Re: Explaining the difference between Electronic Dosimeter

["Sandy Perle" <sandyfl@earthlink.net>]

Mike Lantz, Palo Verde Nuclear Station, called me and provided the 
following additional information. Mike allowed me to post his 
information, to further assist others out there (I'm doing this even 
though Mike and I do disagree on some of the finer points, since they 
are based on a generic situation seen at many facilities, where Mike 
may be assessing based on his own program, which is extremely well 
run). In my opinion, Mike is the foremost expert in EPD response, due 
to his extensive testing, conducted over many years.

Mike wrote the following

Sandy wrote:
I appreciate Sandy's efforts to clarify comparisons between EDs and 
TLDs.  It is sometimes a difficult process, but is ALWAYs better if 
all the issues are on the table, as Sandy has attempted to list.  
I've added supplemental thoughts to his excellent list:


> I receive this type of question often, and decided to post my reply, 
> removing the information regarding the source of the latest question 
> and facility. Since the question is raised often, it is probably a 
> question that others ask themselves, but never put it into writing. 
> If this can be of assistance, then it's worth taking the time to make 
> the dialogue available to everyone. Of course the discussion 
> addresses a TLD and an Electronic Dosimeter. However, the philosophy 
> is valid when comparing two devices of any type, that are different 
> and unique.
> 
> My reply to this particular individual follows:
> 
> There are many differences. Each dosimeter has its own 
> characteristics, and needs to be dealt with independently, even in 
> the manner in which they are calibrated. 

One of the issues here is to stay away from box irradiators.  
Problems with scatter dominate the analyses.  Try to reference both 
dosimeter systems to a single exposure situation such as a Cs-137 
point source on a phantom and then acquire as much information about 
the each dosimeters’ variations in energy, angle and field 
conditions. 

> Some thoughts:
> 
> (1) The TLD accumulates dose from the time it is annealed, prior to 
> shipment, until the time it is processed (heating to 300 degrees C). 
> Therefore, all exposure, occupational, natural background and transit 
> dose (higher natural background) is accumulated. The Control Badge is 
> used to define what is natural and not occupational dose, and is then 
> used to report the net dose on the personnel dosimeter.

In many cases, this can dominate the differences found between TLDs 
and EDs for the majority of workers at low dose.  Great care must be 
taken to ensure that the control TLDs are representative of the 
worker dosimeters.  Don’t store them in cabinets, lead pigs, the 
Dosimetry Supervisor’s Office, store them interspersed with the 
stored personnel badges.

> The EPD only measures exposure while the unit is on.
> 
> (2) The TLD uses an algorithm that allows the device to measure and 
> report dose for all types of radiation, i.e., gamma, beta, x-ray and 
> neutron, and, dose ranges based on specific methods of calibration.
> 
> The EPD is generally calibrated to Cs-137, and highly under-responds 
> to much higher energy photons and low energy x-rays, and will not 
> respond to beta (unless you have Siemens or perhaps, the new MGP 
> EPD). In any event, the SDE is not as accurate as the DDE.
> 
> When a facility is in an outage, and there is much work performed in 
> high energy photon araes, where the average energy is around 1 MeV 
> (due to Co-60), the TLD is highly accurate, where the EPD will under-
> respond to this high energy photon component, by ~ 10 to 15%. This 
> causes the TLD to report a higher dose than the EPD. It is very 
> important to know the fileds one has worked in when attempting to 
> compare a TLD / EPD ratio.

Sandy and I may differ slightly here.  I see the electronic dosimeter 
underresponse in outage and non-outage.  Yes, in a point source 
configuration of Co-60, the EDs underrespond compared to a Cs-137 
exposure, but that doesn’t translate to NPP field underresponse.  Co-
60 is a significant component of our exposures, but so is Co-58 and 
our rad environments are always an isotropic field of highly 
scattered photons.  In my opinion, the underresponse of EDs is 
primarily due to the mass of the dosimeter/battery and the pin diode 
characteristics (e.g., reading 30% low to Cs-137 if backwards).

> (3) Many calibrate their EPD to match the TLD. This is not a correct 
> method. They act characteristically different, and need to be 
> assessed differently. Many EPDs have set points, and how these are 
> manipulated determines how the EPD reports dose. The EPD generally 
> estimates a dose rate and then converts to a cumulative dose reading. 
> If there is a spike in the reading, there may be the potential for an 
> over-response for the cumulative dose.

I want to rephrase this.  Yes, do not calibrate EDs to simply match 
the TLDs.  Isolate the list of real differences as Sandy has done 
here and see how they affect the measurements.  If, at that point, 
you find that the ED’s are generally close, but are 8% low on 
average, increase them by 1.1 to be slightly conservative day to day. 
 Use stringent criteria to then identify reviewable dosimetry 
results.  I’ve been presenting new criteria to the industry which has 
helped quite a bit in finding the individual dosimeter problems that 
are related to system problems – in order to fix/eliminate those 
problems.

