Insights into the polymerization kinetics of some α-silylated thiophene oligomers

ELSEVIER Journal of Electroanalytical Chemistry 435 ( I t ) t ) 7 ) 85-94

Insights into the polymerization kinetics of some ot-silylated thiophene oligomers

P. Hapiot "", L. GaiUon P. Audebert t,.l J.J.E. Moreau c, j..p. L re-Porte c, M. Wong Chi Man

'~ t.b~m'a~ir¢ d'Eh'¢w~¢h~nPie M.l~'.lair~ ' de I't/oh'~,rsif~ ~ Pm'i,~ 7, Unit~ ~h ~ Re~'htr~'he A,~,~o~ i~¢ at# CNICS No, 438. 2 I~hwe ,lus,~i~'u, ,75251 Pari,~* C,~des 03, Fran, ¢

' t/~it~ A,~s.~q¢!*' ¢m ('NRS No, IOOL lh~hP~tf~ M.ml~,lli~'r II, S~'itm'¢s ~'t 7~,~'hniqm,s dt¢ l~#ngtdo¢, phwe lh~bm~ Ilatallhm.

Receh'¢d 20clttlx, i' 1t1911; i'evisctl II Nt)v~lnl~r Itl|)t'~

Abstract

Oxidation o1' several silylthiophcne oligomers has been studied in dichloromethane and acetonitrile b> cyclic voltammclry, [~tailcd meelmnislic studies show that Ihe oxidative coupling of silyloligothiophenes involves the carbon=carbon/oond tormalion between two cation~radicals, as previously observed lot other non=silylated oligothiophenes. The lifetimes of the cationoradicals are fimnd to be dependent on the solvent, indicating a nu¢lcophilic assistance to the C.oSi cleavage, This behavior is in agreement with a reversible coupling step leading to the fiwmation of a dimeric cation followed by the irreversible cleavage of the C-S~ bond. Evidence for the exislence of an equilibrium between the monomeric cation-radical and an associated Ibrm (~dimerizat ion) has been oblained from the variation of Ihe rcdox potenlial with the temperature and concentration fiw ahe mosl stable cationoradical. ~ It)t)7 EIs¢~vier Science S.A.

Kevwords: Thiophcne oligomers; a-Silylatkm; Polymelization; Cyclic vollammelry: Kinelics

1. Introduction

Thiophene oligomers have attracted considerable attention in the las! decade [1=36] and retbrences cited therein) in several research fields including molecular electronics (see for example Refs. [4,11,12] and references cited therein) and also to improve the conjugation and the conducting properties of polyt~,iophene. These properties of polythiophene~ have been associated with the selectivity of earbon~¢arbon bond Ibrmation in the coupling of monomeric or oligomeric unit:; [23] which has led to the development of improved synthetic route.~ to polylhiopheues. Oligothiophenes are also inleresting model molecules for understanding the behavior of polythiophene itself or the mechanisms of electropolymerization, in fact. such studies, until recently, could not be managed successfidly, due to the too high reactivity of the cation-radical of the monomer itself. The way to solve this problem is either to have access by fast electrochemical methods to short lifetime~ [37] or to stabilize the cation-radical chemically [38]. One way that we experimented with ~uccess before in the ca~ of pyrroles [38-40], is to use new fast voltammetric methods with the help of ultramicroelectrodes associated with fitst potential sweep rates. The other way is to stody model molecules such as thiophene oligomers [8= 10.16o~ 18,24,25,34]. in which thiophene units are linked by their ~opositions, and for which [h~ reactivity of the radical cation can be tailored considering the length of the chain and the nature of the aqerminal subs~i~uents. Indeed, a combination of these mcthod'~ can be expected to provide still more convincing insights into Ihe problem. In a detailed mechanistic study of the electrochemical oxidation of several substituted quinque~hiophene~; and mi~'cd pyrro!e + thiophene oligomers, it has ~cJi

* Corresponding author. E-maih hiipiot(=n~paris7.jussieu.l'r. i Also corresponding author.

0022-0728/97/$17.00 ~. 1997 Elsevier Science S.A. All rights reserved. Pil $0022-0728(97)00 ! 16-2

\ P. Hapiot el al /Jourmd af Electrm~ptalytical Chemisto" 435 ¢~9~#7~ 85-~4 86

Me,St ~ ~ ' ~ S i M e ~ (n-,2)I"

Scheme 1. Gc~erM formula lot" the oligomers studied: Mc~Si-,T-Sil~e~.

\ • ~ • . •

demonstrated recently that the carbon-carbon bond fonnation results from the couphng el,two canon-radicals |orated by oxi~tion of the ol igo~rs [25]. T h ~ conclusions are similar to those obtained several year~ago concerning the first steps of the el~tropolymerizati~ of pym~ie. Moreover. it has also been shown that besides ,h',:L~e pogsible o'-dimedzation ( c a r b o n - c ~ n ~r-bond formations), an equilibrium can exist between cation-radicals of oligotlt!$phenc and an associated foxm of two cation-radicals in .~lution [8.-10.13.16.18.24.76] and tht~t ~he~ interactions between c~,,ion-r~dicals can explain

ESR ~reas ing activity during thiophene polymerization [41]. This associated form has been a ~ r i ~ d to the existence of a ' , rMi~r ' by analogy with the fact that other organic ion radicals tepid to ibrm ,r-dimers in solution [10,16]; however, no clear evi~nce couceming the exact nature of this direct ha~ ever bee, o~tained, (For simplicity of t ~ text, we will use Ih, term.~ ,r~dimer m qualify thi~ asm~iated form and ~dimerizatiou by opposition m the (r~dimedzatii,,,n leading to the t'ormati(m of earbon~carbon bonds.) Evidence fiw ~'odimerization has been (~tained with substituted oligodil~phenes which lead m tbe formation of stable cmion-rMicals [8~ 10.13.16.18.24.3~.36]~ but it is likely that "~dime~' coulffMso exist in °!~lymerizable' oligothiophenes even if they h:tve not been reported with such comt~unds. Thet~lbre, ~dimers have been proposed a~ F~ib te inlermediates in the coupling process leading to the intonation of Ille erodimc~ [Ib,36] and such an inlem~diate would be expected to fiwor a radical~radical couplin G railer than a t~action of one catiouoradical with the ~tarting monomer.

More paflieulafly, oligothiophenes have been shown to polymeri~ not only when the ~ temtinal groups are hydrogen atom~, but also in the cam of silylosnbstitutod monomers [6,26~34]. Several groups have r~poned the synthesis of po~thiophone films from silylated precurmrs [26o~M] and the premnce of ~silyl substituents at the thiophene ring has b~n ~own to favor the fi~rmation of highly structured polymers exclusively linked 2.5 throughout [29]. On the basis of Raman and photoluminescence studies, the lx~lymers t , rained from Me~Si-2T-SiMe:~ or Me~Si~3T~SiMe~ (me Scheme I ) appeared to be highly structured, with higher mean conjugation lengths and lower of det~.cts, when compared t~ t~:~lyme~ obtained from the cott~osponding nonosilylated monomers [31].

