Thin Solid Films 517 (2009) 6234–6238

Contents lists available at ScienceDirect

Thin Solid Films

j ourna l homepage: www.e lsev ie r.com/ locate / ts f

An atomic-scale study of hydrogenated silicon cluster deposition on a crystallinesilicon surface

Ning Ning ⁎, Steven M. Rinaldi, Holger VachLPICM, Ecole Polytechnique, CNRS, 91128 Palaiseau, France

⁎ Corresponding author.E-mail address: ning@leonardo.polytechnique.fr (N.

0040-6090/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.tsf.2009.02.086

a b s t r a c t

a r t i c l e i n f o

Available online 21 February 2009

Keywords:

Controlled deposition of clupotential application to tailoout an atomic-scale study to

Molecular dynamics simulationsHydrogenated silicon clustersDeposition mechanismSilicon surface

understand the deposition mechanism. The molecular dynamics approach basedon a modified Tersoff potential is used to simulate the deposition mechanism of hydrogenated silicon clusterson a crystalline silicon surface in detail. The important factors governing the deposition process such asimpact energy and substrate temperature, are investigated for the hydrogenated silicon cluster Si29H24 on a

sters on solid surfaces has attracted lots of attention in recent years, because of itsring the desired electronic properties of the resulting surfaces. We have carried

H-terminated Si(100)-(2x1) surface.© 2009 Elsevier B.V. All rights reserved.

1. Introduction

The controlled deposition of clusters on solid surfaces hasattracted a lot of attention in recent years due to its potentialapplication in tailoring the electronic properties of the resulting thinfilm. In particular, two types of deposition mechanisms areattractive: the destructive deposition of clusters and the soft-landing of cluster. Destructive deposition, e.g., ionized clusterdeposition, holds promise in synthesizing high quality thin filmsat low temperatures. In this mechanism, the impact energy of thecluster is transferred to the fragmentary cluster atoms, thusbecoming migration energy, and leading to high atomic mobilityon the surface even at low temperatures [1]. In soft-landingdeposition, the cluster is not destroyed upon landing on thesubstrate. This type of deposition is equally promising, as clustersexhibit interesting quantum properties [2,3] that are sensitive totheir size and bonding states, and which can therefore be controlled.In both cases, the deposition of clusters instead of individual atomsor molecules attracts industrial applications due to its inherentpossibility of high-speed thin film deposition.

In this work, we used model potential molecular dynamicssimulations to study the controlled deposition of clusters on solidsurfaces in order to understand the growth mechanisms andconditions. We begin our research with the simulation of thedeposition of hydrogenated silicon clusters under low temperatureplasma-enhanced chemical vapor deposition conditions.

Ning).

ll rights reserved.

2. Computational methods

To simulate the nanoparticle-surface interaction, the Ohira–Tersoff(OT) potential was employed as it has been shown to have a goodperformance in describing the structure and properties of amorphous Si(a-Si) bulk phases [4] and c-Si surfaces [5], and because it has beensuccessfully used to study the radical–surface interactions [6–8], a-Sithin film growth mechanisms [9–11], and the H-induced crystallizationof a-Si thin films [12–14]. The potential function form is described in[15]. The performanceof theOTpotential in describing the structure andproperties of c-Si bulk is given inTable 1, which shows that the results ofthe OT potential are in good agreementwith ab initio calculation results.

In this study, the filled fullerene stable configuration Si29H24

shown in Fig. 1 has been chosen to be the impinging cluster, as thestructural and optical properties of this 1 nm nanoparticle have beenstudied and compared to theoretical and experimental findings[17,18]. The H-terminated Si(100)-(2x1) surface is used as thesubstrate, which consists of 288 silicon atoms (see Fig. 2). Thedimension of this substrate is 21.72 Å×21.72 Å×10.86 Å with 2dimensional periodic boundary conditions (PBC) applied in thesurface plane. The top silicon layer is a reconstructed layer consistingof silicon dimers. The bottom two layers are kept rigid in theirequilibrium positions to avoid translation of the substrate in space.

