Electrochemistry with alkyl-linked oligothiophenes

Electrochimica Acta 56 (2011) 3419–3428

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Electrochemistry with alkyl-linked oligothiophenes Markus Pagels a,b,1 , Günther Götz c , Thomas Fischer c , Peter Bäuerle c , Jürgen Heinze a,b,∗ a

Institute for Physical Chemistry, Albert-Ludwigs-University Freiburg, Albertstr. 21, 79104 Freiburg, Germany Freiburg Materials Research Center, Albert-Ludwigs-University Freiburg, Stefan-Meyer Strasse 21, Germany c Institute for Organic Chemistry II and Advanced Materials, University of Ulm, Albert-Einstein-Allee 11, 89081 Ulm, Germany b

a r t i c l e

i n f o

Article history: Received 4 August 2010 Received in revised form 11 December 2010 Accepted 19 December 2010 Available online 28 December 2010 Keywords: Oligothiophenes Electropolymerization Cyclic voltammetry Conductivity Conducting polymers

a b s t r a c t Two pairs of ␤,␤ -didodecyl substituted quinquethiophenes linked via a tri- (3) and tetramethylene spacer (4) at their ␣-positions have been synthesized and their electrochemical behavior investigated with cyclic voltammetry in solution and in solid films. Both compounds can be charged to a tetracation in two twoelectron transfer steps, which are reversible at room temperature but become partially irreversible at low temperatures. Concentration and scan rate dependent measurements support an intramolecular coupling of the oligothiophene units preferentially next to their bridging site with ␴-bond formation. In solution no electropolymerization can be observed. In the solid state upon oxidation intermolecular coupling of the oligothiophene segments takes place. Conductance measurements confirm the voltammetric findings. Obviously, the mechanism of conductance is based on hopping processes within mixed valence states. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction In contrast to alkyl substituted thiophenes, electrochemical studies of alkyl linked thiophenes are rare. For the first time, ␣,␻-bis(thienyl-)alkanes have been used only as pure linkers for polythiophene in order to improve the mechanical properties of this conducting polymer [1,2]. In stress–strain-experiments a behavior similar to vulcanized elastomers has been found and it has been concluded that the polythiophene chains are indeed cross-linked via alkyl groups. Four-probe resistance measurements with dry films, however, have shown decreasing conductivity with increasing cross linking. This has been explained with the argument that cross-linked thiophene chains are prevented from attaining a coplanar configuration which affects conductivity negatively. Further publications about this family of substances mainly describe the synthesis of monomers [3,4], their electropolymerization and the electrochemical properties of the resulting polymers [5,6]. As in all these experiments only simple ␣,␻-bis(thienyl-)alkanes with the coupling site at the 3-position of the thiophene units have been studied, their reactivity has been very high and electrochemical

∗ Corresponding author at: Institute for Physical Chemistry, Albert-LudwigsUniversity Freiburg, Albertstr. 21, 79104 Freiburg, Germany. Tel.: +49 (0)761 203 6202; fax: +49 (0)761 203 6237. E-mail address: [email protected] (J. Heinze). 1 Current address: Schlumberger Technology Corporation, 1935 S Fremont Dr., Salt Lake City, UT 84104, USA. 0013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2010.12.057

oxidation of the monomers has always resulted in the formation of typical polymers. Recently, a series of dimeric quinquethiophenes linked at the 2-position with a di- to a hexamethylene spacer were synthesized and examined as a ␲-dimer model of polythiophene [7]. The authors claim that these compounds upon chemical two-electron oxidation in methylene chloride readily form intramolecular ␲-dimer species, except for the dimethylenelinked dimer that cannot be bent into a ␲-stacked structure. In a subsequent paper, similar results are described for a cyclophanetype of dimeric quinquethiophenes [8]. In order to get further insights into the reactivity of alkyl linked thiophenes we have studied the electrochemistry of two newly synthesized bis(quinquethienyl-)alkanes. Surprisingly, in solution during anodic oxidation of the monomers intramolecular ␴-coupling steps dominate in comparison to intermolecular coupling. As model compounds oligothiophenes 1,3-bis(4 ,3 -didodecyl3 and 2,2 :5 ,2 :5 ,2 :5 ,2 -quinquethien-2-yl)propane 1,4-bis(4 ,3 -didodecyl-2,2 :5 ,2 :5 ,2 :5 ’,2 -quinquethien2-yl) butane 4, in which two quinquethiophene units were linked either by a propyl or a butyl bridge, were synthesized by nickel-catalyzed coupling of ␣-bromoquaterthiophene 1 and the Grignard reagent of 1,3-bis(5-bromothien-2-yl) propane 2a or 1,4-bis(5-bromothien-2-yl) butane 2b, respectively (Scheme 1) [9]. The oligothiophenes are forced to a certain proximity by the alkyl bridges, which should facilitate inter- and intramolecular bond formation. However, the alkyl groups are also spacers, which length has to be considered. The spacer acts as a barrier for mobile

