Electrodeposition of epitaxial Cu(111) thin films on Au(111) using defect-mediated growth.

Thin films of various materials are deposited onto the surface of bulk materials (substrates) to achieve properties unattainable in the substrates alone.1 Among the various types of thin films, those with single crystal quality and in registry with the crystallographic orientation of the substrate are called epitaxial films. They have enormous importance especially in the semiconductor industry, because they exhibit reduced or inhibited side reactions with their surroundings, improved metal properties (e.g., the reduced resistivity and electromigration), and atomically smooth layer.2 Compared to the conventional methods for thin film growth in ultrahigh vacuum (UHV), metal deposition in electrochemical environments has such advantages as remarkably low cost and easy handling. However, electrochemical deposition at room temperature often results in a rough polycrystalline film, making it difficult to adapt the electrodeposition to situations that demand high-quality metal films. One strategy to obtain a smoother film has been the use of additives such as thiourea.3 Although the additives make the film flatter and brighter, the resulting film typically takes the form of a polycrystalline film, sometimes with the void inside the deposit, and the additives may contaminate the film during the course of electrodeposition.4 In their pioneering work, Sieradzki et al. proposed a new and general method to perform metal epitaxy in electrochemical environment, which was named the defect-mediated growth (DMG).5 In their scheme, the metal of interest is co-deposited with a reversibly deposited mediator metal (Pb2+ or Cu2+). The mediator is periodically deposited and stripped from the surface by appropriate cycling of the electrode potential, which spawns a high density of metal clusters on a growing layer and leads to layer-by-layer growth. For Ag/Au(111) and Ag/Ag(111) systems with Pb2+ or Cu2+ as mediators, they could obtain epitaxial films with considerable thickness (250 ML or more). While the lattice mismatch between Ag(111) (metallic radius 0.288 nm) and Au(111) (0.288 nm) is negligibly small, that between Cu(111) (0.256 nm) and Au(111) is as large as 11%; thus, the electrodeposition of Cu on Au(111) generally results in rough polycrystalline films. Hence, we were motivated by the question whether the DMG method would also produce a heteroepitaxial film for the film/substrate system with large lattice mismatch. In this paper we present a study on Pb-mediated Cu electrodeposition on Au(111) using various surface tools such as cyclic voltammetry (CV), in situ scanning tunneling microscopy (STM), X-ray diffraction (XRD), and Auger electron spectroscopy (AES). Cu deposition solution was made from ultrapure water (Modulab, U.S. Filter, > 18 MΩ cm), 1 mM Cu(NO3)2 (Aldrich, 99.999%), 0.1 M HClO4 (Aldrich, double-distilled), and 0.1 M Pb(ClO4)2 (Aldrich, 99.995%). Electrodeposition of Cu was performed on a Au(111) working electrode (Metallhandel Schröer GmbH), which was annealed with the H2 flame, by pulse cycles of the electrode potential between 0.050 and 0.400 V versus Pb2+/0 using an Autolab potentiostat. STM images were obtained in constant current mode with a Topometrix TMX1000 while an electrochemically etched Pt/Ir wire (Molecular Imaging) coated with Apiezon wax was used as the STM tip. All images are presented unfiltered. For the ex situ measurements, Cu thin film was prepared as above, rinsed with ultrapure water, and dried in air. The XRD measurement was carried out using a D/Max-IIIc (Rigaku), and the AES was measured using a PHI4300 (PerkinElmer). Figure 1 shows a temporal sequence of STM images for the Cu electrodeposition on Au(111) using the DMG method. The electrode potential was manually pulsed