Electrochemical ATR-SEIRAS Using Low-Cost ... - ACS Publications

Article Cite This: Anal. Chem. 2017, 89, 11818-11824

pubs.acs.org/ac

Electrochemical ATR-SEIRAS Using Low-Cost, Micromachined Si Wafers Tyler A. Morhart, Bipinlal Unni, Michael J. Lardner, and Ian J. Burgess* Department of Chemistry, University of Saskatchewan, Saskatoon, Saskatchewan S7N 5C9, Canada

Anal. Chem. 2017.89:11818-11824. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/22/19. For personal use only.

S Supporting Information *

ABSTRACT: Thin, micromachined Si wafers, designed as internal reflection elements (IREs) for attenuated total reflectance infrared spectroscopy, are adapted to serve as substrates for electrochemical ATR surface enhanced infrared absorption spectroscopy (ATR-SEIRAS). The 500 μm thick wafer IREs with groove angles of 35° are significantly more transparent at long mid-IR wavelengths as compared to conventional large Si hemisphere IREs. The appeal of greater transparency is mitigated by smaller optical throughput at larger grazing angles and steeper angles of incidence at the reflecting plane that reduce the enhancement factor. Through use of the potential dependent adsorption of 4-methoxypyridine (MOP) as a test system, the microgroove IRE is shown to provide relatively strong electrochemical ATR-SEIRAS responses when the angle of incident radiation is between 50 and 55°, corresponding to refracted angles through the crystal of ∼40°. The higher than expected enhancement is attributed to attenuation of the reflection loss of p-polarized light and multiple reflections within the wafer-based IRE. The micromachined IREs are shown to outperform a 25 mm radius hemisphere in terms of S/N at wavenumbers less than ca. 1400 cm−1 despite the weaker signal enhancement derived from the steeper angle incident on the IRE/sample interface. The high optical transparency of the new IREs allows the spectral observation of displaced water libration bands at ca. 730 cm−1 upon solvent replacement by adsorbed MOP. The results are highly encouraging for the further development of low-cost, Si wafer-based IREs for electrochemical ATR-SEIRAS applications.

E

generated when radiation undergoes total internal reflection at the IRE/solution interface. The strength of the coupling is dependent on the angle of incidence through the IRE, and maximum electric field enhancement values at ∼80° have been calculated by Osawa.9 In principle, other high index materials with even greater IR transparency such as ZnSe10,11 and Ge10−14 can be used as ATR-SEIRAS IREs, but in practice, silicon hemispheres and hemicylinders are far more favored for electrochemical applications due to the wide pH stability of Si and its compatibility with metal derivatization using wet chemical techniques.15 More significantly, metal films exhibit very poor wetting and adhesion on ZnSe and Ge surfaces and have a greater tendency to delaminate during electrochemical experiments in comparison to Si. A significant drawback of silicon IREs is decreased IR transparency below ∼1400 cm−1, which prevents high throughput measurements in the low energy region of the mid infrared spectrum. Efforts to mitigate the poor transparency of Si at long wavelengths have been reported for multiple reflection ATR applications and rely on the use of thin

lectrochemical attenuated total reflectance surface enhanced infrared absorption spectroscopy (ATR-SEIRAS), as pioneered by Osawa and co-workers, is a powerful surfacesensitive tool for probing molecular films at electrified interfaces.1 In ATR-SEIRAS, the deposition of several tens of nanometers thick metallic films on high refractive index infrared prisms results in an enhancement in the infrared spectra of molecules adsorbed, or in the very near vicinity, of the metal surface. Much like surface-enhanced Raman spectroscopy (SERS), the enhancement is typically described in terms of the excitation of broad-band, surface plasmon polaritons (SPPs) in aspherical metal nanoparticles in an effective medium. 2−6 In an attenuated total reflectance (ATR) configuration, incident radiative electromagnetic modes are resonant with SPPs and generate large electric fields at the metal−solution interface.7 The increased local electric field intensity enhances the IR absorption of adsorbed molecules. Both the optical geometry and the characteristics of the metal film play critical roles in SEIRAS. Strongly enhancing gold films can be produced by modifying the surface of internal reflection elements (IREs) by either electroless deposition or metal evaporation methods, and they consist of anisotropic islands supported by a thin continuous, metallic film.8 Similar to surface plasmon resonance spectroscopy, strong excitation of SPPs in the metal film is achieved by the evanescent wave © 2017 American Chemical Society

Received: August 28, 2017 Accepted: October 11, 2017 Published: October 11, 2017 11818

