt different possibilities to do this. Thus, it is possible to modify the hydroxyl functions and carboxyls to isocyanates by treating them with toluene diisocyanate [Ts ~ '~], or to epoxides by reaction with epichlorohydrin [T ~ '31. Concern ~ nt specifically carboxylic acids, it is possible to modify them into acid chloride by action of thionyl chloride [Ts ~ 1. Chloride acid can itself be transformed into acid azide by treatment with sodium azide, this function reor ~ ni ~ nt under the action of heat in isocyanate [T ~ '51.
NCO

~ 0 SOCIz (~
,, (~ COOH- -(1) SOCI7 (~ CON3 ~ (~ NCO

~ a ~ 7a ~ o- ~ al ~ ~ d Figure 2.2: Chemistry of gro ~ e,., In ~ carboxyls pr ~ se.lt ~ on the carbon surface.

21 94 ~ 7 ~ o OCN ~ CH ~

(~ OH c ~ 2a ~ ~ (~ O-CH2 ~
Flgure 2.3: Chemistry of gro ~ e,., Ent ~ hydroxyl pl ~ s ~ ht ~ at 18 carbon surface.
All these functions are capable of reacting with nucleophiles, and in particular with ter nin ~ s polymers by g ~ oup ~ nts amino or hydlo ~ ylés, as illustrated in Figure 2.4, pçnn ~ tt ~ nt thus to obtain carbons grafted with different polymers such as polyethylene glycol (PEG), polypropylene glycol (PPG), polyb ~ lt ~ di ~ -ne glycol (PBG), polyvinyl alcohol (PVA), polyethylene imine (PEI), silicone diol (SDO), .cilirone. ~ min ~. (SDA).

2 ~ ~ 2 - P ~) n V ~ v ~

Figure 2.4: Grafting of nucle ~ p ~ s polymers (Fl = -N, NH, or OH) with the sufface of reactive carbons. .

The ~, essential parameter characterizing these carbons is the polymer grafting rate defined as the ratio of the mass of grafted polymer relative to the mass of carbon, so we have grouped in Table 2.3, the grafting rates obtained for various experimental conditions Table 2.3 Rate of grafting of polymers obtained on the surface of functional carbons Reactive function Polymer Mass of the Rate of on the carJone grafted polymer grafting Reference COO ~ C ~ 8-OH 269 7.4% [PapireR]
COOH / OH PPG-NCO 2000 82.2% ~ rsubokawa2]
PPG 2000 11.2%
PEG 8200 1 8.5%
PBG 1000 27.2%
COCI PVA 66000 41.6% [Tsubokawa4]
S DO 5600 20.4%
SDA 3000 40.7%
PEI 20,000 45.8%
PPG 2000 21.9%
CON3 PEG 6000 29.7% ~ subokawa5]
PVA 66,000 43.4%

1..
PPG 2000 19.5%
SDO 1000 23.6%
SDO 1800 26.3%
Epoxide S DO 3200 30.0% [Tsubokawa3]
SDO 5600 32.2%
PEI 1800 55.6%
SDA 1700 59.5%
SDA 3900 67.2%
Let’s take a closer look at the results obtained by grafting polymers onto a carbon (Neos ~ ecll ~ II) carrying 0.4 meq / gr of group-nt ~ epoxides ITq: L - 31. As shown these, the silicone diol grafting rate increases with the mass of the polymer, inversely the number moles of ch ~ isles grafted per gram of carbon, symbolized by G, decreases with this same mass. It’s likely that as polymer segments graft onto carbon, on the one hand the reactive functions on the surface are masked, and on the other hand the polymer must make their way between the chains already grafted to react with a groupen.cnt of sllrf ~ re, or, this displacement ~ crawling is slower as the length of the chains increases. Number of chains g. ~, Lr ~ s is thus connected to the molar mass ~ c ~ ire (M) of the silicone diol by:
G = K Ma, with K = 0.0468 and a = -0.77 It is in ~ lcssant to note the simili ~ ude of this ~ equation with the law of Mark-Houwink-Sakurada which describes the intrinsic viscosity ~ ue of a given polymer as a function of its mass by:
~ 1 ~] = K Ma On the other hand, it appears that for an equivalent length of chain, the grafting rate with the amino polymers is greater than that obtained with hydroxylated polymers, and thus goes from z 25 to 60% for chains with a mass ~ 2000. This result is in agreement with the least power nucleophile of alcohols relative to amines.
Thanks to polymer grafting, colloidal carbon suspensions can be obtained in the organic solvents relatively stable over time. Thus, for the Neospectra II grafted with silicone ~ i ~ mine and dispersed in THF, 55% of the carbon grafted with 59% of dry polymer ~ iment ~
after 100 hours, and 20% of the carbon is grafted with 67% polymer grafted after the same period of time, while the c ~ l, one non goeffé sedim ~ ntf fully in a few hours.

We will m ~ inten ~ nt we i ~ -le-Gssei to the possibilities of modifying the carbon surface thanks to a polyTn ~ ric ~ tion by anionic, ca ~ ionic, or r ~ r ~ l ~ iTe initiated from this surface.

b) Grafting by anionic polymerisaffon:

Consider the case of grafting via anionic polymerization initiated from groups present on the surface of an oxidized carbon, of the same type as those exposed previous ~ emmen ~ First, Figure 2.5 illustrates the specific modifications to Up ~ tc carboxyles. Even if it is possible to initiate anionic polymers ~ ~ ~ d ~ e ~ n by the carboxylic acid, this function is activated by transforming it into its alkali metal salt ~
tr ~ itemP, nt by a strong base in water [T ~ ~ I On the other hand, it is possible to obtain the ~, of ammonium from the carboxylic acid by treatment of the acid with a hydrox ~, ammonium IT- - ~~]. Transformation of the carboxylic acid into a basic amine effectllée in various ways, starting from the acid chloride by reaction of a compound ~ i ~ mine, ~ j from the isocyanate of the carboxylic acid, either by its hydrolysis, or by reaction with ".
cc ... ~ difonrtionnp ~ l carrying a primary amine, or secon ~ 1 ~ ire ~ and a primary amine, second ~
or tertiary IT ~ bok ~

MOH ~ (~) - COO M + M = U, Na, K, Cs, Rb (~ COOH R4NOH ~ @ ~ coo N ~ R4 OH
SOC12 ~ (~ H2N (cH2) 2NH2 (~ C - N - (CH2) 2 - NH2 (~ NCO H20 ~ (~ NH

R ~ CH3 or HC IN - HERE -, N - (CH2) 2NR ~
Figure 2.5: Chemistry of carboxylic groups to initiate radical polymerizations. -On the other hand, the tr ~ it ~ P.mP, nt of carbon oxidized by butyllithium makes it possible to transform the whole ~ oxygenated functions (carboxyls, hydroxyls, quinones) into liLhium alcoholate [Tsuboka wa12l, horn :: -illustrated in figure 2.6, we can thus access carbons carrying many reaction functions .-:
(~ 2 meq / gr) on their surfaces.

(~ COOH (~ U

(~ OH + BuLi ~ (~ OLi He Bu ~, OR
~ D (~ D
O Bu OR
Figure 2.6: Effect of 13utyllithium on gro ~ en, oxygenated enls.
The butyllithinm allows eg ~ lpmpnt to make ~ ive the surface of non-oxidized carbon contenam .-little g ~ up ~ .-. . .P.nt.C oxygenated on the surface ( <0.01 meq / gr), and particularly that of graphite.
m ~ Ct ~ lli.c ~ tion of grol ~ np ~ tc aromatic emerging on the surface of the carbon, like the illus ~ ~
figure 2.7 [Tsubokawa13 14l - 2 1,941 27 ,. .

(~ BuLi ~ (~ Lj Figure 2.7: Metallization of aromatic cycles by butyllithium.

All these grou ~. .. ~ l ~ ts permpnt to modify the carbon surface, by initiating anionically, either the polymerization by OU ~ lUI ~; cycle of cyclic monom ~ res, or polym ~ .ri.c ~ tion of my vinyl records.

Conce-rn ~ nt the polym ~ -ri. ~ Tion by or ~, ~ "Mon cycle, the grafting of polyesters is obtained by the polymerization of ~ -propiolactone initiated by alkali salts [T ~ 71 and ammonium ITg_L ~ 101 quaternaries of carboxylic acids, or by copolymerization, initiated by salts alkali of carboxylic acids, between an anhydride and, respectively, an epoxide lT ~ 1, OR a alkylene carbonate IT ~ ubokawa8l. On the other hand, the polymerization of y-methyl L-glutamate N-carboxyanhydride initiated by a carbon carrying functions ~ min ~ es allows to graft a polypeptide on its surface ITsubokawa1 ~ I. It is thus possible to obtain grafting rates of between 30 and 150% on the surface of carbon.

If we are interested m ~ inte. ~ Nt vinyl monomers, lithium alcoholates are capable directly initiate the polymerization of acrylamide [Tsubokawa121, acrylonitrile [T5 '''_121, and meth ~ crylates [T5 ~ '' 12 ~ 14l. By cons, the polymerization of ~ -methylstyrene ITsubokawa12l ~ or ITsubokawa12 ~ 14l styrene can only be carried out by activating the lithium alcoholate with a crown ether.
On the other hand, the lithiated phenyl groups can initiate the polymerization of methacrylate.
methyl and styrene [T- '''_13-14l, directly in solvents forming çh ~ your with the lithillm, such as tetramethylethylenedi ~ mine ('I ~ EDA) or ~ eY ~ methylphosphorus tri ~ mi ~ e (HMPT), or in conjunction with a crown ether in toluene. Grafting rates at the surface of carbon between 50 and 250% are obtained under these conditions.

c) Grafting by caffonic polymerization:

The carboxylic acids p- ~ feel at the carbon surface are capable of initiating the polyrn ~ ris ~ t1on by cationic pathway of direct monomers, and among these of N-vinylcarbazole lBiswas1-2, Tsubokawa16-171, N-vinylpyrrolidone IT 16-181, 2-methyl-2-oxazoline [~ 181, from spiro ortho esters, bicyclo ortho esters, and spiro orthocarbonate [T ~ 191. In all these cases, the polym ~ ric ~ tion is initiated by the attack of the acidic proton of the carboxylic acid on the monomer, the carboxylate anion subsequently serving as a counterion during the propagation reaction, the grafting effect ~ l ~ nt in the end by neutr ~ lis ~ tion of the cationic charge by the carboxylate anion, this equivalent to a reaction from It; ....; l- ~ isol-.
It is also possible to transform the carboxylic acids into an electrophile, by treating their acid chloride with a metal halide such as ferric chloride (FeCl3) [TsuLu ~ ~~ 1, O
U the silver perchlorate (AgCl04) [TslJL ~ .. -11, the synthesis of acylium perchlorate being more particularly described in Figure 2.8.
SOCI2 @ ~ AgCI04 (~ CO ~ CIO4 Figure 2.8: Carbon chemistry per ". ~ Lt ~ nl to initiate polymerations ~ cationically.
It is thus possible to initiate the cationic polymerization of the carbon from the surface of the carbon.
styrene [Tsubokawa21, Ohkita1], lactones [Tsubokawa21-221, THF [T8ubokawa20-221, formals cyclical ITe ~]. It is also possible to graft copolymers by copolymerization.
by opening the THF cycle and a cyclic anhydride [Tsubokawa23 ~, an oxepane, and respectively, a cyclic anhydride or epichlorohydrinel TSubokawa22l. The grafting of terpolymer is obtained by the ring-opening polymerization of THF, of a cyclic anhydride, and epichlorohydrin ITs ~ 1. Graft rates typically between 40 and 100% are thus obtained by polym ~ ris ~ tion by anionic.
d) Grafting by radical pol ~ ation:

Even if the carbon surface has radicals, their stabilization by aromatic polycycles limits their reactivity, eliminates ~ nt by the same, the possibility of initiating polymerizations by radical. To do this, it is necessary ~ ire to introduce initiateu ~ s on the carbon surface radicals such as azo groups ITsubokawa24-25l ~ or peroxyesters [Tsubokawa261, such as the illustra ~ m ~ have figure 2.9.

(~ CCOH

(~ OH @ ~ ~ wNt ~ R ' f ~ H -. ~ (~ C - O - O - tBu R = COCI, NCO. '8 or epoxy.
~ 0 Figure 2.9: Introduction of poly initiator, eri ~ ti ~ n radical with carbon sufface.
Thus the pelo ~ els allow to initiate the polyméri.c ~ tion from the carbon surface of different vinyl monomers such as methyl methacrylate, methyl acrylate, acrylonitrile ~ [Tb ~ 6l. Gloupem ~ nts azo allow ~ nt for their part to initiate polym ~ .ri. ~ Tion many vinyl monomers such as N-vinylpyrrolidone, 4-vinylpyridine, 2-hydroxyethyl methacrylate, glycidyl meth ~ crylate, methyl vinyl ketone ITSJL '~ 41, - ~ 1,941 27 . .
styrene, vinyl acetate, m ~ methyl tharrylate, acrylic acid, acrylonitrile lT ~ ~ 4-251, The grafting rate obtained in these different cQn ~ itionc varies between 20 and 90%.

e) Jeffamine grafting:

As we have just seen, the pos ~ ibilities of grafting polymers on the carbon surface are many and varied, both from the point of view of m ~ tkri ~ ll carbon (oxidized carbon, gr ~ phit ~,) that the nature of the polymers. Ceppn ~ l ~ nt ~ grafting via a reacdon of polym ~ ris ~ tion inided from the carbon surface n ~ ss; lellt to develop the operating cont ~ itionC to optimize the rate grafting. Indeed, as in any reaction of polymerization, the nature of the initiator, the polarity of the solvent, the ~ e of the monomer modifies drasd ~ lut; lllelll the course of the polym ~ nc ~ tion, of more, there is a compeddon between the polymerization initiated from the surface of the carbon, and the polymerization taking place in solution.

For our part, the goal is above all to obtain a carbon grafted by a solvating polymer, in order to validate the concept of the modification of the surface of the carbon grains in the composite cathode polymer electrolyte batteries. With this in mind, we have chosen the synthetic route which seemed to us the easiest to implement, and more precisely to graft via a reaction of amidation, segments of poly (ethylene oxide) monoamines from acids carboxylic present on the surface of an oxidized carbon, as illustrated in figure 2.10.

+ H2 ~ CH2 - CH2 - o3 ~ CH3 - H20 ~ (~ C ~ cH2 - CH2 - OtnCH3 Figurc 2.10: Grafting of polymer on a carbon by a ",. ~; On acids pr ~ senl ~ has its sufface.
To do this, we used mono ~ mined POE segments produced by the company Texaco under the name of kffamine ~, and as oxidized carbon the FW 200 manufactured by the company Degussa. This choice is motivated by the fact that FW 200 is oxidized carbon, commercially available, has the most surface acid functions (~ 0.61 meq / gr), on the other hand, we find in the Texaco catalog of monoaminated POE segments with a respective molecular weight of 1000 (Jeff ~ mine M-1000) and 2000 (Jçffamine M-2000). Assuming we can graft the whole acids by POE segm ~ nts, it is possible to achieve grafting rates respectively in the range of 60 and 120%.

The formation of an amide bond between carbon and Jeff ~ min ~ involving the removal of a molecule water, its ~ limin ~ tion as the av ~ ncçm.ont of the reaction promotes it. Also, the first mode of synthesis that we envisaged to carry out this reaction was to chanff ~ r under vacuum at a temperature z 160-180 ~ C, a mixture of Jeffamine M-2000 and carbon. To to obtain a relatively homogeneous paste before tr ~ itPmçnt ~ the carbon is mixed in the acetate ethyl with double the stoichiometric quantity of Jffff mine relative to the quantity presumed carboxylic acid present on the surface of the carbon. After vacuum chqllffage during z 24 hours at a temperature of 180 ~ C, the product obtained is cleaned in ether, then centrifuged, and this in order to remove the excess of Jra, this washing is repeated several times. The same operation is then repeated twice in ~ ltilisqnt a m ~ lqnge 4/1 by volume of ether and ethyl acetate, so to guarantee a complete elimination of the Jeff ~ mine not grafted. The carbon is then dried for - 24 hours at 60 ~ C under vacuum.

The grafting of J ~ ffqmine M-1000 or M-2000 on the surface of the carbon can even, qlemçnt s'çffçctller par la ~ iistill ~ q ~ tion a2éo ~ l0pique of a mixture of carbon and Jeffamine (50% in excess) in toluene.
After ~ 24 hours of reaction, the toluene is evaporated, then the carbon washed with ether, recovered by filtration, this operation being repeated several times. Washing then continues in a m ~ equivolumic lqnge of ether and ethyl acetate, and this plnc; eurs times. The carbon is then dried for ~ 24 hours at 60 ~ C under vacuum. it is interesting to note the difference in visual appearance between the call ~ not grafted with the J ~ ffamine M-1000 and M-2000. The first ~, has a dull complexion, then that the second is shiny, which indicates that the latter carbon is fully coated by the polymer, this di ~ tinctiQn being a common fact in the field of paints.

The microanalysis of the different carbons considered is given in Table 2.4, the results Theories are calculated assuming that the carbon contains 0.61 meq / gr of groupings carboxylic, and from the exact masses of Jeffamines determined by acid-base titration, lespe ~ / el, ellt from 1116 and 2219 gr.
Table 2.4 Results of the analysis of the grafted carbons.
Carbon% C% H% O% N
FW 200 95.98 0.4 2.01 FW200 / J-2000 (Azeotropy) 71.83 4.47 21.83 1.54 FW200 / J-2000 (theoretical) 73.4 5.5 19.5 0.4 FW200 / J-2000 (160 ~ C under vacuum) 73.1 4.9 22.3 1.48 FW200 / J-1000 (Az ~ otl ~ r ~) 74.19 4.37 20.05 1.38 FW200 / J-I000 (theoretical) 79.2 3.9 15.3 0.5 The result for the carbon grafted with Jeffamine M-2000, obtained by azeotropy or by dehydration under vacuum, is in agreement with the theoretical calculation. On the other hand, for carbon grafted with Jprra ~ e M-1000 the difference between theoretical calculation and microanalysis shows that there remains at the carbon surface a significant amount of Jeff ~ qmine not grafted. Even if Jeffamine M-1000 is soluble in m ~ lqnge 4/1 ether / ethyl acetate, this solvent is probably not sllffi ~ nt solvatant to eliminate ~ r Jeffamine adsorbed on the carbon surface, it is therefore appropriate ququmçnter the amount of ethyl acetate to wash the carbon grafted with Jçffamine M-1000. The call, one grafted with Jçffqmine M-2000 does not l ~ worm ~ not a dialysis member, this process should allow to clean pqrfaitçmçnt the grafted carbons, given the solubility of the Jeffamine M-1000 or M-2000 in water.

2194i27 , Figure 2.11 shows the results of the differential scanning calorimetry analysis (Cf Annex), of the grafted carbon, as well as that of the free JPrr; ~ e with reference. As we can the const ~, r, the melting temperature of the Jeff ~ grafted mine is close to that of the free JeffAmine, even if the result for the calbooe grafted with the J ~ rr ~ ED-1000 must be tempered with regard of microanalysis performed on this sample. It may seem surprising to observe such difr ~ .e. ~ e de ~ e of fusion between J ~ rr ~ "; .. ~ M-1000 and M-2000, this fact is explained ~ ic ~ mtont since JPrr ~ o M-2000 contains ~ 30% by weight of PPO units which amorphizes the polymer, the J ~ -rr ~ -; - e M-1000 does not contain ~ 17% rrexacol.

