J. Phys. Chem. B 2008, 112, 6331–6337
6331
In situ Resonance Raman Spectroscopy of Polyazulene on Aluminum Anna Österholm, Beatriz Meana-Esteban, Carita Kvarnström,* and Ari Ivaska Process Chemistry Centre, c/o Laboratory of Analytical Chemistry, Åbo Akademi UniVersity, FI-20500 Turku/Åbo, Finland ReceiVed: August 6, 2007; ReVised Manuscript ReceiVed: February 25, 2008; In Final Form: February 26, 2008
Polyazulene (PAz) has been electrochemically deposited on different electrode substrates. The films were characterized with Raman and UV–vis spectroscopy. The spectroelectrochemical studies were performed in situ during p- and n-doping (electrochemical oxidation and reduction, respectively). The focus of this work was mainly on the charging and discharging reactions of PAz on Al substrates. The results were compared to the corresponding results obtained from PAz on Pt substrates. Three different excitation wavelengths (514, 633, and 780 nm) were used in the Raman experiments and the resonance enhancement effect was observed when changing the wavelength of the excitation line. The vibrational behavior of PAz deposited on Al was very similar to that of PAz deposited on Pt during p-doping. Furthermore, it was found that the vibrational responses during p- and n-doping are different indicating that the electronic structure of PAz is not the same during positive and negative charging. It was concluded that PAz is not reversibly n-doped on Al. The n-doping on Pt was shown to be more reversible. In this paper, the important correlation between UV–vis and Raman spectroscopy is discussed as well as the correlation between doping-induced infrared active bands and Raman bands of neutral PAz. 1. Introduction Raman spectroscopy is a useful tool for obtaining information regarding the structural and electronic properties of polyconjugated materials. It is well known that conjugated polymers can easily be made conductive by chemical or electrochemical doping. The conductivity of conjugated polymers is commonly discussed in terms of solitons, polarons, bipolarons,1,2 and more recently also polaron pairs.3–6 Because the conjugated backbone of the polymer is strongly coupled with the delocalized π-electrons, a modification of the electron distribution, due to introduction of charge carriers, will also cause a modification in the geometry of the conjugated backbone.7 These changes in the electronic structure can easily be monitored in situ with Raman spectroscopy during electrochemical doping. By varying the energy of the excitation line, it is possible due to the resonance effect to activate different segments of the polymer that give rise to optical absorptions with energies that match those of the excitation lines used. Because the introduction of charge carriers will create new electronic levels in the band gap there will also be changes in the UV–vis-NIR spectra of the polymer. Generally, these changes are seen as bleaching of the band arising from the so-called π-π* transition and appearance of new bands due to the introduction and presence of different charge carriers in the polymer film. By studying the in situ UV–vis spectrum, it is possible to anticipate what part of the polymer is resonantly activated with a specific excitation wavelength.8 In the vibrational spectra of polyconjugated molecules there are features related to the existence of delocalized π-electrons. These features can be observed in both Raman and infrared spectra of conjugated polymers. Many theories have been formulated to explain these spectroscopic features. For this reason Zerbi and co-workers developed the so-called effective * Corresponding author: Tel.: +358-2-215-4419. Fax: +358-2-215-4479. E-mail:
[email protected].