3420

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

H25C12 S

S S C12H25

S

+

Br

Ni(dppp)Cl2 BrMg

(CH2)n

S

1

C12H25 S

S S

S

2a (n=3), 2b (n=4) H25C12

S

MgBr

S

S

(CH2)n

S S

C12H25

S S H25C12

3 (n=3), 4 (n=4) Scheme 1. Synthesis of 1,3-bis(4 ,3 -didodecyl-2,2 :5 ,2 :5 ,2  :5  ,2 -quinquethien-2-yl)propan 3 and 1,4-bis(4 ,3 -didodecyl-2,2 :5 ,2 :5 ,2  :5  ,2 -quinquethien-2yl)butane 4.

charge carriers and limits the size of the conjugated domain within the chain-like molecule. The longest possible conjugated length in the used model substances 3 and 4 comprises 10 thiophene units. The occurring conductivity has to derive from hopping processes between these conjugated domains, which are forced to a spatial proximity by covalent cross links between different chains.

from 17.0 g (25.5 mmol) 3,3 -didodecyl-2,2 :5 ,2 :5 ,2 quaterthiophene dissolved in 1 L DMF at 40 ◦ C and 4.6 g (25.9 mmol) N-bromosuccinimide in 50 mL DMF. Purification on a column of silica (32–63 ␮m) with n-hexane followed by recrystallization from the same solvent yielded a yellow solid in 11.0 g (58%). M.p. 56–57 ◦ C; elemental analysis: C40 H57 BrS4 (Mw 746.03), calc.(%) C 64.40, H 7.70, S 17.19, found C 64.68, H 7.87, S 17.22.

2. Experimental 2.1. Chemicals 3,3 -Didodecyl-2,2 :5 ,2 :5 ,2 -quaterthiophene, 1,3-Di(2thienyl)-propane and 1,4-Di(2-thienyl)-butane were synthesized according to literature [3,4,9]. Acetonitrile was bought from Fisher Scientific in HPLC/FarUV quality and used without further purification. The supporting electrolyte tetrabutylammonium hexafluorophosphate (TBAPF6 ) was used as bought from Fluka in electrochemical grade quality and stored in an oven at 80 ◦ C. All experiments were carried out under an argon atmosphere. 2.2. Characterization NMR spectra were recorded on a Bruker ACF 250 spectrometer and AC 250, resp. at 25 ◦ C [1 H NMR: 250 MHz, 13 C NMR: 63 MHz]. (Further details: see Supporting Information) Mass spectra were measured on a Kratos Kompact MALDI 3 V4.0.0 (ETH Zürich). UV/VIS spectra were recorded at a PerkinElmer Lambda 19 spectrometer. Melting points were determined at an Electrothermal Model 9100 and are not corrected. 2.3. Synthesis 2.3.1. General procedure for the synthesis of the bromo-oligothiophenes 1 and 2 In an inert atmosphere and under exclusion of light a solution of N-bromosuccinimide in DMF was added slowly to a solution of the oligothiophene in the same solvent. The mixture was stirred at room temperature for 15 h. Water, twice the volume of DMF in total, was added and the organic layer extracted three times with dichloromethane. The combined organic phases were washed with saturated NaHSO4 and water and dried over MgSO4 . After evaporation of the solvent the crude product was purified by column chromatography. 2.3.1.1. 5-Bromo-3,3 -didodecyl-2,2 :5 ,2 :5 ,2 -quaterthiophene (1). According to the general bromination procedure starting

2.3.1.2. 1,3-Di(5-brom-2-thienyl)-propane (2a ). According to the general bromination procedure starting from 7.0 g (33.6 mmol) 1,3-di(2-thienyl)-propane in 7 mL and 12.1 g (68 mmol) Nbromosuccinimide in 60 mL DMF. The crude product was filtered over silica with n-hexane and afforded 12.0 g (98%) of colourless oil. elemental analysis: C11 H10 Br2 S2 (Mw 366.13), calc.(%) C 36.09, H 2.75, S 17.51, found C 36.36, H 2.50, S 17.39. 2.3.1.3. 1,4-Di(5-brom-2-thienyl)-butane (2b ). According to the general bromination procedure starting from 7.0 g (31.5 mmol) 1,4-di(2-thienyl)-butane in 7 mL and 11.4 g (64 mmol) Nbromosuccinimide in 60 mL DMF. The crude product was filtered over silica with n-hexane and afforded 11.4 g (95%) of colourless, waxy product. M.p. 28 ◦ C; elemental analysis: C12 H12 Br2 S2 (Mw 380.15), calc.(%) C 37.91, H 3.18, Br 42.04, S 16.87, found: decomposition. 2.3.2. General procedure for the Ni-catalyzed cross-coupling to the oligothiophenes to 3 and 4 In a dried apparatus purged with argon the magnesium, ␣,␻bis(5-bromo-2-thienyl)alkane and diethyl ether were mixed and the reaction started with 2–5 droplets of 1,2-dibromoethane. After 2 h of ultrasonication a brown oil has formed, which was heated under reflux for further 2 h. The reagent was cooled to room temperature. 10 mg of NidpppCl2 was added followed by a solution of bromoquaterthiophene 1 in ether in one portion under reflux conditions. Reflux conditions were maintained for 6 days while the residual catalyst was added after 3 days. The reaction mixture was hydrolyzed with 2 mL 2N HCl under cooling and extracted with dichloromethane. Combined organic layers were washed with saturated NaHSO4 and water and dried over MgSO4 . The residue after evaporation of the solvent was purified by column chromatography on silica with a cyclohexane–dichloromethane gradient. A second chromatography on a preparative HPLCcolumn filled with nitrophenyl modified silica (5 ␮m, ID = 20 mm, L = 200 mm) and n-hexane-dichloromethane as eluent furnished a pure product, which was recrystallized from n-hexane.