Figure of DSC of grafted carbon ~ Jeffamine Obtaining two carbon pellets grafted with the JçffAmine M-2000 by compaction under a pressure of 10 tonnes allows the density of the mAt ~ riau to be estimated at 1.24 and 1.26, a value average of 1.25. This density is consistent with that of 1.36 c ~ lc ~ e assuming that we graft 0.61 mmol of JçffAmine M-2000 (density ~ 1.1) on the surface of the FW 200 (density - 2), i.e. with a grafting rate of 135%. According to these hypotheses, we can deduce that the carbon pellet compaçt ~, pl ~ nte porosi ~ r ~, sidual ~ 9%, which is after all a fairly low value. It is possible to de ~ ..in ~ r the electronic resistivity of the grafted carbon compacted by measuring the reCictAnce of this patch using an analog ohmmeter, this leads to a conductivity 2.10-2 Ohm ~ l.cm ~ l, i.e. a conductivity of 50 Ohm.cm, which is comparable to the conductivity observed for many carbon fibers for example ~ E ~ nl ~ l, This composite material coll ~, sl, olld in fact at a volume fraction of 29% of FW 200 in the grafted polymer matrix.

Carbon grafted with Jeffamine M-2000 disperses in water and most solvents polar organic, to give a stable homogeneous suspension of carbon. These dispersions provide evidence of grafting JeffAmine through an amide bond. Indeed, it is equal ~ mPnt possible to obtain a stable dispersion of ungrafted FW 200 by adding Jeff ~ mine which forms an ionic bond with the acids of the surface. Cep ~ n ~ nt, unlike carbon grafted via a amide bond, when the carboxylates are protonated by adding an acid to the solution ~ carbon se ~ feeds largely immediate ~ t ~ mpnt Despite the plesel ~ this polymer on the carbon surface, a weak Agit ~ tion m ~ c ~ nique is the ~ ce-ssAire to dicperser the material, and this in order to unravel the polymer chains which l ~ n ~. ~ t the grains to each other. In addition, a POE 300,000 film col.l, ~ AI ~ 40% by volume of this carbon, obtained by spreading a solution of POE and carbon in the ~ étonit ~ ile, plésellte a particularly homogeneous visual aspect. A good indicator of quality of the ~ isper ~ sion of c ~ l, one in the polymer is the conSel ~ alion of this homogeneity when one t; rr ~; l ~ I a traction of the film, this fact translates the co ~ p ~ I; bilit ~ of the grains with the polymer matrix.

-If we assume that it is thus possible to disperse ~ icérnpnt the carbon grains, it is possible to admit that it is possible to obtain a composite in which the particles are ul ~ ifo ~ P.nt distributed. We can estimate the ~ lict ~ nce separating these grains, assuming the particles spherical, as a function of the volume fraction of carbon, assuming a l ~ p ~ lition in the space given ~ R. In this case, this ~ i ~ t ~ nce (dCC) is calculated for a given grain size (dc), or in the hypothesis of a placing of the grains according to a centered cubic structure, according to equation 1, or face centered cubic, according to equation 2, and this as a function of the volume ratio of calbone to polymer (~). We go back to the volume fraction (~) of carbon in the composite from equation 3.

Central cubic structure: (~) Equation 1 Center-centered cubic structure: (~ 2 Carbon volume fraction: ~ Equation 3 The dic ~ nce separating two grains is calculated on figure 2.12, for a centered cubic structure and cubic face centered, and this for a diameter of 13 nm equal to that of the particles of FW 200 [Degussa], as a function of the volume fraction of carbon in a polymer matrix such as poly (oxide ethylene). Whatever the lép ~ liLion supposed particles, we end up with a ~ list ~ n ~ c between grains of the order of 50 A for a volume fraction around 15%, representative of the yua ~ é of incoming carbon generated ~ l ~ m ~ nt in the composition of the cathode of electrolyte batteries polymers.

80 ~ Centered cubic structure 8 _ \ ~ _ d ~ 60 ~
- \ \ FW 200 / J-2000 _ \ \ compacted 20- Structure ~ \ ~ ' ~ i - face centered cubic ~ ~ -OII ~ 'I ~

Carbon volume fraction (%) Figure 2.12: Distance between carbon grains as a function of crystal structure This ~ ist ~ nce of 50 A is the one from which we can assume that the electrons are moving by tunnel effect between the carbon grains, without it being necessary for the particles to be physically connected to each other. As we can see, the fraction volume co ~ olldant with CalbGI ~ graft compacted, glues ~ olld to ~ lictanr ~ es less than 20 A, this explaining the low resistivity of this m ~ t ~ ri ~ l Under the same conditions, the distance which s ~ pa, ~. ~ .l grains of black Shawinigan, with a diameter of 42.5 nm [Banl1], would be of the order of 200 A. The conducl; vi ~ high of the black of Sha ~ -vinigan can therefore only be explained by the ~ n of carbon grains according to a tri ~ imen ~ ional network according to which the electrons move in the electrode.

This mechanism of displacement of electrons by tunnel effect is recognized to explain the conductivities obtained with carbons with high electrical conductivities such as Ketjenblack (Akzo) [AkZo ~ verhel ~ Nelsonl. This carbon presents a special case due to its slluclu ~ ressr.mbl ~ nt to that of an empty shell, as depicted in Figure 2.13, which increases the probability that two aggregates él ~ menl ~ i.e. either s ~ ffl.c ~ mment close to each other so that the electrons can move by tunnel effect. So, just load a polyethylene film with ~ 2.5% by volume of Ketjenblack to obtain a resistivity # 50 Ohm-l.cm-l [Akzol, identical to that of the FW 200 grafted with Jeffamine M-2000 in the compact state ~.
It ~, '_ .; 7 Figure 2.13: Structure of an aggregate of l ~ tjen ~ l ~ ck, after tAlczo].
This type of carbon due to its high specific surface (-1000 m2 / gr) rapidly increases the viscosity of the charged polymer rVe ~ elstl, which risks rapidly inducing a f ~ l ~ ncing of the m ~ .ri ~ ll On the other hand, it is likely that there is an interest in the joint use of grafted carbon and Ketjenblack in the cathode of polymer electrolyte batteries, the first ensuring homogeneous distribution of electrons on the surface of the insertion m ~ - ~ ri ~ .., and a small amount of second (3 2% by volume) ensuring a high overall electrical conductivity. Such a composite could be described, from a formal point of view, as an interpenetrating network of two types of cond ~, c ~ u ~ éleclr ~, ~ ue.

On the other hand, in terms of technology to manufacture ~ ion ~ we can im ~ iner to obtain such a composite by an extrusion process, or such a process by the co .. ~ es m ~ c ~ ues it involves e, ~ îne the desiuu ~; ~ on aggregates. Extrusion is therefore favorable for optimal dispersion of 2194 ~ 27 -grains of c ~ bone both grafted as well as Ketjenblack. Now, a homogeneous dispersion is the goal ~ h ~.; llé for the grafted carbon, and the condu ~; livi ~ of KPtjenbl ~ rl ~ is little ~ ffçctée by the process of mixture [Akzol. These two types of carbon are therefore complementary both in terms of ~ n ~ nities that of the évenhlellP ~ manufacturing process of the cathodes. This would not be the case, if one iliC ~ it the carbon most used in the dom ~ in-P, of c ~ tteries, the black of Shawinigan, whose résisdvité ~ lug ~ e ~ -t ~ when the aggregates are broken under the effect of co, lll ~ ltes m ~ c ~ ni (read ind ~ es by the process of m ~ l ~ ngP, These grafted carbons have not currently been tested in battery, but it seems clear that the principle of grafting a polymer on the carbon surface can bring clear improvements in the particular field of b ~ tteries with polymer electrolytes, both in terms of the anode than the cathode.

II-2-3 Salt grafting:

As we have previously exposed, the use of polymer electrolytes with grafted anions, and therefore purely cationic transport, allows, a priori, to improve the perform ~ nces of b ~ tteries polymer electrolytes, and especially during calls of pnics ~ nce important. Also, it can be interesting to graft on the surface of carbon grains dissociated anions, which would allow to obtain loc ~ lement both an electronic conductivity, and a unipolar ionic conductivity, after dispersion in a solvating polymer.

The grafting of malononitrile to the surface of a carbon carrying gro - pem ~ nt ~ s carboxyls, such as FW 200 is one of the ways, relatively simple to implement, which should lead to such a material. In addition, as we have seen prérédPmm ~ have, the conductivity of the lithium salt of poly (acetylmalononitrile methacrylate) being close to 10-5 Ohm ~ l.cm ~ l at 50 ~ C, the lithium salt malononitrile grafted onto the carbon surface should therefore be at least as conductive, the aromatic nuclei cond ~ nsés to which is connected the acid being more attractive than a group alkyl. On the other hand, if there can be a doubt as to the stability of the anions containing groul ~ c; ll.ents nitriles with respect to a lithium electrode, the oxidation potential of the anion benzoyl ~ llalononitrile in DMSO - 4.5 Volts g ~ r ~ ntit the electrochemical stability of this salt in a cathode typically operating between 1.5 and 3.5 Volts. All this justifies, in a first time, the choice of grafting the malononitrile on the surface of the FW 200, in order to validate the principle to salify the carbon surface on the ~ s performance of the cathode of electrolyte batteries polymer ~ s.

The reaction allowing the grafting of malononitrile is sch ~ m ~ ticked in Figure 2.14. More precisely 4 gr of FW200 (~ 2.44 meq of carboxylic acid), are treated with 0.5 cc of oxalyl chloride (~ 6 meq) in ~ 10 cc of THF at 0 ~ C, this results in the release of CO2 ~ i 941 27 , - ~
for a few minutes. At the same time, a solution of 322 mg (~ S meq) demalononitrile in # 10 cc of THF at 0 ~ C, is treated with 80 mg of lithium hydride (~ 10 meq).
After 30 ... h .. ~ P ~, the col-tPn ~ nt carbon solution is centrifuged to ~ remove excess chloride of oxalyl, and then the anion of m ~ lonitrilP is added, in solution in THF to the treated carbon. After 2 hours, a little water is added in order to remove the LiH which may have remained in the solution, then the carbon is washed several times with THF. The carbon is then dried for ~ 24 hours under vacuum at room temperature.

(~ COOH> (~ CO ~ LiH / a ~ 2 (CN) 2 ~ (~ HERE - C ~ Li Figure 2.14: Grafting of a deion '~ c ~ with the carbon sufface.
The result of the microanalysis carried out on this carbon is given in Table 2.5, it does not inconsistent with the theoretical calculation assuming that 0.61 meq of lithium salt is grafted malononitrile on the carbon surface. On the other hand, the lithium salts generally being hygroscopic, if we assume that the lithium cation is in the form of a hydrate containing 21 molecules of water per lithium, the theoretical calculation is in good agreement with the result of analysis. The assumption of such an amount of water is not meaningless since unlike a free salt, the salts grafted onto the carbon surface cannot be organized into crystals, ~ ugmçnt ~ nt thus their hydrophilic powers. The excess of lithium can be explained by the presence of traces of LiOH adsorbed on the carbon surface (~ 30 mg / gr).
Table 2.5 Result of analysis of the carbon grafted with rnalononitrile.
Carbon C% H% N% 0% Li%
analysis 80.26 1.8 1.88 / 1.15 theory 97.8 0.3 / / 0.4 Li, 21 H2O 80.4 2 0.6 16.4 0.3 II-2-4 Conclusion:

Even if these grafted carbons have not yet given rise to b ~ ttPries tests, it has been possible to show on the one hand the richness of Fennett's chemistry to modify the carbon surface, and on the other share the po ~ ibilit ~, to graft short segments of poly (ethylene oxide) on the surface of a carbon oxidized commercial. This chemistry may be extended to the grafting of delocalized anions which could limit the development ~ .llell ~ of gr ~ iPntc of concpntration apps ~ 5 ~ iu during the discharge of the battery.

The principle of modifying the carbon surface is not limited to the composite c ~ tho ~ e, this concept pOull ~ it prob ~ hlemPnt be extended to the anode by grafting ~ oegrn ~ nt ~ poly (ethylene oxide) to the surface of c ~ l.one serving as anode in bqtteri ~ ps ~ like grains of glaphile, for b ~ tterips d ~ -no ~ ~ s "rocking chair", in which it plays the role of an insertion electrode.
.

~ 1,946 27 II-3 REDOX SOLVENT POLYMERS:

II-3-1 Introduction:

A redox solvating polymer is a material that contains both segments of a polymer solvating agent, such as poly (ethylene oxide), capable of l. ~ o, ler ions, and an electroactive function allowing the transport of electrons. L ~ m ~ c ~ ni.cmc depl ~ cPment of electrons can be done by successive jumps between an electroactive function in an oxidized state and a function in a state of lower oxidation, among these materials, let us quote the combed or sequenced copolymers of the phenothi ~ 7ine [;::, f ~ l, ocene [W ~ '3 ~ r ~ 1, or methylviologene la'an91 with segments poly (ethylene oxide). The electrons little ~ el ~ t eg ~ l ~ ment move by a mechanism of transport coop ~ if, and this speci ~ l ~ -m ~ nt in the case of polyimides formed by con-lPnc ~ tion of segments of poly (ethylene oxide) ~ min ~ s with dianhydrides of aromatic molecules, within a conduction band resulting from the covering of the 7 ~ orbitals of the lBanl1 aromatic functions 2].
As will be explained below, these m ~ tkri ~ llx appeal to the notion of molecular dlempilpmentc to explain their plupliety of electronic con ~ iuction.

Electroactive functions can be distinguished by different electrochemical criteria, such as by example, the number of electrons exchanged by function, the potential of the redox couple (s), the kinetics of electronic exchange. But from the point of view of chemistry, the essential parameter is the nature of the charge present on the electroactive function.

POLYMER
REDOX
CA ~ ONIQUE ~ + X

Me Me Me M + ~ M +
O = C ~ N ~ C_O -O - C ~ = O O - C ~ = O
~; REDOX

f ~ C ~ N ~ O r ~ C ~ N ~ o ~ C ~ N ~ O;

Figure 2.15: Exeir ~ es de poiy ", eles a mixed conduction an ~ n:, eu and c ~ lioni ~ ue.

~ 1 94 ~ 27 , ..........
As shown in Figure 2.15, methylviologen, for example, exists either in neutral form or c ~ tioni ~ read ~ depending on its oxidation-reduction state. The same is true for ferrocene or phenothi ~ 7ine, too, the redox polymers integrating them into their structures will be classified in the family of redox polymers c ~ tioniqlle ~

InvP ~ r ~ men ~, polymers containing diimides of electroactive aromatic rings, such as diimide of b ~ e (see Figure 2.15), of narht ~ lPnP., or of perylene, being either in neutral form or anionic, they will be considered subsequently as anionic redox polymers.

The chemistry of lithium batteries is a basic chemistry, in fact, the traces of water present in the electrolyte, the chain ends of poly (ethylene oxide) copolymers, or certain polym catalysts ~ ri.cation dor-nPnt strong bases by reaction with lithium, as by example of lithium hydroxide. Consequently, the cationic redox polymers may undergo a nucleophilic attack by bases present in such an environment, limit ~ nt by the same the lifetime of these materials. On the contrary, liquid amines, including liquid ammonia, are solvents of choice for the study of the electrochemical properties of hydrocarbons with characteristics aromatic 1 ~. Indeed, the very weak electrophilic power of the amines makes it possible to observe in these solvents the reversible reduction of aromatic molecules, such as perylene for example. The anionic redox polymers are therefore, a priori, compatible with the chemistry specific to batteries at lithium, in particular to polymer electrolytes, as he will try to prove later.

II-3-2 Cationic redox polymers:

Despite the doubt as to the long-term stability of cationic redox polymers in basic medium, it is still interesting to seek to extend this family of polymers, the applications poten ~ ielles of these materials beyond the field of lithium b ~ tteries. In this context, our interest was focused on the copolymers of ph ~ n ~ 7ine and 4,4'-bipyridine with segments solvents.

The synthesis of polyph ~ n ~ 7ine, to our conn ~ iss ~ nce origin ~ le is described in Figure 2.16. The met ~ tion of the nitrogen cycle increases consider ~ blem ~ nt their nucleophilic power, so the ph ~ a ~ e treated pre ~ l ~ blt-m ~ nt with sodium m ~ t ~ eu in the ~ im ~ thoxyethane allows to obtain, by reaction with alkyl halides, the N-alkylated derivatives of ph ~ n ~ 7ine ['''~]. For our part, we obtained the di-lithiated derivative of ph ~ n ~ 7ine by co-grinding a quantity st ~: chiometric of lithium with ph ~ n ~ 7ine in THF. More specifically 541 mg of ph ~ n ~ 7ine (3 meq) and 42 mg of lithium (6 meq) are weighed in a glove box in a sealed bottle polyethylene, we then add ~ 10 cc of THF, and zirconia rollers. THF takes rapi ~ emen ~ a dark mauve color, typical of monolithic phen ~ 7ine. The bottle, under argon atmosphere, is then rotated on itself, in order to carry out the co-grinding via the zirconia rollers. After ~ 24 hours, one obtains in suspension in THF, which has become colorless again, a pléci ~ orange island of 9,10-di-Li 9,10-dihydroph ~ n ~ 7in ~. 561 mg of 1,2-bis (2-chloroethoxy) ethane (3 meq), the reaction of the dichloro compound is practically immediate ~ te After one hour of grinding in the flask, the solution is centlir ~ légé, and thus the polyph ~ n ~ in ~ insoluble in THF. The polymer thus obtained is then dried for # 24 hours vacuum at 60 ~ C.

co-grinding Li (-CHzOCH2CH2Clk HN N THF ~ UN ~ U THF; --N N - (CH2CH20) 2 - (CH2) 2--Figure 2.16: Synthesis of polyphenazine.
The nitrogen of 4,4'-bipyddine is suffic ~ mmPnt nucleophiles to be quaternized, for example with segments of di-halogenated polymers, following the reaction of MenchustkinlB ~ Uridahl.
Unfortunately there are few di-halogenated derivatives of poly (ethylene oxide) available.
commercially. Even if it is possible to synthesize these materials from poly (ethylene glycol), we preferred to use another synthetic route. Poly (ethylene oxide) being cut by electrophiles, by treating high mass POE with trifluorom ~ th ~ nesulfonic anhydride, we get POE segments functional ~ licensed by trin ~ your ~ this grouping is known for be an excellent group leaving lMarChl. The advantage of such an approach is to be able to have functional segments: ~ read, the average mass of which is determined solely by the quantity anhydride used, Figure 2.17 describes sch ~ m: ltique this synthesis. More specifically, we prepare a solution of 3 gr of POE 9.105 in 20 cc of anhydrous dichloroethane, and this under argon atmosphere. This solution is brought to ~ -20 ~ C by a liquid methano-nitrogen mixture, then 940 mg of trifluorometh ~ nesulfonic anhydride (3.33 meq) are added to the solution, this stoichiometry co.lespond to an average mass of 900 for the POE segments functional ~ li ~ s.
After one night, the solution has returned to room temperature, and we can observe a clear decrease in viscosity of the solution. 521 mg of 4,4'-bipyridine (3.33 meq) are then added in solution in 10 cc of ac ~ to ~ island. After a few hours, the solvent is evaporated, the polymer taken up in acetonillile, then ~~ p- ~ cipitate in ether. The polymer is then dried for ~ 24 hours at 60 ~ C under vacuum.

_ ~. ~, (CF3So2) 2c ~ Dc ~ CF3SO3 ~ a ~ 2a ~ 2 ~ 02SCF3 N
M = 900,000 ~ 20 ~ C m ~ t ~ r. 'r; ~ e -'03SCF3 CF3S03 ~ 2a ~ 2 ~~
; Figure 2.17: Synthesis of polyv ~~ .'Dgène.

