Characterization and significance of the sequence of stages of oxide

Langmuir 1992,8, 658-667

658

Characterization and Significance of the Sequence of Stages of Oxide Film Formation at Platinum Generated by Strong Anodic Polarization G. Tremiliosi-Filho,+G. Jerkiewicz, and B. E. Conway* Chemistry Department, University of Ottawa, 142 Louis Pasteur Street, Ottawa, Ontario K1N 6N5, Canada Received February 28,1991. I n Final Form: October 23, 1991 Investigations on the electrocatalysis of anodic reactions such BB 02 and Clz evolution on Pt require information on the state, and changing state, of the oxide film on which these processes take place. At Pt anodes, subjected to regimes of anodization at high controlled potentials for various periods of time, it is found that at least five stages in the process of oxide film formation can be distinguished,in reduction, using linear-sweep voltammetry. The growth of the oxide film at Pt is initially logarithmic in time but, contrary to earlier reports, no attainment of a finite limit of overall 0 %overage” is reached. However, formation of a distinguishable two-dimensional (2-D) monolayer does reach a limit of two equivalent monolayers of “PtO”. When one or two distinguishablestages of phase oxide formation are resolved in addition to the quasi-2-D species, controlled reduction at the negative end of that peak’s profile (with the other phase oxide species not yet reduced),followed by reimposition of a positive-going sweep, regenerates the 2-D oxide film. Hence the latter process is able to take place independently of the presence of the thicker phase oxide film on the surface; the observed behavior indicates that this re-formed 2-D oxide resides beneath the thicker film on the underlying Pt metal surface. During growth of the oxide film at potentials in the range 2.1-2.3 V, vs RHE, through the various stages identifiable in cyclic voltammetry, small but significant changes in the kinetics of 02 evolution arise. 1. Introduction

At Pt and other noble metals, the initial stages of oxide film formation can be studied in detail, including the important transition from submonolayer, through monolayer oxide, to growth of multilayers, corresponding eventually to quasi-bulk-phaseoxide fonnation.l-l0 Knowledge of the states in which the oxide film on Pt can be formed is also of major importance in various continuous anodic Faradaic reactions since it determines the electrocatalytic surface on which 021596v8J0 C1P evolution, hydrocarbon formation (the Kolbe reactionlZ),and some ionic redox reactions13take place. Hence, for studies of such reactions, the states and electrochemical properties of such oxide films require characterization, e.g. by cyclic volta”etry,lOJ1 by optical m e a n ~ . , ~ Jor~by J ~ XPS.16-19 At submonolayer levels of coverage at Pt, Rh, Ru,and Au, a reversibly depositable and reducible oxygen species + On leave of absence from the Universidade de SHo Paulo, at SHo

Carlos, Brazil. (1)Laitinen, H. A; Enke, C. G. J . Electrochem. SOC.1960,107, 773. (2) Vetter, K.;Schultze,J. W. J.Electroanal. Chem. 1972,34,131,141. (3) Ord, J. L.; Ho, F. C. J . Electrochem. SOC.1971, 118, 46. (4) Reddy, A. K. N.; Genshaw, M.; Bockris, J. O’M.J . Chem. Phys. 1968, 48, 671. (5) Damjanovi6, A.; Ward, A. T. J. Electrochem. SOC.1974,121,113C; 1973,126,593d; 1979,126, 555. (6) DamjanoviC, A.; Birss, V. I. J. Electrochem. SOC.1983,130,1688. (7) Gilroy, D.; Conway, B. E. Can. J . Chem. 1968, 46, 875. (8) Gilroy, D. J . Electroanal. Chem. 1976, 71, 257; see also J . Electroanul. Chem. 1977, 83, 329. (9) Angerstein-Kozlowska,H.; Conway, B. E.; Sharp, W. B. A. J . Electroanul. Chem. 1973, 43, 9. (IO) Conway, B. E.; Liu, T. C. Langmuir 1990, 6,268. (11) Roscoe,S. G.;Conway,B. E. J.Electroanal. Chem. 1987,224,163. (12) Vijh, A. K.; Conway, B. E. Chem. Reu. 1967,67,623. (13) Anson, F. C. J . Am. Chem. SOC.1959,81,1554; Anal. Chem. 1961, 33, 934. (14) Gottesfeld, S.; Conway, B. E. J . Chem. SOC., Faraday Trans. 1 1973,69, 1090. (15) Angerstein-Kozlowska, H.; Conway, B. E.; Hamelin, A.; Stoicoviciu, L. J . Electroanal. Chem. 1987,228,429; see also Electrochim. Acta 1986,31, 1051.

can be distinguished v01ta”etrically~~J~ and optical19l4 from oxide material that is irreversibly reducible due to place exchange,4t9after formation a t higher coverages (but initially still less than a monolayer) and/or after holding the potential constant at an appropriate value for controlled times. Results of various workers on the subject of oxide film formation and reduction at Pt have been substantially a t variance; for example, early work indicated that the charge for oxide reduction at Pt was half that for its formation! The kinetics of oxide film growth at Pt were studied by Gilroy and Conway7and later, in extended ways, by Gilrog and by Biegler and Woods.20 From the latter work, a second discrepancy arises in that, with extensive surface oxidation for long times and/or at elevatedpotentials, oxide film formation had been said by Biegler and Woods20*z1 to attain a so-calledz0monolayer limit (supposedly 2.6 0 atoms per Pt), yet, contrarily, Shibatazz*23 had shown that multilayer oxide film formation can arise at Pt, as also found in works by James,24Balej and S ~ a l e kand , ~ ~more recently Vassilev and Gromyko.= Shibata’swork22showed that a distinguishable phase oxide (virtually a bulk oxide film, designated the “p” state in his papers) could be formed on Pt and was reduced in a cathodic current peak appearing (16) Hammond, J. S.; Winograd, N. J . Electroanal. Chem. 1977, 78, 55. (17) Kim, K. S.; Winograd, N.; Davis, R. E. J . Am. Chem. SOC.1971, 93, 6296. (18) Peuckert, M. Electrochim. Acta 1984,29, 1315. (19) Allen, G. C.;Tucker, P. M.; Capon, A.; Parsons, R. J . Electroanal. Chem. 1974,50, 335. (20) Biegler, T.; Woods, R. J . Electroanal. Chem. 1969; 20, 73; 1971, 29, 269. (21) Rand, D. A. J.; Woods, R. J.Electroanul. Chem. 1972, 35, 209. (22) Shibata, S. J. Electroanal. Chem. 1978,89, 37. (23) Shibata, S.; Sumino, M. Electrochim. Acta 1975, 20, 739. (24) James, S. D. J. Electrochem. SOC.1969, 116, 1681. (25) Balej, J.; Spalek, 0. Collect. Czech. Chem. Commun. 1972, 37, 499. (26) Vassilyev, Y. B.; Bagotzky, V. S.; Khazova, 0. A. J. Electroanal. Chem. 1984,181,219; see also Vassilyev, Y. B.; Bagotsky, V. S.;Gromyko, V. A. J . Electroanal. Chem. 1984,178, 247.

