Dear Joseph,

I´ll try to answer tomorrow. I need some time.

Survey meters used for radiation protection are usually good enough to indicate levels of exposure. (The term exposure has been sufficiently muddled that I intend it generically, not associated with units or particular definitions.) The measurements are indicators of need for protection. The level of protection is easily determined by a competent professional for the particular exposure situation. 

Instead of operational units in Sievert, units of Gray free-in-air are adequate to radiation protection. A radiation protection professional should know the characteristics of the instrument and the radiation field and thereby devise radiation protection.

A calibrating beam for instruments should be well characterized at the calibration location to allow calculating the free-in-air Gray or measured free-in-air Gray by a standardized probe.

The survey instrument manufacturer should be required to document the energy and intensity response of the instrument so the radiation professional can evaluate the appropriateness of a particular instrument for an exposure situation.

It is not productive to calibrate instruments to imaginary quantities in units of Sievert if the calibration has no bearing on actual exposure situations. It is not productive to claim an instrument to be one-size-fits-all with a calibration to Sievert when it is not and cannot.

Dear Joseph,

I first read our old discussion again, in order not to repeat to much.

I think you had a look at my chapter about doses and dose units. The central point there is, you can calibrate a dose meter using exaggerating geometries like enlarged and adjusted fields in special phantoms. That means that your calibration factors contain  attenuation and scattering from the phantom and maximal enlarged fields. If you now add the radiation quality factors Q you indeed can factor in the kind of radiation (light or dense ionizing).

I have strange problems to use the unit Sievert (Sv) for "physical" operational doses used for measurements and body doses like organ dose and effective dose which are used to estimate risks.

For the information of other participants in this question I add again my textbook chapter, where you can find the calibration geometries, phantoms and dose definitions etc.

Dear Hanno

Exactly. An instrument can be calibrated with a phantom and a given radiation field for a position in the phantom. There is no reason to believe that the instrument is calibrated for a different radiation field for the same or a different location. A survey instrument is normally hand-held, free-in-air. Why invent operational quantities different from a phantom and usual calibration geometries? Why have several operational quantities when one will do? Why not, say, relate a free-in-air measurement to AP geometry to the abdominal region of a standard phantom?

An instrument should be characterized free-in-air for all energies and intensities for which it is proposed and how it relates to a point in a standard phantom.

A radiation professional needs to understand how different source geometries affect measurement and dose, at least, in general.

@Roland

These same type definitions, measurement techniques, and instrumentation problems are inherent in all fields of safety. Don't get me started on aerosols, radon, formaldehyde, ...

Dear Joseph,

thats mainly a problem of calibration. And your free in air measurements say for some low energy photons surely exaggerate the "real" dose or dose rate dependant from photon energy absorption and scattering. So far I understood, this effect is wanted in radiation protection because  you want to measure "at the secure side" of error bars.

Since a serious calibration for all particles, photons, energies and intensities is impossible, I prefer the current over estimation or specialised detectors for all special cases. Very often it´s no problem to restrict your radiation field to one kind of quanta, say eg 140 keV photons from 99mTc in nuclear institutes.

Hi Joseph! Thank you for opening this line of discussion.

I do remember that this discussion was very heated also at the begining of the 90s, when these operational quantities were introduced. In order to give my opinion (because "answer" is an inappropriate term in this case) in need to go into a bit longer discussion so please bear with me.

The operational quantities of interest are 3: personal dose equivalent, ambiental dose equivalnet and directional dose equivalent. So let's take them piece-wise and see if we can really calibrate our instruments to measure these quantities.

The personal dose equivalent, Hp(d), is the dose equivalent in ICRU tissue at a

depth d in a human body below the position where an individual dosemeter is

worn. Normally it comes in 3 flavours: Hp(10) for penetrating, Hp(0.07) for skin dose and Hp(3) for eye lense.