21 94 1 2 ~

Analysis by c ~ lnrim ~ differential scanning sorting (see Annex) of the polyviologen, for a speed scanning speed of 10 ~ C / min, allows the determination of a lP- ~ of tr ~ nCition vill ~ use of the polymer -40 ~ C. By cons we do not observe peaks of fusion or recrystall ~ lli.c ~ tion, this indicating the c ~, t ~ - ~; amorphous polyviologen [GUillaudl.

Voltammetry ~. cyclic for two scans succes.cif.c of the polyphenazine on one micfoéle ~ il, ude of 125 llm (see Appendix) is given in figure 2.18. The deposit on the mid ~ roelectrode is made by dipping its end in a dispersion, obtained in acetonitrile, of polyph ~ .n ~ 7.ine in POE 9.105, several times after evaporation of the solvent. The microelectrode is then dried for # 24 hours at 100 ~ C under vacuum, and placed on its support with respect to a POE 5.106 electrolyte containing LiTFSI at a concentration O / Li _ 20/1. The voltal ~ ro, l.étrie is obtained using an EG&G 270 potentiostat for a scanning speed of 2 mV / s between 1.5 and 4 Volts, and this at 80 ~ C.

8T = 80 ~ C v = 2 mV / s Ep ox = 3.65 VA
On ~ o ~ e hvdedePt12511m E ~ 2 = 3.2 V ~

~ ip ox -8 ~~ Ep red = 2.77 V ip red 1.4 1.8 2.2 2.6 3.0 3.4 3.8 4.2 E (Volts) vs. Li + / Li ~
Figure 2.18: Volta ", ~ erui" F: l, ie cyclic at 2 mV / s of poly ~, héna ~ i "e for two successive cycles ~ ~.
As we can see, the flight ~~ elected ~ cyclic lletry of polyph ~ .n ~ 7ine, because ~ cté.n.cti-lue of a quasi-reversible electrochemical system, has two redox couples at 3.2 and 3.8 Volts (this last value being confirmed by another e ~ h ~ nt ~ tiQn) ~ the corresponding redox reactions are given, Pc in Figure 2.19.

[~, ~ E - 3.2 V ~, ~ E = 3.8 V [~

Figure 2.19: Reactions, redox of phenazine.
The polyviologen gave rise to a battery test (see Annex) following a discharge regime intonciost ~ tirlue at a current density of 30 IlA / cm2, and this at a temperature ~ 80 ~ C, the result of this e ~ .nt ~ tion is given on figure 2.20 [Guillaud 2, 8 ~
2, 6 - ~ Polyvlologene _ 2.4 2.2- ~ -1, 8 -1.6--1, 4 0 0.5 1 1.5 2 2.5 e by ".Dléc- ~ 'e Figure 2.20: Déch ~ rye i "tensio ~ ti ~ .Je à 30 ~ Vcm2 du Polyviologene.
To evaluate the redox potentials co ~ ondant to these couples, Figure 2.21 preS ~ ntent the result of the inverse derivative of the potendel as a function of time, we thus obtain a curve similar to that obtained by a cyclic voltammetry. So, as we can see, the polyviologène has two steps to reduce an average potential of 2.6 and 2.15 Volts respectively. These values are in agreement with the results of the ~ Boundahl literature for the dibromo poly (decaviologene) in cyclic voltammetry.
~ I ~ I ~ I ~ I ~ I ~ I -~ ~ r UJ

1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 E (Volts) vs Li + / Li ~
- - Figure 2.21: Inverse derivative of p-, len ~ iel as a function of time (smoothed).
The .echa ~ gcability of this electrode is poor, on the other hand we only discharge 1.5 electrons per viologen pattern instead of 2 in theory, relative to the two couples at 2.6 and 2.15 volts.
The electrolyte used for this test having been uni ~ uelllellt placed in a glove box for a few days to dry it, after evaporating the solvent in air, it is likely that it still contains 4 ~; not negligible strips of water. It can therefore be assumed that the lithium hydroxide generated at the surface of the lithium attacks the viologene, and makes the reacdon high ~ u ~ - irreversible.

.
II-3-2 Anionic redox polymers:

KaptonQ or poly (4 ~ 4 ~ -o ~ ydiphenylèn ~ p ~ cllitimide) is a m ~ qtéri ~ q. ~ L of choice of ~ in (~ ustrie electronic, e ~ nl; ell ~ om ~ n ~ due to its eYc ~ nt ~ held at te-mp ~, cl ~ ih ~ ues agents, and of its resistivity of 10l6 Q.cm. Despite these properties, when the Kapton ~ is plasticized by a solvent, pyrom.qllitimide is able to accept reversibly two electrons. Once dope, the O ~
polymer becomes a semiconductor of ~ N ~ O ~
resistivity from 105 to 107 Ohm.cm IKraUse ~
Recently, the electrochemiflue properties of o O n diimides of aromatic nuclei were added to KAPTON FORMULA
profit for the synthesis of solvating polymers redox obtained by polycondensation of aromatic dianhydrides with poly (oxide) segments ethylene) ~ 1iqmin ~ s [Baril1-2], The cited work esse-ntiell ~ ment on polyimides useful the narhtqlèn- ~. and perylene as the aromatic nucleus.

Given the potential interest of anionic redox polymers, there is an interest in extending this family of materials by studying the possibilities of using other aromatic nuclei than nqphtql ~ ne or perylene, and more particularly ben7 ~ ne, benzophenone, pyrazine, and dinitrobenzene [GUillaudl. But first, it appeared important to assess the possibilities to improve the synthesis of these polyimides.

II-3-2-1 Summary:

The synthesis of polyimides can be carried out by the polycon ~ enc ~ ion of compounds di ~ min ~ s and dianhydrides, as shown in Figure 2.22. The amine is sufficient ~ nucleophile to initiate the opening of the anhydride, this leads to polyamic acid which by dehydration gives the polyimide desired.

OOOOHH
~ N--O ¦ 1 o + H2 N-- --NH2 ~ HOOC-- J - COOH n ~ ~ Acid pol ~ e OO.
~ 2 H20 --N ~ N

We ~ 1 ~ '' Figure 2.22: Synthesis of a p ~ iy; ".._ 'e by condens ~ tion of a diamine and a dianhydride.

In the dom ~ ine of electronics, we directly use a polyamic acid solution, deposited at the spinner on the substrate. After evaporation of the solvent, the polyamic acid is cyclized by a ; lr- ~ llt thPrrnique à ~ 350 ~ C lK ~] or by treating it with an equivol solution ~, pyridine lique and IMaZurl acetic anhydride.

Polly ~ t ~; c ~; we dianhydride n ~ rht ~ lane or perylene with Jeff ~ ED-900 mine in the DMF, in the form of polyamic acid at first, then its cylisation by mixing pyridine /: lcétQnitrile does give electroactive polyimides [Banl11 However, this process has the drawback of using a solvent with a high boiling point (Teb = 153 ~ C) of which it is relatively tliffi ~ island to get rid of & -lasser.

First, it is possible to obtain the polyimides by Jeffamine polycondensation ED-900 or ED-2000 with respectively the dianhydride of ben7 ~ ne, napht ~ l ~ ne, and perylene, direct.oment in cha ~ -ff ~ nt a mixture of polymer and dianhydride under vacuum at 160 ~ C. More specifically the Jeff ~ mine samples are initially degassed and dehydrated, bubbling dry argon for - 2 hours in the molten polymer brought to a temperature 100 ~ C. The exact mass of Jeff ~ min ~ is then determined by determining the amines te rnin ~ lP in water with a standard solution of hydrochloric acid. Subsequently, 10 gr of Jeff ~ ED-2000 mine or ED-900 and the quo ~ itG stoichiometric respectively of ben7 ~ ne ~ naphth ~ l ~ nP, or perylene are weighed in a glass cell. The cell is then placed in an oven, connected to a pump vacuum, then the oven brought to a temperature of 100 ~ C for 2 hours, and overnight at 160 ~ C by the following. There is a marked increase in viscosity over time, indicating that the polycon-lenc ~ tion is carried out under the conditions described above with the dianhydride of benzene and napht ~ l ~ ne ~ It is interesting to note the similarity of this synthetic route with an extrusion to hot of a mixture of Jeff ~ mine and dianhydride, this process could very well be a way inrlllstrielle access to these m ~ téri ~ m ~

On the other hand, the two tests with perylene do not lead to a polycondensation reaction, the Jeff ~ mine remaining in the liquid state, moreover we can still clearly see the grains of the dianhydride perylene. It therefore seems that the low solubility of perylene prevents Jeffamine from reacting with anhydride. Polycondence ~ tion of polyimide with perylene dianhydride is fine ~ l ~ mPnt obtained by bringing to reflux, under argon, a ~ luanlilG éqUimol ~ ire of JeffAmine ED-900 or ED-2000 in the DMF hangs ~ nt ~ 24 hours. The DMF is then evaporated using a rotary evaporator.
The polyimides obtained are then dissolved in ~ celon; I. island, centrifuged to remove impurities p ~ nt ~ s in the solution, then .G ~ rGcipités in ether. These polymers are then vacuum dried at 80 ~ C for ~ 24 hours.

Subsequently, during a new synthesis of these m ~ t ~ ri ~ c at a temperat - re of 180 ~ C, it appeared that the polyimides obtained were insoluble in acetonitrile. Likely ~ hl ~ m, have polyimides 2 ~ 941 27 may crosslink when the cyclisadon temp ~ l ~ lur ~ is too high, probably by attack from Jeffamine oxygen by carboxylic acids, relatively acidic (pKa ~ 2), acid polyamic. Also, another synthetic route by diC-tilI ~ tio ~ azeotropic of Jeffamine and a Dianhydride in toluene has been considered for the synthesis of polyimides. The tests were performed with ben7 ~ dianhydride, respectively with Jeff ~ min ~ q ED-900, ED-600, or ppO ~ min ~ s mass 400. In order to limit the risk of crosslinking, four solutions are added to the solution.
equiv ~ lcnts of triethylamine p ~ r equivalent of dianhydridc. Once the scl .. ~ iGn de Loiuene at reflux, a quantity of water is observed in the condensing finger of the Dean-Stark apparatus, substantially equivalent to the stoichiometric quantity corresponding to the polycon ~ çnc ~ tioll, and this after a period of time ~ l hours. After overnight reflux, the toluene is evaporated, the polymer taken up in the acétQnitriIe, centrifuged, ~ JleCi ~ in ether, and finally dried ~ dried for ~ 24 hours at 60 ~ C under vacuum. With the same starting monomers, polyimides which are soluble in acetonitrile even in the ~ absence of triethylamine, the reaction taking place at a moderate temperature (Teb toluene = ll 1 ~ C) which is not suffic ~ nte to crosslink the material.

The det ~ l ...; .. ~ tion, by size exclusion chromatography (Cf Annex), of the average mass of polyimides makes it possible to compare the influence of their synthesis process on the reaction of polycon ~ encqtion. To this end, Table 2.6 groups together the results obtained for the copolymers of benzene dianhydride with different solvating segments r ~ i ~ min ~ s ~ GUillaudl. From the mass number average, it is possible to determine the average number of component monomer units polyimide ch ~ înPs. Except for the polyimide with Jeffamine ED-2000, this number is between lS and 30 units, which is a good result for a polymer obtained by a reaction of polycorldçnc ~ tion.
Table 2.6 Influence of l "~ t ~ of poly synthesis;, l, des on the polycondensation reaction.
Polymer Mass Mass Number Index Method diamines by weight in number polydispersity of cycling unitsPPO-400 9100 4600 2 15 A
J-600 21400 10300 2.1 27 A

J-900 21800 14200 1.5 20 B
J-2000 16500 9700 1.7 8 C
A: Azi distillation: vtr read in toluene in preseI ~ e of triethylamine.
B: Poly ~; ondensation in acetonitrile at 20 ~ C, then cy., Ii ~ tWn of polyamic racide at 160 ~ C under vacuum.
C: Cyclization at 1 60 ~ C under vWe.
The low number of monomer units with the kff ~ ED-2000 mine can be explained by the process of synthesis. Indeed, we do not start from a homogeneous mixture of aromatic dianhydride and Jçrr ~. ,, i,. &, Which ~ 1iminIle all the ren-lem ~ nt of the polycûntlen ~ qtiorl On the other hand, as illustrated the results ~ tc with the Jeffamine ED-900, we get a polymer sencibleme-nt equivalent as well by (li ~ till ~ n azeotropic, that by hot cyclization of polyamic acid, previously obtained in solution in a ~ ~ tQ ~ ihiIe It therefore appears that the synthesis by azeotropy or by hot cyclization of the polyimides allows to obtain average masses between 10,000 and 20,000, which is s ~ lffls ~ nt for a apply ~ fion in the ~ sb ~ tt ~ ries with polymer electrolytes. Indeed, the polyirnide is in this case charged with carbon grains and insertion material which serve as reinforcement for the polymer. Moreover, he low mass polyimides are unlikely to diffuse into the electrolyte based on poly (o ~ ethylene cyde), poly ~, s of chemical nature dirr ~. ~ ente being generally inr ~ o ~ patible between them. On the other hand, the synthesis of polyimides by azeotropy is the most advantageous to use.
laboratory scale, due to its ease of implementation, and the possibility of synthesizing batches of polymers of one hundred ~ ine grams each time.

Another p ~ m ~ tre ~ cc ~ ntiel to consider is to ~ determine if all of the polymer is well cyclized in the form of polyimide under the conditions of synthesis prece ~ e-mment described. To do this, the the simplest method is to detach a battery (see Appendix) whose cathode is composed of a mixture of carbon and polyimide to be studied, which actually corresponds to carrying out an assay coulometric of the imide functions of the polymer. Figure 2.23 shows the result of the intensiost ~ tick discharge at 30 ~ lA / cm2 of the polyimide obtained by a_eotropy, from the dianhydride benzene and Jçff ~ ED-900 mine (B / J-900), for a battery with a capacity of ~ 0.16 mA.h / cm2.
3.5 1 ~ l 3 BlJ-soo ~ -t_ '2.5 -o 1, 5 -oo, 5 1 1, 5 2 e per molecule Figuro 2.23: Discharge i "lel ~ sio ~ eu at 30 ~ LA / cn ~ of the B / J-900 obtained by azeotropy.
As we can co ~ e ~, the quantity of charge extracted from the glued polyimide:, yond e "~ e ~ nt the equivalent of two electrons per molecule of pyrom ~ .llitimi ~ e, this result validates the process for the synthesis by distill ~ tioll a_éotropique of polyimides. Similarly, trials similar on the polyimide of ben7 dianhydride ~ nP. with the Jerr ~ "; ~ lo ED-900 or ED-2000, obtained by cyc! i ~ tion hot, confirrn.-nt that in these con-litions the whole polymer is well cyclized.

2 ~ 94127 II-3-2-2 Electrochemical test:

~ .cssons we m ~ inten ~ nt to the study of the electrochemical properties of various dianhydrides of aromatic rings (ben ~ e, benzophenone, pyrazine, dinitrobenzene) copolymerized with J ~ rr ~ ç ED-900. The reslllt ~ tc are obtained from batteries (Cf Annex) whose cathode is composed of a m ~ nge of the polyimide studied and of carbon, using a Mac Pile ~ 9 device (Bio-Logic) developed to eLfecluer from flight ~ on ~ slow caliper.

a) Benzene: -Figure 2.24 p, ~ feels the cyclic voltammetry of the polyimide of the dianhydride of ben7 ~ ne with the Jçffamine ED-900, obtained by azeotropic distillation, which we will call more simply by B / J-900. This cyclic voltammetry, typical of a quasi-reversible system, shows that the material pl ~ senle two redox couples well de ~ set respectively at 2.16 and 2.5 volts, both couples being separated by ~ 340 mV.

T = 80 ~ C ~ ~ E ~ 85 mV
POE / LiTFSI 20/1 / \ P

v = 84 mV / hr Epred2 = 1.94 V Epred1 = 2.27 V

1.4 1.6 1.8 2 2.2 2.4 2.6 E (Volts) vs Li + / Li ~
Figure 2.24: Vo / tai "~ é, u", ~ lri ~ cyclic at 84 mV / hdu B / J-900.

The results given in table 2.7 make it possible to co, l, counter the redox potentials of the diimides of ~ l ~ ne, naphthalene [Banl1], and perylene IB ~ nlll with Jçffamine ED-900.

Table 2.7 redox rolerdiel of the diimides of benzene, napl, l Y ~ e, perylene with Jeffamine ED-900.
CoreE ~ l (Volts) E ~ 2 (Volts) ~ E (mV) Method Speed ~ ~ n ~ ene 2.31 2.16 340 Volt ~ "~ rA" e ~ ie slow 84 mV / h Naphthalene2.6 2.35 250 11elecrode 125 llm 2 mV / s Perylene 2.48 / Volt ~ ,, ~, e, u ,,, ét-ie slow 10 mV / h 219ll ~ 27 ._ It is inlél ~, ssanl to note that the difference between the first and the second redox couple tlimin ~ le as and as we ~ ugJnente the size of the aromatic nucleus, it is likely that this is related to the ~ ugm ~ nt ~ tiorl of the ~ ict ~ nce separating the two imide functions. So, for perylene, the two imide functions are sllfflc ~ mment distant so as not to be influenced one relatively to the other, the red ~ ction s'Prrecll ~ e in this case by a m ~ C ~ nicelectronic.
b) Benzophenone:

Cyclic voltammetry of the polyimide of benzophenone dianhydride with Jeff ~ mine ED-900 (BzP / J-900), obtained by azeotropy, is given in Figure 2.25. This polyirnide presents three peaks of reduction of a potential at the top of the peak of approximately 2.1, 1.75, and 1.4 Voltc, in agreement with the r ~ slllt ~ tc of the literature on the polyimide of the benzophenone dianhydride with the di ~ mi ~ ino phenylindole [V ~ hbeck].

BzP / J-900 f O ~

~ -1 O- ~

- 2 o ~ 3 ~ ~ v = 10 mV / h 0.5 11.5 2 2.5 3 3.5 E (Volts) vs Li + / Li ~
Figure 2.25: Voltai "~ eru", cyclic ts ~ rie at 10 mV / h of BzP / J-900.

Unlike these latter results ~ tc, we do not observe oxidation peaks in back sweep in the present case. The redox reaction corresponding to this polyimide is given in Figure 2.26.

--N ~ ON-- --N ~ N--o OO O ' --N ~ N-- ~ --N ~ N--Figure 2.26: F / redox eations of the ben2 ~ henone diimide.

~ 1,941 27 It is inlél ~ cs ~ ll to note that everything happens as if one superimposed the redox properties of the diimide with that specific to benzophel-one. The pl ~ nce of a r ~ ire species with low potential, conjugated to the lelll ~ ule of 80 ~ C cycling, can explain the degradation of m ~ téri ~ u; by cutting a bond c ~ l, one carbon, or by a reaction of collp1age between two species r ~ rli ~ ires by ~ emrle c) Dinitrobenzene:

The dianhydride of dinitroben7 ~ not being commercial, it is synthesized by oxidation of dinitrodurene, as shown in Figure 2.27. A solution in 2 liters of water of 30 gr of dinitrodurene (0.159 moles), 118.1 gr of potassium carbonate (0.855 moles), 211.5 gr of potassium permanganate (1.34 moles), and 20 g of tetrabutylammonium bromide (0.064 moles) is brought to reflux with mechanical agitation for ~ 72 hours. The solution is then filtered in order to ~ eliminate the manganese dioxide formed during the reaction. After having concentrated this solution to 3/4 using a rotary evaporator, it is aci ~ ifi ~ e with 80 ml pure hydrochloric acid (pH ~ 1). Then extracted with five fracdons of ethyl acetate, then three ether fractions. The organic fractions are mixed, s ~ h ~ e. ~ With sodium sulfate, then the whole evaporated using a rotary evaporator. The product obtained (6 gr, 10% yield) melts at a temperature of 230 ~ C, which agrees with the melting point of the acid dinitropyromellitic given in the literature [Smith]. The poor performance obtained can be explained probably by the low solubility of this tetraacid in organic solvents.
NO2 N ~ 2 o N ~ 2 o H3C ~ 1 CH3 HOOC_ 1 COOH
H3 ~ KMnO4 ~ ~ COOH THF ~

N02 N ~ 2 ~ N ~ 2 ~
Figure 2.27: Synthesis of dinitrobenzene dianhydride.
This product is then cyclized to its dianhydride by treating it with two equivalents of anhydride trifluoroacetic in cold THF. After one hour, the THF is evaporated using an evaporator rotary, and the dinitrobenzene dianhydride stored in a glove box. This product allows you to polyimide coll ~, ~ undulating with J ~ ffamine ED-900 (DNB / J-900) by azeotropy.