0 1992 American Chemical Society

Stages of Oxide Film Formation

at a potential less positive than that for the quasi-2-D oxide film, in fact over the H UPD region. One of Shibata’s significant conclusions was22 that reduction of the quasi-2-D film (“OC1” state here) led to deposition of Pt atoms on the outside of the thicker, phase oxide film. In the present paper, among other matters, we provide a quite direct indication in favor of the quasi2-D state lying beneath that film and reducible independently of it. A similar conclusion had been reached by Gottesfeld et al.28based on 0 2 evolution electrocatalysis and through a paper by James.24 Settlement of this question is important for theories of oxide film growth and electrocatalysis in the Cl2 (cf. ref 11)and 0 2 (cf. refs 27 and 28) evolution reactions. Additionally, in the present work, we (a) investigate further the sequence of stages of anodic formation of the oxide film, distinguishable in its reduction behavior, and (b) show that it is the first quasi-2-D state of oxide formation that reaches a limit of extent of formation while a phase oxide can continue to grow independently with increasing time and potential of anodization and also (c) show the time dependence of anodic 0 2 evolution rates during oxide film growth at constant potential. 2. Experimental Section 2.1. Electrode Preparation. For oxide film growth at Pt, conditions of preparation or pretreatment of the electrode surface are important, as found in several earlier works.22,24,26*~ In the present work, the preferred electrode and solution preparation9J0 procedures were as recommended in the article by Angerstein-Kozl~wska~~ with the addition of a final flameannealing of the Pt. BDH Aristar grade HzS04 was used. 2.2. General a n d Electrochemical Procedures. Oxide films in various stages of formation were generated at Pt electrodes by application of a linear, positive-going potential sweep at 50 mV s-l to a constant controlled potential limit, Eh (the “growth potential”; Figure 1) between 1.8 and 2.3 V, RHE, for various times, th, up to 48 h. During this procedure, purified Nz was bubbled to displace anodically evolved 02 which accompanies oxide film growth above 1.23 V, RHE. The potential vs time program employed is illustrated in Figure 1. Controlled-potential growth is the preferred procedure here as controlled-currentB~32 or cycling procedure^^*^^ provide electrochemically less well defined conditions. In order to avoid cathodic currents due to reduction of 02 formed during Pt oxide formation (as probably arose in ref 26) at E h , the negative-going sweep was arrested at the potential of zero current Figure 1) between 1.60 and 1.75 V, RHE, and held for 30 min. with Nz passing. Separate experiments showed that the anodic oxide films on Pt remained unchanged for up to 48 h on open circuit, in the absence of Hz (cf. ref 23). Cathodic reduction of the anodically formed film was characterized by application of a negative-going potential sweep at 50 mV 8-1 which generated current (i) vs potential (E) profiles in which distinguishable stages of reduction could be resolved. It should be emphasizedthat only by means of the linear potential-

(27)Liu, T. C.; Conway, B. E. Proc. R. SOC.London, A 1990,429,375. (28)Gottesfeld, S.;Yaniv, M.; Laser, D.; Srinivasan, S. J. Phys., Colloq. C5 1977,38(suppl. ll),145. See also Gottesfeld, S. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1989; Vol. 15. (29)Burke, L. D.;Roche, M. B. C. J. Electroanal. Chem. 1982,137, 175. (30)Conway,B. E.; Sharp,W. B. A.; Angerstein-Kozlowska,H.; Criddle, E. Anal. Chem. 1973,45,1321. Shibata, S . Electrochim. Acta 1977, 22, 175. (31)Angerstein-Kozlowska, H. In Comprehensive Treatise of Electrochemistry; Yeager, E., Bockris, J. O’M., Conway, B. E., Sarangapani, S., Eds.; Plenum Publishing Corp.: New York, 1984; Vol. 9,Chapter 9. (32) Shibata, S. Electrochim. Acta 1977,22,175. (33) Burke, L. D. In Modern Aspects of Electrochemistry; White, R. E., Bockris, J. OM., Conway, B. E., Eds.; Plenum Publishing Corp.: New York, 1986; Vol. 18,Chapter 4.