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

3421

2.3.2.1. 1,3-Bis(4 ,3 -didodecyl-2,2 :5 ,2 :5 ,2 :5 ,2 quinquethien-2-yl)-propane (3). According to the general coupling procedure starting from 366 mg (1 mmol) 2a , 48.6 mg (2 mmol) magnesium, 3 mL diethyl ether. After formation of the Grignard reagent 14 mg (25.8 ␮mol) NidpppCl2 and 2.0 g (2.68 mmol) 1 dissolved in 10 mL ether were added. Work-up and purification afforded 0.41 g (27%) of product as an orange solid. M.p. 64–70 ◦ C; elemental analysis: C91 H124 S10 (Mw 1538.58), calc.(%) C 71.04, H 8.12, S 20.84, found C 70.90, H 8.31, S 20.94; MALDI-TOF MS: m/e = 1537.8 (M+ ); UV–vis (CH2 Cl2 ): max. (log ε) = 407 nm (6.8). 2.3.2.2. 1,4-Bis(4 ,3 -didodecyl-2,2 :5 ,2 :5 ,2 :5 ,2 quinquethien-2-yl)-butane (4). According to the general coupling procedure starting from 380 mg (1 mmol) 2b , 48.6 mg (2 mmol) magnesium, 3 mL diethyl ether. After formation of the Grignard reagent 14 mg (25.8 ␮mol) NidpppCl2 and 2.0 g (2.68 mmol) 1 dissolved in 10 mL ether were added. After work-up and purification 0.63 g (41%) of product was obtained as an orange solid. M.p. 83.5–86.5 ◦ C; elemental analysis: C92 H126 S10 (Mw 1552.61), calc.(%) C 71.17, H 8.18, S 20.65, found C 71.39, H 8.28, S 20.37; MALDI-TOF MS: m/e = 1551.9 (M+ ); UV–vis (CH2 Cl2 ): max (log ε) = 407 nm (7.3). 2.4. Voltammetric measurements A Pt disc (diameter 1 mm) sealed in a soft glass rod was used as the working electrode; it was polished with diamond polishing paste and then rinsed thoroughly with ethanol and acetone. Pt and Ag wires were used as counter and quasi reference electrodes, respectively. Potentials were measured using a ferrocene/ferrocinium redox couple as internal standard. All data were rescaled versus the saturated Ag/AgCl electrode (0.35 V versus Ag/AgCl). A Jaissle potentiostat PG10 with the IPS software package ELCM was used for electrochemical control and data recording. All measurements were carried out in methylene chloride or acetonitrile with 0.1 M TBAPF6 as supporting electrolyte at different san rates, temperatures and concentrations. 2.5. In situ conductivity In situ conductivity measurements [10] were carried out on a microarray working electrode (5 ␮m gap). The working electrode was separated from the potentiostat by two 1 k resistors. A bias of E = 10 mV was applied to the microarray electrode and a third 1 k resistor at which the potential was measured. The conductivities of the polymers on the microarray electrode were calculated according to the ohmic rules. This setup was controlled by an AMEL 553 potentiostat, and the potential scans were performed with an EG&G/PAR model 175 scan generator. Normally, the conductance was measured within a potentiodynamic experiment at a scan rate of 0.02 V s−1 . 3. Results and discussion 3.1. Electrochemistry of 3 in solution The voltammetric measurements of the monomeric starting compound 3 produced unexpected results. The cyclic voltammogram of a 1.5 × 10−4 M solution of 3, which was measured in methylene chloride/0.1 M TBAPF6 with a scan rate of v = 0.1 V s−1 and at a temperature of 20 ◦ C, shows two well-separated oxidation peaks at potentials of 0.81 V and 1.08 V, respectively (Fig. 1a). In the reverse scan two corresponding reduction peaks at potentials of 0.77 V and 1.04 V, respectively, can be seen. The peak separation of the anodic and cathodic waves Ep for the first redox step is 45 mV and for the second 44 mV, the peak current ratio of both redox steps

Fig. 1. Influence of scan rate and temperature on the cyclic voltammograms of 3 in CH2 Cl2 , c = 1.5 × 10−4 M, (a) T = 20 ◦ C, (b) T = −20 ◦ C.

amounts to 1. Apparently, two independent reversible redox processes at formal potentials of E10 = 0.791 V and E20 = 1.062 V occur in the system. The first wave represents a two-electron transfer in which the second electron is transferred at a potential that is identical or slightly positive to the first, hence the peak separation is smaller than 57 mV. (If the difference between the standard potentials of two redox reactions is less than 100 mV, only a single wave appears in a voltammogram. According to the theory of cyclic voltammetry [11], the Ep separation indicates inter alia the difference between the redox potentials of a two-electron transfer reaction. Ep values smaller than 57 prove that the formal separation of the standard potentials is less than 35 mV. Exact values can be found in the literature. This means, each oligothiophene chain is oxidized to a monocation, the whole molecule to a dication. As the peak current for the second oxidation wave is the same as for the first one, the second oxidation process also reflects a two-electron transfer. Therefore, after the second oxidation peak the molecule is definitely a tetracation. Similar results were obtained at higher scan rates (Fig. 1a) and at higher concentrations (5 × 10−4 M). The experimental data are surprising. Normally, oligothiophenes up to hexathiophene with at least one unsubstituted ␣-position undergo follow-up coupling reactions [12] at the radical cation level. Here, even the tetracation (dication of a single

3422

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

3

-2e 0

E1

3 2+ 0 E2 -2e

3 4+

kf kb

2+ 3cycl 0

E2´ -2e 4+ 3cycl

Scheme 3. Reaction scheme of 3 or 4 during anodic oxidation; 3cycl characterizes the product of intramolecular cyclization.

