Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.elch.chem.msu.ru/rus/ganzha.pdf Äàòà èçìåíåíèÿ: Fri Oct 8 12:18:47 2010 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 20:29:12 2012 Êîäèðîâêà: ISO8859-5

J Solid State Electrochem (2010) 14:1401-1413 DOI 10.1007/s10008-009-0952-9

ORIGINAL PAPER

The influence of the conditions of the anodic formation and the thickness of Ag(I) oxide nanofilm on its semiconductor properties
Alexander Vvedenskii & Svetlana Grushevskaya & Dmitrii Kudryashov & Sergei Ganzha

Received: 2 September 2009 / Revised: 5 October 2009 / Accepted: 6 October 2009 / Published online: 11 November 2009 # Springer-Verlag 2009

Abstract The anodic formation of Ag(I) oxide nanofilms on polycrystalline silver and Ag-Au alloys as well as on low-index single crystals of silver in 0.1 KOH was examined. By the methods of photocurrent iph and photopotential Eph measurements, the n-type conductivity of Ag2O film was established. Since the film (6-120 nm) is thinner than the space charge region, the dependence of photocurrent and photopotential appears on the film thickness L: iph ~L and Eph ~L2. The transition from polycrystalline silver to single crystals as well as the addition of a small amount of gold (XAu 4 at.%) into the silver lattice decreases the degree of deviation from the stoichiometric composition Ag2O. The parameters of Ag2O film (optical absorption coefficient , donor defects concentration ND, space charge region W, and Debye's length of screening LD) depend on the index of a crystal face of silver, volume concentration of gold XAu in the alloy, and film-formation potential E. At = 0.52 V, the sequences of variation of these parameters correlate with the reticular density sequence. The growth of the potential disturbs these sequences. The band gap in Ag2O formed on Agpoly, Aghkl, and Ag-Au is 2.32, 2.23, and 2.19 eV. Flat band potential in Ag(I) oxide, formed on Agpoly in 0.5 M KOH is 0.37 V. The appearance of the clear dependence between the state of the oxide/metal interface and the structure-sensitive parameters of semiconductor Ag(I) oxide phase allows considering the anodic formation of Ag2O on Ag as a result of the primary direct electrochemical
A. Vvedenskii (*) : S. Grushevskaya : D. Kudryashov : S. Ganzha Department of Physical Chemistry, Voronezh State University, Universitetskaya pl. 1, Voronezh 394006, Russia e-mail: alvved@chem.vsu.ru

reaction, not of the precipitation from the near-electrode layer. Keywords Anodic oxide formation . Nanofilm . Silver . Silver-gold alloys . Photocurrent . Photopotential

Introduction The electrode processes, the oxide formation, in particular, are very sensitive to the state of the metal/solution interface predetermined, first of all, by the electrode crystalline structure, its chemical composition, as well as the potential and the nature of the electrolyte. The AgjOHÐ ?H2 Oî system is typical in this context. It has been reliably established [1-8] that Ag2O is the main stable product of the first one-electron stage of silver oxidation. The formation potentials of Ag(I) and Ag(II) oxides noticeably differ. Hence, it is possible, varying the potential, to examine the kinetics of the growth of namely Ag(I) oxide, while AgO or Ag(OH)2 formation is thermodynamically excluded. The role of the index of a crystal face of silver in the OH- adsorption and initial stages of the Ag2O growth is discussed in [9-12]. In accordance with [10], the potential of the beginning of silver oxidation shifts to positive values in the following sequence: [110] < [100] < [111]. Only for the face [110] is the clear peak on the voltammogram revealed. The authors [10] interpret this peak as a pre-phase oxidation not reaching one monolayer. This points to the enhanced reactive activity of this face with respect to -, which correlates with the results [11, 12] on the electroreduction of oxygen in alkaline solution and the data of our quantum-chemical calculations [13, 14] on the highest energy of Ag110-OH(ads)- bond.


1402

J Solid State Electrochem (2010) 14:1401-1413

The role of chemical irregularity of silver surface, modeled by adding to silver the atoms of another metal, has been examined in a less degree. Gold is considered the most suitable metal for a number of reasons: Ag-Au system is a continuous series of solid solutions; the atomic radii of Ag and Au are very close, and gold is thermodynamically stable at the potentials of Ag(I) oxide formation. It was shown [15-18] that the active dissolution of silver from Ag-Au alloys is controlled by interdiffusion of Ag and Au atoms in the alloy surface layer. However, at the potentials of Ag2O formation, the mass transport in the oxide phase becomes the rate-controlling stage [19-21]. The anodic potential of Ag2O formation shifts to positive values with the growth of gold concentration in the alloy. This fact is assumed to be connected with the initial stage of dealloying [21], preceding the oxide formation, resulting in the surface enrichment by gold.1 Hence, the essential role of the crystal structure and chemical composition of the substrate surface in the kinetics of the anodic silver oxidation at the potentials of Ag2O formation is considered to be reliably established. The less-clear question is connected with the degree and the manner of the revealing of the kinetic features of Ag2O formation, caused by the type of crystal face, gold concentration in the alloy, and the formation potential in the oxide properties, first of all, the semiconductor properties measured in situ. Now the problem of "size factor" arises in the first place, especially actual for nanofilms, where the properties change with the thickness [24-27] and hence, with the kinetic characteristics of the process, the conditions of the film growth, etc. These issues are poorly examined because of the difficulties of determining the partial rates of the processes simultaneously proceeding at the film-formation potentials: 1. oxide formation, 2. metal dissolution via the pores of an oxide layer, 3. chemical dissolution of an oxide phase. The list of structure-sensitive in situ methods of investigation of oxide films is also short; among them, the spectroscopy of photocurrent (PC) and photopotential (PP) occupies an important place [28-32]. Some aspects of these investigations connected with the correlation between the film thickness L and space charge region, the correct estimation of the film thickness, and the interpretation of the photopotential relaxation after the polarization switching off, revealed by us earlier [33-35], remained outside the scope of the research. The second, the more serious problem closely connected with the problem discussed above, concerns the mechanism
1 The evidence of enrichment of the surface of Ag-Au alloy with gold during the dealloying are presented, for example, in [22, 23].

of the formation of an oxide phase on the electrode, either by a direct primary electrochemical growth [3, 8] or via the stages of dissolution/precipitation (upon reaching the dissolution limit) [5, 12]. In our opinion, the revealing of a clear interdependence between structure-sensitive parameters of the oxide nanofilm and the substrate characteristics can be a strong argument in favor of the primary oxide growth. The aim of the work is to establish the character of influence of a silver crystal face, the addition of a small amount of gold, the film-formation potential, and the concentration of hydroxide ions in the solution on the properties of the anodic Ag(I) oxide nanofilms and, on this basis, to define more accurately the oxide formation route.

