Effect of thermal annealing on the stress and morphology of deposited nanofilms analyzed using molecular dynamics

Effect of thermal annealing on the stress and morphology of deposited nano

films

analyzed using molecular dynamics

Zheng-Han Hong

, Shiang-Jiun Lin

, Te-Hua Fang

⁎

, Shun-Fa Hwang

a

Department of Mechanical Engineering & Institute of Mechanical and Precision Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 807, Taiwan b_{Department of Mold and Die Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung 807, Taiwan}

c

Department of Mechanical Engineering, National Yunlin University of Science and Technology, Yunlin 640, Taiwan

a b s t r a c t

a r t i c l e i n f o

Article history: Received 17 March 2010

Accepted in revised form 31 January 2011 Available online 20 February 2011 Keywords:

Molecular dynamics (MD) Surface roughness Intermixing Stress

Annealing radial distribution function (RDF) Recrystallization

Molecular dynamics (MD) simulations are used to investigate the surface roughness and intermixing of grownfilms and to calculate the atomic-level stress before and after the annealing process. The tight-binding second-moment approximation (TB-SMA) many-body potential is employed to model the interaction. The results reveal that the morphology of thefilm becomes smoother after 900 K annealing due to the substrate temperature being lower for the as-deposited film. The results indicate that the annealing decreases the number of vacancies and voids for all thinfilms. More Al atoms penetrate into the substrate after the annealing process. The interface mixing of annealedfilm is thus better than that of as-deposited film. The first peak of the radial distribution function (RDF) becomes higher and narrower when the Al–Cu film is cooled from 700 to 300 K. This indicates that new grains nucleate and grow to replace those deformed by internal stresses and that recrystallization may occur for the Al–Cu film. In the atomic-level stress analysis, the average normal stress along the thickness direction and the average mean biaxial stress are considered. Both residual stresses decrease for all layers after the annealing process. This implies that internal stress is relieved within the Al–Cu film during the cooling phase.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Cluster beam or ion beam sputtering (IBS) deposition is a conventionally adopted approach for producing a thin film on a substrate[1–3]. The IBS process is applied to enhance the mobility of the deposited atoms and to promote thefilm growth. It is also used to bombard the surface of a solid substrate to remove its atoms for further deposition[3]. These two processes depend on the incident energy of the ions, the incident angle of ions, and the ion assisted ratio, which is the ratio of ions to neutral atoms. These IBS process parameters affect the quality and morphology of the deposited thin film. Combinations of IBS process parameters lead to various film growth processes, such as epitaxy[4,5],film mixing (the mixing of deposition atoms and substrate atoms)[6–8], and sputtering[9,10]. The mechanical properties of films produced using epitaxy and mixing can be measured after thermal annealing[11–13].

Lee et al.[11]sputter-deposited an Al–Cu thin film on a Si wafer and heat-treated it at 200 and 480 °C. Heat-treatment at 480 °C reduces the resistivity, lattice strain, and surface roughness. Molecular dynamics (MD) are preferable to experiments for investigating the morphology of deposited thinfilms in detail and to understand the

mechanisms of growth. Kim et al. [12] investigated Al atoms deposited on a rough Cu(111) surface during annealing process. After annealing at 700 K, intermixing is observed between Al and Cu and the surface roughness decreased.

The stress associated with the IBS process in a deposited thinfilm on a substrate has been exploited for such applications as ultra-large scale integration (ULSI) in aluminum and aluminum–alloy intercon-nections. MD simulation has been adopted to investigate why stress evolution is critical to the reliability of a deposited_film[14]. Chocyk et al. [15] observed stress evolution in Cu thin films during and after deposition at room temperature using the Lennard–Jones (LJ) potential. Their results indicate a compressive_{–tensile–compressive} evolution of stress in Volmer–Weber (VW) mode [16] at various deposition rates. The simulated stresses were very close to the experi-mental results for thin depositedfilms. Sander et al.[17]examined the mechanisms of the generation of atomic level stresses with increasing the number of atom deposited using the tight-binding potential to model the thin Co–Cu(00i) system. The stress approached 2.9 GPa when strain relaxation led to reduced stress in Co islands.

From the above discussion, a comparison of as-deposited and post-annealedfilm using MD simulation is of interest. In this study, the tight-binding second-moment approximation (TB-SMA) many-body potential is used to model the atomic interactions in the system of Al–Cu film to understand the effect of thermal annealing on the surface roughness, layer coverage function, radial distribution

Surface & Coatings Technology 205 (2011) 3865–3871

⁎ Corresponding author. Tel.: +886 7 3814526 5336. E-mail address:[email protected](T.-H. Fang).