Figure 2.28 shows the intentiostatic discharge curve at 30 ~ A / cm2 of the DNB / J-900. n is surprising to con ~ t ~ ter that the diimide of dinitrobenzene is capable of accepting four electrons in reduction Malheu ~ se ~ .e ~ -t ce m ~ tér ~ pr ~ llte mediocre cyclability. A discharge limited to an electron does not improve the cyclability of the material, this low reversibility of the dil.i ~ rllitimide is not app & lc llent not linked to the fact of illje ~ r four electrons in the molecule.

~ 1,941 27 ,.

3 1 1 1 3.2 DNB / J-900 ~ DNB / J-900 ~
o ~ 3 - _ + 2.5 discharge 2.8 ~

u ~ \ \ 2.6 ~ ~ r discharge 2nd discharge ~ -- ~ 2nd discharge ~~
1.5 1 1 1 ~ 2.4 0 1 2 3 4 0 0.2 0.4 0.6 0.8 e ~ by molecules e ~ by molecules Figure 2.28: Inten ~ io, ~ eu discharge at 30 ~ A / cm2 of the DNB / J-9oo.
The calculation of the inverse derivative of the potential as a function of time makes it possible to simulate a discharge potentiostatic of the polyimide, and thus to specify the potential of the redox couples, as illustrated by Figure 2.29. During the first discharge up to a potential of 1.5 Volts, three peaks in reduction are obtained with a respective potential of approximately 2.7, 2.5, and 1.8 Volts. However, the second discharge only appears a peak at 1.78 Volts, thereby illustrating the degradation of the polyimide. By limiting the discharge to an electron by dinitropyromelli ~ imi ~ e, we still observe a peak at z 2.5 Volts at the second discharge, but the discharged capacity corresponds to only one fifth of the capacity in the first cycle.
~ I ~ I ~ ~ I ~ I ~ I -1 st cycle 1 st cycle ~ / ~
J ~ ~ ~ 2nd cycle , 0 ~, r2ème cycle '~, J

2 3 4 2.2 2.6 3 3.4 3.8 E (Volts) vs Li + / Li ~ E (Volts) vs Li ~ / Li ~
Figure 2.29: Inverse derivative of the potential as a function of time (smoothed).

~ 1 94127 .
To tPrminate with this m ~ tee ~ l Figure 2.30 p- ~ feel a redox reaction that can explain injecting four electrons into the molecule. In fact, it seems that everything is happening, exactly in the case of the benzophenone dimide, as if the redox properties of ~ liimide ~ were posed with that specific to dih ~ i ~ obe .. ~ n ~ ..

~ N ~
_ ~ _ = ~ _ ~

Figure 2.30: ~ éa-, redox dinitropyromellitimide.
d) Pyrazine:

Before studying this material, we had to synthesize its dianhydride, according to the procedure described in Figure 2.31 lCha ~ aWaYl. More specifically, 55 gr of 1,2-phenyl ~ nedi ~ mine (0.51 moles) are dissolved in 500 ml of hot water, then 70 gr are added dihydro ~ y ~ l, ~ sodium te. After complete dissolution of the latter, the solution is filtered, and 100 ml of 36% hydrochloric acid are added to the filtrate. After cooling down, the solution is to again filtered, and the solid obtained recrict ~ e in 700 ml of boiling water with activated carbon.
After hot filtration, it appears to cool ~ emnt the solution of the crystals, which are recovered by filtration and washed with cold water. Quinoxaline-2,3-dicarboxylic acid is obtained with a 38% yield, and has a melting point of 188 ~ C.
OH
Q ~ NH2 HO COO Na + Ha ~ ~ COOH
~ NH2 HO- COO ~ a + H2o l ~ l ~ hd ~ COOH
OH

o ~ O ~ o ~ (CF3CO) 20 X (~ ~ KMnO4 OO
Figure 2.31: Synthesis of the pyrazine dianhydride.

21 941 ~ 7 Subsequently, a solution of 18.6 gr of this acid in 240 ml of 1M potash, added with one liter of KMnO4 at 8%, is brought to reflux for 6 hours. Once the solution has cooled, add rneth ~ nol until the solution discolours. After filtration, the solution is concelllr6e to 200 ml, then acidified to a pH of 3 with 50 ml of SM hydrochloric acid. The precipitate of potassium di-salt tetraacid pyra_ine is recovered by filtration, and washed with absolute alcohol. Salt pot ~ Csillm is then recrict ~ llisé in 100 ml of 20% hydrochloric acid, then with 40 ml of water.
This gives Létlacall acid, oxylic P, ~ h ~ G, with a ren ~ emçnt of 70%. Its anhydride is obtained under the same conditions as for the dianhydride of dinitroben7 ~ ne In this case, it is not possible to obtain the corresponding polyimide by azeotropy, pyrazine seems to degrade during the reaction, probably by a decarboxylation reaction. The polyimide is rather synthesized by reaction of the anhydride with kff ~ mine ED-900 (Py / J-900) in ~ retonitrile ~, then cyclization of polyamic acid with m ~ l ~ npe pyridine / acetic anhydride, followed by reprecipitation in ether. Figure 2.32 gives the result of the first and the second intensiostatic discharge carried out at 30 ~ Vcm2 of this m ~ téri ~
2.8 2, 6 ~ PYlJ-9oo T = 80 ~ C
o_ - POE LiTFSI 20/1 ~ 2.4 - ~ ~ Id = 30 ~ A / cm 2 , 2.2 -.. 2 - \ ~ _ 1.8 - \ 1st discharge 1, 6 - 2nd discharge 1.4 o 0.5 1 1.5 2 2.5 e by ", ~ é ~ lle Figure 2.32: Discharge i "la,) ti ~ sl ~ e from Py / J-900 at 30 ~ J / cn ~.
During the first discharge, two well defined trays appear which can be attributed to rGdox couples specific to the two imide functions, but which represent only half of the ~ pa ~ theoretical polyimide. In addition, after recharging, these two trays can still be distinguished on the discharge curve, but IGprGsGlltent more than half of the capacity in the first cycle For ~ assess the redox potentials corresponding to these couples, Figure 2.33 pr ~ n ~ en ~ the inverse derivative potential as a function of time, we thus obtain a simil ~ ire curve to that obtained by a cyclic voltamp ~ IllGl ~ ie. The potentials at the peaks we attribute to redox couples of the pyrazine diimide are 2.6 and 2.25 Volts, respectively, or approximately 300 mV
above those relating to the diimide of ~ 7 ~ l ~ e.

~ 1 94 ~ 27 .
~. . . . . . . . . . . . . . . . . . .
I cycle DD 1st cycle \ ~
~ / ' ~. I ~. . . I. . . . I. . . . I. . . .
1.5 2 2.5 3 3.5 E (Volts) vs Li + / Li ~
Figure 2.33: Inverse derivative of the potential as a function of time (smoothed).

II-3-2-3 Molecular stacks:

Table 2.8 groups the peak current intensities obtained in cyclic voltammetry over a microelectrode of 25 ~ Lm, for three electroactive species, respectively the diamide of the ferrocene, the diimide of naphthalene, and the diimide of perylene, copolymerized with Jeff ~ mine IBanl1l. These inten.citPs being determined for an identical scanning speed (40 mV / s), at the same ~ t ~ Ire (80 ~ C), and with the same chain length of the ~ egm ~ n ~ solvating agent (Jeff ~ mine ED-900), a ~ Increase in the intensity of current between ferrocene and ~ iimides of nuclei aromatic is surprising to say the least.

Table 2.8 Influence of the electroactive species, copoly ", erected with the Jeffamine ED-900, on the current peak measured on a ", r ~ le_ll., from 25 llm to 80 ~ C, for a speed of 40 mV / s, according to lBanI1].
Redox molecule Ferrocene diamide Naphthalene diimide Perylene diimide ipiC (nA) 0.046 1.3 130 ipic / ipic (ferrocene) 1 283 2830 This increase in current for the iimides of aromatic nuclei can be justified if we admit that the electrons no longer move by activated jump from one active site to another, in correlation with segmP ~ s movements of the macromolecular chain, as in the case of ferrocene, but by a m ~ r ~ n; ~ e of cooperative transportation. Such a m ~ c ~ ni.c.Ae of displacing electrons involves the concept of molecular stacking 18anl1l Indeed, planar molecules po ~ s ~ ed ~ nt a system ~ conjugated orbital, such as diimide ~ aromatic nuclei, are likely to crictallicp ~ r following a quasi-~ midimPn.cional structure which in fact corresponds to an empiIe.mP.nt of the nuclei aromatic. These empilPments are a priori favored in polar lalond environments ~ that are in especially the polymer electrolytes formed by the ~ icsol ~ tion of a salt with anion delocalized in poly (ethylene oxide) 2194t27 I.
Figure 2.34 illustrates this conce ~ l in the particular case of the copolymer obtained from dianhydride n ~ rh ~ lP.nP, and an oegm ~ nt of poly (o ~ yde of ethylene) (li ~ min ~ s e- 'O. .

'. ~ ~ M ~
.
Figure 2.34: Model of e "l ~.;. 'E7, ents i7 ..,' éc ~ for poly (naphthalimide-ethylene co-oxide).

Analysis by differential scanning calorimetry (Cf Annex) of the tllprmi ~ read properties of polyimides according to the length of the chain of the solvating segment, and the nature of the species electroactive. should allow us to confirm the validity of this hypothesis. Figure 2.35 illustrates the results obtained for the copolymers, obtained by azeotropy, of the benzene diimide, with the Jçff ~ rninP, ED-600, 900, 2000, and the PPO di ~ rnin ~ s mass 400 IGUilbUdl.

- 30 ~
T / ~ t = 1 0 ~ C / min - 44 ~ l ~ B / J-600 o ~
~ D. 60 ~ 1 ~ B / J-900 ~ I \
~ 300 \

Temperature Figure 2.35: Influence of the nature of diam sey ", ~" l on the properties tl, e "" ~ es of the copolymers diimide du be, ~ el) e with Jeffamine ED-600 900, and 2000, and PPO 400 diamines after [Guillaud].

. .
It app ~ ul that the tempel ~ tulG of tr ~ n.cition vill ~ se polyimides ~ ngmçnte when we decrease the length of the rh ~ în ~ -s, and thus goes from -60 ~ C with the Jerr ~ ..ine ED-2000, value close to that for pure POE [VaU ~ el, at 3 ~ C with PPO 400. This a ~ lgrn ~ nt ~ til ~ n of te ~ of tr ~ n.
is probably related to the ~ limim1tion of the degrees of freedom of the ch ~ înP ~ constraints to their two e ~ iles. On the other hand, only the polyimide with the Jeff ~ mine ED-2000 recri.ct ~ lli.se, the other three segments perm ~ tt ~ nt to obtain m ~ t ~ .ri ~ amorphous.

As for Table 2.9, it is possible to evaluate the h-fl.) E ~ -ee of the aromatic nucleus on the temperature of glass transition of the copolymers of Jeffamine ED-900, with respectively the diimide of ben7 ~ ne ~ n ~ pht ~ l ~ .n ~., and perylene lGUillaudl.
Table 2.9 Influence of the nucleus aI ~ "~ d ~ Je on Ia te""~ erdl- of l, dnsiti ~ n vitreous copolymers of Jeffamine ED-soo with. ~ ~; tiioment the benzene diimide, na ~ ne, perylene, according to [GuiIIaud].
Aromatic nuclei Benzene Naphthalene Perylene T ~, (~ C) -44 -43 -45 As this table indicates, the glass transition temperature is independent of the nucleus aromatic component of the diimide, it can be concluded that it only depends on the thermal properties of Jeffamine ED-900. This result militates for the hypothesis of the presence of molecular stacking of the diimits of aromatic rings, the polymer segments placed at exteriors s'or ~ ni.c ~ nt independently ~ mment of these.

On the other hand, to complete these results, an X-ray diffraction study of different polyimides was undertaken in collaboration with the physical spectrometry laboratory (LSP) of the University of Grenoble. First, Figure 2.36 shows the X-ray diffraction spectrum of copolymers of diimide of bçn7 ~ n ~, naph ~ ne, and perylene, with the Jeff ~ mine ED-148 lDiuradol.

~ P / D-1 48 ....,, .. ,,, ~, ~ ,, ~, N / J-148 I

~ ~ ~ B / J-148 o 10 20 30 40 50 60 2 ~ (degree) Figure 2.36: X-ray diffraction spectra of copol ~ ", er ~ s of the benzene diimide, naphthalene, etp ~ ne with Jeffamine EO-148, relative to the K "~ line of copper, according to [Djurado].

.
Figure 2.37 compares the X-ray diffraction spectrum between the copolymer of Jerr ~ ED-148 and ED-230, and the diimide of b ~ n7 ~ n ~

J ~ ~ B / J- 230 ~ ~ B / J- 148 o 10 20 30 40 50 60 20 (degree) Figure 2.37: Influence of the chain length on the Jirfict-, lion X spectra of the diimide of benzene with Jeffamine ED-148 and E0-230, relative to the Ka line of copper, according to [Djurado].
As for Figure 2.38, it compares the X-ray diffraction spectrum between the copolymer of Jeff ~ mine ED-900 and ED-2000, and the diimide of ben7 ~ ne 1 ~ '1 1 not a word ~ B / J- 2000 20 (from ~ é) Figure 2.38: Influence of the length of the chain on the X-ray diffraction spectrum of the copolymer of pyrromellitimide with Jeffamine ED-900 and ED-2000, relative to the Ka line of copper, according to [Djurado].

219412 ~

II-3-3 Application as ~ born ~ electron in batteries:

The polymers with mixed conduction devl ~ ienl allow to catalyze the reaction of insertion within the cathode of polymer electrolyte batteries, allowing ~ nt the joint transport of ions and electrons on the surface of the insertion material. One way to highlight this hypothesis is colupa ~ r the discharge of a battery whose cathode is conctit-lée of the mixture of a m ~ t ~ ri ~ u of insertion and of a polymer electrolyte based on poly (ethylene oxide), in the ~ bsPnre of any additive electronic conduction, like carbon particles, with a battery in which substitutes poly (ethylene oxide) with a mixed conduction polymer capable of transporting electrons in the redox state ~ approplié.

More specifically, the equation below describes the reaction corresponding to the joint insertion of a cation lithium and an electron in the ~ u; Lure of the m ~ t ~ ri ~ l insertion (MI):
~ Li + + ~ e ~ + LixMl ~ Lix + ~ MI
The catalysis of this reaction consists in coating the m ~ t ~ ri ~ l) of insertion (MI) with a polymer to mixed conduction (PI) capable of transporting sim - lt ~ né-mPnt the cadon and the electron on its surface. Of a in a way, this amounts to considering that the polymer with cond ~ mixed action pr ~ se ~ t ~ nt a cinédque fast relative to that of the insertion material, it temporarily stores the charges, before it does not fit in the end in the insert material ~ ion. The following squadron illustrates this pdncipe:
~ Li + + oe ~ + PI + LiXMI) Li ~ + PI ~ - + LiXMI) Pl + LiX + ~ MI
If it is possible to carry out such a catalysis, a low coulombic capacity of the polymer to mixed conduction, relatively to that of the insert material, must be sufficient But this principle can only work if the thermodynamic potential of the insert material is greater in a wide range of composition to that of the redox couple, ideally with higher potential, redox polymer, in order to allow the reoxydadon of the latter during the discharge. If the kinetics of the insertion reaction is controlled by the mixed conduction polymer, the battery to the latter's redox potentdel. The goal is obviously to make work the highest possible insertion material to the potendel, in order to optimize the energy density.
.
If we plan to védfier this pdncipe in u ~ n ~ the diimide of benzene which has a first peak of reducdon to 2.27 Volts, molybdenum trioxide (MoO3) whose half-discharge potendel is 2.3 Volts l ~, piésellte a good choice of m ~ t ~ ri ~ u insertion for such a study. This material, from sllu ~: lu ~ m ~ ire ~ and capable of inserting 1.5 lithium atoms per molybdenum into its structure, the subject of numerous studies in the 70s, with a view to an eve-ntuçll ~ applicadon as cathode in lithium anode batteries [Al,, d ~]. This test only makes sense if we compare cathodes with a capacity and an equivalent composition. Also, the two batteries tested pr ~ sçn ~ e ~ -t such c ~ closest possible capacity relative to molybdenum trioxide, in this case included between 14 and 15 C / cm2, considering 1.5 Li / Mo, and are of the same volume composition, ie 40% by volume of MoO3 for 60% by volume of polymer, whatever its nature.

Figure 2.39 shows the result of the discharge at the same rate of 30 IlA / cm2 for these two configurations of boateries. The calculation of the theoretical capacity for that using the polymer to COI-duC!; OI ~ mixed (B / J-900) takes into account the cap ~ i ~ é brought to the cathode by the latter. The simple substitution of poly (ethylene oxide) by poly (benzodiimide-ethylene co-oxide) allows undoubtedly improve the performance of the cathode, and thus move from a capacity use of the method from 28 to 91%. In addition, it app ~ î ~ a well defined plateau at 2.35 Volts, representing ~ 20% of the theoretical capacity of the cathode, which can be attributed to the potential redox corresponding to the start of the first polyimide reduction peak (Cf II-3-2-2).

2.8 ~ Polymer: B / J-900 2.6 - \ --- ~ ---- r ~ c.ly "~ èr ~ :; POE 3.105 t. ~ Insertion material: MoO3 ~
J 2.4 -Id = 30 ~ A / cm2 -2.2 - ~. Em = 2 26 V Em = 2 3 o 2 uJ 1, 8 -1, 6 --C, h = 3.9 nlA.hlcm2 C ~ h = 4.4 mA.h / cm2 1.4 ~ I ~ I I. I, % Theoretical capacity Figure 2.39: Catalysis of li "se, ~ i ~ n of lithium in MoO3 speaks B / J-900.
The validity of these tests is confined by a ~ e-lxth battery test for each configuration, for slightly different capacities, which gives a coulombic discharge yield substantially equivalent to that obtained during the first test corresponding to each configuration ~

For its part, FIG. 2 ~ 40 makes it possible to compare the discharge from the cathode cont ~ n ~ nt the polymer to con ~ uction mixed and molybdenum trioxide, to that relating to what would be the discharge of the only polymer con ~ u ~ tiQn mixed, and this according to the result obtained during the discharge intenctiost ~ tirlue B / J-900 (Cf II-3-2-1, figure 2 ~ 22) for the same current density, and a capacity of the same order of magnitude (0.16 vs ~ 0.26 mA ~ h / cm2). In this case, the theoretical c ~ p ~ eity appo. ~ Ée by polyimide n, pr ~ sellte ~ 6% of the theoretical capacity of molybdenum trioxide, moreover, if one fits ~ .esse ul ~ uenlent to the capacitive neck-espondant to a potential greater than or equal to 2.35 Volts, it only ~ lese ~ more than 0.3% of the c ~ r ~ ç; l ~ practice obtained during the discharge of the cathode.

