Vacuum-annealed undoped polycrystalline CVD diamond: a new electrode material

Electrochimica Acta 49 (2003) 41–49

Vacuum-annealed undoped polycrystalline CVD diamond: a new electrode material Yu.V. Pleskov a,∗ , M.D. Krotova a , V.G. Ralchenko b , A.V. Khomich c , R.A. Khmelnitskiy d a

c

Frumkin Institute of Electrochemistry RAS, Leninsky pr. 31, Moscow 119071, Russia b General Physics Institute RAS, Vavilova Str. 38, Moscow 119991, Russia Institute of Radio Engineering and Electronics RAS, 1 Vvedenskogo square, Fryazino 141190, Russia d Lebedev Physical Institute RAS, Leninsky pr. 53, Moscow 117924, Russia Received 11 November 2002; accepted 29 May 2003

Abstract Vacuum-annealing imparts conductivity to initially insulating undoped polycrystalline chemical–vapor-deposited diamond, thus turning it to a possible electrode material. The diamond film annealed at 1775 K appeared to be practically not conducting. With further increase in the annealing temperature above 1825 K, the film effective resistivity decreased from initial value of 1011 to 1012  cm down to less than 0.1  cm; the differential capacitance increased from ∼10−3 to ∼50 ␮F per 1 cm2 of geometrical surface; the transfer coefficients for electrochemical reactions in the [Fe(CN)6 ]3−/4− redox solution increased from ∼0.2 to 0.5; and the degree of reversibility of the electrochemical reaction increased. The observed changes in the electrode properties are attributed to gradual change in the thickness and/or properties (first and foremost, conductivity) of the nondiamond carbon phase formed along the intercrystallite boundaries upon the annealing; the conducting phase is outcropping at the film surface as an array of microelectrodes (“active sites”). © 2003 Elsevier Ltd. All rights reserved. Keywords: Annealing; Diamond films; Electrochemical kinetics; Electrochemical impedance; Microelectrode

1. Introduction Diamond is an extremely corrosion-stable electrode material, suitable for synthetic, analytical, and environmentaloriented electrochemical applications [1]. To impart electrical conductance to diamond at room temperature, which is necessary for its use in the electrochemistry, diamond can be doped with boron, thus producing a p-type conductor. (Recently, codoping with B and S was suggested to be the method for fabricating n-type semiconductor diamond [2].) However, the doping is not the unique way to make diamond conducting. We recall that the first paper on the electrochemistry of diamond dealt with undoped chemical–vapor-deposited (CVD) films whose conductance was attributed to some unidentified structural defects formed within crystallites in the course of the growth process and playing the role of acceptors [3]. Also attempts were made to ∗

Corresponding author. E-mail address: [email protected] (Yu.V. Pleskov).

0013-4686/$ – see front matter © 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.electacta.2003.05.005

use an ion implantation for making surface layer of dielectric diamond crystal conducting, due to its graphitization [4–6]. An alternative approach to impart conductance to polycrystalline CVD diamond is the high-temperature-annealing in vacuum; thus, a diamond-based conducting material can be prepared, that can be used as electrode material in electrochemistry [7]. Recently, the structural changes in diamond films, annealed in vacuum up to 1900 K, were studied with electron paramagnetic resonance, IR and UV-Vis optical absorption, Raman and photoluminescence spectroscopy, thermal conductivity, electron energy loss spectroscopy and high-resolution electron microscopy (HREM) [8–11]. Using the HREM analysis [11], the formation of turbostratic and/or well crystallized graphite or amorphous carbon layers up to 20 nm thick along the grain boundaries in our thick (a few 100 ␮m) film was visualized (Fig. 1), as well as islands of “graphite-like” material in defective areas inside the crystallites. The amount of the nondiamond phase formed upon the heat treatment depends on the annealing temperature

42

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

Fig. 1. High resolution TEM image showing grain boundary in (a) as grown and (b) annealed at 1675 K polycrystalline diamond film [11]. Note the formation of a (∼11 nm) strip of a graphitic material—turbostratic carbon (shown by arrow) on a grain boundary. A and B are two grains; crystal orientation of the grains is given in brackets.