conjugation coordinate (ECC) model7,9,10 by reformulating the amplitude mode (AM) theory introduced by Horovitz et al.11,12 The ECC model correlates the doping-induced infrared absorption bands to totally symmetric Raman bands by introducing an internal conjugation coordinate (Я) that describes the collective geometrical changes that take place in a polymer when its structure is transformed from the ground state (benzenoid) to the first excited state (quinoid). According to the ECC model, the bands that will appear in a Raman spectrum of a pristine polymer will arise from totally symmetric vibrations. These vibrational modes are silent in the infrared spectrum of the pristine polymer. Introduction of polarons, bipolarons, or polaron pairs into the polymer backbone will cause a break in the symmetry, and the chain will become polarized causing these vibrational modes to become infrared active. Thus, Raman bands arising from doped segments of a polymer film should be weak or unobservable unless resonance enhancement conditions are reached. Although the ECC model and the AM theory can predict the doping-induced bands observed in the infrared spectrum, they do not allow a determination of the type of charge carriers present in the polymer. Furukawa et al.13,14 introduced the charged oligomer approach for identifying the charge carriers on the basis of Raman and UV–vis spectra by taking advantage of the resonance enhancement effect and comparing the spectra of polymers with the spectra of corresponding oligomers. Because polarons and bipolarons are merely radical ions and divalent ions, single or double charged oligomers should give a similar electronic spectrum as doped polymers assuming that the charged oligomers have a similar geometry as the polaron or bipolarons localized on a polymer chain. According to this approach, it should be possible by in situ Raman spectroscopy to distinguish between the types of charge carriers introduced during doping. In this work we have studied the p- and n-doping of polyazulene (PAz) using in situ resonance Raman spectroscopy. PAz is a member of the category of fused-ring-based conjugated
10.1021/jp0762828 CCC: $40.75 2008 American Chemical Society Published on Web 04/29/2008
6332 J. Phys. Chem. B, Vol. 112, No. 20, 2008 polymers. These types of materials have gained a great deal of attention due to their many potential applications and interesting optical and electrical properties. It is expected that the fusedring structure has a reducing effect on the band gap.15 PAz has also shown relatively high conductivities that are most likely due to its planar molecular geometry giving rise to an extended π-conjugation.16,17 A complete in situ Raman characterization of PAz requires registration of the Raman spectra, not only as a function of the potential but also using different excitation wavelengths to be able to enhance the spectral regions where the different electronic transitions occur. In this work, three different excitation lines have been used: 514, 633, and 780 nm. Raman spectra can be obtained from a film deposited on almost any electrode substrate. For many applications, it would be advantageous to use commodity metals as electrode substrates, and therefore we have studied the charging-discharging reactions of PAz on aluminum electrodes. These results have been compared with the results obtained using Pt as the electrode substrate. To the best of our knowledge, there have been no previous reports published on the in situ Raman characterization of PAz on either Al or Pt. The Raman bands appearing in the spectra have not been assigned to any specific vibrations because it is beyond the scope of this work. 2. Experimental Methods 2.1. Chemicals. Azulene (99%) was purchased from Aldrich and used as received. Tetrabutylammonium hexafluorophosphate (TBAPF6, 99%, Fluka) was used as the supporting electrolyte. The electrolyte salt was dried at 80 °C under vacuum for 60 min before use. Acetonitrile (ACN, Riedel de-Häen) was stored over calcium hydride, distilled, and dried over basic alumina (150 mesh, Aldrich) before each experiment. All solutions were purged with nitrogen prior to use. 2.2. Electrochemistry. Azulene was polymerized electrochemically on an Al or a Pt disk (area: 0.07 cm2) in a threeelectrode one compartment electrochemical cell from a solution containing 0.01 M monomer in 0.1 M TBAPF6-ACN. A coiled Pt wire was used as the counter electrode (CE), and a Ag/AgCl wire was used as the quasi-reference electrode (RE) (calibrated versus ferrocene/ferrocenium, Eredox ) 0.36 V). The working electrode (WE) was polished with diamond paste (3 and 1 µm) and Al2O3 (0.3 and 0.05 µm) and rinsed with deionized water and acetone before use. The polyazulene films were synthesized by potentiodynamic cycling between –0.6 and 1.2 V during 12 consecutive cycles at a 50 mV/s scan rate. A more detailed description of the electrochemical synthesis has been reported earlier.18,19 The electrochemical doping was performed potentiostatically by applying a stepwise increasing potential of 0.2 V. In the case of p-doping, the initial potential used was –0.4 V and the final potential used was 1.0 V, whereas n-doping was initiated at 0 V after which the potential was increased stepwise to –2.2 V. The potential was controlled using an Autolab PGSTAT 100 equipped with general-purpose electrochemical system software. 