Fig. 2. Influence of scan rate and temperature on the first redox process in the cyclic voltammograms of 3 in CH2 Cl2 , c = 1.5 × 10−4 M, (a) T = 20 ◦ C, (b) T = −20 ◦ C.

thiophene chain) is stable. In addition, the energetic separations between the redox states of the mono- and dication, and the tri- and tetracation, respectively, are very small (E0 < 0.035 V) [13] indicating that there are, in contrast to expectations, low but attractive and stabilizing forces between the charged thiophene chains. When the cyclic voltammograms are recorded at low temperatures (−20 ◦ C) the peak potentials of the first redox step are shifted about 100 mV to lower potentials at all scan rates, while the peak potentials of the second redox step correspond to those at room temperature (Fig. 1b). The formal redox potentials amount to E10 = 0.703 and E20 = 1.065. Furthermore, both redox processes become chemically irreversible at low temperatures and slow scan rates, which means that a chemical follow-up reaction after the dication formation takes place. Details of the follow-up reaction are presented in the following paragraph. Cyclic voltammograms of 3 involving only measurements of the first redox step up to the dication level are shown in Fig. 2. The upper curves (Fig. 2a) measured at 20 ◦ C show a reversible process; the ratio of the peak heights of the reduction and oxidation waves amounts to 1 at all scan rates. At a temperature of −20 ◦ C the peak current ratio is significantly smaller than 1 for scan rates between 5 and 0.1 V s−1 . In the reverse sweep two partially merged waves at peak potentials of 0.59 and 0.64 V can be seen. The latter one disappears almost completely at a scan rate of 0.1 V s−1 (Fig. 2b) whereas the first wave melts into the second at a scan rate of 10 V s−1 . Clearly the reduction wave at 0.64 V belongs to the

reduction of the dication; the wave at 0.59 V indicates the reduction of a follow-up product. At higher concentrations (5 × 10−4 M) the current in the respective voltammograms increases but the shape of the voltammetric response does not change. This gives unambiguous evidence that the chemical follow-up process is a first-order reaction. In the case of a second-order process, the shape of the voltammograms in the reverse sweep should definitely change but this has never been observed. The fact that this process becomes “visible” at low temperatures is typical for (reversible) dimerization reactions of radical ion species [14,15]. The reason is that especially the thermodynamic properties significantly change at low temperatures whereas the kinetic parameters are scarcely affected. Thus, owing to negative reaction enthalpies and the slipping influence of reaction entropy, the equilibrium of radical ions shifts towards dimers. Additionally, the static dielectric constant ε0 of solvents in the liquid state increases [16], which diminishes repulsive forces between identically charged species. As shown in several preceding papers [14,15,17–21] the coupling step of radical ions results in the formation of covalent ␴-bonds, even in the case of blocked reaction centers. The high negative reaction enthalpies with values between −40 and −90 kJ mol−1 as observed for many of these “dimerization” reactions exclude weak ␲-interactions [14,15]. A kinetic theory describing the coupling rate between identically charged entities as a function of the solvent polarity has been developed by Debye [22]. In principle, there exist two possibilities for a dimerization. A radical cation of 3 may couple intermolecularly at the outer ␣position of a thiophene chain with a second radical cation or it may react intramolecularly at the inner ␣-position with the second thiophene segment of the same molecule. At a first glance, an intramolecular coupling seems rather unlikely because of steric reasons but experimental results are in favor of such an assumption. The first-order kinetics excludes an intermolecular dimerization of radical ions. Moreover, in voltammetric multisweep experiments the intermolecular formation of longer thiophene chains involving additional redox steps has never been observed. Generally, it has been shown that in the case of intermolecular coupling steps oligomeric thiophene chains (n = 2–8 units) form longer redoxactive species the lowest oxidation potentials of which possess smaller values than those of the starting oligomers [12]. Consequently, only an intramolecular coupling reaction satisfies the condition of a first-order reaction. A realistic pathway is the coupling between both the cationic thiophene chains of 3 at the inner ␣,␣ -positions which leads to the formation of a stress-free cyclopentyl ring and an intramolecular ␴-bond between the thiophene chains (Scheme 2). In contrast to normal thiophene coupling intermediates where proton elimination produces elongated neutral thiophene chains, the intramolecularly generated dicationic ␴-intermediates are stable within a dynamic equilibrium between the closed cyclopentyl and the open propyl structure. The starting molecule is regenerated after the reduction of the system. A mechanistic scheme which involves all the redox steps and the chemical follow-up reactions is shown in Scheme 3. It should be noted that the equilibrium between the closed and the open structure also exists at room temperature though connected with a smaller equi-

