Experiment-Simulation comparing count values in MCNP

[gkcn]

TL;DR Summary

NaI(Tl)detector-MCNP

I made some measurements using a NaI(Tl). Then I simulated this measurement using MCNP. Now I want to compare the count values. However, when the simulation count values are very low, the ones I get from the detector are high. What steps do I need to follow to set this up?

 

[PSRB191921]

you must normalized the results of MCNP (e.g. take into account the activity of the source )

 

[gkcn]

Do I need to divide Count values into activity?

 

[Alex A]

This depends a lot on the input, what the experiment is, how you did the measurement and a lot else.

Most answers in MCNP are per source particle. If you asked for a gamma flux tally the answer will be per source particle. So if you multiply the tally by the activity of the source the result should be similar to the detected activity.

There are a lot of reasons why this might fail on a first try. Good luck!

 

[gxa]

Counting efficiency=137Cs,0.01 microsilverperhour
Count per minute=40
Source activity=87357
I've been struggling for a few weeks but I couldn't find out exactly what I'm missing. Unfortunately I can't add the tables and graphs here is some information I have. I would appreciate it if you could give a little more detail on how to use them.

 

[Grelbr42]

Background

The usual way MCNP works is to give you tallies on the basis of "per particle started." So, whatever you are tallying, say energy deposited in a target, you get the result per particle started. Then your normalize by multiplying by the number of particle per second in your source. Your source appears to be 87357 particles per second.

I *think* the counting efficiency reports something about how many counts per second the detector can accommodate before it saturates. The detector acts on the basis of ionizing radiation increasing electric current flow. It might be in a crystal or a detector tube or some such. It has to recover between each event. The efficiency seems to report that it can handle 137 counts per second before saturating. So your count rate of 40 seems to be well within the capability of the counter. But you should check the user manual for your detector to be sure that is correct.

The next number seems to be reporting that you are observing 0.01 micro-Sieverts per hour. You have written microsilverperhour. So your detector reports dose in micro-Sieverts per hour. A Sievert is a measure of equivalent damage to tissue. Different types and energy of radiation produces different levels of damage to tissue.

https://en.wikipedia.org/wiki/Sievert

H = Q x D

Dose (H) is quality (Q) times energy deposited (D).

Doing Counts

It looks like your detector can report counts.

One thing you can do is just count particles that get to your detectors. There is a tally in MCNP that counts particles crossing a surface. Possibly your detector will have an information sheet that tells you the range of energies it is sensitive to and the size of the detector active part. (A tube or a crystal or some such.) So you can then count particles that make it to your detector from your source, and filter that according to the energy the detector is sensitive to. Probably you can use cutoffs in MCNP so that any particle that goes below the sensitive range of the detector is just deleted. That will save you some CPU time.

MCNP will then give you counts per particle started. You can then multiply that by your source activity, and result in counts per second. The numbers you wrote were counts = 40, and activity 87357. So, your MCNP tally should report 4.57E-4 counts per particle started so that when you multiply that by 87357 you get 40.

You can do a rough-and-ready self check for this. Get the size of the detector. Then just do an inverse square law to estimate that fraction of particles that should get from your source to your detector, without any of them getting absorbed by anything between. You have a detector with an area of (for example) 1 square cm. What distance away should it be so that the 1 cm^2 area is the correct fraction of the sphere to receive the correct fraction of particles? At a distance of 10 cm, for example, the sphere has an area of 1256 cm^2, so the 1 cm^2 detector sees 1/1256 of the total, or 7.96E-4. So I'm guessing you are holding your detector about 12 cm from the source.

Doing Dose in Micro-Sieverts

First you need the energy deposited per particle. This is straight forward in MCNP. You just tally the energy that gets deposited. It's an F6 tally if I recall correctly.

But then there is a quality multiplier. This takes into account the fact that different energy radiation produces different degree of harm to tissue. So, visible light is pretty much harmless. Soft x-ray does one level of damage. Hard x-ray another. If it were neutrons, another, etc.

One way to do this is to apply an average value for the range of radiation you are working with. That might be something you can look up in a radiation handbook that covers your source. Or you may have an information sheet for your source.

If you want to get really official with MCNP, you can apply a dose multiplier table. If you can get yourself a radiation dosimetry handbook, you can find such a table. That way you can multiply the quality factor right in MCNP. You get a table of multiplier values at each energy range, and apply that to the tally. That is the FM card in MCNP.

Again, the result will be per-particle-started. So you have a source of 87357 per second, or 3.1449E8 particles per hour.