Now, for Hp(10) the calibration can be done for individual dosemeters with relative ease and is dose on a routine basis. All one needs is a good electrometer and a special Hp(10) ion chamber. But then comes the problem of real measurement, because Hp(10) is not only energy dependent, but also angle dependent. The energy dependence can be solved with relative ease if you use different filtrations, but for the angle dependence one must from the very begining define a combination of angles that would fit as close as possible the real incidencies of the worker and include also this factor into your calibration. Ideally I should do this for every workplace monitored which means that every single dosemeter will have a different calibration factor and for a dosimetry service monitoring a few thousand people this is completely not feasible. So what is the answer for this? Well, the services use a standard combination of incidencies for calibration (in the ideal case) or simply take into account only the normal incidence, hoping that the measurement will still be within the requirement fo the standard, i.e. -100% to + 50% error. Hp(0.07) is even worse, as the angular dependence is worse, tne energy deoendence is a nightmare, so most people do not even measure it. Hp(3) simply cannot be measured as defined because I will not be able to put a dosemeter in front of the crystaline. Attempts have been made to use glasses with tranparent detectors on the lenses but people find them cumbersome and do not wear them

So, first conclusion? Hp(d) is not a very practical choice. Hp(10) can be used, with some care, Hp(0.07) and Hp(3) are totally unpractical.

My next comment will be on H*(10)

Now, for H*(10): it is defined in 10 cm in the ICRU sphere, in an expanaded and aligned field. As Hanno stated, most standard laboratories can create these conditions and therefore one can calibrate a survey meter in ambient dose equivalent. For a good meter, the detector will be energy compensated and thus the energy dependence of the response should be kept within +/- 15% of the reference energy and a good calibration certificate will also state the responses at different energies. So, from my point of view, due to the fact that a good laboratory will be able to incorporate all these influences in a calibration factor, among the operational quantities this qould be the most practical because a meter calibrated in H*(10) can be used with enough confidence that the indication will be, within reasonable limits, the ambient dose equivalent as defined in ICRU.

For the directional dose, however, the H'(d, omega), I do not see any hope. The angle dependence makes not only the measurement impossible with any degree of reasonable accuracy, but also the calibration, in my opinion.

And now we come to the Sv and the interpretation of the results and this is, in my opinion, the biggest problem of them all. So I will write a separate problem on it.

Dear Hanno

Yes. When there is moderate over-response, that is preferred. I am concerned about extreme over-response and any amount of under-response. I have observed ion chamber under-response for 99mTc.

A hospital that uses nuclear medicine, x-ray machines, and accelerators will have a broad range of radiation energies and types.It is important to have the appropriate instrument. 

Accidents and clean up require strict allocation of worker time. A large amount of over or under response may have serious consequences.

...sorry, I meant a separate comment.

We go back to the discussion you opened previously. As both you and Hanno stated, using the Sv both for equivalent and for effective dose makes no sense and creates confusion. The operational quantities - practical or not - are meant to be exctly that: operational, i.e. they should be an instrument for mid-level dosimetry technician that will do some measurements, get a reading, compare it with the administrative limit and move on to taking the measures prescribed by the standards if the case.

The problem is that the Sv is also a unit of risk when use for effective dose and therefore many people will not interprete the readings only as they should do (i.e. a shielding is correctly made or I followed or not the  radiation protection procedures to avoid un-necessary exposure) but they tend to interprete it in terms of risk, especially if they saw some paper saying "so many Sv.man mean this increase in cancer incidence". In my experience, no matter how hard I tried to explain to the laymen that the operational quantities must not be directly related to risk, but they rather need a further interpretation, I was not succesful. So yes, as both of you have proposed before, a new unit is needed for the operational quantities to separate them clearly from the effective dose.

That's it, I apologise for beating arround the bush for so long.

Just a small correction to Radu, H*(10) or Hp(19)  means 10 mm not cm depth in phantom.

A short reply to dear Joseph.

