ACCELERATED ARTICLE pubs.acs.org/ac
Synchrotron Infrared Radiation for Electrochemical External Reflection Spectroscopy: A Case Study Using Ferrocyanide Scott M. Rosendahl,† Ferenc Borondics,‡ Tim E. May,‡ Tor M. Pedersen,‡ and Ian J. Burgess*,† † ‡
Department of Chemistry, University of Saskatchewan, Saskatoon, Saskatchewan, S7N 5C9 Canada Canadian Light Source, Saskatoon, Saskatchewan, S7N 0X4 Canada ABSTRACT: Synchrotron infrared radiation has been successfully coupled through an infrared (IR) microscope to a thin-cavity external reflectance cell to study the diffusion controlled redox of a ferrocyanide solution. Excellent signal-to-noise ratios were achieved even at aperture settings close to the diffraction limit. Comparisons of noise levels as a function of aperture size demonstrate that this can be attributed to the high brilliance of synchrotron radiation relative to a conventional thermal source. Time resolved spectroscopic studies of diffusion controlled redox behavior have been measured and compared to purely electrochemical responses of the thin-cavity cell. Marked differences between the two measurements have been explained by analyzing diffusion in both the axial (linear) and radial dimensions. Whereas both terms contribute to the measured current and charge, only species that originate in the volume element above the electrode and diffuse in the direction perpendicular to the electrode surface are interrogated by IR radiation. Implications for the use of ultramicroelectrodes and synchrotron IR (SIR) to study electrochemical processes in the submillisecond time domain are discussed.
I
nfrared spectroelectrochemistry has been an area of sustained interest among electrochemists for the better part of thirty years. Almost from the method’s inception, two competing technical approaches have been advanced: (i) external reflection spectroscopy (ERS) and (ii) internal reflection spectroscopy (IRS). It is generally agreed among contemporary users that ERS is advantageous as it allows the use of well-defined single crystals but suffers from the disadvantage of restricted mass transport owing to the thin-electrolyte layers required to prevent complete IR attenuation by the solvent. An additional advantage of ERS is the use of photon polarization to differentiate between solution and surface adsorbed species.1 On the other hand, IRS, particularly the surface enhanced infrared absorption spectroscopy (SEIRAS) technique pioneered by Osawa and co-workers,2 offers nearly exclusive sensitivity to surface species without the need of photon polarization. Furthermore, as the incident radiation in SEIRAS does not need to pass through the electrolyte, a SEIRAS electrochemical cell can easily be constructed which does not suffer from impeded mass transfer. SEIRAS advantages are offset by the need to directly modify the internal reflective element with a metallic film, thus preventing the use of well-defined single crystals. Both external and internal reflection IR spectroelectrochemistry typically employ large electrodes to facilitate high photon throughput. Unfortunately, using large electrodes to improve photon throughput compromises the minimum achievable time scale of an electrochemical experiment. Establishment of the desired potential at the electrode|electrolyte interface is determined by the double layer charging time, which in turn is the r 2011 American Chemical Society
product of the electrolyte resistance (R) and the working electrode’s capacitance (C). As the latter is directly proportional to the electrode area, only very slow processes (on the order of seconds) can be meaningfully studied with large electrodes. RC charging times are further slowed in thin-cavity ERS cells by hindered mass transfer which increases the electrolyte resistance.3 Double layer charging restrictions are one of the major reasons that the overwhelming majority of electrochemical IR measurements in external reflectance cells are nondynamic in nature, as they are performed by acquiring spectra only after equilibration of the electrified interface. Examples include subtractively normalized interfacial Fourier-transform IR spectroscopy (SNIFTIRS) and photoelastic modulation IR reflection absorption spectroscopy (PM-IRRAS) which have emerged as the dominant external reflection methods in IR spectroelectrochemistry.4 Nonstatic methods, such as electrochemically modulated infrared spectroscopy (EMIRS),5 use ac modulation of the applied potential to greatly reduce 1/f noise and provide dynamic information. However, with notable exceptions using step-scan interferometers,610 the success of EMIRS in modern FT-IR spectrometers is limited by both slow diffusion rates and large charging times. One of the great successes of SEIRAS has been its elimination of highly resistive electrolyte layers which has permitted the study of kinetic processes on the ∼150 ms time scale.1113 Nevertheless, decreasing the dimension of the Received: January 28, 2011 Accepted: April 12, 2011 Published: April 12, 2011 3632
dx.doi.org/10.1021/ac200250s | Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry
ACCELERATED ARTICLE
concerning the feasibility of the use of small UMEs for future SIREERS studies.
