Electrochemical Study of Charge Transfer in Conducting Polymers

Aspectos cinético e termodinâmico da transferência de carga em polímeros condutores foram investigados. Um modelo de difusão restrita foi impregado para interpretar os da dos obtidos por cronopotenciometria com pulso. Em filmes de poli 3-metiltiofeno, o coeficiente de difusão aparente diminuiu em função da espessura do filme mas ficou constante com a concentração eletrolítica. Em filmes finos, usando-se voltametria cíclica, caracterizou-se mudanças termodinâmicas, como em par tic u lar, a deconvolução das duas transferências eletrônicas envolvidas no processo. Usando-se um modelo nernstiano incluindo potenciais de interação, conseguiu-se descrever a reação para as duas etapas eletroquímicas. Finalmente, como esperado, uma correlação en tre o raio iônico dos dopantes e o coeficiente de difusão aparente foi estabelecida em filmes de polipirrol.


In tro duc tion
Con ducting poly mers be long to a class of quite ver satile 1 ma te ri als of in ter est be cause of their pros pect in techno log i cal ap pli ca tions 2 . The in su la tor/con duc tor tran si tion plays a fun da men tal role which mo ti vates the study of the re dox re ac tion in these ma te ri als (charge trans fer and ac cumu la tion), in or der to de ter mine pa ram e ters such as dop ing level, re vers ibil ity, or ki netic fac tors.
The dif fi culty in un der stand ing the prop er ties of these ma te ri als co mes from the fact that the work is usu ally carried out on poly mers with vari able struc tures, us ing dif ferent char ac ter iza tion tech niques, some of which are not very suit able for these pur poses. It is gen er ally rec og nized that in all kinds of con duct ing poly mers the oxido-reduction reac tion is equiv a lent to a two-step elec tro chem i cal pro cess that does not fol low the Nernst model. It is com monly at-trib uted to high intersite in ter ac tions and other non-ideality fac tors 3 . On the other hand, it is also known that a sim ple ki netic model of dif fu sion does not ex actly fit the dop ing pro cess of the electroactive films. This is mainly be cause the poly mer suf fers a large phase tran si tion dur ing the re action (con for ma tion, elec tri cal con duc tiv ity) which moves the reactional in ter face from the elec trode/poly mer re gion to the poly mer/elec tro lyte one, bring ing about in a sig nif icant change in the dimensionality of the dif fu sion 4 .
This work aims to in ves ti gate the ther mo dy namic and ki netic as pects of the re dox re ac tion of con duct ing polymers, vary ing the struc ture and the dop ing con di tions and choos ing suit able meth ods of anal y sis. Polypyrrole (PPy) and poly 3-methyl (P-3MeT) were cho sen as the ma te ri als to be stud ied in aque ous and or ganic me dia, re spec tively. The poly mer iza tion was per formed in dif fer ent ways, namely, galvanostatically, potentiostatically, and potentiodynamically. Films of var i ous thick nesses were syn the sized, and the con cen tra tion and the na ture of the elec tro lyte salt were changed.
Low speed cy clic voltammetry was used as a tool for de ter min ing the for mal po ten tial, the elec tro chem i cal doping level, and the reversibilty grade of the re ac tion. In ter action po ten tials were cal cu lated in some spe cific cases us ing an in ter ac tion re dox model 5 . An ap par ent dif fu sion co ef ficient (Dap) was de duced from higher speed ex per i ments, plot ting i vs. v 1/2 . The mor pho log i cal as pect of the films was con trolled by elec tronic mi cros copy, and the dop ing level was also eval u ated by microprobe ex per i ments. A small am pli tude method, pulse chronopotentiometry, was peformed on re duced films to mea sure the ap par ent dif fusion co ef fi cient with out dis turb ing the struc ture of the poly mer. A re stricted lin ear dif fu sion model was used to fit the po ten tial-time plots.