In a previous preliminary study [34], we have detem~ined the Ill, times and oxidation potentials fiw a series of mono~ and diosilyl~ted monomers. We found that the reactivity of these ~i t~;ene oligomers is afl~ted stnmgly by the premnce of trimethy!silyl groups compa~d to what is t~served with other olig~thiophenes (with hydrogen atoms or other non-silylated substituent~ or, aot~rminal position~). The~ results enabled us to realize an electrocl~emical study of the oxidation of these silykdigothiophencs to m~derstand the particular eflL'ct of silyl suhstituents anti more generally to gain new insights ¢o~cemin,, the ~activity of ollgothiophenes.

The diligent o l i g o t h i ~ e ~ s have been synthesized according to previously publis~d procedures [31,32]. The solvents u~toni~rib~ (Meek, Uva,~l), toluene (Prolabo, analytical grade) or dichloromethane (Sigma, HPLC grade). The

~ n g el~trolyte was tctrabutylammonium pcrchlorate NBu~CIO 4 (Fluka~ puriss) which was u~d at a concentration of 0~ I mot I ~ in all esl~rimems,

2,2, Celt ~ ele~'rr~tes

The counter electtx~l~ u s a Pt wire and the referen~'e ¢l~,t~le an aqueous ~turated calomel electra;de ( l: "~ vs. s e e =/:'~ vs, SHE ~ 0,2412 V) with a ~l t bridge containing solvent ~ d the supporting electrolyte. The working electrode was either a 3ram dmtnc~r glassy ~rbon disk, or a I mm diameler goid ¢1¢¢tr~¢ or 5 to 17p~m diameter gold ultramicroelectrodes ~hk:h w~e¢ ~ as dc~ril~d p~¢viously [42,43], ' ~ cell ~ s thetmostated by outside circulation allowing the control of the iempoaiur~ of th¢ solution [24], The referelrce ¢l~trt~ae was as de~ribed previously [24] thermostated at 20°C by an i~poa~¢a{ ~in:ulation oi" water, As di~ussed previously, such an arrangement leads to a negligible contribution (less than 0,03 mV K ) of the. thermal junction potential to the variation of the measured vedox potential with the temperature. The potenti~ of the reference electrode, was checked at 20N~ against the fcrrocene/ferricinium couple (E ° = + 0.405 V (SCE)) befme ~lmi ~ e~-h experiment, All potentials are reported according to this standard.

P. Hapiot et al./ Jo:u~ol of filectr~mmd, vth,al Chemistn." 433 f 1997~ 85- 04 87

2.3. b~strumentarion and procedures

Electrochemical instrumentation consisted of a PAR Model 175 universal programmer and a i~ome-built potentiostat equipped with a positive feedback compensation device [44]. The cathodic and anodic peaks were measured on a 310 Nicolet digital oscilloscope. From repetitive measurements, the error made on the relative determinations of E t /2 during one set of temperature variations was found to be better than ~ 5 mV, except for the measurement made with the lowest concentration (error ± 8 mV) because of the low value of the current.

For high scan rate cyclic voltammetry, the ultramicroelectrode was a gold wire sealed in soft glass [43]. The signal generator was a Hewlett Packard 3314A and the curves were recorded with a 450 Nicolet oscilloscope with a minimum acquisition time of 5ns per point. The values of the lifetimes were obtained by comparison of the ex~rimenta | voltammograms with simulated curves at several scan rates. The simulated voltammograms and the R-A curves characterizing the C R - C R limiting mechanisms were obtained by numerical computation according to the Schmidt method [45] of the preceding sets of partial derivative equations with initial and boundary conditions using the defined dimenskmless variables. Calculations were performed on a l~,,-type microcomputer 486DX50 alter intnnluction in the c~lculations of the redox potential E ° for the oligomer/oligomer cationontdical couple and of the inversion potential of the scan of the experimenl.

3. Results aud discussion

3, L ('yclh' volumu~let~3, ¢!f the si!~'h~ligothiophenes itl avetonitrih, am/di~'h/otwmethatw

The different oligothiophenes corresponding to the general formula displayed in Scheme I have beet~ studied by cyclic voitammetry both in acetonitrile and dichlommethane at different sweep rates, in acetonitrile and at low scan rates (in the 0,1 V s~ t hinge or lower), the voltammograms for the oxidation of the silyloligomers presented in Scheme I are completely or partially irreversible. An ¢xatnple of such a voltammogram is presented in Fig. I for the oxidation of Me~Si~3T-SiMe:~,

At low scan rate (t, < 0.5 V s i ), on the backward scan. a cathodic peak is visible corresponding to the redtiction of the cation~radical of higher oligomers (mainly the cation-radical of Me~Si-6T-SiMe~) produced during the forward scan (see Fig. I(a)), When increasing the scan rate. a there positive peak i~; visible during the backward scan which is due to the reduction of the cation-radical of Me~Si-3T-SiMe~ indicating that its cation-radical is stable dtnfng the time of the cyclic voltantmetric experiment (see Fig. I ( b ) ) . These results are in agreement with an oxidative coupling of Me~Si~3T~SiMe~ leading to the fomlation of higher oligomers.

The electrochemical behavior observed for the bi- and quaterthiopllene detwative~ is similar, hnt~ a~ expected, tile ,,can rate required to observe the change of behavior depends on the leilgth of the oligomer: for Me,St, 4T~SiMe~. tile wave i,, completely t~versible liar scan rates lower than I V s t in MeCN and, on the c,mtrary, for Me~Si~,2T~SiMe~ ,,~can rates higher than 20ff0V s- i are required to observe a beginning of reversibility [34], Notably. if the same general features were found in acetonitrile and dichloromethane, showing in both cases the oceutxence of a coupling reaction, the lifetimes of the different cation-radicals depend strongly on the nature ot' the solvent [34], This effect, which seems to ~ specific to this kind of oligomer (see below), enabled us to obtain a better characterization of the cation-radical prior to the carbon---carbon bond formation.