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

H

H

S +

H

H H

H

H

+

S H

H H

H H

H S

H

H

S H

S

S H

3423

H

H

H

H

H

S

S

S

S

H H

H

H

H H Scheme 2. Possible product of an intramolecular ␴-coupling of 3 at the inner ␣,␣ -positions of the thiophene chains (shown without the dodecyl chains for clarity).

librium constant [14]. A consequence of this behavior is that a significant fraction of molecules (10–20%) forms a closed structure even at room temperature. Of course, the question arises why the ␴-bond formation occurs at the inner substituted ␣-position of the thiophene chain but not at the outer ␣-position which is not sterically hindered. The answer is simple; the accessible concentration of 3 in solution is low (<10−4 M), a small second-order rate constant as given for a quinquethiophene prevents measurable dimerization rates. On the other site, the intramolecular distance between the two thiophene chains in 3 is extremely small which formally corresponds to a very high concentration of oligothiophene moieties and favors coupling reactions even under the condition of a small rate constant. A qualitative estimation of the rate constant at −20 ◦ C with the aid of working curves gives a value between 2 and 8 s−1 [23]. In order to quantify the voltammetric measurements digital simulations (Digisim® ) of 3 were carried out at different scan rates always using the same set of basic parameters. A very good agreement between experimental and simulated curves was achieved for a rate constant k of 6 s−1 , which shows that the intramolecular coupling step is a slow reaction. Using digital simulations (Digisim® ) and working curves, an estimation of the equilibrium constant K of the intramolecular coupling yielded an approximate value of 102 at −20 ◦ C. It lowers drastically at higher temperatures. The second redox process of 3 at a potential of E2 0 = 1.065 V leading to a tetracation is apparently independent of the previous coupling reaction, as the currents of the two anodic waves for the di- and tetracation formation change in the same manner at decreasing temperatures (Fig. 1b). Obviously, the electronic structures of the “cyclic” and the “linear” tetracation are almost identical and, therefore, a perfect additive superimposition of the respective redox components takes place. This finding again excludes a theoretically possible coupling of the thiophene chains at the outer ␣-positions because a chain with 10 conjugated thiophene units would shift both redox steps to significantly lower potentials which is not the case.

3.2. Film characterization Compound 3 could not be precipitated onto an electrode by electrochemical polymerization because of the good solubility in dichloromethane and the tendency for intramolecular coupling rather than intermolecular network formation. In order to generate a thin solid film of 3 on an electrode, a liquid layer of the monomer was deposited by dip-coating from a CH2 Cl2 solution on a microelectrode array, and after drying in vacuo it was studied electrochemically in an acetonitrile/TBAPF6 (0.1 M) electrolyte system in which 3 was insoluble. Fig. 3 shows a multisweep voltammogram of a film of 3 at room temperature at a scan rate of 0.05 V s−1 . The intensity of the signal decreases with every cycle, which is probably caused by the par-

Fig. 3. Solid state multisweep voltammogram of 3 in ACN/TBAPF6 (0.1 M); v = 50 mV s−1 , T = 20 ◦ C.

3424

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

tial dissolution of oxidized monomer at room temperature. From the 9th cycle on, the film seems to be stable. In the first cycle at a potential of about 0.9 V the current signal starts to rise strongly. The peak potential of this oxidation lies at 1.02 V. A second oxidation peak appears at a potential of 1.26 V. In between these two peaks, the signal does not decrease but forms a plateau. This additional current signal in the range between 1.0 and 1.2 V appears in all solid state voltammograms. As it does not influence the charging/discharging balance in any of these measurements we assume that an electroactive impurity such as spurious water in the system causes this effect. After reversing the potential scan there is a broad reduction wave at a potential of 1.07 V and a reduction peak at a potential of 0.7 V with a shoulder at 0.79 V. The first anodic scan up to a potential of 1.1 V represents the oxidation of thiophene segments of the monomer. The molecules are charged to the dication state at which only intramolecular coupling steps occur. Up to the switching potential of 1.3 V tetracations and intermolecular coupling products are generated. As the concentration of the monomers is very high compared to the experiments in solution, coupling reactions between neighboring molecules certainly take place and partially insoluble oligomers with decathiophene chains are formed. At 20 ◦ C an essential fraction of the starting molecules is dissolved into solution during repeated scans. There is a strong memory effect between the first and the second cycle (Fig. 3). In the second cycle an oxidation peak appears now at a potential of 0.86 V and a second one at a potential of 1.21 V. The reduction waves remain unchanged regarding potentials and current values. The nature of the memory effect is still unclear. Different models have been proposed to explain the phenomena [24,25]. As explained below we assume that processes at the film/electrode interface are causal for this effect. Subsequently to the memory effect in additional voltammetric scans, the current values continuously decline. Finally from the ninth cycle on, the voltammetric response remains stable for many scans. Fig. 4 shows the 10th cycle in which a new reversible redox process can be seen with an oxidation peak at a potential of 0.70 V and a reduction peak at a potential of 0.63 V indicating a redox potential of 0.665 V. As this redox potential