2.3. In situ Resonance Raman Spectroscopy. The in situ resonance Raman spectra were recorded directly on the Al or Pt working electrode at room temperature (23 ( 2 °C) with a Renishaw Ramascope (system 1000 B) equipped with a Leica DMLM microscope and connected to a charge-coupled device camera detector. The spectra were collected at a 90° angle to the excitation beam. The charging-discharging reactions were studied in a monomer-free solution of 0.1 M TBAPF6-ACN in a three electrode spectroelectrochemical cell, using Pt or Al as
Österholm et al. the WE, a Pt wire as the CE, and a Ag/AgCl wire as the RE. A detailed description of the cell has been reported earlier.20 The spectra were recorded at different potentials in the same solution by increasing the potential stepwise as described above. At each potential, the film was allowed to stabilize for 60 s before the spectrum was recorded. The potential was applied for a total of 100 s. The excitation wavelengths used in this work were provided by three lasers with λexc ) 780 nm (Renishaw, NIR diode laser), 633 nm (Renishaw, HeNe laser), and an argon laser with λexc ) 514 nm (LaserPhysics). The spectrometer was calibrated using a Si standard before the measurements. When spectra were not recorded, the laser was switched off to avoid degradation of the film. Some Raman bands were divided into component Lorentzian peaks using a curve fit program available with the GRAMS32 software. 2.4. In situ UV–vis Spectroscopy. In the UV–vis experiments, tin oxide (TO) glass was used as the WE. PAz was deposited by potentiodynamic cycling between -0.6 and 1.2 V during six consecutive cycles at a 50 mV/s scan rate. The measurements were made in a 1 cm path length quartz cuvette, and the TO glass was rinsed with deionized water and acetone prior to use. The same CE and RE were used as in the electrochemical experiments. A background spectrum was always recorded using two blank TO glasses in the electrolyte solution. The spectra were recorded with a Hitatchi U-2001 spectrophotometer. The wavelength was scanned between 1100 and 325 nm with 800 nm/min scan rate. The potential was increased stepwise with an interval of 0.1 V in the case of p-doping and with 0.2 V in the case of n-doping (due to the larger potential range). The potential was kept at a chosen level for 120 s before the spectrum was recorded. 3. Results and Discussion 3.1. Correlation between the in situ Raman Spectra and the in situ UV–vis Spectra of PAz. The in situ UV–vis spectra of PAz on TO glass during p- and n-doping in 0.1 M TBAPF6ACN can be seen in Figure 1, panels a and b, respectively. The lines indicate the excitation wavelengths (λexc) used in the Raman experiments. The in situ UV–vis response of PAz during p- and n-doping has been discussed in detail in our previous work.18,19 In summary, it can be said that the π–π* transition for a neutral PAz film has an absorption maximum at approximately 420 nm. When the potential is increased three absorption bands can be observed at approximately 450, 650, and 900 nm, the latter extending into the NIR region. All three bands have been related to the introduction and presence of different types of charge carriers.18,21 According to the UV–vis spectra, λexc ) 514 nm is located between the π-π* transition and the doping-induced absorption bands, whereas the λexc ) 633 nm is close to the maximum of the optical absorption band at 650 nm. As can be seen, λexc ) 780 nm is close to the electronic transition that gives rise to the absorption band that extends into the NIR. The Raman spectra of PAz in pristine form in electrolyte solution obtained with the three different lasers can be seen in Figure 2. The spectra were obtained at open circuit potential. The wavenumbers of the Raman bands in the spectrum obtained with λexc ) 514 nm (spectrum A) differs from the other two spectra obtained with the λexc ) 633 nm (spectrum B) and the λexc ) 780 nm (spectrum C). Frequency dispersion with excitation wavelength has been explained for some polymers as selective excitation of polymer segments with different conjugation lengths. Frequency dispersion has also been explained as the enhancement of vibrational modes due to insertion
Resonance Raman Spectroscopy of Polyazulene on Aluminum
J. Phys. Chem. B, Vol. 112, No. 20, 2008 6333
Figure 3. In situ Raman spectra of PAz excited by λexc ) 514 nm recorded at different potentials during p-doping in 0.1 M TBAPF6ACN solution. The inserted wavenumbers are taken from the spectra obtained at –0.4 V.
Figure 1. In situ UV–vis absorption spectra of PAz in 0.1 M TBAPF6ACN recorded at different potentials during (a) p-doping and (b) n-doping. The Raman excitation lines are inserted in the spectra.
Figure 2. Raman spectra of PAz at open circuit potential in 0.1 M TBAPF6-ACN excited by λexc ) 514 nm (solid line, A), 633 nm (dotted line, B), and 780 nm (dashed line, C).
