Bidimensional Spectroelectrochemistry: application of a new device in the study of a o-vanillin-copper(II) complex

A new bidimensional spectroelectrochemistry setup for UV-Vis absorption measurements has been developed. The new device has been used to follow electrochemical reactions using two different arrangements: 1) a near-normal configuration that supplies information about the processes taking place both on the electrode surface and in the solution adjacent to it, and 2) a long-optical-pathway


Introduction
Spectroelectrochemistry (SEC) is a powerful instrumental technique that combines two classical analytical techniques, such as electrochemistry and spectroscopy, to obtain in situ chemical information [1][2][3][4][5][6] about the reactions taking place during an electrochemical experiment. Usually, the electrochemical technique controls the chemical process, while spectroscopy provides molecular information about chemical compounds involved in the process, being complementary to the electrochemical information. Since 1964, when Kuwana proposed this technique [7], many approaches using different experimental setups [8][9][10][11][12][13][14] have been proposed to improve the quality and significance of the spectroelectrochemical experiments, trying to shed more light on the processes involved in the electrochemical reactions.
UV-Vis absorption spectroelectrochemistry can be performed in two different optical configurations, taking into account the position of the light beam with respect to the working electrode: normal [15,16] and parallel [17][18][19] arrangement. On the one hand, in normal configuration the light beam goes through the diffusion layer perpendicularly to the working electrode surface. In this case, a transmission [20,21] or a reflectance [22,23] experimental setup can be used. On the other hand, in parallel configuration, also known as long-optical-pathway configuration, the light beam passes parallel to the electrode surface, with the light beam sampling the adjacent solution to the electrode surface [19,24,25]. Each configuration provides different type of information about the processes occurring at the electrode surface. Normal configuration mainly supplies information about the whole diffusion layer, including the compounds adsorbed or crystalized on the electrode surface, but also a small contribution of species in the solution near the electrode can be obtained. However, parallel configuration only provides information about the processes occurring in the solution close to the electrode. Besides, this second arrangement guarantees a longer optical pathway than in normal configuration. The main drawback of parallel configuration is the lack of sensitivity to adsorption/crystallization phenomena taking place on the electrode surface.
Usually, only one of these configurations is chosen depending on the chemical problem, but in many cases both kind of information are necessary to fully understand the studied chemical system. In this respect, bidimensional spectroelectrochemistry (BSEC) [26] allows us to perform simultaneously measurements in the two configurations, obtaining in only one experiment a more complete picture on the electrode processes. The main disadvantage of BSEC is the complexity of the spectroelectrochemical cells to get reproducible and good enough experimental results. Recently this problem has been minimized using optical fibers, simplifying the correct alignment of the light beam and improving the reproducibility of the experiment [27]. In this work, another setup with specific improvements is proposed for BSEC. On one hand we use a simple three electrode system placed in a spectrophotometric cuvette together with a reflection probe for normal configuration measurements. On the other, we use piezoelectric positioners to control the position of a slit that allows us to place the light beam adjacent to the electrode at controlled distances. Thus, although parallel beam alignment is usually very complex and troublesome, by using this new approach, previously tested with liquid/liquid interfaces [20], the experiments can be performed in a very simple and easy way.
Firstly, o-tolidine has been selected as reference system to prove the good performance of this new spectroelectrochemical setup [26,28,29]. Finally, BSEC has been used to study the electrochemical mechanism of reaction of, [Cu(o-Va) 2 Copper can be coordinated by ligands with therapeutic properties, often with a synergic effect on the activity, as the metal center improves the mobility and bio availability of the agent. The studied compound, Cu-o-Va is an antimutagenic, anticarcinogenic agent and exhibits superoxide dismutase (SOD) mimic properties. In particular, the ligand ovanillin, as other related hydroxyaldehydes, has demonstrated antimutagenic and carcinogenesis inhibitory activities and is also an antioxidant agent and scavenger of free radicals. It has been employed in the syntheses of many active poly-functional ligands, including those belonging to the Schiff Bases family, that form many active copper complexes, among other metals.
In this work we report the usefulness of the new BSEC spectroelectrochemical setup proposed to understand the complex electrochemistry of [Cu(o-Va) 2

