Shani G. Radiation Dosimetry: Instrumentation and Methods

Подождите немного. Документ загружается.

250 Radiation Dosimetry: Instrumentation and Methods

241

Am source. [22] The energy of the alpha particles was

degraded with thin mylar foils of different thicknesses.

The net glow curve is shown in Figure 4.40.

The TL glow curves were deconvoluted using the ﬁrst-

order TL kinetics model according to the ‘‘Podgorsak

approximation” method:

(4.22)

where

T  temperature (K), I  intensity of the TL glow

signal at the temperature T, I

 the maximum height

of the glow peak, E  trap depth of the glow peak (eV),

 temperature at the maximum peak height (K), and

k  Boltzmann constant 8.62  10

5

eV/K. The trap

depth E (eV) is a linear function of the maximum peak

temperature T

; therefore, the following approximation

is valid:

(4.23)

The responses of extruded TLD-600 and TLD-700

ribbons (Solon Technologies, Ltd., Solon, OH), 3.175



3.175  0.889 mm

, were calibrated by McParland and

Munshi [23] against the thermal neutron ﬂuence in a mod-

erated neutron ﬁeld created in a tank of water (60  60

 40 cm

) containing a

252

Cf spontaneous ﬁssion source.

FIGURE 4.36 Effect of nitrogen gas ﬂow on the standard deviation and linearity of large (a) and small (b) LiF TLD chips exposed

to the



-rays of a 10-mg Ra Eq

137

Cs tube. The relative sensitivity is deﬁned here as the ratio of the sensitivity (TL/cGy) for a given

dose to the sensitivity of the TLDs with the absorbed dose of 22 cGy for large chips and 35 cGy for small chips and located at the

same source-to-detector distance. (From Reference [20]. With permission.)

IT() I

1 ET( T )k

1

()[exp

ET( T

1

2

()]exp

E 25kT



Ch-04.fm Page 250 Friday, November 10, 2000 12:01 PM

Thermoluminescent Dosimetry 251

The source intensity was 2.1  10

neutrons s

1

, and the

TLDs were placed 15 cm from the source.

The thermal neutron response of the TLD-700 was ini-

tially investigated by comparing the response of the bare

TLD-700 in the moderated neutron/photon ﬂux with that

with the dosimeter encapsulated in one of a number of

thermal neutron-absorbing shields. The absorbers consid-

ered were TLD-600 (LiF with 95.62%

Li and 4.38%

Li),

cadmium, and indium. A natural tin shield was also used.

The response of the TLD-700 was unaffected by the place-

ment of the TLD-600 absorber about it, whereas, in con-

trast, cadmium, which has a thermal neutron absorption

cross section about 2.5 times that of

Li, elevated the TLD-

700 response by a factor of almost 2. The indium led to an

even greater increase in the TLD-700 response, whereas the

tin shielding did not cause a signiﬁcant elevation.

TLD-700 chips were irradiated [24] with the fast neu-

trons from

241

Am/Be and

252

Cf sources, energy-degraded

neutrons from the

241

Am/Be source moderated by light

water, and gamma-rays from a

137

Cs source. The net area

under the deconvoluted high-temperature (300°C) peak

of neutron-irradiated chips was observed to increase with

the average neutron energy. The high-temperature part

(250–340°C) of the TL glow curve was deconvoluted using

FIGURE 4.37 Effect of nitrogen gas ﬂow on the linearity of large (a) and small (b) LiF TLD. The relative sensitivity is deﬁned as

the ratio of the sensitivity (TL/cGy) for a given dose to the sensitivity for 50 Gy. (From Reference [20]. With permission.)

Ch-04.fm Page 251 Friday, November 10, 2000 12:01 PM

252 Radiation Dosimetry: Instrumentation and Methods

the ﬁrst-order TL kinetics model according to the “Podgor-

sak approximation” method (Equation (4.22)). [25]

All glow curves were normalized to the highest peak

height at T

 220°C (Figure 4.41). It is apparent that the

area under the deconvoluted high-temperature glow peak,

A(HT), increases with the average neutron energy. It is

evident that the high-temperature (300°C) peak of TLD-

700 dosimeters is sensitive to gamma-rays as well, but to a

lesser extent than to fast neutrons.

