A problem about carriers drifting in Sentaurus TCAD

[Huazhen]

Hi experts,

Recently I am using "HeavyIon" in Sentaurus TCAD to simulate a high energy particle going pass through a device and how the carriers drift inside the device. I found an interesting result: the cloud of carriers generated is stretched along the electric field direction, and it becomes a elliptical shape (or in another word, the diffusion seems to be stronger along the e-field direction). I can not quite understand this, because the number of e-h pairs generated is quite small (the carrier-carrier interaction should be weak) and the time of carrier generation is negligible...

Do someone know the possible cause of this?

Cheers
Huazhen

 

[dick_freebird]

The "funnel" is the extension of built-in junction fields along the
plasma track. There is a lot of literature on single event effects in
IEEE NPSS Trans Nuc Sci and conference papers / posters for the
space electronics industry.

Charge-pair density is not low when you have one of the heavier ions.
The charge collection impulse often will dwarf explicit operating
currents or local charge storage.

 

[Huazhen]

The "funnel" is the extension of built-in junction fields along the
plasma track. There is a lot of literature on single event effects in
IEEE NPSS Trans Nuc Sci and conference papers / posters for the
space electronics industry.

Charge-pair density is not low when you have one of the heavier ions.
The charge collection impulse often will dwarf explicit operating
currents or local charge storage.

Thank you very much for your reply!

So if I am understanding correctly, The field along the plasma track will still be affected by the heavy ion, so the carriers will not only face the field caused by electrodes?

If so, is there any method to avoid this? Because we only want to simulate the induced current on electrodes caused by the drift of fixed number of charge pairs generated at fixed time and position... Can sentaurus TCAD somehow keep the field fixed while doing the transient simulation? Or, instead of using "HeavyIon", is there any solution to introduced only some charge pairs?

Sorry for so many questions, I am currently new to this effect. Thanks for your patience.

Cheers
Huazhen

 

[dick_freebird]

I don't know much about TCAD beyond the superficial and nothing
at all Sentaurus specific.

An ion traversing (presumably) silicon has a braking dE/dx which is
known, and this can be calculated (w/ material properties) into a
freed ehp density at t=0 (you can call evolution delay in a relativistic
heavy ion through a few microns or a hundred microns, "sub-timestep").
This (I imagine) would be a step-function free carrier bump along a
zero radius track oriented however you say (at least for space part
interests, omnidirectional must be assumed - a fixed detector:source
relation might let you play more games).

It would not be surprising if Sentaurus recognized the interest in such
analyses and maybe had some app notes in pocket. Universities (or
certain ones) love this stuff because you don't need to wait for silicon
or test time to be donated but you can get academic licenses because
"chumming".