and time, as well as on the diamond film “quality” [10–12], the more defective films being graphitized easier. These amorphous carbon and/or “graphite-like” layers compose a continuous conducting network making the primarily insulating material conducting. The outcroppings of intercrystallite boundaries to the film surface may play the role of microelectrodes, thus facilitating the charge transfer at the electrode/electrolyte interface. For simplicity, in what follows the above-listed phases are designated by a single term “graphite-like carbon”. We bear in mind that the transformed sp2 phase is under high compression from surrounding diamond matrix (due to the difference in density between diamond and carbon and incompressibility of the diamond grains) and therefore may possess properties somewhat different from ordinary turbostratic carbon or graphite. Some difference in electrochemical behavior may also be expected. In particular, one may hope that its outcroppings to the electrode surface will be more

corrosion-stable than graphite proper. This point deserves special studies. In the present work, we studied the effects of the vacuum-annealing of undoped polycrystalline CVD diamond films on its electrochemical behavior, by measuring the electrochemical impedance in the indifferent electrolyte solutions and taking potentiodynamic curves in Fe(CN)6 3−/4− redox system.

2. Experimental A 400 ␮m-thick diamond film was grown in a methane/ hydrogen/oxygen mixture on a 60 mm-diameter mirrorpolished silicon substrate using a 5 kW microwave plasma-enhanced CVD system (an Astex PDS19 model) as described elsewhere [13]. After the deposition the Si substrate was chemically etched-off to obtain a free-standing

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

film, which was then laser-cut to 6 mm × 6 mm pieces. The samples were dark gray in color, yet being partially translucent. The concentration of nitrogen impurities in the form of single substitutional NS0 atoms in the samples under study varied from 1 to 2 ppm, as estimated from the amplitude of a 4.6 eV (270 nm) absorption band [14]. Raman spectra of the as-grown samples exhibited a strong diamond line at 1332 cm−1 , and only negligible, if any, sign of amorphous carbon inclusions. The average thickness of residual SiC layer at the nucleation side of the films, upon etching-off the Si substrate, does not exceed 1 nm (the detection limit of IR spectroscopy), as estimated from reflectance spectra in the 600–1000 cm−1 range [15]. Typically, the resistivity of the films was in the range of 1011 to 1012  cm. The samples were annealed in a graphite-wall furnace at temperatures Tann between 1725 and 1925 K (accuracy 10 K) for 1 h in vacuum of 10−5 Torr. To remove any surface graphite, the samples after annealing were exposed to a boiling H2 SO4 + K2 Cr 2 O7 solution. This procedure has been proved to remove all residual surface graphite from natural diamond crystals annealed under the same conditions at 1925 K, so the observed darkening of the annealed samples was not caused by the graphitization of their surface. Optical spectroscopy measurements were carried out using a UV-Vis spectrometer “Specord M400” (Carl Zeiss, Jena, Germany) in a 0.185–0.9 ␮m wavelength range. In the electrochemical experiments, both the faceted growth side (the grains sized 20–60 ␮m) and the smooth nucleation side with submicron-sized grains were studied for each sample. For electrochemical measurements at, e.g. growth surfaces, the electrical connection to the rear side of the samples was made with silver epoxy. The current lead was isolated with a layer of high-purity paraffin. On performing a full set of impedance and kinetic measurements on the growth side, the paraffin layer was removed, the silver contact washed off from the nucleation surface, and

43

the silver epoxy was applied to the growth surface, which now served as an electrical connection in the subsequent measurements on the nucleation side. A three-compartment glass electrochemical cell was used. A normal Ag/AgCl electrode was the reference electrode (in what follows, all potential values are given against this electrode); the auxiliary electrode was made of platinum. The current and differential capacitance values are related to 1 cm2 of geometrical surface. The electrodes’ differential capacitance was measured in a 20 Hz to 200 kHz frequency range using an R-5021 ac bridge (“TOCHELEKTROPRIBOR”, Ukraine). The measurements were performed in 2.5 M H2 SO4 indifferent electrolyte solution. Potentiodynamic curves were taken in 2.5 M H2 SO4 + 0.01 M Fe(CN)6 3− [or Fe(CN)6 4− ] solutions at a linear potential scanning rate v ranging from 5 to 100 mV s−1 , by using a PI-50-1 potentiostat equipped with a PR-8 programming unit and a PDA-1 x–y recorder (“ZIP”, Byelorussia).

3. Results 3.1. Samples Characterization The diamond film annealed at 1775 K appeared to be practically not conducting. With further increase in the annealing temperature above 1825 K, the film effective (that is, averaged over the entire bulk) resistivity abruptly decreased from ∼3300  cm (Tann = 1825 K) down to ∼50  cm (1885 K) and eventually fell below that of the electrolyte solution (∼0.1  cm). The anneal runs at temperatures higher than 1700 K resulted in a marked increase of absorption in IR and UV-Vis ranges. The differential absorption spectra (Fig. 2) characterise the average amount and composition of the

Fig. 2. The UV-Vis difference absorption spectra (i.e. the spectrum of a sample annealed at temperature Tann (shown at the curves) minus the spectrum of the same sample prior to the annealing).

