Photoelectrochemical kinetics on a copper electrode

ELSEVIER

Journal of Electroanalytical

Chemistry

415 (1996)

161-163

Short Communication

Photoelectrochemical kinetics on a copper electrode V.S. Kublanovsky, Institute

of General

and Inorganic

Chemistry,

Ukrainian Received

G.Ya. Kolbasov, V.N. Belinskii National

2 October

Academy 1995: revised

of Sciences,

32-34

19 March

1996

Pulladin

Ave., 252680 Kiev 142, Ukraine

Abstract The relaxation kinetics of the photoeffect on a Cu electrode coated with oxide film in an alkaline solution have been studied using N, laser pulse excitation (ri = 15ns, Pi = IO4 Wcm-*). Two relaxation regions corresponding to the negative and positive photovoltage (A E) values have been found. The negative A E relaxation region is attributed to the effect of photogenerated majority charge carrier (hole) capture to surface OH- groups bound in the film as the Cu(OH); complex. The positive relaxation region is attributed to the effect at the Cu,Olelectrolyte interface involving minority carriers (electrons). Keywords:

Photoelectrochemistry;

Copper

oxides

1. Introduction It is known that passivating oxide layers are formed on copper in alkaline and weakly acidic solutions. Detailed investigations of these passive layers by electrochemical [l-4], surface analytical (XPS, low energy ion backscattering (ISS) [5]) and ellipsometric methods [6] showed that they may consist of Cu,O or a duplex Cu,O/CuO, Cu(OH), layer depending on formation conditions, and at high potentials a Cu,O, layer is additionally formed. Some papers also report photoelectrochemical properties of copper oxides and passive films on Cu. Schbppel and Gerischer [7] showedthat a cathodic photocurrent reducing Cu,O to Cu occurred on Cu,O crystals. Some authors used this result in the analysis of cathodic photoprocesses on passive films on Cu [S-11]. In Refs. [12-151 the photoelectrochemical properties of oxides formed on Cu during electropolishing in a H,PO, solution were studied. The spectraof the photoeffect on Cu,O films were studied in Refs. [4,12,16]. Ref. [4] showed that on a Cu electrode covered with a Cu,O/CuO film a cathodic photocurrent gives rise to the reduction of CuO to Cu,O. In the same study it was found that the passivating layer, consisting mainly of Cu,O, on Cu in an alkali is a p-type semiconductor with an energy gapwidth of es= 2.3 eV and a flat band potential of Efb = -0.28 V(SHE). In the potential range where CuO formation is possible, Z&, was shown to be - 0.05 V on the basis of the dependenceof the square 0022-0728/96/$15.00 PII SOO22-0728(96)04659-

of the photocurrent on potential, the appearance of a cathodic photocurrent indicating that in this particular case CuO also has a p-type conduction. The fact that Cu,O and CuO are generally p-type oxides was also pointed out in earlier papers[ 17,181.Ref. 1191showedthat a barrier layer is formed on Cu,O due to H,O chemisorption, SO:- and Cl- adsorption decreasingthe barrier height. In addition to this, Ref. [15] revealed an increase in photocurrent, which was coupled with Cl- ions. The aim of this work was to study the kinetics of the photoelectrochemical process in the case of exposure to short laser light pulseson a copper oxide film at an early stage of its formation on polycrystalline and monocrystalline (surface orientation (I 11) and (3 11)) Cu electrodes in an alkaline solution.

2. Experimental The electrode surface was previously polished, degreased and etched in dilute nitric acid. The working electrolyte was an alkali, to which reductants (formaldehyde, Na-hypophosphite) or an oxidant (H202) were added in some experiments. The thickness of the electrochemically grown oxide film was 1 to 3 nm according to coulometric data. Photoelectrochemical kinetics were studied using a pulsed nitrogen laser LGI-21 (h = 337nm, r,, = 15ns); the maximum specific radiant pulse power was

Copyright 0 1996 Elsevier Science S.A. All tights reserved. 1

162

V.S. Kuhlanoosk?:

er ui./Journal

o~Elecrroctna/~r~ctrl

Chemistry

approximately lo4 W cm-’ and was changed with light filters. The measuring device allowed us to study relaxation processes with a time constant of 7 2 80ns. An estimation of the effect of electrode heating during the action of the laser pulse showed, according to Refs. [20,21], that the photovoltage observed experimentally is much higher than the magnitude of the photoeffect caused by heating. The contribution of electron photoemission at negative potentials is also small due to the absence of special electron acceptors in the solution.