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824

Article

Analytical Chemistry (ca. 500 μm) micromachined16−19 or cleaved20 silicon wafers. Thin Si wafers have also been employed for single-reflection electrochemical ATR-SEIRAS studies. For example, Shao et al.21 used a ZnSe prism to couple ATR radiation to a thin, platinum-coated silicon wafer to study the generation of superoxide anions during oxygen reduction. Xue et al.15 improved upon this approach by sandwiching an ultrathin water layer between a ZnSe prism and a 200 μm thick Si wafer to improve optical coupling. Nevertheless, the number of subsequent reports utilizing either of these approaches is very small,22,23 which is likely due to the difficulty in achieving uniform contact between the ZnSe and the Si wafer. Through use of the equations developed by Ulrich for the theory of prism-film coupling in thin-film waveguides,24,25 it is evident that the transmission into the Si wafer at an angle of incidence of 70° is only 50% for an air gap thickness of 100 nm. The difficulty in effecting a uniform gap of several tens of nanometers over large areas (several cm2) is a well-documented problem that afflicts the Otto configuration26 in surface plasmon resonance spectroscopy. Furthermore, large refractive index mismatches between the IRE and the Si wafer results in significant refraction of the incident radiation away from the Sisolution interface. A recent alternative to multiple reflection silicon wafers are the microstructured single-reflection elements reported by Schumacher et al.27 These substrates, referred to herein as μ-groove IREs, are formed through anisotropic etching of a Si (110) wafer, providing an array of v-shaped grooves with Si {111} sidewalls. As well as decreasing the path length, μ-groove IREs do not require the stringent focusing and collimation needed to couple IR light into a multiple reflection Si IRE.27 Herein, we evaluate μ-groove IREs optimized for ATR-FTIR spectroscopy as platforms for electrochemical ATRSEIRAS. We report the effect of angle of incidence on SEIRAS enhancement and compare the performance of the μ-groove IREs to a large Si hemisphere typical of commonly used IREs in electrochemical ATR-SEIRAS experiments.

SEIRAS experiments were performed using an Ag/AgCl (saturated KCl) reference electrode. 4-Methoxypyridine (97%, Aldrich) was used as received. Au Film Cleaning. The Au film electrode was cleaned electrochemically in 50 mM KClO4 by cycling the potential at 50 mV s−1 vs an Ag pseudo reference electrode. An Ag wire pseudoreference electrode was only used during electropolishing to minimize chloride-induced Au dissolution during the large potential excursions required for cleaning. The film was initially scanned from the open circuit potential (ocp, ∼−200 mV vs Ag) between −200 and +200 mV vs Ag for three cycles. The potential window was widened in 100 mV increments after every three cycles. Finally, the potential was scanned between −400 to +900 mV for five cycles. The high potential limit represents the Au oxidation peak; mild excursions into electrochemical oxidation and subsequent reduction of the Au clean and texture the film to increase SEIRAS enhancement. However, applying potentials beyond the peak can delaminate the Au layer, so care must be taken in cleaning. The same electropolishing procedure was used for all substrates as enhancement factors are empirically found to be dependent on the extent of potential cycling at large potentials. MOP Electrochemistry. An aliquot of 0.1 M MOP was spiked into the spectroelectrochemical cell containing deaerated 50 mM KClO4 such that the final MOP concentration was 0.1 mM. The solution was sparged with Ar to ensure mixing, and the experiments were performed under a continuous blanket of Ar. The potential was scanned to +200 mV vs Ag/AgCl and held for 1 min to ensure complete adsorption. Then, the sample spectrum was collected. The potential was then scanned to −700 mV to completely desorb the MOP, and a reference spectrum was collected immediately. This was repeated for all angles. FTIR Measurements. Infrared measurements were performed on a Bruker Vertex 70 FTIR spectrometer using a globar and an MCT/A detector. A Pike Technologies VeeMAX III ATR accessory provided control over angle of incidence. All spectra using the μ-groove IREs were collected at 4 cm−1 resolution and 40 kHz scanning velocity with 2 mm aperture.

■

EXPERIMENTAL SECTION Au Layer Preparation. The IRE (25 mm diameter Si hemisphere, ISP Optics, or μ-groove wafer, IRUBIS GmbH (formerly ATR-elements.com)) was polished for 20 min each with 3 and 0.5 μm diamond grit polish. The IRE was rinsed with ultrapure water before, between, and after polishing steps and then dried under flowing Ar. Au films were prepared by sputtering in a Denton Desk Vacuum IV as described elsewhere.28 Effort was taken to ensure the gold film deposition procedure was the same for both types of IREs. The procedure used here yields gold layers of roughly 20 nm. The excellent suitability of the Au film and the ATR-SEIRAS cell for electrochemical purposes is evidenced by the high-quality voltammetry recorded in 10 mM K3Fe(CN)6 in 100 mM NaF (see Figure S1). Electrochemistry. The ATR optic was assembled into a home-built spectroelectrochemical cell.28 Minor modifications were made to the cell to adapt it to the μ-groove IRE. Electrical contact was made through Au-coated spring-loaded pins (MillMax Mfg Corp). An Autolab PGSTAT302N potentiostat provided potential control. The electrolyte was 50 mM KClO4 (Aldrich) in ultrapure water. KClO4 was recrystallized before use and stored under vacuum. Oxygen was purged from the electrolyte solution by Ar bubbling for at least 20 min prior to experiments; an Ar blanket was maintained over the solution throughout. The counter electrode was a coiled Au wire. All

