MCNP Modelling Detector Energy Response Vs Empirical Results

[jjames_gunn]

Hello All,

I'm trying to simulate in MCNP the energy response of a PIN diode. To do this, I have modelled a "slab" of silicon in an epoxy case at 2cm away from the source and with the F8 tally set to 25keV bin increments to 1MeV, I do as follows:

1. Set the source energy to 33keV
2. Run the simulation,
3. Record the gross total pulse height distribution (PHD),
4. Subtract the 1e-5 (Sigma) bin from the total PHD,
5. Subtract counts below 20keV,
6. Record the net PHD for that energy and normalise to Cs-137 (662keV),
7. Repeat for energies of 48keV, 65keV, 83keV, 118keV, 161keV, 205keV, 248keV, 662keV and 1.25MeV

Below is the MCNP Simulated Plot Vs the Empirical Response.

To clarify the empirical experiment, I have a PIN diode connected to a charge sensitive amplifier and gaussian shaping circuit, I have a comparator set to 20keV, any pulse that goes above this threshold gets counted and displayed in counts per second. The dose rate for the empirical experiment was set to 10mSv/h.

Can anybody explain the discrepancy at lower energies? I believe MCNP results are "per photon emission from the source" meaning that the dose rate presented to the diode would vary as each photon carries a different "weight", and the empirical response compensated for this; I've tried compensating the simulated results by reducing the response by the flux needed to keep the dose rate the same but to no avail.

Below is my MCNP code:

Energy Response UnCompensated
c **********************************************************************
c Cell Cards
c **********************************************************************
c
c Pin Diode Silicon
00001 1 -2.32 -1
c Aluminium
00002 4 -2.7 -2
c Epoxy Housing
00003 2 -1.3 1 -3
c Room
00004 3 -0.0012 1 2 3 -4
c Explicit Blackhole/Universe
00005 0 4
c **********************************************************************
c SurfaceCards
c **********************************************************************

c Diode
1 BOX -0.14 2.00 -0.14 0.27 0.00 0.00 0.00 0.03 0.00 0.00 0.00 0.27
c
c Aluminium
2 BOX -0.23 1.91 -0.20 0.45 0.00 0.00 0.00 0.02 0.00 0.00 0.00 0.40
c
c Epoxy Housing
3 BOX -0.23 1.93 -0.20 0.45 0.00 0.00 0.00 0.20 0.00 0.00 0.00 0.40
c
c Room
4 BOX -1.00 -1.00 -1.00 2.00 0.00 0.00 0.00 5.00 0.00 0.00 0.00 2.00
c
c **********************************************************************
c Data Cards
c **********************************************************************

MODE P
IMP:P 1 1 1 1 0
SDEF POS=0 0 0 ERG=D1 VEC=0 1 0 DIR=D2
SI1 L 0.033 $ Source Energy (SI Source Information)
SP1 D 1 $ Source Probability
SB2 -31 100 $ Source Bias
F8:P 1 $ Energy distribution of pulses (photons) created in a detector
E8 0 1e-5 0.025 38i 1.0 $ 50keV bins to 1MeV
c **********************************************************************
c Materials
c **********************************************************************
c
c Silicon
M1 14000 1
c Epoxy
M2 6000 -0.706
8000 -0.171
1000 -0.085
7000 -0.032
c Air
M3 1001 -0.00111967185672092
1002 -3.58077068509439E-7
2003 -7.01300681482183E-13
2004 -6.93002257732448E-7
6000 -1.269899924930E-4
7014 -0.744226018468389
7015 -0.002930937363427
8016 -0.238274132603651
8017 -6.338532733037E-4
10020 -1.188009929771E-5
18036 -3.8063224269139E-5
18038 -7.57346759599989E-6
18040 -0.0126264625590237
36078 -9.80896820559013E-9
36080 -6.47380864579173E-8
36082 -3.3656973211545E-7
36083 -3.37936500433013E-7
36084 -1.69483116165613E-6
36086 -5.26125375559769E-7
54124 -3.55783232684075E-10
54126 -3.37976986796745E-10
54128 -7.36915932270391E-9
54129 -1.02645376834607E-7
54130 -1.59506563063874E-8
54131 -8.38324527235628E-8
54132 -1.07054801524348E-7
54134 -4.21479276142738E-8
54136 -3.63078420449797E-8
MT3 LWTR.20T
c Aluminium
M4 13000 1
c
NPS 100000000
PRINT

TIA

 

[Alex A]

Hi @jjames_gunn,
Welcome to Physics Forums. Cool experiment. A photodiode with no scintillator. So the material needs to generate a photoelectron from a photon and then totally stop it inside the depletion region of the device. This depends on the photoelectric effect cross section of the material and it's stopping power to the resulting electron. For silicon of thickness 300um these events become unfavourable over 100kev so results may look a bit garbage. Photoelectrons that escape result in a noise floor to the spectrum.

How are you calibrating your experiment? Am-241 is quite common. Are you sure the spectrum shown is from a 33kev source? What sources did you use in the real experiment? What is your PIN diode and does it have any sort of metal backing? What reverse voltage are you running it at and what is the noise level of your system?

 

[jjames_gunn]

Good Morning, Alex

Thank you for the reply. The diode in question is a BPW34, it does have a metallic backing that I have not modelled and it's attached to a FR4 PCB. I have modelled the PCB in a prior experiment, it didn't make much of a difference but I'm willing to reinstate it given a good argument.

The 33keV was derived from an X-ray machine in compliance with ISO 4037 Narrow Spectrum Series, N-40 at the UK Health Security Agency in Harwell, Oxfordshire - a prestigious testing facility.

The noise floor of the circuit was measured to be 16.29keV, this was done by setting the comparator threshold in increments until no pulses were detected; away from any source of radiation. However, a comparator voltage of 20keV was chosen to allow some overhead. I am reverse biasing the BPW34 with 25V DC.

Kind Regards

Jonathan

 

[Alex A]

Hi Jonathan,

These are my guesses. I would suggest switching to mode pe. The thickness of the PIN diode silicon is probably right, but the depletion region the signal comes from is probably less than 100um thick. The circuit board is probably not important but the metal behind the silicon might be. I don't understand the SB card. I guess the intention is to fan out the beam a bit, is the use of that card right? That you have a lower bin of 25kev but a cut off of 20kev bothers me a bit, but like the noise level, that is what it is, as they say.

It might be worth looking at the actual spectra - the raw numbers in whatever form they are. How did you calibrate the amplifier? One of the lower energy peaks? I'm having a bit of difficulty wrapping my head around what the final response function is, absorbed dose? It seems too flat to be per photon.