Fig. 4. Solid state voltammogram of 3 in ACN/TBAPF6 (0.1 M), v = 50 mV s−1 , T = 20 ◦ C; 10th cycle from Fig. 3.

is more negative than that in solution, these data must represent charging and discharging of a larger conjugated system than the quinquethiophene chain in 3. It is obvious that this redox process belongs to the oxidation and reduction of decithiophene segments which have been formed by slow irreversible coupling of two quinquethiophene chains at the respective unsubstituted ␣-positions involving the formation of a ␴-bond and immediate twofold proton elimination. Nevertheless, the original redox waves of the quinquethiophene dications do not disappear which indicates that the coupling reactions between the starting molecules preferably end with a simple dimerization. Notably, a high anodic current in the range between 1 and 1.2 V can be still observed after the 10th voltammetric cycle. As shape and intensity of the voltammetric response do not change during additional sweeps

Fig. 5. Cyclic voltammogram (full line) and in situ conductance (dotted line) of a film of 3 in ACN/TBAPF6 (0.1 M), v = 20 mV s−1 , T = 20 ◦ C.

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

3425

this anodic signal must result from an unknown impurity in solution. 3.3. In situ conductance of 3 In situ conductance measurements of a film of 3 within a potentiodynamic experiment are shown in Fig. 5. As can be seen the conductance starts to rise after the first voltammetric oxidation peak at 0.82 V and reaches a maximum just before the second oxidation peak at 1.1 V. Then a very steep decrease of conductance occurs in the range between 1.1 and 1.2 V. In the reverse scan conductance increases again and reaches a maximum at a potential of 1.01 V. This maximum is significantly higher than in the forward scan. After passing a small shoulder the conductance signal reaches the detection limit at a potential of 0.70 V. The conductance values reflect electrochemical and chemical processes in the film occurring during charging and discharging. The first redox step generates a dication involving a partial intramolecular coupling between the two thiophene segments of the monomer, which even at room temperature results in the formation of a ␴-bond and the generation of localized charges (Scheme 2). That means that in dependency on the coupling equilibrium constant the number of mobile charge carriers is significantly less than for undisturbed redox states and no visible conductance appears in this potential range during charging (Fig. 5). By contrast, the conductance curve in the range of the tri- and tetracation formation reveals that the number of mobile charge carriers is considerably higher than those for the dicationic redox state. However, it is surprising that at the end of the redox charging the conductance rapidly decreases. Generally, the conduction mechanism in conjugated chain-like systems is described by the bipolaron model, which is based on a band model [26]. The bipolaron model explains conductance in conjugated polymers with the transport of first locally stabilized polarons followed by the transport of bipolarons and not of electrons. According to these predictions, the conductance should persist at the tetracation level. Very recently, Heinze and co-workers presented data which support the view that conductance or more generally conductivity in polymers containing hexathiophene segments [27] or bis(arylamine) systems [28] results from hopping processes within mixed valence states. Using this model, charge transport in these electroactive materials takes place via intermolecular electron hopping events between isoenergetic redox centers. The model which is in perfect agreement with the approach of Chidsey and Murray [29] developed for redox polymers predicts maximum conductivity when exactly half of the redox states at the formal redox potential E0 are charged, because the number of mobile charge carriers hopping between these isoenergetic states approaches its highest value. As the tri- and tetracation redox states of 3 are almost degenerated the conductivity reaches its maximum at the trication level. At the tetracation level all sites of the redox state become empty and consequently conductivity drops down. In the reverse scan, when electrons as charge carriers are injected, hopping processes start again and conductance rises. A very interesting point is that the maximum value of conductance is higher in the reverse scan than in the forward scan. A plausible explanation may be that at the beginning of oxidation film compartments are charged which are directly connected with the platinum digits of the array whereas areas in between the digits are only scarcely charged and, therefore, a charge gradient involving a high resistance is formed. In the reverse scan the accumulated charges are uniformly distributed within the film which leads to a lower resistance and consequently to a higher conductance of the film, which can be observed up to the end of the discharging process. When within a voltammetric multisweep experiment the switching potential is gradually increased, the transformation from

Fig. 6. Multicycle in situ experiment with 3 in ACN/TBAPF6 (0.1 M); v = 20 mV s−1 , T = 20 ◦ C; (a) cyclic voltammogram; (b) conductance.

an oligomeric film to a crosslinked polymer can be observed (Fig. 6). During the first cycles the waves of the two redox processes in which the alkyl bridged quinquethiophenes are oxidized and reduced can be seen. With gradual rise of the switching potential the current values of these oxidation and reduction peaks decrease until there is only a plateau left which is typical for conducting polymers [12,25]. At the beginning of these additional electrochemical charging/discharging processes a small quantity of the material may be dissolved into solution. However, the increasing irreversibility of the redox processes at the di- and tetracation level unambiguously demonstrates that charges are consumed for additional coupling steps between the oligomers. At a first sight surprisingly, the conductance in the range of 0.8–1.2 V increases within succeeding voltammetric cycles from 3.5 × 10−4 to 6 × 10−4 S at rising switching potentials. Using the hopping/mixed valence model (HMV), it can be excluded that the number of mobile charge carriers significantly increases during successive electropolymerization steps. However, it is very reliable that the density of mobile electrons grows due to the cross-linking of the chains. The cross-linking may diminish