■

RESULTS AND DISCUSSION μ-Groove IREs for ATR-FTIR Spectroscopy. As detailed by Osawa,1,8 the evanescent wave produced by total internal reflection at the reflecting interface of an ATR crystal can excite localized surface plasmon polaritons in thin metallic films with thicknesses much smaller than the wavelength of the incident light. Stronger surface plasmon coupling and higher electric field enhancement factors are found for increasingly grazing angles of incidence.9 Therefore, before considering the ATRSEIRAS performance of the μ-groove IRE, it is informative to first discuss how incident radiation is refracted through the crystal. Figure 1 shows a schematic of the 500 μm thick ATR IRE, which is grooved on one side and polished smooth on the other. The groove angle is equal to the angle between Si {110} and {111} planes, i.e. 35°. Light at a given angle of incidence (AOI) is incident on the slopes of the grooves at an angle γ = AOI − 35° with respect to the interface normal. Light refracted at the Si/air interface travels through the crystal at an angle ⎡ n sin γ ⎤ with respect to the interface normal φ = sin−1⎣⎢ airn ⎦⎥where Si nair and nSi are the refractive index of air and silicon respectively (NBboth γ and φ are negative quantities for AOI < 35°). The angle of the refracted light with respect to the laboratory 11819

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824

Article

Analytical Chemistry

Figure 2. Infrared transparency of a 25 mm radius Si hemisphere (black line) and a 10 × 10 mm μ-groove IRE using an angle of incidence of 55°. Spectra were the result of 36 co-added scans using 4 cm−1 resolution and 2 mm aperture. The spectra are referenced to the reflected response of a gold mirror of approximately the same area as the IREs. Figure 1. Schematic representation of the μ-groove IRE with a thin metal film deposited on the smooth surface to serve jointly as the working electrode and enhancing surface for electrochemical ATRSEIRAS (top). A magnified cross section of a single groove, and the definition of the angles of incidence of light are also shown (bottom).

explained as a combination band of the 1108 cm−1 Si−O−Si asymmetric stretch and transverse acoustic and optical phonon modes.33 However, recent evidence suggests that this band may be better explained as a combination mode that does not involve the asymmetric Si−O−Si stretch.34 The optical transparency of silicon can be improved somewhat by employing more expensive FZ rather than CZ silicon, but it is clear from Figure 2 that the intrinsic character of silicon requires smaller optical path lengths to prevent near total extinction of light with wavelengths longer than ca. 10 μm. This advantage is provided with the μ-groove IRE as shown by the second spectrum shown in Figure 2. Below 1500 cm−1, the thin ATR substrate is significantly more transparent compared to the 12.5 mm radius hemisphere. The transflectance profile of the μ-grooved wafers provides a moderately strong 1107 cm−1 absorption band, but only weak phonon resonances are present between 650 and 1000 cm−1. In this frequency range, the throughput of the wafer is between 8 and 20 times larger than that of the hemisphere. This is further illustrated in Figure 3,

normal is the sum of the groove angle and φ. It should be noted that the experimental incident beam is not collimated and has an estimated angular spread of ∼10°. Thus, the algebra above represents the average angles of incidence and refraction. An advantage of the μ-groove IRE is apparent upon consideration of the optical properties of silicon. Below 1500 cm −1 , the throughput of infrared radiation in silicon significantly decreases due to bulk phonon and interstitial SiOx absorptions. Phonon modes in Si arise from various combinations of two transverse optical modes, one longitudinal optical mode, two transverse acoustic modes, and a longitudinal acoustic mode.29 Oxygen impurities originating from crucibles are responsible for the presence of interstitial oxygen in silicon produced using the Czochralski (CZ) method but are greatly reduced in float-zone (FZ) silicon.30 An ATR throughput comparison of μ-grooved wafers and a 12.5 mm radius Si hemisphere is shown in Figure 2 using an angle of incidence of 55°. The throughput is shown as the transmitted light after reflection (transflection) relative to a gold mirror of approximately the same footprint placed at the sample plane of the specular reflectance accessory. The Si hemisphere has an average transmission of ∼0.4 above 1500 cm−1, which is slightly lower than that expected after accounting for the reflectance losses at each Si/air interface (R = 0.30 for normal incidence and unpolarized light). The slightly lower than expected reflectance is most likely due to the hemisphere presenting a marginally smaller reflecting area compared to the gold mirror used as the reference. The hemisphere exhibits prominent absorption features at 1719, 1446, 1378, 1298, and 1108 cm−1. On the basis of the compiled data provided by Medernach,29 multiphonon resonances in the silicon crystal account for the 1446, 1298, and 1378 cm−1 bands. The band at 1108 cm−1 originates from the asymmetric stretching of Si−O−Si arising from interstitial oxygen.31 The absorption feature observed at 1719 cm−1 is also strongly indicative of the presence of interstitial oxygen defects in the silicon.32 It is typically