Experimental Single silver crystals and polycrystalline alloys with 1-15 at.% of gold were prepared from Ag (99.99 wt.%) and Au (99.99 wt.%). The composition and structure of electrodes were monitored by X-ray fluorescence and metallographic analysis. Single crystalline electrodes were grown in a horizontally moving furnace (2 mm/h) at temperatures from 1,273 to 673 K with the subsequent cooling during 24 h. The surface of the electrodes was mechanically ground on grain paper with subsequently decreasing size of the abrasive, polished with water-MgO suspension on suede and rinsed with double-distilled water. The state of the surface was monitored using the Scanning Electron Microscopy (SEM) (JSM-6380LV) and Atomic Force Microscopy (AFM) (Solver P47-PRO) methods. The chemical composition of the alloy surface layer was monitored by XP-block of the Auger microscope -100 (Vacuum Generators, GB) with use of Al-anode (1,486.6 eV; radiation intensity of 200 W) and the source of Argon ions AG62 with accelerating voltage of 2.3 kV and current of 60 A cm-2. X-ray Photoelectron Microscopy (XPS) measurements showed that the surface composition practically does not change as compared with volume one after the treatment described above. Indeed, after the Ag4Au alloy surface had been prepared for the experiment, the concentrations of Ag and Au were 96 and 4 at.%, respectively. Single crystals were mechanically polished with water- MgO suspension and, additionally, chemically polished with HCl-containing saturated chromic acid solution. Before the beginning of the measurements, the state of the electrode surface was electrochemically standardized by a cathodic activation at E = -0.2 V. The solutions 0.1 and 0.5 M reagent-grade KOH were prepared from double-distilled water and deoxygenated in the cell by passing the chemically pure argon. The electrochemical cell was added with a quartz window at the bottom (1 mm thick). In the electrochemical


J Solid State Electrochem (2010) 14:1401-1413

1403

measurements the potentiostat IPC-compact was used. Counter Pt electrode and Ag/AgCl/Cl- reference electrode were used. All potentials are referred to the standard hydrogen electrode scale. In photoelectrochemical measurements, t he lightemitting diodes ?NICHIAË and ?LIGITEKË (l = 385- 875 nm) were used as a source of rectangular light pulses with the duration t = 2 ms and the repetition frequency = 5 Hz. The use of the earthed work electrode, a low-noise amplifier with an active filter of the fifth-order frequency and a digital signal processing (PowerGraph 2.0) made it possible to reduce the noise level to 1-2 V. Photocurrent was measured in the regime of potentiostatic polarization; photopotential was measured in the regime of an open circuit. The sensibility of measurements was about 10 nA for photocurrent and 2-3 V for photopotential. The specific feature of the environment [36, 37] for PC and PP measurements in the regime discrete spectroscopy was the possibility of instantaneous and independent control of the light pulse parameters (t , , and density of light flow 0). We calculated the values of 0 by the density of the optical power P linearly connected with the current of power supply of a light-emitting diode. The bolometer IMO-2 N was used to perform the relevance calibration.
Fig. 1 Cycle voltammograms of silver a and Ag-Au alloys b

Results and discussion Anodic formation of Ag(I) oxide on single silver crystals, polycrystalline silver, and Ag-Au alloys The role of a crystal face of the silver electrode is revealed already in the shape of voltammograms (Fig. 1a). At transition from polycrystalline silver to single crystals, the peak (A) of Ag (I) oxide formation is shifted to positive values by 10 ? 30 mV (it depends on the crystal face); the maximal currents decrease slightly. Besides, the anodic prepeak (A') and the relevant cathodic peak (C') are more clearly revealed on single crystals as compared with Agpoly. The fragment in Fig. 1a demonstrates A' and C' peaks. In [21-23] the appearance of the prepeak is accounted for the bilayer structure of the anodic film. The anodic process can be described in the simplest form: Ag ? OH Ð ! AgOH
Ð ad

! AgOH

ad

?e

Ð

? 1î ? 2î

2AgOHad !2AgOH !Ag2 O ? H2 O

The Eq. 2 combines the stages of 2D- and (or) 3Dnucleation, the growth of nuclei, and the formation of AgOH phase with the subsequent dehydration to Ag2O. The addition of Au into Ag crystal lattice results in the ennoblement of and ' growing with gold atomic share

XAu (Fig. 1b). The shift of peak potentials can be attributed to the alloying, resulting in the decrease of the thermodynamic activity of silver aAg in the volume, hence, at the surface of Ag-Au alloy. The second possible reason of the shift is the decay of XAgs during a short stage of dealloying, preceding t he oxide formation and resulting i n t he enrichment of the surface with gold. Indeed, the results of XPS measurements on Ag4Au alloy after the preliminary anodic oxidation at the potentials of Ag2O formation points to slight enrichment of the surface with gold when the oxide is removed by the argon ions (t 120 s; Fig. 2). It is interesting that gold is fixed at the initial stage of etching. In our opinion, it can be caused by the islet structure of silver oxide, resulting in the appearance of free parts of silver surface. Note that the correlation XO/XAg in the oxide is not constant but monotonously decreases from 0.46 (in accordance with the stoichiometry of Ag2O) to zero. However, since the film is not continuous, XPS signal most probably contains the response of silver substrate. Hence, it is impossible to consider these data as an evidence for the decreasing of the Ag-O correlation during the oxide formation. The SEM data (Fig. 3a, c, d) are in agreement with AFM data (Fig. 3b); they confirm that the silver oxide anodically formed on silver and alloys is porous and consists of the


1404

J Solid State Electrochem (2010) 14:1401-1413

Fig. 2 The results of XPS analyze of the Ag4Au alloy surface after the anodic oxidation at E = 0.60 V. The dotted lines show Ag and Au concentrations in the alloy without anodic oxidation after Ar+ etching during 60 s

separate islets no more than 1 m. The number and size of the islets decrease with the growth of gold concentration in the alloy. The probability of self-dissolution of Ag(I) oxide in an alkaline solution [5, 6] and of silver dissolution through the defects or pores of the oxide film [7] makes it impossible to estimate the film thickness coulometrically, if there are no the data on the current efficiency y of oxide formation. In order to determine the current efficiency on a stationary Ag electrode after the anodic dissolution the cathodic reduction was carried out in a "fresh" solution without Ag+ ions. It was found that the growth of polarization time of Agpoly electrode (hence, the specific anodic charge q and the film thickness L) results in the initial sharp growth of the current efficiency and subsequent stabilization at the certain level not depending on the potential of oxide formation (Fig. 4a). The transition to single crystals does not change the shape of y -q dependences, however, the values of y slightly