III. LIF:Mg, Cu, P DOSIMETER

The development of high sensitivity TLD by doping LiF crys-

tals with Mg, Cu, and P was ﬁrst done by Nakajima et al. [25]

The sensitivity of the new TLD was more than 20 times

higher than that of LiF:Mg,Ti. Wu et al. [26] showed that

LiF:Mg,Cu,P (LiF(MCP) maintains its sensitivity during

prepared reuse cycles. The TL characteristics of

LiF:Mg,Cu,P include, in addition to high sensitivity, almost

ﬂat photon energy response, low fading rate, and linear dose

response. The sensitivity and glow-curve shape are both

dependent on the maximum readout temperature and the

pre-irradiation annealing parameters. Short low-tempera-

ture annealing of 165°C for 10 s prior to readout is capable

of removing most of the low-temperature peaks. As a result,

there is very little fading up to two months or more at room

temperature. High-temperature annealing at 400°C results

in irreversable elimination of the main dosimetric peak and

causes some increase in the high-temperature peaks.

Advantages of LiF:Mg,Cu,P include high sensitivity

as compared to LiF:Mg,Ti, almost ﬂat photon energy

response, low fading rate, and linear dose response. The

lack of supralinearity at higher dose levels is particularly

useful for accident dosimetry and eliminates the source of

error usually associated with the application of supralin-

earity corrections. The main drawbacks are still the rela-

tively high residual signal and the loss of sensitivity for

high-readout temperatures. LiF:Mg,Cu,P is interesting in

low-dose measurements due to its high sensitivity and its

good tissue equivalence.

The glow curve of LiF(MCP) consists of several over-

lapping glow peaks. The main peak at approximately

220°C, known as peak 4, is the one used for dosimetry

applications (the “dosimetric peak”). The rest of the glow

curve consists of a low-temperature part in the range of

approximately 70–160°C (peaks 1, 2, and 3), and a high-

temperature peak at approximately 300°C (peak 5). There

is evidence that the glow curve of this material is even

more complicated where peaks 4 and 5 are each composed

of two overlapping peaks.

The high sensitivity, combined with its tissue equiv-

alence, is the main advantage of this material in per-

sonal dosimetry applications. The sensitivity of

LiF(MCP) is approximately 25 times higher than that

of LiF:Mg,Ti (TLD-100). It is important to note, how-

ever, that the measured sensitivity depends not only on

the TL properties of the material itself, but also on the

spectral response of the light detection system. Both

LiF(MCP) and LiF:Mg,Ti have the same effective

atomic number (8.2) and could therefore be expected

to have a similar photon energy response in reality. The

over-response of LiF:Mg,Ti at 30 keV is approximately

35% (relative to 662 keV), as compared to only 6% for

LiF(MCP).

The sensitivity of LiF:Mg,Cu,P was studied by Furetta

et al. [27] as a function of the annealing temperature and

of the repeated cycles of annealing-irradiation-readout. A

fading study was carried out over a period of 40 days with

the purpose of checking the stability of the stored dosim-

etric information as a function of different annealing tem-

peratures. 10 LiF:Mg,Cu,P phosphors were cycled 10

times according to the following sequence: annealing, irra-

diation, readout.

FIGURE 4.38 Dose build-up in solid water for a 6-MV x-ray

beam, ﬁeld size 10

 10 cm

. Three TLD chips of different

thicknesses (nominal thickness nc: 0.89 mm, tc: 0.39 mm, and

xtc: 0.14 mm) were placed at different depths in solid water. The

thickness of the chips was taken into account, assuming the

active center of the chip to be in its physical center. Error bars

show the range of uncertainty (

2 SD) for the ﬁrst chip of each

type. (From Reference [21]. With permission.)

Ch-04.fm Page 252 Friday, November 10, 2000 12:01 PM

Thermoluminescent Dosimetry 253

Figure 4.42 shows the TL emissions obtained after

various thermal procedures. Figure 4.42a (annealing at

220°C) displays the dominance of peak 3 over peak 4,

which appears as a shoulder on the descending part of

peak 3. At 240°C (Figure 4.42b), the TL emission appears

as usually observed: peak 4 becomes the main peak and

peak 3 is now a shoulder on the ascending side of the

fourth peak. For temperatures larger than 240°C, peak 3

tends to become smaller and smaller, so that the main

peak becomes narrower (Figure 4.42c, at 270°C). It was

noted that the shape of the TL emission, through the 10

repeated cycles, remains unchanged for each tempera-

ture. The plots of the peak 3 and 4 TL integrals, for each

temperature and as a function of the repeated cycles, are

reported in Figure. 4.43. The TL responses were normal-

ized to that relative to the ﬁrst cycle, at 240°C. The best

reproducibility is obtained at 240°C. At the beginning,

the TL emission at 220 and 230°C is a little larger than

at 240°C, but in the next cycles, the decrease of the TL

is obvious and reaches the 30% level in the tenth cycle

at 230°C. The situation is worst for temperatures larger

than 240°C. Temperatures higher than 240°C could

produce a sort of trap disactivation which increases

with the number of cycles. In this sense, a high anneal-

ing temperature ( 240°C) seems to produce a progres-

sive quench of the TL emission.