’ EXPERIMENTAL SECTION
Figure 1. Diagram of the in situ spectroelectrochemical cell. The thin cavity electrochemical cell was prepared by compressing a nominally 5 μm thick PTFE gasket between an IR transparent window (CaF2, 1 mm thick and 25 mm diameter) and the main PVC body. The working electrode (WE) and counter electrode (CE) were made from 0.5 mm diameter gold wire, and the quasi-reference electrode (RE) was from 0.5 mm diameter silver wire. The cell was designed to be used as a flowthrough cell allowing the thin cavity to be filled after assembly.
working electrode in an external reflectance thin-cavity cell to micrometer size could, in principle, further decrease the minimum time scale of electrochemical IR experiments by at least an order of magnitude. Sun and co-workers have already made advances in this regard by coupling FT-IR microspectroscopy with relatively large ultramicroelectrodes (UMEs; radius ca. 200 μm).1419 They have reported studies of processes occurring on electrode surfaces on the time scale of tens to hundreds of microseconds.14,15 Importantly, as the size of the electrode decreases, signal-to-noise ratios (S/N) will be increasingly limited by the instrument’s ability to focus large photon intensity onto very small illumination areas and nonconventional sources of IR radiation may be required. The goals of the present study are (1) evaluation of the feasibility of coupling synchrotron generated radiation through an IR microscope into a thin-cavity external reflectance cell and (2) contrasting the time-resolved electrochemical and spectroscopic responses. To the best of our knowledge, the only previous reports of coupling synchrotron IR radiation with electrochemistry are the nondynamic studies of Melendres and co-workers. These authors have primarily used far-IR synchrotron radiation to study processes related to copper oxidation2025 and ion adsorption on gold,26,27 silver,28 and platinum electrodes.25,29 In this work, we have chosen a simple redox probe (ferrocyanide) to dynamically study the diffusion processes occurring at a relatively small electrode (500 μm diameter) in a thin cavity cell. Timeresolved experimental techniques have been employed that allow the use of synchrotron infrared radiation (SIR) for electrochemical external reflectance spectroscopy (SIR-EERS) studies. We demonstrate that SIR outperforms a traditional globar source when electrode dimensions require aperture settings less than 40 μm 40 μm. We also outline how our results have important implications
Reagents and Solutions. Potassium hexacyanoferrate(II) trihydrate (K4Fe(CN)6 3 3H2O, g99.99% trace metals basis) and sodium perchlorate monohydrate (NaClO4 3 H2O, g99.0% trace metals basis) were purchased from Sigma Aldrich and were used as received. All aqueous solutions were prepared from MilliQ (>18.2 MΩ cm1) water. Experimental Cell. The main body of the in situ spectroelectrochemical (SEC) cell (Figure 1) was constructed to fit on the stage of the IR microscope and was made out of PVC plastic for its rigidity and electrical insulating properties. Similar designs for in situ SEC cells have been previously reported.30,31 A thin cavity electrochemical cell was prepared by compressing a nominally 5 μm thick PTFE gasket between an IR transparent window (CaF2, 1 mm thick and 25 mm diameter) and the main body of the cell. A conventional three electrode arrangement was used with the working electrode (WE) and the counter electrode (CE) made from 0.5 mm diameter gold wire (Alfa Aesar, 99.95%) and a quasi-reference electrode (RE) from 0.5 mm diameter silver wire (Alfra Aesar, 99%). The electrodes were held in the main body using an epoxy, and the top (reflecting) surface was polished flat and to a mirror finish using successively finer grade diamond suspensions (Leco Corporation, St. Joseph MI, US) down to 0.5 μm. The cell was designed as a flow-through cell to allow the electrolyte into and out of the thin cavity through two holes placed on extreme edges of the cavity after dry assembly. Using electrochemical impedance spectroscopy, the solution resistance and working electrode capacity in 0.5 M NaClO4 were determined to be 3.4 kΩ and 76 nF, respectively. The time constant of the cell is ∼0.25 ms, and the time to establish the desired potential at the interface is ∼1.25 ms (5RC). From interference fringes obtained in the dry in situ SEC cell and the equation b = n/(2Δυ), where b is the path length, n is the number of fringes, and Δυ is the wavenumber spacing of the fringes, the path length of the thin layer was determined to be 13.7 μm. As the incident radiation converges on the focal point about a mean angle of 30 from the surface normal, the actual cavity thickness is estimated to be 12 μm. The volume of the cell is therefore ∼5 μL. FT-IR Measurements. Fourier transform infrared (FT-IR) spectroscopy measurements were collected using the mid-IR beamline facilities located at the Canadian Light Source (Beamline 01B101, Canadian Light Source, Saskatoon, SK, Canada). The end station consists of a Bruker Optics IFS66 v/S Spectrometer (with Rapid Scan) coupled to a Hyperion 2000 IR Microscope (Bruker Optics, Billerica, MA, USA). Light was focused and collected, in reflectance mode, onto the working electrode using a 36 Schwarzschild objective (NA 0.52) and measured using a narrowband 100 μm mercury cadmium telluride (MCT) (liquid nitrogen cooled) detector. Rapid Scan FT-IR Measurements. The moving mirror in the time-resolved rapid scan experiments was driven at 100 kHz (in relation to the reference HeNe laser wavelength of 632.8 nm) at a resolution of 8 cm1 enabling the collection of one double-sided forward/backward interferogram every 120 ms. Before each time-resolved experiment, a spectrum (consisting of 100 scans) was collected at the reference potential (0.2 V) using the same mirror velocity. This spectrum was then used as the reference spectrum in a subtractively normalized calculation for each 3633