Fig. 3 SEM images (a, c, d) and AFM image (b) of the surface of Ag 120 nm formed at E = 0.56 (a, b), 0.60 (b), and 0.77 V(d)

poly

(a, b), Ag4Au (c), and Ag15Au (d) covered by Ag(I) oxide with L =


J Solid State Electrochem (2010) 14:1401-1413

1405

visible electrode surface (square centimeters); is the density of Ag2O (7.14 g cm-3). The values of average thickness of Ag2O on silver as well as on Ag-Au alloys did not exceed 150 nm (Fig. 4b). However, L is noticeably less in fact, since in calculations by (3) we did not take into account the roughness factor of the polished surface of silver and alloys. The latter is usually equal to 1.7-1.9 [38, 39]. The appearance of y -q dependence causes the nonlinearity of L-q dependence, which often is out of scope in coulometric analysis of the average thickness of thin oxide layers. This effect is especially significant for small specific charge and hence, for small thickness (Fig. 4b). When the value of q grows and the value of y stabilizes, the dependence of the average thickness on the charge becomes practically linear (Fig. 4b, fragment). Photocurrent and photopotential in anodically formed Ag (I) oxide The equations for photocurrent and photopotential processing are usually presented for semiconductor structures with thickness L exceeding the space charge region W [30]. The problem of stationary diffusion-migration distribution of the majority and minority carriers in an n-type semiconductor is solved under the following assumptions: donors are completely ionized, a semiconductor zone beyond the space charge region is quasi-neutral, the distribution of n and p in the space charge region in equilibrium state satisfies Boltzmann law, the recombination of carriers in the space charge region during the radiation does not occur. In the case of a thick semiconductor film (L W) the initial system of equations has the form [20]: ! d dp?xî dy ? xî ? p? xî Ì Ð a a1 e dx dx dx ! d dn?xî dy ? xî ? n? xî Ì Ð a a2 e dx dx dx

Fig. 4 The influence of anodic charge on the current efficiency of Ag (I) oxide formation on Agpoly (a) and the film thickness (b)

decrease in general, which is more clearly revealed for small values of q (Table 1). On the alloys, the current efficiency is noticeably less as compared with silver. Nevertheless, a tendency of the gradual increase of the current efficiency with the anodic charge retains, pointing most probably to the decrease of porosity with the thickness. We calculated the average thickness of the anodic film (centimeters) by the formula: AAg2 O Ò Q AAg2 O LÌy Ò q; Ìy zF r zFS r ? 3î

Ða x

?0 x Wî;

? 4î

Where Q is overall charge; AAg2 O is molar weight of Ag2O (232 g mol-1); z =2; F = 96,500 C mol-1; S is the

Ða x

?0 x Wî;

? 5î

Table 1 Current efficiency y of Ag(I) oxide formation, and the film thickness L q/mCcm-
2

= 0.56V Ag
poly

Ag1Au Ag
100

poly

( = 0.57V)

Ag4Au

poly

( = 0.60V)

Ag15Au

poly

( = 0.77V)

Ag

110

Ag

111

2 3 4 5 7

71/2.4 73/3.7 79/5.3 83/6.7 85/10.0

50/1.7 58/2.9 76/5.1 89/7.5 90/10.6

47/1.6 56/2.8 63/4.2 69/5.8 77/9.1

60/2.0 70/3.5 76/5.1 82/6.9 83/9.8

33/1.1 41/2.1 60/4.0 65/5.5 72/8.5

45/1.5 64/3.2 65/4.4 66/5.6 70/8.2

40/1.3 47/2.4 46/3.1 58/4.9 64/7.5

Current efficiency values are presented as percentage of Ag(I) oxide formation (nominator) and the film thickness L, in nanometers (denominator)


1406
d2 y ?xî dx2

J Solid State Electrochem (2010) 14:1401-1413

ÌÐ

e2 h pe ""0 kT 0

Ðy ?xî

? N D Ð n0 e

y ? xî

i ÐNA ?0 x Wî; ? 6î

outer boundary [41]. The relevant refraction coefficients are denoted by Rout and Rin . Dimensionless electric potential ref ref is given by the equation: e8 ? xî 1 ÌÐ 2 y ? xî Ì kT 2LD &
?xÐWî2 L !W ?xÐLî2 ??W2 ÐL2 î; L < W

d2 p? xî Ì Ðaa1 e dx2

Ða x

p?xîÐp0 ? ?W x Lî: L2 p

? 7î

? 8î

Here n0 and p0 are intrinsic concentration of electrons and holes; ND is a volume concentration of donor defects; is a coefficient of optical absorption; a1 Ì h0 =Dp , a2 Ì h0 =Dn , where Dp and Dn are the concentration of diffusion ofÑelectrons and holes; 0 is a light flux density; Ð 1=2 Lp Ì Dp t p is a diffusion length; t p is an averag e out life t ime o f hole s . T he parameter h Ì hf 1 Ð Rref 1 ? Rin eÐaL is e xpre ssed v ia the i nn er ntum qua ref k efficiency and coefficient f Ì~ec = ~ec ? t Ð1 , taking k p into account the difference in the rates of recombination of holes and their consumption in the electrochemical reaction at the oxide/solution interface, characterized by the rate constant ~ec [40]. Besides, h takes into account the possik bility of the light flux refraction from the inner oxide boundary (with respect to the substrate) as well as from the 2 Eph Ì Ð kT 4 ln 1 ? e hf 0 1 Ð Rout 1 ? Rin e ref ref ND D
n

where 8(x; t) is a local electric potential in the oxide film; 1=2 LD Ì ?""0 kT=e2 ND î is Debye's length and is a dielectric constant of the oxide. The properties of the oxide are considered constant along its thickness. For thin films with a thickness less than a space charge region (L < W) the Eq. 7 is redundant and the solution of Eqs. 4-6 is processed in an interval 0 x L. Let us show the final equations obtained in [20] for photocurrent and photopotential in a thin film with a high enough level of light absorption, when LD 1:2
iph Ì ehf 0 1 Ð Rout 1 ? Rin e ref ref
Ða L