I : .... b

iii ....... i i1, : ̧ 0.6 1,2 0,~ |,~,

Fig. I. Cyclic voltammetry of ti;e ~xidation of Me ~Si~-~YF-SiMc ~ (v'~ l0 ~ me11 ' ) in ,rct,ttitdle ,m a ~ mm ~t; ,nctcr gl~t~,,~y c~tb~t~ ,:~e~:|¢~J~, ~ ~t, ~w v~-IVs -t (a), 20Vs -t (b): T= 20°C.

g~ P. It.plot et aL / Jounud ¢~f Elecotumalytical Chemistry 435 (1997) 85-9~

32, ¢roDimerization reacthm

As mentioned before, it was shown that stable cation-radicals generated from substituted terthiophenes and especially from quaterthiophene derivatives, are in rapid equilibrium with the '~r-dimer" of two cation-radicals and that in the millimolar range concentration and at room temperature the "¢r-dimer" is the major species [8-10,13,16,18,23,35,36]. It is likely tim| "¢r-dimers' exist also between chemically unstable cation-radicals of oligothiophene, and these were proposed as a possible intermediate in the coupling mechanism [16]. On the contrary, on the basis of the temperature dependencies of

d i ~ z a t i o n rates during the oxidation of unsubstituted long-chain oligothiophenes, other authors suggested that the carbon-carbon bond focmation occurs directly between the two cation-radicals and does not involve the intervention of pce~complexati~ (in this case ~ "¢r-dimer') t! 7]. However. in a previous publication concerning the oxidation of a series of ot-c~'.bi~trimethylsilyl)-substituted oligothiophenes, the authors exclude ~r~dimerization due to the steric hindrance by the ~bsfitueats [6]. hut no temperature-dependent investigations were undertaken. On the contrary, with solid-state ~xithiop~nes. coplanar ~:~mfigurations allowing maximal ¢r-'n'-imeractions between two oligome~ were ~bund even for mol~ule:~ wi~ ethyl or irimethylsilyl suhstitaents, indicating no significant effect du~ to steric hindra.~ce [46]. This Ig~bi l t |y enabl~J us to check whether ~uch ~-dimcrization exists prior to the din~r/~ation with silylated oligomer and the lon~, lifetime of the eationoradical of Me~Si~4T=SiMe~ makes it a g~agl candidate for such study.

All these gtudies are ha~d on temperature and concentralion variations, moditication of the UVosp¢c|r;| or of the ~dox i~)len(ial. The ~¢ond method is Jess precise d|an the sF.cctroscopi¢ investigation but has the advantage of i~¢ing usable Ibr a le~g stable catioaoradical, a~ in the case of Me~Si~4T~SiMe~ [I 3.24]. To improve the solubility of Me~Si~4~ ~SiM¢~ al low temperatu~ in ace|out|rile a mixture of ace|out|rile 30'~ and toluene 70~ was used instead of the pu~. so. ~nt. In this mixture, reversible cyclic vottammograms lot the fi~t oxidation were obtained for scat) rates around 0.2 V s ~ Ther~qb~. cyclic v~tammetry allows the determination of the redox rg~lent ia l for the oxidation of the oligomer, by sin)ply considering Ihe mean of the forward aM backward I~ak l~)tentials E~/: ~ (b~. + 1~) /2 . This ~versible potential de~nds on the possible exlsten¢¢ of the ~rodimerization equilibrium (which has been shown to be fast) which can ~ demonstrated by its tem~rature and subs|rate concentration dependencies [13.24]. We have explored the E~/: variation of Me~Si~4T-SiMe~ w ~ n varying the temperature of the solution t?om as low as 230 K up to 330 K~ As sl~swn previously, the redes I~lenlial measured during such experiments is the E ~ of the T / T * couple only in tl~¢ tem~ratare range (high leml~rature) where no ~dimefization occurs, otherwise the measured potential is in fuel assigned to the monomer /~d imer system. Fig. 2(a) sinews the plot of the E~/~ !~)temial vs. the temperatu~. As found previously for the oxidation of non°sil~.~la|ed ~ h s t i t u l c d oligothiophenes, two straight lines ar~ ob~ain~ co~spondi~xg to the lwo tempenm|r¢ gk~mains. The slol~S of E~/~ ~.f(7 ~) rcpte~nt the respective standa|xl enlropics of each pl~dominant electrochemical reaction and the AS ass~-iated with the two reactions T - ¢ ~ T ' and 2T ~ 2e ~ '¢rodimer" entropies are differenc As expected, the s lo~ at low temperature is higher because the entropy of lbmlation of the dimer is larger than the entropy Ibr the formaUon of the single radical cation. From these values, the enthalpy and enux~py values attd the equilibrium constants at n~om temperature tbr the ~ac|i~m of tbcmation of |he dimer from two r~dicabcations (~ x T ' c * '~dimer ' ) can h¢ ext,|clod [13.24]. From this a~lysis, ~ estimated tl~ equilibcium constant a! r,~m tem~Pature ibr the 'incliner' of Mo~Si~4T~SiM¢~ as /f 10 ~ I reel: ~ (AS ~ ~ 13)J K ~ reel: ~ and ~ H ~ ~ f~)kJ reel ~ ), 1% confirm the exis|ea¢¢ of ~h¢ ~d imer in ihe low ~mperaiur¢ domain, we Sso investigat~ ti~ variation of the E~/: with the initial concemraUon of oligomcr c ° in the two don~ins of teml~r~ur~, T'ne~ vaciations a~ d i spby~ in the Fig, 2(b). At 36°C, a very small variation is observed wilh the ¢~w~'enlra~io~. ¢~a~fitming that in this case that th¢ ~ / ~ iS ~ua! |o i~ of ~ T / T ~ and lhus that the ca|ten-radical exits in the fi~m of a s id le s ~ i e s , On the contt~ary, the ~m¢ ex~dment performed at - I 7'~C " gives a linear variation with a slope

l J l 1 ~ ~ ~ 2 ~ ~ 4 5 ~ ~ 5 ,3 • 2 5

~ K |~c7m~,i L ~

~ 2.. ~ ¢ h c x ~ r ~ r y ~ ~ ¢ S i - 4 T - S i ~ ~ ~ 3 Into di~n'~cer gla.~y c',uflon d ~ t r o ~ in an ac~toniLrile + toluene mixture (30 /70) . (a) Varia;ion of ~ ~ ~ ~,'~h ~h¢ ¢k~=so~ ~en~pcr,~luce: ~-~ ~ IO- ~ n ~ | - I; r = 0.2 %" ~ - ~. ~(b) Vail|toll o f E j / : wi~h ~he inLqal concentration c ° / m o l | - ~ o f Me~Si -4T-S iMe ~:

P. Hapiot et al. / Jounud oJ" Electregma!vticai Chemistry 435 (1997) 85-94 ~9

around - 2 3 m V per tenlbld increase of c °, confirming in this case the existence of the associated form of t~,o cation-radicals. The value of the slope was also in agreement with the theoretical s i o f lbr the reversible |brmation of dimer: - 2.3RT/2 F = - 25 mV/Iog(c °) at - 17°C [24].

The equilibrium constant estimated for this silylated oligothiophene is in the expected] range for such compounds, compared to the value previously determined for oligothiophenes with the same length but with some other substituent~ [8-10,14,18,24,36], This indicates that the presence e r a bulk3 substituent does not change greatly the ~r-dimer~,. ~,m equilibrium and that this value depends more on the chain length than t~n the nature of the substituent, in agreement ~, ' h previously published results [24,36]. Therefore, as has already been concluded by other authors [16], the ~r-dimerizatm, process tbund for several oligothiophenes and MeaSi-4T-SiMe 3 seems to be rather universal.

