Difference in the drift_time_histo between GenericPropagation and SimpleTransfer modules

Simulation / Configuration Questions
1

Dear experts,

I just have a small question that is puzzling me without being able to answer it.
What is the difference between the drift_time_histo we get from the GenericPropagation module and the one we get from the SimpleTransfer module? Which one should I use? Since the results are different.
I looked into the source code but couldn’t know.

Excuse me for the naive question.

Best regards,

Hasan

2

Hi @hdarwish

  • GenericPropagation will note the drift time of any charge carrier until its end-of-motion state. This might be reaching the sensor surface (HALTED), but also being trapped or recombined (TRAPPED, RECOMBINED), or even the ones which are still in motion at the end of the integration (MOTION).
  • SimpleTransfer only notes the drift times of charge carriers which it adds to its output signal, i.e. not the ones too deep in the sensor.
  • “Which one should I use?” is a difficult question, because for analysis purposes I would rather suggest storing the final information of your simulation to file and analyze it with a script to your liking

Cheers,
Simon

3

Hi @simonspa

Thanks a lot for your reply, point taken.
I just have another small question, I checked the manual but it didn’t clarify my confusion.

There are two parameters for monolithic sensors: chip_thickness and sensor_thickness:
I am confused by the difference. Especially that for ALPIDE, as defined by default in allpix, the chip_thickness is 25um and the sensor_thickness is 25um, but I think the sensor_thickness should be 50um for ALPIDE, as it is the thickness of the active volume.
So where am I mistaken?

Thank you and best regards,

Hasan

4

Hi @hdarwish

this is indeed a valid question, maybe @pschutze could clarify as the author of the detector model.

/Simon

5

Hi @hdarwish ,

the reference publication we used for parameterising the ALPIDE model claims that the sensor is 50 um thick, while the epi-layer makes up for 25 of that. Choosing chip_thickness = 25um and sensor_thickness = 25 um should reflect this situation, where only half the detector is sensitive.

If you would like to have a model where the whole depth can be used as an active volume, you can copy and change this model and simply change the detector type in the geometry configuration accordingly.

Cheers
Paul

6

Hi @pschutze ,

Thanks a lot for your clarification.
Just to be sure:
Isn’t the substrate also simulated?
Despite that it’s highly doped, to my understanding, it can contribute to the signal.
Thus, it should be included in the sensor_thickness parameter, which results in more than 25um as sensitive volume.

Please correct me if I am wrong, and thanks for clarifying my confusion.

Best regards,

Hasan

7

Hi @hdarwish ,

of course you can also simulate the substrate, simply copy/change the model file and use other parameters. To clarify: when we set sensor and chip thickness to 25/25 um, it will not be simulated. Any energy deposited in the chip is ignored as this is not part of the sensitive volume. If you change the thicknesses, you can simulate the substrate.

We made this approximation since the number of charge carriers coming from the substrate is not very large. We neither have information on the doping concentration nor on the electric field, hence we were to choose between to approximations: using the substrate leads to an overestimation of charge, declaring it as “chip” leads to an underestimation. If you have information on doping and electric fields, of course you can gain a lot in terms of precision.

Cheers
Paul

1 Like