Good time of day, dear users!

I got a question.

Is there any way to simulate an electron cascade in silicium created by high energy electron?

I don't get how to introduce electrons moving around atoms and their interaction between each other and atoms.

Thank your for your attention.

Good time of day, dear users!

I got a question.

Is there any way to simulate an electron cascade in silicium created by high energy electron?

through quantum mechanics?

I don't get how to introduce electrons moving around atoms and their interaction between each other and atoms.

ever heard of the schroedinger or dirac equation?

cheers,

axel.

To wit, LAMMPS is a classical MD code. So unless you

implement an empirical potential that describes how

an electron interacts with atoms, you can't do it. You

can check out the eFF potential, recently added to

LAMMPS, but I don't know whether it is suitable for

what you want to model. You'd have to read the papers

about it.

Steve

There is a two-temperature model implemented in LAMMPS through the AtC package (I believe there is also a fix two_temperature, but I'm not as familiar with it). If the physics in that model are appropriate, you could try that.

Jeremy

some more comments...

as jeremy mentioned, it depends a _lot_ on what kind

of information you want to extract from the simulation.

the key terms here are "time scales" and "correlation".

a "hot" electron is a very fast moving object so that you

typically have to look at sub femtosecond time scales

and solve the time dependent schroedinger equation

to describe it properly and need to do very expensive

quantum chemistry calculations.

at the same time, these processes tend to be so fast,

that the effect on atoms is for all practical purposes

decorrelated. a very popular approach is to use wave

packets.

there are also people i know that do so-called ehrenfest

dynamics, where you propagate the total wavefunction

using DFT (or TD-DFT) and the atoms. this is somewhat

of an approximation (particularly TD-DFT), but you can

look at shot time scales with dynamics of small systems.

if however, the focus is more on the overall structure,

you can treat the system with a more approximate

description of electrons, by using DFT with a free

energy functional and "hot" electrons (but all of them).

this would be a good approximation for material under

strong laser excitation.

andres' eFF model, that steve mentioned, is making

some assumptions about the wavefunctions involved

and allows to treat some kind of systems, e.g. plasmas

very well.

...and just considering the injected electron as a

(localized) heat source, would be the most non-quantum

approximation that you can get.

in any case, life is not as simple as "i have a computer,

i have a software, i start a simulation, i get a result". i would

claim that to do a good simulation, you will spend _more_

time planning and analysing than actually doing the simulations,

and a "just slab everything together and wait what happens"

is a very costly approach. just look at D.E. Shaw and his

"anton" computer, that has produced massive trajectories,

yet very little new insight (so far).

cheers,

axel.