of charged species, assuming that the excitation is close to a specific electronic transition of the inserted species.9,14 In this analysis, we will not take into account the dispersion due to differences in conjugation lengths in the polymer. As discussed above in connection with the in situ UV–vis spectra, the λexc ) 514 nm is located close to the π-π*
transition, therefore the Raman bands observed should mainly originate from neutral segments in the film. The two other lasers will, due to the resonance effect, enhance vibrations arising from doped segments of the film. Since the films are held at open circuit potential the films in spectra B and C will be slightly excited since the excitation wavelengths coincide with electronic transitions at around 650 and 900 nm in the UV–vis spectra (Figure 1). The Raman spectrum A exhibits two bands of high intensity at 1562 and 1528 cm-1 (a doublet) and additional bands of weaker intensity can be found at 1400, 1254, 1215, and 1047 cm-1. In spectra B and C, Raman bands are found at approximately 1510, 1355, 1190, and 1000 cm-1. Comparing spectrum B with spectrum C, it is evident that they are very similar. The bands found at 1510 and 1355 cm-1 in spectra B and C are shifted approximately 50 cm-1 toward lower energies when compared with spectrum A. Comparing the lower intensity bands observed in spectrum A with those observed in spectra B and C, it is clear that they show a very similar pattern but are shifted to slightly lower frequencies in spectra B and C. The doublet appearing at 1562 and 1528 cm-1 and the doublet at 1254 and 1215 cm-1 in spectrum A probably merge when lower excitation energies are used and are therefore seen as two broader bands in spectra B and C at approximately 1510 and 1190 cm-1. When comparing spectrum B with spectrum C, a difference in the intensity ratio of the two strongest bands at 1510 and 1355 cm-1 is observed. In spectrum C, the band at 1355 cm-1 is of higher intensity, whereas the band at 1510 cm-1 has a higher intensity in spectrum B. The wavelength-dependence on the intensity dispersion is discussed further in Section 3.2. Raman bands from the solvent appear at approximately 920 and 1375 cm-1; the latter band will appear at similar wavenumbers as bands originating from the film. However, the intensity of the solvent band is significantly lower and thus its contribution to the total intensity should be fairly small. 3.2. In situ Resonance Raman Spectroscopy of PAz. As mentioned in the previous section, the λexc ) 514 nm is close enough to the π-π* transition and to enhance the Raman bands originating mainly from neutral segments in the polymer. The in situ Raman spectra of PAz obtained during p-doping on Al can be seen in Figure 3. The spectra have been separated for a
6334 J. Phys. Chem. B, Vol. 112, No. 20, 2008 clearer comparison. As the potential is increased, the Raman bands gradually start to broaden and decrease in intensity. This suggests that the amount of undoped polymer decreases as the doping level increases. This is common behavior for most conducting polymers during doping when using an excitation wavelength that mainly resonantly enhances Raman bands arising from undoped segments in the film.8,9 At higher doping levels, starting at 0.8 V, a band is observed at 1471 cm-1. This band is even more well defined at 1.0 V. There are several possible origins for this band. One possibility is that the band becomes visible due to broadening and a decrease in intensity of the bands at 1562, 1528, and 1400 cm-1. Another possibility is that it is a new doping-induced band suggesting that also doped segments are resonantly enhanced using this excitation wavelength; this is possible since the excitation wavelength is found between the π-π* transition and the doping-induced absorption bands. According to the ECC model, the Raman spectra of doped polymers in general may show new features at high doping levels but these features usually originate from residual segments that have avoided interaction with a dopant.22 However, Lefrant et al.23 and Furukawa et al.14,24 have reported that Raman spectra of highly doped polymers may contain new doping-induced Raman bands due to the resonance enhancement effect. This is not accounted for in the ECC model. Whether this 1471 cm-1 is a dopinginduced Raman band or simply a band appearing due to the decrease in intensity of the strong doublet cannot be unequivocally concluded just from this spectra. At the highest doping level (1.0 V), the doublet at 1562 and 1528 cm-1 is still visible whereas the Raman band at 1400 cm-1 is hardly observable. The bands between 1250 and 900 cm-1 are still visible at 1.0 V. Because of a high degree of fluorescence a clear in situ Raman spectra of PAz on Al during reduction (n-doping) could not be obtained. It has been shown that also the azulene monomer exhibits anomalous fluorescence behavior.25 The in situ Raman spectra obtained with the λexc ) 780 nm during p- and n-doping of PAz on Al can be seen in Figure 4a, and the corresponding spectra of PAz on Pt can be seen in Figure 4b. The band originating from the solvent is marked with an asterisk (*). The spectra comparing the highest obtained p- and n-doping levels with a spectrum obtained at a potential where the film should be in a pristine state on both substrates can be seen in Figure 4c. The spectra in the figures have been separated for a clearer comparison. The spectra obtained during p- and n-doping are not of the same film but are placed in the same figure to facilitate the comparison. Because of the broadness of the bands, an exact assignment of wavenumbers is difficult; thus, the wavenumbers given are only indicative. The changes observed in the in situ Raman spectra during p-doping (Figure 4a and 4b) will be discussed first. As in the case of the in situ Raman spectra obtained with the λexc ) 514 nm (Figure 3), the bands broaden and decrease in intensity with increased doping levels. Broadening, intensity dispersion, and frequency dispersion due to applied potential is, according to the ECC model, the expected behavior for polymers during doping.7 Higher doping levels are probably achieved when PAz is deposited on Pt (Figure 4b) than on Al because the bands in the spectrum obtained at 1.0 V have almost completely disappeared. The same conclusion was drawn from the FTIRATR experiments performed on PAz deposited on the two different substrates.19 The main changes in the spectra are observed between 1600 and 1300 cm-1. At high p-doping levels, a band appears with a maximum at approximately 1460 cm-1 and with a shoulder slightly above 1500 cm-1. This can clearly
Österholm et al.