I rember different detectors for medical work of course dependant on the different fields of application. And if we needed accuracy we bought instruments with clearly defined and restricted application and error bars. I think if some layman detects "radiation" say with a simple GM, immediately a specialist has to appear and has to analyse and perform quantitative measurements of radiation kind and quality and doses or dose rates. And only after this procedure one could try to define some risks by interpreting exposition of persons by Sieverts.

@Hanno: yes, you are right, sorry, it's a mistake from fast typing.

I agree with Hanno on the measurements made by laymen. I had these problems time and again with people using dosemeters and geting scarred when the indication went up from, say, 90 nSv/h to 110 nSv/h. And it is very difficult to explain to people that risk is, at the end, a matter of statistics and one cannot relate directly any operational quantity to the risk of a single person to get cancer or to any another stochastic effect. In my experience, people tend to see the stochastic effects as delayed deterministic effects and this scares them a lot.

Dear Radu

I developed an instrument that was capable of measuring AP H(10), (0.07), and (3) for photons and electrons, including angular dependence and energy dependence. Electronics at the time (1980s) prevented a long battery life and a weight less than 3 kg. I also developed personal dosimeters that did the same with a weight comparable to 2-chip TLD.

My opinion is that the operational units are useless, in particular ambient dose equivalent. We have AP, PA, ROT, ... defined. Use one of those. If you are building a chicken house if may be useful to start with a spherical chicken to develop the concept. At some point, you will have to progress to, at least, an oblate spheroid with legs, if you intend to be practical.

Some standards laboratories have gone to great expense to approximate an expanded, aligned field of a single energy. Yes, the errors are small. It is practical for some detectors of prescribed sizes. These are useful for inter-comparison and standardization, especially free-in-air. It ends there. AP H(10) is a calculation. Next comes spherical chickens.

Joseph, in my opinion the usefulness comes precisely from the fact that we can compare measurements without worrying - I mean for H*(10) about AP, PA or other geometries. Aligning the field does just that and, for radiation protection measurements, if I am far enough from the source, I can easily consider that the field is expanded and aligned within a reasonable error.

 The danger comes from the interp

Sorry, for some reason it was not posted completely. What I meant is the advantage comes from the fact that the measurement can be performed by any technician, but the danger comes from the interpretation of the results. The interpretation should be done only by a well trained physicist or engineer.

Radu

It is easy to specify conditions for comparison without the requirement for imposing a spherical chicken. Nothing, as in not one thing, is added by the chicken. What is the advantage to imposing an imaginary or actual sphere to a point in space of a sufficiently aligned and and expanded field? I already agreed that the error is small making a measurement in an aligned and expanded field. No one has defined (to my knowledge) what is reasonable, an essential requirement for establishing a standard measurement,.

What type measurement are you proposing to be made by a technician for interpretation by an appropriate expert?

I do not believe that the analogy to the spherical chicken is correct. The correct analogy would be to the average chicken because the expanded and aligned field does just that: it averages the conditions from the measuring point over the entire detector volume.

But we should probably just agree to disagree on that, because I do not believe there is a right or wrong opinion on this. In the end, if one wants to evaluate the risk, one has to calculate the effective doses. And it really does not matter wherefrom you start, exposure, absorbed dose, dose equivalent, eauivalent dose etc. - all that matters is that equivalent conditions should lead to equivalent risk calculations.

I don´t like the chicken and agree with Radu. Find a good starting point, be consequent and in some step of your procedure, integrate risk factors.

It sounds like you are agreeing with me when you state "all that matters is that equivalent conditions should lead to equivalent risk calculations." The spherical chicken is the ICRU sphere. An expanded and aligned field should give average conditions over the average volume of the detector. Nevertheless, it is necessary to specify what constitutes an acceptable expanded and aligned field. I agree to agree.

Here is the next part of the problem. Why effective dose? No one (not even ME) can measure effective dose with a survey instrument. Why claim to calibrate to effective dose? If dose is assigned using a survey instrument, someone must do an analysis. A survey instrument measures free-in-air. We should have a free-in-air standard, not effective dose of some esoteric flavor.