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry interferogram in a given time-resolved experiment. In this manner, any possible spectrometer drift throughout the course of the experiment would be minimized. Each complete rapid-scan experiment lasted ∼6 s, collecting 52 such reference corrected single beam spectra. However, there is an instrumental trade-off between the high mirror velocity and the low signal-to-noise ratio of a given spectrum. This necessarily requires that a given set of time-resolved experiments must be measured and signal averaged numerous times (in this case, 256 replicates for the 20 20 μm aperture size and 1024 for the 8 8 μm aperture) for high S/N results. Electrochemical Measurements. All electrochemical measurements were performed in the previously mentioned in situ SEC cell using a HEKA PG 590 Potentiostat (HEKA, Mahone Bay, NS, Canada). Cyclic voltammetry (CV) and double-step chronocoulometry (CC) experiments were performed using software written in the LabVIEW (National Instruments Corporation, Austin, TX, USA) programming environment. Data acquisition, signal generation, and triggering for the in situ SEC experiments were done using a National Instruments multifunctional Data Acquisition Card (DAQ) PCI 6251 M Series. A second National Instruments DAQ card was used (PCI 6251 X Series) to measure the current transients during the in situ SEC experiments. This second card was necessary to acquire the full 6 s of rapid-scan data as the first DAQ card’s data buffer became full after only 2 s of the complete rapid-scan measurement. To ensure the synchronized start of the potential waveform and the IR measurements, a much higher sampling rate was required on the first card filling its buffer before the measurement was done. A much lower sampling rate was used on the second card to ensure acquisition over the duration of the 6 s current transient. Interfacing Hardware and Software. The major technical challenge in performing these experiments was interfacing the FT-IR spectrometer and the potentiostat for consistent, synchronized measurements. The primary objective was to be able to start the acquisition of the IR spectra in synch with the application of a particular potential waveform by the potentiostat and repeat this process numerous times. This required a series of hardware and software triggering signals to ensure proper timing, communication, and data flow between the instruments. Briefly, a sequence of transistor-transistor logic (TTL) hardware trigger signals were generated and received between the FT-IR spectrometer and a computer controlling the potentiostat (through the LabView software environment, National Instruments). These two TTL signals worked in parallel to gate the start and stop of the acquisition of the rapid-scan FT-IR spectra (and electrochemical data) reproducibly with the application of the potential waveform to the working electrode. This two-way communication ensured that the proper sequence of events could be repeatably achieved. A more rigorous explanation of the instrumentation, including hardware/software interfacing, will be discussed in a future contribution.
’ RESULTS AND DISCUSSION Diffusion Considerations. Before discussing the results of
our SIR-EERS experiments, it is important to first describe the electrochemical response of our spectroelectrochemical cell. The theory of diffusion currents in restricted geometry, thin-layer cells under voltammetric conditions has been treated by several authors.3234 Although solutions for such cells under constant current conditions have been reported by Oglesby et al.,35 Anson and co-workers,36,37 and more recently by Leddy and Zoski,38
ACCELERATED ARTICLE
there does not seem to be an exact analytical solution for the corresponding diffusion controlled chronoamperometric experiment. For a disk electrode of radius, re, and a thin-layer cavity thickness h along the z direction, a rigorous treatment of the problem must include radial and axial diffusion terms (i.e., cylindrical diffusion) " # DCðr, z, tÞ D2 Cðr, z, tÞ 1 DCðr, z, tÞ D2 Cðr, z, tÞ ¼D þ þ Dt Dr 2 r Dr Dz2 ð1Þ with the following initial and boundary conditions Cðr, z, 0Þ ¼ C limr f ¥ Cðr, z, tÞ ¼ C Cðr, 0, tÞ ¼ 0 limz f h
ðr e re , t > 0Þ
DCðr, z, tÞ ¼0 Dz
where D and C*are the diffusion coefficient and the initial concentration of the redox species being consumed. The spectroelectrochemical relevance of this problem has inspired several models which invoke simplifying assumptions to reduce the dimensionality of the system.3941 For example, Micka et al. made the assumption that, for sufficiently large enough cavity thicknesses, radial diffusion can be neglected entirely and only linear diffusion along the axis (henceforth denoted as axial diffusion) is of consequence. The corresponding mass-transfer limited, chronoamperometric response is39 " # 2nFADC ¥ 1 2 π 2 iðtÞ ¼ exp n þ Dt ð2Þ h 2 h n¼0
∑
where A = πre2 is the electrode area for a perfectly smooth disk, F is the Faraday constant, and n is the number of electrons transferred. A different expression under the same boundary conditions and assumptions was provided by Oglesby et al.35 but seems to lead to essentially identical simulated current responses. For completeness, we note that Oglesby et al.’s equations are more robust for calculating concentration profiles at short times. A complementary scenario arises if mass transfer in the axial direction for r e re is assumed sufficiently rapid such that there is nearly immediate exhaustion of the material initially in the electrolyte volume above the electrode. This would be especially true for very small values of h and large values of D. Under such conditions, the problem reduces to semi-infinite radial diffusion to a cylinder of area 2hπre, and the approximation of Szabo et al.42 is applicable iðtÞ ¼
"
# 1=2 2exp½0:05ðπτÞ 1 2hπnFDC þ lnð5:2945 þ 0:7493τ1=2 Þ ðπτÞ1=2 ð3Þ where τ = 4Dt/re2. For small values of τ (short times), only the first term in eq 3 is significant, the exponential approaches unity, and a modified Cottrell-like relationship results 3634
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry
ACCELERATED ARTICLE
Figure 2. Simulated diffusion-controlled responses in a thin-cavity, spectroelectrochemical cell. Current densities (main figure) have been plotted for finite-volume linear diffusion (solid black line) and semi-infinite radial diffusion (dotted red line). Corresponding charge densities are shown in the inset. The time domain has been transformed for both sets of transients to allow easy comparison with Cottrell-like behavior. Simulation parameters: C* = 1 mM, h = 20 μm, re = 0.025 cm, D = 6.4 104 cm2 s1.