Figure 3. ATR spectra of neat 4-methoxypyridine (MOP) using 25 mm radius ZnSe (top) or Si (bottom) hemispherical IREs. Refer to Table 1 for designations of labeled peaks. 11820

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824

Article

Analytical Chemistry Table 1. Assignment of ATR Infrared Vibrational Bands for MOPa peak label

description

symmetry class

measured

literatureb

calculatedc

13 12 11

ring vibration (υ8a) ring vibration (υ8b) ring vibration (υ19a) CH3 bend ring vibration (υ18a) CH3 bend CH3 bend ring vibration (υ19b) COC (asym stretch) weak, A′ (COC asym stretch) ring vibration (υ9a) υCl−O from perchlorate aniond COC (sym stretch + ring mode) ring breathing (υ1) ring vibration (υ11) ring vibration (υ5)

A′ A′ A′ A″ A′ A′ A′ A′ A′ A′ A′

1591 1567

A′ A′ A″ A′

1024 989 817 800

1597,s 1574,m 1507,m n/a n/a 1465,w 1445,w 1423,w 1289,s 1244,w 1215,s 105338 1029,s 990,m 818,s 802,m

1593 1562 1504 1500 1493 n/a 1455 1422 1268 1243 1218 n/a 982 966 815 776

10 9 8 7 6 5 4 3 2 1

1501 1461 1442 1418 1281 1241 1210

a

s = strong; m = medium; w = weak; sh = shoulder; n/a = not observed. bAssignments are made on the basis of reference 35 unless otherwise indicated. cCalculated frequencies have been multiplied by a factor of 0.985 to account for harmonic approximation. dOnly observed in electrochemical ATR-SEIRAS experiments. See Figure 4.

An advantage of large ATR crystals such as Si hemispheres and hemicylinders is that their bigger reflecting surfaces can better accommodate the larger beamspots produced at very high angles in commercial specular reflection accessories. Nevertheless, Figure S4 shows that maximum throughout using the 12.5 mm radius Si hemisphere occurs below 70° AOI, which is typical of the angle used in ATR-SEIRAS studies.8 Several factors need to be considered when choosing the optimal AOI for ATR-SEIRAS using the μ-groove IREs. First, an AOI of 35° means normal incidence upon the Si groove surface and minimum average reflection losses for unpolarised light. The reflection losses at the air/Si interface are plotted in Figure S5 as a function of both AOI and γ. The angles obtainable with the commercial specular reflection accessory employed in this study cover a spread of AOI between 30−80°, which equates to γ = 20 ± 25°. In this range, the average reflection losses are largely independent of incidence angle. However, only p-polarized light produces ATR-SEIRAS enhancement, and AOI = 35° is the maximum for reflection losses of p-polarized light. Although the Brewster angle for the air/Si interface would be at a geometrically unobtainable angle of γ ∼ 74°, it is apparent from Figure S5 that AOIs greater than 35° increase the relative intensity of transmitted, p-polarized light. Second, as discussed above, the angle of the light traveling through the crystal, φ, should be maximized to provide the strongest electric field enhancement at the Au/solution interface. However, the fact that the groove geometry maps the 50° range of incident angles to a much smaller range of refracted angles through the crystal (34° ≤ (35° + φ) ≤ 47°) should mitigate this effect. With these considerations in mind, a series of electrochemical ATR-SEIRAS experiments were performed as a function of the angle of incidence. Figure 4 shows the potential difference MOP ATR-SEIRA spectra for different angles. In contrast to the ATR results of Figure 3, the spectra shown in Figure 4 measure a monolayer of MOP adsorbed at the Ausolution interface. All experiments were performed on the same film and are presented as absorbance changes as per eq 1