Radu

What type measurement are you proposing to be made by a technician for interpretation by an appropriate expert?

Hanno, Radu

What do you have against chickens?

As I understand the operational quantities first proposed by ICRU was to overcome the difficulties in estimating the organ doses as per ICRP-27 issued in 1977. ICRP-27 mentioned that the organ doses should be estimated by constructing a sphere of 30 cm diameter of unit density and the organ at the centre. This resulted in confusions of overlapping, non-additivity etc. Also, there were different maximum permissible limits making difficult to implement. To overcome these difficulties new operational quantities called ambient dose-equivalent, directional dose equivalent, personal dose equivalent were defined. Realising the difficulties in the direct measurement of these quantities ICRU published conversion factors from air-kerma measurements to arrive at these quantities. With these quantities a single quantity of whole body dose equivalent simplified implementation of dose limits.

I attended an IAEA/RCA workshop in 1994 at JAERI, Japan mainly on the topic of calibration of survey meters and personal monitors in terms of these quantities. The IAEA coordinator Dr. R.V. Griffith when asked about establishing the aligned and expanded field required for the measurement of these quantities mentioned that at a source to detector distances of 2 meters and more the field approximates to these definitions. One important change brought by these quantities is in the definition of energy response of detectors. Compared to earlier flat energy response the energy response should follow the changes of the conversion factors with energies. For example, for 60 keV Am-241 photons the factor is ~1.75 and for Cs-137 it is ~1.20.

I think BIPM, Paris is the only laboratory with facilities specially designed to measure these operational quantities experimentally.

Thanks Ganesan

Historical information is helpful in establishing how we got to where we are. If there was a decision on aligned and expanded it should be documented. Nevertheless, there must be a defined, measurably-acceptable definition of aligned and expanded.

Thanks Joseph

I think the following reference available online will help you.

"Bases for the selection of operational radiation protection quantities", A.Allisy, Radiation Protection Dosimetry, Vol.29,No.1/2 Pages 135-138, 1989.

Effective dose is based on radiation energy, organ weighting factors, and direction of irradiation. Organ weighting factors are subject to change (the data supporting them is sparse) and must be applied after determining the irradiating conditions. The reference suggested by Ganesan states, "the underlying data ... permit the determination of particular dose equivalents received by exposed individuals when further information on irradiation conditions is available." Calibrating an instrument to an arbitrary dose equivalent force fits the instrument response to a meaningless quantity. One needs further information on irradiation conditions in order to estimate dose equivalent to an exposed individual. Why calibrate to an arbitrary definition of effective dose if it has no meaning? How does this aid designing radiation protection?

Dear Joseph, nothing against chickens per se, but I prefer the average conditions as already stated. ;-)

Dear Joseph,

I try to understand your last comment "Calibrating an instrument to an arbitrary dose equivalent force fits the instrument response to a meaningless quantity."

Of course you have to input measured or estimated doses into an algorithm to derive effective dose. The reason is, no measured quantity knows risk factors because of different factors depending on the exposed body region.

So we must measure a helpful dose quantity and use these measured results to do risk estimates later on. What I really regret is the use or misuse of measured Sv as effective dose. Thats totally wrong.

Dear Joseph,

I did not say that we should measure effective dose, because this is not possible. I was just saying that, in the current models, the effective dose is the measure of risk so one should calculate the effective dose starting from a measured quantity. As Hanno just said above, in my opinion, too, we should concentrate on eliminating the confusion between the equivalent dose Sv and the effective dose Sv - here lies the problem.

Now, coming to the ICRU sphere, I do not see any difficulty related to it. Hp(10) is anyway defined in 10 mm in the human body, so nothing to do with the sphere, whilst the H*(10) is defined at 10 mm depth in the sphere but for all practical calibration purposes I can use a sphere, a slab phantom of the same material, or whatever geometrical form I like provided my phantom will give me the same backscatter as the ICRU sphere and it will be made from a material with the same Zeff.