3 1 1 1 lll ------- MoO3 + B / J-900 ~ 2.5 - "B / J-900 only t ~ ~ --------------------_ _ Id-30 ~ A / cm 2 ~ ~. -C ~ h = 0.26mA. h / cm C, h = 0.26 (B / J-900) +4, 1 7 (MoO3) = 4.43mA. h / cm Time (in hours) Figure 2.40: Theoretical discharge of the polyimide relative to the MoO3 / B-J900 cathode.
It follows from all of these r ~ s - 1tqtc, that the term of electronic catalysis is perfectly app, op.ié, in this case, to qualify the role of electronic m ~ iqtellr of polyimide with respect of the reaction for inserting lithium into molybdenum trioxide It is of course obvious that these tests deserve to be deepened further, but at this stage, it is already clear that the concept of electronic catalysis, during the discharge of batteries, by polymers with anionic mixed conduction, and this at the microscopic level, opens the way, at least for primary systems, a wide choice of cathode materials the use of which has not present not retained in practice, due to their slow kinetics, or their electronic resistivity too high, defects which could therefore be partyPll ~ m ~ n ~ overcome.

On the other hand, these results suggest the importance of finding polymers with mixed conduction anionic, whose potential would be greater than that of the insertion material, and would therefore be ide ~ l ~ mPnt in a range of potential located between 3 and 4 Volts, and which would allow for their part to catalyze the reaction of disconnection when recharging the battery.

II-3-4 Application as a reserve of pl ~; cs ~ rce in batteries: -For some ~ nnpes ~ there is a renewed interest in the energy systems of pnic ~ cqnce ~ originally initiated by American military programs aimed at developing guns electrom ~ gr ~ ticks and pulsed lasers. At the time; ~ you ~ Plle, the potential applications of these systems d ~ pq ~ nl largely this framework, and we consider di ~ er ~ s utilic ~ tions of these components for the electrot ~ hnique, and more particulièlGIllellt in the dom ~ ine of the automobile.
Indeed, direct electrical functions in current automobiles, such as starting up heat engines, airbag triggering, prech ~ nff ~ ge catalytic converters, or 21 94 ~ 27 .
defrost n ~ these $; slow to have a significant license (~ 2 kW) for a period of time fairly short. The problem is even more crucial in electric vehicles for which only the on-board ieS bq-tte fn-rniccent energy perrnettq-nt to propel them, and all the more in veh ~ lles hybrid electric traction, ie those where we couple a g ~ n ~ r ~ qtrir, e thermal with bqtteriPs To fix the orders of grqnde ~ lrs, table 2.10 summarizes the energy needs supplied by b ~ qttpries for these three types of vehicles in terms of energy content, puiCcqnce, and potential of qlimerlt ~ tion of the engine [Degayllo ~. Assuming these vehicles are equipped with b ~ q-tte- ~ içs with a specific energy of 100 Wh ~ kg, this table also indicates ~
weight of head ~ ieS n ~ stop ~ i ~ e, and the specific pnic ~ cqnce which they must deliver under these conditions.
Table 2.10 Energy requirement of a hybrid and electric vehicle.
Power and Energy Current vehicle Hybrid vehicle Electric vehicle kW 2 50 100 kWh 0.5 5 50 Vol ts 26 300-500 500-700 battery weight (kg) 5 50 500 W / kg battery 400 1000 200 IndépPndq-mmP-nt of the nature of the bq ~ ttç ~ ies used, it is often antagonistic to optimize them with the times to mqlrimicer the specific energy, and to deliver p ~ iss ~ qnres important, this lltilicq, tion r ~ sult ~ nt in a premature degradation of battçriPs. The coupling of the battery with a component of pniccqnce is currently the envisaged solution to have both a battery which one can optimize the geometry in order to optimize its energy density, and which limits the pnicsqnce which is required of it at a relatively low value (~ 20 W / kg).

Studies on pllissq-nre components focus on capabilities that implement the use of the electrochemical double layer (~ 20 IlF / cm2) developing on the surface of carbon with a large specific surface (-1000 m2 / gr). At present, in lltilicq-nt electrolytes organic, we obtain capacities of a specific energy -SWh / kg and a pnicsqnce specific from 1 to 10 kW / kg, capable of performing at least 100,000 charge / discharge cycles.
Other avenues of study relate to the use of ruthenium oxide electrodes in an acid medium sulfuric, we take advantage in this case the rapid insertion of the proton in RuO2, it is actually a pseudo-c phenomenon ~ p ~ ce It is églqlpmpnt envisaged to develop symmetric systems utiliQqntc of electrodes based on electronic cond ~ c ~ u ~ polymers, in organic medium, of which one of the electrodes would undergo a doping of type "p", so that the other would be subjected to a doping of type "n", it is again a ph ~ nomane pseudo-capacitance, since we are working not on the surface, but in the volume of m ~ t ~ riqll of electrode.

Cepen ~ qnt the coupling of the batteries with components of puicsq-nce nécçsc ~ t to use a clc., lrollique de p ~ s ~ nce expensive. On the other hand, it is possible to envisage the use of polymers redox solvents to directly introduce this functionality into batteries. Indeed, .
suppose that the polymer electrolyte composing the cathode is replaced by a redox polymer solvating. During a call of p ~ liQ-Q ~ r ~ e important, the kinetics of the redox polymer being faster than that of the m ~ n ~ l ~ insertion, we reduce eQ ~ QentiPll ~ m ~ nt the redox polymer. As long as the potential of the m ~ t ~ n ~ l ~ insertion is ~ u ~ linker to that of the redox polymer in the reduced state, when the dem ~ n ~ e of puiQs ~ nce stops the insertion material re-oxidizes the redox polymer, and it is again available as a reserve of p ~ -; c5 ~ ce in the b ~ tPrie. This concept is summarized in Figure 2.41, in ut ~ Q Int the example of a cathode composed of m ~ l ~ nge of poly (benzodiimide-ethylene co-oxide), carbon, and v ~ n ~ m oxide Cathode in Control Control Kinetic pseudo-equilibrium thermody "alr,: ~ e App ~ 3 n ~ lon ECath = EMI

ECath = PPE / ~ 5 ECath = EMI ~ D.

power stationary slalionnaire time o O o O Li + r ~ O
Y N--> - N ~ EMI> PPE ~ N--- Li yVOx Li yVOx Li y + ~ VOx Figure 2.41: Principle of using a redox polymer as a power reserve in a battery.

In order to better define the orders of magnitude, suppose that we are interested in a useful battery a cathode with a capacity of 5 C / cm2, consisting of a volume fraction of 40% in oxide of v ~ n ~ inm, 1'0% by volume of carbon, and a redox polymer, e ~ l ~ se., l ~ nt 60% of the volume of the electrode. Sarh ~ nt that we insert 0.6 Li per atom of v ~ n ~ lm ~ and conn ~ i ~ s ~ nt the density and the molar mass of each of these m ~ t ~ ri ~ llx, Table 2.11 gives the thickness and weight per cm2 of battery film, for s ~ lnc comprising lithium (Li) in excess relative to the cap ~ i ~ of the cathode (3 times), electrolyte (l ~ lec), cathode, and coll ~ cte-lrs of aluminum current (Al).

, Table 2.11 Weight and surface area of the various co, -l ~ osa "ts of a battery film.
Material ~ o Volume Weight (mg / cm2) thickness (llm) V ~ x (5 C / cm2) 40 7.85 23.4 Polymer 10 3.5 29.2 Carbon 50 1.16 5.8 Al / Li (3x) / Elec (30 ~ 1) / Cathode / Al (8 ~) mbatt = 21.5mg / cm2 ebatt = 124.6 ~ 1 Such a confi ~ l ~ ation coll ~ corresponds to 3.5 mg of redox polymer per cm2 of battery film The study will be limited to poly (benzodiimide-ethylene co-oxide). In this weight, the salt of LiTFSI assumed in the present case at a concentration O / Li = 30/1. Under these conditions, it is possible to determine the density of mass and volume energy provided by the polyimide, such as this is illustrated in Figure 2.42. This calculation is Pffectllé in c ~ ncidér ~ nt the two electrons of the benzodiimide, and in assuming that the average potential for use of this rn ~ t ~ is 2 Volts. So if we are interested a polyimide consisting of a POE di ~ min ~ s with a mass of 600, we have a density of energy # 20Wh / kg of battery film, in addition to that of the insertion material of 160Wh / kg. he is possible to estimate the energy and the pui ~ s ~ nr, e at 60% of this value for a complete system of battery, i.e. ~ 12 Wh / kg and 100 W / kg. Using the hypothesis of a 100 kW puicc ~ nce for an electric vehicle, and 500kg on-board batteries, this corresponds to a reserve of power ~ 360 s, which is entirely compatible with the time of an acceleration (~ 1 min).
. ~, I. I.
~ Wh / l 40 - ~ _ ~~ \
a ~ ~

c 20 -~,,. Wh / lcg O ~ I ~ I. I, Mass of diamined POE
Figure 2.42: Density of energy supplied by the poly.;) .: ~ c ~ ") ~ osanl the cathode of the battery film.
This principle is only viable if the redox polymer is capable of performing 50,000 to 100,000 cycles charge / discharge, while retaining the cyclability of the insertion m ~ t ~ -ri ~ l. This assumption is not meaningless. Indeed, during a call of ~ nce important, the potential of the cathode can suffics ~ P ~ minl) er to reach the stability limit of the insertion m ~ tkri ~ ll. For example vanadium oxide degrades when the potential of the cathode becomes less than ~ 1.5 Volts.
On the other hand, in the hypothesis where the redox polymer absorbs the peak of pllics ~ nce ~ the potential of the el ~; trode remains at values compatible with the m ~ t ~ ri ~ u insertion. The v ~ tion of this princess . ~.
dt ~ pqc ~ c ~ tlu within the framework of the studies within our reach, we cont ~ ntt ~ s to assess the cyclability and the p ~.; cc ~ l ~ tt ~ e ~ tcce ~: ble of a battery (Cf Annex) utilitt ~ qnt the copolymer obtained by azeotropy from Jerr; ~ ED-900 and dianhydride of bte- ~ tt7 ~ ie (B / J-900), as matt ~ r ~ cathode. So, t the cathode of the battery tested contains 3.4 mg / cm2 of polyimide, a value close to that envisaged prt ~ redemmt ~ nt which colltt ~, lays at a capacilt ~ of 0.57 C / cm2.

Figure 2.43 shows the cycling result for a current density of 30tuA / cm2 of this drums. As shown in this figure, the entire capacity of the material is ~ ccessihle to the first decl ~ ge, but it decreases rapidly in the first ten cycles to stabilize by the following tt- 80% of the initial cztp ~ citt ~. It is likely that this loss of ca ~ acile is related to the deg ~ d ~ tion a part of the electroactive functions in the reduced state, by the lithium hydroxide formed by reaction residual water on the surface of the lithium. In addition, the coulombic yield expressed by the ratio of the recharge capacity over that in discharge (C / D) is one for the majority of cycles, which indicate the ~ losence of ph ~ nom ~ m ~ s parasites during recharging, and in particular of den ~ ri ~ çc 12 0 ~ 1 1, 2 1 oo ~ I ~ IJ ~ I ~) ~ L ~ nJ ll ll ll l ~ JJ ~ 1 ~ _ Power test 3 80- ~ 0.8 . ~ q 60 - Ic = 6011A / cm '~ - 0.6 c ~
Ic = 3011A / c m2 ~
4 0 - Id = 301lA / cm2 Id = 60 ~ lA / cm2 ~ 0 4 I Umin = 1, SV
20 ~ ¦umax = 2 ~ 5v .- 0.2 OIIIIII o 0 1 o 20 30 40 50 60 70 N ~, II, r ~ of cycles Figure 2.43: Battery cycling at 30 ~ LA / crn2.
The capacity being perfectly stabilized after sixty cycles, it was possible to carry out a ~ nce on this battery by regularly increasing the discharge current from 15 to 3840 IlA / cm2. The charging current is chosen at half the value of the discharging current up to 120 ILA / cm2, then set at 120 ~ LA / cm2 thereafter, to guarantee efficient recharging of the c ~ thocle after a die, the quick d ~ ge.

Figure 2.44 ~ pr thus the fraction of the available capacity as a function of the density of current, ~ ull ~ lle ~ lt says p ~ liss ~ nce dçm ~ ndée, l ~ Çélellcée at a char ~ ity of 100% after 59 cycles (Csg) for convenience 11 is au ~ ssi indicated the discharge regime in Cltd, which actually expresses the time in hours (td) which coll ~, spondrait, for a given current, to the discharge of the battery, in assuming use of 100% of the capa. ; l ~

o I ~ 1 9412 ~ 7 120 ~ 2.2 100 ~ C / 4 ~ 2.1 ~ - C / 2 ~ 80 f ~ C - ~ 2 - ~ \ -._ - ~ C ~ D
._ 60 ~ 4C ~ _ ~ 1.9-- 8C ~
40 _ 16C \ - ~ --c 1.8 -rL ~ \
2 0 - 32C - 1, 7 -o IIIIII 1 1.6 o 0.5 1 1.5 2 2.5 3 3.5 4 o 0.5 1 1.5 2 2.5 3 3.5 4 Current density (mA / cm2) Current density (mAlcm2) Figure 2.44: Influence of the discharge regime Figure 2.45: Influence of the discharge regime surb capacity of the cathode, I ~, Sf ~ ~ Se has C59 discharge on the p. ~ lenliel means of the cathode.

The average potential, calculated by integrating the discharge curve, is given on the Figure 2.45. This potential evolving in a relatively linear fashion with the current density up to 960 IlA / cm2, we can deduce an equivalent ohmic resis ~ nce # 350 Ohm.cm2. This value is relatively high, but it should be remembered that the battery uses an electrolyte approximately ten times thicker than that which would be used in an optimized system.

Figure 2.46 allows to assess the hysteresis between the potential during recharging relatively to that of ~ lllPillé in discharge for the 59th cycle of the battery. It follows that for a slow discharge and charge regime (30 ~ lA / cm2), the recharge potential is only higher # 50 mV to that in landfill, which is after all relalivelllent weak.
2.8 1, I
2, 6 - Cycle N ~ 59 _ 2.4 ~ Id = lc = 30 ~ lA / cm2 2, 2 ~
~ 2- ~ ~ _ _ uJ 1, 8--1.6-1.40- 20 40 60 80 100 %Use Figure 2.46: D ~ cl) aryd and charge i "lensi ~ '~ eu at 30 ~ / crri2 from B / J-900 to 59 ~ me cycle.

2 1,941 27 -Figure 2.47 gives the decha, ~, es curves of the battery for the different currents (indicated in ~ A / cm2) used during the pniss test ~ nce It is it, t ~ ssant to conct ~ ter that the loss of capacity is çcoentiell.ompnt due to the clensi ~ me red ~ lction of ~ mo ~ 1iimi (3rd, which translates into less kinetics of the reaction relative to the pl ~ ", iè, t; ré ~ uction of the polyimide.

3,840 1,920,960 %Use Figure 2.47: Influence of the current density (in l ~ A / cn ~) on the discharge potential.

It is common practice to characterize an energy system by plotting on a scale logarithmic, the power density as a function of the energy density. In other words, this , epr ~ sel-t ~ tion given on figure 2.48, called Ragone curve, makes it possible to highlight the kinetic limits of battery operation. The establishment of this curve assumes to make hypotheses on the geometry of the system, which correspond in the present case to a thickness of 86 µm, and a weight of 12 mg per cm2 of battery film, including lithium, the electrolyte, the cathode, and two aluminum current collectors. Calculation of energy density takes into account the average discharge potential of the battery for a given current density.

103 '''''''' I
~ e (AI) = 2x811m = 1 611m ~
3 ~ 280W / I _ ~ e (Li) = 1 0 ~ 1m 200W / kg ~ e (Elec) = 3011m "1 o2 ~ e (Cath) = 30) 1m ~ t ~ e = 8611m . ~ ~ -~ 1 01 ~
._ o,,. ..... I,. . .....

Energy density (Wh / l) Figure 2.48: Ragone curve of the studied battery.

The electrolyte which we assume with a thickness of 30 μm for the calculation is in reality of a thickness ~ 200 to 300 ~ Lm in the stu ~ e battery. Consequently, the density of power ~ nce evolving at the inverse of the square of the thickness of ~ s ~ c, the pui.cs ~ nce calculated is an underestimate ~ tion of the puiSs ~ nce which would be obtained with a thin electrolyte. In addition, for a given penetration, the density of available energy is é ~ le.ment undervalued because of the pol ~ ris ~ tior ~ Anyway, this test proves that it is possible to obtain peaks ~ l ~ these i ~ pollutants ~ 1 kW / l for a density of energy ~ lOWhA, which is completely co ... p ~ lible with the ~ app ~; we are considering.

The other essential parameter to be reconciled is the cyclability of these materials for a load regime and rapid discharge in relation to the role ~ csign ~ e redox polymers in the b ~ tterie- Figure 2.49 shows the first cycles obtained for a current density of 500 ~ A / cm2, which corresponds approximately to a ~ n ~ e of 100 Wh / l and an energy density of 20 Wh / l. The cdpaciled in discharge depends on the recharging potential of the battery, the thickness of the electrolyte film requires slightly increasing the recharge potential to limit the effect of the gradients co ~ nt ~ tiOns. n is quite illt ~. ~ ssdnt of collar ~ t ~ r e ~ cçll ~ ntP. cyclability of the redox polymer, its c ~ p ~ c; l ~ d '~ ltili.c ~ tion remaining stable for more than 300 charge / drop cycles; as well as ~ hsP.nre of ~ P.nl1rites when recharging. This ~ hspnre of ~ len ~ ritec POUIl ~ could be put on the account of the low cycled lithium capacity, about 0.25 C / cm2, which corresponds to a thickness of ~ 0.3 ~ m / cm2, with regard to the thickness of the electrolyte film. Based on experience with other battery tests, this effect should rather be attributed to the homogeneity of the cathode resulting ~ nt from the macromolecular nature of cathode material, which guarantees a homogeneous dissolution and deposit of the lithillm 1 2 0 ~ 1, 2 100 - ~: 00 ~ 00 0 0 o ~ to o O ~ O 1 c 80 _ ¦ Id = 500 ~ lA / cm2 - 0.8 ~ I Ic = 500 ~ lA / cm2 .tn 60 - ~ ~ f 0.6 c ~
40 _Umax = 2.6V ¦ Umax = 2.8V ¦ Umax = 2.6V o 4 - Umin = 1.5V
20-- - 0.2 OIIIIII o 1 oo 1 50 200 250 300 350 400 Number of cycles Figure 2.49: Cycling at 500JJA / cn ~ of the battery,% Use r ~ f ~ r ~ nc ~ at ~ h ~ at Cs ~
The test continued at a current density of 1 mA / cm2, approximately at a pniss ~ nce of 200 W / l and an energy density of 15 Wh / l. To simplify, capacity is a again normal to that obtained after 59 charge / discharge cycles. The result obtained in these con ~ ition ~ is given in Figure 2.50.