Ð

1Ðe

Ða L

Ñ

? 9î

eEph Ñ' &Ð eND Dn 1 Ð eÐ kT e E Ð Efb ? pffiffiffi Ð pffiffiffi Ñ exp Ð ; kT 2LD F L= 2LD

Ða L



?1 Ð e

Ða L

î

Ñ '3 &Ð pffiffiffi e EÐEfb L 5Ñ LD 2F pffiffiffi exp kT 2L D

?10î

Here, is the oxide-formation potential, fb is flat band potential, and F(u) is an integral of Doson. Let us assume that it is not the appearance of photopotential but the generation of electron-hole pairs that makes a main contribution to the photocurrent. Let us consider Rin 1, ref and the second term in brackets under the logarithm is smallpffiffiffi compared with unity. Taking into account F(u) u as at L= 2LD < 1, one can obtain under obvious assumptions:
Ð iph Ì ehf 0 1 Ð Rout 1 Ð e ref
Ð2aL

Ñ

Ìi

max ph

Ð

1 Ð eÐ

2aL

Ñ

% 2ai

max ph

L;

?11î hf 0 1 Ð Rout aL2 ref ND Dn Ñ' &Ð e E Ð Efb exp Ñ kT ?12î It follows from Eqs. 11 and 12 that PC and PP dependence on 0 is linear at low enough light intensity. However, the influence of the film thickness on the photosignal parameters for thin film alters: iph ~L and Eph ~L2. Note that, for thick films, such an influence is absent.

Eph

2kT Ò ÌÐ e

Photocurrent A typical chronoammogram of Ag2O formation on polycrystalline silver is characterized by a fast decay of the anodic current (Fig. 5). In the initial period of potentiostatic polarization (5-7 s), the photocurrent is not registered or does not exceed the noise level. However, starting from a certain time, corresponding to the formation of the oxide film with a significant thickness, the pulse illumination results in the synchronous arising of pulses of the anodic photocurrent (Fig. 5, fragment). The appearance of the anodic photocurrent at the potential greater than the flat band potential3 points to n-type conductivity of semiconductor oxide on Agpoly [28, 29]. The existence of free electrons can be attributed to the prevalence of superstoichiometric Ag atoms or oxygen vacancies in Ag2O crystal lattice. According to the data in [19], the Ag+-ion transport through Ag2O film determines the kinetics of
2 The numerical analysis shows that in thin films with a low level of light absorption (LD < < 1), PP and PC values are negligibly small. 3 In a special set of experiments, the dependence of the capacity of Agpoly/0.5 M KOH boundary on the potential was obtained. The extrapolation in the coordinates of Mott-Schottky gives the value of flat band potential for Ag2O Efb = 0.37 V in accordance with [42], where Efb = 0.33 V


J Solid State Electrochem (2010) 14:1401-1413

1407

Fig. 5 Chronoammogram of Ag E = 0.56 V

poly

electrode in 0.1 M KOH at

oxide formation. Therefore, it is reasonable to assume, that the super-stoichiometric Ag atoms determine the defectiveness of the oxide. Hence, the composition of the oxide can be described by the formula (AgAg)2(Agi)OO. In accordance with Eq. 11, the photocurrent amplitude increases with the growth of charge and with the thickening of the oxide film, reaching the limiting level imax Ì ph ehf 0 1 Ð Rout (Fig. 6a). The shape of iph-L dependence ref does not change in general with a variation of the Ag2O formation potential. Nevertheless, iph slightly increases with E, which is especially noticeable at peak potential (0.56 V). Theoretical curves, calculated by Eq. 11 (dotted lines in Fig. 6), are in agreement with the experimental data, which makes it possible to calculate the coefficient and product f(1-Rout ) (Table 2). The values of W are obtained by ref extrapolat ing i ph-t dependence to the value of imax , ph corresponding to the lack of influence of thickness on the photocurrent and to the condition iph(L)/imax =0.99. Assumph ing that "Ag2 O = 11 for thin Ag2O films[43], we obtained ND Ð Ñ1=2 by the f ormu la W Ì 2""0 EÐEfb =eND and t hen calculated LD and LD. The values of imax and f(1-Rout ) do not practically ref ph depend on the potential of Ag2O formation (Table 2). If we assume = 1 and f = 1, then we formally obtain a very high value of Rout ~0.9926, which is hardly probable. Most ref likely f < < 1 because of a low rate of the electrochemical process of hole assimilation at the oxide/solution interface as compared with the rate of their volume or surface recombination. It follows from Fig. 6 and Table 2 that L < W. However, since LD is 3 ? 4 times less than the space charge region, the correlation between L and LD can be both greater and less than unity, depending on the film thickness. The latter is compared with the depth of optical absorption -1, which means that the entire oxide phase is optically active. At the same film thickness and formation potential, the photocurrent in Ag2O on single Ag crystals is always less than the photocurrent in Ag2O on Agpoly substrate (Fig. 6b). The dependences of photocurrent on the thickness of Ag2O

Fig. 6 The dependence of photocurrent on the Ag(I) oxide thickness on Agpoly (a), Ag100 (b), and Ag-Au alloys (c); the relevant theoretical dependence, calculated using Eq. 11, is presented as a dotted line

films, formed at = const on different faces of Ag single crystals, are similar in shape, if the films are thin. However, their character changes with the film-formation potential. At low potentials ( = 0.52 ? 0.54 V), the photocurrent monotonously grows with Ag2O film thickness. At higher potentials ( = 0.55 ? 0.56 V) the initial growth of photocurTable 2 The dependence of structural and optical characteristics of Ag2O on the potential of oxide formation on Agpoly at l = 470 nm and 0 = 3.56 ç 1015 photon s-1 cm-2 Parameter 0.52V
2

0.53V 4.2 73 0.90 111 256 2.98 72.5 0.65

0.54V 4.3 75 1.08 93 213 4.51 58.9 0.63

0.55V 4.0 70 1.10 91 209 4.90 56.5 0.62

0.56V 4.4 77 2.30 43.5 100 22.4 26.4 0.61

imax /A cm- ph Rout ref

f(1- )ç104 -5 ç10 /cm-1 -1/nm W/nm ND ç10-15/cm-3 LD/nm LD

4.4 77 0.90 111 256 2.83 74.3 0.67


1408 Table 3 The parameter f(1-Rout ), the absorption ref coefficient (at l = 470 nm), and the space charge region in Ag(I) oxide formed on single crystalline silver E/V f(1-Rout )ç 104 ref Ag 0.52 0.53 0.54 0.55 0.56
110

