A couple of valid answers for this question, the first I remember reading a few years ago and is covered reasonably well here http://hps.org/publicinformation/ate/q9194.html  The second; meter should have come with an energy response graph such as that seen in the following manual http://www.ludlums.com/images/stories/product_manuals/M44-7.pdf this gives you a clear indication of how the meter responds to the incoming particle, as you can see from the included graph, for the meter in question, it may not be a simple "linear" relationship.

Thanks for your answer.

Just to clarify:

My impression is that the geiger counter has no grasp of the energy of the particle. So if it detects 100 cps it can only translate that count rate to a dose rate if you know the energy from other means. After all, dose rate is J/kg. And 100 100 keV photons should correspond to the same dose as 50 photons with 200 keV each.

However, I read that geiger counters have an instrument specific count to dose rate conversion factor. And that factor is apparently more or less independent on the energy of the particles.

So, what I get from this is:

The geiger counter cannot distinguish if it was hit by 100 photons/s with 100 keV each or 50/s with 200 keV each. But it does know that it was hit by a total of 10000 keV/s?

That would mean even if the geiger counter was calibrated with Cs-137 it can work reasonably well with U-238 sources?

Ahhh gosh...... I should always read questions before answering..............Starting again its not Cs-131 but Cs-137 yes I know you mentioned ~662keV and U-238 50&110keV but one reads over that! I am afraid that a GM tube can "see a difference" between these energies unless it has energy compensation http://en.wikipedia.org/wiki/File:GM_tube_compensation.gif and this is because of the interactions that occur in matter. The photoelectric effect depends on the energy and the material in the GM tube and that is the prominent effect up to about 50-60keV then the Compton effect starts. I would urge caution in using an uncompensated tube "calibrated" for Cs-137 to measure lower energy gammas, even with compensation you can see that as the energy gets lower response diminishes. The example from wikipedia even suggests that there is a difference between 50 and 110 keV!

If your U-238 source hasn't been chemically separated from the Th-234 daughter, the most significant photons contributing to the dose rate will actually be coming from Pa-234m. This is a decay product two down from U-238 in the decay series (U-238->Th-234->Pa-234m). Important Pa-234m decay photons are 743, 766, 1001, ... keV. In which case, you should just be able to use the displayed dose rate on the Geiger counter as an estimate for the U-238 sample's dose rate. But this needs to be checked for the details of the instrument - read the manual! + everything else Lindsay has said

If you have a calibrated U-238 source, i.e. you know the activity in Bq, then you can simply calculate its dose using the dose coefficient [mSv/Bq]. Then you can set this in relation with the dose the counter is displaying.

E. Strub: Your answer ist correct, but you have to add up over all photon energies (as Alex does, saying that the major photons have energies near the 662 keV of  Cs137). What you call the calibration is essentially what the Geiger counter gives you (scaled to the full sphere).

So it's probably easier to get a dosemeter that estimates the energy of the particles. The usual electronic personal dosemeters with a diode as detecor are linear from 60 keV to 3 MeV, some go down to 20 keV. They give a good indication of the person dose in 10 mm depth (Hp(10)).

Florian: Where do you know from that the source is U238? Don't trust paper labels on old glasses or hearsay, always consider the possibilty of a wrongly labeled neutron source.

There is no simple way unless you have a compensated GM. The dose rate ratio is not equal to the energy ratio. I suggest you read Measurement of Effective Dose on my RG page. You will find a discussion of different detector types. If you have questions after reading, please do not hesitate to ask.

Thanks for all the answers.

edit: Joseph, I read your message just now, will read the text and probably get back with some question.

The U-238 was a hypothetical sample. Let's maybe get even more hypothetical to make clear what I mean:

There is a source of 100 keV gammas where the exact activity is known. Using that source I can calibrate a geiger counter to display the exact dose rate (within instrument accuracy) for that specific source.

Now if I replace the 100 keV source with a 50 keV source, the geiger will still display a dose rate.

My question is, how accurate will that dose rate be? Originally I thought the geiger counter would display a dose rate 100% too high since it assumes each count has 100 keV while it actually has 50 keV now. But I'm not so sure anymore, as it seems geiger counter can be build to compensate for that difference in energy, so the displayed dose rate would be more or less acurate?

You will likely get more counts from 50 keV (depending on the detector) so that a dose rate reading at 50 keV will be nearly correct. What is correct will depend upon the definition of dose you use.

I try to be quantitative ..

The doserate dD/dt (dose D per time t) deposited by a source of activity A in a distance R is:

   dD/dt = Gamma * A/R^2

Gamma is the so called "exposure rate coefficient" (or "Dosisleistungskonstante" in german). Gamma is proportional to the photon energy and the mass absorption coeffient mu. So the ratio of the doserates of two sources 1, 2 is simply

    (dD1/dt) / (dD2/dt) = (E1 mu1 A1 / R1^2) / (E2 mu2 A2 / R2^2) / 

The display of your Geiger counter is proportional to (A/R^2), so you can *measure* this quantity. The energies E you seem to you know. So just look up the mass absorption coeffcients in water or air (almost the same) and you are done.

Be careful with the units. If you look for the ratios only they should not matter. Otherwise units in dosimetry are a nightmare, take some time to get them correctly. The details you'll find in any textbook, e.g. in Krieger.

@Patrik

You are missing the characteristics of the detector and the translation to effective dose. You do state that care must be taken with units, but it is not that simple. If the GM were calibrated to dose at 662 keV, the likely calibration is Hp(10). Hp(10) does not follow mass absorption coefficients.

Detector response characteristics are highly energy dependent at lower energies.

As previously, I suggest "Measurement of Effective Dose" as it discusses the dosimetry units and compares the units to detector response.

If the actual dose rate were needed for a situation, say for dose reconstruction, considerable effort is needed to determine the response of the detector. If the dose rate were needed to design radiation protection, most situations allow reasonable estimates of characteristics and additional safety factors.

To simplify this job which dear Patrick defines, here is a private copy of a chapter about dose rate constants.

The initial question was answered, thanks again.

I had a discussion with a collegue today, however, where the following point came up:

Not-compensated GM tubes generally give a much higher response to low energy gammas. So you would expect the dose rate to be much too high. But is that really the case?

Say 100 50 keV gammas hit a 1 kg mass. Then the dose rate won't be 5000keV/kg since not all gammas will be absorbed, correct? But the absorption rate will be higher than that of 600 keV gammas, I think. So isn't the larger count rate of the 50 keV gammas somewhat justified? Because the same stronger interaction that falsifies the tube's readings happen in the body.

So for correct dose rate determination, isn't it also important to know what fraction of a gamma of a given energy is absorbed?

 No. The GM does not react like the body. Please see this and Figure 20.