Best way to adjust Linear Electric Field to experiment parameters

Simulation / Configuration Questions
1

Hi,
I am simulating He-Ions being detected with a underdepleted Timepix detector, and measuring the deposited energy by looking at the collected charge per hit pixel at detector [energy].
The amount of charge that is collected is around half of what would be expected from experimental data.
I have been trying to adjust the electric field inside the sensor with the following (experiment) given values:

[energy]
sensor_thickness = 300 um
bias_voltage = -7.4 V (reverse bias)
depletion_depth =ca. 210 um OR depletion_voltage =ca. 40 V

I have checked the formulas used in allpix to calculate the linear electric field from bias and either depth/depletion_voltage.
When comparing the (experiment) depletion_depth with the one allpix calculates when given -7.4 bias/ 40 depletion_voltage, I see that the depth from allpix is shorter.
It seems to me like the sensor in the real world has a larger active/depleted volume at -7.4 bias, or a higher electrical field than allpix assumes. The bias is the only value that is known with high certainty.

I sadly do not have access to any TCAD field sim.

Could I use other infos about the sensor material (for example the doping concentration) to improve my electric field ?

Thanks !

As I cant attach my configs right now, here are the relevant modules:

detector:
…
[energy]
type = “timepix”
position = 0 0 357.00mm
orientation = 0 0 0
sensor_thickness = 300um
chip_thickness = 100um
…

global:
[DepositionGeant4]
physics_list = QBBC_LIV #FTFP_BERT_LIV
particle_type = “ion/2/4/2/0/false”
source_energy = 1MeV
source_energy_spread = 0MeV
source_position = 0 0 -1mm
beam_direction = 0 0 1
beam_size = 0mm
number_of_particles = 1
record_all_tracks = false

[ElectricFieldReader]
name = “energy”
model = “linear”
bias_voltage = -7.4V
depletion_depth = 210um #depletion_voltage = 40V

[GenericPropagation]
temperature = 293K
charge_per_step = 50
[PulseTransfer]
[DefaultDigitizer]
electronics_noise = 0e
threshold = 0e
threshold_smearing = 0e
qdc_smearing = 0e
[CSADigitizer]
name = “energy”
model = “csa”
integration_time = 0.5e-6s
sigma_noise = 0.0e-3V
threshold = 10e-3V
feedback_capacitance = 10e-15C/V
input_capacitance = 100e-15C/V
krummenacher_current = 25e-9C/s
amp_output_capacitance = 15e-15C/V
transconductance = 50e-6C/s/V
temperature = 293.15K
clock_bin_tot = 1ns
ignore_polarity = true

2

Hi @Linus

One parameter which you do not seem to set but which strongly influenced how much charge you see in a detector that is not fully depleted is

[GenericPropagation]
intergration_time = 25ns

Here with the default value. This means, signal formation will be stopped after 25ns, no matter what. However, part of your signal will likely come from diffusion in the undepleted volume and therefore be slower. By adjusting the integration time you can match simulation to data.

When increasing the integration time, another problem arises. When setting it very high, at some point you will see almost the entire charges because you just continue diffusing them until they reach the depleted part. So what would be necessary for a simulation that is independent of tuning to data, you would have to set a (rough) doping concentration for the sensor and then activate a recombination model, e.g. the combined SRH-Auger model.

I hope this helps!
Simon