Mat. Res. Soc. Symp. Proc. Vol.

800 2004 Materials Research Society

AA7.7.1

Nanoscale Modeling of Shock-Induced Deformation of Diamond S. V. Zybin1, I. I. Oleinik2, M. L. Elert3, C. T. White4

1 2

The George Washington University, Washington D.C. 20052 University of South Florida, Tampa, FL 33620 3 U.S. Naval Academy, Annapolis, MD, 21402 4 Naval Research Laboratory, Washington D.C. 20375 ABSTRACT Molecular dynamics (MD) simulations of shock-induced deformations in diamond were performed using a reactive bond order (REBO) potential. A splitting of shock wave structure into elastic and crystal deformation fronts was observed in the [110] and [111] crystallographic directions above piston velocity thresholds of up 1.8 and 2.5 km/s, respectively. The crystal lattice response in a split two-wave regime consists of the relative movement of {111} planes in the diamond crystal and has different structural character for [110] and [111] shock waves. The strain produced by a [110] shock wave occurs only along one of the transverse crystalline directions, whereas in the [111] case crystal deformation involves the movement of the atoms in both transverse directions. To gain insight into the atomistic mechanisms of orientational dependence of shock compression of crystals, we have investigated in detail the constitutive stress-strain relationships under static uniaxial compression. The REBO potential gives a reasonably good description of stresses and energetics under moderate uniaxial compressions corresponding to an elastic shock wave regime. However, under compressions higher than 10% ([110] case) and 20% ([111] case) the REBO potential shows deficiencies in the quantitative description of stress response that might affect the MD picture of shock wave deformations in diamond. INTRODUCTION Microscopic study of the orientational dependence of shock-induced transformations of crystal lattices and splitting of shock wave fronts in solids is of paramount importance for understanding materials response under shock compression at the nanoscale. Such structural transformations are dependent on the crystallographic orientation of shock wave propagation as well as the details of interatomic interactions that govern the energetics and dynamics of the underlying atomic scale processes. To elucidate the internal structure of shock-compressed crystals and to develop robust constitutive models, a better understanding of atomic scale mechanisms is highly desirable. Molecular dynamics (MD) simulations of shock waves in solids at the atomistic level have been successfully applied to study plastic deformations [1-3], melting [4,5], amorphization [6], and shock wave splitting due to phase transitions [7,8]. The interatomic potentials are at the heart of MD simulations and their ability to describe quantitatively the fundamental physics and chemistry at the atomic scale is the key to achieving reliable and meaningful results. Simple Lennard-Jones (LJ) pairwise interatomic potentials have already revealed complex mechanisms of shock-induced lattice transformation, such as emission of stacking fault arrays and defect nucleation [9], martensitic transformation [10] and appearance of twinning-like chevron band

AA7.7.2

patterns during a structural fcc to hcp transition [11]. It is reasonable to expect even more complex phenomena in shock compressed covalently bonded solids where the angular character of interatomic interactions adds an additional complexity into many-body interatomic interactions. We report MD simulation of shock wave propagation in the diamond lattice using the reactive bond-order (REBO) potential that was originally developed for MD studies of diamond chemical vapor deposition [12,13], and later successfully employed for MD modeling of other systems [14-16]. In this work we report results of preliminary studies aimed at establishing a connection between the simulated shock wave structure in a diamond lattice and the properties of the REBO potential. RESULTS AND DISCUSSION Depending on shock wave strength we have observed various regimes of material response in MD simulations of shock compression of REBO diamond: (i) a pure elastic shock wave, (ii) shock-induced two-wave splitting accompanied by substantial crystal lattice movement, and (iii) amorphization into highly compressed carbon with partial rehybridization to sp2-bonds [17]. Importantly, the simulations revealed an orientational dependence of shock induced crystal deformations including appearance of particular plane slipping directions, shear stress relaxation, and -bond chemistry (rehybridization). Because planar shock wave compression of solids is purely uniaxial, insight into the mechanisms of lattice deformation can be gotten from the shear stress strain relationship obtained by performing a separate simulation of static uniaxial compression of the crystal. In particular, the transition to a metastable overheated state in shockinduced melting in a LJ solid [4,5] along the [100] direction of a fcc crystal was explained by the non-monotonic character of shear stress-strain relationship under static uniaxial compression in that direction. In the case of weak shock waves in perfect crystals below the so-called Hugoniot elastic limit [18], the material response is entirely elastic, i.e. there is no shear stress relaxation by irreversible plastic deformation. In the case of MD simulations using the REBO potential, we have observed pure elastic shocks at piston velocities up and volumetric strains = (1 V / V0 ) below the following thresholds: up < 1.8 km/s, < 0.09 for the [110] direction, and up < 2.5 km/s, < 0.11 for the [111] direction. Above these limits for both [110] and [111] crystallographic directions we observe a two-wave splitting of the shock wave front resulting from a slower deformation wave lagging behind a faster elastic precursor. Figure 1 shows a typical snapshot of a two-wave shock structure propagating from left to right along the [110] crystallographic direction (z axis in the figure). The compression ratio in the lagging wave is 1 0.82. The bulk deformation of the lattice develops by shear of {111} planes along [112] and [112] directions leading to clearly visible diamond-shaped regions within the y-z plane in the figure. Only a few (110) atomic planes are projected into y-z plane of the figure, and the regions of higher compression are highlighted by the lighter color of the atoms. We found that no strain occurs in x-z planes of the {001} family that are normal to the plane of the figure. As a result, only shear stress in the y-z plane yz = ( zz yy ) is relaxed by the corresponding deformation of the latticethe shear stress in the x-z plane has not been relieved. (In fact, xz even increases as a result of lattice distortions.)