So you tally energy deposited per particle started, and multiply by your quality factor to get micro-Sieverts dose per particle started. Watch those units because MCNP gives you the energy in MeV. And then you multiply by your particles per hour to get micro-Sieverts per hour. If that 0.01 micro-Sieverts per hour is your target, you should get something like 3.18E-11 micro-Sieverts per particle started.

 

[Alex A]

"Counting efficiency" ought to be the number of detected events for a given flux. If 40 isn't the calibration of the detector then that line is missing a value. If we knew the dimensions of the detector that would help confirm, but if the calibration is given in uSv/h it sounds like the crystal isn't part of the simulation. Unless it is and the conversion is also supposed to be done within MCNP. Which MCNP is capable of doing as an energy deposition tally.

So much within radiation protection is a very carefully judged fudge. I can't find an easy way to turn a 662kev based flux into a Sv based flux.

We're just not being told enough of the problem.

 

[gxa]

Thank you very much for the information you provide.These are detector information.I added the data I received and the excel file I prepared for comparison.These are the data I get when I put the source(60 Co) 5 cm away.This source source activity 87357.2 becquerel

 

[Grelbr42]

You really need to get yourself one of the standard handbooks on radiation protection. Maybe you can get it from your local university library if you don't want to pay for it, or if your company does not want to pay for it. Possibly if you Google search for standard handbook of radiation protection you will find something.

60Co has a keen conversion from source intensity to dose. I don't happen to have it on hand because I don't do much radiation protection. In 15 years using MCNP I have only done it twice, and borrowed a coworker's copy.

But because 60Co is such a well-known isotope, there is a standard conversion between dose at a fixed distance and intensity of the source. So-and-so many decays per second produces such-and-such many micro-Sieverts per hour at 1 meter. That would be very helpful for you to get a good check on your numbers.

 

[gxa]

Thank you, but what I want to do in my study is to make the count values and energy values obtained by simulation similar to these values obtained from the experiment.

 

[Alex A]

Ok, is this MCBEND not MCNP?

In 60Co-5cm-MCBEND-graph.xlsx the blue line which seems to be W is generated from E 'C.P.S.', what exactly is this and why does it only have one high energy peak?

The scintillator is 13mm x 13mm x 20mm ?
What shape is it?

Is the scintillator modeled inside MCBEND? If so how, ie what exactly is being counted?

 

[gxa]

Actually I first asked this question for MCNP and cylindrical NaI(Tl) detector. However, I lost the usb with some of these data and I need to repeat some experiments. These are my old data. This detector is rectangular prism and the data is for MCBEND.I runned this code for 50000000 particle history so I tried to multiplied count values and 50000000.I sent right graph and MCBEND code.The scintillator shape is rectangular prism.

 

[Alex A]

This is a CsI detector not an NaI detector? I do not know MCBEND so I will trust that result is correct.

Do not multiply by 50'000'000. Multiply by the activity. If the result needs to be for 1 minute multiply by activity*60.

Because the scintillator peaks are broad and the simulation peaks are sharp the heights will not match directly.

The heights of the peaks should match the areas of the peaks for the real experiment. Just sum the columns around the peaks.

 

[gxa]

This result is for CsI(Tl) detector.If I multiply the MCBEND count values and 60*activity,these results don't match but I don't multiply 60 these results almost match.

 

[gxa]

The detector counts for 1 minute, the same way it runs for one minute on MCBEND. Do the results still need to be multiplied by 60?

 

[Alex A]

These results are now for Cs-137 not Co-60, is the real scintillator CsI(Tl) or is it NaI(Tl)?

The results look okay to me, if I do
=SUM(C871:C1180)
The answer I get is 2754
I would not copy that, I'm doing this quickly by eye.

MCBEND*activity*60 is,
3.49e3
(It is not spread out so I do not have to sum)

Does that sound right?

 

[gxa]

Yes these values are right.This greaph for Cs-137 and CsI(Tl)

 

[gxa]

These results are now for Cs-137 not Co-60, is the real scintillator CsI(Tl) or is it NaI(Tl)?

The results look okay to me, if I do
=SUM(C871:C1180)
The answer I get is 2754
I would not copy that, I'm doing this quickly by eye.

MCBEND*activity*60 is,
3.49e3
(It is not spread out so I do not have to sum)

Does that sound right?

One is the results I got with the simulation and the other is the results I got with the real experiment, is it logical for these graphics to be like this? Since my background is not physics, I do not fully master these issues, so I apologise if my questions are absurd.