Dear Miguel. The way we do it at our lab, and in other groups that simulate glasses, is to generate the amorphous material by placing the atoms randomly in the simulation box and then use a melt-and-quench approach: MD simulations at high temperatures (say 3000K) followed by a series of NVT simulations at lower temperatures until T = 300K. I would suggest you looking at some of the papers of Antonio Tilocca (Chemistry UCL)

Not that I know, but it will be certainly interesting. I will follow the answers.

Dear Miguel. The way we do it at our lab, and in other groups that simulate glasses, is to generate the amorphous material by placing the atoms randomly in the simulation box and then use a melt-and-quench approach: MD simulations at high temperatures (say 3000K) followed by a series of NVT simulations at lower temperatures until T = 300K. I would suggest you looking at some of the papers of Antonio Tilocca (Chemistry UCL)

Dear Miguel,

I don't know any specific database bu maybe the link bellow may be a start.

http://www.crystalmaker.com/library/. Maybe the software can also be of help.

Dear Miguel, I agree with Devis Di Tommaso. Besides the valuable papers of Antonio Tilocca, I would suggest you lokking at the works of Cristina Menziani, University of Modena and Reggio Emilia, http://personale.unimore.it/rubrica/pubblicazioni/menziani.

We work on semi-glass structures which are useful in electrical eng. usages. we first use MonteCarlo algorithm to generate the random place of atoms in a box, however sometimes give it a path to be similar to crystals at the beginning. This would be similar to sputtering effects on nano crystals. Then calculating MD simulation. However recently we try to improve Classical MD to use CP method in molecular dynamics....unfortunately i can not find any database for at least materials useful in electronic.

If you are trying to do Discrete Fourier Transform, I would like to make a few suggestions. You should try to use the Fast Fourier Transform algorithm (Cooley and Tukey) as that would give very efficient computation. Also, you could look at options like exploring the use of GPGPUs (General Purpose Graphics Processing Units like the ones built by NVIDIA) to increase the speed of your computation as they are designed to increase the throughput for inherently parallel algorithms like FFT.

Dear all, thank you for your advice. They are very usefull for beginning our work.

Sincerely

Miguel Prado

Of there is no database available, I suggest another option. You can create a glass using reactive force field molecular dynamics at high temperature. Check the work about SiO glass by van Duin

The below attachment is help for your question.

Thank you Xiangfu Wang