iðtÞ ¼
2hre π1=2 nFD1=2 C t 1=2
ð4Þ
On the other hand, at sufficiently long enough times, the first term approaches zero and a quasi-steady state current is expected 4hπnFDC ð5Þ iqss ¼ ln τ Overall, currents arising from axial diffusion will dominate at short times as long as the radius of the disk and the depletion layer thickness above the electrode are such that the area over which radial diffusion occurs is much smaller compared to the actual area of the electrode. However, regardless of disk radius, once the depletion layer normal to the electrode surface extends across the full thickness of the cavity, a quasi-steady state current arising solely from radial diffusion will be present at sufficiently long times. Thus, in the absence of an exact solution to the diffusion problem at hand, a reasonable estimation is to assume that the measured current is largely dictated by eq 2 at short times and solely by eq 3 at long times. Figure 2 provides simulated current (main figure) and charge (inset) densities for C* = 1 mM, h = 20 μm, re = 0.025 cm, and D = 6.4 104 cm2 s1 using eqs 2 and 3. Under these conditions, both currents are initially linearly dependent on 1/t1/2 (but with different slopes) and cross after ∼0.8 s as the axial current quenches. The axial diffusion charge versus t1/2 curve is linear (Cottrell-like) at early times but plateaus after less than a second as the electroactive species initially in the layer is completely consumed. In comparison, the charge curve for radially diffusing species continues to grow with increasing time. The validity of our assumption that eq 2 determines the current (and consequently the amount of electrogenerated species) at short times must now be considered for different sized electrodes. With a fixed cavity thickness, inspection of eqs 2 and 4 reveals that the diffusion current for the former is proportional to the square of the electrode radius while for the latter it is only linearly dependent on re. Axial diffusion is therefore dominant for large electrodes and Micka et al.’s result will describe the current behavior over the majority of the experimental time scale when re is significantly larger than the cavity thickness. However, as the electrode size
Figure 3. Experimental electrochemistry of 1.13 mM Fe(CN)64 plus 0.5 M NaClO4 for a 0.5 mm diameter gold electrode in the thin-cavity cell. Cyclic voltammetry is shown in inset (a) whereas the double potential step sequence used is shown in inset (b). The main body shows the charge transient (green line for the positive step, blue line for the return step) obtained from numeric integration of the measured current during the double-step.
decreases, the radial contribution becomes more significant at increasingly shorter times. It is important to note that the overwhelming majority of species electrogenerated at the perimeter of the electrode as a result of radial mass transport will diffuse away from the electrode into the semi-infinite space parallel to the surface of the disk. An important consequence is that, in a thin cavity SIR-EERS electrolysis experiment, the spectroscopic signal should report primarily only on the concentration of species electrochemically formed from the material initially in the volume element defined by the electrode area and the cavity thickness. For relatively large electrodes, this will correlate with the measured electrochemical signal at short times but decreasing the electrode radius will lead to increasingly greater deviation between the IR signal and the cumulative charge passed. If the size of the electrode is reduced to the dimension of an ultramicroelectrode (re < ∼25 μm), the mass-transport problem is identical to diffusion to a scanning electrochemical microscope (SECM) tip above an insulating substrate.43 This diffusion problem has been treated with both numerical simulation methods43,44 and approximations similar in nature to those we have described above.4446 Electrochemical Results. Inset (a) in Figure 3 shows the cyclic voltammogram (scan rate 50 mV s1) for 1.13 mM Fe(CN)64 plus 0.5 M NaClO4 supporting electrolyte within our spectroelectrochemical cell. An apparent formal potential of ∼0.15 V is observed, and both the reduction and oxidation peaks display nonsteady state, diffusional tailing. Prior to our SIR-EERS measurements, the cell was filled with a fresh volume of electrolyte solution, the potential was held at 0.2 V for 30 s, and current transients were recorded during the double potential step depicted by inset (b) in Figure 3. The integrated current transients for the forward step (0.2 to 0.5 V: green line) as well as the backward step (0.5 to 0.2 V: blue line) are shown as a function of time in the main body of Figure 3. Both halves of the transient show a rapid initial change in the charge followed by a much more slowly changing tail which is consistent with the argument above that axial diffusion dominates at short times and radial diffusion at long times. Interestingly, there is a net positive charge passed in the total transient, and the tail is appreciably flatter in the backward step compared to that of the forward step. 3635
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry
Figure 4. Comparison of the noise levels in the SIR-EERS cell using synchrotron produced IR radiation (red lines) and globar (black lines) for (a) 40 μm 40 μm, (b) 20 μm 20 μm, and (c) 5 μm 5 μm aperture settings. These “100% line” spectra were determined from the ratio of two successive, single interferograms (1 cm1 resolution). The cell contained an H2O solution of 0.5 M NaClO4.