3.3, Carbon-carbon Inmd,hmnathm

As discussed before, an important question concerns the nature of the carbon~-carbon I~md formation and two mechanisms can be considered [25,39,40], i.e. coupling between two cation,radicals (CR~CR) [47] or attack of one cationoradical on the starting oligomer (CR~S). We have Ibund that in the case of the electrochemical oxidation of quinquethiophenes, the |eaction is of the (CR~CR)qype [25]. However, it is importanl Io check the validity of thi.~ mechanism Ibr silyloligomers and to consider both possibilities in mechanistic aualy.,,i,..

C CR .... R:

( Me,St ~ n'l'~SiMe ~ )

or CR-S:

+ ( Me ~Si.~n'l'~SiMe ~ ) " -* (Me ~Si- nTooSiMc ~ )~'

(Me~Si-n'l'-SiMe~) ~ + Me~Si~nT-SiMe~

( Me :~Si-nT-SnMe ~ )

Rearomatization:

( Me~Si ~nTo.S~Me ~ )~

Dimer oxidation:

~ (Me~Si-nT-SiMe~),

+ (Me~Si-nl"-SiMe~)' -o (Me~Sn--nT-S~Me~)~ + Me~Si~T,~SiMe~

+ 2 e ~ Me,St--( nT),-SiMe~

(o)

(i)

(1')

(1")

{2)

Me~Si~(nT)~-SiMe~ + (Me~Si~nT--SiMe~) ' Z e ,,,* Me,St--( nT)~-SiMe~ + Me ~Si~n'l'-SiMe~ (3)

In these two possible mechanisms, the reaction in Eq. (2) after the coupling is the reammalization of the dimeric dicafion corresponding to the cleavage of the C~Si bond which is more likely to occur viii nucl¢ophilic ,'Jltack o~1 the silicon by the solvent itself than by direct elimination of an 'Me,St ~ " silyl cation [33,48,49]. The last step (the reaction in Fzt, (3)) corresponds to Ihe oxidation of the dimer which hasa lower potential than the starting oligomer [17~ 19].

The range of scan rates (experimental times) where the reversibility appears tier the oxidation of Me~Si-3T=SiMe~ and Me~Si~4T~SiMe~ is reasonably low and allows us to pertbma detailed mechanistic analy,~is of the ~r~dimerization and more particularly of the coupling step leading to the formation of the C=C bond. Unfortunately, such experiment~ are mor~ difficult to pertbrm with Me~Si~2T~SiMe~ due to the high reactivity of its cationot~dical and the fast lbrmation of a layer of polymer (or precipitation of higher oligomers) on the electrode surface. Based on our preliminary result~ of the exploratory cyclic voltammetric experiments, quantitative analysis o1' the tr-dimerization reaction mechanism has been done by considering the reversibility of the voltammogram. The principle of the method consists ia the determination of the ratio R between the peak currents corresponding to the oxidation of the oligothiophene during the Ibrward ,~,can !~., and the reduction of its cation-radical during the backward scan It,,. as a function of the different experimental parameler,~ (lhe potential scan rate, initial concentration of monomer, etc.). The theoretical variation~ can be calculated for the different postulated mechanisms, and the comparison ~tween the predicted and the experimental curves usually allow~ one m select the actual mechanism, in principle, this method of analysis is very similar to whal we have done in the ,~t~udy of line oxidation mechanism of pyrrole [40] or of quinquethiophenes [25] by double step potential chronoampcmmelry [471, :~nd tx~th techniques can provide the same kind of information on the chemical steps fb!lowing the initial e!ecmm trampler, However, in this case, we preferred to use cyclic voltammetry because the oxidized dimer~, of the.~e oligo,~ilyilbiopheoe~ (which are tbrmed during the voltammetry scan) and the starting oligomer have clo,~e redox poten|ial~ which m~|ke m~re difficult the choice of the step potential.

P. Hapit~# et al. /,hmrmd o]" Electrolmailvtical Chemisit3' 435 11097185-94

-4 -2 0 2

0.6 *

• S .~ - I

iotl~c"v '/lull I , ' V " sl

-4 I

0.1

0,4

li.Z

0

-~ 0 2

~b)

-? ..~ -3 -I llltlt¢'i '/til~ L" II s ' St

O,O

11,4

ll, l

li ot

• 4 ,1 U .4 . t 0

t1,11

ll, tt

0 ,4

li, l

o3 . ! I .S -.~ . I I

Iog le" l 71ol I . ' ¥ ' s ) I og t¢ " i " l i ~ l l Io' V ' s l

F ig ,! (Tl'tqt~: ~t l ( t , i l l l l l l l¢ l f~ Of Iht~ o l ida ihm ol oli~rtli~qf~¢n¢. Variation Ill r R g'tih the ¢ tpCt i lV I t ' l l l l i l pai+ i l t l l¢ l i ' l ' i°++ll (Io~./¢! ~I11¢1 |iti" Ih~' oli~o~il l l thiopl~il¢~: In) II,Ie~Si=JT:SiM¢~ in CI t :C I , c <'~ 4.2 × I 0 ~ i l l id l+ i ( i 1. 2.1 × I IY ! illelt i ' ( l l l l t . 1.0.~ × i l l i n u l l i ( i k (h i M¢ IS i - . I I ~ i M e i in i~¢toil l i f i lc, c< '~ , t .2x I l l ~niol l I l i l t , 2.1 × I t t ~nl~lll i ( t l t l , I.l;15~ I l l ~ n l ~ t t ( l l | , . I¢l M¢~Si-4T-SiMc~ in C l t & ' t : , ,.' ~1}.~ × i i t ~ m o l l i ( i l l ; (d) Mc~Si..-IT-SiM¢~ ill ~ll;.¢lonilfil¢, ~<' ','+ 11>.16 × I l l ~ i l l , i l l i I l L O.tGir, d l ' i ( i l l ; ~ 2{I~C ( ' l i l l l lP iUJtq l l l ~i lh the lh~o¢iltctil I~hl ivi~l l i l t the CRo('R ( } aml Ihi; CR :S lilei:hlllii'~ln I . . . . . ) (tippet ~ ~¢Me: h : t i ig l l~ l t .v D