44

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

Fig. 3. (a) ln C vs. Tann dependence taken in 2.5 M H2 SO4 solution at steady-state potential, 0.4 V vs. Ag/AgCl ((䊊) growth side; (䊉) nucleation side). (b) Equivalent circuit of electrodes: Rs is the series (ohmic) resistance, RF is the resistance of the interfacial charge transfer, C is the differential capacitance.

nondiamond phase formed. The UV absorption peak at 255 nm is characteristic of “graphite-like” materials; it is generally assigned to the interband transitions between ␲ bands near the M point in the Brillouin zone [16]. Annealing at higher temperatures resulted in a shift of this peak to higher wavelengths, probably due to the compressive stress experienced by the grain boundaries and/or changes in the structure and composition of the intercrystallite-boundary phase [9,10]. Basing on the increase in absorption, we estimated the total thickness of the “graphite-like” material formed by the annealing. According to [12], a 100 cm−1 -increment in the absorption coefficient α∗ corresponds to formation of a 0.07–0.5-␮m-thick graphite layer. With the diamond wafer thickness of 400 ␮m, we concluded from Fig. 3b that upon annealing, e.g. at 1925 K, ca. 0.1 vol.% of diamond is converted to a “graphite-like” material. The obtained electrodes showed low background current (∼0.1 ␮A per 1 cm2 of geometrical surface, over the −0.5 to +1.5 V potential range) which did not interfere with the impedance and potentiodynamic curves measurements. 3.2. Differential capacitance With the annealing temperature Tann increasing from 1825 to 1915 K, the differential capacitance measured in the indifferent electrolyte at a steady-state potential, 0.4 V (Ag/AgCl) increased from ∼10−3 ␮F cm−2 , a value characteristic of poor conductors, to ∼50 ␮F cm−2 , showing a “metal-like” behavior of the electrodes (Fig. 3a). In calculating the capacitance C, we approximated the impedance spectra by the

Randles’ equivalent circuit (Fig. 3b) constituted by the differential capacitance of the interface C, the interfacial charge transfer resistance RF and a series resistance Rs related to the film bulk.1 Because of rather wide (by several orders of magnitude) capacitance variation range, the real and imaginary components of impedance Z also vary widely; therefore, the impedance spectra taken for different annealing temperature can be conveniently compared on the complex plane in bi-logarithmic, rather than linear, coordinates. Fig. 4a gives this presentation for the growth sides of the films; in Fig. 4b, a Bode plot for the films is shown. Some common features of the curves are noteworthy. With increasing annealing temperature, the log |ImZ| versus log ReZ curves shift toward lower impedances. The “high-temperature” curves are as if continuations of the “low-temperature” ones, all curves make a narrow “band” on the complex plane, whose width is probably a merit of scatter of experimental points. The intercepts at the ReZ axis cut off by the high-frequency (“vertical”) segments obviously give the film bulk resistance Rs ; with increase in the annealing temperature Tann , this resistance regularly decreases. And the Bode curves (Fig. 4b) for the strongly annealed samples (1885 < Tann < 1910) are dominated by the high-frequency resistance Rs (see the circuit of Fig. 3c). Whereas the films annealed at lower frequencies are dominated by the ascending part of the semicircle related to the RF C chain of the circuit. Shown in Fig. 5a is the 1000/Tann -dependence for the logarithm of differential capacitance C. The data scatter is probably due to some inhomogeneity of the diamond wafer: the samples cut from its different areas somewhat differed in their properties. The scatter did not allow us to distinguish between the capacitance values for the growth and nucleation sides. In particular, the same “activation energy”2 Ea ∼ 1500 kJ mol−1 was calculated from the slope of the line for both surfaces, by using the formula C = C0 exp(−Ea /RT), where R is the universal gas constant, T is the absolute temperature. For comparison purposes we give in Fig. 5b, for the same set of samples, a 1000/Tann -dependence for the logarithm of increment in the IR absorption coefficient, α∗ (α∗ is the absorption coefficient for the wavelength of 800 nm). The “activation energy” here is Ea ∼ 300 kJ mol−1 .