3426

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

H H

H H

H H

S

H

H

H +

S

H

H

S

H

H

+

H

S

H

H H H

S H

S

S

H H

H

H

S

S

H

H H

S

H

H H

H Scheme 4. Possible product of an intramolecular ␴-coupling of 4 (shown without the dodecyl chains for clarity).

the volume of the film thus generating shorter hopping distances. 3.4. Cyclic voltammetry with 4 in solution The prolongation of the alkyl spacer in compound 3 by one CH2 group, which leads to homologue 4, causes small changes of the electrochemical properties. On first sight, as expected, there is no significant difference between the cyclic voltammograms of 3 and 4 in solution. The redox potentials are slightly shifted to lower values. Again, reversible two electron oxidations are observed at redox potentials of 0.54 and 0.81 V. The small peak separation of 50 mV indicates that the redox states are almost degenerated. When the experiments are carried out at a low temperature of −20 ◦ C, the first pair of redox waves similar to compound 3 shifts into negative direction but both redox steps up to the tetracation formation remain reversible for scan rates higher than 0.1 V s−1 . But when the temperature is lowered to the solubility limit – at the given concentration (7 × 10−5 M) 4 starts to precipitate at −24 ◦ C – low scan rate voltammetry indicates a follow-up reaction (Fig. 7a). Measurements at −30 ◦ C in a supersaturated solution show in the reverse sweep of the first redox step two closely separated reduction waves one of which at the lower potential belongs to a follow-up product (Fig. 7b). As with 3 we assume that the follow-up reaction is a slow intramolecular reversible ␴-dimerization during which a cyclohexyl ring is formed (Scheme 4). Unfortunately, at this low temperature no systematic experiments could be performed as 4 slowly precipitated. Therefore, the concentration of the monomer at −30 ◦ C is rather low. In order to distinguish between interand intramolecular ␴-bond formation, again voltammetric experiments with a higher monomer concentration were performed. But at a monomer concentration of 1.8 × 10−4 mol/L there are no changes in the voltammetric response at lower temperatures com-

pared to the response at room temperature. Thus, in the case of 4 in solution also an intramolecular bond formation takes place. The follow-up reaction of 4 appears at a lower temperature than that of 3. This indicates that the equilibrium constant of the reversible intramolecular coupling reaction is smaller than in the case of 3. 3.5. Solid state voltammetry with 4 In order to perform electrochemical solid state experiments with 4, a thin film was generated by dip-coating from a methylene chloride solution containing the monomer 4. Fig. 8 shows the cyclic voltammogram of such a film in acetonitrile with TBAPF6 (0.1 M) as supporting electrolyte at a temperature of −10 ◦ C and a scan rate of 0.01 V s−1 . At higher temperatures the film slowly dissolves. In the first cycle there is a sharp oxidation peak with a peak potential of about 1.03 V. In the back scan three reduction waves are observed at potentials of about 1.11 V, 0.93 V and 0.64 V. In the second cycle a new oxidation wave appears at a peak potential of 0.85 V and two smaller oxidation waves at potentials of 1.0 V and 1.13 V which are in the range of the sharp peak of the first cycle. Additional redox processes which proves the formation of intermolecular conjugated coupling products are not visible. Similarly to compound 3, the data reflect a typical memory effect that normally arises in polymeric films after an extended relaxation time [25]. In the very first cycle immediately after the dip-coating step, the film is oxidized within one sharp wave up to its tetracation, in the reverse sweep the system is totally reduced involving three separated redox steps, and a neutral film is regenerated. In the subsequent cycles the oxidation starts with the formation of the dication followed by separated tri- and tetracation redox steps, the voltammetric response during discharging is almost identical with that of the first cycle. However, after a waiting time of 48 h in

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

3427

Fig. 8. Solid state voltammograms of 4 in ACN/0.1 M TBAPF6 involving a memory effect after deposition (full line) and after 48 h relaxation (dotted line); v = 10 mV s−1 , T = −10 ◦ C; 1: first cycle, 2: second cycle.

supports the view that here also a surface effect steers the rate of the heterogeneous electron transfer. It is improbable that deposition in an organic solvent generates a passivating layer. However, it is very reliable that the “contact” between an electroactive film and a metal electrode improves during charge transfer processes. The correct alignment of redox molecules in a film to the electrode surface is an important prerequisite for a fast electron transfer. At the beginning of a voltammetric experiment the charge transfer resistance in a disordered film may be high; it decreases during charging because molecules at the electrode surface get an optimal position for a heterogeneous electron transfer. 4. Conclusions

Fig. 7. (a) Cyclic voltammograms of 4 in CH2 Cl2 at different scan rates; T = −24 ◦ C, c = 7 × 10−5 mol/L; (b) first redox process with 4 in CH2 Cl2 at T = −30 ◦ C, v = 0.1 V s−1 .