J Solid State Electrochem (2010) 14:1401-1413 ç10-5/cm Ag
111 -1

W/nm
100

Ag

100

Ag

110

Ag

Ag

111

Ag

110

Ag

100

Ag

111

70 77 72 75 77

74 77 75 77 75

72 75 74 75 79

0.08 0.11 0.06 0.38 0.7

0.12 0.30 0.35 0.8 1.4

0.13 0.13 0.18 0. 8 1.0

2,875 2,091 3,833 605 329

1,917 767 657 288 164

1,769 1,769 1,278 288 230

rent is followed by its decay.4 The maximum of PC shifts to a greater thickness in the sequence Ag100 Ag111 Ag110. In this case, the initial section of iph-L dependence was used for a numerical analysis of and imax by Eq. 11 ph for single crystals. The results of calculation are presented in Tables 3 and 4. The parameter f(1-Rout ) practically does not depend on ref the face of single Ag crystal and the film-formation potential. Moreover, the values of this parameter coincide with that presented in Table 2 for polycrystalline silver. At the same time, the optical absorption coefficient in Ag(I) oxide, formed on a single crystalline substrate, is sufficiently less as compared with a polycrystalline substrate. The growth of the potential in an interval 0.52 ? 0.54 V slightly influences . Nevertheless, the further increase of the potential by 10 mV causes a sharp enough increase of the absorption coefficient, proving a significant structural change of the forming oxide and indirectly confirming the bilayer structure of the oxide film. The space charge region W in Ag(I) oxide, obtained at low potentials on single crystals, is sufficiently, almost by an order, greater than in the case of a polycrystalline substrate. The decrease of W with the oxide-formation potential, revealed for a polycrystalline substrate, retains and becomes clearer. The concentration of donor defects in Ag(I)oxide, formed on single silver crystals, is almost ten times less than in Ag(I) oxide, formed on polycrystalline silver. Respectively, the degree of deviation from the stoichiometric composition Ag2O decreases from = 120ç10-8 for Ag2O|Agpoly to = 1 ? 45ç10-8 for Ag2O|Aghkl systems. The gr owth of the anodic potential of Ag(I) oxi de formation leads to the increase of the degree of structure disordering of the oxide phase. The analogous tendency characterizes the change of Debye's length: independent of the crystal face, the values of LD decrease in general with the growth of . At low potentials, LD in the oxide on single crystals is noticeably greater as compared with a polycrystalline substrate. It is demonstrative that the

pr oduct LD, slightly decreasing with the potential, coincides with that obtained for Ag(I) oxide formed on Agpoly (Table 2). Thus, the parameters of anodically formed Ag(I) oxide depend not only on the face of a single Ag crystal but on the film-formation potential as well. Only at = 0.52 V, corresponding to the least (among the reviewed) shift of the potential from Eeq jAg2 OjOH Ð , we can set the following Ag sequences:
a?Ag2 O=Ag110 î < a?Ag2 O=Ag100 î < a?Ag2 O=Ag111 î ND ?Ag2 O=Ag110 î < ND ?Ag2 O=Ag100 î < ND ?Ag2 O=Ag111 î W ?Ag2 O=Ag110 î > W ?Ag2 O=Ag100 î > W ?Ag2 O=Ag111 î LD ?Ag2 O=Ag110 î > LD ?Ag2 O=Ag100 î > LD ?Ag2 O=Ag111 î;

?13î that correlate with the change of a reticular density of crystal faces: [110] < [100] < [111]. The growth of the potential disturbs these sequences:
a?Ag2 O=Ag100 î < a?Ag2 O=Ag111 î < a?Ag2 O=Ag110 î ND ?Ag2 O=Ag100 î < ND ?Ag2 O=Ag111 î < ND ?Ag2 O=Ag110 î W ?Ag2 O=Ag100 î > W ?Ag2 O=Ag111 î > W ?Ag2 O=Ag110 î LD ?Ag2 O=Ag100 î > LD ?Ag2 O=Ag111 î > LD ?Ag2 O=Ag110 î;

?14î Chemical composition of the substrate is also a sufficient factor, influencing the properties of Ag(I) oxide. Ag (I) oxide films formed on Ag-Au alloys at the peak potentials of i-E(t) dependence are characterized, in general, by lower amplitudes of photocurrent, decreasing with the growth of gold concentration in the alloy (Fig. 6c). The character of the photocurrent-thickness dependence is rather complicated. Hence the optical and structural parameters of Ag(I) oxide formed on the Ag-Au alloys were calculated by the initial section of iph-L dependence (Table 5), as was the case in the experiments with single crystals. The growth of XAu results in a sharp decrease of coefficient, the space charge region significantly increases and the concentration of donor defects in the oxide film noticeably decays (Table 5). It means that Ag2O film formed on Ag-Au alloys is more stoichiometric than the film formed on pure silver, which is rather unexpected. The degree of deviation from the stoichiometric composition decreases to d Ì 0:3 Ô 2:2 Ò 10Ð8 .

4

The problem of iph decay in Ag(I) oxide on Aghkl after the film thickness achieves 50 nm needs further investigation in detail.


J Solid State Electrochem (2010) 14:1401-1413 Table 4 Donor defects concentration, Debye's length of screening, and the parameter LD in Ag(I) oxide formed on single crystalline silver E/V ND ç10-14/cm Ag 0.52 0.53 0.54 0.55 0.56
110 -3

1409 LD/nm Ag
111

L Ag
100

D

Ag

100

Ag

110

Ag

111

Ag

110

Ag

100

Ag

111

0.224 0.445 0.139 5.85 20.7

0.50 3.31 4.73 25.9 83.0

0.591 0.622 1.25 25.9 42.3

837 593 1061 163 87

559 217 182 78 43

515 502 354 78 61

0.67 0.65 0.64 0.62 0.61

0.67 0.65 0.64 0.60 0.60

0.67 0.65 0.64 0.62 0.61

The parameter f(1-Rouft ) is the same as for Ag2O films re formed on poly-silver crystals and single silver crystals. Debye's length LD, being less than W, noticeably grows with XAu. The product LD remains close to unity, which makes it possible to consider the film formed on alloys as a semiconductor with a high absorption of light at l = 470 nm. Photopotential After switching off the polarization, the illumination of Ag2O films of different thickness, potentiostatically formed on Agpoly electrode, gives a negative photopotential (Fig. 7), pointing to an n-type conductivity. Independently of the formation potential, illumination intensity and the value of the anodic charge in the interval L = 2 ? 200 nm, the photopotential decreases in time. The mos t par t of Eph-t dependence is lin ear ized in the coordinates (Fig. 8), corresponding to a formal-kinetic equation of irreversible reaction of the first order: h i h i! ln Eph ?tî Ð Est Ì ln Eph ?0îÐ Est Ð k t ph ph ?15î