Langmuir, Vol. 8, No. 2, 1992 659

1YIV

1YIv

OD5V

M5V

Figure 1. Potentiodynamic and potentiostatic conditioning programs for oxide film formation and reduction at Pt electrodes: &,polarization (holding)potential; t h polarization (holding)t h e ; Ei-0, zero current potential (Nz bubbling). sweep method (cf. refs 34-36) can major, or more subtle, differences in the states of such oxide films be properly distinguished, since the procedure generates a differentialresponse, i = C(dE/dt), where C is the interfacial capacitance comprising the pseudocapacitance due to oxide reduction and the smaller double-layer capacitance. The distinction of various stages of film reduction arises, in part, for kinetic reasons3’ and we have referred to this as analogous to‘the opening of a Japanese fan”.16 Such effects are involved in the present work where various stages in the reduction of previously formed oxide films can be clearly distinguished in a reproducible way (cf. refs 15,22, and 26). In fact, only by means of linear-sweep voltammetry can such resolution of oxide film reduction behavior be realized. 2.3. Instrumentation. Standard instrumentation was employed for recording linear-sweep voltammograms generated in response to programs of controlled potential (cf. refs 9, 30, 31, and 38) as in Figure 1. 2.4. Charge Evaluation a n d Real Surface Areas. Evaluation of reduction charges, Q, was made by integration of the i vs E profiles, allowing for the double-layercharging contribution and basing the Q values on that for H accommodation (cf. ref 20); overlapping peaks were approximately deconvoluted. Large t h leads to relatively thick films which, upon reduction, generate an increased real-to-apparent surface area ratio (R)of the electrode,sometimes as large as 3.25times the originalrealsurface area (Table I). Since substantial R values can arise, it is most informative to calculate the reduction charges per cmZ,based on the original H accommodation since it is only after the oxide reduction that values of R substantially greater than the original value (ca. 1.59) are generated. Interestingly, the surface willrelax back to within 5% of its original real area after further anodic/ cathodic cycling (referred to as “electrochemicalannealing”) for some time between 0.05 and 1.40 V, RHE, as found in earlier in this laboratory. The changes in R are shown in Table I together with changes in the ratios of principal peak currents ~H.&A~ and iHCl/iHC2 (designated in ref 9) in the UPD H region, indicating surface restructuring.

3. Results and Discussion 3.1. Introductory Remarks. It will be convenient to present the results comparatively, showing a selection of voltammograms which illustrate the principal aspecta of behavior observed a t a sequence of growth potentials, &, from E h = 1.80 to 2.30 V, RHE, and growth times, th, up to 48 h. The cathodic current peaks (or shoulders) are annotated in the order that the reduction peaks arise in the negative-going sweep, viz. OC1, OC2, OC3, etc. 3.2. Oxide Film Growth as a Functionof Potential and Time, and Resolution of Distinguishable Stages of the Reduction Process. For reasons explained in section 2.2, the use of linear-sweep voltammetry enables stages in the reduction of the film to be distinguished in (34)Will, F. G.; Knorr, C. A. Z . Elektrochem. 1960,64,258, 270. (35)Engelsman, K.; Lorenz, W. J.; Schmidt, E. J.ElectroanaZ. Chem. 1980,114, 1. (36)Angerstein-Kozlowska, H.; Conway, B. E.; Hamelin, A,; StoicioViciu, L. Electrochim. Acta 1986,31, 1051. (37)Tilak, B. V.; Angerstein-Kozlowska, H.; Conway, B. E. J. Electroanal. Chem. 1973,48,1. (38)Tellefsen, K. Ph.D. Thesis, University of Ottawa, 1988.

660 Langmuir, Vol. 8, No. 2, 1992

Tremiliosi-Filho et al.

Table I. Real/Apparent Area Ratios (R),Currents and Changes of UPD H Profiles due to Changes of Crystallographic Orientation upon Cycling and Thick Oxide Film Formation electrode description ~ H A ~ I ~ H~ HAC~ I I ~ H CR~ fresh Pt electrode surface, prior to film growth and cycling 2.06 2.52 1.23 1.21 1.56 1.59 electrode cycled for 174 h between 0.05 and 1.4 V RHE (electrochemically ‘annealed”) Pt electrode after multiple oxidationireduction, 1.06 1.32 1.87 followed by prolonged cycling Pt electrode after extended oxide growth and reduction, 1.11 1.92 2.64 but before surface regeneration (2.1V for 12 h; yellow oxide) Pt electrode dter extended oxide growth and reduction 1.16 1.76 5.18 but before surface regeneration (2.0V for 48 h; brown oxide)

Figure 2. Cyclic voltammograms for a Pt electrode; sweep rate = 50 mV 8-1 (298 K). (1) Original profile between 0.05 and 1.40 V. (2) Profile after polarization at 1.8 V for 24 h 57 min; the negative-going sweep is arrested at E,-o (ca. 1.70 v, RHE) for 30 min with N2bubbling, followed by continuation of the negativegoing sweep to 0.05 V; the profile reveals two peaks OC1 and 0c2. (3) As for Profile 2 but for Polarization for 49 h; the Oc2 peak increases in charge and shiftstoward less positive potentials.

a way not achievable by other methods with comparable sensitivity. The general reduction behavior observed, following surface oxidation for various th a t six selected Eh values, 1.8-2.3 V, RHE,is as follows (see following Figures 2-7): the peak OCl (Figure 21, commonly observed,9137 is the first resolvable feature. With extended oxidation, ita charge becomes increased (see Section 3.3 below) and ita peak potential moves to less positive value^.^ Depending criticallyon th and& other stages of reduction, designated as OC2 and OC3 + OC4, are resolved, the latter two being equivalent, it seems, to that designated ‘@” by Shibata.22723Additionally, a new and eventually major broad-shoulder (BS)feature develops (denoted a1 by Shibata22p23).The features of the reduction profiles thus developed will be described sequentially in section 3.5 and, for brevity, are recorded in Table 11. Additionally, the logarithmic oxide film-growth rates (cf. refs. 7 and 361, dQ/d log th, are also recorded in Table I1 for each of the Eh Values. 3.3. Limit to Extent of Formation of the OC1 State. Before proceeding further, we examine Biegler and Woodsm observation that Pt surface oxidation reaches a so-called

feature regular anodic/cathodic i va E profile small changes in anodic and cathodic i va E profiles small changes in distribution of currents in the UPD H profiles distribution of currents in H UPD profiles changes significantly distribution of currents in H UPD profiles changes significantly

t ,,/sec

Figure 3. Oxide film growth plots in log t h at Pt electrodes in 0.5 M aqueous H2S04at 298 K for various polarization potentials (cathodic total charge vs log t h plots): (A) the overall plots; (B) the linear parts of the plots,