the discharged state the subsequent voltammogram looks like the very first one. In contrast to other examples, the memory effect in this film is very strong involving a shift of the first oxidation peak of about 200 mV. Because the material consists of relatively short oligomers explanations for this effect, which are based on the properties of polymers [30–32], are not applicable. Due to the strong influence of the scan rate on the magnitude of the memory effect, Odin and Nechtschein have suggested that the heterogeneous rate constant at the interface should be time-dependent [33]. Very recently, Heinze and Rasche reported a strong memory effect in solution [34]. According to this hypothesis, after discharging of the film a passivating layer of the “solvent” may be formed at the electrode surface, which hinders the electron transfer between electrode and electroactive species (film). This passivation reaction may be based on a polymerization of the solvent or the adsorption of solvent molecules at the surface. During electrooxidation of the redox active compound, this passivating layer is rapidly destroyed so that the electron transfer rate accelerates to a normal rate. The fact that a freshly deposited film of 4 exhibits a strong memory effect during the initial charging/discharging cycles, which returns within 48 h,

The electrochemistry of alkyl-linked quinquethiophenes 3 and 4 has been studied. Both compounds are rather stable upon electrooxidation and do not undergo intermolecular coupling processes at low concentrations. However, measurements at low temperatures reveal that intramolecular coupling processes occur leading to a cyclopentyl or cyclohexyl structure with a ␴-bond between both the quinquethiophene segments. The equilibrium between the closed and open form shifts to the closed form at decreasing temperatures, whereby the equilibrium constant is higher for 3 than for 4. In situ conductance measurements at thin films show that the closed form which is generated at the level of the dication formation has a significant lower conductance than the open form and succeeding redox states due to the fact that the ␴-bond between the thiophene oligomers produces sp3 -centers and, therefore, strongly localized charges. The experimental results support the view that charge transport in these oligomers is based on hopping processes within mixed valence states. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.electacta.2010.12.057. References [1] K. Kaeriyama, S. Satira, H. Masuda, Makromol. Chem., Rapid Commun. 11 (1990) 37. [2] Q. Pei, O. Inganäs, Synth. Met. 55–57 (1993) 3724.

3428 [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

M. Pagels et al. / Electrochimica Acta 56 (2011) 3419–3428

G. Koßmehl, F.D. Hoppe, B. Hirsch, Z. Naturforsch. 48b (1993) 826. K.J. Hoffmann, E.J. Samuelsen, P.H.J. Carlsen, Synth. Met. 113 (2000) 161. K.M. Mangold, K. Morgenschweis, K. Jüttner, Electrochim. Acta 44 (1999) 1865. K.M. Mangold, K. Jüttner, Synth. Met. 101 (1999) 71. T. Satou, T. Sakai, T. Kaikawa, K. Takimiya, T. Otsubo, Y. Aso, Org. Lett. 6 (2004) 998. T. Sakai, T. Satou, T. Kaikawa, K. Takimiya, T. Otsubo, Y. Aso, J. Am. Chem. Soc. 127 (2005) 8082. T. Fischer, Dissertation, 1998, Universität Würzburg. A. Smie, A. Synowczyk, J. Heinze, R. Alle, P. Tschuncky, G. Götz, P. Bäuerle, J. Electroanal. Chem. 452 (1998) 87. J. Heinze, Angew. Chem., Int. Ed. 23 (1984) 831. K. Meerholz, J. Heinze, Electrochim. Acta 41 (1996) 1839. J.B. Flanagan, S. Margel, A.J. Bard, F.C. Anson, J. Am. Chem. Soc. 100 (1978) 4248. A. Smie, J. Heinze, Angew. Chem. Int. Ed. 36 (1997) 363. P. Tschuncky, J. Heinze, A. Smie, G. Engelmann, G. Kossmehl, J. Electroanal. Chem. 433 (1997) 223. R.F. Grant, D.W. Davidson, P. Gray, J. Chem. Phys. 33 (1960) 1713. P. Huebler, J. Heinze, Ber. Bunsen-Ges. 102 (1998) 1506.

[18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34]

H. El-Desoky, J. Heinze, M.M. Ghoneim, Electrochem. Commun. 3 (2001) 697. J. Heinze, C. Willmann, P. Bäuerle, Angew. Chem., Int. Ed. 40 (2001) 2861. J. Heinze, A. Rasche, Electrochem. Commun. 5 (2003) 776. A. Rasche, J. Heinze, Electrochim. Acta 53 (2008) 3812. P. Debye, Trans. Electrochem. Soc. 82 (1942) 265. R.S. Nicholson, I. Shain, Anal. Chem. 36 (1964) 706. B. Villeret, M. Nechtschein, Phys. Rev. Lett. 63 (1989) 1285. J. Heinze, B. Frontana-Uribe, S. Ludwigs, Chem. Rev. 110 (2010) 4724. J.L. Brédas, G.B. Street, Acc. Chem. Res. 18 (1985) 308. H. John, R. Bauer, P. Espindola, P. Sonar, J. Heinze, K. Müllen, Angew. Chem. Int. Ed. 44 (2005) 2447. O. Yurchenko, J. Heinze, S. Ludwigs, ChemPhysChem 11 (2010) 1637. C.E.D. Chidsey, R. Murray, J. Phys. Chem. 90 (1986) 1479. G. Inzelt, Electrochim. Acta 34 (1989) 83. G. Zotti, G. Schiavon, Synth. Met. 31 (1989) 347. T.F. Otero, I. Boyano, J. Phys. Chem. B 107 (2003) 6730. C. Odin, M. Nechtschein, Phys. Rev. Lett. 67 (1991) 1114. J. Heinze, A. Rasche, J. Solid State Electrochem. 10 (2006) 48.