Figure 4.44 shows the plots obtained in a fading exper-

iment over 40 days of storage. All the results are normalized

to the TL response of peaks 34 obtained after annealing

at 240°C for “zero” days (reference value). The stability of

the TL response is perfect at 240°C annealing, while a

decrease is observed at any other temperatures. It is obvious

that the trap stability is compromised at temperatures differ-

ent from the reference value. The maximum loss is observed

after 40 days for the group initially annealed at 270°C: the

remaining TL is only 43% of the reference TL value.

The response of LiF:Mg,Cu,P thermoluminescence

dosimeters to high-energy electron beams used in radio-

therapy was investigated by Bartolotta et al. [28] They

found that LiF:Mg,Cu,P phosphor is a suitable candidate

for quality control of in vivo dosimetry in electron-beam

therapy. The TL chips (4.5 mm in diameter, 0.8 mm thick)

used were produced by Radiation Detector Works

(Beijing) and are commercially known as GR-200A.

FIGURE 4.39 (a) Dose buildup in LiF as determined with extra-thin (xtc) and thin (tc) TLD chips stacked on top of each other on

the surface solid water. The depth is given as the physical depth of the center of each chip. The reading of a single normal TLD chip

(nc) placed on the surface is indicated by a horizontal line. (b) Dose build-up in LiF as determined with extra-thin (xtc) and thin (tc) TLD

chips stacked on top of each other in solid water. The highest chip is Bush with the solid water surface. The depth is given as the physical

depth of the center of each chip. The reading of a single normal TLD chip (nc) at the surface is indicated by horizontal line. (From Reference

[21]. With permission.)

Ch-04.fm Page 253 Friday, November 10, 2000 12:01 PM

254 Radiation Dosimetry: Instrumentation and Methods

Figure 4.45 shows the TL signal of the GR-200A

dosimeters vs. absorbed dose in the entire investigated

dose range. Each data point corresponds to the mean

of nine readings. The error bars represent the overall

uncertainty of the mean to a level of conﬁdence of

95%.

The sensitivity of GR-200A dosimeters to electrons was

found to be about 13% less than that of

Co gamma-rays,

in agreement with similar results already known for

LiF:Mg,Ti (TLD-100) dosimeters. Variance analysis of the

data showed the dependence on electron-beam energy rela-

tive to each other to be nil or buried under the measurement

uncertainty. When used for dose measurements in high-

energy electron beams, it is adequate to calibrate the

LiF:Mg,Cu,P dosimeters with any one electron-beam energy

without need for energy correction, but if they are calibrated

with a

Co beam, an appropriate energy correction must be

applied.

Mailed dosimetry for radiotherapy is one of the most

demanding applications of thermoluminescence dosime-

try (TLD). Ideally, the uncertainty of a mailed system

should be comparable to that of the ionometric methods

employed for the calibration of the radiation beams, i.e.,

around 1% (1



), and certainly lower than 5%. [29]

The simple dose dependence of GR-200 is in con-

trast to the supralinear dependence of TLD-100. The

proportionality between TL and dose found for GR-200

is in favor of this material, as calibration can be sim-

pliﬁed requiring fewer calibration points than with

TLD-100.

Very good reproducibility was found for the individ-

ual sensitivity factor of every dosimeter. Throughout the

experiment and for 20 groups of dosimeters, the worst

value found for the standard deviation of 10 determina-

tions of the individual sensitivity factor of a dosimeter was

0.65%. Values of 0.2–0.4% were the most frequent.

Sets of dosimeters were stored at room temperature

(18–24°C) for periods of 7, 15, 30, and 60 days. Every

set was composed of three groups of ﬁve GR-200 dosim-

eters irradiated to a dose of 2 Gy at different points of the

storage—at the beginning, in the middle, and at the end,

designated SA (storage after irradiation), M, and SB (stor-

age before irradiation), respectively. [29] Peak 4, the main

peak in the GR-200 glow curve, increases in intensity

during the ﬁrst days of storage, stabilizing afterwards.

Peak 3, a weak peak, decreases monotonically with time

at room temperature. The evolution of the peaks reveals

that peak 4 experiences mainly trap effects, leading to an

increase in the number of traps available after storage.

Peak 3 presents the two effects, trap effect (decreasing the

number of traps) and fading. A clear anticorrelation was

observed in the evolution of peaks 3 and 4, rendering the

sum of these two peaks a more constant dose estimator

than peak 4 alone for delayed measurements. Taking these

two peaks together, the response change is limited to only

2–4% relative to the prompt sensitivity.