The ratio R is nonllMit.ed wilh ~spccl to the value (#~,~ > i~.,),~,, it would have ill Ihe ilb~ence of I~ilhlw~up chemical re.l ions by ¢a!¢utating the quantity l ~ I #t,~ll,, )t( li,.tl,,, )dill IO show Ill Ill R pliS~S from unity fi~r a Iotally reversible pr~g-¢ss and iend~ to mn~ when the chemical i~a¢lion becomes faster and faster. The Illeoreiical Kinetic analysis of two Iimilin+ ~ h a n i s m s shows thai the electrochemical ~havior depends on only one single parilmeler A o~ k++."~{ RT/Fv) {see A ~ n d i s At. wher¢ ~, is II~ dilnerizalioil rate ¢onslarll. c <~ Ihe inUial coueemralion of oligolhiophene iiad v tile scau rate. A.~ i l~t lcd ~ t ~ i ~ [ ~ , 3 1 ~ , 4 0 L lh¢ :4aille va lue ~ff R is o b i l i i n e d ~ h c a the r a l i o ~'<~Iv is k c p i ~;~liisltuil i i i l i c ¢ o r d w i l h e i l l l c r Ibe

C R : C R o¢ the CR~S mechanisms: I ~ l ~ v e r , the var ia tkm of R with Ihis paranl¢ler ¢~'tI' is ditTereut for the two i l ' ~ ¢ ~ i s m s . T h e ottsi.-rvation of the e s~r in len ta l data and the craves #¢:~'°iv {Fig. 3) shows thai a g l ~ d agreement is f~m~ ~i ih ttl¢ Ihei~atieal curves thai eharacierim a C R - C R nlechanisnl ralher than a C R - S i . ng 'chanism, either for and in ACN ia~ ~¢11 as ili CH ~Cl ~,

M t v e r , h~¢ the lit of ihe cxpcriraenlal dala wiih the theoretical curves for Ih¢ (CR-CR) mechauism, the kinetic dil~ii~iili~wl rilt¢ ¢onslants of the deiemlining step can he ~iermined and die results are sununarized ill the Table I. Assuming ~ . , a ~ l h a r i i s m , the values for MejSi-2T=SiMes can he determined from the scan rate required to observe the ~:'~amlg-e of a reversible voltamlrmgram.

This kil~.¢lic analysis s~ws that the c a i n - c a r b o n bond formation occurs by coupling I~tween two cation-radicals, as f~w the Iirst steps of |ml).nleri~alion of pymde or substituted pyrroles [39.40]. the oxidative dimerization of SOlne

Tahte 1

~S~-_"I'~,.~Me~ ACN 5 x IO s CH:CI: 2 × I0 ~

~ S ~ 3 T - S~M~e~ ACN i.3 x IO s C H s C | ~ | . O x | 0 4

~ i T ~ S ~ M ¢ ~ AC tTq " 3.2 × IO: CH ,CI, 75

p. lt,¢,ot i,~ ~1. f J~uirlud fg" Electr¢lt~ludytical ¢ l len t i s to ~ 435 £ ! 997J 8 5 - 9 4 91

non-silylated oligothiapheaes or mixed pyrrole + thiophene oligomers [25], a~,~.d some general conclusions could be drawn about the nature of the mechanisms (CR-CR) involved in the coupling of thiophenes or pyrroles, in the two solvents, the dimerization rate constants for the silylated oligothiophenes are much lower than the value found for the unsubstituted oligothiophenes. For example, recent detemlinations made by photochemical [50] on" electrochemical [51] investigations gave values of k~ = (2-3) × IO s Imol ~ t s a in the case of the unsubstituted terthiophene. Also, the dilnefization kinetic rate constants for this series of silyloligothiophenes are strongly dependent on the nature of the solvent, and this effect becomes larger with the more reactive caLl|n-radical (near three orders of magnitude for Me~Si-2T-SiMe.O.

The first possible explanation is that different salvation modifies the kinetics at the level of the coupling step. The stabilization of an organic cation radical in dichioromethane, compared to other organic solvents like acetonitfile, is a well known phenomena (see fi~r example Ref. [52]) and the coulornbic r,".pulsion between the two cation-radicals during the coupling step can be lower in a more polar solvent like acetonitri!e. However, we did not observe this effect with unsubstituted aligothiophenes or with oligothiophenes bearing ~-substitaents other than a silyl group [34]. For examph:, the same experiments performed with oligothiophenes containing no silyl substituen| (Br-2T-Br, H~3T-OCtlr~, tt-4T~.H) show only .small solvent effects, of the artier of two or lower, on the lifetimes of the correslx~nding cation|radicals. So, the most surprising evidence is that this stabilization seems to be specific to the t~-silylated compounds.

The second pitssihle explanation is that the change of solvent nl,xiifies anouller reaction step in tile din neritation process, namely the rearamalizalian reaction. We should relllelnber |hal no silyl suhstituenls are found in the final palythiophene afi:er palymerizaiknl of silylierlhtophelrle and tl)at the desilylation reactian occurs during this proce.~s [33,34]~ This last retleliall iS specific Io these kinds ~,1' cl!mp~lnnd. We have proposed thiit the chemical reactivity of acelonitrile could he reSl~nlsibie for this phenomenon, most probably due Io a ilucleol~hilic assistance for Ihe ctcava~e of Ihe C~Si bond [34],

i,,Ik~;is in forms of a classical CR.. CR mechanism wllet'e lhc irrever~il, le Ttlerefare, it looks difficult to cxphiin Illcsc solvcnl , ' " " dinl~ri lal ion ~lwet~ll Ihn Iwll caiionorlidi¢iils is Ih¢ rule ilelernlillint~ <~lep ~!l" Ihe reaclion killelic~0 On Ilie c0onlrary, ii indicates thai the desilylalion reaction is hlvalved in the global kinetics of polymerization. Fol lowing tills hypothesis, we can first consider the possibility thai the relielion i l l desilylation occurs before the coupling step at the level ~ff the ealian-radical leading l | the l'omlaiion of a neutral radical, It is likely i:hai such a reaction, if ii exists, should be irreversible. resnlihl I hi a first-order decay for the cation-radical which is in total disagreenlent with the observed dependency of the R values on c° /c . This conclusion falls in line with the increase of selectivity of the coupling reaction thiil can be inllerpreled in lerms of ii sliibili#.alion of an hilerlnedial¢ silyl cationic species which iissuines thaL the silyl subsiiiueni is presenl on lhe inierniediale during the coaplhig step.

