Nanoscale studies of Ag electrodeposition on HOPG (0001)

Electrochimuo Acto. Vol. 40, No. 10, PP. 1469-1474, 1995 Copyright 0 1995&evier science Ltd. Printed in GreatBritain. All rightsreserved 0013-4686195 $9.50 + 0.00 ‘IPergamon zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA NANOSCALE R. T. STUDIES P~TZSCHKE,* OF Ag ELECTRODEPOSITION HOPG (0001) C. A. GERVASI,? S. VINZELBERG, G. STAIKOV and W. J. ON LORENZ Institute of Physical Chemistry and Electrochemistry, University of Karlsruhe, D-76131 Karlsruhe, Germany zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP (Received 20 December 1994) zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ Abstract-The electrodeposition of Ag on HOPG (0001) substrate is studied using conventional electrochemical methods (cyclic voltammetry, chronoamperometry) and in situ scanning tunnelling microscopy (STM). Due to a weak substrate-deposit interaction the deposition of Ag occurs according to a VolmerWeber island growth mechanism. At relatively low overpotentials the electrodeposition reaction is mainly initiated on steps and other surface defects. The initial deposition kinetics can be described by a model involving progressive nucleation on active sites and diffusion-controlled 3-D growth. The analysis of experimental results shows that the Catomic Ag cluster represents the critical nucleus in the overpotential interval - 35 mV < q < - 10mV. It is demonstrated that local nanoscale deposition of Ag on flat HOPG (0001) terraces can be induced using STM technique and specific pulse polarisation routines. Key words: STM, opd, nanoscale electrodeposition, nucleation 3-D is land growth. 1. INTRODUCTION 2. EXPERIMENTAL In addition to the determination of surface topography and structure, the local probe techniques find an expanding application field in the nanometerscale surface modificationc I, 23. The electrochemical metal deposition represents a suitable and straightforward approach for local surface structuring and modification. Several studies have shown the feasibility of locally obtaining metal deposits with nanometer-scale dimensions[3-51. An excellent substrate for such studies is the cleavaged (0001) surface of highly oriented pyrolytic graphite (HOPG) which exhibits large atomically flat terraces. The problem of nanoscale surface modification by electrodeposition of metals is directly related to the formation, growth and stability of small metal clusters having quite different properties from those of the metal bulk phase. However, there is at present some degree of controversy between the available experimental evidence and the theoretical understanding of the processes. The aim of this work is to present a comparative study of electrodeposition of Ag on HOPG (OOOl), correlating the morphology of the deposits generated and imaged by scanning tunnelling microscopy (STM) and the kinetics of the deposition process as investigated by means of electrochemical methods. *Author to whom correspondence should be addressed. TPermanent address: Centro de Investigation rollo en Tecnologia de Pinturas (CIDEPINT), Plata, Argentina. CIC Researcher. y Desar1900 La The experiments were carried out in the system HOPG (0001)/10-2 M AgClO, + 1 M HClO,(T = 298 K). The electrolyte solution was prepared from suprapure chemicals and fourfold-distilled water and deaerated by purified nitrogen. Samples of HOPG (Union Carbide) have been used as substrates. A clean (0001) surface was produced by cleaving the HOPG with a piece of sticky tape. Electrode potentials were measured against a Ag reference electrode placed in close vicinity of the HOPG substrate. The ohmic drop was negligibly small in all experiments. In situ STM studies were carried out with a standard Nanoscope III (Digital Instruments) using Pt-Ir tips. The whole STM assembly was placed on an air table to isolate external vibrations and screened by a steel shield grounded at a common point. The electrochemical STM cell and the potentiostatic control of the substrate and the tip have been described in detail previously[6]. 