Instrumentation
A new experimental setup ( Figure 1 and Figure SI1) has been developed to perform BSEC experiments using a mobile slit to locate the desired position of the light beam in the parallel arrangement. The experimental setup used for BSEC measurements to control the motion of the slit are based on previous works [18], where a detailed description of the device can be found. In the present work, the previous liquid/liquid interface is replaced by a solid electrode. The mobile slit device consists of a rigid block supporting the slit and the lenses, attached to the positioner in order to transmit the motorized actuator movement to the whole block. The slit is located between two collimating lenses with fitted optical fibers. Normal arrangement is based on a nearnormal incident reflection spectroelectrochemical configuration (NNIRS) [32,33].
Spectroelectrochemical experiments were carried out using a PGSTAT 302N potentiostat (Eco Chemie B.V) coupled to two QE65000 Spectrometers The last and central part of this setup is the spectroelectrochemical cell ( Figure 1). It consist of a three-electrode system with a glassy carbon disk as working electrode (WE), a platinum wire as auxiliary electrode (AE) and a home-made Ag/AgCl/3M KCl reference electrode (RE). The three-electrode system is placed in a modified commercial quartz spectrophotometric cuvette 110-QS (Hellma) to properly perform simultaneously the experiment in the two spectroscopic configurations. Thus, the bottom part of the quartz cuvette was cut to fill it with the corresponding solution and to place the reflection probe and the reference and auxiliary electrodes. The WE was placed in the cuvette hole usually used to fill the cuvette, sealing the electrode with Teflon tape to avoid any leakage of the solution when the cuvette is placed in an inverted position during a BSEC experiment. Next, a Suba-Seal septa with five drilled holes is used to close the upper part of cuvette and to properly fix the reference and the auxiliary electrode, the reflection probe, the nitrogen inlet and outlet, and to avoid any interference with the light beam in parallel configuration.
All the experiments were performed in a semiinfinite diffusion regime. Actually, this is the most used diffusion regime in electrochemistry. Although thin-layer cells are commonly used in spectroelectrochemistry to obtain formal potentials and number of electrons of redox couples, time resolved spectroelectrochemistry is more useful to follow diffusive processes, providing molecular information on the reactants and/or products involved in the electron transfer process.
Absorbance data from all spectroelectrochemical experiments were calculated taking as reference spectrum the one at the starting potential of each individual experiment.
Mass spectrometry experiments were carried out using a Micromass AutoSpec (WATERS).

Computational methods
The geometry of all the species under study were optimized using the B3LYP hybrid density functional [34,35]. The cc-pVDZ [35] basis set was used for H and C, whereas the aug-cc-pVDZ [35,36] basis set was utilized for O. On the other hand, the cc-pVDZ-PP [37,38] basis set, which includes a pseudopotential to represent inner electrons up to the 2p sub-shell, was used for Cu.
To facilitate the comparison of theoretical results with experimental data, geometry optimizations and calculation of properties were performed including solvent effects (DMSO) through the Polarizable Continuum Model [39,40].
The Hessian matrix of the total energy with respect to the nuclear coordinates of every molecule was calculated at the same level of theory and was diagonalized to verify whether they are local minima or saddle points on the corresponding potential energy surfaces.
The electronic spectra of all the species in DMSO were calculated using the time dependent density functional theory (TDDFT) [41]. All the calculations were carried out with the Gaussian 09 package [42].