AA7.7.3

[112]

elastic wave

deformation wave
y[001] x[110] z[110] Figure 1. Snapshot of [110] two-wave shock splitting in REBO diamond for piston velocity up = 3.28 km/s and shock velocities us 20.5 and 18 km/s for elastic and deformation waves, respectively. Atoms are shaded according to potential energy.

elastic wave

deformation wave

[211] x[011] z[111] Figure 2. Snapshot of [111] two-wave shock splitting in REBO diamond for piston velocity up = 3.96 km/s and shock velocities us 22.8 and 18.8 km/s for elastic and deformation waves, respectively. Atoms are shaded according to potential energy.

AA7.7.4

On the other hand, in the case of shock wave propagation in the [111] crystallographic direction the bulk deformation develops in both [211] and [011] transverse directions within the y-z and the x-z planes. Figure 2 shows the structure of a shock wave front propagating along the [111] direction in diamond at the compression ratio 1 0.79. The strain occurs through the shear of {111} planes not only in the [211] direction, but also in the [011] direction resulting in clearly visible ripples within the (011) plane. Again, the higher compression in the lagging shock wave is displayed by the lighter color of the atoms having higher potential energy. Since the deformation in this case develops by the movement of {111} planes normal to the shock direction, the shear stresses xz , yz are not substantially relaxed. To understand the relationship between the deformation pattern and the underlying interatomic interactions included in simulation via the REBO potential, we investigated the shear stress behavior under uniaxial compression by calculating the stress tensor components as a function of static uniaxial strain. (These calculations did not include the effect of temperature.) The size of the sample crystal cell along the z-axis was scaled by an appropriate compression ratio and the total energy of the crystal was minimized using a steepest descent algorithm to relax atomic positions under the fixed strain of the lattice. Figure 3 displays the shear stress strain dependence for static uniaxial compression along the [110] and [111] directions. As can be seen from this figure, both cases exhibit non-monotonic shear stress dependence. Although the shear stresses are equal in the [111] case ( xz = yz ), they have different behavior under compression in the [110] direction. The non-monotonic character makes it possible to decrease the shear stress at higher strains, thus opening new possibilities for developing deformation patterns different from those at smaller strains. Because the strain rate under shock loading is very high, intermediate steps during compression have little influence on the final result. In particular, the absence of strain within x-z planes in the [110] shock wave might be related to early decrease of the shear stress yz as compared to xz . It is worth noting that similar non-monotonic behavior of the shear stress was also observed for uniaxial compression of silicon along the [100] direction [19]. Although the lattice deformation in diamond occurs by movement of {111} planes along directions expected for deformation twinning, no actual rebonding of the planes was detected
8.0 8.0
[110] shear stresses [111] shear stresses

Mbar

6.0 4.0 2.0 0.0

2xz 2yz

6.0 4.0 2.0 0.0 0.70 0.75 0.80 0.85

2xz = 2yz

2 x Shear stress

Compression ratio 1- Compression ratio 1- Figure 3. Twice the shear stress vs. strain for cold uniaxial compression of REBO diamond along [110] and [111] crystallographic directions.

0.75 0.80 0.85 0.90 0.95 1.00

0.90

0.95

1.00

AA7.7.5

Mbar

8
[110] compression

Etot
6

-5.0 -6.0 -7.0 6

[111] compression

Etot

-5.0 -6.0 -7.0 -8.0 -9.0

zz ,

zz

zz

Longitudinal stress

Etot

zz
0.75 0.80 0.85 0.90 0.95

-8.0 -9.0

0 0.70

1.00

0 0.70

0.75

0.80

0.85

0.90

0.95

1.00

Compression ratio 1- Compression ratio 1- Figure 4. Longitudinal stress component zz and total energy per atom Epot vs. strain for cold uniaxial compression of REBO diamond. Maximum of zz is reached at compression ratios of 0.85 and 0.788 for [110] and [111] directions, respectively.