1 A closer inspection of the spectra showed that thus calculated capacitance shows frequency dispersion, even if moderately low. It would be well to replace C in the equivalent circuit by, e.g. a constant phase element. However, the approximation we used has little or no effect on the qualitative conclusions we have drawn. 2 We use the term “activation energy” in a conditional way because the data presented in Fig. 5 are not truly kinetic in nature. We cannot define the particular process which results in the increasing capacitance or kinetic properties. Thus, the Ea values are but effective characteristics for the intensity of the carbon phase transformation affecting the measured electrochemical properties. Yet the term “activation energy” is acceptable for description of optical absorption evolution because this value is proportional to the amount of transformed (graphitic) material.

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

Fig. 4. (a) The log |ImZ| vs. log ReZ dependence (the growth side) for different Tann (shown at the curves). (b) The Bode plot.

45

46

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

Fig. 5. The 1000/Tann dependences for (a) ln C (C is the differential capacitance measured in 2.5 M H2 SO4 at steady-state potential, 0.4 V vs. Ag/AgCl) ((䊊) growth side, (䊉) nucleation side). (b) (䊐) log α∗ (α∗ is the absorption coefficient for λ = 800 nm).

3.3. Potentiodynamic curves Typical anodic and cathodic potentiodynamic current (I) versus potential (E) curves, taken in a 0.01 M Fe(CN)6 4− or Fe(CN)6 3− solution, respectively (at a background of 2.5 M H2 SO4 ), at different potential scan rates are shown in Fig. 6. The curves measured for samples annealed at temperatures exceeding some critical value (Tann ∼ 1865 K) show a current peak, which evidences a relatively rapid electrochemical reaction: the rate of charge transfer at the electrode/solution interface and the rate of the reactant mass transfer in solution are comparable [17]. (For samples annealed at lower temperatures, the current is much lower and probably of kinetic nature; no peaks were observed on the I versus E curves.) The peak potentials for the anodic and cathodic current Ep depend on the potential scanning rate v, which points to an irreversible nature of the reaction. From the slope of Ep versus log v dependencies we calculated the transfer coefficients for the cathodic (α) and anodic (β) reactions in the Fe(CN)6 3−/4− redox system. With increase in the annealing temperature, the transfer coefficients α and β increased, in the studied Tann range, from ∼0.1–0.2 to a value of ∼0.5; the peak magnitude of

Fig. 6. Potentiodynamic curves: (a) anodic in 2.5 M H2 SO4 + 0.01 M Fe(CN)6 4− solution; (b) cathodic in 2.5 M H2 SO4 + 0.01 M Fe(CN)6 3− solution. The potential scanning rate v (mV s−1 ): (1) 5, (2) 10, (3) 20, (4) 50, (5) 100. Film annealed at 1910 K, its geometrical surface area 0.15 cm2 .

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

Fig. 7. Dependence of transfer coefficients for (a) Fe(CN)6 3− reduction (α) and (b) Fe(CN)6 4− oxidation (β) on Tann ((䊊) growth side, (䊉) nucleation side).

the anodic and cathodic current peaks also grew gradually. In Fig. 7 we give the α and β versus Tann dependence; in Fig. 8, the Tann dependence of the peak currents Ipred for Fe(CN)6 3− reduction and Ipox for Fe(CN)6 4− oxidation. We can judge on the degree of irreversibility of electrochemical reaction by the difference Ep of potentials of the anodic and cathodic current peaks on a cyclic voltammogram taken in a mixed Fe(CN)6 4− +Fe(CN)6 3− solution (Fig. 9a). In particular, the larger is Ep , the more irreversible (slow) the reaction. With increasing Tann , Ep decreases as Fig. 9b shows. For the higher Tann , Ep approaches its “theoretical” value of 57 mV characteristic of reversible reactions [17].