This can be rationalized by considering the initial conditions before each component of the experiment. Prior to the step from 0.2 V, only Fe(CN)64 is present in the cell and it is homogeneously distributed in both the axial and radial directions. Upon stepping the potential to 0.5 V, only the Fe(CN)64 above the disk is initially converted to Fe(CN)63 and the oxidized product will largely be retained within this volume. Radial diffusion will begin to contribute more significantly as the layer becomes depleted of Fe(CN)64, but as mentioned above, Fe(CN)63 generated at the electrode perimeter will diffuse away from the electrode surface. Upon stepping the potential back to 0.2 V, a significant portion of this material will not be recollected at the electrode within the measured time scale of the transient. Semi-infinite radial diffusion conditions, as well as the 30 s potential hold at 0.2 V prior to the next double step sequence prevent any complications from the slow accumulation of ferricyanide in the electrolyte in the vicinity of the working electrode. In fact, in our IR measurements (vide infra), we performed over 1000 consecutive iterations of this potential step profile (without flowing new electrolyte into the cell) and the first and last transients are identical to one another within experimental error.
ACCELERATED ARTICLE
Synchrotron Advantage. The subject of synchrotron versus thermal (globar) source has been addressed elsewhere,47 but it is prudent to provide a brief overview for the electrochemical community. Flux is a common measure of source energy output and is defined as the photon density per source unit area per unit time. A typical commercial globar actually has higher flux than SIR radiation produced under normal synchrotron operating conditions. However, focusing thermal source radiation leads to significant losses because a globar emits photons in all directions whereas the highly directed SIR photons do not suffer such high magnitude losses and lead to 2 to 3 orders of magnitude higher brilliance (flux density). Thus, it is only when IR radiation needs to be focused onto small spatial regions that SIR is truly advantageous. This is illustrated in Figure 4 where we provide 100% lines collected from our electrolyte-filled cell using SIR (solid red lines) and the spectrometer’s thermal source (black lines) for different microscope aperture settings. These spectra are the ratio of two successive single interferograms measured at 1 cm1 resolution. As shown in panel (a), the globar and SIR provide comparable noise levels for a 40 μm 40 μm square aperture for frequencies above 2400 cm1. Below this frequency, the SIR signal is actually nosier than the globar due to absorption from the beamline’s diamond window. At 20 μm 20 μm aperture (panel b), the SIR and globar noise levels are very comparable between 1800 and 2400 cm1 but favor the SIR outside the region of diamond interference. Most impressively, panel (c) shows that, for a 5 μm 5 μm aperture, SIR noise is significantly smaller across the entire region of interest. Clearly, SIR noise levels become increasingly better relative to those of the globar as the aperture size is decreased. Unfortunately, for this particular case study, it should be noted that the attenuation of the SIR signal due to the diamond window absorption overlaps with the CtN stretches of both forms of our redox couple. Nevertheless, even after these losses, there is still about 2-fold smaller noise levels when SIR is used to study the electrochemistry of ferri/ferrocyanide at the chosen aperture setting of 8 μm 8 μm. A more judicious choice of redox couple would see an even larger SIR advantage. SIR-EERS Results. It is well-established that the ferri/ferrocyanide redox couple is particularly well suited for IR spectroelectrochemistry7,9,10,48,49 owing to the clear spectral shift in the cyanide stretches of the reduced and oxidized forms which are both frequency resolved from water absorption bands. In Figure 5, we present a subtractively normalized curve, ΔS/S = (Ssample Sref)/Sref, for the oxidation of 1.13 mM ferrocyanide in 0.5 M NaClO4. The single beam reference spectrum was obtained by holding the potential at 0.2 V, and the sample spectrum was obtained at þ0.5 V after waiting for 60 s. The positive peak at 2038 cm1 is assigned to the loss of ferrocyanide whereas the smaller, downward band at 2115 cm1 is assigned to the formation of ferricyanide in the thin layer. As no other peaks or shoulders were observed even after several hours of repetitive potential steps, there is no evidence of the formation of adsorbed polymeric hexacyanoferrate (HCF) complex.10,48 Although our surface sensitivity is poor due to the small angle of IR incidence, it is unlikely that adsorbed HCF is formed below detectable levels because of the relatively mild potential perturbations used.10 This was confirmed by the consistency in cyclic voltammograms run before and after our long duration time-resolved studies. Rapid scanning, time-resolved SIR-EERS spectroelectrochemistry was performed using freshly injected ferrocyanide solution and the same potential perturbation depicted in Figure 3b. With 8 cm1 resolution, a double-sided interferogram could be 3636
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry
ACCELERATED ARTICLE
Figure 5. Subtractively normalized IR signal (256 scans) for ferrocyanide oxidation. The sample spectrum was measured after holding the potential at þ0.5 V for 30 s. The reference potential was 0.2 V. The strong upward band at 2040 cm1 corresponds to the loss of ferrocyanide whereas the weaker, downward band at 2115 cm1 is attributable to the electroformation of ferricyanide ions.
