Galvanic metal plating is a widely used industrial method for the formation of metal structures with sub-micrometre dimensions. Its applications include dual-damascene plated ULSI microchips, magnetic storage devices, and micromechanical devices (LIGA process). The continuing need for smaller feature sizes demands even better control of these electrochemical deposition processes and hence an improved fundamental understanding of the underlying elementary steps. A central challenge is to clarify the relationship between the atomic-scale structure of the solid-liquid interface, the growth behaviour and the resulting surface morphology, which requires direct structure-sensitive investigations during the growth process.

To provide a deeper insight into this complex interplay, metal electroplating from aqueous electrolyte solutions was investigated in situ by surface X-ray diffraction, using Au(001) homoepitaxial growth in 0.1 M HCl as an example. The results demonstrate that studies of the kinetic growth behaviour at a solid-liquid interface are possible by this technique and reveal a pronounced influence of the applied electrode potential on the growth process.

Experiments were performed at beamline ID32 (EPh = 18-20 keV) using a “hanging meniscus” transmission electrochemical cell with minimised electrical cell resistance and nearly unrestricted mass transport, which allows the combination of in situ surface X-ray diffraction studies of rapid structural changes with high quality electrochemical measurements [1]. The electrodeposition process was initiated by exchanging the Au-free solution via a remote controlled pump system by electrolytes containing 0.1 to 0.5 mM HAuCl4, resulting in diffusion-limited (i.e., potential-independent) deposition rates up to four monolayers (ML) per minute.

To study the growth behaviour, the sample potential was first kept at 0.6 VAg/AgCl, where the very high surface mobility results in a rapid smoothening even of rough surfaces, and then stepped to a more negative value while the intensity (I) of the scattered X-rays was monitored as a function of time at selected reciprocal space positions along specular and nonspecular crystal truncation rods (CTR). In an analogous way, as in MBE growth studies under UHV conditions, different kinetic growth modes become manifest in characteristic I(t) curves. Specifically, in a wide potential regime intensity oscillations (Figure 90) indicating layer-by-layer growth were observed (typically only 3-4 oscillations due to imperfect 2D growth). The oscillation period is in very good agreement with the deposition time per Au monolayer obtained from parallel electrochemical measurements, supporting the interpretation of these data as growth oscillations.

Fig. 90: Growth oscillations indicating layer-by-layer growth for Au electrodeposition on Au(001) in 0.1 M HCl + 0.5 mM HAuCl4, after potential steps into the regime of the unreconstructed (1x1) (blue line) as well as the “hex” reconstructed (red line) Au surface. The insets show high-resolution STM images of both surface structures.

As illustrated by the kinetic growth mode diagram in Figure 91, the growth mode depends strongly on the potential as well as on the Au deposition rate. With decreasing potential, regimes of step-flow growth, layer-by-layer growth, multilayer growth, and a “reentrant” layer-by-layer growth are observed. This behaviour can be rationalised within the framework of kinetic growth theory by the pronounced influence of the potential on the interface structure and consequently the surface transport processes. The potential-dependent growth behaviour in the positive potential regime, where the Au(001) surface is unreconstructed, is in accordance with the known substantial increase in Au adatom diffusion with increasing potential, caused by the change in the electric field at the interface and the influence of coadsorbed anions [2]. The crossover from 3D to 2D growth at negative potentials can be attributed to the formation of the “hex” reconstruction on the Au(001) surface, which results in an enhanced Au surface mobility. The formation of the “hex” reconstruction directly after the potential step also leads to an initial increase in the scattered intensity, resulting in the noticeable “phase shift” of the curve with respect to that observed on the unreconstructed surface at 0.35 VAg/AgCl.

Fig. 91: Kinetic growth mode diagram for Au electrodeposition on Au(001) in 0.1 M HCl, showing the dependence of the kinetic growth mode on potential and deposition rate.

In summary, these experiments demonstrate that by surface X-ray scattering in transmission geometry direct in situ studies of the kinetic growth mode are possible not only for MBE, but also for solid-liquid interfaces. The results obtained for homoepitaxial deposition on Au(100) revealed that the surface mobility and consequently the growth behaviour is strongly affected by the complex solid-liquid interface structure, i.e. the presence of the surface reconstruction, the electric field at the interface, as well as the coadsorbed species on the solution side of the interface. Since this method affords studies at deposition rates that approach those used in practical applications, it is a promising new tool for future fundamental and applied studies of galvanic deposition processes.

 

References

[1] A.H. Ayyad, J. Stettner, O.M. Magnussen, Phys. Rev. Lett. 94, 066106 (2005).
[2] M. Giesen, et al., Surf. Sci. 595, 127 (2005).

Principal Publication and Authors

K. Krug, J.Stettner, O.M. Magnussen, Phys. Rev. Lett., 96, 246101 (2006).
Institut für Experimentelle und Angewandte Physik, Universität Kiel (Germany)