4. Discussion Generally, we can conclude from the measured annealingtemperature dependencies of the electrode characteristics that with increasing Tann , the samples’ properties change from those characteristic of “poor conductor” to “metal-like”. Indeed, samples annealed at higher temperatures have differential capacitance (∼50 ␮F cm−2 ) and transfer coefficients (∼0.5) characteristic of metal electrodes, rather than semiconductor ones. And judging from

47

Fig. 8. Dependence of the peak current density for the potentiodynamic curves of (a) Fe(CN)6 3− reduction and (b) Fe(CN)6 4− oxidation on Tann ((䊊) growth side, (䊉) nucleation side). The potential scanning rate 5 mV s−1 .

the Ep value, the electrochemical reaction becomes reversible. We relate the annealing-temperature effects on the electrochemical characteristics of diamond films to the formation in the films, upon their annealing, of a conducting nondiamond phase, first and foremost along the intercrystallite boundaries. However, as shown in [10,11], in addition to the intercrystallite boundaries, this phase appears also at structural defects in the bulk of diamond crystallites. We may speculate that in well defective crystallites, upon overcoming the percolation threshold, the “graphitized” areas can form continuous conducting paths for electrical current.3 Thus, the quantities we measured ( α∗ ; C; Ip ; α and β;

Ep ) are the merit of the intensity at which the new phase forms.

3 Annealing of a natural diamond single crystal under the conditions described in the Section 2 resulted in the “grafitization” of its surface, whereas the bulk remained dielectric. Yet, we cannot fully disprove the emergence of conductivity within diamond crystallites constituting the defective CVD films.

48

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

Fig. 9. (a) Cyclic voltammogram in 0.01 M Fe(CN)6 4− + 0.01 M Fe(CN)6 3− + 2.5 M H2 SO4 solution for the growth side of the film annealed at Tann = 1910 K. The geometrical surface area 0.15 cm2 ; the potential scanning rate 5 mV s−1 . (b) Ep vs. Tann ((䊊) growth side, (䊉) nucleation side); dashed line shows the theoretical Ep value for reversible electrode reactions.

When discussing the effects of annealing on the electrode properties of polycrystalline diamond, we are aware of the obvious nonuniformity of the electrode surface which must contain active areas (the outcroppings of conducting intercrystallite boundaries) spaced by the dielectric diamond crystallites. The active sites probably occupy but small fraction of electrode surface; nonetheless, we give the values of differential capacitance and current density per 1 cm2 of the geometrical surface. The annealing affects different quantities we have measured through different mechanisms. For example, the differential capacitance obviously is proportional to the fraction of electroactive surface; at the same time, the capacitance depends on the conductivity of the active areas (more precisely, the concentration of charge carriers therein). This complex annealing effect may explain the difference in the “activation energies” found for the differential capacitance (Fig. 5a) and the absorption coefficient (Fig. 5b). The current density for the electrochemical reaction proper (i.e. interfacial charge transfer) is also proportional to the active surface area. However, for relatively rapid reactions the diffusion limitations in solution are brought to the fore, and the situation becomes more complicated. The “graphitized” intercrystallite boundaries are rather narrow (a few nanometer wide), hence, each conducting intercrystallite boundary outcropped at a dielectric diamond surface can be thought of as a band-shaped microelectrode. The annealed polycrystalline diamond surface contains an ensem-

ble of such microelectrodes which interplay in the diffusion process. In the steady-state measurements, the diffusion to each intercrystallite boundary cannot be thought of as independent: the diffusion fronts of separate “microelectrodes” merge and the array of microelectrodes thus operates as a macroelectrode with uniformly active surface, to which linear diffusion takes place. Therefore, the measured current must depend but slightly on the true active area. That is also the reason why the difference in electrochemical behavior between the two sides of the film is found to be not so strong as could be expected when taking into account the much larger total area of “graphitized” grain boundaries on the fine-grained nucleation side as compared to the coarse-grained growth side. For lower Tann , the measured current is small because of low conductivity of the “graphitic” phase formed; and the kinetic control is passing from diffusion to the interfacial charge transfer. The transfer coefficients α and β, as well as Ep , must not depend on the fraction of the active surface. With increase in the conductivity of the “graphitic” phase, the electrode properties improve. Hence, with rising Tann , α and β gradually increased and approached ∼0.5 (Fig. 7); Ep decreased and approached ∼57 mV, a value characteristic of a reversible metal electrode (Fig. 9b). The obtained results in aggregate enable us to conclude that increase in the annealing temperature results both in growth of the amount of nondiamond (“graphitic”)

Yu.V. Pleskov et al. / Electrochimica Acta 49 (2003) 41–49

phase and in change of its properties, first and foremost, conductivity. As the effective activation energy of graphitization of grain boundaries depends on the degree of disorder in these boundaries [10], the required annealing temperatures probably can be reduced for the films grown intentionally with large abundance of defects and incoherent grain boundaries. In this connection we note that an extreme example of highly defective diamond—nanocrystalline (undoped) diamond films, with a dense network of grain boundaries, conductive even without annealing procedure, have been demonstrated recently [18].