0257-8972/$– see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2011.01.066

Contents lists available atScienceDirect

Surface & Coatings Technology

function (RDF), and residual stress induced in as-deposited and post-annealed_film.

2. Molecular dynamics method

In recent years, many-body potentials have become commonly adopted to accurately describe the interaction between atoms in MD. In this investigation, the TB-SMA many-body potential[18]is employed to model the atomic interactions between Al and Cu atoms. The TB-SMA potential energyU(rij), which contains a repulsive pair potential

and a cohesive band energy term, is expressed as follows:

U rij
 =∑N i = 1 − ∑j ξ 2 exp −2q rij r0−1    !1 2 + A exp−p rij r0−1     2 4 3 5_ð1Þ

where rijis the separation distance between atoms i and j,ξ is the effective

hopping integral, N is the number of atoms, and r0is thefirst-neighbor

distance. The parametersξ, A, q, p, and r0are obtained fromfittings to the

experimental values of cohesive energy, lattice parameter, bulk modulus, and two shear elastic constants, respectively[18]. The values of these parameters for Al–Cu atoms are listed inTable 1. The parameters of the cross potential were obtained using an interpolation scheme[19,20].

The stress distribution of a thinfilm after deposition is sometimes important. The stress in a thinfilm can be evaluated as the atomic stress, which is calculated from the volume Viof atom i, and consists of

two parts: the kinetic energy of the atom and the interatomic forces. The definition of atomic stress by Basinski, Duesbery, and Tayo (BDT stress,σ)[21]can be expressed as:

σ =_V1 i mivi⊗vi+ 1 2∑ N j≠i rij⊗fij " # ð2Þ Table 1

Parameters for TB-SMA potential.

Parameter A(eV) ξ(eV) p q r0(nm)

Cu–Cu 0.0855 1.224 10.96 2.278 0.255

Al–Al 0.1211 1.316 8.612 2.516 0.325

Al–Cu 0.1017 1.269 9.786 2.397 0.290

Fig. 1. Simulation model after the deposition of Al atoms on the Cu substrate.

Fig. 2. Relationship between temperature and time step for thermal annealing.

Fig. 3. Deposition morphology at an incident energy of 0.5 eV, an incident angle of 0°, and a deposition rate of 4 atoms/ps before and after annealing: (a) as-deposited, (b) heated up to 700 K, (c) 300-K-annealedfilm.

0.352 for higher substrate temperature during the annealing process. The Poisson ratio increases with increasing temperature.

4. Conclusion

This study investigated the change of atomistic properties of Al–Cu film due to the annealing process. The RMS roughness is not affected by the annealing process when the incident energy and/or substrate temperature are low for the as-depositedfilm. However, the morphol-ogy of 900-K-annealedfilm becomes smoother when the substrate temperature is low for the as-depositedfilm. The results indicate that annealing decreases the number of vacancies and voids for all thinfilms. The intermixing at deposited layers occurs with increasing substrate temperature. However, more Al atoms penetrate into the substrate after the annealing process. Hence, the interface mixing of post-annealed_film is better than that of as-depositedfilm. After annealing, the first peak of the RDF becomes lower and wider as the Al–Cu film temperature is increased from 300 to 700 K. This may indicate increased damage when the Al–Cu temperature becomes high. In contrast, the first peak of the RDF becomes higher and narrower when the Al–Cu film is cooled down from 700 to 300 K. This indicates that new grains nucleate and grow to replace those deformed by internal stresses and that recrystallization may occur for the Al–Cu film. For atomic-level stress, the results indicate that both the average normal stress along the thickness direction and the average mean biaxial stress of the substrate layers are compressive due to the deposition of incident atoms. Both the average stresses of the substrate layers oscillate, but there is no clear oscillation at the deposited layers. Both the residual stresses decrease for all layers after the annealing process. This implies that the internal stress is relieved within the Al–Cu film during the cooling phase.

Acknowledgement

This work was partially supported by the National Science Council of Taiwan under grants NSC 098-2811-E-150-002 and NSC 96-2628-E-150-005-MY3.

References

[1] R. Hühne, S. Fähler, B. Holzapfel, Appl. Phys. Lett. 85 (2004) 2744. [2] T.H. Fang, W.J. Chang, C.M. Lin, W.C. Lien, Appl. Surf. Sci. 254 (2008) 3436. [3] T.H. Fang, S.H. Kang, J.H. Liao, Appl. Surf. Sci. 256 (2009) 1395.