The substrate temperature is controlled by using a Berendsenthermostat [19] in the three layers above the rigid layers. These threelayers also act as a heat reservoir which absorbs the heat during thecollision. We used two substrate temperatures which represent thetwo extreme values employed in low temperature PECVD: 373 K and573 K. The surface was heated up to the target temperature. Then, thesurface temperature was maintained until the surface structurereached its equilibrium state, which took about 10 ps. Afterwards,the deposition of the Si29H24 cluster on the substratewas initiated. The

Table 1Comparison of the energies and structures of c-Si and H-terminated Si(100)-(2x1)predicted by the OT potential with ab initio results.

c-Si OT ab initio [16]

Bulk cohesive energy (eV/atom) 4.6296 4.63Lattice parameter (Å) 5.4320 5.43

H-terminated Si(100)-(2x1) OT ab initio [6]

Dimer bond length (Å) 2.44 2.4H–Si–Si bond angle (°) 110.2 114.7

6235N. Ning et al. / Thin Solid Films 517 (2009) 6234–6238

impact energies were selected from our results using the hard surfacecollision model [20]. We only employed the hard surface collisionmodel to simulate the elastic scattering of the cluster from a hardsurface, in order to obtain quantitative information about the clusterdeformation energy.

3. Results and discussion

Guided by the results of our hard surface collision simulations, weselected three different impact energies: 0.3 eV/atom, 2 eV/atom and5 eV/atom.

Fig. 1. The initial configuration of the hydrogenated silicon cluster Si29H24. Big ballsrepresent silicon atoms while small ones represent hydrogen atoms.

Fig. 2. The initial configuration of the H-terminated Si(100)-(2x1) surface. The big ballsrepresent silicon atoms while the small ones represent hydrogen atoms.

Figs. 3 and 4 demonstrate the results of collisions for differentimpact energies at substrate temperatures of 373 K and 573 K,respectively. The deposition mechanisms are quite similar for thosetwo substrate temperatures. When the impact energy is low (0.3 eV/atom), the cluster remains intact on the surface; with a mediumimpact energy (2 eV/atom), the cluster partially dissociates andspreads out on the substrate; while at a high impact energy (5 eV/atom), not only does the cluster dissociate, but it also penetrates intothe surface. At this high impact energy, the deposition mechanismsare slightly different for the two substrate temperatures: at the highersubstrate temperature, the cluster penetrates deeper into thesubstrate.

Biswas et al. [21] suggested that epitaxial growth requires thateach impinging cluster dissociates and spreads in a uniform layer onthe substrate. Therefore, we chose the collision at a medium impactenergy where the cluster deforms and spreads out on the surface formore detailed study. As there is no significant difference at themedium impact energy for the two investigated temperatures, the lowtemperature case was chosen for illustration. Figs. 5 and 6 show theresults obtained from the cluster-surface collision at an impact energyof 2 eV/atom and a surface temperature of 373 K. Fig. 5 demonstratesthe structural evolution of the cluster during the collision. Fig. 6(a)displays the time evolution of the instantaneous temperature of thefirst silicon layer and Fig. 6(b) shows the temporal spread-out of thekinetic energy for this layer. Combining the results of Figs. 5 and 6(a),we see that the temperature of the first silicon layer increases

6236 N. Ning et al. / Thin Solid Films 517 (2009) 6234–6238

promptly at 0.5 ps (Fig. 5(b)), and that the translation energy of thecluster is converted to vibrational energy of the surface atoms. At theend of the collision (Fig. 5(e)), the local surface temperature obtainsits maximum value which is approximately the melting point ofsilicon. Thereafter, the temperature decreases to 373 K due to thethermostat layers. We can obtain more details about the heat transferfrom Fig. 6(b). First, the translation energy of the incoming clusterconverts to kinetic energy of the surface atoms only in the impactregion. Then, the heat spreads out from the impact region on apicosecond time scale. In this manner, the whole surface heats up toits maximum temperature.With the thermostat inside the surface, the

Fig. 3. Snapshots of the atomic configurations at 10 ps after the cluster surface impactwith a surface temperature of 373 K and various impact energies.

Fig. 4. Snapshots of the atomic configurations at 10 ps after the cluster surface impactwith a surface temperature of 573 K and various impact energies.

additional heat is dissipated and the temperature decreases. There-fore, we would expect to see incorporation of the deposited clusterinto the crystalline structure of the surface after subsequent additionalcluster depositions due to annealing.

4. Conclusions

We studied the deposition mechanism of the hydrogenated siliconcluster Si29H24 on a H-terminated Si(100)-(2x1) surfacewith differentsubstrate temperatures and various impact energies. At low and

Fig. 5. Snapshots of the atomic configurations at various times during the cluster surface collision with an impact energy of 2 eV/atom and a surface temperature of 373 K.