monolayer limit of2.6 0 atoms per Pt atom (but evidently not corresponding actually to an 0 monolayer) with increased th. Qualitatively,this is clearly incorrect as Shib a t a ’ ~and ~ ~our , ~ own ~ earlier work7 indicate that oxide film formation continuously increases with log time. The first resolvable feature in oxide film reduction is the commonly 0bserved~9~~ OC1 peak; ita charge becomes increased (Table 111)with th and ita peak potential moves to less positive value^.^.^ Beyond Eh = 2.0 V, the apparent charges for OC1 do increase well beyond the true or the nominal “monolayer” limit20of 2.6 0 atoms per Pt atom, (-a. 1140 pC per real cm2). However, under such conditions, other stages of oxide film formation arise and, upon reduction, give rise to substantial changes of real area as determined by accommodation for H, as noted by Gottesfeld et aL28 Then, when the apparent OC1reduction charges are corrected for these increases of real area, those charges are found to remain essentially constant (Table 111, Figure 81, and close to 880 pC cm-2, independent of Eh in the range 1.8 to 2.3 V, a result which clarifies the hitherto controversial aspect of Biegler and Wood‘s results, vis-a-vis those of refs 8, 22, and 23. Thus it is only the OC12-D film component that reaches (understandably) a limit of ca. 880 pC per real cm2 even up to 2.2 or 2.3 V


Langmuir, Vol. 8, No. 2,1992 661

ii

I

Figure 6. Cyclic voltammograms for a Pt electrode polarized at 2.2 V for different time periods: (A) 2 h 35 min; (B) 4 h; (C) 6 h 30 min. Upon the increase of the polarization time, the total charge for reduction of oxide states does not change significantly but the OC3 state increases at the expense of OC4, which decreases.

11 L 500

oca

LT0c3+0c4

Figure 4. Cyclic voltammograms for a Pt electrode; sweep rate = 50 mV s-l (298 K). (1)Original profile between 0.05 and 1.40 V. (2) Profile after polarization at 2.0 V for 24 h; the negativegoing sweep is arrested at E,=o(ca. 1.70 V, RHE) for 30 min with Nzbubbling, followed by continuation of the negative-going sweep to 0.05 V; the profile reveals peaks for OC1 and OC3 OC4 states, the broad shoulder, BS, and a peak/shoulder with a maximum around 1.52 V (peroxy compounds formed?). Highest curve shows area increase.

+

t

tI l/ I!

ii 2 1/

I ‘

[

2 50

1 --

!

--I ~. .,

-4

I

l

l

..

Figure 5. Cyclic voltammograms for a Pt electrode; sweep rate = 50 mV ~ ~ ( 2 K). 9 8 (1)Original profile between 0.05 and 1.40 V. (2) Profile after polarization at 2.1 V for 18 h 50 min; the negative-going sweep is arrested at Ei-0 (ca. 1.65 V) for 30 min with Nz bubbling, followed by continuation of the negative-going sweep to 0.05 V; the profile reveals peaks OC1 and OC3 + OC4 and the broad shoulder BS. growth potential while concurrent growth processes, leading t o t h e quasi-3-D or a 3-Doxide film, provide the observed continuous, logarithmic increase of oxide film formation with th.

’hOC4 oc2 Figure 7. Cyclic voltammograms for a Pt electrode; sweep rate = 50 mV s-l (298 K). (1)Original profile between 0.05 and 1.40 V. (2) Profile after polarization at 2.3 V for 1h; the negativegoing sweep is arrested at E,,0 (ca. 1.71 V) for 30 min with Nz bubbling, followed by continuation of the negative-going sweep to 0.05 V; the profile reveals peaks OCl,OC2, and OC4 and the broad shoulder BS.

3.4. “Electrochemical Annealing” of Enhanced Area Surfaces Generated following Reconstruction after Strong Surface Oxidation. After reduction of t h e thick oxide film produced at high Eh and large t h , R for the resulting Pt surface can be up to 3.25 times its initial value (see Table I and Figure 4). As was noted by Sharp39in earlier work in this laboratory in collaboration with Kozlowska, it was found, interestingly, that t h e roughened surface could be “electrochemically annealedw back to ita original R (ca. 1.59) by repetitive cycling at 50 mV s-l between 0.05 and 1.40V, RHE, for 12-24 h. Similar effects were noted by Gottesfeld et al.28and must arise from restructuring of a high energy surface of Pt metal (39)Sharp, W.B. A. Ph.D. Thesis, University of Ottawa, 1976.

Table 11. Characteristic Behavior of Reductively Resolved States of Pt S u r f a c e Oxide F i l m s following their Anodic Formation at Various Potentials (Eh) f o r Various Times ~

EhIV

RHE 1.80

features of i v8 E profiles and states resolved OCIonly OC2 appears as a shoulder on OC1 and continues its growth

OC2 becomes a main cathodic peak but is completed positive to H UPD region 1.90

figure no. for i vs E profile 2(cunre 2)

2(curve 3)

reduction peak potentials1 V, RHE Ep(0Cl) = 0.58 Ep(OC1) = 0.55 peak potential becomes less positive with increasing Q; OC2 develops on less pmitive side of OC1 Ep(0C2) = 0.37

Ep(OC1) = 0.55 peak potential becomes less positive with increasing Q

OC1 only; then

Ep(0C3) = 0.27 Ep(0C3) = 0.26 BS over potential range 1.10-0.70

OC3 appears as a sharp peak broad shoulder (BS) features similar to those for 1.90 V but changes develop in shorter times OC1 only; then

BS begins its growth and remains a major feature OC3 begins ita growth; OCI remains a major feature new broad small peak between 1.7 and 1.4 V (peroxy compounds or a new oxide state);OC3 continues ita growth

4

Ep(OC1)= 0.51 peak potential becomes less positive with increasing Q BS over potential range 1.10-0.70 Ep(0C3) = 0.22 1.7e1.40

0 monolayer (decade th)-'

-7.0 h 24.5 h

47

0.106

49.8 h

ca. 4ooo; no limiting value for Q O ~(Fig Z 3)

ca. 9.1

64

ca. 0.147

lo" s

0-4h

84 (for t h < 500 8 ) ;