which shows the pure ATR spectrum of neat 4-methoxypyridine (MOP) using 12.5 mm radius ZnSe and Si hemispheres. There is no surface enhancing Au film on either IRE in these experiments, and the measured signal arises from molecules within the depth of penetration of the evanescent wave (several hundreds of nanometers) as it extends into the liquid MOP. The ZnSe IRE is optically transparent above 600 cm−1, and eight prominent bands and five weak bands are present in the resulting ATR spectrum. Quantum mechanical calculations28 and previous reports35 were used to help make the band assignments listed in Table 1. If the ZnSe infrared element is exchanged for a Si hemisphere, the resulting ATR spectrum is qualitatively very similar above 1000 cm−1. However, the weak throughput of the Si hemisphere at long wavelengths lowers the S/N of peak 3 at 1000 cm−1 relative to the ZnSe ATR spectrum, and bands 1 and 2 at 750 and 800 cm−1 are completely absent. The ATR spectrum of neat MOP using the μ-groove IRE (see Figure S2) is very similar to that obtained using the ZnSe hemisphere. This indicates that using the thin ATR wafers in an electrochemical ATR-SEIRAS experiment should allow the observation of new spectral information in comparison to previous work on the potential dependent adsorption of MOP using Si hemispheres. μ-Groove IREs for Electrochemical ATR-SEIRAS. To test the suitability of the μ-groove IRE for electrochemical ATRSEIRAS, we used the potential dependent adsorption of MOP as an exemplary system. MOP was chosen because we have intensively characterized its potential-dependent adsorption on Au in recent years.28,36,37 Briefly, this pyridine derivative does not undergo faradaic processes but rather, as described by the voltammetry shown in the Supporting Information (SI) (Figure S3), it adsorbs as a monolayer (maximum surface coverage of 6.6 × 10−10 moles cm−2) on polycrystalline gold electrodes through the nitrogen nonbonding electron pair in the potential range −0.6 V < E < 0.5 V. At potentials more negative than −0.6 V, it is completely desorbed from the electrode surface. For electrochemical application, a gold film was deposited on the smooth side of the μ-groove IRE, and experiments were performed to determine the optimal angle of incidence for signal enhancement. 11821

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824

Article

Analytical Chemistry

Figure 5. Signal-to-noise ratio (S/N) for the MOP absorption peak at ∼1310 cm−1 (peak 7) as a function of AOI on the μ-groove IRE. The solid line is to guide the eye.

Figure 4. Dependence of the electrochemical ATR-SEIRAS response of MOP on angle of incidence (AOI) using a μ-groove IRE. Spectra were collected on the same film at different angles using single beam spectra at −0.70 and +0.20 V as reference and sample spectra, respectively. Spectra are the result of the co-addition of 128 interferograms.

ΔAbs = − log

and increased noise levels. In addition, the position of the beam on the reflecting plane of the specular accessory is dependent on the AOI. At low angles, this effect is quite small, but at increasing angles, the beamspot moves several millimeters offcenter. The position of the μ-groove IRE is fixed at the center position in the current configuration of the spectroelectrochemical cell, and consequently, there is a notable drop in signal. At optimized AOIs, the S/N levels are sufficiently high in the electrochemical SEIRA spectra to allow the identification of additional positive-going bands at 1562 cm−1 (peak 12), 1237 cm−1 (peak 6), 1210 cm−1 (peak 5), 1056 cm−1 (most likely the υCl−O stretch from coadsorbed perchlorate ions),38 1028 cm−1 (peak 4), and possibly a very weak signal at 736 cm−1. The latter band may be peak 1, but this assignment is tentative at best. The replacement of adsorbed and oriented water at −0.70 V with MOP at +0.20 V results in a negative-going signal at 1650 cm−1 due to a loss of the water bending mode. A broad downward band is also observed centered at ∼725 cm−1 and can be assigned to water librations39,40 associated with the frustrated rotations of hydrogen-bonded water molecules.41 The transition dipole moments of MOP vibrations are either in the plane (A′) or out of the plane (A″) of the aromatic ring. MOP is adsorbed on polycrystalline gold films through the ring nitrogen atom at applied potentials larger than −0.70 V vs Ag/ AgCl, and consequently, only A′ modes with dynamic dipole moments oriented along the molecular axis should be infrared active on the basis of SEIRAS selection rules.5 Peaks 8 and 2 are not observed because they possess A″ symmetry. In the case of peaks 3 and 12, a large angle between the dynamic dipole and the molecular axis renders the modes largely infrared inactive for vertically adsorbed MOP. Nevertheless, the selection rules cannot explain the absence (or very weak signal) of peak 1, which according to the DFT calculations has a transition dipole moment along the molecular axis and provides a large absorption feature near 800 cm−1 in the ATR experiment (see Figures 3 and S2). The fact that the water libration mode is observed is evidence that the μ-groove IREs provide SEIRAS sensitivity at low wavenumbers, but we do not presently have a satisfactory explanation for why the low frequency MOP mode is not more strongly SEIRAS active. Future studies are in progress examining other systems with IR features in this spectral region to see if the result is unique to the MOP system.

