Anal. Chem. 2004, 76, 5236-5240
Two-Dimensional Correlation Analysis of Spectroelectrochemical Data for p-Benzoquinone Reduction in Acetonitrile Young-Ok Kim, Young Mee Jung, Seung Bin Kim, and Su-Moon Park*
Department of Chemistry and Center for Integrated Molecular Systems, Pohang University of Science and Technology, Pohang 790-784, Korea
Two-dimensional (2D) spectral correlation analysis has been employed to interpret the complex spectroelectrochemical data obtained from an electrochemical system undergoing following reactions after electron transfer. The system used was electrochemical reduction of p-benzoquinone (p-BQ) in acetonitrile, which produces anion radicals and dianions at its first and second reduction potentials. The dianions undergo a fast comproporationation reaction with neutral p-BQ molecules to produce anion radicals back, complicating the spectral analysis. Upon application of 2D correlation analysis in conjunction with the self-modeling curve resolution technique, we were able not only to resolve the spectra and determine the sequence of spectral emergence but also to extract the individual spectra. The techniques offer a very powerful tool for interpreting highly convoluted spectra obtained from a system where a series of chemical reactions occur following the electron transfer at the electrode/electrolyte interface. The spectroelectrochemical technique has been widely used for studies of a variety of electrochemical reactions since Kuwana et al. first introduced it.1 Most spectroelectrochemical experiments described in earlier reports used a potentiostatic mode, and time was the only external perturbation under such conditions. One of the major advances made for a better interpretation of the spectroelectrochemical data include the derivative cyclic voltabsorptometry (DCVA), in which derivative absorbance data are plotted as a function of potential at a given wavelength;2,3 the DCVA curves are directly related to the corresponding voltammetric currents. Also, the technology advances made in spectrographs using detectors such as charge-coupled device (CCD) array detectors and polychromators allowed the spectral recording to be made in a much shorter time, and many recent spectroelectrochemical experiments were run in real time during the potential sweep.4 As a result, another external perturbation, i.e., * To whom correspondence should be addressed. E-mail: smpark@ postech.edu. Phone: +82-54-279-2102. Fax: +82-54-279-3399. (1) Kuwana, T.; Darlington, R. K.; Leedy, D. W. Anal. Chem. 1964, 36, 2023. (2) Bancroft, E. E.; Sidwell, J. S.; Blount, H. N. Anal. Chem.1981, 53, 1390. (3) (a) Zhang, C.; Park, S.-M. Anal. Chem. 1988, 60, 1639. (b) Zhang, C.; Park, S.-M. Bull. Korean Chem. Soc. 1989, 10, 302. (4) See, for example: (a) Shim, H.-J.; Yeo, I.-H.; Park, S.-M. Anal. Chem. 2002, 74, 3540. (b) Lee, H. J.; Cui, S.-Y.; Park, S.-M. J. Electrochem. Soc. 2001, 148, D139.
5236 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
potential, could be introduced to the experiments, and a large body of spectral data thus obtained provide a wealth of information provided that there are no complicating reactions. The spectroelectrochemical data can often be too complex, particularly when the data are obtained in a cell, where the following chemical reactions are allowed to occur between the electrogenerated species and its environment after the electron transfer. This situation demands a new data treatment technique for better analysis of the complex spectral data. Generalized two-dimensional (2D) correlation spectroscopy has become one of standard analytical techniques to interpret spectral data sets obtained during the observation of spectra with an external perturbation.5 In 2D correlation spectroscopy, the spectral intensity is plotted as a function of two independent spectral variables such as wavelength, frequency, or wavenumber. The two orthogonal axes of spectral variables define the 2D spectral plane, and the spectral intensity is obtained along the third axis normal to the spectral plane. Some of the most notable features of 2D correlation spectra are the following: simplification of complex spectra consisting of many overlapped peaks; enhancement of spectral resolution by spreading peaks along the second dimension; establishment of unambiguous assignments through the correlation of bands of selectively coupled by various interaction mechanisms; and determination of the sequence of the spectral peak emergence.5 The intensity of a synchronous 2D correlation spectrum represents the simultaneous or coincidental changes of spectral intensity variations measured