6237N. Ning et al. / Thin Solid Films 517 (2009) 6234–6238

medium impact energies, there is no significant difference in thedeposition mechanism when the substrate temperature is increasedfrom 373 K to 573 K; at relatively high impact energies, however, the

cluster penetrates deeper into the substrate at the higher temperature.At the lower temperature of 373 K with the medium impact energy of2 eV/atom, the surface atom temperature achieves the melting point

Fig. 6. Time evolution of the instantaneous temperature for (a) first, reconstructed siliconlayer, (b) two regions in the reconstructed silicon layer. The initial surface temperature was373 K and the impact energy was 2 eV/atom.

6238 N. Ning et al. / Thin Solid Films 517 (2009) 6234–6238

of silicon due to the cluster impact. Therefore, we suggest tentativelythat epitaxial growth accompanied by hydrogen evaporation mayoccur by the subsequent deposition of additional clusters under thesame conditions.

Acknowledgments

N.N. is the recipient of a PhD studentship from the MESR. Theauthors are also grateful to the French National Computer Center IDRISand to M. Labrune and Dr. P. Roca i Cabarrocas for discussions of theexperimental PECVD processes.

References

[1] H. Haberland, M. Karrais, M. Mall, Y. Thurner, J. Vac. Sci. Technol. A 10 (1992) 3266.[2] G. Belomoin, J. Therrien, A. Smith, S. Rao, R. Twesten, S. Chaieb, M.H. Nayfeh, L.

Wagner, L. Mitas, Appl. Phys. Lett. 80 (2002) 841.[3] S.E. Baltazar, Mario DeMenech, Ulf Saalmann, Aldo H. Romero , M.E. Garcia, J. Appl.

Phys. 103 (2008) 023714.[4] T. Ohira, T. Inamuro, T. Adachi, Mat. Res. Soc. Symp. Proc. 336 (1994) 177.[5] T. Ohira, O. Ukai, T. Adachi, Y. Takeyuchi, M. Murata, Phys. Rev. B 52 (1995) 8283.[6] S. Ramalingam, J. Appl. Phys. 84 (1998) 3895.[7] S. Ramalingam, P. Mahalingam, E.S. Aydil, D. Maroudas, J. Appl. Phys. 86 (1999)

5497.[8] S. Ramalingam, D. Maroudas, E.S. Aydil, J. Appl. Phys. 86 (1999) 2872.[9] S.P. Walch, S. Ramalingam, S. Sriraman, D. Maroudas, Chem. Phys. Lett. 344 (2001)

249.[10] S. Sriraman, E.S. Aydil, D. Maroudas, J. Appl. Phys. 92 (2002) 842.[11] S. Sriraman, E.S. Aydil, D. Maroudas, J. Appl. Phys. 95 (2004) 1792.[12] S. Sriraman, S. Agarwal, E.S. Aydil, D. Maroudas, Nature 418 (2002) 62.[13] M.S. Valipa, S. Sriraman, E.S. Aydil, D. Maroudas, J. Appl. Phys. 100 (2006) 053514.[14] M.S. Valipa, S. Sriraman, E.S. Aydil, D. Maroudas, J. Appl. Phys. 100 (2006) 053515.[15] N. Ning, G. Dolgonos, W. Morscheidt, A. Michau, K. Hassouni, H. Vach, Comput.

Meth. Sci. Eng. 1 (2007) 224.[16] D. Alfe, M.J. Gillan, M.D. Towler, R.J. Needs, Phys. Rev. B 70 (2004) 214102.[17] L. Mitas, J. Therrien, B. Twesten, G. Belomoin, M. Nayfeh, Appl. Phys. Lett. 78 (2001)

1918.[18] S. Rao, K. Mantey, J. Therrien, A. Smith, M. Nayfeh, Phys. Rev. B 76 (2007) 155316.[19] H.J.C. Berendsen, J.P.M. Postma, W.F. van Gunsteren, A. Dinola, J.R. Haak, J. Chem.

Phys. 81 (8) (1984) 3684.[20] H. Vach, Phys. Rev. B 61 (1999) 2310.[21] R. Biswas, G.S. Grest, C.M. Soukoulis, Phys. Rev. B 38 (1988) 8154.