[4] R.W. Smith, D.J. Srolovitz, J. Appl. Phys. 79 (1996) 1448. [5] L. Dond, R.W. Smith, D.J. Srolovitz, J. Appl. Phys. 80 (1996) 5682.

[6] S.P. Kim, Y.C. Chung, S.C. Lee, K.R. Lee, K.H. Lee, J. Appl. Phys. 93 (2003) 8564. [7] S.G. Lee, Y.C. Chung, J. Appl. Phys. 100 (2006) 074905.

[8] Z.H. Hong, S.F. Hwang, T.H. Fang, J. Appl. Phys. 104 (2008) 1.

[9] G.V. Kornich, G. Betz, V. Zaporojtchenko, A.I. Bazhin, F. Faupel, Nucl. Instr. Meth. 227 (2005) 41.

[10] R. Smith, D. Ramasawmy, S.D. Kenny, Nucl. Instr. Meth. 228 (2005) 330. [11] D.H. Lee, D.M. Jeon, S.Y. Yoon, J.P. Lee, B.G. Kim, S.J. Suh, Thin Solid Films 435

(2003) 170.

[12] S.-R. Kim, K.-R. Lee, Y.-C. Chung, M. Sahashi, Y.K. Kim, J. Appl. Phys. 105 (2009) 114312.

[13] T.H. Fang, C.D. Wu, W.J. Chang, S.S. Chi, Appl. Surf. Sci. 255 (2009) 6043. [14] I. Rosenblum, J. Adler, Phys. Rev. B 62 (2000) 4117.

[15] D. Chocyk, T. Zientarski, A. Proszynski, T. Pienkos, L. Gladyszewski, G. Gladyszewski, Cryst. Res. Technol. 40 (2005) 509.

[16] G.H. Gilmer, M.H. Grabow, A.F. Bakker, Mater. Sci. Eng. 6 (1990) 101.

[17] D. Sander, S. Ouazi, V.S. Stepanyuk, D.I. Bazhanov, J. Kirschner, Surf. Sci. 512 (2002) 281.

[18] F. Cleri, V. Rosato, Phys. Rev. B 48 (1993) 22. [19] P. Süle, J. Chem. Phys. 128 (2008) 134708.

[20] M. Imafuku, Y. Sasajima, R. Yamamoto, M. Doyama, J. Phys. F. Met. Phys. 16 (1986) 823.

[21] Z.S. Basinski, M.S. Duesberry, R. Taylor, Can. J. Phys. 49 (1971) 2160.

[22] J.M. Haile, Molecular Dynamics Simulation: Elementary Methods, A Wiley-Interscience Publication, New York, 1992, p. 199.

[23] Z.H. Hong, S.F. Hwang, T.H. Fang, Comput. Mater. Sci. 41 (2007) 70. [24] Z.H. Hong, S.F. Hwang, T.H. Fang, Cryst. Growth Des. 8 (2008) 1191.

[25] J.A. Zimmerman, C.L. Kelchner, P.A. Klein, J.C. Hamilton, S.M. Foiles, Phys. Rev. Lett. 87 (2001) 165507.

[26] G.V. Kornich, G. Betz, V. Zaporojtchenko, A.I. Bazhin, F. Faupel, Nucl. Instr. Meth. 227 (2005) 41.

[27] C.W. Pao, D.J. Srolovitz, Phys. Rev. Lett. 96 (2006) 186103.

[28] D. Sander, Z. Tian, J. Kirschner, J. Phys. Condens. Matter 21 (2009) 134015. [29] M. Pletea, R. Koch, H. Wendrock, R. Kaltofen, O.G. Schmidt, J. Phys. Condens.

Matter 21 (2009) 225008. -1 0 1 2 3 4 5 6 7 0 2 4 6 8 10 12 14 16 18 20 22 24 -8 -6 -4 -2 0 2 4 6 8 10 0 2 4 6 8 10 12 14 16 18 20 22 24 As-deposited Heated-up 700K Annealed at 300 K As-deposited Heated-up 700K Annealed at 300 K

Stress (GPa)

Layer

a

b

c

d

Fig. 11. Layer stress before and after annealing for various deposition conditions: (a) average mean biaxial stress with an incident energy of 0.5 eV and a substrate temperature of 300 K, (b) average normal stress with an incident energy of 0.5 eV and a substrate temperature of 300 K, (c) average mean biaxial stress with an incident energy of 1 eV and a substrate temperature of 900 K, and (d) average normal stress with an incident energy of 1 eV and a substrate temperature of 900 K.

3871 Z.-H. Hong et al. / Surface & Coatings Technology 205 (2011) 3865–3871