Application of a Diamond Thin Film as a Transparent Electrode for In Situ Infrared Spectroelectrochemistry

Figure 1 contains a schematic of the ATR electrochemical cell. The ATR cell consists of a single reflection, ZnSe internal reflection element ( Harrick Scientific) with an electrochemical cell pressed onto its upper surface. Sandwiched between the electrochemical cell and the ZnSe crystal is a conductive diamond film deposited onto a thin (∼50 μm) silicon wafer. The boron-doped, polycrystalline diamond films are grown by hot-filament assisted chemical vapor deposition using a 1% in mixture with trimethyl boron and BN as the boron sources. Reference ¹⁵ summarizes further growth details. The substrates are double-sided polished silicon wafers (Virginia Semiconductor) doped to a conductivity level of 10-20 Ω cm (p-type). Prior to growth, we spray the silicon wafer with a 0.1 μm diamond powder slurry and then grow for 24 h at a filament temperature of 2050°C. This procedure forms a heavily doped (∼6000 ppm) film approximately 4-6 μm thick. Because of the compressive stress of the film, the wafer becomes deformed with the diamond side of the wafer in a convex shape. This deformed shaped introduces no obvious complications for the ATR measurements, however, because the wafer remains flexible enough to be pressed flat during measurements.

Zoom In Zoom Out Reset image size

The diamond-coated silicon wafer is the working electrode in the electrochemical cell. The acrylic cell presses the wafer onto the ATR prism for good optical contact and a leaktight seal with the electrolyte; the O-ring makes the only contact to the diamond (Fig. 1). The exposed circular area of the ZnSe crystal is ∼1.5 cm². A hand screw mounted onto the assembly provides the pressure to make the seal without breaking the diamond-coated wafer. The electrical connection is copper tape attached to the diamond with silver conductive epoxy; the tape is masked where necessary to avoid grounding. Point-contacting these wafers would shatter them; therefore, the copper tape connection is kept relaxed. A second, larger diameter O-ring on the acrylic cell presses down on the tape, so that the diamond will not be affected if the tape were pulled.

The counter electrode (Pt wire) and the reference electrode (a dynamic hydrogen reference) entered the electrolyte via angled holes in the acrylic cell. The electrolyte was 0.5 M (Fisher) and was inserted into the cell by pipette through the angled holes. In this work, the diamond electrode was conditioned with cyclic voltammetry (CV) using an EG&G PAR model 283 potentiostat; the potential was swept between −1.5 and +3.0 V vs. standard hydrogen electrode (SHE) at 200 mV/s for at least 10 min. During infrared measurements on the conditioned electrode, the potential was held at some constant value between these extremes. In all cases, the maximum current was 1 mA. Figure 2 contains a sample voltammogram for the diamond film electrode after performing at least 10 min of CV.

Zoom In Zoom Out Reset image size

Figure 1 also shows a schematic of the optical path through the system. The IR light enters the ZnSe crystal and undergoes a series of refractions until it internally reflects at the diamond-electrolyte surface (angle of incidence 58° while diamond-water critical angle 33°). The Fourier transform infrared (FTIR) spectrometer is a Bomem 151 interferometer operated at 8 cm⁻¹ resolution using a mercury cadmium telluride (MCT) detector. Spectra are obtained as interferograms using GRAMS/386 software (Galactic Industries) and then Fourier transformed to yield single beam spectra between 550 and 4000 cm⁻¹; the ZnSe crystal sets the lower wavenumber limit. In a typical FTIR measurement, we polarized the conditioned diamond electrode to 0.0 V vs. SHE (point 1 in Fig. 2), and then collected 500 scans (∼6 min) of the ATR reflectance as a reference spectrum. We then stepwise polarize the electrode to the desired voltage (e.g., points 2-5 in Fig. 2) and then ratio the desired spectrum to the reference spectrum (point 1).

Figure 3a shows the infrared spectra of water and electrolyte when placed directly onto the ATR prism (no film in optical path). In this figure, the reference spectrum is the air-ZnSe interface. Both of these spectra show the characteristic absorption features of water: OH scissoring at 1635 cm⁻¹ and the symmetric and asymmetric O–H stretches at ∼3320 cm⁻¹. In the electrolyte spectrum, two new features arose due to the S=O stretches (1050 and 1190 cm⁻¹) of the sulfuric acid in the electrolyte. Figure 3b shows the infrared spectrum taken at the diamond-electrolyte interface when the diamond-coated wafer is pressed onto the ATR prism. The electrolyte spectrum displays the same characteristic water and sulfuric acid features as for electrolyte placed directly on the ATR crystal (Fig. 3a). As a practical matter, we increased the pressure on the diamond wafer until the electrolyte bands appeared in the spectrum; this indicates that sufficient optical contact was established for unimpeded IR transmission to diamond-electrolyte interface. Note that there is a difference in the magnitudes of the IR absorption features in Fig. 3a and b, due to differences of the angle of incidence in the two experiments. In any ATR experiment, the penetration depth of the evanescent IR wave decreases as the angle of incidence increases.¹⁷ In the electrolyte-ZnSe experiment, the angle of incidence is 45°, whereas in the electrolyte-diamond measurement the angle is 58° (Fig. 1). Thus, this explains why the absorption features of Fig. 3a are larger than those of Fig. 3b.