An investigation of the thermal stability of LiF:Mg,

Cu,P (GR-200) compared to the more traditional LiF:Mg,

Ti (TLD-100) at 40°C and 70°C was presented by Alves

et al. [30] Samples of both varieties of dosimeters were

stored irradiated or un-irradiated in order to evaluate the

relative importance of the temperature/storage-induced

effects on either traps or trapped charges. The measured

glow curves were analyzed using the deconvolution pro-

grams developed at CIEMAT. These techniques allow the

detailed characterization of the evolution pattern followed

FIGURE 4.40 The glow curve of a TLD-600 chip irradiated with non-degraded alpha particles (E

 5.5 MeV) from a 560-Bq

241

Am source. The glow curve was recorded with a data logger at a sampling rate of 2 readings per second. The high temperature

peak

P(280°C) was deconvoluted using the “Podgorsak approximation” of the ﬁrst-order TL kinetics model. (From Reference [22].

With permission.)

Ch-04.fm Page 254 Friday, November 10, 2000 12:01 PM

Thermoluminescent Dosimetry 255

FIGURE 4.41 Glow curves of TLD-700 dosimeter encapsulated in cadmium box and irradiated by fast neutrons with average energy

of 5.1 MeV (a), 4.0 MeV (b), 3.2 MeV (c), and 1.9 MeV (d), as well as 662-keV gamma-rays (e) from a

137

Cs source. Each data

point represents the average value of thermoluminescence output from ﬁve readouts from the same TLD chips at a particular

temperature. The high-temperature (250–340°C) region of the glow curve was deconvoluted using the ‘‘Podgorsak approximation”

of the ﬁrst-order TL kinetics model. The glow curves are normalized to the highest value of the TL output at

T  220°C. The area

under the deconvoluted peak is represented by

A(HT). (From Reference [24]. With permission.)

Ch-04.fm Page 255 Friday, November 10, 2000 12:01 PM

256 Radiation Dosimetry: Instrumentation and Methods

by each individual peak during the experiment. The results

contained conﬁrm that, for both varieties of LiF phosphor,

the process affecting their respective main dosimetric

peaks is not lading, understood as the spontaneous release

of trapped charges. On the contrary, the observed varia-

tions in the TL response should be addressed to the mod-

iﬁcations experienced by the trap system during storage.

However, some slight differences between the evolution

of TLD-100 and GR-200 have been found in this compar-

ison. [30]

The decrease of the TL yield observed after exposure

periods to environmental radiation has been traditionally

interpreted as due to the spontaneous leakage of trapped

charges, usually called Randall-Wilkins fading.

The reading cycles used were linear at 7°C s

1

in a N

atmosphere up to 300°C for TLD-100 and, in the case of

GR-200, up to a maximum temperature placed 10°C above

the detected position of peak 4. Once reached (around 240°C),

this temperature was maintained for 5 s. In this way, better

reproducibility with GR-200 has been consistently

obtained, reducing the tendency of this material to decrease

in sensitivity with reuse to almost negligible levels. Natural

cooling down inside the reader was always allowed, mean-

ing a rather rapid and reproducible cooling. No additional

annealing other than the readout itself was employed

before reuse with both materials.

Figure 4.46 presents the evolution with storage time

at 40°C of the relative areas of each individual peak, TLD-

100 on the left-hand side and GR-200 on the right-hand

side. Closed symbols represent the evolution of irradiated

dosimeters (SA, storage after irradiation) and open sym-

bols dosimeters stored un-irradiated (SB, storage before

irradiation). The peaks of TLD-100 (as those of GR-200)

present a different evolution with time. While the area of

peak 5 relative to the prompt area presents a slow tendency

to increase independently for irradiated and un-irradiated

dosimeters, the area of peak 4, after a small increase, starts

to diminish after the ﬁfth day, reaching saturation in 15 days.

FIGURE 4.42 TL emissions obtained after various thermal procedures of annealing. (From Reference [27]. With permission.)

Ch-04.fm Page 256 Friday, November 10, 2000 12:01 PM

Thermoluminescent Dosimetry 257

Peak 3, on the contrary, decreases quickly from the ﬁrst

day to the tenth and then mildly moves also to a saturation

level. In the case of GR-200, the area of peak 4 also shows

a slow tendency to increase from the beginning of the

experiment, whereas peak 3 diminishes even more quickly

than the corresponding peak 3 in TLD-100. Peak 4 seems

to reach a saturation level at approximately the same time

as peak 3 and also presents a softer decrease to a saturation

level.