Elec tro chem i cal mea sure ments
All elec tro chem i cal ex per i ments were car ried out at room tem per a ture un der ar gon at mo sphere us ing equipment from Autolab/Ecochemie. A one com part ment Metrohm-type elec tro chem i cal cell was used with a sat urated cal o mel elec trode (SCE) as a ref er ence (equipped with a sa line bridge for or ganic me dia), a plat i num wire as an aux il iary elec trode, and a plat i num work ing elec trode with an ap par ent sur face of 8 x 10 -3 cm 2 . A con cen tra tion of 0.1 mol/L was used in stan dard ex per i ments for both monomer and elec tro lytes. For or ganic so lu tions, af ter prep a ration, the so lu tion was dried with mo lec u lar sieves (3 Å) and fil tered be fore use. Be fore each ex per i ment, the work ing elec trode was pol ished with neu tral alu mina and washed with dis tilled wa ter. Electrosynthesis of PPy and P3MeT films was per formed galvanostatically us ing a cur rent density of 1 mA/cm 2 .The PPy films were pre pared in aque ous so lu tions con tain ing so dium dodecylsulfate (DSNa) as the elec tro lyte, and P3MeTh in acetonitrile so lu tions con taining LiClO 4 as elec tro lyte. Af ter electrosynthesis, a po tential of 0 V was ap plied to the PPy film and 0.7 V to the P3MeTh film un til the cur rent de creased be low 1µA; then the poly mer was rinsed sev eral times with wa ter (PPy) or acetonitrile (P3MeT) and dipped in the work ing elec trolyte. With re gard to the com pu ta tional as pects, sim u la tions were per formed on an IBM RISC/6000, model 3CT, for kinetic data, and on a 486 DX2 PC for ther mo dy namic data.

Re sults and Dis cus sions
Mono mer re ac tiv ity The voltammetric ox i da tion of some mono mers is presented in Fig. 1. As ex pected, the re ac tion is ir re vers ible at this po ten tial sweep ing speed; the an odic dis place ment of the peak po ten tial from the terthiophene to the thiophene is re lated to the length of the con ju ga tion, which is in versely pro por tional to the ion iza tion po ten tial 6 . For ex per i men tal con ve nience, pyrrole and 3-Me thiophene were cho sen for a study in aque ous and or ganic me dia, re spec tively.

Poly mer iza tion modes
For the same poly mer iza tion speed, the potentiostatic and galvanostatic modes are roughly equiv a lent. The lat ter was sys tem at i cally used in this study in or der to achieve better con trol of the charge den sity. Fur ther more, small cur rent den si ties (less than 1 mA cm -2 ) were pre ferred, since ear lier stud ies have shown that these mild free monomer dif fu sion con di tions pro vide denser and more reg u lar films 7 . Pulsed cur rent poly mer iza tion was also ap plied, but the re sults are not pre sented here. Actually, as this mode of syn the sis un doubt edly leads to more com pact and ad her ent films 7 , it was found that the re vers ibil ity of the re dox process of the films de creased and the thick ness could not be sim ply re lated to the poly mer iza tion charge 8 .

