Answer to Question #10100 Submitted to "Ask the Experts"

We use LiF TLD-100 and a Harshaw 3500 TLD reader in our hospital for various applications of radiation dosimetry. It’s been our experience for the last 10 years that the measurement from the TLDs (thermoluminescent dosimeters) is accurate for a certain number of exposures. After that, radiation measurements with these TLDs underestimate the amount of exposure. However, once the TLDs are recalibrated we again get better results from the TLDs. What property of the TLD-100 reduces the sensitivity after few uses? What should be the frequency of calibration for the TLDs? Is there any publication on this aspect?

While the general use of TLDs for routine dosimetry purposes is quite straightforward, simplifications and conveniences are often implemented that make operations easier for the users but may not be consistent with the optimization of TLD performance and stability. TLD-100 (LiF[Mg:Ti]) is among the most commonly used TL materials but it is one of the more complex materials in terms of its characteristics especially with regard to the electron-hole trapping centers.

There are numerous traps of varying depth, each representing potentially different energies that may be observed in the glow curve structure when the dosimeter is read out following irradiation. The detailed glow curve for TLD-100 exhibits at least 13 peaks. The major dosimetry peak used in routine photon dosimetry applications is a major peak with its usual maximum at about 210^o C. In routine use the trapping centers are sufficiently stable so that the variance associated with numerous sequential uses is acceptable. Unfortunately, however, the distribution, frequency, and availability of trapping centers available for radiation dosimetry are subject to various influencing factors.

The various trapping centers are defined by different energy depths (or temperature), but they are not all thermally independent, so that there may be transitions among traps such that one energy trap may deplete while another is enhanced, depending on thermal effects. One of the most important influencing factors is the annealing history to which the dosimeters have been subject. Different annealing cycles can result in changes in dosimeter sensitivity. For example, if the user depends on the typical readout cycle of the reader to anneal the dosimeter, the dosimeter does not get the complete anneal that provides reasonable assurance that the dosimeter will be returned to the same state of sensitivity (i.e., same trap distribution), and changes in sensitivity may be observed with time. The recommended full anneal for LiF(Mg:Ti) is 400^o C for one hour followed by a longer anneal at a lower temperature—e.g., two hours at 100^o C, or 24 hours at 80⁰ C, or 18 hours at 75^o C—have all been used with success. Because of the time involved in this process, especially if many dosimeters are being used and rapid turnarounds are necessary, users often opt to allow the TLD reader cycle to suffice for annealing purposes. Often the annealing process is not adequate, and ultimate decreases in sensitivity may result.Even when the full anneal is applied, one must settle on a consistent way of carrying it out. For example, if the cool-down period between the high temperature (400^o C) anneal and the lower temperature anneal is very short, as might be the case if the TLDs are removed from the oven and placed on a heat conducting surface, thermal defects produced in the crystalline material appear to be enhanced; these defects lead to more trapping centers, which in turn result in higher TLD sensitivity. When a slower cool-down period is used, as when the TLDs are allowed to cool to the lower temperature in the oven, fewer such defects are produced, and the sensitivity is relatively lower. Many users prefer the latter approach. In any event, whichever technique is used, it is important to be consistent and use the same method every time that the TLDs are annealed.

Even with care in the annealing process many users recommend that the TLDs not be used more than about 50 times, but this is quite subjective, depending partly on what level of dose uncertainty the user is willing to accept.

In addition to the above considerations that bear on changes in sensitivity some other factors may be influential in reducing sensitivity. One physical effect is the presence of dirt, grease, etc. that might accumulate on the surface of the TLD, and reduce the amount of light getting out of the dosimeter when it is read out. This can usually be minimized by careful handling, storage, and cleaning. Also, if high doses (usually >100 Gy) are delivered to the TLD, damage to the crystalline structure may result in reduced response. Finally, the elevated temperature anneal, when carried out in air can cause some oxidative damage at the surface of the TLD, and this can also produce reduced sensitivity; the effect seems to be most noticeable if low penetrating radiations are being measured.

If TLDs are being used for personnel dosimetry in which the TLDs provide the doses of record for legal purposes, the laboratory/facility that processes the dosimeters must demonstrate an acceptable level of performance. Typically, such demonstration requires the laboratory to participate in an accepted accreditation program such as the NVLAP program.

If the TLDs are being used internally within your facility and not for purposes of providing personnel doses of record, such formal involvements are not generally required, and the users often establish their own program to confirm that the TLDs are performing as required. This involves calibrating and recalibrating the dosimeters as needed. For dosimeters that are not heavily used, an annual calibration may be adequate. For dosimeters that are used frequently or if they receive unusually high doses such that the dosimeters are operating out of the usual linear response region or are not being annealed sufficiently to ensure removal of residual dose (when high doses have been encountered), more frequent calibrations for the conditions of use may be warranted. Similarly, if routine use has not included a full or consistent annealing process, sensitivity may change with time, and more frequent calibrations may be required, as you have found.

You should have a quality assurance program in place that routinely checks the performance of the reader and the TLDs. If you are using the TLDs for critical measurements, such as monitoring the output of a medical therapy unit, it is also desirable to participate in testing and quality assurance comparisons with other facilities. These can be very helpful in identifying systematic biases that may be present.

There are numerous references available through an Internet search. There are several American National Standards Institute (ANSI) standards that relate to TLD performance and testing; among these are ANSI N13.37 Performance, Testing, and Procedural Specifications for Thermoluminescence Dosimetry (Environmental Applications) and N13.11-2009, Personnel Dosimetry Performance – Criteria for Testing. There are also a number of technical/textbooks dealing with the theory and the applications of thermoluminescence. An early, seminal text is Thermoluminescent Dosimetry by J.R. Cameron, et al., Univ. of Wisconsin Press, 1968. It is dated in many respects, but still offers some insightful and understandable explanations of various aspects of the subject, especially with respect to LiF. Numerous other texts are available; you can find many of them online—e.g., type "thermoluminescent" into the search box on Amazon.com (in Books) to bring up several; typing "thermoluminescence" will also bring up some others.

I wish you well with your continuing use of thermoluminescent dosimetry.

George Chabot, PhD, CHP