photopotential amplitude is lower in the whole interval of time as compared with the photopotential in Ag2O formed on a polycrystalline substrate. Meanwhile, the transition from Agpoly to Aghkl practically does not change the value of ~ It is clear, if one takes into account that the dissolution k. of Ag(I) oxide proceeds at the film/electrolyte interface, hence, does not depend on the substrate properties. Independently of the crystal face, formation potential, light intensity, and wave length, a sufficient influence of Ag2O film thickness on the photosignal amplitude is observed. The experimental Eph(0)-L2 dependence is linear (Fig. 9a) in accordance with Eq. 12; the same is valid for Eph(0)-0 dependence (Fig. 9b). Presenting Eq. 12 in another form 2e0 hf 1 Ð Rout aL ref s
n 2

Eph % Ð

Ñ' &Ð e E Ð Efb exp kT

?16î

Here Eph(0) and Est are the values of photopotential at ph the moment of switching off the polarization and of the reaching the quasi-stationary level, ~ is a rate constant of k the process, causing the photopotential decay. Taking into account the dependence of photopotential on the anodic film thickness by Eq. 12 and the results obtained by us earlier in the experiments with RRDE [44], we can confidently assume that the decay of Eph in time is caused by a thinning of the film in the course of its chem ical dissolut ion in accordance with the kinetic ! equation dL/dt = - k L. The negative photopotential is also valid for Ag(I) oxide formed on single Ag100, Ag110, and Ag111 crystals. The shape of Eph-t dependence retains (Fig. 8), though, the
Table 5 Optical and structural parameters of Ag2O oxide formed on Ag-Au alloys ç10-5/cm 0.3 0.1 0.009
-1

and using optical and structural parameters of the oxide film, obtained in photocurrent experiments, we calculated the partial electron photoconductivity n = enn enND from the slope of Eph-L2 dependence at E = const and then the electron mobility n in Ag2O film (Table 6). In the oxide of a constant thickness formed on polycrystalline silver, the partial electron photoconductivity n slightly increases with the film-formation potential. Since n changes less significantly and non-systematically, we can assume that a slight growth of n with is exactly connected with the increase of the concentration of donor defec t s. Note, i n t his c onnec t ion , th at, i n g eneral, the photoconductivity of an n-type oxide is bipolar s Ì s p ? s n . However, the minority carriers (holes in our case) make the main contribution to the overall migration flux in the space charge region under illumination. However, the calculation by Eq. 16 results in the

Au/at.% 1 4 15

E/V 0.58 0.60 0.77

W/nm 767 2,300 25,556

ND ç10

-12

/cm

-3

f(1-Rout )ç104 ref 70 72 72

LD/nm 194 570 4,870

L

D

414 49.7 0.66

0.58 0.57 0.44


1410

J Solid State Electrochem (2010) 14:1401-1413

Fig. 7 Chronopotentiogram of Agpoly electrode in 0.1 M KOH at q = 7 mC cm-2, = 470 nm, and 0 = 3.56 ç 1015 photon s-1 cm-2

values of n, incomparable in principal with a volume conductivity of Ag(I) oxide, determined by the majority carriers. The transition from Ag(I) oxides formed at = 0.56 V on a polycrystalline substrate to the oxides of the same thickness but formed on single crystals shows that partial electron photoconductivity slightly decreases and the electron mobility in the space charge region of the oxide increases. The order of changing of n and n at = 0.53 ? 0.56 V coincides with that established above for , ND, W, and LD (the sequences (14)): s n ?Ag2 OjAg100 î > s n ?Ag2 OjAg111 î > s n ?Ag2 OjAg110 î mn ?Ag2 OjAg100 î < mn ?Ag2 OjAg111 î < mn ?Ag2 OjAg110 îÑ ?17î Silver oxide on Ag-Au alloys for photoelectrochemical investigations was formed at the potential of a main peak in the voltammograms. This potential shifted to positive values with the growth of gold concentration. In all cases, the photopotential is negative. The exponential character of a photopotential dependence on time after switching off the

Fig. 9 The dependence of photopotential on Ag2O thickness (a) and the light intensity (b)

polarization retains as well. The values of n and n are calculated from the linear part of Eph-L2 dependence (Table 6). An increase of gold concentration in a silver crystal to 4 at.% results in a decrease of the partial electron photoconductivity n. A sharp increase of n in the oxide on Ag15Au is hardly demonstrative, since, at the filmformation potential, = 0.77 V, the formation of AgO is possible. The influence of OH- ion concentration on the photopotential In order to define more accurately the route of the oxide formation on silver, the photopotential was measured in the solutions with different KOH concentrations. Preliminary voltammetric investigations show that the peak potentials (EA' and EA) on the anodic branch shift to negative values with the growth of COH-. However, the shift approximately coincides with the difference of equilibrium potentials of Ag2O formation; hence, the overpotential of Ag(I) oxide formation practically does not depend on the concentration of alkaline solution. This conclusion results from a more precise analysis of the influence of OH- concentration (0.01; 0.02; 0.05; 0.07; 0.10; and 0.20 M) on the peak potential in the nitrate-alkaline solution with the constant ionic force of 0.2 mol/dm3. It was found that the slope of the linear dependence of the peak potential on the concentration of OH ions dEA =d lg COH Ð = 0.059 V . ich is in accordance with the Nernst slope wh eq dEAg O=Ag d lg COHÐ .
2

Fig. 8 Photopotential decay in time after the polarization switching off


J Solid State Electrochem (2010) 14:1401-1413

1411

Nevertheless, in the overall i-E dependence, the currents increase with concentration. In order to establish the reason for current growth, either the increase of the rate of the anodic oxide formation or the intensification of silver dissolution with the formation of soluble products, the values of current efficiency in KOH solutions with different concentrations were determined. It has appeared that, at all potentials, the values of noticeably decrease with the growth of OH- ion concentration. Indeed, for the oxide formation in 0.1 M KOH at the potential of the main anodic peak EA = 0.56 V, the current efficiency is equal to 83% (Table 1), while in 0.5 KOH at the potential of the main anodic peak EA = 0.53 V, this value is reduced more than twice and is equal to 38%. Note that the specific charge was the same q = 5 mC cm-2. With the growth of the anodic charge from 5 to 14 mC cm-2, the values of y in 0.5 M KOH increase from 38% to 53%, remaining low enough. One can assume that the growth of alkali concentration results in the formation of a more porous oxide layer and hence, the increase of the rate of the active silver dissolution. The data obtained are important for the solution of the f ol lowing problem: does t he photopotential in anodic Ag(I) oxide depend on COH-? The analysis of Fig. 10 shows that such dependence seems to exist: independent of the wave length the value of Eph noticeably decreases with the growth of electrolyte concentration. Meanwhile, if the main parameters of Ag(I) oxide do not change with COH-, then combining Eqs. 3 and 12, one can obtain the correlation: Eph ?C1 î y ?C1 î Ì y ?C2 î Eph ?C2 î !2 ?18î