Ssample Sref

(1)

where Ssample and Sref are single-beam spectra collected at +0.20 V (potential of maximum MOP adsorption) and −0.70 V (potential of complete MOP desorption; MOP replaced by interfacial water). The six spectra provide strong signals for peaks 13 (1617 cm−1), 11 (1508 cm−1), and 7 (1310 cm−1). These features were evident in our previous report using Si hemispheres and originate from A′ vibrations that shift to higher wavenumbers upon adsorption to gold surfaces. A more complete analysis of the spectra is provided below, but it is first informative to discuss the dependence of the strong absorption signals on the angle of incidence. Despite the expectation of maximum throughput at AOI close to the groove angle, rather weak SEIRAS signal is observed at AOI = 40°. Relatively large band intensities are seen for 45° ≤ AOI ≤ 60°, but at AOI greater than 60°, the spectral quality deteriorates, and the spectra become significantly noisier. At 70°, there is no detectable signal (data not shown). The optimal angle of incidence for electrochemical ATR-SEIRAS response using the μ-groove IREs was quantified by evaluating the S/N of the peak near 1310 cm−1. The standard deviation of the baseline between 1320 and 1400 cm−1 was used to measure the noise, and the S/N ratios are plotted in Figure 5. The strongest signal and the maximum S/N are reached using an AOI between 50 and 55°. A maximum in the angular dependence is not intuitive as consideration of polarization-dependent reflection losses, and refracted angles through the crystal would predict the SEIRAS enhancement to increase with larger AOI (vide supra). Furthermore, geometric ray tracing (see Figures S6−S8) indicates that increasing AOI results in more incident rays undergoing double reflections within the IRE, which would increase the SEIRAS intensity. The poorer than expected S/N at high angles can be explained by considering the angular dependence of the beamspot size and position. In the SI, it is shown that at AOI greater than ∼55°, the beamspot size becomes larger than the grooved area on the IRE (see Figure S9). This results in greater light scatter from the delrin holder 11822

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824

Article

Analytical Chemistry

substrates using a commercial specular reflection accessory depends on the aggregate effects of (1) accessory throughput; (2) reflection losses at the surface of incidence; (3) refracted angle through the crystal and (4) the number of internal reflections. We have empirically shown that maximum electrochemical ATR-SEIRAS performance is achieved using an AOI of ∼50−55° for a μ-groove IRE with a groove angle of 35°. Our analysis indicates that better performance should be achievable at higher angles if the position of the μ-groove IRE can be moved to track the position of the beamspot. We are currently adapting our specular reflectance accessory accordingly. A S/N comparison of the μ-groove IRE shows that it outperforms a 12.5 mm radius Si hemisphere below ∼1450 cm−1 at optimized angles of incidence due to its much greater transparency, smaller reflection losses of p-polarized light, and the occurrence of multiple internal reflections. These effects are amplified at increasingly longer wavelengths, and the μ-groove IRE allows the observation of water librations at ∼725 cm−1, which cannot be observed using the large Si hemisphere. It should be noted that Si remains IR transparent into far IR frequencies and, with appropriate choice of detector, μ-groove wafers could easily be extended to far-IR ATR-SEIRAS measurements although the dimensions of the periodic groove structure and the wafer thickness would need to be considered to avoid diffraction and channeling effects. The results of the comparison between μ-groove wafers and hemispheres are reversed at frequencies where silicon is largely transparent (above ∼1440 cm−1) as the higher angles in the Si hemisphere provide greater SEIRAS enhancement. The performance of the μ-groove IRE is rather remarkable given that it provides much smaller refracted angles through the crystal as compared to the hemisphere. Utilization of μ-groove IREs with larger groove angles could further improve the electrochemical ATR-SEIRAS performance. The μ-groove IREs represent an appealing option for electrochemical ATR-SEIRAS. As well as outperforming conventional Si IREs at longer mid-IR wavelengths, they are significantly cheaper. This renders them amenable to applications where multiplexing strategies might be preferable. The small geometric footprint of the wafers would allow multiple substrates to be simultaneously derivitized with metal films. Furthermore, polishing Si IREs is a labor intensive process and is required prior to every metal film preparation. The substrates described in this report could eventually be regarded as disposable IREs for studies employing harsh chemical environments or toxic/pathological samples.