Therefore, a more likely possibility is l | consider Ihal the ciilion-radical=cation-riidical coupling .~lep (the reaction in l~q. " C l i b ) ) can he reversible and Ihai the reverse reaclion (k i ) has to be taken into a¢coi i l i t (( l - R ,,,, ). leversihle

dimerizaiions tire ofleri observed when the COllpling position of the inlernlediale is silbsliluled by bulky subsliluetli~, and several siniilar sillliilians have already hecn described (see for cxanlpl¢ ~tle oxidalive diln+,l'iZaliiln ill' 4.snbsliluled 2,fl-di-ler/-buiylphenols in btlsic nledia I.s,l-$f l or the I'edl~c;tivc dinleri#aiion of 4,/ertobiJlylp)'idininni M in Orp'l l l l t '

solveni). This silualion is likely ia occur with Me~Si which i~ at ii COillparablc ~i#,e lit tile tc, r/oi!li~,i .,,uil.,,iiiUClll. We could nlodify the previous 2R-(?R ineehanisin i | gel Ike CR =(711,~, nleehanisnl whieli is de: eribed by the fol lowin~

reaction scheme:

Me~Si-:onT=SiMe~ ~-~ (Me~Si-n'I'-SiMe~) ' + e (Ob)

(Me,Si -nT-SiMe, ) ' +(Me~Si=n'r~.SiMe,) ' ~ (Me;Si--nT-SiMe,)~' (Ib)

(Me~Si-.~nT..SiMe~)~ -:> Me~Si.-(,T).,~SiMe~ (2b)

Me~Si-( nT)2-SiMe~ + (Me~Si-nT-SiMe~) ' ~ Me,St-( nT)~-SiMe~ + Me~Si-~IT-SiMc< (3b)

Ill this mechanism, the dimeric cation is in fast equilibrium with the cation radical (Eq. ( Ib) ) and the ghtbal kinetics are directly dependent on the rate of carbon-silicon bond cleavage (Eq. (2b)). in this very particular case. it is p{~s.~ible IO show (see Appendix A) that the s:ime mathematical ext~ression is Ii~und whatever the reversibility of the coupling step and that the response obtained in cyclic wlltammetry depen~ls as before oil one kinetic paramelcr A = [k~kJ(k ~+ k:)]c°(RT/Ft'X meaning that the same time and concentration dependency is expect, d for reversible or irreversible dimerization. (Thi:~ i~ the same parameter as in the CR-CR,, case, wbere k~ i.~ replaced b) k~k~/(k a + k~).) Two limiting situa|b.~n~ can immediately be deduced from this analysis depending on the competi~i~}n between the reaction of di~,~ciati~m of (Me3Si-nT-SiMe.0~ + and its desilylation: if k, ~ k ~ the dimerization is irreversible and we fi~und the ¢las,~icai CR-CR,~, mechanism A = kic°(RT/Fc)(the kinetics do not low, lye the desilylation step); on the conlrary, if k ~ ~ k: the coupling reaction acts as a pre-equilibrium before the desilylation reaction, A becomes directly proportional to the rate

t~ 2 P. H~piut et aL / Journal ¢~f Ele~'tn~malyticui Chemistry 435 (1997) 85-~4

of the C-St bond cleavage A = Kk:c°(RT/Fv) (where K is the equilibrium constant for the tbrmation of the dimeric catty).

This mechanism is clearly in agreement with all the variations found for the variation of R with cO or ~,. Moreover, it can also explain the variations of the apparent dimerization rate constant with the solvent through the dependence of k~ even if the coupling reaction (Eq. ( I b)) is not influenced strongly.

Another indinect prcK~f for this mechanism is brought by the behavior of the silylated cation-radicals in the presence of fluoride ions. Becau~ of the well known reactivity of the fluorides towards silyl compounds to favor the desilylation reactions [53] considerable effects could be expected to occur on the polymerization reaction. The addition of 2 × IO "~ tool I ~ a of telraethylammonit|m fluoride to a solution of Me~Si-4T-SiMe~ in CH 2CI2 decreases the lifetime of the cation, radical by a factor around l0 4. Similar effects are ob~rved with Me~Si-3T-SiMe~, for which the cyclic voltamn~gram hecomes totally irreversible. However. no peaks due to the formation of the direct are now visible, i~icating a change of mechanism, which must involve ~mly direct attack of F ~ on the cation-radical before the reversible coupling s~po acctm~panied by ~he tiwmation of a neutral radical. Thi.~ indicates that, when there is I~m~ation of a neutral ~ iea t° thi~ d~¢~ not result i~ dimer (or I~dynr~zr) fi~rmation, and also that the dimcrizatiou of the cation-radical is a p~qui~ite to ~he l:~lymer formation.

The last remaining question i~ Whelhe~ Ih¢ intermediate dimeric cation invo!ved in [~|. (I b) is in fact the same '~dimer ' t~i~etv~ fiw most of ehe ~tahle catk~n~radical~ of oligothiophene. We have no cleat" answer to this question, and. unlbffunately, Ihe high ~ t i v i t y ot the cationoradicals (or the dimeric cation) from short oligomer~ impedes the study and the char~terization of a reversible coupling prior to the desilylation reaction, However, we can make some analogies with the re~ults lbund with <~r~'odisuhslituted oligothiophcnes: Ihey indicate that Ih~? ~dimerization equilibrium constant increa~s with the oligomer length, meaning that the ~r-dimer tbrm was unfavorable tbr bithioi~henes ~nd, on the contrary, the major ~pc¢ie~ were the ~dimer fi~m~ for the quaterlhiophene derivatives at r;~m tem~rature and fiw millimohu' concentratkms [8~ 10,13, I fi, 18,24.36]. Naturally, it must be kept in mind that the desilylation occurs on ¢r-dimers and of co~a~ not on ~dimcrs. in the framework of our analysis. However, it should not be excluded that there may be a continuous pa.~s~¢ Ih~m the ~-dimer to the a-dimer, and, therefi~re, that the Iormation kinetics o1' the erodimer are in fact clo~ly related to the ~dimer equilibrium. Reasoning this way in the ca~ of silyloligolhiophenes, means tha! the condition ko ~ :~ k, is more likely to be fulfilled lbr bithiopherp~s than fiw quaterlhiophcues and, therolbre, that the largest solvem effect is e x i t e d to bc less tbr sbort oligomers, Remarkably, Ihis is exactly what it is observed (see Table I ) when passing from Me~Si~2T-SiMe~ to Me~Si~4T=SiMo;,

4, Concl~km

In this work, we have shown thai the oxidative coupling of silyloligolhiophcnes in~.oives the carbon-carbon hond formation ~twccn two cationoradic~ls, as was found for other nou-silylated oligolhiophenes or the first steps of th~ polymcri~ation of pyrroles, itowever, tbr silyloligothk~p~ncs the iil'~fimes of Ihe cation~radicals ~re dependent on the ~lvent, which can b0 explain~ by a ~uclc~philic assistance to the C-St cleavage. This electrochemical behavior is in ~ n t ~ith a rcvc~ihle c~mp~h~,, step ie~ing to the t'(wmation of a din'~ric cation followed hy the irreversible cleavage ot" the C~$i bond~ Prior to the carbon~cadxm bond Ibmtation, when the cation-radical is stable enough, the existence of an ~uilibrium b e t ~ n the i ~ o m e r i c cation-radical and an associated form (~r~dimerizatiou) can be seen by the variation of the ~ x p~ential and the temperature,