> (4) TLD / EPD correlations. The TLD only reports dose > MRD, and in 
> ICN's case, as is the industry standard, we report dose > 10 mrem. 
> The EPD generally measures dose rate in 0.1 mr/hr (so they say) and 
> will report a 1 mrem cumulative dose. The real issue is in how the 
> facility issues and reads EPD dose for personnel. Do you give a 
> person an EPD and they wear it for the whole day, a certain time 
> period, or do they exchange for a different EPD on each entry into 
> the radiation controlled area? If there is a new EPD used for each 
> entry, there is the issue of EPD accuracy and precision, and, the 
> small incremental dose that is received, and rounded up on each 
> entry. This could lead to higher EPD over-response.

This is generally a small problem overall, but there are lots of 
little scenarios here related to how the access control software 
stores the Ed measurements:  (TLD = 13, ED = 1), eg people who make 
lots of entries where the TLD accumulates all the dose and the Ed 
software rounds 0.4’s to zero or truncates 0.9’s to zero.  (TLD = 9, 
ED = 23) people who make 10-12 hr, low dose rate entries where they 
accumulate some natural bkg on the ED and a little plant dose where 
0.4 or 0.5 + nat bkg is always converted to 1.0 mrem on ED.  But this 
is not the big problems I see in the industry. But remember, the LLD 
for an ED is generally 0.5 mrem PER ENTRY. 
 
> (5) The EPD is susceptible to external factors, such as RF, cell 
> phone etc. These do cause spikes in many EPDs used in the power 
> reactor world. This is getting better, but not there yet.

If you read my article a few years ago, you know that this is the 
side of the ED problems.  Catch these spikes.  Simply add a feature 
to your access control to download histograms whenever an ED has a 
max dose rate greater than say, 1 R/hr.  Those kind of RFI spikes 
cause false readings and you can correct the files.


> (6) The TLD using LiF is tissue equivalent. Therefore, the dose is 
> easier to measure and report. The EPD is not tissue equivalent, and 
> is only as good as the source used to perform the calibration. As 
> individuals move through the plant, they are exposed to all kinds of 
> radiation, and energies, due to primary and scattered radiation. The 
> TLD algorithm accommodates this. The EPD can not, and will assume all 
> response due to Cs-137 equivalent energy, 0.662 MeV. All of the low 
> energy x-ray will under-respond, and the high energy gamma, > 0.662 
> MeV will under-respond. How the TLD / EPD ratio works at the end of a 
> monitoring period is dependent on these mixtures and time frames of 
> exposure.

Exactly right.  The TLD is complicated because it can be with several 
elements and filters.  However, if we were simply to calibrate it to 
Cs-137 and write no algorithm, it would do better than Eds in a much 
wider set of environments.

> (7) The EPD is highly geometry dependent, and the batteries do skew 
> the data sometimes. The TLD is more tolerant to geometry, and is not 
> nearly as much a problem.

Yes.
 
> (8) EPD accumulated dose in ROM does not fade. TLD dose will fade 
> based on the TL material used in the dosimeter. If it is a Panasonic 
> TLD, which incorporates both Lithium Borate and Calcium Sulfate 
> elements, each element fades independently of the other. If it is a 
> Lithium Fluoride TLD, it too has a specific fade characteristic 
> (similar to Lithium Borate). We incorporate fading correction in our 
> algorithm. In that we can not determine the specific fade for each 
> specific dose received by the individual, we use the mid-point of the 
> wear period for calculating the fade correction factor. In other 
> words, for a quarterly wear period, the fade used is based on 45 
> days. The variance around that point is not significant. What is 
> significant is when the individual receives ALL of the dose at the 
> very beginning of the monitoring period, or at the very end of the 
> period. In the first case, the TLD under-responds, since the dose 
> requires a fade correction of 90 days, and we used a 45 day fade. For 
> the latter case, the fade to be used may only be on the order of a 
> few days, but the 45 day fade will cause an over-response. These are 
> issues that can be addressed between the processor and facility PRIOR 
> to the dose calculation process.

Again, yes.  Work with your processor with any and all information you have.  
Cooperation is the key here.




------------------------------------------------------------------------
Sandy Perle					Tel:(714) 545-0100 / (800) 548-5100   				    	
Director, Technical				Extension 2306 				     	
ICN Worldwide Dosimetry Division		Fax:(714) 668-3149 	                   		    
ICN Biomedicals, Inc.				E-Mail: sandyfl@earthlink.net 				                           
ICN Plaza, 3300 Hyland Avenue  		E-Mail: sperle@icnpharm.com          	          
Costa Mesa, CA 92626                                      

Personal Website:  http://www.geocities.com/capecanaveral/1205
ICN Worldwide Dosimetry Website: http://www.dosimetry.com

************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html