| Abstract:
Al(110) homoepitaxy exhibits nanoscale self-assembly into huts with well-defined (100) and (111) facets [1,2]. Describing hut formation in this system is a challenging, multiscale problem, in which nanohuts form because of the atomic-scale processes by which adatoms ascend/descend steps and facets on Al(110). We developed a model of hut formation in this system. To properly describe the thermodynamics of hut formation, we quantified interactions between atoms on Al(110), Al(100), and Al(111) using first-principles density-functional theory (DFT) based on the VASP code [3,4]. Interestingly, pair interactions on Al(110) are attractive only for the nearest [1 1 0] neighbors and cannot explain hut formation [3]. We showed that higher-order, many-body interactions provide the [0 0 1] attraction needed for huts. To describe interactions, we propose the Connector Model [4], whose central idea is to combine groups of many-body interactions into important structural units (e.g., step edges) that have a single interaction energy. The Connector Model is accurate and efficient compared to the traditional lattice-gas approach and it may be useful in other systems.
We found the kinetic processes underlying hut formation using accelerated ab initio MD simulations and transition-state searches [5, 6]. In addition to finding new energetically preferred mechanisms for Al adatom diffusion on Al(110), we found several new mechanisms for adatoms to ascend/descend steps and facets. These mechanisms play a key role in dictating hut geometries observed experimentally in assembly.
We incorporated all the kinetic mechanisms and interactions that we found into KMC simulations to describe hut formation. Similar to the experiments, we predict that huts form, but we do not achieve precise hut placement. Recently, Kalyanaraman and colleagues used thermal-field-directed assembly (TFDA) to achieve ordered ripples during the deposition of Ag and Co on Si(100), with a simultaneous pulsed two-beam laser interference irradiation of the substrate. The ripples occur because adatoms diffuse more slowly in cold regions of the substrate than in hot regions, so huts nucleate in the cold regions. We simulated growth on surfaces with thermal gradients consistent with two-beam laser interference and achieved groove morphologies in both the [1 1 0] and [0 0 1] directions. By simulating growth on surfaces with temperature gradients consistent with four-beam laser interference, we nucleated a square array with a hut in each cold spot. Our results suggest that TFDA may be fruitfully employed to achieve the alignment of nanostructures in other systems, for example, in III/V semiconductor heteroepitaxy, where quantum dots form.
1. F. Buatier de Mongeot, W. Zhu, A. Molle, R. Buzio, C. Boragno, U. Valbusa, E. G. Wang, and Z. Zhang, Phys. Rev. Lett. 91, 016102, 2003.
2. W. Zhu, F. Buatier de Mongeot, U. Valbusa, E. Gg. Wang, and Z. Zhang, Phys. Rev. Lett. 92, 106102, 2004.
3. Y. Tiwary and K. A. Fichthorn, Phys. Rev. B 75, 235451, 2007.
4. Y. Tiwary and K. A. Fichthorn, Phys. Rev. B 78, 205818, 2008.
5. K. A. Fichthorn, R. A. Miron, Y. Wang, and Y. Tiwary, J. Phys. Cond. Matt. 21, 084214, 2009.
6. Y. Tiwary and K. A. Fichthorn, Phys. Rev. B 81, 195421, 2010.
7. C. Zhang and R. Kalyanaraman, Appl. Phys. Lett. 83, 4827, 2003.
8. C. Zhang and R. Kalyanaraman, J. Mater. Res. 19, 595, 2004.
9. W. Zhang, C. Zhang, and R. Kalyanaraman J. Vac. Sci. Technol. B 23, L5, 2005. |