between 0.050 and 0.400 V (vs Pb/0.1 M Pb2+). Initially at 0.600 V, which is the just negative of the Cu upd, Figure 1a exhibits the typical Au(111) terrace on which a number of Cu upd islands are deposited. The Cu upd takes the form of an island because its deposition is extremely slow in perchlorate, where the formation of a full monolayer takes 10-20 min.6 After Figure 1a is acquired, the voltage was changed to 0.400 V. During the acquisition of Figure 1b, a 1-sec potential pulse to 0.050 V is applied to the electrode, which deposits the Pb upd adlayer that acts as a mediator. Then, a potential step back to 0.400 V induces the stripping of the Pb upd adlayer and the continued growth of the Cu adlayer. It is seen that the Cu island density is fairly increased by the deposition and stripping of the mediating Pb upd adlayer. Figure 1b shows that the first Cu adlayer is almost completely deposited and the second one begins to nucleate after another 1-s potential pulse to 0.050 V followed by the Cu deposition at 0.400 V. After application of several more pulses, the Cu adlayers enlarge and grow layer by layer. (Figure 1c) Thicker Cu film is grown, after retracting the STM tip from the surface to minimize tip shielding effect, by repetitive potential pulse cycles (0.050 V for 0.5 s and 0.400 V for 2 s). In Figure 1d, the STM image after 2000 potential pulse cycles shows that the Cu film has grown layer-by-layer with a terrace size of ca. 80 nm. Note the triangular shape of the Cu terraces which is as expected from the three-fold symmetry of the epitaxial Cu(111) thin film. The overall STM observation indicates that the role of the mediating Pb upd adlayer is to spawn a large density of Cu islands,7 which then enlarge and merge together to construct the full Cu adlayer. This is similar to the layer-by-layer growth of Ag on Au(111) with a Pb mediator, in which it was proposed that the deposited Ag atom undergoes interlayer exchange with the underlying upd Pb atom and, after stripping off the Pb upd adlayer, myriads of small Ag clusters remain and continue to grow.5 Also, in the UHV analogue of the surfactant-mediated growth, the homoepitaxy of Cu on Cu(111) surface which is covered with a full layer of Pb resulted in nucleation of a large density of Cu islands on the terrace and layer-by-layer growth.8 To confirm the formation of the epitaxial Cu(111) thin film on the Au(111) substrate, XRD measurements are carried out. * To whom correspondence should be addressed. E-mail: jhkwak@cais. kaist.ac.kr. (1) Smith, D. L. Thin-Film Deposition; McGraw-Hill: New York, 1997. (2) (a) Murarka, S. P. Metallization: Theory and Practice for VLSI and ULSI; Butterworth-Heinemann: Stoneham, 1997. (b) Hwang, R. Q.; Bartelt, M. C. Chem. ReV. 1997, 97, 1063. (c) Neamen, D. A. Semiconductor Physics and DeVices, 2nd ed.; Irwin: Chicago, 1996. (3) (a) Schmidt, W. U.; Alkire, R. C.; Gewirth, A. A. J. Electrochem. Soc. 1996, 143, 3122. (b) Hölzle, M. H.; Apsel, C. W.; Will, T.; Kolb, D. M. J. Electrochem. Soc. 1995, 142, 3741. (4) Plieth, W. Electrochim. Acta 1992, 37, 2115. (5) Sieradzki, K.; Brankovic, S. R.; Dimitrov, N. Science 1999, 284, 138. (6) Schneeweiss, M. A.; Kolb, D. M. Phys. Status Solidi A 1999, 173, 51. (7) Because the Pb upd on Cu(111) shows slow kinetics, (see Brisard, G. M.; Zanati, E.; Gasteriger, H. A.; Markovic, N. M.; Ross, P. N., Jr. Langmuir 1995, 11, 2221) it is possible that the full upd Pb adlayer is not yet deposited within the time scale of deposition (0.5 s). If this is the case, the step edge of the immature Pb adlayer can be another site where Cu deposits preferentially nucleate. (8) Camarero, J.; Ferrón, J.; Cros, V.; Gómez, L.; Vázquez de Parga, A. L.; Gallego, J. M.; Prieto, J. E.; de Miguel, J. J.; Miranda, R. Phys. ReV. Lett. 1998, 81, 850. 7176 J. Am. Chem. Soc. 2001, 123, 7176-7177