measured every 120 ms which is sufficient to follow the masstransfer controlled redox process. Although our working electrode (500 μm diameter) was much larger than a UME, we reduced the aperture setting to 8 8 μm to mimic the magnitude of the IR signal that could be collected from an ultramicroelectrode. To offset the decrease in photon throughput caused by decreasing the illumination area, we coadded a relatively large number of timeresolved spectra to achieve high signal-to-noise ratios. This was accomplished by repeating the double potential step 1024 times. The S/N of the ferrocyanide peak was ∼60 at the end of each half of the double step which was comparable to the S/N level after the coaddition of 256 interferograms collected with a 20 20 μm aperture size. The time-resolved IR absorbance data are presented in Figure 6a. The ordinate shows the change in absorbance relative to the IR signal after holding the potential at 0.2 V for 30 s. There is a rapid change in the IR signal for both forms of the redox species upon stepping to the more positive potential. However, in less than one second, both signals reach a plateau and remain invariant in time until the return step to 0.2 V at which point they return to their prestep values. The last point is particularly important as it reveals that the IR radiation only interrogates ions that remain close enough to the electrode to be oxidized and then rereduced within the time scale of the experiment. Compared to the experimental charge transients, there is no evidence of a longer-time tail which we have attributed to the contribution of radial diffusion to the overall electrochemical signal. In fact, the shape of the ferricyanide optical transient in Figure 6a is qualitatively identical to the charge transients simulated for pure axial diffusion. To further illustrate this point, we wished to quantitatively compare an axial diffusion simulation with our SIR-EERS experimental results. In order to do so, the molar absorptivity, ε, of both ferrocyanide and ferricyanide was first determined by measuring the IR absorbance of solutions of known Fe(CN)63/4 concentration in our spectroelectrochemical cell. The determined values of ε for ferrocyanide and ferricyanide were 3.9 103 M1 cm1 and 8.8 102 M1 cm1, respectively. While the former is very close to the value reported by Drew, the latter differs by roughly 40%.50 The cause of this discrepancy is unclear. The optical data in Figure 6a was converted to the change in species concentration using Beer’s law, our measured molar absortivities, and the path length of our cell. A
Figure 6. (a) Time-resolved, changes in IR absorbance of ferrocyanide (9) and ferricyanide (red O) during the double potential step sequence shown in Figure 3b. Each point represents the results of 1024 coadded scans at 8 cm1 resolution. (b) Experimental concentration changes of ferro/ferricyanide (points) extracted from the forward step data and calculated concentration changes (lines) from simulated charge transients.
charge transient for ferrocyanide oxidation was simulated using the integrated form of eq 2, with h = 12 μm, C* = 1.13 mM, re = 0.5 mm, and D = 6.4 106 cm2 s1.51 Charges were converted into concentrations of ferrocyanide and ferricyanide using the Faraday constant and dividing by the volume of the cavity (hπre2). Figure 6b reveals excellent agreement between the two methods (charge simulation and SIR-EERS experiment) confirming that the SIR-EERS technique is insensitive to species arriving at the electrode via radial diffusion. This last point has important implications for future studies using IR microspectroscopy and ultramicroelectrodes. First, the discrepancy between electrochemical and spectroscopic data will become much more greatly pronounced with decreasing electrode dimension. Second, even though much faster time scales become accessible with UME-based SIR-EERS, the electrogenerated spectroscopic signal will always be limited by the amount of species formed from axial diffusion within the period of the potential perturbation. For linear diffusion, the thickness of the depletion layer, zd, after 1 μs can be estimated using Zd = (4Dt)1/2 to be on the order of 50 nm. In other words, assuming a gap thickness of 10 μm, less than 0.5% of the optical path propagates space that contains analytical information. Clearly, the high brilliance of SIR will be greatly advantageous in offsetting diminished signal levels. One must also consider that the minimum time resolution of rapid-scan FT-IR instruments is on the order of 10 ms. Kinetic studies using ultramicroelectrodes and dc potential steps will require much faster spectroscopic time resolution which is achievable with step/scan interferometers. However, the discrete 3637