in the strained region within the second shock wave structure. The deformed lattice exhibits lenticular-shaped planar patterns that stay intact during the simulation time. This phenomenon can be understood in terms of the longitudinal component of the stress tensor component zz displayed in the Figure 4. The decrease in zz at higher compression ratios suggests the onset of a polymorphic phase transition. Indeed, we found that the diamond lattice transforms into a dense graphite-like structure at large uniaxial compressions. Obviously, the result of the simulations is the direct consequence of the interatomic interactions as described by REBO. To understand the behavior of the REBO potential under large uniaxial compressions, we calculated the stress uniaxial strain relationship by density functional theory. The preliminary results indicate that while the REBO total energy per atom follows reasonably well the first principles results, the longitudinal stress in both cases shows substantial deviation at strains greater than 0.1 0.2. This deficiency of REBO potential is likely related to the fact that no high-pressure data was included into fitting the REBO potential. In particular, such behavior might be related to an excessive softness of its repulsion term at small interatomic distances. To further clarify the structure of shock compressed diamond in the region of two-wave shock splitting the REBO potential must be improved to describe properly the stress-strain relationship at large uniaxial compression. CONCLUSIONS We have performed MD simulations of shock-induced deformations in diamond using a reactive bond order potential (REBO). We observed the splitting of shock wave structure into elastic and crystal deformation fronts in [110] and [111] crystallographic directions above piston velocity thresholds of up 1.8 and 2.5 km/s, respectively. The crystal lattice transformation in a split two-wave case consists of the movement of {111} planes and has different structural character for [110] and [111] shock waves. The strain produced by the [110] shock wave occurs only along one of the transverse crystalline directions, whereas in the [111] case crystal

Total energy per atom Etot , eV

AA7.7.6

deformation involves the movement of the atoms in both transverse directions. We have studied static strain-stress relationships under uniaxial compression to check the behavior of REBO at large deformations. Our preliminary results indicate that at stains higher than 10% (for [110] compression) and 20% (for [111] compression) the REBO potential show deficiencies in the quantitative description of stress response. This might affect the MD picture of shock wave deformations in diamond crystal. Further efforts are required to develop robust interatomic potentials for MD modeling of shock wave physics and chemistry at the nanoscale. ACKNOWLEDGEMENTS This work was supported by the U.S. Office of Naval Research (ONR) both directly and through the Naval Research Laboratory (NRL). IIO thanks the ONR/NRL for the support of his summer faculty fellowship at NRL Theoretical Chemistry Section. REFERENCES 1. V. V. Zhakhovski, S. V. Zybin, K. Nishihara, and S. I. Anisimov, Phys. Rev. Lett., 83, 1175 (1999); in Proceedings of Symposium on Shock Waves98, Aoyama Gakuin Univ., (Tokyo, 1999), pp. 241244. D. H. Robertson, D. W. Brenner, and C. T. White, in High Pressure Shock Compression of Solids III, edited by L. Davison and M. Shahinpoor (Springer, 1998), pp. 37-57. T.C. Germann, B.L. Holian, P.S. Lomdahl, and R. Ravelo, Phys. Rev. Lett., 84, 5351 (2000). V. V. Zhakhovski, S. V. Zybin, K. Nishihara, and S. I. Anisimov, Progr. Theor. Phys. Suppl., 138, 223 (2000). S. V. Zybin, V. V. Zhakhovskii, M. L. Elert, and C. T. White, in Shock Compression of Condensed Matter - 2003, edited by M. Furnish et al, AIP Conf. Proc. (in press). S. L. Chaplot and S. K. Sikka, Phys. Rev. B, 61, 11205 (2000). D. H. Robertson, D. W. Brenner, and C. T. White, Phys. Rev. Lett., 67, 3132 (1991). K. Kadau, T. C. Germann, P. S. Lomdahl, and B. L. Holian, Science, 296, 1681 (2002). B.L. Holian and P.S. Lomdahl, Science, 280, 2085 (1998). J. P. Hirth, R. G. Hoagland, B. L. Holian, and T. C. Germann, Acta Mater., 47, 2409 (1999). J.-B. Maillet, M. Mareschal, L. Soulard, R. Ravelo, P. S. Lomdahl, T. C. Germann, and B. L. Holian, Phys. Rev. E, 63, 016121-1 (2000). D. W. Brenner, Phys. Rev. B, 42, 9458 (1990). D. W. Brenner, J. A. Harrison, C. T. White, and R. J. Colton, Thin Solid Films, 206, 220 (1991). D. W. Brenner, O. A. Shenderova, J. A. Harrison, S. J. Stuart, B. Ni and S. B. Sinnott, J. Phys. Cond. Matt., 14, 783 (2002). B. I. Dunlap, D. W. Brenner, J. W. Mintmire, R. C. Mowrey, and C. T. White, J. Phys. Chem., 95, 5763 (1991). J. A. Harrison, C. T. White, R. J. Colton, and D. W. Brenner, Phys. Rev. B, 46, 9700 (1992). S. V. Zybin, M. L. Elert, and C. T. White, Phys. Rev. B, 66, 220102-1(R) (2002). K. Kondo and T. J. Ahrens, Geophys. Res. Lett., 10, 281 (1983). D. C. Swift and G. J. Ackland, LANL internal Report No. LAUR-00-2654, 2000.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.