Figure 4. In situ Raman spectra of PAz in 0.1 M TBAPF6-ACN solution excited by λexc ) 780 nm recorded at different potentials during (a) p- and n-doping on Al, (b) p- and n-doping on Pt (c) comparison of spectra of highest obtained p- and n-doping levels with spectrum of pristine PAz on Al (solid line) and Pt (dashed line).
be seen in Figure 4a. Once again there are two possible explanations: either this is a Raman band that becomes observable due the decrease in intensity of the bands at 1510 and 1355 cm-1 or it is a band originating from a new species formed during doping. In any case, the shoulder observed is most likely related to the band originally found at 1510 cm-1. Comparing the results of the p-doping of PAz on Al to similar
Resonance Raman Spectroscopy of Polyazulene on Aluminum measurements performed on Pt, it can be concluded that the response is the same regardless of the substrate. The vibrational response during p-doping was reversible in both cases. A wavenumber shift to lower energies is observed during n-doping of PAz on both Al and Pt. The band at 1510 cm-1 gradually shifts to 1485 cm-1 whereas the band at 1355 shifts to 1348 cm-1. In addition, intensity dispersion can also be observed during n-doping. The band found at 1510 cm-1 observed in the spectrum of the pristine polymer has a slightly lower intensity than the band at 1355 cm-1. As the negative potential is increased, the band at 1510 cm-1 begins to grow at the expense of the band at 1355 cm-1. In contrast to p-doping, the bands are still relatively well defined at high doping levels, that is, at high negative potentials. The same behavior was also observed in the Raman spectra of PAz films that were n-doped on Pt electrodes, see Figure 4b. The doping level achieved during n-doping is lower than during p-doping and the charge carriers are also more localized. This might be a reason for the more well-defined bands observed during n-doping. Another reason for the different behavior might be differences in polarizabilities during p-and n-doping. It has been shown that n-doping of some aromatic polymers, like poly(p-phenylene) (PPP), leads to a marked increase in linear polarizability whereas p-doping leads to a decrease in polarizability due to the extraction of highly polarizable π-electrons.26 If this is the case for polyazulene, a similar vibrational behavior would not be expected during p- and n-doping. In studies by Neugebauer et al.,27–29 on some thiophene-based low band gap polymers, it was shown that these materials exhibit different electronic structures during p- and n-doping. Also Lefrant et al. observed a symmetry breaking between positive and negative charge carriers in poly(p-phenylene vinylene) (PPV).30–32 In Figure 4a, at high negative potentials between –1.8 and –2.2 V a new band is observed with a maximum approximately at 1085 cm-1. This band was not observed during p-doping. The same band was also found during n-doping of PAz on Pt approximately at the same frequency; see Figure 4b. As the potential was stepped back from -2.2 V to the initial potential, it was observed that only the intensity dispersion showed some reversibility when PAz was n-doped on Al. A shift back to the original wavenumbers was not observed during discharging. This suggests that the reduction reaction on Al is not completely reversible. The n-doping on Pt was found to be more reversible; the bands shifted back to higher wavenumbers and also the intensity dispersion showed some reversibility during discharging. The intensity ratio between the bands at 1510 and 1355 cm-1 did not return to its initial value but rather to a level comparable to low n-doping levels. This is in good agreement with the in situ UV–vis experiment.19 The band found at 1085 cm-1 did not disappear during discharging on either substrate. This band may be due to a defect or, for example, charge trapping, because it is well known that negative charge carriers are not as mobile as their positive counterparts. When Pt was used as the electrode substrate this new band began to appear at lower doping levels. This might be due to the morphology of the film or that there is more material on the Pt electrode than on the Al electrode, as discussed in our previous work.19 Another reason might be that it is easier to dope PAz on Pt than on Al. In any case, the appearance of a new band clearly contradicts the ECC model. Because of overlapping and the rather weak intensities of the vibration bands found between 1200 and 900 cm-1 (in Figure 4c), the wavenumber of the maximum of each band is taken from the corresponding Lorentzian curve fit. These bands show some frequency disper-
J. Phys. Chem. B, Vol. 112, No. 20, 2008 6335
Figure 5. In situ Raman spectra of PAz in TBAPF6-ACN solution excited by λexc ) 633 nm recorded at different potentials during (a) pand n-doping, (b) the wavenumber range 1700–900 cm-1 decomposed into components of Lorentzian profiles.