Appendix A

The reversible CR-CR,~,,. mechanism can I~ represented by the fi)llowing scheme:

k~ 2 B ~ C

C ~ D

D + B ~ E + A

P. H~piot et aL I Joun;al oj Es~,rtr~malyth'al Chemistr)" 435 ¢ 1997~ 85- 94 93

Using the classical dimensionless scheme:

= - - + A~bd ~r 33 ,2

~b ~ b ~- - 2Aib 2 + 2A_ i t ' - A~bd

ar ~3,2

ac ~'~c = +h~b "~-h ~c-A2c'

ad it~ d = - - + h~c - A~bd

with the boundary conditions

,~aliablcs, the Ibliowing set of differential equations is obtained flora the chemical

The current ! flowing in the electf{.~de i~ given by ,/i ~ I# l /~ / ( I.S¢."D I/~) ,~ (i~a/;)),),, will; 7- .... I/RT/Io'r. where r i~ the scan rate. R is Ihe gas constant. F is Faraday's consiam. T is the absolute lemperature, and the dimensionless', eleclr{Rle

potential .~ ~ F I R T ( E ~ I:;~); y ~= or(DRTIFv) '/~, {linear diffusion, x is the distance fi'om the electrode. D is the diffusion coefficient, assumed to be the same for all intervening species): a = [ , T ] / ¢ °, b = [nT~ ]. d = [oxidized dimer]~ Aj = klc°RT'lirr, A I --"/"~ I R T / F r , A~ = k 2 RTIFv , A.~ = k3c°RT/Fr.

We assume that the deprolonalion and the oxidation reactions are faster than the dimerization reaction and that tile oxidation of Ihe dimer is monoelectronic. This last point has been discussed in detail elsewhere and shown to not modify lhe shape of the RooA curves [25]. It follows from the preceding that |he dimcr D and Ihe dimer l~efiwe rearomaliz,'gion C do not accumulate in the solution, i.e. their concemralions obey the steady slate assumption.

aa ~3"a A I A:

~1"~= it). ~ + A ~ + A , b "

ilb ;}~b A I A 2 ~ ;~r 2 3 A ~ +A~ b2

it follows that the electrochemical response for such an electrochemical mechanism depends on only one kinetic parameter A ~ [ k l k J ( k . . s+ k.~)]c°(RT/Fv). The set of equations has the same mathematical expression as the classical CR-CR, , mechanism with an irreversible dimerizaUon. This means that the same working curve R-A is obtained with the same dependence as a function of the initial concentration of oligomer and scan rate.

Two subcascs can be considered according to the rates of the backward dimerization k , compared |o the C~ Si cleavage reaction k~.

if k ~ i ~' k:: A '~ ( k l / k 1)k2c"( RT/Fc ) reversible dirnerization

if k I << k~: h ~ k j c ° ( R T / F r ) irreversible dimerization

References

[I] B. Xn, D. Fichou, G. Horowitz. 17, Gamier. J. Electroanal. Crhem. 267 (1989) 339. [2] D. Fichou. G. tlorowilz, F. Gamier, Syath. Met. 3{) (1990) 125. [3] D. Fiehou, G. Horowitz, B. Xu, F. Garnier, Synlh. Met. 39 (1990) 243. [4] H. Akilishi. K. Waragai. S. Hotta, H. Kano. H. Sakkaki. Appl. Phys. Lett. 58 (1991) 15(X). [5] J. Guay, A. Diaz, R. Wu. J.M. Tour. L.H. Dao. Chem. M'ger. 4 (1992) 254. [6] J. Guay. P. Kasai, A. Diaz, R. Wu, J.M. Tour. L.H. Dao, Chem. Mater. 4 (1992) lID7. [7] P. B[iuerle, Adv. Mater. 4 (1902) IO2. [8] M.G. Hill, K.R. Mann, L.L. Miller, J.-F. Penncau, B. Zinger. Chem. M~iter. 4 (1992) I 1(16. [9] M.G. Hill, K.R. Mann. L.L. Miller. B. Zinger, Chem. Mater. 4 11992) II 13.

P, Hapiol el aL / J+~urnul qf E/¢ctr+ultm/ytica/ ChemL+l+)' 435 ( 1¢)97J 85-!M

[iO] M+G Hill, K,R+ M~nn+ L,L. Miller, J.-F. Pcmlcau, J, Am. Chem. S(~. I i4 (11~2) 2728. [I I] A. Yil~,iiir. D. Dehil~mgli.~. M. Hmyci~. B+ Ne.s+sakh. G. Hor(Iwilz. F. Gamier Adv. Mater. 4 (1<<~21 4~). [12] A. Yit++~r. M. Hmyer~. F. Deloffre. R. Hajhioui. G. Horowilz. F. Gamier. Symh. Met. 54 (1993) 435. 113] P. A ~ . P. H ~ o l . J.-M. Pemaul. P. Garcia. J. Electroanal. Cllem. 361 (1993) 21.13. [14] P, ltthl~le, U. Segelbt~hcr, K.-U. Gaudl, D, Huttenlocher. M. Mehring. Angew. Chem. Int. Ed. Engl. 32 (1~-)3) 76. [15} U+ Segi~llmllcher. N.S+ Saricilici, A. Gmpp. P. Biiuerle. M, Mehring, Synlh. Mel. 57 (lOtJ3~ ,1728 [16] P. Bflu~le, U, Segethaehct. A Maier, M. Mehfing, J. Am. Chem. Soc. 115 (It)O3) IO21 +7 [l?] G+ Zlltti, G, Sehiavoll+ A. Berlin. G. Pagani. Chem. Maler..% (1993) 430. [1t4] G. Z~mi. G. Schiawln. A. Berlin. G+ P-'lgani+ Chem. Maler. 5 (I0¢13) f+2;). [19] P. Giircia+ J.+M+ Penmm, P. Hapiot. V, Winlgen~. P. Val;zlt. D, Delal~mgli~. F. Gamier. J, Phys. Chem. 97 (It>~3) 51.t. [20] C, Alemiln. E, Bnlla~. A,G, Davies, L+ Fajari, D. Gir6. L. Juli=i+ J.J. Pere#~, J. Rius+ J. Org. Chem. 58 (19q3) 3091. [21] G, 7+otti, G, Schiavon, A. Berlin. G. PaganL Synth. Mel+ 61 (1993) 81. [22] Y, Wei+ C,oC, Chan. J. T i~ . G.-W. Jang, K.F. ltsueh, Clam, Mater. 3 (19~} ~ +.~-: [23] J. Roilca]i. Chem. Re+ ++ 92 (1~2) 71 I, and ret+crences cited Ihcrein. [24] P, Aiitkd~fi+ P. Haph+ll0 K, M~mnier. J.+M. Pcrnaut, P. Garcia~ Chem. Mater. 6 (10941 1540, [ ~ P, A ~ + J+~M, Cillel+ G, [~ Coli~m~r+ V+ l ) l i~ae l , P, lhpi~t, J+ Pily~ ( 'hem ~) (Iti~IS) I I02] l ~ M> |ella+m+ W< Bli+hii+r, R+ Ga~mu, A,I+J+ lii+iynh, A. Guy+ J+ ll~mcali, J+ l'+i~+~'Irlmltal+ Chcm +14+ (i~i+~O+ ~<}~I 1241 J+ R,~tit++ A+ (+my+ M+ I+m+ire. R, Gam;aa, A,II, lhlynh0 J+ Elocimanal, Chenl+ M2 ( i + H ) ++77+ 13+~ S+l~+ lli++, R+E+ Nor+re, Ch+yIB Mm++. + ( 1~2 ) i472, [ ~ H+ Ma.+udm, Y+ Tailiki, K+ Ka+flyama, J, Polym, S¢i, Pall A; ill llOti2l If+f+7+ [~') H M+uda+ Y Tmfiki+ K Kaeriyamm, Synth, M++'I+ +~ ll~+OJit 1241++. 13| J+l~, Si~uv~joL C Ch~li~, J,~P~ I,,~¢++ l+ml,e, R+J,P+ Cofflu+ J,J+E M+~iva~i, P. i~li~i~il+ M. Wong Cili Man. Sy+llh+ MeL h2 (lqO4) 2]?+, 1412 IIJ.P. Cm~'lu. J.J.l+. Mot%~illi. P, Th+lm+l. M. Won~ ( 'hi Mail. C. CliliIT+t. J.+P. I+++P, me. I lL. S+nlV~!lill. ('lit+me Me+tier I~ { lDtl41 el<IlL [JJ M. ltou,ii¢llrln+. J.+l++ I.~+.re.l+tiiite+ JJ.E. Moi+e~lu+ M+ Wlll l~ Chi Man..I. Miller. ('lit+m. ~ ( l i~)+l +tD"P. [+1.t P+ Hmptot. I.+ (]al l lo,. P. /~.ml~.~n. lJ. l ! . Mili%.ail. J+l +. I++re.lii+lle. M. Won~ Chi M'm. Synlh+ Mel. ?~ (l l)O+l I}O. 134 i t NP+~+iikh. (i. thliowii.~. F. Gaillier. IL I~hlf l i 'e. P. Sriva+~tava. A. ii.l~+~i+lr..I+ I{li.i'lr~lanlll. Chenl+ Jltq ( 109.%i 97<