,,.
~ 161 2194 ~ 27 60 1 1 1 1 1 1.2 5 ~ ~ 1 ~ ~~ ~ V ~ V - 1 0 4 0 ~ O, 8 , ¦Ic = lmA / cm2 ~ _ ~ - 0.6--20 - Id = lmA / cm '~, _ 0 4 - I Umin = 1.5V
10 - IU max = 3V - 0, 2 O lllll O

Number of cycles (in thousands) Figure 2.50: ~ yclage at 1 mA / cm2 of the battery,% Usage ~ rér ~ nce at C69.
As we can see, after - 60,000 charge / discharge cycles, the battery delivers still ~ 30% of the capacity obtained at the start of the cycling test, for this discharge regime of 1 mA / cm2. L '- tili.c ~ tion of the battery capacity for some discharges carried out at lS ~ LA / cm2 during this cycling is given in table 2.12. It is clear from these results that the decrease observed capacity is not linked for esse ~ tiel to a degradation of the redox polymer, but probably a ~ lgment ~ tion of the r ~ sict ~ nce interface between the polymer electrolyte and lithium.
Table 2.12 Eissai of discharge at low current (15 ~ Vcm2) of the battery during cycling at 1 mA / cm2.
Number of cycles 34,761 51,231 57,565 % Uttl. 15 ~ A / cm2 76 73 67 ft r ~ ncé at 100% for the ~; t é at the 59th cycle.
This is confirmed by the decrease in potential 2 means obtained during the discharge of the battery at Umin = 1.6VlUmax = 3V
cycling, as shown in Figure 2.51, 1.9 ~ Ic = ld = 1mA / cm2_ this decreasing - 250 mV after 60,000 cycles at a regime of 1 mA / cm2, which corresponds to an o_ resistance of S00 Ohm.cm2, which corresponds to t-- 1 ~ 8 probably for the most part, to the evolution of in ~ rfa ~ e between lithium and the polymer electrolyte E 1, 7 -In addition, according to the reslllt ~ c of table xx, the loss 1, 6 -capacity after ~ 57,000 cycles at a speed of 1mA / cm2 discharge is ~ 35%, i.e. a loss of 1.5 lllll capacity -6.10-4% per cycle, This result is all at o 1 o 20 30 40 50 60 ", ~ Jo." Bre of cycles (in thousands) falt remarkable, especially considering that 1l s aglt there Figure 2.51 ~ evolution of the pot ~ i "iel iel of a laboratory test for which the different brs of b decharye of the battery ~ 1 mA / cm2.

21 9 ~ 1 ~ 7 parameters are not controlled at best, such as for example the drying of m ~ t ~, ri ~ u ~, the state of surface of the lithinm, which it is well known that they notably influence the performance of élé ~ ocl ~ liques systems integrating them, and part ~; uli ~. ~, .llenl on their lifespan.

After the cycling test at a discharge rate of 1 mA / cm2, cycling continued at a slow discharge regime of 30 ILAlcm2, the result is given in Figure 2.52. Even if the c ~ u ~
slightly decreases during the first twenty cycles, it then stabilizes at ~ 60% of the capacity at S9 ~ me cycle until 64 ~ me cycle. Decreasing the discharge current to 15 ~ lA / cm2 makes it possible to further improve the use of the cathode z 10%.
1 00 ~ 1.2 8 0 e ~ rl J ~ ~ ~ ml ~,. . l. ~
~ 60 ~ 0.8 - ~ o 4 ~ ~ I ic = ld = 30 ~ A / cm ¦ Ic = ld = 15 ~ A / cm '0.6 20- ¦Uminx = li56V ~ 0 ~ 4 0 1 1 1 1 1 1 0.2 Number of cycles Figure 2.52: Low current cycling test after 57500 cycles,% U ~ .dlion, referenced at Csg.
Figure 2 ~ 53 allows vi.c- ~ li.cer the low current discharge curves of the battery during cycling ~ Apart from the loss of capacity, there is no visual clue which would indicate a major degradation of the cathode material, even after z 57,000 charge / discharge cycles ~
~ 3 - Ic = ld = 30 ~ Ucm2 (59.57505) J 2.5 ~ Ic = ld = 15 ~ Vcm2 (34761,51231,57565) ~ 1.5 c ~ l , ~ u) ~

. %Use Figure 2.53: Low current cycling test after 57500 cycles,% Usage ~ rt7nc ~ at C59.

._ To finish with this cycling test, Figure 2.54 gives the evolution of the energy density and the cumulative energy density for all cycles carried out. After ~ 60,000 cycles, the density residual energy, for a discharge regime of 1 mA / cm2, is still ~ 4 Wh / l. The element more interesting to note is the value of the cumulative energy ~ 450 kWh / l. Comparatively, a elemPnt of lithium battery, u tilis ~ nt a m ~ t ~ .ri ~ u insertion as cathode, rarely performs more 1000 charge / discharge cycles with an average energy density of 100 Wh / l, i.e. a density cumulative energy of 100 kWh / l, it should be noted that this col. ~ ond calculation at a high ectim ~ tion of pe ~ ro ~ ncPs of such a b ~ tteriP .. In these con ~ itionc ~ plése "the battery used ~ nt a redox polymer stores 4.5 times more energy than a lithium battery made from insertion material.
1 0 0 ~ 5 0 0 , 1 0 _ '/ ~ 300 ~
- 200 ~.
4Wh / ll - 100 o Number of cycles (in thousands) Figure 2.54: Energy density and cumulative energy density of the battery.

The performance of the lithium electrode of a battery is generally defined by its factor of merit, expressed by the ratio of the amount of cumulative charge. during cycling relative to the charge equivalent to the amount of lithium initi ~ lemen ~ present in the b ~ t ~ rie In the present case, this calculation does not make sense, the lithium electrode being thick ~ 100 llm which corresponds to an excess of lithium on the order of 100. It is more interesting to define the merit factor of redox polymer which is particularly high, since of a value ~ 12000, relatively to the capacity of the first battery discharge.

This result validates the hypothesis according to which the electroactive species exict ~ n ~ in neutral form or anionic are compatible with the chemistry specific to lithium batteries and fully justify ~
interest and study of this materials class.

, Therefore, it is possible to consider developing batteries of which the cathode would be composed only mel ~ nge of a polyimide with carbon grains, to achieve either batteries of pl ~ iss ~ nce ~ stationary batteries for the management of an electrical network, or batteries rescue for example.

i.

For example, Figure 2.55 shows the density energy density of a battery ~ ltilic ~ nt un poly (benzodiimide-ethylene co-oxide) as n ~ tee uq-- of cathode, assuming a thickness of fixed cathode of 5011m, electrolyte of 20 ~ Lm, aluminum collectors of 1611m, and a lithium thickness corresponding to three times the excess relative to the cqp ~ city of the cathode. Of more, we consider that the cathode works on two electrons with an average potential of 2 Volts.

1 2 0 - \
~ "1 0 0 - \ -c 80--~ 7 ~

._ Jeffamine mass Figure 2.55: Density of volume energy as a function of the mass of Jeffamine.
The calculation is performed for a mass of Jeffamine between 400 and 1000, a mass of 400 also making it possible to obtain a polymer soluble in organic solvents, such as acetone, acetonitrile, or THF for example, and thus obtain homogeneous cathodes. The result is an approximation, close to reality, since the presence of carbon is not taken into account in the composite, nor of the mass of lithium salt dissolved in the polyimide.

It is thus possible to obtain a volume energy density of 100 Wh / l, for a battery film, using a Jeffamine of mass -540. In this case, the fraction of solvating segments still represents 74% by weight of the material, and 83% of its volume fraction, which allows to consider obtaining an ionic conductivity close to that obtained for POE in a state amorphous. This energy density is certainly two times lower than that for a battery using an insertion material, but this disadvantage may be offset by the service life of the drums. In fact, the interest of such a battery is dependent on the intended application.

In particular, it is difficult to conceive of the production of an electric vehicle with such a battery, the energy density being incllffic ~ nte to guarantee ~ -tonomy appreciated ~ q ~ ble ~ By cons, in the case of a hybrid vehicle, the energy requirement supplied by the b ~ tteries being ten times less (Cf II-3 ~ 1, table 2.10), and the pnics ~ nce l- ~ Sce ~ s ~ it; more important, it is possible that such batteries p. ~ 3e.l ~
a very special interest. Finally, an impol lifespan ~ - this is the parameter p,; nr; p ~ ql ~ to consider for stationary battery systems, such as inverters, or storage power plants ~ - 21 941 27 say ~, ~, s ~ es used for the storage of current peaks of electrical networks, for example, for which it makes more sense to r ~ i ~ onn.o, r in terms of cumulative energy.

21 941 ~ 1 II-3-5 Applications to electrochromic glazing:
Polyin ~ ide-Polyether Mixed Conductors As Switchable Materials For Electrochromic Devices i C. Michot, D. Baril and M. Armand.

Laboratoire ~ d'lonique et Ele ~ chi ~ ue d ~ Solid ENSEEG Ins ~ inlt Na ~ ional Polytechnique de Grenoble Domaine Universitaire, BP 75, 38402 Saint-Mar ~ in d'Hères, France.
1 - Introduction:
Electrochromism is under active scrutiny by both the scientific and industrial communities. Tea principle of ch ~ nging the coloration of a material by an elec ~ ... ic ~ l reaction was origin ~ lly studied for applications in displays for the watch industry tl], but extrapolation to the video r ~ erlce (~ 25 Hz renewal) soon appeared elusive. The main en ~ ph ~ cic is now on visible light and IR modnl ~ tion for windows in bnildingc and glare attenuation in rear view mirrors for automobiles tl-5] For the former application, the response time can be much longer and there is anyhow a compromise bel ~, n the thickness, thus conductance, of the ITO or similar conductor destined to ensure a uniform potential distribution across the whole surface and the tr ~ ncmi.ssion of visible light in the bleached state. While elegant all thin-film systems obtained using various deposition technologies (sputtering, CVD ...) have been shown to be feasible [7,8], many technical difficulties may arise when dealing with large surfaces. A plastic solid electrolyte is desirable for systems which then can be lamin ~ ted. Thoughts, the pursuit of high conductivities with gelled systems, ie a solvent thir ~ Pned by a small fraction of high polymer may still result in the long term to viscous flow and creep for systems operating in the vertical position.
As electrochromic materials, transition metal oxides have the favor of many research groups, especially WO3 due to its high coloration efficiency and (relative) stability at pHs <3 or in aprotic systems. This is in fact returning to the first m ~ tenal for which electrochromism was described t9].
Photochromism may however play a deleterious role when this semi-conductor is under illumination [10] Conjugated polymers offer in principle more opportunities for tuning the adsorption in the desired spectral domain and several, like poly (aniline) and poly (thiophene), are straightforwardly obtained as adherent films by electropolymerization [11] However, their cycle life is presently limited. A possible explanation for the lack of stability is the fragility of the positively charged moieties which are formed upon "p" doping and are very se "silive to nucleophiles.
basic species are created in the electrolyte by electrochemical side reactions, including hydrogen evolution from trace water which locally creates OH- speciP-s On the other hand, polyimides are widely used in the electronic industry for their thermal stability, in ~ ln ~ ling high Tg's and in ~ nl ~ ting plupellies, all p ~ pellies dependent upon the use of rigid aromatic subunits in the backbone. It came as a surprise when these m ~ teri ~ lc were reported to become electronically conducting upon electrochemic ~ l reduction, owing to the relatively large electron affinity of the imide group [12-14] There is however no eYten ~ ion of the conjugation beyond these centers, and the conductivity results from electron hoping ~ Iween neighboring moi ~ ptips ~ RPS; deS, the diffusion of the required charge-con ~ c ~ ting ionic species (cations) is excee-iingly slow in the ~ bsPnc e of a solvent acdng as swelling agent or p ~; c; 7f, ~. To O ~; GIllC thi ~ s riiffi ~ ulty, a new family of mixed-conduclivily co-polymer in which imide redox-active groups are separated by solvadng PEO blocks have been de ~ igne ~ (iiffering from the already described PEO-based nelwoll ~ bearing .

2194i27 "
hang redox groups like ferrocene and phenothi ~ ine [15]. In the following, the preparation, electr ~ hPmic ~ l and optica ~ p, ~ ies of these novel m ~ tPri ~ lS are des ~ ibe-l 2 ~ Experimental:
2.1 ~ Synth ~ sic:
The polymers were pl ~, pa ~ d by polycond ~ nQ ~ tion reaction b ~, ~ 2 ~ n ~ t_t ~ d ~ Aylic dianhydride of a, c ~ -diamino oligopolyethers with various aromatic bis (cyclic anhydrides). The later whose formula is depicted inFig. 1 wereobtained from commercialsources (Aldrich, ~ ~
Fluka), together with LiC104, with the highest available naph -'enetetracarboxylicdianhydride purity and were used as received. Samples of Li (CF3SO2) 2N (LiTFSI) were given by Hydro-Québec c ~
(Canada). ~ ~ O
O, O'-Bis (2-aminopropyl) polyethylene glycols:
perylene t ~ t ~ dcdll ~ A ~ dianhydride H2NCH (CH3) CH2 [0CH2CH2] nOCH2CH (CH3) NH2 3_ ~~
with central PEO sequence of l ~ IW 500, 800 and 1900 (JeffaminesZD ED 600, ED 900 & ED 2000 resp.) Hereafter referred as J-600, J-900 and J-2000 were a generous gift benZOpl ~ enOnet ~ lldcdll ~ uAylicdlanhydnde from Texaco-France. ~ _ ~
'Samples of O, O'-Bis (3-aminopropyl) poly (THF) were c 1 ~' ~ L ~, ~
simil ~ rly don ~ ted by BASF (~ w 750, 1100). Prior to the co polycondensation reaction, the oligomers were dehydrated Fig. 1: dcvel ~ ped fonmulae of the starting by bubbling anhydrous argon for two hours while kept at bisanhydrides used as precursors for the 80 C. The amino functionality numbers were obtained by pr ~ pdjdtion of poly;. ,, 'es The corres-acid-base titration in water. Solvents (HPLC or anhydrous B, N, P 8 BZO respectively.
grade) from SDS were kept on activated neutral alumina granules. All h ~ ndling of oxygen- or moisture-sensitive m ~ leri ~ lc was done in a dry box ( <lvpm ~ 2, H20).
2.2 - Polymer characterization:
The GPC analysis was performed on a Waters 590 GPC, through two lOOA UltrastyragelB ~
columns. A Netzsch STA 409 differential sc ~ nning calorimeter (DSC) was used to determine the glass transition temperatures Tg, and the m ~ l ~ ing points Tm. Thermograms were recorded at a rate of 10 ~ C.mn-1 in helium. Each sample was first quçn ~ hed to -100 ~ C and then heated to 150 ~ C, allowed to cool and at last heated again to 150 ~ C.
The thermal stability range was investig ~ t ~ od by heating the sample in air at the rate of 10 ~ C.mn-l from room Llllpe.a ~ ul ~ up to 400 ~ C.
The optical absorbance spectra were recorded on a Perkin-Elmer (Lambda 7) UV-visible s ~ ec ~ u ~
2.3-Ele.; Tloçl. ~ ..; c ~ l char ~.; 7 ~; on:
Conductivity ûf films samples were sandwiched bet ~ ns ~ inless-steel electrodes. AC conductivity measu G ... ents were carried out under dynamic va ~; uulll using a HP 4192A impedance analyzer, from room temperature up to 110 ~ C, over the rlG ~ lu ~ ncy range S Hz - 13 MHz .
The redox plupellies were tested by cyclovoltAmmet-y- using a computer-controlled microelectrode (0 either 25 or 125 ~ lm) set-up built in the laboratory with a resolution of 1 pA. The cell comprised the c ~ f ~; r ~ l chain:
Li / PE0-Li + TFSI- (C104-) / Pt microelectrode (TFSI- = (CF3S02) 2N -) -el) ~, ultra slow-scan vQ ~ y (MacPile ~ 19) was used with 2 cm2 button cells:
Li / PE0-Li + TFSI- / redox polymer / stAinless steel In both cases, the redox polymers were used in thick films (~ 50 ~ Lm). For the button cell ex ~. ;. ~ PntC, the theoretical capacity could be c ~ qrlculqted from the weight of the sample and, otherwise specified, 5% v / v carbon black was added.
No attempt was made to use electrolytes other than linear PE0 to reduce the operating temperature [15].
3 - Results and .liecns ~; on:
3.1 - Synthesis and characterization:
The formation of imide linkage from reaction of a primary amine group on a cyclic anhydride is known to proceed in two steps via an interrnerli ~ qte amic acid (Fig. 2).
O

H OH!
O + N ~ _ _ H20 + ~ l ~ r H ~ HY ~
OO
Fig. 2: reaction scheme for the formation of the imide linkage from a cyclic anhydride and a primary amine group.
In the case of the polyethers used as soft segment ~ s, the reaction of stoichiometric quantities of ~ IiAmine and dianhydride in a common solvent proceeds readily at room tempeli, l ,. ~ to the polyamic in ~ erme ~ iAte; cyclization is achieved chemically with the well known acetic anhydride / pyridine mixture. More simply, heating of the neat components at 200 ~ C achieves similar result. We also found recently that Dean-Stark azeotropic rlistillq, tion in toluene cleanly and rapidly yields the polymers from the simple mixture of the starting ingredientc In all cases, precipitation into diethyl ether from the reaction medium yields the solid polymers. pllrifieqtion can be $ imilArly accomplished by dissolution in dichloromçthAne or acetonitrile and pf ~ cip; ~ tion with ether. The appeq -, qnce and concict ~ nre of the materials varies widely depending upon polyether segment length, from waxy semi-crystalline mq ~ eriq1 (JçrrAt ~ e 2000, poly (FHF)) to amorphous and l ~ spa ~ e, -l (Jeffqmine 600 and 900). Similarly, the imide segment imposes its color (benzene: colorless; narht ~ len ~: light yellow-brown; perylene: dark red).
GPC of the materials in THF indicates molecular weight close to 2.104 and a broad distribution cpe. ~; ~ y ~ 2) as e ~ l) ected from a polycon ~ en ~ qtion reacdon.