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry wavenumber sampling inherent to the step/scan technique when coupled with weak signal levels means tens of thousands of repeated potential steps and very long overall experiment times would be required. While feasible, such studies would be limited to very stable and reversible electrochemical processes. As an alternative to using static dc potential steps, it should be noted that decreasing the time constant of the cell to 0.1 ms or less will allow the use of ac potential perturbations at frequencies 10 times greater than the highest scanning frequency of an FT-IR instrument (∼1 kHz)52 without requiring step/scan interferometry. Finally, in the limit of extremely fast potential perturbations, the electrochemical response of the system becomes increasingly isolated to species on or very near the electrode surface. Sun and co-workers have been successful in the use of Pt and Ru ultramicroelectrodes that exhibit surface enhancement to study bound carbon monoxide with conventional-sourced microspectroscopy.1419
’ SUMMARY AND CONCLUSIONS This work has provided a basis for the use of synchrotron infrared radiation to study electrochemical kinetic processes. Using a ultramicroelectrode, we have shown that SIR radiation can be successfully coupled with an external reflectance electrochemical cell. Excellent signal-to-noise ratios have been obtained from time-resolved measurements and an illumination area less than 10 μm10 μm. Whereas purely electrochemical responses in the thin-cavity cell include contributions from both axial and radial diffusion, we have demonstrated that the IR spectroscopic signal arises only from species originating in the thin-cavity between the electrode surface and the IR window. Simulations of linear diffusion in a finite volume element are in excellent agreement with the time dependence of the experimentally measured IR absorbance. Furthermore, for aperture settings less than 40 μm 40 μm, SIR provides significantly higher S/N compared to a conventional thermal source. Optimization of conditions, particularly using different redox couples that absorb at frequencies removed from beamline interferences, would result in further improvement. Although we have not used an ultramicroelectrode in the current study, we have discussed how coupling UMEs with high SIR brilliance could be especially advantageous for studying very fast electrochemical processes. In future studies, both the dimension of the electrode and the experimental time scale will be decreased. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT This work was supported by a grant from the Natural Sciences and Engineering Research Council of Canada (NSERC). S.M.R. acknowledges NSERC for graduate funding through a postgraduate scholarship (PGS-D). The authors are grateful to Luca Quaroni for initial discussions regarding this work. Research described in this paper was performed at the Mid-IR beamline of the Canadian Light Source, which is supported by the Natural Sciences and Engineering Research Council of Canada, the National Research Council Canada, the Canadian Institutes of Health Research, the Province of Saskatchewan, Western Economic Diversification Canada, and the University of Saskatchewan.
ACCELERATED ARTICLE
’ REFERENCES (1) Foley, J. K.; Pons, S. Anal. Chem. 1985, 57 (8), 945A–956A. (2) Osawa, M. In Diffraction and Spectroscopic Methods in Electrochemistry; Alkire, R. C., Kolb, D., Ross, P. N., Lipkowski, J., Eds.; WileyVCH: Weinheim, 2006; pp 269314. (3) Nichols, R. J. IR . In Adsorption of Molecules at Metal Electrodes; Lipkowski, J., Ross, P. N., Eds.; VCH: New York, 1992. (4) Zamlynny, V.; Lipkowski, J. In Diffraction and Spectroscopic Methods in Electrochemistry; Alkire, R. C., Kolb, D. M., Lipkowski, J., Ross, P. N., Eds.; Wiley-VCH: Weinheim, 2006; Vol. 9, pp 315376. (5) Bewick, A.; Kunimatsu, K.; Pons, B. S.; Russell, J. W. J. Electroanal. Chem. 1984, 160 (12), 47–61. (6) Ataka, K.; Hara, Y.; Osawa, M. J. Electroanal. Chem. 1999, 473 (1,2), 34–42. (7) Brevnov, D. A.; Hutter, E.; Fendler, J. H. Appl. Spectrosc. 