[++~1 cP, ~mi+ .~ , P+ H:+i,,+,. I++M+ S.++am, CIIt~nl. I cy . ~(t ( l i ~ i l 'P~J. l;+ll r+ A,, l+t+~. L. (+iUyal%l+ M> N+upen Dinh An. P. Hapiol. M. Challlia. C> COllll~lla~.+ A+ ]'hiehal+lh. J. t+le¢lr<i-uial. Clicni. 41ff I I~++l+l If~+!. I~+1 e+P. AndPi+lil+ P. Audel+Pl+ P+ tlapl.ol. J.-M. Sav~<aill+ J. Ai l l . Ch¢ill. Soc. 112 ( l+~ l l l 24;lq+ [~ol C.P. Aildrieux+ P. Aud+l~i l . P. I lapiol. I .+M Sav~<alll. J. Phy~+ Ch+ni. 94~ ( l i + l i I( l l : lg. tell M+ ~h l l t l i . H+ lies+~. G. Hartmke. W+ Berliner. K.W. Bla#.ey. K.A+ Mtiller. Silringer ~+fie++ il l Solid ~l~lll" ~l~iel|¢l+'~. Xpi+hlgc.r+ViPrlag. lie.rim. Vii i. 714.

19118, p, 377, [42 I)+ Gm~au, P, ltapiot, JM, b i v ~ l , J+ Ele~lr~mnal+ Cll~nl, 27? (1010}l I, [4:1 C.P. Amlrieux. D. GaPi~au. " tlal+iOi. J+ Pin+oB. J.*M. Sa~'+anl. J. I!k~:ln~'utalr (:helll.. 2.t.l l l91481 ~21 144 17). (hlrrellu. J.+M. Sav+anL j. +{N.¢u~i~nal. Ch¢lil> .1~ (1972) .liP+ 14+ il l~ Cl+ai+lk+ Math~lllliic'.~ Of DIIT~<i~+. O~l~ml Ihii+vr+ity Ih~.s~. I+lsltdOli. ll+j(++ 14f~ A. Ya++ai. ! + ̀G~miei'+ F. I~+1olt+ .. i ih~Pmili~.+ l . Ri+++td. Adx.. M~ll~+f. f+ 111~44) f r i l l I~L? C, An la l l~ , ILl+ (iltltYt~llll, M, I-lai~w~l. J ilin~lm. J.oM+ ~ili+'~ill'+l, ], I%,%+lroanal CIk'i+l I t4 l l98.~l I (.ttl II+LP, CimVi+, M+ l{+iill<++L+iml. ], tl+'+~ <+iilt.l Cltplll< 74 ( t'+74l i. l i t ) t I l . t +, ( 'oixhi. C. Gliettll. l . l . l ; . Mm-+a+i. t+i+..~i+it+ltc.l+ll+< I$ (I~i144i 4.1, [~ i V ViBl!+n~+ P. Vi~lal+ 1 +; Gat+~liei. I PI'I)+ . t ~i~.lll q~ ( l++- l l 2214. tsI P+ Aiid~l~lt. ],-M, Catet+ G lee C;~l~unic . ~. t.ltt+t~nei. P itapiol. Ill Ill+parlllioll. [$21 tt. L l l t l . ill: It. LiiP~l. ISIAI. I lat lo i ll+'~dl+l+ t++.l+< ili.t ̀ I~tc¢ii~+henli.~tl\v. Mabel I-~kkcr. Ni.m Yl+it.. +lrd +llititm. l~l~ll, p..102. 154'+ RJ,P+ C~Pim+ Ill.E+ I~ +" ' , ' - ++BI. ( . ( lePta. in: S +++it+++ Z Ilat+l~q~+++l lind,.). I"14 ('h+misir~+. of (hltanic of Sili¢i+il (+Olill~llllids. Wile). Nero York. 19811.

1~] 17LT. +.K+~4+++, LV+ ~hi+yaki, l% ++ (~m. , RPV. 87 t pellet I t l ] . l++.sl D-+H t++~+m+. P.I. llmen+P.. M J. K+~lly. 1. I~ l~ t i~ im l ('bent. If~] (11184) 14~. i~+1 P+ I t l l o t . J. Pmmm, J+ EK~llV~lilal. China..tit? 1 It)t!3) 2+?, I+?] D, Tl+lw~', ;1+ Ik~n'i~+lin, E, Kh'l+w+i+oEisn~++', J ElectPtmllal, Clk, m, I+llt t 1982l 1,11,