The same situation described for storage at 40°C occurs

at 70°C but on a shorter time scale; see Figure 4.47. Peak 5

in TLD-100 and peak 4 in GR-200 present an initial quick

rise, observed in the ﬁrst few hours. Peak 4 in TLD-100 also

decreases more rapidly to a saturation level attained earlier

than at 40°C. The decrease of peak 3 in both LiFs is also far

more rapid than at 40°C, and in GR-200 no peak 3 is detected

after two days for SA detectors. This appears to prevent the

growth of peak 4, since as soon as peak 3 disappears, peak

FIGURE 4.43 Peak 3 and 4 TL integrals for each annealing temperature and as a function of repeated cycles. (From Reference [27].

With permission.)

FIGURE 4.44 Plots obtained in fading experiments over 40 days of storage. (From Reference [27]. With permission.)

Ch-04.fm Page 257 Friday, November 10, 2000 12:01 PM

258 Radiation Dosimetry: Instrumentation and Methods

4 seems to stabilize. For all the peaks studied, the same ordered

behavior for ﬁlled and empty traps is preserved at 70°C.

A fully automatic computer program for the decon-

volution of LiF:Mg, Cu, P glow curves was described

by Gomez Ros et al. [31] This program permits the sub-

traction of the residual contribution of the high-tempera-

ture peaks, producing the best ﬁtted values for the kinetic

parameters and the areas of the dosimetric peaks.

FIGURE 4.45 The TL signal (arbitrary units) vs. absorbed dose for the GR-200A dosimeters irradiated with the 15-MeV electron

beam. The error bars represent the standard error (95% conﬁdence level). The full line is the interpolation polynomial curve. The

inset shows the linear ﬁt up to 10 Gy. (From Reference [28]. With permission.)

FIGURE 4.46 (a) Individual evolution of peaks 5, 4, and 3 of TLD-100. (b) Individual evolution of peaks 4(a  b) and 3 of

GR-200 after different storage intervals at 40°C. Closed symbols represent storage after irradiation (SA) and open symbols

represent storage before irradiation (SB). Error bars are typical 3



values. (From Reference [30]. With permission.)

Ch-04.fm Page 258 Friday, November 10, 2000 12:01 PM

Thermoluminescent Dosimetry 259

Figure 4.48a presents some LiF:Mg,Cu,P glow curves

(ii) obtained with a linear/plateau heating proﬁle (i),

together with the corresponding second readout measure-

ments (X10) (iii). Only peaks appearing during the linear

part of the heating cycle can be separated by glow-curve

ﬁtting, but the contribution of the highest temperature signal

cannot be ignored, and it is one of the most important

problems in practical dosimetry with LiF:Mg,Cu,P, espe-

cially at low doses. A procedure to subtract the residual

signal is illustrated in Figure 4.48b, in which the residual

contribution of the high-temperature peaks to the region

where dosimetric peaks appear is approximated by a

straight line. This “linear background” is identiﬁed and

subtracted in each glow curve, producing a net TL curve

where the individual peaks can be ﬁtted.

The analytical expression used to ﬁt the glow peaks

depends on the kinetic model considered for the processes

involved. In the case of ﬁrst-order kinetics the differential

equation describing the variation of the peak intensity I(t)

with time is:

(4.24)

where

n(t) is the trapped charges density, E is the activa-

tion energy, s is the frequency factor, k is Boltzmann’s

constant, and T(t) is the temperature. The solution when

a linear heating proﬁle

T(t)T





is applied can be

written as:

(4.25)

where

and I

are the temperature and intensity of the

maximum, respectively, and E

is the second exponential

integral function deﬁned as :

(4.26)

for positive values of z, in particular for z EkT

and z 

E kT. Function E

(z) cannot be evaluated analytically, but

a very accurate rational approximation can be used for

arguments greater than 1 if we deﬁne the function

(4.27)

FIGURE 4.47 (a) Individual evolution of peaks 5, 4, and 3 of TLD-100. (b) Individual evolution of peaks 4(a  b) and 3 of GR-

200 after different storage intervals at 70°C. Closed symbols represent storage after irradiation (SA) and open symbols represent

storage before irradiation (SB). Error bars are typical 3



values. (From Reference [30]. With permission.)

It()

nt()d

------------

 s

kT t()

-------------







nt()exp

It() I

----------

------







----------





exp







expexp

 E

----------





-------





------













Z() 

zt

---------





Z 0()



z() e

z()

 1



---------------------------------------------------------------------

z()

Ch-04.fm Page 259 Friday, November 10, 2000 12:01 PM