2004, 58 (2), 184–192. (8) Budevska, B. O.; Griffiths, P. R. Anal. Chem. 1993, 65 (21), 2963–2971. (9) Pharr, C. M.; Griffiths, P. R. Anal. Chem. 1997, 69 (22), 4665–4672. (10) Pharr, C. M.; Griffiths, P. R. Anal. Chem. 1997, 69 (22), 4673–4679. (11) Ataka, K.; Nishina, G.; Cai, W. B.; Sun, S. G.; Osawa, M. Electrochem. Commun. 2000, 2 (6), 417–421. (12) Pronkin, S.; Wandlowski, T. J. Electroanal. Chem. 2003, 550551, 131–147. (13) Samjeske, G.; Osawa, M. Angew. Chem., Int. Ed. 2005, 44 (35), 5694–5698. (14) Zhou, Z. Y.; Sun, S. G. Electrochim. Acta 2005, 50 (2526), 5163–5171. (15) Zhou, Z. Y.; Lin, S. C.; Chen, S. P.; Sun, S. G. Electrochem. Commun. 2005, 7 (5), 490–495. (16) Zhou, Z. Y.; Tian, N.; Chen, Y. J.; Chen, S. P.; Sun, S. G. J. Electroanal. Chem. 2004, 573 (1), 111–119. (17) Chen, Y. J.; Sun, S. G.; Chen, S. P.; Li, J. T.; Gong, H. Langmuir 2004, 20 (23), 9920–9925. (18) Gong, H.; Sun, S. G.; Chen, Y. J.; Chen, S. P. J. Phys. Chem. B 2004, 108 (31), 11575–11584. (19) Gong, H.; Sun, S. G.; Li, J. T.; Chen, Y. J.; Chen, S. P. Electrochim. Acta 2003, 48 (2022), 2933–2942. (20) Hahn, F.; Melendres, C. A. J. Synchrotron Radiat. 2010, 17 (1), 81–85. (21) Hahn, F.; Mathis, Y. L.; Bonnefont, A.; Maillard, F.; Melendres, C. A. Infrared Phys. Technol. 2008, 51 (5), 446–449. (22) Hahn, F.; Mathis, Y. L.; Bonnefont, A.; Maillard, F.; Melendres, C. A. J. Synchrotron Radiat. 2007, 14 (5), 446–448. (23) Melendres, C. A.; Bowmaker, G. A.; Leger, J. M.; Beden, B. J. Electroanal. Chem. 1998, 449 (12), 215–218. (24) Melendres, C. A. Proc. - Electrochem. Soc. 1998, 97 (26), 913–924. (25) Melendres, C. A.; Bowmaker, G. A.; Leger, J. M.; Beden, B. Proc. - Electrochem. Soc. 1996, 96 (18), 280–291. (26) Melendres, C. A.; Hahn, F.; Bowmaker, G. A. Electrochim. Acta 2000, 46 (1), 9–13. (27) Melendres, C. A.; Hahn, F. J. Electroanal. Chem. 1999, 463 (2), 258–261. (28) Bowmaker, G. A.; Leger, J. M.; Le Rille, A.; Melendres, C. A.; Tadjeddine, A. J. Chem. Soc., Faraday Trans. 1998, 94 (9), 1309–1313. (29) Melendres, C. A.; Beden, B.; Bowmaker, G. A. J. Electroanal. Chem. 1995, 383 (12), 191–193. (30) Borg, S. J.; Best, S. P. J. Electroanal. Chem. 2002, 535 (12), 57–64. (31) Salsman, J. C.; Kubiak, C. P. In Spectroelectrochemistry; Kaim, W., Klein, A., Eds.; Royal Society of Chemistry: Cambridge, 2008; pp 123144. (32) Hubbard, A. T.; Anson, F. C. Anal. Chem. 1966, 38 (1), 58–61. (33) Keller, H. E.; Reinmuth, W. H. Anal. Chem. 1972, 44 (3), 434–442. 3638
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
Analytical Chemistry
ACCELERATED ARTICLE
(34) Aoki, K.; Tokuda, K.; Matsuda, H. J. Electroanal. Chem. 1983, 146 (2), 417–424. (35) Oglesby, D. M.; Omang, S. H.; Reilley, C. N. Anal. Chem. 1965, 37 (11), 1312–1316. (36) Christensen, C. R.; Anson, F. C. Anal. Chem. 1963, 35 (2), 205–209. (37) Hubbard, A. T.; Anson, F. C. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1970; p 129. (38) Leddy, J.; Zoski, C. G. J. Electroanal. Chem. 2003, 543 (1), 13–22. (39) Micka, K.; Kratochvilova, K.; Klíma, J. Electrochim. Acta 1997, 42 (6), 1005–1010. (40) Rousar, I.; Anastasijevic, N. A. Electrochim. Acta 1988, 33 (8), 1157–1160. (41) Anastasijevic, N. A.; Rousar, I. Electrochim. Acta 1989, 34 (6), 887–888. (42) Szabo, A.; Cope, D. K.; Tallman, D. E.; Kovach, P. M.; Wightman, R. M. J. Electroanal. Chem. Interfacial Electrochem. 1987, 217 (2), 417–423. (43) Mirkin, M. V. In Scanning Electrochemical Microscopy; Bard, A. J., Mirkin, M. V., Eds.; Marcel Dekker: New York, 2001; pp 145200. (44) Bard, A. J.; Denuault, G.; Friesner, R. A.; Dornblaser, B. C.; Tuckerman, L. S. Anal. Chem. 1991, 63 (13), 1282–1288. (45) Gabrielli, C.; Keddam, M.; Portail, N.; Rousseau, P.; Takenouti, H.; Vivier, V. J. Phys. Chem. B 2006, 110 (41), 20478–20485. (46) Remita, E.; Boughrara, D.; Tribollet, B.; Vivier, V.; Sutter, E.; Ropital, F.; Kittel, J. J. Phys. Chem. C 2008, 112 (12), 4626–4634. (47) Miller, L. M.; Smith, R. J. Vib. Spectrosc. 2005, 38 (12), 237–240. (48) Korzeniewski, C.; Severson, M. W.; Schmidt, P. P.; Pons, S.; Fleischmann, M. J. Phys. Chem. 1987, 91 (22), 5568–5573. (49) Pons, S. J. Electroanal. Chem. 1984, 160 (12), 369–376. (50) Drew, D. M. Anal. Chem. 1973, 45 (14), 2423–2424. (51) Bard, A. J. Electrochemical methods: fundamentals and applications; Bard, A. J., Faulkner, L. R., Eds.; Wiley: New York, 1980. (52) Nafie, L. A.; Vidrine, V. W. In Fourier transform spectroscopy; Ferraro, J. R., Basile, L. J., Eds.; Academic Press: Orlando, 1982; Vol. 3, pp 83123.
3639
dx.doi.org/10.1021/ac200250s |Anal. Chem. 2011, 83, 3632–3639