Zoom In Zoom Out Reset image size

Changes to the as-grown (unconditioned) diamond surface occur as soon as we apply a potential of 0.0 V vs. SHE to the electrode (point 1 in Fig. 2). Ratioing this 0.0 V spectrum against the spectrum of the unpolarized electrolyte on the diamond (Fig. 3c) clearly shows that the amount of sulfate in the interface region has increased; in addition, a peak at 3240 cm⁻¹ appears. Although this peak overlaps the O–H stretch of the water in the electrolyte (Fig. 3a), the corresponding OH scissoring peak of water at 1635 cm⁻¹ did not appear. Thus, the 3240 cm⁻¹ feature is not due to the –OH in the water, and we have assigned this peak to adsorbed OH. Note the change in scale between Fig. 3a,b and Fig. 3c.

Previous work has shown that the first voltammogram on the hydrophobic as-grown diamond electrode is usually different than later scans (when the electrode is hydrophilic).^{14 15} Although understanding this phenomenon is one objective of this research, we have focused our initial experiments on diamond electrodes that have been preconditioned with at least 10 min of CV to remove this effect. In this way, we can show a proof of principle for our method and qualitatively identify changes in the diamond electrode due to polarization effects with less complication from the one-time changes in the surface properties of the electrode.

After conditioning the electrode, we have applied a series of stepwise polarization potentials to the diamond electrode in the order indicated in Fig. 2. The potentials are chosen based on voltammetry features of typical diamond electrodes, including the redox couple at +1.9 V which is consistent with surface oxidation at the grain boundaries.^{14 15} Figure 4 shows the resulting spectra corresponding to points 2-5 in Fig. 2; each spectrum is ratioed to the spectrum taken on the conditioned electrode at 0.0 V (point 1 in Fig. 2). The spectra are offset for clarity. The spectra clearly show two new features at 3240 and 1100 cm⁻¹ that change from points 2 to 5. [The other sharp features are due to atmospheric changes in the (2350 and water vapor (3720 cm⁻¹ and 1600 cm⁻¹) surrounding our FTIR.] These features change in tandem with each other indicating that they are different modes of a common adsorbed species or are different species linked by a common chemistry. The features grow as the potential increases to +3.0 V (toward evolution), but, interestingly, the largest features appear at +1.5 V when the potential has been stepped back down.

Zoom In Zoom Out Reset image size

Again we assign the feature at 3240 cm⁻¹ to the O–H stretch of adsorbed hydroxyl. Based on the observation that the C-O stretch of alcohols lies in the range of 1000-1200 cm⁻¹, we assign the 1100 cm⁻¹ feature to the C–O stretch of adsorbed OH on the diamond surface. Both of these assignments are consistent with other measurements of water adsorbed on diamond: Ando et al.^{18 19 20} observed O–H in the region 3500-3000 cm⁻¹ while Struck and D'Evelyn⁴ observed the C–O stretch near 1125 cm⁻¹ for water adsorbed on diamond (100). In addition, the above interpretation is consistent with measurements using XPS, which show oxygen addition to the surface after polarizing the electrode.^{15 16} On the other hand, compared to other oxidized diamond surfaces,^{18 19 20 21} there is no observable C=O stretch nor ether (C–O–C) stretch, but the C=O feature may be hidden underneath the vapor bands (1600-1800 cm⁻¹). There is no observable C–H stretch in the spectra at ∼2900 cm⁻¹, but even if it were present, this band is expected to be very weak.^{4 5} The spectra also do not show any changes in the sulfate features as observed in Fig. 3c.

Although polarizing the electrode is critical to this oxygen addition, time probably also plays a significant role. IR measurements of a preconditioned electrode taken at 0.0 V before and after additional CV show little change in the 3240 and 1100 cm⁻¹ features observed in Fig. 4, so some critical polarization voltage is required to activate the –OH addition. On the other hand, the amount of adsorbed –OH does not necessarily increase with the magnitude of applied potential since the diamond electrode at +1.5 V on the reverse step has more –OH than does +3.0 V. This experimental fact indicates that the amount of adsorbed –OH also depends on the amount of time the electrode is in contact with the electrolyte. Thus, +1.5 V (point 5) has larger features because it has been in contact with electrolyte longer than the previous points. (Note that each of the marked points in Fig. 2 requires ∼6 min of scan time on the FTIR.) More experiments are required to elucidate this phenomenon, however.

This paper reports research on the application of diamond thin films as transparent electrodes for in situ infrared measurements during electrochemical polarization. The electrode consists of a heavily boron-doped diamond film (4-6 μm) deposited on a thin silicon wafer (∼50 μm). The electrode is then pressed onto a ZnSe prism, and electrolyte is placed onto its surface. As proof of principle, we then performed a series of attenuated total reflectance experiments to examine how the infrared spectrum of the diamond-electrolyte interface changes as the electrochemical potential varies. For an unpolarized electrode, the IR absorption bands of the water and sulfate in the electrolyte are clearly observable through the electrode; polarizing the electrode also affects the IR absorption spectrum of the diamond-electrolyte interface. After conditioning the electrode for at least 10 min with cyclic voltammetry, the IR spectra recorded at different potentials show two new spectral features attributed to the O–H stretch (3240 cm⁻¹) and the C–O stretch (1100 cm⁻¹). These features change with potential and time.

Because of their wide potential window for electrochemical experiments and wide spectral transparency (ultraviolet to infrared), diamond electrodes should be useful transparent electrodes in many other spectroelectrochemical studies.

This research was supported by a National Science Foundation grant (CTS-9625095) (P.W.M.) and the Case School of Engineering (H.B.M.). The authors also acknowledge the experimental assistance of Wayne Schmidt, Cliff Hayman, Ram Raghavan, Dave Csuy, and Tammie Rupnik as well as useful discussions with John C. Angus and Uziel Landau.

Case Western Reserve University assisted in meeting the publication costs of this article.