Simple Test for H*(10) returns weird output for Am-241

Dear FLUKA users,

I was just trying to do a really simple reference simulation to show the compatibility of FLUKA generated data, the formula H*(10)=Γ*A/r² to estimate dose rates and a measurement with a dose rate meter.

Basically what I did is set up a point source and created 3 detectors (1cm³ of water each) in 10cm, 20cm and 40cm distance respectively. I used Co-60, Cs-137, Eu-152 and Am-241 and 2E5 primaries on 10 spawns but only 1 cycle each, but I do not think that is the problem here.
For the first 3 isotopes the results deviate by a maximum of 15% between the 3 versions. But for the Americium FLUKA gives a value that is 15 times above the result of the formula and up to 30 times of the measurement.

From now on I am only looking at Am-241.
My sample has an activity of 1.69E+8 Bq (originally 1.86E8 Bq on 31.12.1967) and I am assuming a distance of 20 cm between detector and probe.

So my three results are
measurement: 13.95 µSv/h
formula: 30.9 µSv/h (I am using 7.30E-03 mSv*m²/GBq/h as dose rate coefficent for Am241)
FLUKA: 411 µSv/h

So FLUKA is way off the other two results. Let me explain how I arrived at the FLUKA result:
My FLUKA-Simulation results in a Dose of 6.755 E-4 pSv per primary. As my detector-regions have a volume of 1 cm³ I only have to multiply this by the number of primaries right?
Instead of using the number of primaries (and that is what I did with all the other isotopes as well) I multiplied by the number of decays happening in one hour. So in my understanding the resulting quantity should be pSv/h or µSv/h in this case:

6.755 E-4 pSv * (1.69E+8 Bq * 3600s/h) = 411 µSv/h

Is this assumption of connecting primaries and activity wrong?

I also changed from that Americium source to a bare 60 keV isotropic photonbeam, but the result was still too high at 58.3 µSv/h.

Can someone enlighten me and point out the error?

Attached you can find my inputfile. There is a potential shield in there, because I wanted to block everything that is not photon.
Punktquelle.flair (3.6 KB)
Punktquelle.inp (2.4 KB)

Also one other Question only related to Am-241: Why does it take so much longer to simulate a primary of Am-241 compared to other isotopes? 0.04ms per Cs-137 vs 1.56ms per Am-241 on my machine.

Edits: spelling and specification.

I have some partial answers:

  1. I am not surprised by the measurement result. Dose rate meters are calibrated with 137-Cs (662 keV gamma), many of them have a lower sensitivity at 60 keV because of their casing, attenuating low-energy photons.
  2. ambient dose equivalent H*(10) is defined in vacuum, you should use vacuum or air volumes to score fluence to calculate it, not water
  3. The branching ratio for the 60 keV gamma transition in 241-Am is 36%, see below the plot from ICRP Publication 107. This fact is included in the Gamma-coefficent in the formula.

Number 1 answers the discrepancy between formula and measurement.
Number 2 and 3 are not enough to solve the one between FLUKA and formula.

I then scored the photon spectrum of 241-Am in Fluka :

And I compare it with the spectrum in ICRP Publication 107, probably the best collection of decay radiations, energies and intensities for radiation protection.
Am-241-ICRP107.pdf (175.1 KB)
Comparison of these two plots indicates that they are not very similar. Maybe a bug in Fluka’s decay database ?
Best regards, @totto

Hi Both,

What you show Thomas is also what I obtain when I run a 241Am isotope source and measure the radiation product yield out of the semi-analogue decay.

You can indeed see the 241Am decay products (59keV gamma, 13.9 keV XR, etc.), however, keep in mind that 237Np, the daughter of 241Am is itself unstable (and so on), and you will proceed down the whole decay chain until you reach stability, thereby emitting decay products from a series of nuclides, not just the ones directly coming from the alpha decay of 241Am. I should note that this is perfectly expected behaviour as the semi-analogue mode gives products integrated over infinite time.

The three other cases you investigated, Tim, are all decaying directly to a stable nuclide, reason why you did not have the same issue arising there.

We also checked the FLUKA radiation product database, and I can report that we do not see any issue with any of the isotopes involved in that decay chain.

Cheers, hope that clears things out

1 Like

Thanks Philippe,
I was not aware that with Tim’s settings the whole decay chain is produced and scored.
Is there a way to score only the single isotope, without daughter products ?
Cheers, @totto

Thanks for both your answers @totto and @pschoofs.

@totto your answers of 1 and 3 were partially on my mind. The discrepancy between measurement and formula was no problem on my side, I just wanted to point out the whole setting. The branching ratio, which I did not pay attention to while my 60 keV photon test, will take my Photon result quite close the result of the formula.

Regarding you answer 2, I remembered H*(10) is defined for the ICRP sphere and its tissue like material, which is not too far from water. Actually, I could probably just use the “Tissue soft (ICRP)” by my argument. But I will take a look into that.

@pschoofs Well, I did not think about the infinite time integration. Although I should have thought about that when I was getting alpha particles out of an 152Eu decay. AND it explains the longer simulation time for 241Am.

At last I would like to second @totto’s question on how to simulate a single radioactive nuclide. I have already been thinking about using DCYTIMES but could not get a grip on how to get it going with the semi-analogue mode, which, as I understand it, is necessary for the simulation of radionuclides as primary particles.

Cheers, Tim

Hello Tim,

H*(10) is defined in a so-called ICRU sphere, a fictitious, 30 cm diameter sphere made from ICRU 4-component tissue, irradiated in an expanded and aligned radiation field. H*(10) is the absorbed dose in 10 mm depth, multiplied by a quality factor depending on LET (linear energy transfer). This definition cannot be readily applied in either measurement or simulation:

  1. ICRU 4-component tissue would be chemically unstable, that’s why we don’t see dosimeters in the form of 30-cm spheres on the market apart from that they would have a mass of 14.1 kg. (Neutron Rem-counters are often spheres of about 30 cm diameter, but for a different reason).
  2. Expandend and aligned radiation field means that, in a multi-directional radiation field, all radiation “vectors” are bundled and impinge on the sphere along the axis which connects the scoring point in 10 mm depth with the sphere’s centre.

As a result, the ICRU sphere is only used to calculate conversion coefficients from fluence or kerma to H*(10). In the calculation, the sphere is in vacuum. Results are published in ICRU Report 57, which is identical to ICRP Publication 74. The conversion coefficients can then be used to score H*(10) directly from fluence, without the detour of the sphere and the expanded and aligned field.

Due to the calculation method, the conversion coefficients can strictly speaking only be applied in vacuum, in air is also possible without too much error, at least for photons and neutrons.
Often, one sees simulated colour maps of H*(10) though a shielding wall, strictly speaking this is not correct, but it remains a useful illustration for the gradual attenuation of radiation intensity of all particles types by the wall.

In your simulation, just setting the scoring volumes to the same material as the surrounding VOID, either AIR or VACUUM, will give correct results.

Best regards, Thomas (@totto)


Hello Thomas,

Thank you very much for your explanation. Seems like I misunderstood quite a bit about H*(10).

I will correct my scoring volumes to AIR.

Best Regards, Tim


For the answer to the question relative to the isolation of decay products from only the 241Am part.