Re dox prop er ties of the poly mer films
The ther mo dy namic as pect As pointed out in the In tro duc tion, these data come from low speed cy clic voltammetry where dif fu sion is assumed to not in ter fere. The re dox pro cess is as sim i lated to two monoelectronic trans fers be tween three spe cies, neutral (PP), cat ion rad i cal (PP + ), and dication (PP ++ ), each one chem i cally equiv a lent to six mono mer moi eties.
The mod i fied Nernst equa tion is 5 : where the in ter ac tion po ten tials have a clas sic lin ear depend ence on con cen tra tion: where C T = C(PP) + C(PP + ) + C(PP ++ ). The equa tion system was solved us ing New ton's al go rithm, and the program for gen er at ing sim u lated voltammograms was writ ten in Ba sic/Win dows. Some typ i cal re sults are shown in Fig. 5.
The ki netic as pect Data from pulse-chronopotentiometry may be, in princi ple, more re li able than oth ers from large am pli tude methods (for ex am ple dou ble-step chronoamperometry). This oc curs ow ing to the fact that the pulse is ap plied on a Vol. 8, No. 2, 1997 Charge Trans fer in Con ducting Poly mers 177  potentiostatically equil i brated film in its re duced state. This means that the poly mer prop er ties does not change sig nif i cantly dur ing the in jec tion of a small quan tity of charge and be haves as a re dox poly mer 9 in its re duced state. In this case, it can be as sumed that the ki net ics of change trans fer is of the diffusional type and that it is con trolled by ionic mo tion rather than self-exchange elec trons 10 . Another im por tant point is that the mea sure ment is per formed in an open cir cuit, so that it can be as sumed that no mi gration con tri bu tion is in cluded in the ionic mo bil ity. The method con sists of in ject ing a small cur rent pulse into the re duced film in or der to cre ate some ox i dized sites, and then to fol low the open cir cuit po ten tial (E oc ) as a function of time. The tran sient con cen tra tion at the elec trode (C(0,t))-time can be cal cu lated from fi nite lin ear dif fu sion equa tions. The ini tial pro file, C(x,t), is cal cu lated as sum ing a pla nar source of dif fus ing spe cies, since the in ter face is lo cated at the elec trode 11 where τ is the pulse du ra tion, i p the cur rent den sity of the pulse, n the num ber of elec trons (1e / PP), and F the far a day con stant. Af ter ter mi na tion of the cur rent pulse, the dis tri bu tion of the dif fu sion spe cies at the elec trode sur face as a function of time is given by 11 : where L is the thick ness of the film, and the first term is the fi nal uni form con cen tra tion given by Ip τ / FL. These equations were solved by stan dard nu mer i cal in te gra tions, using x = 0 for the last one. The pro gram for gen er at ing sim u lated chronopotentiograms was writ ten in FOR TRAN 77. On av er age, the pa ram e ters used were 1000 in ter vals for the dis tance, 150 terms for the sum ma tion, and 200 time in cre ments. The pulse and the cur rent pa ram e ters are known, and the thick ness, L, is eval u ated from the poly mer iza tion charge, Q p , as 12 : P3MeT: L(cm) = 2 x 10 -4 .Qp (C cm -2 ) PPy: L(cm) = 4 x 10 -4 .Qp (C cm -2 ) The open cir cuit po ten tial Eoc-time ex per i men tal transients were fit ted us ing an em pir i cal po tential-concentration re la tion de rived from low speed cy clic voltammograms as fol lows: the ini tial an odic cur rent curve was in te grated in a 200 mV range (Qa), the ca pac i tive contri bu tion (QC) was substracted as sum ing 11 an av er age double layer ca pac i tance (C d ) of 35 µF cm -2 , and the spe cies con cen tra tion C was de duced from the faradic re sult ing change (Q F ): where v is the po ten tial sweep ing speed and E F -E i is equal to 200mV.
Finally, the ap par ent dif fu sion co ef fi cient, D ap , was obtained by match ing ex per i men tal and sim u lated tran sients (see, for ex am ple, Fig. 6).

Doping level eval u a tion
First, it is im por tant to know that this de ter mi na tion is del i cate be cause it can be re lated to the ionic dop ant con tent or to the cyclable dop ant con cen tra tion, which is not nec essar ily the same quan tity. In ad di tion, the lat ter in cludes a low trapped ionic frac tion (ca pac i tive) and a deeper trapped one (faradic) 13 . Two dis tinct at tempts were performed in or der to ap proach this prob lem: an elec tro chem ical method (in situ) in te grat ing the low speed voltammogram, and an other one us ing an elec tronic microprobe (ex situ) to dose Cs, S, and Cl atom con tents in Cs-doped films. The lat ter method in di cated a dop ant concen tra tion vary ing from 30% in the ox i dized state (film equil i brated at 0.8 V) to 20% in the re duced state (-0.8 V), while the elec tro chem i cal charge var ied in the 20% range. This re sult shows the dif fi culty in iden ti fy ing the ac tual elec tro chem i cal spe cies, but sug gests the ex is tence of two ionic dop ants be sides the im mo bi lized cat ion in the re dox pro cess. More over, it ap pears that only a frac tion of the dop ant con tent is in volved elec tro chem i cally, as pre viously pub lished 14 .