3. RESULTS AND DISCUSSION Figure 1 shows typical cyclic voltammograms for the first and second cycles obtained on HOPG (0001) in the potential range - 120mV < E - ‘ %,,A,+ < 90mV. The absence of underpotential deposition (UPD) is apparent in the voltammograms 1469 1470 R. T. PGTZSCHKE ef al. - q---j---AE- I ‘1 0.5 AE - I 0.3 7 E c;“ E 0.1 0.1 : < . .- . -0.1 E .- -0.1 -0.: -0.3 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA -“$ls j I E-E I I 0.05 -0.05 AgIAg+ .5 0Ill! 5 E-E / v Fig. 1. Cyclic voltammograms for Ag deposition and dissolution. System: HOPG (OO01)/10-2M AgCIO, + 1 M HCIO,, T = 298 K, IdE/dt[ = 20mV s- ‘. (- - - -) first cycle; (. . .) second cycle. indicating a weak deposit-substrate interaction. Furthermore, the rate of overpotential deposition (OPD) of Ag is larger for the second cathodic scan as compared to the one of the first cycle. Moreover, the plating charge density remains larger than the corresponding stripping charge density. This behaviour is similar to that observed by other authors[7, 81 and can be explained in terms of residual Ag deposit on the HOPG surface after the initial potential cycle. Relatively high anodic potentials are required to strip this residual deposit as illustrated in Fig. 2. The cyclic voltammogram for the second cycle shown in Fig. 2 was recorded after an anodic prepolarisation of the HOPG electrode at E - EAg,Ag+= 400mV for t = 5 min. As seen in this case the rate of Ag OPD for the second cathodic scan is smaller than that of the first cycle indicating that during the anodic prepolarisation a complete stripping of the residual Ag deposit is achieved while some of the initial nucleation centers became deactivated. During the next cycles the deposition rate remains unchanged if this prepolarisation routine is applied prior to each cycle. Consequently, with this pretreatment it is possible to bring the electrode surface to a reproducible state. In situ STM observations show that at low overpotentials the deposition process on the atomically flat terraces of HOPG (0001) is extremely inhibited and Ag is deposited preferentially at step edges and other surface defects. Figure 3 shows typical in situ, STM images of a HOPG (Oool) surface before (a) and after (b) a short polarisation in the low overpotential range. As seen three-dimensional (3-D) clusters are formed at surface defects during the initial stages of Ag deposition. Figure 4 shows in situ STM images of a stepped HOPG (0001) surface after a pol&isation in the UPD and OPD ranges at potentials corresponding to the anodic (a) and cathodic (b) limits of the cyclic voltammograms in Figs 1 and 2. The changes in the surface profile according to the I -0.05 0.05 AgIAg+ C .5 J v Fig. 2. Cyclic voltammograms for Ag deposition and dissolution. System: HOPG (OOOl)/lO-* M AgClO, + 1 M HClO,, T = 298 K, IdE/dtl = 20mV s-‘. (-) first cycle; (. .. .. .) second cycle after anodic prepolarisation at DE = 400mV for 5 min. plating/stripping sequence can be clearly seen in the corresponding cross-sections of these STM images. Deposited silver fills the stepped gaps on the initial surface while nucleation and growth of islands on flat terraces is inhibited in this potential range. The lack of VPD phenomena in the voltammetric experiments and the observed morphology of growth indicate that the early stages of the electrochemical deposition of Ag on HOPG (0001) correspond to an island (Volmer-Weber) growth mechanism[7-91. The nucleation and growth kinetics were studied using potential pulse methods. The current transients were recorded after a step of the substrate potential from the UPD range (AE = E - EAgIAl+> 0) into the low OPD range (q = E - EAEiAs+< 0). Prior to each pulse experiment, a reproducible state of the HOPG (0001) surface was established by polarising the electrode at AE = 400 mV for about 5 min. Figure 5 