at two different spectral variables, ν1 and ν2. Autopeaks located at the diagonal line represent the overall susceptibility of the corresponding spectral region to change in spectral intensities as an external perturbation is applied to the system. The external perturbations include changes in potential, electrolysis time, or temperature during spectroelectrochemical experiments. Cross peaks located at the off-diagonal positions represent simultaneous or coincidental changes of spectral intensities observed at ν1 and ν2. The sign of synchronous cross peaks becomes positive if the spectral intensities at the two spectral variables corresponding to the coordinates of the cross peak are either increasing or decreasing together as functions of the external variable t during the observation interval. However, (5) (a) Noda, I. Appl. Spectrosc. 1993, 47, 1329. (b) Noda, I.; Dowrey, A. E.; Marcott, C.; Story, G. M.; Ozaki, Y. Appl. Spectrosc. 2000, 54, 236A. (c) Noda, I. Appl. Spectrosc. 2000, 54, 994. 10.1021/ac049587g CCC: $27.50
© 2004 American Chemical Society Published on Web 07/29/2004
the negative sign of cross peaks indicates that one of the spectral intensities is increasing while the other is decreasing. The intensity of an asynchronous 2D correlation spectrum represents sequential or successive changes of spectral intensities measured at ν1 and ν2. An asynchronous cross peak develops only if the intensities of two spectral features change out of phase (i.e., delayed or accelerated) with each other. The sign of an asynchronous cross peak becomes positive if the intensity change at ν1 occurs predominantly before ν2 in the sequential order of t. It becomes negative, on the other hand, if the change occurs after ν2. This rule, however, is reversed if synchronous peak is negative. To extract the absorption spectrum of each species generated during electrochemical reduction, one can use the self-modeling curve resolution (SMCR) analysis.6 Jung et al.7 have introduced the use of 2D correlation spectroscopy in conjunction with alternating least-squares (ALS)-based SMCR analysis of the spectral data set to obtain a set of concentration profiles and spectra of pure components from a set of unknown mixture spectra without a prior knowledge about the system. In the present study, we explore the use of the 2D correlation analysis technique in conjunction with the SMCR analysis for a complete interpretation of the highly convoluted in situ spectroelectrochemical data obtained during p-BQ reduction. The techniques provide a complete interpretation of the complex spectroelectrochemical data obtained with the potential used as an external perturbation, and also, the result obtained thereof offers a new insight into the p-BQ reduction reaction as well. EXPERIMENTAL METHOD p-Benzoquinone (Aldrich, 98%) was used after recrystallized twice from ethanol and dried overnight. Tetrabutylammonium hexafluorophosphate (TBAPF6, Fluka, >99.0%) used as an electrolyte was also recrystallized twice from ethanol and dried overnight under reduced pressure at 100 °C. Acetonitrile (Aldrich, 99.8%, anhydrous) was dried by three vacuum distillations over P2O5. Oxygen was removed by a few freeze-pump-thaw cycles before the dry acetonitrile was moved to a glovebox under an argon atmosphere. Electrochemical experiments were performed using an EG&G PAR model 273A potentiostat/galvanostat. A glassy carbon electrode (area, 0.2 cm2) was polished successively with 1.0-, 0.3-, and 0.05-µm alumina slurries (Fisher) and then cleaned ultrasonically with doubly distilled, deionized water before it was used as a working electrode. A platinum spiral wire and a silver wire were used as counter and pseudo-reference electrodes, respectively. The cell preparation was done in a glovebox filled with argon gas. In situ UV-visible absorption spectra were taken with an Oriel InstaSpec IV spectrometer with a CCD array detector, which was configured in a near-normal incidence reflectance mode using a bifurcated quartz optical fiber.3,8 A xenon lamp (75 W) was used as a light source. The wavelength of the spectrograph was calibrated using a small mercury lamp. The 2D correlation analysis was performed using an algorithm based on the numerical method developed by Noda.5 The 2D correlation analysis was carried out after the baseline correction (6) (a) Kvalheim, O. M.; Liang, Y. Anal. Chem. 1992, 64, 936. (b) Gemperline, P. J. Anal. Chem. 1999, 71, 5398. (c) Grand, B.; Manne, R. Chemom. Intell. Lab. Syst. 2000, 50, 19. (7) Jung, Y. M.; Noda, I.; Kim, S. B. Appl. Spectrosc. 2003, 57, 1376. (8) Pyun, C.-H.; Park, S.-M. Anal. Chem. 1986, 58, 251.