0.191

beyond this point the increase becomes more rapid 2h >4 h 12 h

ca. 36000; no limiting value for Q

ca. 82

____

other features and features not shown in any of the presented figures but revealed in other experiments up to th N 15 h only the normal cathodic peak, OC1, arises; growth is linearly logarithmic in t h up to ca. 2000 s; beyond this int, the oxide growth rate K o m e a more rapid (Fig 3); OC1 tends to reach a limit; there is no limiting value for Q

up to t h 6 h0nlyOC1 arises; ita growth is linearly logarithmic in t h up to ca. 900s; again,beyond this point, the oxide growth rate becomes more rapid (Fig 3); OC1 reaches a limit, but there is no limit in the growth of OC3; the growth rate of BS is low (compared with values a t higher potential) Fig 4 refers to polarization for t h = 24 h and rev& all mentioned features; the broad small peak between 1.70 and 1.40 V is not due to physically adsorbed or occluded 0 2 since it could not be diminished either by extended NZbubbling or ultraeonifcation of the electrode; it is either a new state of oxide or an adsorbed peroxy species; OC1 reaches a limit again but there is no limiting value for Q;the sub recorded H UPD px:?& substantial increase of the surface area (2X) and some redietribution of the H oxidation peaks indicating changes of crystal surface orientation (Pi 4) polarization for up to 48 h foes'not reveal the growth of OC4

2 ?

2.10

2.20

5

BS over potential

1.5 h

5

range 1.10.65 Ep(0C4) = 0.15

5h

1.70-1.40

6h

Ep(0C3) = 0.22

16.5 h

5

OC3 develops and remains a major feature up to t h = 6 h BS begins its growth and remains a major feature OC4 appears on the less positive side of OC3 and continues its growth it seems to grow a t the expense of OC3 which disappears (at ca. 6 h) broad small peak again between (peroxy compounds) OC3 reappears and continues its growth

more complex behavior and similar to that a t 2.1 V OC1 only OC3 appears and continues its growth OC4 appears and OC3 disappears BS begins its growth and remains a major feature OC3 regrows together with OC4

+

OC3 OC4 reach a limit in Q; OC3 continues its growth but a t the expense of OC4 (Fig 4); BS region becomes again further developed (ca.18% of total Q) 2.30

(t1.5 h

5

Ep(OC1) = 0.47 peak potential hecomes less positive with increasing Q Ep(OC3)= 0.22

OC1 only and remains a major component

similar behavior to that for 2.20 v OC1 retains a limiting charge but is difficult to resolve due to overlap by OC2 (see below) BS continues to be an increasingly dominant feature OC4 becomes main species reducible a t the least positive potential in H UPD region OC2 develops as an extpded shoulder on OC1 which is thereby obscured

5

6 6

0-50 min

Ep(0C4) = 0.14 over potential range 1.10.60 Ep(0C3) = 0.22 Ep(0C4) = 0.14

2h 2h

Ep(OC1)= 0.55 Ep(0C2) = 0.46

7

BS over potential

7 7

S 1 0 0 min

range 1.10.65 Ep(0C4) = 0.12

the OC1 growth is linearly logarithmic up to ca. 200 s; beyond thii point the oxide growth becomes more rapid (Fig 3); again. OC1 tends to a limit; however, there is no limit for Q even though it reveals an inflection; the broad small peak between 1.70 and 1.40 V is not shown in Fig 5 since N2 bubbling was conducted in the region of its reduction and the species was reduced

Q reaches a maximum (Qy 11900pC cm-9 and mflecta a t ca. th = 11h; beyond c a 28 h it continues growth but a t a lower rate (Fig 3)

130 (for t h < 70 8); beyond this point the increase is rapid

ca. 0.295

OC1 growth is linearly logmithpic up to c a 70 s; beyond this pomt there is a region of rapid oxide growth (Fig 3); BS region becomes much more developed with increased polarization time and there seems to be no limit to ita growth

2.6 h 6.5 h

7

ca. 0.242

1.54h

Ep(OC1) = 0.47 Ep(0C3) = 0.22

6

107 (for t h < 200 8); rapid increase beyond this point

inflection in Q vs log th a t Q lo900 pC cm-2, after fast growth, a t ca. 2.5 h

150 (forth < 300 8); beyond this point the increase is rapid 0.5 h 0.5 h

lh

Q reaches a limit of ca. lo900 pC cm-2 and inflects a t ca. 24 h

ca. 0.340

OC3 starts to P O W ( t h 17 h) and OC4 right after; further growth of OC4 Continues a t the expense of OC3 which practically disappears; when OC2 state has formed, OC3 begins to regrow a t the expense of OC4 ( t h = 2 h); the maximum Q is for t h 4 h, since further increase of t h does not change totd charge, (Fig 3) but only changes %e charge distribution among OC1. OC2,OC3,0C4, and BS;after reduction of the thick oxide grown for t h = 1h. the surface area increases ca.1.7X; ale0 H UPD current components are redistributed