The PPy case
Voltammetric be hav ior is sen si tive to the na ture of the dop ant 13 . It has al ready been dem on strated that for large size an ionic dop ants, the first pro cess is as so ci ated with the cat ion mo tion 14 . For the al ka line se ries used in this work, it seems that the cat ion af fects the ther mo dy nam ics of the reac tion (see Fig. 2), pos si bly due to a con tri bu tion to site inter ac tions 15 , as the na ture of the dop ant ap pears to have lit tle ef fect on the elec tronic struc ture, as re ported elsewhere 16 .
The gen eral equa tion of the re dox re ac tion can be written as fol lows 17  where PP is the ac tive site of PPy as de fined be fore, DS is the tensio-active an ion (dodecylsulfate), M the al ka line cat ion, and A the an ion of the cy cling elec tro lyte (gen er ally Cl -). The ki net ics of the first dop ing pro cess was in ves tigated by pulse chronopotentiometry and treated as seen in a pre vi ous sec tion, as sum ing that the rate of the charge transport is con trolled by cat ion mo tion 10 . As ex pected, the larger the hydratation ra dius of the ion, the lower / the diffu sion co ef fi cient. This fea ture is ex em pli fied in Fig. 3 for two dif fer ent cat ions of the al ka line se ries. Suprisingly, low val ues of the dif fu sion co ef fi cient were ob tained as com pared to the lit er a ture val ues, but it is im por tant to remem ber that we char ac ter ized the diffusional pro cess in the re duced poly mer, which is not gen er ally the case when large am pli tude meth ods are used. We char ac ter ized a distinct phe nom ena. Cal cu lating D ap in the NaCl case from the slope of i pa -v 1/2 (voltammograms be tween 50 and 500 mV/s), a value of 1.3 x 10 -10 cm 2 s -1 was ef fec tively deduced.

The P3MeT case
It has been sug gested that the elec tro lyte con cen tra tion may have an in flu ence on the re dox prop er ties of the conduct ing poly mers 12 . As can be seen in Fig. 6, the an odic poten tial changed sig nif i cant from 0.65 V at 0.1 mol/L to 0.79 Vol. 8, No. 2, 1997 Charge Trans fer in Con ducting Poly mers 179 V at 6 x 10 -3 mol/L (note that the cath odic coun ter part moves too, so that it is not a con duc tiv ity ef fect). Ki netic mea sure ments were also per formed to de tect an even tual con cen tra tion ef fect. How ever, ex cept for the dis place ment of the equi lib rium po ten tial, prob a bly due to the de crease of the elec tro lyte con duc tiv ity, no sig nif i cant change in the dif fu sion co ef fi cient was de tected by chronopotentiometric anal y sis (Dap re mained in the range of 10 -12 to 10 -11 cm 2 s -1 ). This sug gests that the re sid ual ionic con tent of the film would have lit tle ef fect on the ionic mo bil ity of the dop ant.
The thick ness de pend ence of the prop er ties was in ves tigated, since it is a con tro ver sial sub ject 1 8 . The voltammetric re sults cleary showed the ex is tence of two redox steps in thin films. Fitting these data us ing an in ter action re dox model 5 , for mal po ten tials were found to be 0.25 and 0.65 V for the cat ion and dication, re spec tively, us ing only low in ter ac tion po ten tials (10-50 mV) (Fig. 5). This is prob a bly due to the low mo lec u lar weight of the poly mer, which pro vides more ideal re dox be hav ior. It was found that the ap par ent dif fu sion co ef fi cient de creases ini tially with the thick ness, and then sta bi lizes above a cer tain value. Ac cord ing to Mar tin et al. 11 , this could re sult from the densification of the poly mer/elec trode in ter face due to late re act ing mono mer. This point is not in con tra dic tion with the in crease of po ros ity of the thicker films, since in this kind of ex per i ment only the lim it ing step near the electrode is de ter min ing.

Con clu sion
The re dox prop er ties of some con duct ing poly mers were in ves ti gated, and PPy and P3-MeT were cho sen as exam ples for this study. A spe cial small am pli tude method was used to char ac ter ize the ki netic lim i ta tion of the re dox pro cess, while cy clic voltammetry al lowed the in ves ti gation of the ther mo dy namic changes. It was con cluded that un der our ex per i men tal con di tions the ionic dif fu sion co effi cient de creases with the thick ness of the poly mer film and it is slightly in flu enced by the elec tro lyte con cen tra tion. Fo cusing on the cationic dop ing of PPy, it was pos si ble to ob serve a change in the dif fu sion speed as a func tion of the ionic ra dius of the dop ant. Work is in prog ress on the ox idized form of the poly mer, which be haves elec tro chem ically as a po rous semi con duct ing elec trode.