sion, but because of the difficulty and uncertainty of the wavenumber assignment they will not be discussed further. The in situ Raman spectra obtained with the λexc ) 633 nm during p- and n-doping of PAz on Al can be seen in Figure 5a. The spectra have been separated for a clearer comparison. The peak originating from the solvent is marked with an asterisk (*). The spectrum obtained at 1.0 V contain undefined, overlapping bands between 1530 and 1340 cm-1. These bands can be decomposed into components of Lorentzian profiles; the spectrum including the Lorentzian peaks can be seen in Figure 5b. The film at -0.4 V is assumed to be in a pristine state. As the positive potential is increased, the bands broaden and decrease in intensity, showing a similar behavior as when PAz was excited with λexc ) 780 nm. In contrast to the spectra obtained with the λexc ) 780 nm, the band at approximately 1510 cm-1 is always of higher intensity than the band at approximately 1350 cm-1. At the highest p-doping potentials, there appears to be three bands between 1600 and 1300 cm-1. This is probably the same phenomenon as observed in Figure 4a (λexc ) 780 nm), but in this case the relative intensity between the three bands is different. There is also a slight difference in some of the band positions when comparing Figure 5a with Figure 4a. The band originally found at around 1510 cm-1 (λexc ) 780 nm and λexc ) 633 nm) is more defined at high doping levels when excited
6336 J. Phys. Chem. B, Vol. 112, No. 20, 2008
Figure 6. Raman spectrum (λexc ) 514 nm, solid line) of neutral PAz film on Al and IRAV (dashed line) spectrum of p-doped PAz film on Al in 0.1 M TBAPF6-ACN solution.
with λexc ) 633 nm compared with λexc ) 780 nm, whereas the band 1460 cm-1/1440 cm-1 (λexc ) 780 nm/λexc ) 633 nm, respectively) is more defined using the latter excitation line. The wavelength-dependence on the relative Raman band intensities in PPV has been explained as a coexistence of different types of charge carriers that are enhanced differently depending on the laser excitation wavelength.31,32 The PAz deposited on Pt showed the same behavior (spectra not shown here). The spectra obtained with λexc ) 633 nm during n-doping on Al was quite similar to that obtained during n-doping on Al with λexc ) 780. Using the former excitation line, the intensity ratio between the two peaks at approximately 1507 and 1350 cm-1 is larger and continues to increase as the negative potential is increased. The band at 1085 cm-1 observed during n-doping when using λexc ) 780 nm is hardly visible when using λexc ) 633. The p-doping was found to be rather reversible on both substrates whereas the n-doping reaction was not reversible on Al. Figure 6 shows the comparison between the doping-induced IRAV bands from a p-doped PAz film and the Raman bands of a pristine PAz film, both on Al electrodes and in 0.1 M TBAPF6ACN solution. The experimental setup and results of the infrared measurements have been presented in our previous work.19 Both films have been electrodeposited using the same parameters as described in Section 2.2. Because of a larger surface area of the electrode in the infrared experiment, the number of cycles during polymerization was twice the number of cycles in the Raman experiments. The excitation wavelength used in the Raman experiment was 514 nm. As can be seen, the IRAV bands show a similar pattern as the Raman bands but they are shifted to lower wavenumbers. These experimental results are in good agreement with the ECC model. 4. Conclusions In this work, the redox reactions of PAz have been studied by in situ Raman spectroscopy by using three different excitation lines, λexc ) 514, 633, and 780 nm. The 514 nm excitation line mainly enhanced vibrations in the neutral form of PAz, whereas the λexc ) 633 and 780 nm enhanced vibrations of doped segments in the film. The PAz film in electrolyte solution showed more sensitivity toward the 514 nm excitation line than