2 P


664 Langmuir, Vol. 8, No. 2, 1992

Table 111. Approximate Evaluation of Charges Associated with Partially Resolved Peaks in Negative-Going Potential Sweeps by Deconvolution charge charge charge charge charge electrode description under BS, under OC1,b under OC2 under OC3 under OC4 features and (reference to figure) pC cm-2 pC cm-2 ~ L cm-2 C pC cm-2 pC cm-2 peak potentials 654 OC1- 0.58 V polar. at 1.8 V for 6 h, 57 min 837 621 OC1- 0.55 V, OC2 0.47 V polar. at 1.8 V for 24 h, 57 min (Fig. 2 curve 2) 879 1210 OC1- 0.52 V, OC2 0.37 V polar. at 1.8 V for 49 h, 47 min (Fig 2 curve 3) 894 726 OC1- 0.55 V, OC3 0.27 V polar. at 1.9 V for 10 h (Fig 5, curve 5) 896 864 OC1- 0.55 V, OC3 0.26 V polar. at 1.9 V for 12 h 6200 polar. at 2.0 V for 24 h (Fig 4) 1410 9570 peroxy species ca. 390 pC 0311-2, (OC3 + OC4) deconvolution of OC3 and OC4 impossible due to their superimposition; the roughness factor of the electrode surface increased ca. 2.0X with respect to its initial value.c OC1- 0.51 V, OC3 0.22 V, OC4 0.17 V 14000 polar. at 2.1 V for 18 h, 50 min 2320 1220 6500 peroxy species ca. 16 pC cm-2; deconvolution of OC3 and OC4 (Fig 5) carries an error of up to 10 % ; the roughness factor of the electrode increased ca. 2.6X with respect to its initial value. OC1- 0.47 V, OC3 0.22 V, OC4 0.15 V 1850 2940 deconvolution of OC3 and OC4 polar. at 2.2 V for 6 h 5400 (OC3 + OC4) impossible; the roughness factor of the electrode increased ca. 3.3X with respect to ita initial value. OC1- 0.47 V, OC3 0.22 V, OC4 0.14 V 970” 2270 4330 polar. at 2.3 V for 1h (Fig 7) peroxy species ca. 250 pC cm-2; ( O C l + OC2) the roughness factor of the electrode increases ca. 1.7X with respect to ita initial value. OC1- 0.55 V, OC2 0.46 V, OC4 0.12 V a BS overlaps the region where peroxy compounds are reduced so the calculated BS charges have a significant error (uncertainty) of up to 10%. * As mentioned in section 3 of the Results and Discussion, the OC1 tends to a limit of ca. 880 pC cm-2 (2.0 0 per Pt). The values of the OC1 charge for different Eh and th, shown in this table, are not contrary to the discussion in the text since drastic increase of the roughnees of the Pt electrode causes the increase of OC1charges up to 2940 pC cm-2. The initial roughness factor of the Pt electrode means the roughness it achieved after prolonged cycling between 0.05 and 1.40 V.

-

-

-

1

10’

I

103

102

104

c

th/S@C

Figure 8. Time dependence of the OC1 charge during oxide film growth at Pt (2.0 V), corrected for the real area change that occurs during extended film formation. Polarization at 298 K in 0.5 M aqueous H2SOd.

that is left after reduction of the multilayer film; that surface is evidently microcrystalline (cf. ref 40) rather than amorphoussince submonolayer multiple states of H chemisorption are still clearly resolvable. Special experiments (40) Angerstein-Kozlowska,H.; Conway, B. E.; Barnett, B.; Mozota, J. J. Electroanal. Chem. 1979, 100, 417.

-

-

-

-

-

-

-

showed that this effect is not due to redeposition of dissolved21Pt. 3.5. Sequence of Stages of Formation of the Oxide Film in Relation to Stabilities. The sequence in which stages of oxide film formation are revealed in the negativegoing sweep follows, in general, the previously observed trendgal which, however, requires comment: contrary to what would be expected thermodynamically, oxide species formed a t potentials more positive than those for the submonolayer reversible region (ref 9) are, in fact, reduced only at progressively less positive potentials, e.g. as for the OC1 peak in Figure 2. This behavior was attributede to atotally irreversible process of place-exchange,i.e. once the original 2-D adlayer has become reconstructed, it is reducible only at potentials much less positive than those for ita formation, even at very small reduction rates. For the OC2,OC3 + OC4 reduction stages (Figures 2 and 4-71, this effect is even stronger. Thus, the sequence of reduction processes in the negative-going sweep does not follow the thermodynamic expectations that the higher oxide is reduced at the more positive potential. The distinguishable peaks OC1, OC2,OC3, and OC4 cannot, therefore] represent simply reductions of oxide species in which Pt, as ions, is in different oxidation states. The observed behavior must therefore be interpreted in terms of kinetic effecta41 related probably to the stabilities of whatever oxide structures are sequentially developed.


At Eh = 2.1,2.2, and 2.3 V a complex type of behavior is observed: followinginitial appearance of the OC1 peak, the OC3 peak develops first as a shoulder and continues its growth. However, the OC3 peak diminishes as the OC4 peak begins its growth, and eventually practically disappears (OC4 becoming the new dominant sharp peak). The OC4 reduction charge tends to a limiting value and, after having attained it, the OC3 peak reappears and continues its growth at the expense of OC4 (the total charge, Q, remains almost constant upon increase of th). The thvalues at which these features appear depend on the polarization potential, Eh, and are indicated in Table 11. The peak, OC4, being very sharp, approaches the reduction behavior expected for a single 3-D phase, i.e. reduction at only at singular, thermodynamicallydefined potential. Up to 50 equivalent 0 monolayers are eventually formed, corresponding virtually to a bulk 3-D phase. 3.6. Significance of Distinguishabilityof the Stages of Reduction of Thick Films: Kinetic and Other Effects. The distinction between the OC1 and the other “species” OC2,OC3, and OC4 seems qualitatively easy to make; this is because the OC1 species has the characteristics of a chemisorbed film, reaching a limit in its extent of formation (Figure 81, while the others grow quasiand have bulk-type behavior (cf. independently of it24128 refs 22-25). Also, the OC1 and other thick film species are distinguishable optically.28 However, it is not at all easy to see why the bulk-type film exhibits three distinguishable stages of reduction, separated maximally in average Gibbs energy by ca. 0.36 eV, with at least two being simultaneously present (in addition to the OC1 state) under some conditions (Figures 4,5, and 7 and Table 111). In order to examine the possibility that the above distinguishability of the OC2,OC3, and OC4 stages arises from kinetic effects, i vs Ereduction profiles were recorded at five sweep rates between 5 and 200 mV s-l at films formed for 2 h at 2.3 V. Curves like those in Figure 7 resulted, shiftsof the peaks to less positive potentials arose, as with increase of sweep rate but the point of significance with respect to the distinction between the stages of reduction was that the separation between the two main cathodic peaks (e.g. as in Figure 7) remained almost constant at 362 f 17 mV. Hence this separation (and probably those between other resolved stages) does not seem to be an arbitrary result of kinetic effects but reflects a significant and real difference of “state” of the species being reduced in the sweeps, independent of the rate of change of E. The above behavior observed at Pt could arise for several reasons: (a) polycrystallinity, suggested by grain SEM EDS mapping (section 3.10), although there is apparently (see Figure 6) some uinterconuersion” between the distinguishable stages, depending on th and E h ; (b) as the film grows, there is changing surface-to-volume ratio, that could, speculatively, determine the Gibbs energy of its formation and the energy of activation for reduction. We have made ESCA analyses (cf. refs 1 6 1 9 and 43 and section 3.9) on the state of Pt in the oxide films after various stages of film development but there is no indication that, e.g. in the film that exhibits the OC4 peak, Pt is in a different average oxidation state from that corresponding to OC3 or OC2. A common oxidation state of Pt (+IV) is the main observation. (41) Mozota, J.; Conway, B. E. J. Chem. Soc., Faradoy Trona. 1 1982,

78, 1717.