Validation
The experimental setup for BSEC measurements ( Figure 1 and Figure SI1) was validated using o-tolidine. This well-known reference system for spectroelectrochemistry was chosen because it exhibits a fast two electron transfer and a large molar absorption coefficient in water, ɛ = 60670 M -1 cm -1 at 438 nm [43].  nm is observed. Figure 3.d shows the VA at 510 nm obtained during the first cycle for both configurations. As can be observed, during the reduction process, the absorbance in A N at 510 nm is significantly higher than the absorbance in A P at the same wavelength although the optical path-length is longer in parallel than in normal configuration. Thus, different processes are observed in the two configurations. These differences can be explained in terms of an electrodeposition process taking place on the electrode surface. Therefore, we can conclude that during the cathodic scan at potentials lower than -0.45 V, the reduction from Cu(II) to Cu(0) is taking place, while the reduction to Cu(I) is not observed because no absorption bands are observed in parallel configuration related to these processes. If the reduction reaction took place only in solution, A P should be higher than A N because of the difference of the optical pathways.
During the anodic scan, the CV (Figure 3. nm, similar to that obtained experimentally, indicating a good correlation between experimental and theoretical data. Finally, it should be mentioned that at the end of the first potential scan, a reduction peak is observed around 0.00 V in the voltammogram, but any change in the spectral signals is observed. Inspecting spectroelectrochemical responses during the second potential cycle a more complete map of the electrochemical reaction can be obtained. Two additional reduction peaks are observed during the second scan in the CV (Figure 3.a, red line). The first one at 0.00 and the second one at -0.30 V, together with the ill-defined peak around -0.50 V, previously observed in the first cycle. During this second cycle, in the backward scan the reoxidation of Cu(0) at +0.05 V is clearly detected together with a post-peak shoulder that could be related to the reduction peak appearing at 0.00 V. Also, the oxidation peak of the o-Va ligand of the copper complex is observed in the two cycles.
Moreover, in the second cycle contour plots of spectral responses (Figures 3.b and 3.c) show some relevant changes compared with the first cycle. The more significant change is appreciated in spectra evolution in normal configuration (Figure 3.b) where a very broad band around 720 nm is observed at potentials lower than 0.00 V. This absorption band disappears at potential higher than 0.00 V. However, this absorption band at 720 nm is not observed in the spectra registered in parallel configuration. Therefore, this broad band peaking at 720 nm has to be ascribed to the generation of Cu(0) on the electrode, with the metal being subsequently redissolved, as can be deduced from the sharp peak in the voltammogram at +0.05 V.
Consequently, from the spectra in normal configuration we can deduce that two In the previous experiment the concentration of Cu-o-Va was too high to observe spectral changes below 475 nm, therefore we decide to decrease this value to 5·10 -4 M.
In addition, the preceding study was performed starting the experiment through the cathodic direction. To determine if this experimental condition controls the spectroelectrochemical behaviour explained above, now we present a similar experiment but scanning first the potential through anodic values. Thus, we carried out a cyclic voltammetry in a DMSO solution 5·10 -4 M Cu-o-Va and 0.1 M TBAPF 6, scanning the potential at 0.01 V s -1 from +0.60 V to +1.20 V and back to -0.70 V, finishing at the starting potential, +0.60 V. Figure 4 shows the three spectroelectrochemical responses obtained during this experiment. The CV (Figure 4.a) in this case only shows the irreversible oxidation peak at +1.10 V, the reduction peak at 0.00 V, another ill-defined peak around -0.50 V and the oxidation peak at +0.05 V. As in the former experiment, the sharp stripping oxidation peak at +0.05 V is overlapped with a less intense diffusive peak. The evolution of the spectra in normal configuration with the potential applied (Figure 4.b) confirms that the copper complex is oxidized at +1.10 V, as can be deduced from the band evolving at 510 nm. Additionally, a narrow and well-defined absorption band emerges at 425 nm at potentials where no electrochemical reaction is taking place. This experimental result must be explained as a charge redistribution within the complex of the oxidized Cu-o-Va complex, as will be explained bellow, that is subsequently reduced at 0.00 V. Contour plot of spectra in parallel configuration (Figure 4.c) confirms this hypothesis because both bands, 510 nm and 425 nm, are also observed indicating that both processes, oxidation of the complex and the following charge redistribution, take place in solution. The main difference appreciated between the two optical responses is observed in the cathodic region. As in the experiment shown in Figure 4, the electrodeposition of Cu(0) is only observed in normal arrangement ( Figure   4.b), yielding a Cu(0) film on the electrode that is redissolved around +0.05 V , as evidenced by the sharp peak in the backward scan of the CV (Figure 4.a).
The reduction peak at 0.00 V is attributed in the bibliography to a quasi-reversible electron transfer process [30] for the redox couple Cu II /Cu I . Joining this assignment to our results, we can suggest that the pristine Cu-o-Va complex can only be reduced directly to Cu(0) (first cycle in Figure 4) that implies the breakdown of the complex.
However, when the Cu-o-Va complex is oxidized, it takes place over the phenolic oxygen of the ligand. The new complex right after, can be reduced from Cu(II) to Cu(I).
The absence of the reduction peak at -0.30 V in this experiment, indicates that the presence of this peak in the second cycle of the CV in the previous experiment (

c). This band decreases
during the reduction of Cu(II) to Cu(I) and increases reversibly at the end of the scan when the initial potential is reached. Therefore, it can be undoubtedly assigned to this reversible process. Moreover, a last BSEC experiment was carried out right after the experiment showed in Figure 5. In this case a cyclic voltammetry between +0.30 V and -0.10 V at 0.005 V s -1 was performed. The comparison between the CV and the DCVA at 285 nm ( Figure 6) evidences that both signals are related exactly with the same process, in this case, the reversible reduction of Cu(II)/Cu(I) complex.
The most intriguing point of these results is related to the new complex generated in solution. As it was explained above, changes in the spectra can be ascribed to a shifting Calculations, including geometry optimization, have been carried out for the oxidized and reduced complex species. Electronic spectra were calculated in both cases considering the redox process with no change in the original geometry of the pristine complex and also considering the optimized geometry. The redox processes can be analyzed in two steps; the first involving the electron transfer at fixed geometry and the second considering the relaxation of the geometry.
In the oxidation process slight changes of geometry are predicted after relaxation of the system. Nevertheless, as a consequence of changes in the charge distribution of the oxidized ligands, some spectral changes were observed in the calculated spectra. In