Österholm et al. toward the other two lasers. The Raman bands exhibited intensity and frequency dispersion when changing the excitation wavelength. Different vibrational behavior was observed during p- and n-doping. This may be due to differences in polarizabilities of the polymer during the p- and n-doping reactions.27 PAz may exhibit different electronic structures with positive and negative charging, as has been shown for some low band gap thiophenes and PPV. The degree of localization of positive and negative charge carriers may also be different.27–32 The vibrational response of PAz on Al is very similar to that of PAz on Pt. This indicates that electroactive PAz films with reversible electrochemical properties can also be synthesized on Al. The p-doping reaction was found to be reversible on both substrates whereas the n-doping was not reversible on Al, it was possible to charge the films but not discharge them. The same behavior was observed earlier with FTIR.19 Because in situ Raman study of PAz has not been previously reported, the bands appearing in the Raman spectra have not been assigned to any specific vibrations. It is beyond the scope of this work to assign the observed Raman bands, therefore only some general conclusions regarding their origin will be made. It is well known that Raman spectra of pristine polyconjugated molecules show the strongest characteristic lines near 1500–1400 cm-1, that is, in the C)C stretching range and the C-C stretch and C-H wag can be found in the wavenumber range 1200–900 cm-1.33 These are probably the vibrations that give rise to the bands found in the Raman spectrum of PAz. It has been shown that Я can be defined for several polyaromatic and polyheteroaromatic systems like PPP, PPV, PPy, PTh, and rylenes.9 On the other hand, in the case of condensed aromatic systems like anthracene and pentacene, a specific or implicit definition of Я could not be made due to the topologies of the pz orbitals. Therefore the Raman spectra of these molecules are not accounted for in the ECC model. Because of its nonalternating, unsymmetrical structure, it is difficult to determine on a purely experimental basis if PAz can be discussed within the ECC framework. The correlation between the Raman bands of neutral PAz and IRAV bands of doped PAz is in good agreement with the ECC model, which is not the case for all fused-ring, and condensed aromatic systems. However, there are some features in the Raman spectra of doped PAz that are not in accordance with the ECC model. As in the case of PAz, t-butyl-capped rylenes show many bands in the infrared spectra but only few lines in the Raman spectra. These molecules are planar and rigid thus favoring a large overlap of the pz orbitals and causing a large degree of intramolecular delocalization of the π-electrons thus giving rise to a large electron–phonon coupling. A similar case is found with low band gap polymers; these systems show large hyperpolarizabilities due to strong electron–phonon coupling and thus only a few Raman modes can be observed in their spectra. Rylenes also show large hyperpolarizabilities making them more similar to polyenes than phenylenes since the planar geometry forces delocalization of electrons in the plane of the molecule.9 This may be the reason why Raman spectra of PAz show fewer bands than similar spectra of, for examle, polynaphthtalene and polyparaphenylene.34,35 Acknowledgment. Financial support from the Graduate School in Chemical Engineering is gratefully acknowledged (A.Ö.). This work forms part of the activities being pursued at the Åbo Akademi Process Chemistry Centre, within the Finnish Centre of Excellence Programme (2000–2011) of the Academy of Finland.