Figure 1. Cyclic voltammogram of BQ for reduction of 2.0 mM BQ recorded at a GC electrode in dried CH3CN containing 0.1 M TBAPF6 as a supporting electrolyte. The scan rate was 50 mV/s.
of the UV-visible spectra. A subroutine named KG2D9 written in Array BASIC language (GRAMS/386; Galactic Inc., Salem, NH) was employed for the 2D correlation analyses. RESULTS AND DISCUSSION Figure 1 shows a typical cyclic voltammogram (CV) for p-BQ reduction at a glassy carbon electrode in dry CH3CN with 0.10 M TBAPF6 used as a supporting electrolyte in a bulk cell. The electrochemical behavior shown here is in excellent agreement with those reported in the literature,10 and two steps of electron transfer, in which the anion radical and the dianion are produced at the first and second CV peaks, respectively, are seen clearly. In the bulk cell, the electrochemically generated products at the working electrode are continuously exposed to the flux of reactants. Figure 2a shows a series of spectra recorded during the potential sweep between -0.30 and -1.8 V and back to -0.30 V. As the potential approaches the first CV peak, p-BQ is reduced to produce its anion radical (Q•-), and the corresponding spectra, which are in excellent agreement with those reported in the literature,11,12 are observed. However, the spectra recorded beyond the second CV peak do not appear to have an absorption band assignable to the dianion (Q2-). Singling out a few spectra at potentials corresponding to Q•- production (Figure 2b-2) and to Q2- generation (Figure 2b-3) does not help much as can be seen in Figure 2b except that some small spectral shifts, which are not readily noticeable in Figure 2a, are observed. We attribute the lack of the absorption band assignable to Q2- to the heavy (9) The program can be downloaded from the homepage of Prof. Yukihiro Ozaki of Kwansei Gakuin University, Sanda, Japan. (http://science.kwansei.ac.jp/∼ozaki/). (10) See, for example: Zhao, X,; Imahori, H.; Zhan, C.-Z.; Sakata, Y.; Iwata, S.; Kitakawa, T. J. Phys. Chem. A 1997, 101, 622. (11) (a) Pyun, C.-H.; Park, S.-M. J. Electrochem. Soc. 1985, 132, 2426. (b) Shim, Y.-B.; Park, S.-M. J. Electroanal. Chem. 1997, 425, 201. (12) Gamage, R.; Umapathy, S.; McQuillan, A. J. J. Electroanal. Chem. 1990, 284, 229.
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5237
Figure 3. (a) Synchronous and (b) asynchronous 2D correlation spectra derived from the spectra recorded between -0.3 and -1.1 V (Figure 2a). Solid and dotted lines indicate regions of positive and negative correlation, respectively. Figure 2. Spectra recorded concurrently with the CV shown in Figure 1; (b) spectra singled out at -0.3 (1, -), -0.8 (2, - -), and -1.6 V (3, ‚‚‚), respectively.
overlapping of absorption bands due to products from the following chemical reactions after the electron transfer (vide infra), which clutters the overall spectrum. However, the spectra shown in the optically transparent thin-layer electrochemical (OTTLE) cell clearly displayed that the dianion had its own absorption band at ∼370 nm,12 which was not identified in our spectra. The recording of Q2- spectrum was possible in the OTTLE cell because exhaustive electrolysis was achieved in it and no neutral p-BQ molecules were available for the following reaction. To help understand the process, we briefly review the electrochemistry of p-BQ, which undergoes a two-electron-transfer reaction,10,13
Q + e- h Q-•
(1)
Q-• + e- h Q2-
(2)
than generally thought in the region where the second electron transfer takes place due to the fast comproportionation reaction,14
Q + Q2- h 2Q-•
(3)
whose equilibrium constant is calculated to be 2.7 × 1012 from the respective reduction potentials for reactions 1 and 2. For this reason, the absorption band of a large amount of Q•- might have buried those resulting from Q2- and/or other species, which makes the spectroelectrochemical data uninterpretable by a simple examination of the spectra when the data are obtained in a bulk cell, where reaction 3 is allowed to occur. To interpret the spectra shown in Figure 2, we performed the 2D correlation analysis on them. Figure 3 shows the 2D correlation spectra for those shown in Figure 2. Figure 3a shows a synchronous 2D correlation spectrum obtained during the potential sweep between -0.3 and -1.1 V. Autopeaks shown at 320, 416, and 442 nm indicate that these peaks undergo changes in their intensities upon scanning the potential. Synchronous cross peaks at (320, 416), (320, 442) and (416, 442) nm show that the changes in
where Q represents p-BQ. p-BQ reduction is much more complex (13) Patai, S.; Rappoport, Z. The Chemistry of the Quinonoid Compounds; Wiley: New York, 1974.