21941 ~ 7 The DSC traces, with some shown in Fig. 3 indicate ~ 31O ~ T / at = 5 ~ Cmn ~
that all m ~~ s made from Jeffamine '~ 9 ED 600 or ~ ~ B / J 600ED 900 which are both molten at room leM ~ Iatu ~
yield amorphous polymers ~ ~ N / J- 900 e ~ hibiting well detect ~ lDle ~ - ~
Tg's. With Jeffamine0 ~ ~ ~ 30o ED 2000, an endothermic ~ P / J- 2000 event is detected at 30 ~ C, ~ ~ no reflecting the crystallinity of the starting materials.
These results are to be ex- '' B / PP ~ 400 pected when considering a st ~ tictir ~ l (butnon-random) 1Illllllllllllll llll ~ lllllll ~ ll llllll lllll lllll lllll distribution of crystallinity- -100 -50 o 50 100 150 breaking aromatic groups t ~ r "~ -e, dture along the PEO chains [17]. Fig. 3: DSC traces for some polyconden n polymers.
The poly (THF) samples showed a similar trend and PPO-co ~ ining polymers were completely amorphous -as expected-.
0.1 ~ Under helium, no significant weight loss is , ~ apparent up to 300 ~ C on thermogravimetric B-J900 f analysis, temperature beyond which the aliphatic polyether segm ~ onts decomposes.
0.05 ~
3.2 - Electrochemir ~ l characle ~ ion:
J In the neutral state the solvating power of the E 0 ~, ~ _ polyether segment is kept, as expected. All the r, materials prepared could form complexes with low lattice energy salts like LiCl04 or Li (CF3SO2) 2N.
Conductivities were found to be comparable with 005 _ \ II _ those of conventional amorphous polymer V ¦ electrolytes with ~ 10-5 S.cm-1 at or slightly J above room temperature. The melting point of the materials containing semi-crystalline J-2000 -0.1 segments was no longer visible at EO / Li = 8 to 12 1.6 1.8 2 2.2 2.4 ratio as seen on DSC traces.
Volts vs Li + / Li ~ All materials tested showed redox activity. Tea Fig. 4: sbw scan (~ i o9 ~ "", ~ t "~ scan of slow scan voltammogram for benzenetetracarboxylic acid di-imide co-Jeff ~ mine 900 (B-benzene t ~ lla ~ .ln, Ayllc-Jeffamme ED 900 polymer; scan rate = 84 mV / h, T = 80OC. J 900) is shown on Fig. 4; two well defined reversible l-electron reduction / oxidation waves are app ~, nt on the forward and reverse scans. The quantity of charge passed co.l ~, sponds to the entire theoretical capacity, suggesting an effirien ~ clccl. ~ n hopping mech ~ l-is ... between the relatively remote sites.
N ~ pht ~ lene and perylene-based polymers behave simil ~ rly, with a ~ nr ~ ency for the two l-electron waves to merge into a broader l-electron one, especi ~ lly appa ~ nt for the highly delocalized perylene moiety. This is shown on the CV scans obtained on Pt microelectrode on Pig. 5. It is notewo ~ y that the much faster scans of the microelectrode may also result in smaller peak sep ~; on This is to be brought together with the example of caroviologenes [18] where the two pyri ~ inillm rings are separated by conjugated - (CH = CH) o- spacers, showing a t ~ n (lency low ~ ds displ., pol ~ ionation of the intenne ~ i ~ te radical species into neutral and di ~ tionic forms with increasing conjugation length not. Perylene does howe.er sustain larger current il.t ~ .C; I; l'c oq,.,., ~
~ NJ 900 PJ 900 l ~ \

-0.3 ~ '~' ~ '~''''' 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 E / V vs. Li + / Li ~
Fig. 5: microele ~, t ,. de cyclovoltammetry traces using Pt microelecl, ode (0 = 25 llm) for naphtalene and perylene - Jeffamine 900 copoly -.- t: - ~ (a: NJ 900; b: PJ 900); T = 80 oc, LiC104, O / Li = 9/1.
PPO or Poly (THF) -based mate-rials equally showed redox activity 8 at comparable potentials, with a tendency for peaks to become 6.
broader. This behavior is tenta-: / ~
tively attributed to the lower ionic 4 - / ~ --_ conductivity of the base polyether. ~ z: /, ~
Besides, their weaker solvating 3 - ~ /
propel lies, as compared with PEO, c O f may result in saltdepletion to the,: / ¦
advantage of PEO when the latter ~ -2. ~
is used as electrolyte. Similarly, 4 - / ¦ _ the benzophenone-based polymer: / ¦
where the conjugation of the redox -6. L_ I ~
molecule is less pronounced than for the others, is reduced at less -8 negative potentials and the scans 1 ~ 5 2 2.5 3 3.5 showmediocrereversibility. E vS. Li / Li (Volts), ~. . . Fig. 6: slow scan (MacPile ~ E9) voltammetry traces of ben ~ ene tetra-Strong ln lc ~ tlon for the eY ~ s ~ llce carboxylic dianhydride - Jeffamine 900 copolymer without carbon;
of a mixed conductivity, ie the T = 80 ~ C, LiTFSI, O / Li = 20/1, scan rate = 10 mWh.
mobility of both ions and electrons .

in the reduced states is given when the volt ~ mmetry scans are performed in the absence of carbon black additive; while the shape of the traces, as shown in Fig. 6, are distorted, the reduction being observed at more negadve potentials with an abrupt rise in current corresponding to the onset of electronic con ~ luc!; vily, reversibility and total av ~ ility of the redox sites spread over the 4511m of the film, is m ~ int ~ in ~ A

20 - '''' I '''' I '''' I '''' We found interesting the result from the a 'electrochemical screening of a 15- b,; ~ / ~ copolymer in which three of the aromatic moieties having shown to 10 - behave reversibly (B, N & P) were c ~ / ~ _ co-ir, col ~, o.dted in the same backbone of J-900 polyether as shown in Fig 7.
~ ~ Again, the electrochemical activity of ~ 5 ~ each component of the copolymer is - c \ / preserved, the first wave being -10. \ f \ / - attributed to P and N (first electron a, b a ') the second (b, b') to a combination 2 2 3 of the N second electron and B first ~ ~ 5. 3-5 electron. The last reduction step is in c, Volts vs Li + / Ll ~ c 'for the B moieties. Electron mediat-Fig. 7: Slowscan (MacFi'e ~) v ~ "-"", ~ ytracesofbenzene, ion throughout the system obviously naphthalene and perylene tetracarboxylic dianhydride- occurs, as in such copolymer each Jeffamine 900 copolymers; T - 80 ~ C, LiTFSI, individual entity B, N, P, is now highly diluted in the PEO matrlx.
3.3 - Optical plupellies:
In the reduced state, all polyimides showed intense color changes. The water solubility of the longer chains PEO allowed a simple color test using sodium dithionite as reducing agent in carbonate-buffered solution. The results are summarized in table l.
aromatic ring neutral state colorless yellow-brown ceep red reduced state greenish blue deep purple-brown ~ eep purple Table 1: color ~; l ~ nges induced upon S2O42- re ~ lu ~ - ~ n of water solution of poly :. "~ es The UV / visible absol ~ lion spectra for ben ~ ne and n ~ rht ~ ne derivatives in water solution are shown in Fig. 8. The concentration was chosen to give the equivalent of a ~ 1.2 llm thick layer of dry polymer, a reasonable value for an electrochromic device.
The molar absoll, allces were c ~ tc ~ ted in the visible region (~ g = 26500 mole-llcm-l at 545 nm;
f ~ N = 15200 mole ~ llcm ~ l at 545 nm). In addition, the colors obtained by combining the two aromatic molecules were found to be compl ~ mPnt ~ ry: either a Illib ~ lUle of the B and N homopolymers, or their copolymers resulted in a neutral brown hue.

.
2 ~ 5 1 ~ 5 ._ 1 ' 0 ~ 5 ~
gold ~~

~ Wave number (nm) hg. 8: abso, Ldnces vs. Iight wave length for benzene and na ~ l 'na derivatives of Jeffamine 900. Data were r ~ o.Jed in water solution at a conc ~ rltldtion of 0.1 mg.cm-3 using excess Na2S2O4 as reducing agent with addition of Na2CO3 as buffer.
4- Conclusions:
The polyimides described here are members of a new and broad family of redox-active polymers.
Both the intrinsic properties of the starting polyether (ion conduction) and of the bis (imide) moieties (redox center) are kept in the block a ~ hilr.; l. ~ c; obtained from polycon ~ len.c ~ ti ~ n All the redox centers in such materials are readily acceccihle irrespective of the polyether spacer length, suggesting both a fast electron-exchange mech ~ nism between anion radicals and a tendency for self ~ cspmhly (ct ~ lring) of the planar aromatic groups. This type of chemistry allows for easy manipulation of the polymers, including simple copolymers obtained by from mixtures of either the anhydrides or the starting amines. Similarly, groups allowing for further crosclinkin ~ can be incorporated.
Though no test on actual windows were made, the relatively high ~ values should translate into good coloring efficiencies. The stability of the radical anions formed during reductdon was found excellent, as shown in battery tests to be reported elsewhere. Under similar conditions, polymers incorporating viologen moietdes performed poorly.
These materials appears as promising c: ~ n ~ es for electrochromic devices, the ease of casdng thin films on transparent conducting glass or plastdc substrate being an obvious advantage. Work is in progress for redox polymers showing redox activity upon oxi (~ ti ~ n.

Acknowledgments:
. All the authors are indebted to Hydro-Quebec for fin ~ nci ~ l support of this work and one of us (CM) for partial fellowship from CNRS. We are grateful to F. Dervaux from Texaco-France for the Jeff ~ mine samples and useful information relative to these m ~ ri ~ ls. We thank P. Mailley and L. Burel for l ~ cording the absol ~ ce spectra.

~ 1,941 27 The new developed anions introduced in the docu ~, nt pr ~ cited are stable anions of faiblcs b ~ Ci ~ it ~ s perfluor ~ s or not whose cost is a priori less than those currently udlisé ~.
The ~ locun ~ ent dc thesis makes r ~ f ~, cncc to several family ~ of local ~ anions ~ little basic described below.

Family of anions:
a) The ~ anions based on various substitution chemistry of ~ perfluorom ~ thanesulfonamide (RFS02NH2), or ~ possibly polyalkoxysulfor ~ nide (RFCH20SO ~ NH2), an example of which is dc born above Jessvus;
RFS ~ 2NH2 ~ Base I YSO2X 3 RFSO2NHSO2Y
~ R ~ ~ p Cscntdnt generally ~ g ~ u ~, c.nent pe.lluor ~ of the type (cn ~ 2n + l) ~
~ X representing an appropriate electrofuge (halogen, imid ~ ole. ..,) ~ Y can ~ be a function group of any kind whatsoever and more particularly ~ rc.lJellt - a ~, lo., pe ~ n ~ organic alkyl or aUcenyl, aryl, aL ~ yl-aryl or alkenyl-aryl, containing or not subst; ~ ts OXA OR aza, slmplc or perhAlogén6. As eA ~ plc pa ~ lculier let us quote the ~ déri ~ és obtained from haloO ~ sulfonyl nules ~ poly ~ enc ~
- halogen - A pse ~ do ~ ogene, ie a radical el ~ ih. P ~ dcte ~ -l monovalent such as OCP3 - a ~ vupe''lcnt alkoxy, ~ vinyloxy or dialkylamino, diQlkèny1Amino.
- a ~ Juycnlc, nt chir ~ l.
The ~ g- ~ upc- ~ e ~ s containing unsaturations such as for example styr ~ nes, actylnte ~, ether ~
~ inylics, allyles, vinyl ~,.,. are pardculi ~ r ~ nlo ~ t adapt ~ s to obtain polyéIectrolytes or to lcurs incG ~ .s to a network n Rr ~ lolecular by any chemical bonding mode pc. ~ ant to form a network ~, nol ~ cular.

-One can also conci ~ -er chemistry ~ deri ~ r ~ nt of 1 ~ re ~ c ~ ion of perfluoroslllfinate salts as for example the pc.lluor ~ sodium tfinates (RFS02Na) on leg chlomres dc sulfon ~ mide of the type YSOzNHCl. Y being defined as com ~ e prtScé ~ nt.
This chemistry extends from hcto to that of aut ~ c ~ flnion ~ nucleophileg pcrfluor ~ s oo non, the malononitrile (CH2 (CN) 2) or cy ~ n ~ m; de (H2NCN) are examples ~ representative of prvduits as described Ci ~ 5S ~, Y se ~ tSfin issant with the same general ~ as p ~ c ~ den ~ -nent.
YSO2CHC (CN) 2 YS02NHCN YCOCH (CN) 2 YCONHCN
b) The cyclic anio ~ s derived from the family of barbitulic acids of general formula:

Yt ~ 'Z''' N, Y2 O ~ c ~ x ~ '' o ~ Z represents ~ feels ~ nt a ~ grouQel - ent carbonyl (-CO-) or sulfonyl (-SO2) -.
~ X r ~ pr ~ Sc ~ nt a group -CH (RW) - or a group -NH-. The group R can be ~ nothing, a carbon ~, rle or a suIfonyle and W d ~ f ~ nissdnt itself co ~ me - an organic group ~ Ikyle or alkenyl, aryl, alkyl-aryl or alkenyl-aryl, containing or not substitutes oxa or aza, simple o ~ l perhalogenated.
- halogen - a pse ~ oh ~ lQgene, ie an éleclroat radical ~ monov reactor ~ lellt such as CN, NO2, SCN, OCF3, SCE'3, ...
- a ~, ~ or ~ l ~ slow alkox ~ Inyloxy or dialkylamino, dialkénylamino.
~ Yl, Y2 identical or different represe ~ tfln ~ essentially a gl ~ up orgAniquo, alkyl or alkenyl aLlcyl-aryle, containing or not substitutants oxa or aza, simplc or perhalog ~ born.
Let us cite more particularly the CF3C ~ 12 groups, or the ~ groupen-ent ~ pol ~ mérisables or crosslinkable allyles. vinyl5, CH ~~ CH20 (CH2) n-c) Anions h ~ t ~ .oc ~ clicks of general formula ~

~ 1 ~ 2 Z5 ~ z3 Z representing either A nitrogen atom ~ N, goit a radical ~ CY, a group-l, ent ~ S ~ = O) Q or = PQ'Q ":

; .:

-Q, Q'and Q "identical or different ~ e ~ resentant:
- A ~ srollpcn ~ ent organic alkyl or alkenyl, aryl, alkyl-aryl or alkenyl-aryl, co ~ lt ~ -n ~ nt or not substinl ~ n ~ s oxa or aza, simple or perhalogenated, in particular pçrfll ~ rp-type ore ~
~ prCsu ~ t - a gro ~ nel t Q
- a halogen P, ~ 1, Br - a pseudo-halogen, that is to say a rfldical <~ lec ~ oat ~ .ac ~ r monovalent such as CN, SCN, N3, CE'3, OCP3, 0CnE ~ 2n ~ 1d OC2E74H.
SC ~ 3, SCnP2n + 1, SC2F4H. ...
- a ~ 5 ~ uupe.n ~ nt XC ~ or XSO2 in which X, t; p- ~ is either a fluorine atom, a groupemellt Q, a gro ~, pe.l. en ~ alkoxy, vinyloxy or dialkylamino, diaL ~ cènyl an ~ o.
The ~ -uupeQ ~ ent Q or X can be incGryo ~ és to a polymer frame, in particular by opening of doubk ~ on <contained in X or Q, either by any other chemical bonding mode pc ~ .l.etl ~ nt the f ~ ~ ion of a r ~ se ~ u ~ l ~ fi., - oll) c ~ léc ~ .laire.
Dcux ~ 3 ~ Y ~., N ~ also be part of a cyclc inco.l.ul ~ nt of 4 to 8 atoms calbone, or of hCI ~ rc ~ loops containing ~ n- ~ O, N ~ S. This cycle can in turn have ~ ~
substituents on lcs free sites of the same type as those described previously ~ me ~ (halogen, pseudo-halûg ~ ne, Q. RF. a XCO or X ~ 02 group, or be an integral part of a poly-m ~ rique ~.
l:) ans te cas where only ~ one or two nitrogen atoms are present dan ~ the cycle, it is nécéss ~ ire, for obtain good conductivities, that the group ~ slents CY completing the cycle contain ~ minus ~
two g, oupe. ~ n ~ fc "~ ement alt- ~ c ~ c ~, s of electrons, in particular, Cl, CN or CP3 or OCF3.
ef ~ and been shown by Wright and A1 (Solid-Sate Tonic ~ 1993) that the conductl ~ ity of sodium salts derivatives of in ~, ol ~ or b ~ n7; lni ~ 1 ~ 7ole dissolved dan ~ poly (ethylene oxide) was uès weak.
notnalne d'nppllcntlon:
Let us consider Ic ~ ionic compounds of formula g ~ n ~ rales M ~ X ~ an8 these materials M ~ can represent for example the cations: H ', (Li +. ~ a +, K ~, ~ g ~, Ca' ~ a + ~, ...), or a cation organ ~ such as a, .., llon; -, l,., a ~ mi ~ inium ~ a gu? ni ~ linium, a pyridinium, a nitrosyl, a sulfonium. a phosphonium, an iodonium ".. these examples are not 1 ~ mi ~; fc and ~ 'register in general. For its p ~ rt the anion X- as described ~ p ~ c ~ e ~ f ~ t has a negative charge lisée pe.-n ~ thus being to form M + X- salts of low energy retirul ~ ire ~
These con ~ posh ioniqucs have applications ~ tions in several areas essenl; çllç ~ erlt as electrolyte solutes. These areas include:
~ E ~ electrochemical systems ~ .nes such as batteries ~ or supercapacitors for energy storage élce ~ uc, as well as your ~ o ~ :(; f ~ ~ lect ~ r ~ .nes pe ~ nr to mo ~ er your light. The solutions ~
electrolydes used in these doln ~ ines can 8tre obtained by ~ en ~ rl ~ by dissoluffon of salt M ~ X- in a protic or apro ~ ic liquid, or a foolish polymer ~ as well as their ~ mixtures ,::

to form ~ by extension of the gel ~ in u ~ ilis ~ n ~ the appropriate combinations to do this.
~ Let’s also add the possibility of getting molten scls, in u ~ i1i59r ~ combinations suitable for this, for example prodquc in the case of binary salt mixtures ~
imidazolium X- and i ~ a7Qle or aprotlque for example in the case of alkyl salts in ~ id ~ .olium of X- and their ~ n ~ mixture ~ With liquids or polymers.
The ~ polym ~ res conductive ~ ~ niqucs whose conductivity is obtained by doping called "p" is say by oxidation of the macromol ~ cellular frame and insertion of anions X- in the network ~ l. As example we pc-lt cite the ~ déri ~ lrés of the family of polypyroles, polythiophene ~, or polyanilines, In this context ~, it is a ~ advantageous to be able to use the salts M ~ X ~ comm ~ source ion ~ X- in electrolytic solutions ~ during the synthase of these m ~ - ~ r ~ Y, or to do call to agents OAYdd ~ lS chemical ~ the polyn ~ is ~ ior of ~~ onv ~ s and the doping of polymer thus obtained where Ic dopa ~ ge on the pdymèrc d6jà. separate by another method.
~ Lcs polyn ~ tiQns ca!: O ~ u ~ s de poly ,. ~ es r ~ ches en ~ electrons like ~ poxydes ou ~ c ~
vinyl ethers, aziridine ~, ~ vinyls pni ~ es, isobulylenc. some hydfoc ~ bu ~ e3. The HX acids can be catalysts of this mode of polymerization either in free form, itself ~
. SOIIS latent form like salts of ~~ .onlulll or pyr ~ nillm ~ usceptibles to release these ac ~ des by heating as well as sou ~ the form of salts M + X-, M ~ representing for example a trlaryl sulfonillm or a diaryl iodonium ~ usceplible to release HX acids under the effect of r ~ or ~
~ ktinique.
~ The catalyst for reacting organic ch3mle through the use of cat ~ ons .qcdfs in this application such as lanth ~ ne, aluminum, rare earths, ...

:;

2 1 ~ 4 1 2 7 CONCLUSIONS

Originally, this work aimed at the study of cap ~; l ~ s useful ~ nt the phenomenon of double layer el.; lrocl ~, ~ llique developing on the carbon surface of large specific surface area, these components relatively old widely used in electronics, for example for memory storage computers, being considered for a few years as sources of energy of plliC ~ c ~ nce for various appli ~ tiom to the ele ~; lo ~ cl ~ ue.

In an attempt to improve the performance of current systems, it seemed wise to add short polymer segments on the surface of carbon grains, to promote ion exchange between the electrolyte and the surface of the grain, and on the other hand to facilitate the dispersion of the grains in the matrix polymer. The results obtained show that a dispersion of these carbons in poly (oxide ethylene) achieves a capacity in accordance with the geometric surface of the grains of carbon, which is after all a promising result. However, we must judge this result relative to the numerous studies of an industrial nature currently in progress ~ qll.om ~ nt, which selllPm ~ n ~
few years have made significant progress on the performance of these components both from the point of view of their energetic Cdpa ~ and their manufacturing technology.