(42) Angerstein-Kozlowska, H.; Klinger, J.; Conway, B. E. J. Electroanol. Chem. 1977, 75,45. (43) Conway, B. E.; Jerkiewicz, G.; Tremiliosi-Filho, G.;Underhill, R.; Lazier, B. In preparation.

Langmuir, Vol. 8, No. 2, 1992 665

240r

24ot

I I I

I saull;

14nu

QdYsk

1IOY-2mvzlsav

Uw&L

iau+ m u

*,4v

-aiSv

-0mv

+14v

-14nv

600

Figure 9. Cyclic voltammograms for Pt oxide formation and reduction at 50 mV s-’ (2’ = 298 K). (1) Regular cyclic voltammogram taken from 0.05 to 1.40 V, RHE, showing state OC1 in reduction. (2) Extension of potential sweep to 2.0 V, RHE, then held at 2.0 for 12 h, followed by a negative-going sweep arrested at Ei-0 (ca. 1.7 V, RHE) for 30 min with NZbubbling, followed by continuation of the negative-going sweep to 0.33 V but with reversal at that potential giving positive-sweep (continuationof curve 2 but in positive direction) revealing almost normal quasi2-Doxide film formation over potential range 0.8-1.4 V, RHE. (3) Continuation of voltammetry by immediate reapplication of negative-goingsweep from 1.4V, RHE,givingrenewed OC1 peak but with continuation revealing existence of large OC3 + OC4 peaks over the region 0.4-0.1 V, RHE, for reduction of still remaining stable, quasi-3-Doxide states. Continuation of curve 3 in the positive direction (after sweep-reversal 0.05 V, RHE) revealing re-formation of 2-Doxide, similar to that in curves 1 and 2 over the same potential range (0.8-1.4 V, RHE).

3.7. Independent Deposition of the 2-DOxide Film beneath the 3-D Film, following Reduction of the Former. With respect to earlier sections, the aspect of independence of formation of the OC1 species from that of OC2-OC4 (cf. refs 22, 24, and 281, is one of the most interesting results of the present work. An oxide film that would have exhibited both the OC1 and the OC2 or OC3 + OC4 peaks in a reduction sweep was formed by film growth at appropriate E h and th values. If, for such a prepared film, the next reduction sweep was reversed at +0.33 V (see Figure 91, Le. before reduction in the stages OC3 + OC4 had commenced, then this immediately subsequent anodic sweep would trace out more or less the normal i vs E profile for the 2-0 monolayer oxide film having the characteristic fine structureg that would be observed on an initially oxide-free Pt surface. The generation of such an i vs E profile was, however, at a preoxidized Pt surface which, being incompletely reduced beyond the OC1 stage, still retained (in this example) an intact oxide film of some 10.2 equivalent 0 layers corresponding to the OC3 + OC4 peaks that are eventually revealed, e.g. in the profiles of Figure 9, for completion of the reduction in a subsequent sweep taken to +0.05 V, RHE. This result must indicate that once the 3-D type of thick phase oxide has been formed, and the OC1 state “in it” reduced, a renewed 2-D film of OH and/or 0 on Pt can be anodically generated on the Pt metal surface. Since the

666 Langmuir, Vol. 8, No. 2,1992


Figure 10. SEM EDS mapping analysis micrographs: (a, top left) different grains within a Pt polycrystalline sheet covered with Pt oxide; (b (top right), c (bottom left), and d (bottom right)) EDS mapping analysis micrographs after sputtering with 5.0-keV Ar+ ions Pt and bright ones Pt oxide. (Different thickness [different numbers of for 0. 20. and 40 s. remectivelv. Dark mots remesent * monolayers] at different grains”are indicated.)

i vs E profile for this re-formation process shows all the characteristics (cf. ref 9) of a fiirst-sweepprofiie on a initially oxide-free Pt metal surface, it seems that the 2-D film must lie beneath rather than above the thick oxide film. This conclusionwas also reached indirectly by Gottesfeld et al.,28 based on suppositions about the relation of 0 2 evolution electrocatalysisto the properties of the external surface of the oxide film, and more directly from their ellipsometry results.28 The external, thick film must therefore presumably be porous to H20 and protons (cf. ref 29) so that the 2-D film can separately be formed at the inner interface of the thick film with the metal. Shibata had concluded that reduction of his ‘‘a”component (OC1 here) left Pt atoms on the outside of the P-phase thick oxide film, contrary to the indications of the above experiment. Thus, the conclusion presented here is just the opposite of Shibat a ’ but ~ ~consistent ~ with that envisaged by Gottesfeld et

growth law extends to the highest growth potential (& = 2.3 V) used in the present work and its slope, dQ/d log th, continues to increase with Eh except for smaller Q values.11~44~45 At sufficiently long th values, however, the oxide growth rate changes rather suddenly to much larger values (some 400 times; Figure 3), corresponding to appearance of the OC3 + OC4 peaks. This suggests that a new mechanism of oxide film growth takes over for such conditions. This could be the result of the change of electrical conductivity of the film, observed by Shibata,32 so that a high field can become established over the initial oxide film formed to an extent of ca. 1800 pC cm-2. The to be at Q = ca. onset of increased resistivity was 2000 pC cmP2,i.e. similar in magnitude to the film reduction charge associated with the onset of rapid film growth observed in the present work. It is puzzling that thick-oxide growth requires minutes or even hours but such thick films can be reduced within

al.28

3.8. Kinetics of the Oxide Film Growth Processes. Initially, except at coverages below ca. 0.25 (as (?OH),the oxide film growth is linearly logarithmic in th.7 The main point to be stressed here is that the direct logarithmic