5238
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
(14) (a) Babaei, A.; Connor, P. A.; Mcquillan, A. J.; Umapathy, S. J. Chem. Edu. 1997, 74, 1200. (b) Russel, C.; Jaenicke, W. J. Electroanal. Chem. 1986, 199, 139.
Figure 4. (a) Synchronous and (b) asynchronous 2D correlation spectra obtained from the spectra with the potential scan between -1.1 and -1.8 V (Figure 2a). Solid and dotted lines indicate regions of positive and negative correlation, respectively.
spectral intensities of these bands at 416 and 442 nm are strongly interrelated with each other. The analysis of the asynchronous 2D correlation spectrum (Figure 3b) shows the following sequence of changes in spectral intensities: 320 f 328 f (416 and 442) nm. What appeared to be a single peak at ∼320 nm (Figure 2b-2) is clearly resolved in the asynchronous correlation spectrum (Figure 3b), and two distinctive asynchronous cross peaks indicate that a new absorption peak at 328 nm begins to show up after the 320-nm peak. This is an observation not reported thus far in the literature. The 2D correlation spectra obtained in the potential region between -1.1 and -1.8 V, where Q•- is reduced to Q2-, are shown in Figure 4. The analysis of 2D correlation spectra reveals that the broad autopeak at 342 nm in the synchronous 2D correlation spectrum is clearly resolved into two peaks, one at 330 nm and the other at 351 nm in the asynchronous 2D correlation spectrum. The band at 351 nm is influenced by the band at 330 nm, and it appears to have originated from Q2-, which is not readily noticeable in the 1D spectra (Figure 2b-3). We also determine the order of peak emergence of (416 and 442) f 330 f 351 nm from the analysis of the asynchronous 2D correlation spectrum shown in Figure 4b. Our observations described thus far may be summarized as follows: (1) the band at 320 nm corresponding to the absorption
of Q•- grows first along with those at 416 and 442 nm upon generation of the anion radical, (2) a new band begins to show up at 328 nm even before the potential reaches the second CV peak region, (3) both the 320- and 328/330-nm bands accompany the 416- and 442-nm bands, and (4) the 351-nm band begins to show up as the potential moves into the second CV peak region where the dianions are produced. This series of spectral emergence is somewhat consistent with the sequence observed in the OTTLE cell,12 in which exhausitive electrolysis is done. In the OTTLE cell, the anion radicals are observed in the first CV peak region and only the dianions are observed in the second peak region as all the anion radicals undergo further reduction to dianions (no neutral BQ molecules are available for reaction 3). The differences between the spectra obtained in the OTTLE cell12 and our observations are that (1) the same bands are observed at shorter wavelengths in our work than those observed in the OTTLE cell by 5-14 nm and (2) the new band emergence observed in our work at 328/330 nm was not observed in the OTTLE cell.12 The differences in wavelengths for the same bands in the two works are attributed to different experimental conditions such as solvents and spectrographs. To confirm that the band at 351 nm observed by 2D correlation analysis indeed arises from Q2-, we employed SMCR analysis6,7 to extract individual spectra from the spectral data obtained between -1.1 and -1.8 V. For successful fitting of the spectral data by ALS iteration, of particular importance is the initial selection of the pure variables. As an initial pick of the pure variables, the dominant cross peaks in the asynchronous 2D correlation spectrum (Figure 4b) representing the pure variable bands, i.e., (330, 351), (351, 416), and (351, 442) nm, were selected; the cross peaks shown in an asynchronous 2D correlation spectrum represent the completely revolved spectral peaks. The intensity of pure variables was used as the initial guess for the score matrix comprising the information on the concentration profile of pure components. We then use the iterative ALS regression to obtain the concentration profiles and pure component spectra that satisfactorily describe the experimentally obtained dynamic spectral data matrix. The scores and loading vectors of the ALS representing the concentration profiles (C) and the spectra of individual chemical components (S) were obtained. Details of the theoretical background of the algorithm and the iteration procedure are described elsewhere.7 The spectra of individual chemical components obtained from the SMCR analysis and their concentration profiles are shown in Figure 5a and b, respectively. The SMCR-estimated spectra thus obtained show well-resolved spectra of three species from what appeared to be a single spectrum. What was believed to be a single band arising from Q•- actually consists of two bands as shown in Figure 5a as solid and dotted spectra. This explains why the band at ∼320 nm underwent red shifts and the growth ratio of the 320 nm to that at 416/442 nm varied over time after electrolysis.11 The band at longer wavelengths always grew faster than that at shorter wavelengths in the presence of proton donors, which is now readily explained by the spectral shapes of each species shown in Figure 5a. The 351-nm band assigned to BQ2- in the 2D correlation analysis actually corresponds to the broad band at ∼356 nm in the SMCR-extracted spectrum. A reasonable set of concentration profiles during the potential sweep between -1.1 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
5239
the dianion is not even produced. Also, this species is already generated well before Q2- is generated as pointed out above. We therefore conclude that the species represented by the dotted spectrum is protonated Q•- or QH• produced from the reaction of Q•-,
Q•- + H+ h QH•
(4)
with a trace amount of H+ in acetonitrile. The protonated free radical thus produced would absorb at a smaller energy than the anion radical would with its excited states almost the same as Q•-. Our result obtained in the bulk cell using the 2D and SMCR analysis techniques is in excellent agreement with that obtained in the OTTLE cell in that the dianion spectrum was observed and extracted. However, the spectrum from a third species (QH•) was also extracted by the SMCR technique, which had not been possible from a simple examination of the spectra obtained in the OTTLE cell.