Resonance Raman Spectroscopy of Polyazulene on Aluminum References and Notes (1) Su, W. P.; Schrieffer, J. R.; Heeger, A. J. Phys. ReV. B 1980, 22, 2099–2111. (2) Brédas, J.-L.; Chance, R. R.; Silbey, R. Phys. ReV. B 1982, 26, 5843–5854. (3) Horowitz, G.; Abderrahim, Y.; von Bardeleben; H. J, Synth. Met. 1994, 62, 245–252. (4) van Haare, J. A. E. H.; Havinga, E. E.; van Dongen, J. L. J.; Janssen, R. A. J.; Cornil, J.; Brédas, J.-L. Chem.-Eur. J. 1998, 4, 1509–1522. (5) Hill, M. G.; Mann, K. R.; Miller, L. L.; Penneau, J.-F. J. Am. Chem. Soc. 1992, 114, 2728–2730. (6) Bäuerle, P.; Segelbacher, U.; Maier, A.; Mehring, M. J. Am. Chem. Soc. 1993, 115, 10217–10223. (7) Zerbi, G.; Gussoni, M.; Castiglioni, C. In Conjugated Polymers; Brédas, J. L., Silbey, R., Eds.; Kluwer Academic Publishers: Dordrecht, 1991; pp 435–507. (8) Zagórska, M.; Prón, A.; Lefrant, S. In Handbook of Organic ConductiVe Molecules and Polymers; Nalwa, H. S., Ed.; John Wiley & Sons: London, 1997; Vol. 3, Chapter 4. (9) See for instance: Del Zoppo, M.; Castiglioni, C.; Zuliani, P.; Zerbi, G. In Handbook of Conducting Polymers, 2nd ed.; Skotheim, T. A., Elsenbaumer, R. L., Reynolds, J. R., Eds.; Marcel Dekker: New York, 1998; Chapter 28 and references therein. (10) Castiglioni, C.; Del Zoppo, M.; Zerbi, G. J. Raman Spectrosc. 1993, 24, 485–494. (11) Horovitz, B. Solid State Commun. 1982, 41, 729–734. (12) Ehrenfreund, E.; Vardeny, Z. V.; Brafman, O; Horovitz, B. Phys. ReV. B 1987, 36, 1535–1553. (13) Furukawa, Y.; Tazawa, S.; Fuji, Y.; Haradi, I. Synth. Met. 1988, 24, 329–341. (14) Furukawa, Y.; Tasumi, M. In Modern Polymer Spectroscopy; Zerbi, G., Ed.; Wiley-VCH Verlag Gmbh: Weineim, Germany, 1999; Chapter 4 and references therein. (15) Rasmussen, S. C.; Pomerantz, M. In Handbook of Conducting Polymers, 3rd ed.; Skotheim, T. A., Reynolds, J. R., Eds.; CRC Press: Boca Raton, Florida, 2007; Chapter 12. (16) Hino, S.; Iwasaki, K.; Matsumoto, K. Synth. Met. 1994, 64, 259– 264.
J. Phys. Chem. B, Vol. 112, No. 20, 2008 6337 (17) Brédas, J. L. In Handbook of Conducting Polymers; Skotheim, T. A., Ed.; Marcel Dekker: New York, 1986; Vol. 2, Chapter 25. (18) Lete, C.; Meana-Esteban, B.; Kvarnström, C.; Razus, A. C.; Ivaska, A .Electrochm. Acta 2007, 52, 6476–6483. (19) Österholm, A.; Meana Esteban, B.; Kvarnström, C.; Ivaska, A. J. Electroanal. Chem. 2008, 613, 160–170. (20) Damlin, P.; Kvarnström, C.; Petr, A.; Ek, P.; Dunsch, L.; Ivaska, A. J. Solid State Electrochem. 2002, 6, 291–301. (21) Shim, Y.-B; Park, S.-M. J. Electrochem. Soc. 1997, 144, 3027– 3033. (22) Castiglioni, C.; Rabaioli, R.; Zerbi, G. Synth. Met. 1991, 41–43, 3453–3456. (23) Lefrant, S.; Mulazzi, E.; Mathis, C. Phys. ReV. B 1994, 49, 13400– 13407. (24) Furukawa, Y. J. Phys. Chem. 1996, 100, 15644–15653. (25) Kozlowski, P. M.; Rauhut, G.; Pulay, P. J. Chem. Phys. 1995, 103, 5650–5661. (26) Brédas, J. L.; Dory, M.; Thémans, B.; Delhalle, J.; André, J. M. Synth. Met. 1989, 28, D533–542. (27) Cravino, A.; Neugebauer, H.; Luzzati, S.; Catellani, M.; Petr, A.; Dunsch, L.; Sariciftci, N. S. J. Phys. Chem. B 2002, 106, 3583–3591. (28) Neugebauer, H.; Cravino, A.; Luzzati, S.; Catellani, M.; Petr, A.; Dunsch, L.; Sariciftci, N. S. Synth. Met. 2003, 139, 747–750. (29) Neugebauer, H. J. Electroanal. Chem. 2004, 563, 153–159. (30) Baïtoul, M; Wéry, J.; Lefrant, S.; Faulques, E.; Buisson, J.-P.; Chauvet, O. Phys. ReV. B 2003, 68, 195203-1–195203-6. (31) Orion, I.; Buisson, J.-P.; Lefrant, S. Phys. ReV. B 1998, 57, 7050– 7065. (32) Baïtoul, M.; Wéry, J.; Buisson, J.-P.; Arbuckle, G.; Shah, H.; Lefrant, S.; Hamdoume, M. Polymer 2000, 41, 6955–6964. (33) Socrates G. In Infrared and Raman characteristic group frequencies; John Wiley & Sons Ltd: New York, 2001. (34) Zhang, J.; Liu, C.; Shi, G.; Zhao, Y. J. Polym. Sci., Part B: Polym. Phys. 2005, 43, 241–251. (35) Dubois, M.; Froyer, G.; Louarn, G.; Billaud, D. Spectrochim. Acta, Part A 2003, 59, 1849–1856.
JP0762828