And here comes also the answer to the "technician measurement" I measured and the reason why I think the operational quantities make the life easy on these technicians. The administrative limits are always realted to the Hp(10). The measurement of the H*(10) can be very easily related to Hp(10). All the technician has to do is to measure H*(10) with an appropriate instrument and then, via some very easy arithmetic, he can say if the shielding of a room is in agreement with the regulations or not. If one wants t estimate the risk, then he should take the measurements, feed them into a nice mathematical model together with a lot of other parameters, calculate the effective dose, then feed that result in another mathematical model for risk, and finally get some numbers. In the end, those numbers will have to be used carefully, because it will not help if I tell the radiographer that the doses he will get carry a cancer risk of , say, 1 in 10 billions if he does not know how does this compare to the risk of him getting ill or even dying from other reasons.

And finally, to answer your last question: I have nothing against chicken especially when well cooked. And I can tell you that I make a de-boned chicken which is almost spherical and is delicious, despite his strange geometry... ;-)

Radu

First the technician question. It was not clear if you meant a measurement in a calibration facility or in an exposure situation. The units reported by the technician are not important. They must be consistent. Today, I can show you instruments in use that are calibrated (?) in R, rem, rad, cpm, cps, Gy, Sv, and multiples of those units. As you say and as we all do, we make the rules and assign doses based on what is reported. The units do not matter. What matters is that we know what the mean and how to interpret them. It would be much better if we settled on one unit that we could know how it was calibrated. That is my point. That is your point. No disagreement. It should be that way.

Hanno, Radu

We all agree that the Sievert with all its variants is unfortunate. Hanno said, "So we must measure a helpful dose quantity and use these measured results to do risk estimates later on." I would amend this to: So we must measure a helpful dose quantity free-in-air or on a phantom and use these measured results to do risk estimates later on.

Calibration of an instrument or dosimeter requires specified conditions as single energy, aligned and expanded field. There should be specified acceptable limits on the calibrating conditions. The field intensity at the calibration point determines the free-in-air or on phantom dose value. There is no reason to impose, suppose, or calculate the placement of an ICRU sphere. There is no reason to invent a quantity for the sphere. The calibration unit should be Gray.

Yes. Many do the ICRU sphere calibration. That is no reason to keep it. Everyone who uses the ICRU concept must characterize the field. Why add the sphere and three sub units? 

Radu. I am sure your spherical chicken tastes better than the ICRU spherical chicken. To me it is too large a pill to swallow.

All of this makes me wonder WHY a chicken would EVER cross the road!  It seems that the best that it can expect is to be killed and eaten!  ha ha

On another note, the dilemma of Sv vs. Gy is rooted in the scientifically unsupported Linear No Threshold Theory (LNT).  what would this discussion look like if that confusion were eliminated?

Mark

It would not matter. All you have to do is change the weighting factors or when and how to calculate them. The question here is what came first, the Sievert or the ICRU sphere (nothing to do with gallinaceous fowl and an edible oblate spheroid)?

Hanno

"Calibrating an instrument to an arbitrary dose equivalent force fits the instrument response to a meaningless quantity."

Suppose you have an ion chamber with a flat in-air response from 50 to 600 keV. You calibrate the instrument to 100 keV x-rays in Sv. What does this calibration (or calibration to any other energy) mean for the instrument when measuring another energy?

Joseph,

I wouldn´t use this "calibrated" chamber to measure Sv in another radiation field than the calibration told me. To get estimation I would analyse first the energy transfer coefficients if measuring photons mutr and then analyse entry windows for attenuation and absorption of the radiation field. And I must repeat it, I only would measure equivalent dose but never the other radiation doses with unit Sv, organ dose Ht or effective dose E.

Hanno

No disagreement. I hope all the people in the medical field are doing the same as you. Your knowledge and attention to detail are not common in radiation protection. 

That´s the reason dear Joseph why I still are giving lectures.

A question rather than an answer. What do you recommend for measurable quantities in space and aviation radiation fields? These are mixed fields with 5 or more significant components and are omnidirectional (often assumed isotropic).

