Sure, depending on the formalism you want to use.

Already included in the potentials/ dir in the LAMMPS distribution is a Tersoff-style potential [Comp. Mat. Sci., 39 (2007) 334].

Of course there's the ReaxFF potential, for which also Si/O parameters exist. I think the latest version is [Phys. Rev. Lett. 96 (2006) 095505]. But you can best contact Adri van Duin to be sure.

There's also the COMB potential, for which Si/O parameters are already supplied with LAMMPS [Phys. Rev. B 82 (2010) 235302].

Which one's best? I don't know: it'll depend on your specific system, so you'll have to verify the accuracy yourself. In our group, people got some good results for Si surface/nanowire oxidation with ReaxFF. See Khalilov et al.'s work here at RG: https://www.researchgate.net/profile/Umedjon_Khalilov/publications

Thank you,

ok. Thank you

ReaxFF is ok

I say it depends on the application. From my experience, the current parametrization of COMB for Si/O is good at predicting adhesion energies for crystalline silica. I was never able to "melt-and-quench" with COMB to generate amorphous silica. I have more experience with ReaxFF and its parametrization for Si/O in combination with metals (Cu, Ag). We have been able to generate realistic amorphous structures with good densities,  intrinsic defect formation energies, and other mechanical properties.

There is a parametrization for ReaxFF (Si/O) published, if you are interested let me know, I can get you the force field file.

Regards, 

Thank you!