shows current-time transients measured at different overpotential AE. The deposition rate increases strongly with increasing 1. The analysis was restricted to the initial parts of the transients at short times in order to exclude overlapping effects of the growing clusters. The cathodic transient current density under these conditions is defined as ifree. In Fig. 6 (L) 2’3 is plotted vs. t. The obtained linear relationship corresponds to a nucleation and growth model including progressive nucleation and hemispherical diffusion to the growing 3-D clusters[lO111. According to this model the ifree- t dependence is given by zyxwvutsrqponmlkjihgfedcbaZYXWVUT lfree = -zF ; 7rv;‘J(2D~,,,+)~‘* x [I -exp~~)-J’*(r-LJ3/* (1) Nanoscale studies of Ag electrodeposition on HOPG (0001) 1471 Fig. 3. In situ STM images of a HOPG (0001) substrate surface. System: HOPG (0001)/10~2 M AgCIO, T = 298 K, tunnelling current I, = 5nA, Pt-Ir tip. (a) Bare substrate surface at + 1M HCIO,, AE = 100 mV; (b) 3-D Ag clusters deposited on the substrate at q = - 20mV for 10 s. where v, is the molar volume D the diffusion coefhcient of metal ion bulk concentration, period. The nucleation rate defined by[12-151 J(u) = K(rl) exp i of the bulk metal Me, Me’+ ions, cMez+the and t, the induction J in equation (1) is AG,h) k~ 1 where AG,(rl) is the Gibbs free energy of formation of the critical cluster (nucleus) containing N, atoms. The pre-exponential factor K(q) is only weakly dependent on rl and includes the number of active sites where the nucleation proceeds and the attachment frequency of single atoms to the nucleus. The nucleation rate of Ag was calculated according to (1) from the slopes of the linear (ifrcej2,Jvs. t plots in Fig. 6 using v, = 10.27cm3 mall’ and D = 10m5cm’s_ ‘. The overpotential dependence of the nucleation rate is shown in Fig. 7. The number of atoms N, in the critical nucleus can be obtained using the equation[ 163 In equation (3) the overpotential dependence of the pre-exponential term K(q) in (2) is neglected since it affects N, only by less than one atom. From the slope of the linear log J vs. 1q 1 dependence shown in Fig. 7 a value of N, = 4 was derived using equation (3). This small number of atoms in the critical nucleus and the linear relationship between log J and lrl[ show that in this case the nucleation process can be described by the so-called atomistic model[12-151. It is worth mentioning that a 4-atomic critical nucleus was also found by Milchev et aI.[14, 171 for Ag electrodeposition on glassy carbon in the same overpotential interval as in this work indicating that the nucleation sites on both substrates have approximately the same activity. 1472 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA R. T. P~TZSCHKEet al. (a) @I AE=lOOmV rl=- zyxwvutsrqponmlkjihgfedcbaZYXWVUTS 125 mV Fig. 4. In situ STM images of a stepped HOPG (0001)/10-2 M AgCIO, + 1 M HCIO,, T = 298 K, tunneling current I, = 5 nA, Pt-Ir tip: (a) At the anodic potential limit of the voltammograms in Figs 1 and 2; (b) After Ag deposition at the cathodic potential limit of the voltammograms in Figs 1 and 2. zyxwvutsrqponmlk 0.8 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA HOPG(0001) / Ag+ 0.2 0 0 2 4 6 8 Fig. 5. Current transients for Ag zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA opd on HOPG (0001) at different q. System: HOPG (OOOl)/10-2M AgCIO, + 1 M HCIO,, T = 298 K, initial underpotential ASi = 1OOmV. Nanoscale studies of Ag electrodeposition on HOPG HOPG I473 zyxwvutsrq (0001) (OOOlUAg’ E 2 -3 m’ 0 t 2 Fig. 6. $,/a vs. 3 4 t plots of the initial parts of the transients 5 shown linear regression. The fact that the nucleation process on flat HOPG (0001) terraces is strongly inhibited makes the system under consideration very appropriate for localised electrodeposition of Ag in the nanometer range using STM techniques. A polarisation routine similar to the one used by Penner et al.