415 (1996) I, /mA

161-163

cnY2 80

m”

s,

-1

3. Results and discussion Typical kinetics of the photoeffect relaxation on a Cu electrode are shown in Fig. 1. Two relaxation regions were observed: region I corresponds to the negative photovoltage value and region II to the positive one. For region I the photovoltage rise time ~f,~ did not exceed the resolving time of the measuring device (80ns), the fall time was $2 T 150 to 200 ns; for region II T;,~ = 200 to 300 ns, r,,* - 20 to 50 l.~s. A change from 10 kR to 50 Ilt in load resistance (for an electrode area of around 0.05 cm2 1 had practically no effect on the characteristics of region I and decreased the photovoltage relaxation amplitude and relaxation time in region II at R, < 400 s2. The dependence of A E on illumination intensity did not reach saturation and was sublinear. When the potential was shifted from the stationary value ( Estat = - 0.366 V(SCE)) towards more negative values, the amplitude AE increased in region I (Fig. 2, curve 2) and decreased in region II, then both signals decreased strongly at the same time. A similar picture was observed in the more positive potential region. The photovoltage change AE corresponds to surface copper oxide oxidation or reduction processes. To ascertain these changes, photovoltage values were compared with copper oxide formation potentials in voltammetric

AE/mV

Fio 1. AE photovoltage relaxation kinetics for a monocrystalline Cu eleOdtrode (surface (I II)) covered by an oxide film at E = - 0.25 V(SCE) as a function of solution composition: (1) 7.5molI’ NaOH, (2) 7.5molI-’ NaOH+4Omll~’ 20% CH,O, (3) 7.5moll-’ NaOH+ 46mll-’ 10% H,O,.

I,/

mA cmm2

Fig. 2. Cyclic voltammograms for a monocrystalline copper electrode (surface (3 I 1)) (I) and dependence of negative and positive photovoltage amplitude on potential (2); solution 7Smoll’ NaOH.

curves. Cyclic voltammetric curves exhibit three current peaks in the anodic branch and one peak in the cathodic branch (Fig. 2). In accordance with Refs. [3,22], the first anodic peak corresponds to the formation of monovalent copper oxides and hydroxides; after this the formation of CuO and then Cu,O, is observed, on which oxygenbegins to evolve. The electrode surface orientation had a slight effect on the position of the oxidation and reduction current peaks, but did not affect the character of the photovoltage variation. It is known that for the Cu,O and CuO oxides, which are p-type semiconductors, illumination must shift the stationary potential towards more positive values. The positive A E value in region II of the relaxation curve corresponds to this shift. A comparison of the potential values at which a photoeffect is observed with those corresponding to characteristic peaks in the polarization curve shows the beginning of the appearance of a photovoltage in region II, corresponding to a slight deviation of potential from its stationary value to more positive values where the formation of Cu,O is probable. The appearance of a negative photovoltage (region I in Fig. 1) for a p-type semiconductor (Cu,O) may be accounted for by trapping of minority change carriers (holes) which have an excess of energy in the valence zone of Cu,O and move against the electric field in a film [16,23]. The participation of “hot” charge carriers in photoelectrochemical processes was also discussed for other semiconductor electrodes. As an example, in Ref. [24] it was shown that hot electrons take part in the reduction of p-nitrobenzene on p-b&‘; Ref. [25] points out the contribution of hot electrons to the cathodic photoprocess on n-GaAs. Moreover, according to Ref. [3] the hole-blocking

VS. Kublunovs~

et al./ Journal

oj’Electroanulytica1

layer, which is formed at the Cu,O[copper interface, can make some contribution to the A E value (in this case the negative A E value must decrease with decreasing light wavelength due to an increase in the absorption of light passing through the film). Spectral measurements of AE carried out by us under synchronous detection conditions showed the signal phase to reverse the sign with decreasing light wavelength, indicating an increase in the contribution of negative photo-e.m.f. to the magnitude of the photoeffect. Thus, the negative photovoltage formation mechanism involving majority charge trapped by, for example, surface OH- groups seems to be more probable. The fact that the A E amplitude decreases to zero in region I and increases by a factor of 20 to 30 in region II when H,O,, which oxidizes Cu to Cu,O, is added to the electrolyte and the electrode is held at a negative potential of - 0.30 to - 0.45 V argues in favour of the participation of surface OH- groups in AE formation. That is, since the chemical oxidation of Cu to Cu,O is not accompanied by the implantation of OH- groups in the film, no hole-trapping centres are formed, unlike the anodic oxidation of copper, where this process is very probable. At potentials of E = -0.2 to -0.4V, OH- groups can be bound in the film as the Cu(OH); complex [22] when the following reactions proceed: Cu + OH-+ Cu(OH) +e2Cu(OH) + Cu,O + H,O When OHCu(OH)