I would try to use different detectors. There are methods to distinguish radiation fields. In space and aviation you don´t have the financial problem to enlarge your equipment.

I agree with Hanno, but would say use the appropriate detector for the type of radiation and energy range. The number of detector types may be limited, so there is a need for discriminating for energy and type. The measurement should be in Gray 'in-air' for the pressure and mix of atmosphere. Conversion to risk requires model estimation.

Dear Joseph,

so silent here. Is everything discussed? I think we should really try to restrict the Sv to effective dose. I´m completely astonished how many users mix up the different dose definitions.

Hello Gentlemen,

I have also been silent, after stiring this up - I apologise. From all this discussion, my conclusion are the following:

1. it does not really matter what dosimetric quantity you measure, what really matters is to follow some simple steps: choose the correct detector, take care it is appropriately calibrated, measure consistently and then use with care and consistently the models that will allow you to estimate the risk 

2. most mid level users tend to confuse the equivalent dose Sv with the effective dose Sv and thus draw incorrect conclusions about the risk. If we stay with operational quantities (and I have nothing against this) we should change, at least, the measuring units from Sv to something else (if we stick to the operational quantities, than we cannot revert to Gy either, we need a new unit - let's call it, provisionally, "equivalent Gy" - and if we give up to operational quantities, like Joseph propose, things are even simpler because we can get back to Gy)

Bottom line is: what is the proper line of action for proposing ICRU and ICRP to change the measuring unit of the equivalent dose from Sv to something else?

Hanno

The problem with understanding and using the dose definitions has more causes than there are definitions. Here in the USA we have the problem of regulations that are written in rads and rems, but many requirements expressed in Gy and Sv.

The problem starts with the ICRP. The definitions and practical units are purely academic (necessary) with no recommendations for implementation. It is important that there is a method to apply risk as interpreted by the LNT as the basis for radiation protection. Nevertheless, risk should be applied to the final assignment of energy deposited and not to measurements and measurement methods.

Radu stated that most European countries have accepted and developed practices to calibrate to ambient dose equivalent. The problem is there is no instrument that can perform an energy independent measurement of ambient dose equivalent under the best of circumstances. All fail miserably in practical situations. (It is the responsibility of radiation protection to work with the measurements to ensure safe working conditions.)

The question by C.S. Dyer illustrates the problem, except it is extreme compared to usual dose situations. How do you get to risk if you do not know the composition and direction of the radiation field? How do you get to risk if you do not  know the response of your instrument? ICRP's solution seems to be to invent a lots of units and call them all Sievert.

Yes Joseph and Radu,

all the proposals for calibrations assume a known radiation field. So as I stated earlier, you must use different dosemeters depending on known or suggested field components. And it´s better in radiation protection to overestimate an equivalent dose than to underestimate it. But besides all, no one can measure effective dose with the supposed and implemented risk. You even can´t calculate the effective dose without great efforts because of unknown corpus data of all possibly exposed persons.

So delete the Sv for the operational doses, and we must not discuss LNT or any other models during our measurements.

Hanno and Radu

Sorry I can offer only one up vote to your last answers.

Public consultation of the draft ICRP-ICRU joint report "Operational Quantities for External Radiation Exposure" has just started, which will end on 3 November 2017. Proposed changes include the operational quantities in absorbed dose for the ocular lens and local skin.

http://www.icrp.org/page.asp?id=355

With respect to Hanno's point above, for cosmic radiation (and its products in the atmosphere) you would need a dosemeter for every known radiation particle as well as all the unknown ones.

The draft report "The Use of Effective Dose as a Radiological Protection Quantity" prepared by ICRP Task Group 79 is now open for public consultation, which will end on 3 August 2018. The draft is attached FYI.

http://www.icrp.org/page.asp?id=382

Thank you Dr. Hamada for attached the draft of ICRP Task Group 79

"The Use of Effective Dose as a Radiological Protection Quantity"

The answers have been awesome.

More details coming.... compiling