5. Conclusions We showed that initially insulating polycrystalline diamond films can be transformed into a conducting material by the high-temperature annealing in vacuum. The conductivity occurs via thin pathways, mainly along graphitized grain boundaries. The strong changes in electrochemical properties of the heat-treated films are observed at the annealing temperatures above ∼1825 K. The sort of composite diamond–“graphite” films can be of interest for electrochemical applications as an alternative both to semiconductor diamond and other carbonaceous-material electrodes, provided it retains in great part the corrosion stability characteristic of diamond materials. Further studies of their properties and long-term tests of stability are in progress.

Acknowledgements This work was carried out financed in part by NEDO International Joint Research Grant Program (Project 01MB9) and Russian Foundation for Basic Research (Projects 01-02-16826 and 01-03-32045).

49

References [1] Yu.V. Pleskov, in: R.C. Alkire, D.M. Kolb (Eds.), Advances in Electrochemical Science and Engineering, Wiley-VCH, Weiheim, 2003, p. 209. [2] S. Eaton, A.B. Anderson, J.C. Angus, Yu.E. Evstefeeva, Yu.V. Pleskov, Electrochem. Solid-State Lett. 5 (2002) G65. [3] Yu.V. Pleskov, A.Ya. Sakharova, M.D. Krotova, L.L. Bouilov, B.V. Spitsyn, J. Electroanal. Chem. 228 (1987) 19. [4] M. Iwaki, S. Sato, K. Takahashi, H. Sakairi, Nucl. Instrum. Methods Phys. Res. 209–210 (1982) 1129. [5] B. Miller, R. Kalish, L.C. Feldman, A. Katz, N. Moriya, K. Short, A.E. White, J. Electrochem. Soc. 141 (1994) L41. [6] C. Reuben, E. Galun, H. Cohen, R. Tenne, R. Kalish, Y. Muraki, K. Hashimoto, A. Fujishima, J.F. Butler, C. Levy-Clement, J. Electroanal. Chem. 396 (1995) 233. [7] Yu.V. Pleskov, M.D. Krotova, V.G. Ralchenko, A.V. Khomich, R.A. Khmelnitskiy, Book of abstracts, in: Proceedings of the 8th International Conference on New Diamond Science and Technology, July 22–26, 2002, Melbourne, Australia, p. 150. [8] D.F. Talbot-Ponsonby, M.E. Newton, J.M. Barker, G.A. Scarsbrook, R.S. Sussmann, A.J. Whitehead, Phys. Rev. B 57 (1998) 2302. [9] A. Khomich, V. Ralchenko, L. Nistor, I. Vlasov, R. Khmelnitskii, Phys. Stat. Sol. (a) 181 (2000) 37. [10] A.V. Khomich, V.G. Ralchenko, A.V. Vlasov, R.A. Khmelnitskiy, I.I. Vlasov, V.I. Konov, Diamond Related Mater. 10 (2001) 546. [11] L. Nistor, V. Ralchenko, I. Vlasov, A. Khomich, R. Khmelnitskii, P. Potapov, J. Van Landuyt, Phys. Stat. Sol. (a) 186 (2001) 207. [12] V. Ralchenko, A. Khomich, R. Khmelnitskii, A. Vlasov, in: N. Veziroglu et al. (Eds.), Hydrogen Materials Science and Chemistry of Metal Hydrides, Kluwer Academic Publishers, Dordrecht, 2002, p. 203. [13] V.G. Ralchenko, A.A. Smolin, V.I. Konov, K.F. Sergeichev, I.A. Sychov, I.I. Vlasov, V.V. Migulin, S.V. Voronina, A.V. Khomich, Diamond Relat. Mater. 6 (1997) 417. [14] L. Nistor, M. Stefan, V. Ralchenko, A. Khomich, D. Schoemaker, J. Appl. Phys. 87 (2000) 8741. [15] A.V. Khomich, V.G. Ralchenko, A.A. Smolin, V.V. Migulin, S.M. Pimenov, I.I. Vlasov, V.I. Konov, J. Chem. Vapor Deposit. 5 (1997) 361. [16] L.G. Johnson, G. Dresselhaus, Phys. Rev. B. 7 (1973) 2273. [17] P. Delahay, New Instrumental Methods in Electrochemistry, Interscience Publishers, New York, 1954. [18] B. Fausett, M.C. Granger, M.L. Hupert, J. Wang, G.M. Swain, D.M. Gruen, Electroanalysis 12 (2000) 7.