ions are in excess, the reaction

+ OH--,

Cu(OH),

is possible. Addition of reductants to the electrolyte resulted mainly in a slight increase in AE amplitude in region II; their effect may consist in the reduction of some amount of divalent copper, forming part of the anodic oxide [8], to monovalent copper, the influence of the reductant on Cu(OH), being small, judging from the value of the A E amplitude in region I. In terms of the model considered, the fast A E rise in region I characterizes hole trapping by surface OH- groups. The fall of A E(T) may be caused by various reasons, e.g. by their recombination with electrons which reached the interface or with electrolyte ions (charge transfer); besides, a compensation of negative photovoltage by a positive effect in the film is probable. Owing to the small oxide film thickness, photogenerated holes may have no time to be thermalized in the valence zone, making possible hot-hole trapping by surface OH- groups, which lie energetically below the upper bound of the Cu,O valence zone. This is similar to the case described for thick Cu,O films [4]. For instance, the time of charge carrier delivery to the interface at a Cu,O film thickness of d-lnmis 7= (d/UT)

= lo-l4

s

where uT = lO’cms- I is the thermal hole velocity. This time is much shorter than the time of Cu,O lattice vibra-

Chemistry

415 (1996)

161-163

163

tion scattering of energy involving optical-phonon emission, 7s = lo- l3 s (in this particular case it is the most efficient energy relaxation channel [16,26]). That is, for a thin Cu,O film the major part of the photogenerated holes having an excess energy can reach, in principle, the interface and subsequently take part in electrochemical processes. Minority charge carriers (electrons) have time to be thermalized in the Cu,O conduction zone, judging from the positive photoeffect rise time, which can take place when there are no hot electron trapping centres at the interface.

Acknowledgements The idea of carrying out this work was suggested by the late Professor A.V. Gorodysky.

References 111B. Miller, J. Electrochem. Sot., 121D.D. Macdonald, J. Electrochem.

116 (1969) 1675. Sot., 121 (1974) 65 1. Ateya, J. Electroanal. Chem.,

117 131 S.M. Abdel Haleem and B.G. (1981) 309. J. Electroanal. Chem., 210 (1986) 141 U. Collisi and H.-H. Strehblow, 213. and B. Titze, Electrochim. Acta, 25 (1980) 839. El H.-H. Strehblow 161J.L. Ord, D.J. DeSmet and Z.Q. Huand, J. Electrochem. Sot., 134 ( 1987) 826. 171 H.R. Schiippel and H. Gerischer, Ber. Bunsenges. Phys. Chem., 75 (1971) 1237. IS1 K. Hauffe and K. Reinhold, Ber. Bunsenges. Phys. Chem., 77 (1973) 310. 191 W. Paatsch, Ber. Bunsenges. Phys. Chem., 81 (1977) 645. 1101M. Yamashita, K. Omura and D. Hirayama, Surf. Sci., 96, (1980) 443. L.M. Abrantes, L.M. Castillo, C. Norman and L.M. Peter, J. Electroanal. Chem., 163 (1984) 209. 1121B. Pointu, M. Braizaz, P. Ponce& J. Rousseau and N. Muhlstein, J. Electroanal. Chem., 122 (1981) 111. Chem., 210 (1986) 229. 1131 M. Novak and A. Szucs, J. Electroanal. Chem., 210 (1986) 237. [141 M. Novak and A. Szucs, J. Electroanal. 1151A. Szucs and M. Novak, J. Electroanal. Chem., 210 (1986) 247. 1161L.V. Belyakov, D.N. Goryachev. A.V. Sachenko and O.M. Sreseli, Fizika and Technika Poluprovodnikov, 220 (1986) 213 (in Russian). [171 C. Wagner and M. Hummten, Z. Phys. Chem., B40, (1938) 197. [I81 I. Blaem, Philips Res. Rep., 13 (1958) 167. 17 (1981) 1830 (in Russian). [191 A.S. Valeev, Elektrokhimiya, DO1 V.A. Benderskii and G.I. Velichko, J. Electroanal. Chem., 140 (1982) 1. [21] SW. Feldberg, J. Electroanal. Chem., 185 (1985) 209. [22] M. Pourbaix, Atlas of Elec&ochemical Equilibria in Aqueous Media, Pergamon Press, New York, 1986, p. 304. [23] A.V. Gorodysky, V.S. Kublanovsky, G.Ya. Kolbasov, V.N. Belinskii and A.N. Antishko, Dokl. Akad. Nauk Ukr., l(35) (1989) (in Russian). 1241 C. Cooper, J.A. Turner and B.A. Parkinson, J. Appl. Phys., 54 (1983) 6463. [25] N.L. Dmitruk, G.Ya. Kolbasov and N.I. Taranenko. Ukr. Khim. Zh., 55 (1983) 1295 (in Russian). [26] J.W. Hodly, TX. Jenkis, C. Schwab, H. Tamura and D. Trivich, J. Phys. C: Solid State Phys., 9 (1976) 1429.