Extensive modeling of the SEIRAS enhancement mechanism is probably warranted to calculate the expected enhancement factors in this low energy region of the IR spectrum. Figure 6 provides a direct comparison of electrochemical ATR-SEIRAS performed using the 12.5 mm radius Si

Figure 6. Comparison of electrochemical ATR-SEIRAS response (reference and sample spectra as defined for Figure 4) for the 25 mm radius Si hemisphere (black line) and the μ-groove IRE (red line). The AOI for the μ-groove IRE was 55°, and the AOI for the hemisphere was 68°.

hemisphere and the μ-groove IRE at respective optimized angles of incidence. Above 1200 cm−1, the hemisphere outperforms the wafer in terms of signal amplitude by roughly a factor of 2−3. This is very comparable to the reported performance of a 200 μm thick Si wafer coupled to a ZnSe IRE relative to a conventional Si hemicylindrical internal reflection element.15 However, due to higher phonon absorption, the spectrum collected using the hemisphere is noticeably noisier around the 1310 cm−1 peak, and the calculated S/N ratio is actually 33% lower in comparison to the same peak in the spectrum collected using the μ-groove IRE. The spectral quality of the spectrum obtained using the hemisphere degrades even more significantly below 1200 cm−1. Although peak 4 at 1040 cm−1 is distinguishable, the perchlorate peak at slightly higher frequencies is buried in the interference caused by the strong absorption by the Si hemisphere. The noise level is so high as to make the assignment of the signal at 1040 cm−1 to MOP ambiguous in the absence of an ATR spectrum for the known compound. Such uncertainty in this spectral range would greatly obfuscate the identification of surface-confined intermediates produced during electrochemical reactions and represents a pronounced disadvantage of the large hemisphere IREs. In comparison, below ∼1450 cm−1, the spectral quality obtained using the μ-groove IRE provides a far lower noise level and allows unambiguous assignment of molecular peaks such as those arising from the displacement of water by adsorbed MOP.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b03509. Pure ATR spectrum of neat MOP on a microgroove IRE; measurements of Si hemisphere throughput over the angular range reported in the paper; calculations of the reflection losses at the Si/air interface; a discussion of the spot size at focus; and geometric ray tracing of the μgroove IRE (PDF)

■

■

CONCLUSIONS It can be succinctly concluded that μ-groove IREs are manifestly suitable for electrochemical ATR-SEIRAS applications. The optimal angle of incidence for the μ-groove

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; Phone: +1-306-966-4722. 11823

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824

Article

Analytical Chemistry ORCID

(26) Otto, A. Z. Phys. A: Hadrons Nucl. 1968, 216, 398−410. (27) Schumacher, H.; Kuenzelmann, U.; Vasilev, B.; Eichhorn, K.-J.; Bartha, J. W. Appl. Spectrosc. 2010, 64, 1022−1027. (28) Quirk, A.; Unni, B.; Burgess, I. J. Langmuir 2016, 32, 2184− 2191. (29) Medernach, J. W. In Handbook of Vibrational Spectroscopy; Chalmers, J. M., Griffiths, P. R., Eds.; John Wiley & Sons: Hoboken, NJ, 2006. (30) Ohba, T.; Ikawa, S. i. J. Appl. Phys. 1988, 64, 4141−4143. (31) Kaiser, W.; Keck, P. H.; Lange, C. F. Phys. Rev. 1956, 101, 1264−1268. (32) Karabudak, E.; Yuce, E.; Schlautmann, S.; Hansen, O.; Mul, G.; Gardeniers, H. Phys. Chem. Chem. Phys. 2012, 14, 10882−10885. (33) Pajot, B.; Stein, H. J.; Cales, B.; Naud, C. J. Electrochem. Soc. 1985, 132, 3034−3037. (34) Susarrey-Arce, A.; Tiggelaar, R. M.; Sanders, R. G. P.; Geerdink, B.; Lefferts, L.; Gardeniers, J. G. E.; van Houselt, A. J. Phys. Chem. C 2013, 117, 21936−21942. (35) Spinner, E.; White, J. C. B. J. Chem. Soc. 1962, 3115−3118. (36) Morhart, T. A.; Quirk, A.; Lardner, M. J.; May, T. E.; Rosendahl, S. M.; Burgess, I. J. Anal. Chem. 2016, 88, 9351−9354. (37) Unni, B.; Simon, S.; Burgess, I. J. Langmuir 2015, 31, 9882− 9888. (38) Zhumaev, U. E.; Lai, A. S.; Pobelov, I. V.; Kuzume, A.; Rudnev, A. V.; Wandlowski, T. Electrochim. Acta 2014, 146, 112−118. (39) Bertie, J. E.; Lan, Z. Appl. Spectrosc. 1996, 50, 1047−1057. (40) Maréchal, Y. J. Mol. Struct. 2011, 1004, 146−155. (41) Bergonzi, I.; Mercury, L.; Brubach, J.-B.; Roy, P. Phys. Chem. Chem. Phys. 2014, 16, 24830−24840.