Figure 5. SMCR results obtained from the spectra of BQ shown in Figure 2a between -1.1 and -1.8 V: (a) spectra of individual chemical components and (b) their concentration profiles. Solid, dashed, and dotted lines represent BQ•-, BQ2-, and BQH•, respectively.
and -1.8 V was also obtained (Figure 5b). It is seen from this figure that the intensity of the Q•- band at 320 nm (solid) stays nearly constant, while both that for Q2- at 356 nm (dashed) and the absorption at 330 nm (dotted), which is very similar to that of Q•-, increases until both of them reach equilibrium concentrations at the increasingly negative potential. Note that the 330-nm band already has a fairly high concentration at -1.1 V (Figure 5b, dotted line). This rather steady-state concentration profile of the anion radical (solid line) absorbing at 320 nm in Figure 5b is not consistent with that expected from reaction 3, which predicts a rapid increase in its concentration. This observation, when considered along with the spectra shown in Figure 5a, suggests that both the Q2- (dashed) and another absorption at 330 nm (dotted) evolve at the expense of Q•- (solid) during the electrolytic generation of Q2-. The concentration of Q•- would reach a steadystate value if the rate of decay of Q•- is about the same as its rate of generation via reaction 3, or reaction 3 competes with some other reaction for decay of Q2- resulting in less efficient generation of Q•- than reaction 3 does. Here, we conclude that the effective decay of Q•- via the protonation reaction is responsible for the steady-state concentration of Q•- because the concentration of the species absorbing at 330 nm is already fairly high at -1.1 V where 5240
Analytical Chemistry, Vol. 76, No. 17, September 1, 2004
CONCLUSIONS We have demonstrated from the spectroelectrochemical study of p-benzoquinone reduction that the 2D correlation analysis of the highly convoluted spectra due to the overlapped absorption resulting from the products of the following chemical reactions after the electron transfer provides a complete analysis of the spectra. The analysis allows us to resolve the spectra, and the SMCR technique offers a method to separate individual components from the highly convoluted spectra obtained. Employing the 2D correlation and SMCR analyses, the spectrelectrochemical data obtained during p-BQ reduction have been resolved; the individual spectra responsible for absorption by Q•-, Q2-, and QH•, as well as their concentration profiles, have been obtained. To our knowledge, it is first time that the QH• is identified by a spectroscopic method during the reduction of p-BQ, which explains many observations described in the literature. This species has been described in the literature from indirect evidence obtained by various transient electrochemical experiments,10-14 and our present result explains the observation of this species well. Our study also demonstrated that the progression of the reaction can be followed as shown in Figure 5b. Quantitative information about the reaction kinetics may be obtained by fitting the concentration profile by computer provided the proton concentration is known. The 2D correlation analysis used along with the SMCR analysis offers a powerful technique for complete analysis of complex spectroelectrochemical data due to following chemical reactions. ACKNOWLEDGMENT This work was supported by the Korea Science and Engineering Foundation through its grant to the Center for Integrated Molecular Systems at Pohang University of Science and Technology. The graduate stipends were provided by the BK21 program of the Korea Research Foundation.
Received for review March 18, 2004. Accepted June 11, 2004. AC049587G