(44) Conway, B. E.; Tilak, B. V.; Barnett, B.; Angerstein-Kozlowska, H. J. Chem. Phys. 1990,93,8361. (45) Hadzi-Jordanov, S.; Conway, B. E.; Angerstein-Kozlowska, H.; Vukovic, M. J.Electrochem. SOC.1975,60,359; see also J.Electrochem. SOC.1978,125,1477.


a few or several seconds in a single sweep at 50 mV s-1. This implies that the activation energy of the bulk-type oxide growth is much higher than that for its reduction but the relevant (but unknown) reversible potentials for the processes involved have to be taken into account. We have found, however, that the thick oxide can also be formed on a single sweep but only if much higher polarization potentials are applied at an elevated temperature. Both these observations reflect the effect of electrode potential in the kinetically irreversible formation and reduction of the oxide. 3.9. ESCA Examination. ESCA examination of the oxidation state of Pt in the oxide film (cf. refs 16-19,43) showed that Pt is in the +IV oxidation state when the film gives rise to the OC2, OC3, and 0C443 peaks in reduction. This clearly indicates that the latter peaks do not differ on account of differences of oxidation state of Pt (oxides of different Pt to 0 ratio) so that these stages of oxide reduction must differ, rather, on account of their different ~tabilities.37~~~~~6 3.10. EDS Mapping. EDS mapping analysis results for Pt and 0 show that for the oxide formed on a polycrystalline Pt plate, the film has different thickness at different grains (Figure 10). Sputtering with 5.0-keV Ar+ ions gradually removes the oxide, but differentially, revealing that different grains are covered with oxide to different thicknesses, probably due to different rates of film growth at grains having various exposed surfaces. 3.11. Time Dependence of 0 2 Evolution Rates in Relation to State of Formation of the Pt Oxide Film. During the extended periods of oxide film formation, parallel anodic evolution of 02,of course, takes place.475 The 02 evolution behavior in time is of interest in relation to the changing electrocatalytic properties of the oxide film as it grows and changes its s t a t e as indicated in Figures 2-7 and in relation to the nature of its exposed outer surface.28 The current-densities, io,, for 02 evolution were measured on a Y-time recorder in an attempt to relate them to the course of oxide film development, the currents for the latter being 2.5 hlooks substantial. However, a decrease of activation energy of only ca. 5.5 kJ mol-’ at 298 K would be sufficient to effect the observed maximum increase. Hence the changes in electrocatalytic activity of the oxide surface are really hardly significant.

Langmuir, Vol. 8, No. 2, 1992 667 b

?

22v

23V

E +

200

22v 22v

100

I

P

P

IK, , 0

2

4

,

2.’v

,

,

6

t h / hwrs

, 8

,

-’ ,,, ,

IO”

15

,

, 3S

,

, 55

Figure 11. Rates of 02 evolution, measured as current densities, during an extended period of oxide film formation at various holdingpotentials, (i vs t h ) . Annotationsto curves are as follows: (a) the peak OC3 develops; (b) the peak OC4 develops; (c) the peak OC3 regrows at the expense of OC4 and the total reduction charge remains practically constant.

4. Summary and Conclusions

3. Thick oxide films formed at Pt are stable (in the absence of reducing agents) and remain unchanged for lengthy periods under open-circuit conditions. 4. The distinguishable reduction peaks, OCl,OC2,OC3, and OC4, do not represent simply oxide species in which Pt is in different oxidation states. The observed behavior is to be interpreted in terms both of stability of the oxide species and kinetic effects associated with irreversibility between oxide formation and reduction. 5. When the thick, quasi-3-D oxide film has been formed, the 2-D oxide film can be independently reduced and re-formed at the already oxidized surface with all the characteristics of 2-D film formation on an initially bare Pt surface. This implies that the 2-D film exists and can be re-formed on the Pt metal surface beneath the thick oxide film. 6. Although formation of increasingquantities of surface oxide is a logarithmically slow process, extending over many minutes or hours, the oxide film thus developed can be reduced in a single sweep over a time of ca. 33 s or less. 7. ESCA examination indicates a +IV oxidation state of Pt in films which exhibit the OC2,OC3 OC4 stages of reduction. 8. EDS mapping analyses indicate differential extents of oxide film growth on distinguishable crystal grains a t the surface of smooth polycrystalline Pt. 9. Rates of anodic 02 evolution (at a given potential) depend significantly but only to a relatively small extent on the state of the oxide on which the 0 2 evolution occurs.

1. Potentiostatic anodization of Pt electrodes leads to formation of oxide film which exhibit up to five distinguishable stages (OC1-OC4 and BS) in their reduction. The sequence of development of various stages depends On Eh and th. 2. The overall extent of formation of the oxide film at Pt does not attain any limit; however, the OC1 state does reach such a limit of ca. 880 pC cm-2 which corresponds to 2 nominal monolayers of “PtO” and rationalizes the controversial previous conclusions of Biegler and Woods vis-84s Shibata’s.

Acknowledgment. Grateful acknowledgment is made to the Natural Sciences and Engineering Research Council of Canada for support of this work. We thank Dr. V. I. Birss for communicatingto us some results on the kinetics of platinum oxide film reduction during the course of preparation of this paper. G. Jerkiewicz gratefully acknowledges a Noranda/Bradfield Graduate Fellowship during the tenure of which this and other work was carried out. We also thank Drs. Lazier and Underhill of Alcan International Laboratory, Kingston, for carrying out the ESCA analyses.

(46) Horanyi, G.;Rizmayer, E. M.; Joo,P.J. ElectroanaL Chem. 1983, 152, 211; see also Horanyi, G.; Salt, J.; Nagy, F. J . Electroanal. Chem. 1971, 31, 95.

+

Registry No. Pt, 7440-06-4;PtO,, 11129-89-8;PtO, 1203582-4; HzS04, 7664-93-9;0 2 , 7782-44-7.

Characterization and significance of the sequence of stages of oxide

Recommend Documents