Fig. 10 Photopotential in Ag(I) oxide formed on Agpoly with different KOH concentration at = 470 (a) and 525 nm (b); q = 5 mC cm-2

If C1 =0.1 and C2 = 0.5 M KOH, then the right side of Eq. 18 is equal to 4.4. Fig. 10 shows that Eph(C1)/Eph(C2) is equal in average to 3.5 and 2.9 at l = 470 and 525 nm, respectively (the specific charge in all experiments q = 5 mC cm-2). Therefore, one can conclude that the decay of Eph with the growth of COH- seems, resulting from the decrease of the film thickness because of the current efficiency decrease. In fact, the Eph does not depend on the electrolyte concentration. This fact allows one to assume that photoresponse arises exactly in the oxide phase not at the oxide/solution interface.

Note that the rate constant ~ of the oxide chemical k dissolution, obtained in the experiments with different l (400 ? 629 nm), practically does not depend on the concentration of the alkaline solution, being equal to 0.008 Á 0 .001 and 0:008 Ö 0:002sÐ1 atCOH Ð Ì 0:1 a n d 0.5 M OH-, respectively. Thus, the experimentally established influence of the silver crystal face, its alloying with gold and the film thickness on the photoelectrochemical parameters in Ag (I) anodic oxide as well as the lack of such influence on the part of the electrolyte concentration show that the photopotential and photocurrent are connected with the volume but not the s ur face energy levels. Besides , the presence of a clear-cut dependence between the state of the oxide/solution interface and the values of all structure-sensitive parameters of Ag2O semiconductor phase as well as the lack of interrelations between OH-

Table 6 Partial electron photoconductivity n and electron mobility n in Ag(I) oxide formed on different electrodes

Electrode E/V n ç105/Ohm-1 cm- n ç102/cm2 s-1 V-1

Ag

poly

Ag 0.55 3.1 3.96 0.56 9.8 2.73

110

Ag

100

Ag

111

Ag1Au 0.58 2.19 33.1

Ag4Au 0.60 0.43 53.7

Ag15Au 0.77 222 -

1

0.54 2.3 3.21

0.56 4.8 14.5

0.58 2.19 33.1

0.60 0.43 53.7


1412

J Solid State Electrochem (2010) 14:1401-1413

ion concentration and Ag2O parameters demonstrate that the anodic formation of Ag(I) oxide is mainly a result of the direct primary electrochemical growth and not the result of super-saturation of the near-electrode layer with respect to AgOH with a subsequent dehydration of the precipitate. Photocurrent and photopotenial spectroscopy Since photocurrent and photopotential are proportional to the optical absorption coefficient by Eqs. 11 and 12, one can obtain the similar PC and PP spectral dependence at L = const, E =const, and 0 = const, taking into account - correlation [30]: Ð Ñ2=m Ð Ñ iph hn ÌC1 L2=m hn Ð Ebg ?19î Ð Ñ
2=m

Eph hn

ÌC2 L4

=m

Ð

hn Ð E

Ñ
bg

?20î

Here Ebg is a band gap, 1 and 2 are the coefficients, the parameter m is equal to 1 or 4 for direct or indirect optical transition. The dependences of photocurrent on the film thickness at l = 385 ? 875 nm, = 0.56 V, and 0 = 3.56ç1015 photon s-1 cm-2 are similar, which makes it possible to estimate the value of coefficient and to build its spectral dependence (Fig. 11). Most probably, the middle peak (l = 470 nm) corresponds to the range of intrinsic conductivity; the nature of additional peaks is not clear. Band gap Ebg in Ag2O oxide formed on polycrystalline silver is 2.32 eV for direct optical transitions. Spectral dependences of photocurrent generated in Ag2O| Aghkl and Ag2O|Ag-Au systems are similar in shape, and they have three peaks. In the coordinates (19), they show a better linear dependence in the case of direct transitions. Ebg for Ag(I) oxide on Ag-Au alloys is 2.19 eV (Fig. 12a) and does not depend on the alloy composition (in the frame of

Fig. 12 Spectral dependence of photocurrent in Ag(I) oxide on Ag- Au alloys (a) and photopotential in Ag(I) oxide single crystals (b) in the coordinates criterial for direct optical transitions

experimental mistake).5 Band gap for direct optical transitions in Ag2O oxide formed on Aghkl is 2.23 eV, the role of the crystal face practically is not revealed. Spectral dependence of photopotential in Ag2O formed on polycrystalline silver and its single crystals is also characterized by three peaks (Fig. 12b) with the same positions as for photocurrent spectra. The graphic processing of Eph-h dependence in criterial coordinates (20) shows that direct optical transitions prevail; however, the value of Ebg slightly differs from that obtained by photocurrent measurement. Band gap is equal to 2.09 eV independently from the solution concentration.

Conclusions 1. The p redomina nt rout e o f A g(I) oxide anodic formation on silver is not the precipitation from the near-electrode layer, but mainly the di rect electrochemical reaction. 2. The face of a single silver crystal, the chemical composition of Ag-Au alloy and the thickness of the nanofilm of the n-type semiconductor Ag(I) oxide

5

Fig. 11 Spectral dependence of the absorption coefficient in Ag(I) oxide on Agpoly

Note that the accuracy of determination of band gap in Ag2O oxide Á 0.1 eV is low enough to draw conclusions on the influence of the crystal face and alloy composition on Ebg.