Consequently, the initial objective of this work was less justified, both from a fundamental point of view and practical, due to the difficulty of achieving the results already obtained on industrial prototypes, and by the inability to embark on a technological study at the laboratory level. Of more, it is interesting to emphasize that there is no need to develop these capacities by solid environment, unlike lithium batteries with polymer electrolyte, the systems u ~ ilic ~ nts of liquid electrolytes which have proved their worth with lifetimes of more than 100,000 cycles, both for components marketed over the past twenty years for electronics, only for the prototypes currently developed and tested as part of research on power components for electric vehicles. However, it remains certainly to deepen the study of gelled electrolytes to optimize the performance of these components.

In the end, it seemed more reasonable to reorient the present work in a more prospective concel.lant different ~ l ~ mentc. components the sy ~ t ~ my electrochillical storage energy, and more particularly batteries with polymer electrolytes.

So the pale link of interest in this study is the field of polymer electrolytes, as well for proton conductors than for alr ~ lin conductors ~ The study on conductors proton motivated by the conduction g ~ n ~ r ~ lempnt high of these, in view of réali ~ r sources energy of strong puics ~ nce ~ allowed to obtain protonic polymer electrolytes based on the m ~ l ~ n ~ e of a molten salt with a m ~ trice of poly (ethylene oxide). These molten salts based on the doping with a strong acid of imidazole are just one example of such electrolytes, the plincipe is close to m ~ c ~ nism ~ s of conduclivi ~ e of the proton in the ~ y; ~ Biological lemes. The polymer electrolytes studied in this context are açtllPllPment conducleul ~ anhydrous proton having the best res ~ llt ~ t ~ in terms of conducliviles, and this particuli ~ re., lcnt at low ~ lllpélalu ~ e ~ a conductivity of 10-5 Ohm ~ l.cm-l being obtained at -20 ~ C in the case mPillPllr.

The alkaline conduits are of undeniable practical interest for the development of systems electrochemicals in thin films, and more particularly of lithium batteries Poly (oxide ethylene) or its derivatives being currently ~ known as the best choice of polymer solvent for such an application, the essential point remaining to be improved conce.lle the chemistry specific to the anions which constitutes the lithium salts used in the electrolyte; the anions currently used drifting for most of the fluorine chemistry whose cost is relatively high for the intended application.
An analysis of the physicochemical data on the acidity and the stability of different anions allowed to define relatively simple criteria allowing to arrive at dissociated and hardly oxidizable salts in dipolar aprotic solvents, and particularly polymeric solvents. It was thus possible to identify different families of anions meeting these criteria, which have been confirmed interest by measuring their conductivity in poly (ethylene oxide), and especially to show that probably existed substitutes that were not fluorinated or based on fluorine chemistry easily a ~ cPscihle to actuPIlPmPnt salts used, and therefore a subst ~ ntiPll ~ nPnt cost less.

The second part of this work focused on the study of polymers with mixed conduction. The grafting of polymer segments on the carbon surface developed for the capacities is not without of interest in this context, since directly transposable to lithium electrolyte batteries polymer, the insertion reaction in their cathode necessi ~ nt a homogeneous transport of ions and electrons, gold, these carbons allow to localize at the nanoscopic level a conduction electronics and ion conduction.

The study of polymers with mixed conduction has also been extended to the copolymer composed of a species electroactive and a segment of solvating polymer, and more pardculi ~ .ll ~, nl polyimides obtained from aromatic molecules dianhydrides with sc ~ -llents of poly (ethylene oxide) min ~ s, the study of which was initiated aupaldvdnt au labolatoil ~. it has been shown that one could extend the possible choices of aromadic nuclei, but that the synthesis or the reversibility ele ~; lr ~; ... ique of some of them could be questioned. However, the more extensive tests on the poly (benzodiimide-co-oxide of ethylene) shows all the interest of this family of materials for i . ' the ~ ~

_ improve the kinetics of insertion of lithium in the mat ~ riallx of insertion, allowing to perform a homogeneous catalysis of electronic transport on their surface, and gives hope important applications of these m ~ teria ~ x in b ~ tteries with polymer electrolytes. But the the most interesting result in our opinion is to have been able to show the exceptional stability of these polyimides as m ~ téri ~ u cathode during a long-term cycling test. This result co, -rir ... e la n ~ ceSsité de s'~ .ess ~ r more particuL; ~ au ~ redox species compati ~ them with Basic chemistry specific to lithium-ion batteries On the other hand, the lifespan of this ~ s mat ~ ri ~ llY and performance ~ s, both in terms of energy density and pllicc ~ nce ~ that we can e ~ -er for batteries ~ ltilic ~ ntc such mat ~ r ~ x shows that it is possible to consider the development of either b ~ lithium tteries of nits ~ nce which are in fact an el ~ mçnt of response to the objective which was fixed initi ~ lPmçnt to this work, or of b ~ ttçries being able to carry out a number of very important cycles for applications to stationary electrical energy storage systems. In terms of chemistry, these results tend to show all the interest of being deeply interested in the study of materials o ~ a, li ~ read electroactive in the context of work on ele ~ ~ hillliques generators.

The interest of the materials studied in this work is not limited to the field of generators éle ~ ochi.niquçs, but can extend in particular to that of éleclloch glazing ~ mes which PrésPntçnt also a potential outlet for these. More generally, it is better to group all of these studies within the framework of the chemistry of aprotic environments.

To conclude, the results gathered in this work confirm our choice to have favored a a priori prospective research, rather than the specific study of a particular point. Despite the shortcomings inherent in such a choice, this option ended us ~ lempnt allowed us to meet the initial objectives ~ lemçnt envisaged, and to extend both the concepts and the palette of mat ~ ri ~ u for the development of various electrochemical systems, as well as contributing to the knowledge of aspects fondamPnt ~ related to this area ~ of science VS

~ 194127 Annex A: Experimental methods Al Thermal characterization of samples:

The c ~ lo ~ m ~ tri ~ - diLrélen ~ ielle scanning, or DSC (Dirr ~ .en iel Camping Calorimetry), is an oudl of choice for Cd ~ t ~ tion of thermal pr ~ pfi ~ t ~ s as well as, in this case, salts proton fuses as polymer electrolytes. Its principle consists in the analysis of the dirr ~ this enthalpy recorded between an e ~ h ~ ntillon and a l ~ f ~ .el ~ ce inert which undergo the same heating process. This gives access to the physical properties of such binary mixtures, and in particular to the glass transition temperature Tv (below which no more movement of reorganization does not exist at the microscopic level in the m ~ t ~ ri ~ ll), at the temperature of recrict ~ llic ~ tion Tc, and at different melting temperatures Tf. Figure 1 gives a typical example of such thermogld.,. ", e obtained by DSC.

ABCD
r x CD ~ T, ~

Tg You "er ~ ture>
Figure1: Example of ll, er ".og, dn""e typical of a blenary mixture obtained by DSC.
The tempe, ~ lul ~ of glass transition (zone A) is thus caract ~ risCe by an allgmlont ~ tion of heat specific of the material in the form of a "step", we chose to determine it as the temperature in the middle of it. Zone B presents an exothermic phenomenon attributed to the recrict ~ llis ~ tion of all or part of the amorphous phase. The first endoth ~ rical peak (zone C) characterizes the temperature from which a liquid phase (or amorphous for polymers) appears in the material, and col, ~ ond to that of the bearing eu ~ ectic Te. When at the second point of fusion (zoneD), it indicates the temperature where the whole of the m ~ téri ~ u is melted.

These data make it possible to construct the phase diagram of the materials, by studying the thermal properties of ech ~ ntillor c of different concentrations, as illustrated in Figure 2.

L ~ _____. TCD

Tf ~ - - - / L ~ CD
EA + L ~ \ /
~ Te - A ~ ._ Tg _ _ _ - - - - E ~ D

A Co "" ~ ositi ~ n CD B
Figure 2: Example of phase diagram of a binary system.

In the present case, the sample studied is placed in a sealed anodized alminium capsule.
(Dupont), and placed on the measuring head of the DSC device (Netzsch STA 409), an empty capsule serving as a reference element. We then perform two temperature increases succe ~ cives for a scanning speed of 10 ~ C / min, each preceded by cooling to -100 ~ C (- -30 ~ C / min).

It is important to note that the Netzsch STA 409 device does not perform a DSC measurement, which would consist in de ~ rminin the heat flow nece ~ ss ~ ire to balance the temperature of the sample to that of the reference, but a differential thermal analysis, in other words the measurement of the temperature difference between the sample and the reference, which is subsequently converted to that would give a real DS analysis (~

A.2 Conductivity of electrolytes:

The electrolyte films, with a thickness of 100-200 ~ m, are placed between two electrodes stainless steel blockers. If we assume that we can symbolize the electrolyte by a circuit electric parallel RC, determination by impedance spectroscopy of the system impedance r ~ allows to determine the r ~ e ~ ist ~ nee of the electrolyte for a tempél ~ lul ~ given (RéleC) ~ and this from of the ~ eyl ~ sentaLion of impé-iance in the Nyquist plane, as illustrated in Figure 3.

Thereafter, the conn ~ icslnce of the geometric factor of the sample makes it possible to determine the conductivity of the electrolyte for a given temperature. In this case, the measurements are performed using a Hewlett-Packard impe ~ l ~ nce spectrometer, model HP 4192 A, coupled to a microor (lin ~ tor, in the frequency range S Hz-13 MHz.

.

~ /
Q

Zr (Ohms) Relec Figure 3: Typical example of an impedance spectrum in the Nyquist plane The description of the procedure for preparing the samples for the measurement of conductivity of proton electrolytes having already been described (Cf ChapI, II-2-3), it will be limited in the present case with electrolytes ~ lc ~ linc. Polymer electrolyte films are obtained by casting in a glass ring placed on a Teflon plate, a solution in acetonitrile, previously degassed, the salt studied and poly (ethylene oxide). After evaporating the solvent to ambient temperature, the films are placed in a glove box at least one week ~ inp before carrying out the conductivity measurement. Note that the polymer is dispersed with m ~ th ~ nol before adding acetonitrile (5-10% by volume relative to acetonitrile) in order to accelerate the solubilization of polymer. POE 5.106 is commonly used to make electrolyte films with good mechanical properties which facilitate their subsequent handling. However, POE 5.106 has a mass too high to allow it to flow into the conductivity cell, and ensure thus a homogeneous bonding of the electrolyte on the electrodes. Also, the electrolyte is composed of a weight mixture of 60% POE S. 106, and 40% POE 9.105, thus allows ~ nt to guarantee the contact between the electrolyte and the electrodes. It has been verified that such a mixture does not affect the value of conductivity obtained, and this for the lithium salt of dicyanotriazole.
.

In order to carry out the conductivity measurement, the electrolyte films are placed on a support composed of two steel cylinders with a diameter of 16 mm, and a Teflon guide allowing ~ nt guarantee the ~ lignPmerlt of the two steel electrodes. This support is then introduced into a glass conductivity designed to be placed later in a Buchi TO-50 oven, all these operations are carried out under an argon atmosphere in a glove box. Once introduced in the oven, the cell is connected to a vacuum pump, and brought to a t ~ mp ~ ldlu ~ of 120 ~ C during minus fifteen hours. The measurements are then made from 120 ~ C up to the temperature ambient, at three hour intervals typiquPmPnt The thickness of the electrolyte is det ~ rminPe at the end measurements using a mechanical micrometer. It is important to note that to ensure the accuracy of this measurement, the surf ~ these two steel cylinders were rectified ~ s on a lathe.
j! ' III

Finally, Figure 4 shows the conductivity obtained in ~ decreasing ture for LiTFSI at O / Li = 8/1, for two measurements made by different e ~.

1 o 2 ~ - LiTFSI O / Li = 8/1 E ~ Decreasing temperature -E 10-3 - o C ~ J
O ~
'~ D ~
C ~ 10 4 - ~ _ ~ POE 6.106, [Sylla]
C ~ ~ ~ ------ POE 9.105, [Hammouche]

Temperature (~ C) Figure 4: Comparison of the conductivity for two fl / ms of POE / LITFSI at O / Li = 8/1.

This confirms that it is entirely possible to carry out conductivity measurements, under the conditions ~ ccessihles in the laboratory, with quite correct precision, the difference observed between these two curves corresponds in fact to a maximum dispersion of +/- 2% of the values, except for the two extreme points which correspond to a dispersion of the order of +/- 5%. Of more, it should be noted that the difference for a temperature ~ 90 ~ C probably corresponds to a slight A.3 Electrochemical characterization using microelectrode:

The support allowing perform the tests in microelectrode is described on Figure 5. As a rule we are preparing a film of material to study, or one deposits directly on the electrode this product from a solution, or a dispersion of the latter in a solvent. The potentiostat used to control the measurement a been done at lab, ~ lo ~ etzschl A.4 Character ~; electrolytic position ~ .lique from battery:

IV

-The characterization of materials from a battery makes it possible to provide numerous inform ~ tione, in particular to obtain the CuulbCS of decl ~ & ge int ~ miost ~ tick ~ and flight ~ étrie slow permp ~ tt ~ nt of c ~ dct ~; ser, from a Mac Pile potentiostat ~ (BioLogic) developed for this effect. The ep ~ icse ~ typical of the different components of such a battery are given on the Figure 6. In addition, conl. ~ i. ~, ... ~ a ~ t at the microelectrode, the capa ~; coulombic island of m ~ t ~ r ~ on which s ~ erç ~ read the electrochemical test is known, by weighing the quality of m ~ t ~ n ~ ud ~ pos ~ on the cathode e ~ 25-50 ~ m steel pellet with an area of 2 cm2, serving _ Cl ~ t ~ l ~ to e ~ 200-300 ~ m supporting the cathode. It is therefore possible to Flgur ~ 6 ~ p ~ typical issuer ~ 3 of the constituents of define the theoretical capacity of the battery (Cth) ~ It ~ r; ~ used for ~ chemical tests.
expressed g ~ n ~ ralem ~ nt in Coulomb number by cm2 or in mA.h per cm2, which is obtained by dividing the capacity in Coulomb by a factor equal to 3.6.
In no case is the capacity of the anode taken into account, this being due to the thickness of the sheet of lithium is approximately 100 times that of the cathode.

Thereafter, during reduction or oxidation of the cathode m ~ t ~ n ~ u, the capacity is determined practice obtained for a given current or speed of scanning in potential, and it is thus possible to define the percentage of use of the electroactive material (% Use) by the Mpport of the pMtic capacity to the theoretical capacity. During a battery cycling test, the evolution of this per ~ meter perrnet of qu ~ ntifier degradation of the cathode material during cycling. In addition, the ratio of the capacity obtained in discharge, on the capacity in recharge, which one expresses by C / D ratio makes it possible to obtain the faradaic yield which is equal to 1 in the absence of phenomenon parasite. If when recharging the battery, a dendrite appears which grows up to the cathode, the coulombic yield becomes greater than 1, which characterizes this kind of phenomenon well.

On the other hand, the intensiostatic discharge test of a battery for different discharge currents allows to assess the influence of the discharge regime on the available capacity of the cathode. It is then possible to deduce the energy and power density for a weight or a volume of given battery, generally from a geometry corresponding to an optimized system.
,,.
Finally, note that the tests are mostly carried out in a cell placed in a oven in which circulates 1 'argon at a temp ~, dluie of 80 ~ C, the sealed boxes used being waterproof only about lS days at such a temperature ~ .alu- ~ ,. For tests which are not carried out under argon, we check at the end of the test by dismantling the battery that the lithium is still bright and therefore not oxidized or nitrided, which guarantees the validity of the test.
A.5 Characterization by NMR:

~ 1,941 ~ 7 Nuclear magnetic resonance (NMR) makes it possible to study the structure and composition of synth ~ ti $ ~ s ~ organic products to ensure their purity. It is performed either on a device Brucker working at 80 MHz (carbon and proton), or on a device of the same brand working ~ n ~
at 200 MHz (carbon, proton, fluorine). The reference for fluorine being CFCl3 in chloroform, it is n ~ ceCc ~; ~ de sou ~ e 3.87 ppm to the value obtained for the ~ f ~. ~ ,. this. à l'n ~ élon; l,; le A.6 Characterization by spe ~; infrared lroscopy:

The, pe; ll ~, s infrared are obtained on a Fourier transform apparatus Nicolet ~ -l ~ 410.

Reference:

[I b "", ~ uche] Common, l ~ ~ 'n per ~ onne "e [Netzsch] Notice t ~ l, r,, _e Netzsch STA 409.
[Sylla] Sylla, S. Doctoral thesis of the University of Grenoble, INPG, 1992.

,,. ~

VI

Claims (7)

The embodiments of the invention, about which an exclusive right of property or lien is claimed, are defined as follows: 1. Delocalized anions useful as electrolyte solutes of formula:

RFSO2NH2 + Base + YSO2X ~ RFSO2NHSO2Y

~ RF generally representing a perfluorinated group of the type (CnF2n + 1) ~ X representing an appropriate electrofuge (halogen, imidazole, ...) ~ Y can be a functional group of any kind and more particularly:
- an organic alkyl or alkenyl, aryl, alkyl-aryl or alkenyl-aryl, containing or not containing oxa or aza substituents, simple or perhalogenated. As a particular example, let us quote the derivatives obtained at from sulfonyl halides from polystyrene, - halogen, - a pseudo-halogen, that is to say a monovalent electron-withdrawing radical such as OCF3, an alkoxy, vinyloxy or dialkylamino, dialkenylamino group, - a chiral group. 2. Cyclic anions derived from the family of barbituric acids of formula general:

~ Z representing a carbonyl (-CO-) or sulfonyl (-SO2) - group.
~ X representing a group -CH (RW) - or a group -NH-. The group R can be nothing, a carbonyl or a sulfonyl and W is defining itself as:

- an organic alkyl or alkenyl, aryl, alkyl-aryl or alkenyl-aryl, containing or not containing oxa or aza substituents, simple or perhalogenated, - halogen, - a pseudo-halogen, that is to say a monovalent electron-withdrawing radical such as CN, NO2, SCN, OCF3, SCF3, ...
an alkoxy, vinyloxy or dialkylamino, dialkenylamino group, ~ Y1, Y2 identical or different essentially representing a group organic, alkyl or alkenyl alkyl-aryl, containing or not containing oxa substituents or aza, simple or perhalogenated. Let us cite more particularly the groupings CF3CH2, or the polymerizable or crosslinkable groups allyl, vinyl, CH2 = CH2O (CH2) n. 3. Heterocyclic anions of general formula:

Z representing either a nitrogen atom ~ N, or a radical ~ CY, a group ~ S (= O) Q or ~ PQ'Q ":
Q, Q 'and Q "identical or different representing:
- an organic alkyl or alkenyl, aryl, alkyl-aryl or alkenyl-aryl, whether or not containing oxa or aza substituents, simple or perhalogenated, in particular perfluorinated of the RF type.
Representative:
- a group Q, - a halogen F, Cl, Br, - a pseudo-halogen, that is to say an electron-withdrawing radical monovalent such as CN, SCN, N3, CF3, OCF3, OCnF2n + 1, OC2F4H, SCF3, SCnF2n + 1, SC2F4H ..., - an XCO- or XSO2- group in which X represents either a fluorine atom, a Q group, an alkoxy group, vinyloxy or dialkylamino, dialkenyl amino. 4. Use of the anions according to claims 1, 2 or 3 in systems electrochemicals such as batteries or supercapacitors for energy storage electric, as well as the electrochromic devices allowing to modulate the light. 5. Use of the anions according to one of claims 1, 2 or 3, as source of ions in electrolytic solutions during the synthesis of electronic conductive polymers. 6. Use of the anions according to one of claims 1, 2 or 3 as cationic polymerization catalysts for electron-rich polymers such as epoxides or vinyl ethers, aziridines, vinyl amides, isobutylene and certain hydrocarbons. 7. Use of the anions according to one of claims 1, 2 or 3, in the catalyzes organic chemistry reactions involving active cations like lanthanum, aluminum, rare earths and others.