Ian J. Burgess: 0000-0001-9611-1431 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was funded by a grant from the Natural Science and Engineering Research Council (NSERC) of Canada. T.A.M. also acknowledges funding from the NSERC PGS-D program. The authors would like to thank the staff of the machine shop in the Physics Department at the University of Saskatchewan for constructing the ATR-SEIRAS cell. The authors thank Lorenz Sykora from IRUBIS GmbH for several detailed discussions on the microgroove internal reflection elements.

■

REFERENCES

(1) Osawa, M. In Handbook of Vibrational Spectroscopy; Chalmers, J. M., Griffiths, P. R., Eds.; John Wiley & Sons: Hoboken, NJ, 2006. (2) Vasan, G.; Chen, Y.; Erbe, A. J. Phys. Chem. C 2011, 115, 3025− 3033. (3) Hartstein, A.; Kirtley, J. R.; Tsang, J. C. Phys. Rev. Lett. 1980, 45, 201−204. (4) Tolstoy, V. P.; Chernyshova, I. V.; Skryshevsky, V. A. In Handbook of Infrared Spectroscopy of Ultrathin Films; John Wiley & Sons: Hoboken, NJ, 2003; pp 416−475. (5) Aroca, R. F.; Ross, D. J.; Domingo, C. Appl. Spectrosc. 2004, 58, 324A−338A. (6) Jensen, T. R.; Van Duyne, R. P.; Johnson, S. A.; Maroni, V. A. Appl. Spectrosc. 2000, 54, 371−377. (7) Le Ru, E. C.; Etchegoin, P. G. In Principles of Surface-Enhanced Raman Spectroscopy; Elsevier: Amsterdam, The Netherlands, 2009; pp 299−365. (8) Osawa, M. In Advances in Electrochemical Science and Engineering; Wiley-VCH Verlag GmbH: Weinheim, Germany, 2008; Volume 9, Chapter 8, pp 269−314. (9) Suzuki, Y.; Osawa, M.; Hatta, A.; Suetaka, W. Appl. Surf. Sci. 1988, 33−34, 875−881. (10) Killian, M. M.; Villa-Aleman, E.; Sun, Z.; Crittenden, S.; Leverette, C. L. Appl. Spectrosc. 2011, 65, 272−283. (11) Kellner, R.; Mizaikoff, B.; Jakusch, M.; Wanzenbock, H. D.; Weissenbacher, N. Appl. Spectrosc. 1997, 51, 495−503. (12) Fasasi, A.; Griffiths, P. R.; Scudiero, L. Appl. Spectrosc. 2011, 65, 750−755. (13) Delgado, J. M.; Rodes, A.; Orts, J. M. J. Phys. Chem. C 2007, 111, 14476−14483. (14) Delgado, J. M.; Orts, J. M.; Rodes, A. Electrochim. Acta 2007, 52, 4605−4613. (15) Xue, X.-K.; Wang, J.-Y.; Li, Q.-X.; Yan-Gang; Liu, J.-H.; Cai, W.B. Anal. Chem. 2008, 80, 166−171. (16) Karabudak, E. Electrophoresis 2014, 35, 236−244. (17) Karabudak, E.; Mojet, B. L.; Schlautmann, S.; Mul, G.; Gardeniers, H. J. G. E. Anal. Chem. 2012, 84, 3132−3137. (18) Herzig-Marx, R.; Queeney, K. T.; Jackman, R. J.; Schmidt, M. A.; Jensen, K. F. Anal. Chem. 2004, 76, 6476−6483. (19) Queeney, K. T.; Fukidome, H.; Chaban, E. E.; Chabal, Y. J. J. Phys. Chem. B 2001, 105, 3903−3907. (20) Karabudak, E.; Kas, R.; Ogieglo, W.; Rafieian, D.; Schlautmann, S.; Lammertink, R. G. H.; Gardeniers, H. J. G. E.; Mul, G. Anal. Chem. 2013, 85, 33−38. (21) Shao, M.-h.; Liu, P.; Adzic, R. R. J. Am. Chem. Soc. 2006, 128, 7408−7409. (22) Shi, F.; Ross, P. N.; Somorjai, G. A.; Komvopoulos, K. J. Phys. Chem. C 2017, 121, 14476−14483. (23) Hanawa, H.; Kunimatsu, K.; Watanabe, M.; Uchida, H. J. Phys. Chem. C 2012, 116, 21401−21406. (24) Ulrich, R. J. Opt. Soc. Am. 1970, 60, 1337−1350. (25) Tien, P. K.; Ulrich, R. J. Opt. Soc. Am. 1970, 60, 1325−1337. 11824

DOI: 10.1021/acs.analchem.7b03509 Anal. Chem. 2017, 89, 11818−11824