J Solid State Electrochem (2010) 14:1401-1413

1413 9. Droog JMM (1980) J Electroanal Chem 115:225 10. Doubova LM, Daolio S, Pagura C, De Battisti A, Trasatti S (2002) Russian J Electrochem 38:20 11. Savinova ER (2006) Razmernye i strukturnye effekty v elektrokatalize. Dissertation, Novosibirsk, Russia 12. Savinova E, Zemlyanov D, Pettinger B, Scheybal A, Schlogl R, Doblhofer K (2000) Electrochim Acta 46:175 13. Nechaev IV, Vvedenskii AV (2008) Sorbtsionnye I khromatograficheskiye protsessy 8:753 14. Nechaev IV, Vvedenskii AV (2009) Prot Met Phys Chem Surf 45:137 15. Marshakov IK, Vvedenskii AV, Kondrashin VY, Bokov GA (1988) Anodnoye Rastvoreniye i Selektivnaya Korroziya Splavov. VSU, Voronezh 16. Pickering HW (1968) J Electrochem Soc 115:143 17. Kaiser H. (1987) Chem Industries, Corrosion Mechanism. New York, Basel, 28: 85 18. Kaesche H (1979) Die Korrosion der Metalle. Springer Verlag, Berlin, Heidelberg, New York 19. Kuznetsova ,Flegel' EV, Vvedenskii AV (2002) Prot Met 38:333 20. Kudryashov DA (2008) Anodnoye formirovaniye i svoistva nanoplyonki oksida Ag(I) na poli-, monokristallakh serebra i Ag, Au-splavakh. Dissertation, Voronezh, Russia 21. Vvedenskii , Grushevskaya S, Kudryashov D, Kuznetsova T (2007) Corros Sci 49:4523 22. Forty AI, Rowlands G (1981) Phyl Mag 43A:171 23. Poate IM (1980) Gold Bull 14:2 24. Roldugin VI (2008) Fizikokhimiya poverkhnosti. Intellect, Moscow 25. Kapusta S, Hackerman N (1980) Electrochim Acta 25:1001 26. McAleer JF, Peter LM (1980) Farad Discuss Chem Soc 70:67 27. Collisi U, Strehblow HH (1990) J Electroanal Chem 284:85 28. Bard AJ, Stratmann M, Licht S (eds) (2002) Encyclopedia of electrochemistry. V.6: Semiconductor electrodes and photoelectrochemistry. Wiley-VCH, Weinheim 29. Zoski CG (ed) (2007) Handbook of electrochemistry. Elsevier, New Mexico State University 30. Pleskov YV, Gurevich YY (1986) Semiconductor photoelectrochemistry. Consultant Bureau, New York 31. Oshe EK, Rozenfel'd IL (1978) Itogi nauki I tekhniki. VINITI. Korroziya i zaschita ot korrozii: 111 32. Finklea HO (ed) (1988) Semiconductor electrodes. Elsevier, New York 33. Kudryashov DA, Grushevskaya SN, Vvedenskii AV (2007) Kondensirovannye sredy I mezhfaznye granitsy 9:53 34. Kudryashov DA, Grushevskaya SN, Vvedenskii V (2007) Prot Met Phys Chem Surf 43:591 35. Vvedenskii A, Grushevskaya S, Kudryashov D, Ganzha S (2008) Surf Interface Anal 40:636 36. Luck'yanchilov AN, Grushevskaya SN, Kudryashov DA, Vvedenskii AV (2006) The environment for the photopotential measurement, Patent no. 55988 RF. Bulletin "Invitations. Useful models" 24: 3 37. Vvedenskii AV, Grushevskaya SN, Kudryashov DA, Luck'yanchilov AN (2007) The environment for the photocurrent measurement, Patent no. 66052 RF. Bulletin "Invitations. Useful models" 24: 3 38. Scheblykina GE, Bobrinskaya EV, Vvedenskii AV (1998) Prot Met 34:6 39. Kozaderov OA, Vvedenskii AV (2005) Prot Met 41:211 40. Wilson RH (1977) J Appl Phys 48:4292 41. Sutter EMM, Millet B, Fiaud C, Lincolt D (1995) J Electroanal Chem 386:101 42. Jiang ZY, Huang SY, Qian B (1994) Electrochim Acta 16:2465 43. Chatterjee K, Banerjee S, Chakravorty D (2002) Phys Rev B 66:854211 44. Kudryashov DA, Grushevskaya SN, Vvedenskii AV (2008) Prot Met Phys Chem Surf 44:301

3.

4.

5.

6.

sufficiently influence the parameters of the photosignal generated not at the electrode/solution interface, but in the volume of the oxide phase. The transition from polycrystalline silver to single crystals as well as the alloying of silver with gold up to 4 at.% result in a decrease of the degree of deviation from the stoichiometric composition Ag2O. The growth of the potential has an opposite effect. When the potential of oxide formation is low, its structure-sensitive parameters (optical absorption coefficient , donor defects concentration ND, space charge region W, and Debye's length LD) change in accordance with the reticular density of atoms on the crystal face. The growth of overpotential of oxide formation disturbs these sequences. The band gap in Ag(I) oxide formed on Agpoly is 2.32 eV, in Ag(I) oxide formed on Aghkl is 2.23 eV, in Ag(I) oxide formed on Ag-Au alloys is 2.19 eV independently from the alloy composition (1, 4, or 15 at.% of Au). The direct optical transitions dominate in all systems at UV irradiation. Flat band potential in Ag(I) oxide formed on Agpoly in 0.5 M KOH is 0.37 V. The initial part of the dependence of the average thickness of Ag(I) oxide film on the anodic charge is not linear because of an essential role of silver active dissolution from the open parts of surface. The decay of photopotential in Ag(I) oxide observed at the growth of the solution concentration from 0.1 to 0.5 M seems because of the noticeable decrease of the current efficiency of the oxide formation.

Acknowledgements We are grateful to Professor Leonid Kazanskiy (The Institute of Physical Chemistry and Electrochemistry of Russian Academy of Science) for the assistance in the XPS investigations. This work is supported by Russian Foundation of Basic Research (project 09-03-00554-a).

References
1. Hampson NA, Lee JB, Morley JR (1971) Electrochim Acta 16:637 2. Clarke TG, Hampson NA, Lee JB, Morley JR, Scanlon B (1969) Ber Buns Phys Chem 73:279 3. Becerra JG, Salvarezza RC, Arvia AJ (1990) Electrochim Acta 35:595 4. Alonso C, Salvarezza RC, Vara JM, Arvia AJ (1990) Electrochim Acta 35:489 5. Ambrose J, Barradas RG (1974) ElectrochimActa 19:781 6. Teijelo ML, Vilche JR, Arvia AJ (1984) J Electroanal Chem 162:207 7. Tilak BV, Perkins RS, Kozlovska HA, Conway BE (1972) Electrochim Acta 17:1447 8. Vidovich GL, Leikis DI, Kabanov BN (1962) Doklady Akademii Nauk SSSR 142:109