[3] involving a tip positive bias step in a two-electrode STM-cell configuration was applied to produce nanoscale Ag clusters at a desired position on the HOPG surface. Figure 8 shows the substrate surface without modification (a) and after induced local deposition of two Ag clusters on a flat terrace (b). The deposition of each cluster was produced by applying a single tippositive 6V bias pulse with 0.1 ms duration. A cross section in Fig. 8 depicts the depth profile of these two disc-shaped clusters ca. OSnm in height and 1Onm in diameter. In accordance with the observations of Penner et aI.[3] the local deposition process is most probably preceded by formation of a pit on the graphite substrate which acts as an active site for the nucleation process. HOPG(0001) Fig. 7. Nucleation 6 7 in Fig. 5: a ( .) experimental (-) 4. CONCLUSIONS Due to a weak substrate-deposit interaction the of Ag occurs on UP&unmodified HOPG (0001) according to a Volmer-Weber island growth mechanism. At relatively low overpotentials (q > - 150mV) the electrodeposition reaction is mainly initiated on steps and other surface defects. High underpotentials (AE > 400mV) are required to strip the Ag deposit. The initial deposition kinetics can be described by a model involving progressive nucleation on active sites and diffusion-controlled 3-D growth. The analysis of experimental results shows that the 4-atomic Ag cluster is a critical nucleus in the overpotential interval -35 mV < q < - 10mV. Local nanoscale deposition of Ag on flat HOPG (0001) terraces can be induced using two-electrode STM-cell configuration and polarisation routines involving a tip positive bias step. zyxwvutsrqponmlkjihgfedcbaZYXW OPD ! Ag’ rate, J, determined data; from data in Fig. 6. as a function of TV 1474 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA R. T. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH P~TZSCHKE et al. (b) (a) STM Image Heic G-1 8 2.00 50.0 nn 0 50.0 Data tlJpe z range rlr Section n* Height 2.00 nn Analysis Fig. 8. In situ STM images of a HOPG (0001) substrate surface. System: HOPG (0001)/10-2 M AgCIO, + 1 M HClO,, T = 298 K, tunneling current I, = 5nA, Pt-Ir tip: (a) Bare substrate surface; (b) Ag clusters locally deposited on a flat substrate terrasse applying tip-positive 6V bias with 0.1 ms duration; (c) Cross section of(b), showing the shape of the formed Ag clusters. REFERENCES 1. C. F. Quate, in Scanning Tunneling M icroscopy Related M ethods (Edited by R. J. Behm, N. Garcia 2. 3. 4. 5. 6. 7. 8. K. Itaya and E. Tomita, 9. G. Staikov, K. Jiittner, and and Dor- Surf Sci. 201, L507 (1988). W. J. Lorenz and E. Budevski, Hectrochim. Acta 39, 1019 (1994). 10. J. A. Harrison and H. R. Thirsk, in Electroanaly tica H. Rohrer), p. 281. Kluwer Academic Publishers, Chemistry (Edited by A Bard), Vol. 5, p. 67. Marcel Dekker, New York (1971). drecht, Boston, London (1990). zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Il. H. Bort, K. Jiittner, W. J. Lorenz, G. Staikov and E. 10 Years of STM (Edited by P. Descouts and H. Budevski, Electrochim. Acta 28,985 (1983). Siegenthaler). North Holland, Amsterdam (1992). W. Li, J. A. Virtanen and R. M. Penner, Appl. Phys. 12. A. Milchev, S. Stoyanov and R. Kaischew, Thin Solid Films 22, 255 (1974). Lett. 60, 1181 (1992). and G. Staikov, Ann. Rev. W. Li, J. A. Virtanen and R. M. Penner, J. Phys. Chem. 13. E. Budevski, V. Bostanov M ater. Sci. 10, 85 (1980). % , 6529 ( 1992). R. Ullmann, T. Will and D. M. Kolb, Chem. Phys. Letl. 14. A. Milchev, Contemp. Phys. 32, 321 (1991). 209,238 (1993). 15. G. Staikov, in Proc. NATO ASI Nanoscale Probes of the Solid/Liquid Interface. Sophia Antipolis, France, M. Hiipfner, W. Obretenov, K. Jiittner, W. J. Lorenz, July lo-20 (1993), in press. G. Staikov, V. Bostanov and E. Budevski, Surface Sci. 248,225 (1991). 16. D. Kashchiev. J. Chem. Phys. 76,5098 (1982). R. Sonnenfield and B. Schardt, Appl. Phys. Left. 49, 17. A. Milchev and E. Vassileva, J. electroanal. Chem. 107, 337 (1980). 1172 (1986).