Shani G. Radiation Dosimetry: Instrumentation and Methods

Подождите немного. Документ загружается.

370 Radiation Dosimetry: Instrumentation and Methods

Using a diamond detector for the dosimetry of brachy-

therapy sources was investigated by Rustgi. [10] A high-

activity

192

Ir source was selected for this purpose. The

dosimetric characteristics measured included the photon

ﬂuence anisotropy in air, transverse dose proﬁles in planes

parallel to the plane containing the HDR source, and iso-

dose distributions. The “in-air” anisotropy of the photon

ﬂuence relative to seed orientation was measured at 5 and

10 cm from the source.

A type 60003 diamond detector manufactured by

PTW Freiburg was used by Rustgi. The radiation-sensitive

region of the diamond detector had a volume of 1.8 mm

The low-impurity diamond crystal of unknown shape had

a thickness of a 0.26 mm. The crystal was enclosed in a

plastic housing, with the surface of the crystal displaced

1 mm below the housing front surface. All measurements

were made with an external bias of 100 V, as recom-

mended by the manufacturer, to achieve near-total ion-pair

collection.

The EGS4 Monte Carlo code has been used by Mobit

and Sandison [11] to investigate the response of a PTW/

diamond detector irradiated in both clinical and monoener-

getic megavoltage electron beams ranging in energy from

5 to 20 MeV. The sensitive volume of the PTW/diamond

detector simulated has a thickness of 0.4 mm and a diam-

eter of 4.4 mm. The results show that the PTW/diamond

detector has a constant response (within 1.0%) in electron

beams if irradiated at depths closed to d

max

, and its

response is almost independent of irradiation depth or

incident electron energy (within 3%). The encapsulation

of the bare diamond detector with low-Z epoxy and poly-

styrene wall material does not affect its response in elec-

tron beams.

Figure 8.16 shows the schematic of the simulated

PTW/diamond detector (model T60003) designed to mea-

sure relative dose distributions in high-energy photon and

electron beams. The sensitive volume of the PTW/dia-

mond detector is a disk made from natural diamond (den-

sity 3.51 g/cm

) with a radius ranging from 1.0 to 2.2 mm

and a thickness ranging from 0.2 to 0.4 mm.

When the PTW/diamond detector is irradiated in a water

phantom at given point p, the absorbed dose to the water

phantom at this point D

(p), following the Spencer-Attix

cavity equation and the AAPM-TG21 protocol, is as fol-

lows:

(8.5)

where is the average absorbed dose in the sensitive

volume, of the PTW/diamond detector, is the Spencer-

Attrix mass collision stopping power ratio of water to

diamond (graphite), is the perturbation correction

factor due to the non-water equivalence of the encapsu-

lating wall material of the PTW/diamond detector, and

is the perturbation correction factor resulting from

the insertion of the diamond-sensitive volume into the water

medium, which can be considered as the perturbation due

to the departure of the wall-less (bare) detector from non-

ideal Bragg-Gray (Spencer-Attix) conditions. Here,

in the TG21 formulation is regarded as the product of two

perturbation correction factors, the electron ﬂuence per-

turbation correction factor and the dose gradient pertur-

bation correction factor.

Since the sensitive volume of the PTW/diamond

detector is very small (6.1 mm

), one can assume that the

FIGURE 8.15 Response of diamond detector at photon energies of 4, 6, 15, and 25 MV ( , right axis). For comparison, the average

mass stopping power ratio of water/carbon and of water/air (left axis) is drawn in. (From Reference [9]. With permission.)



p() D

PTW

w,dia

 P

cav

PTW

 P

wall

PTW

 N

PTW



PTW

w,dia

wall

PTW

cav

PTW

cav

PTW

Ch-08.fm Page 370 Friday, November 10, 2000 12:03 PM

Solid-State Dosimeters 371

absorbed dose in the water phantom at point p is equal to

the absorbed dose to the water phantom averaged over the

same volume as the cavity. This means D

(p)  , and

Equation (8.5) can be written as

(8.6)

where is the average absorbed dose in water of the

same volume, as the wall-less diamond detector placed at

the same point p is the uniform water phantom.

The PTW/diamond detector response depends on

the average absorbed dose ratio of water to diamond

, and this ratio can be used to determine if the

PTW/diamond detector response depends on the depth of

irradiation or the incident electron beam energy. Table 8.2

shows the average absorbed dose ratio of water to the

sensitive volume of the PTW/diamond detector (

) for incident monoenergetic electron beams.

For wall-less detectors like TLDs, the deviation of the

absorbed dose at a point p in a uniform medium, from the

Spencer-Attix cavity equation, can be expressed as:

(8.7)

where is the perturbation correction factor,

is the average absorbed dose ratio of the

medium to cavity, and is the Spencer-Attix mass

collision stopping power ratio. However, for an encapsu-

lated diamond detector like the PTW/diamond detector,

the perturbation effects are from two different sources: (1) the

encapsulation and (2) the perturbation due to the sensitive

volume of the detector (air for ion chambers and dia-

mond for the PTW/diamond dosimeter) not acting as a

perfect Bragg-Gray cavity [Equation (8.7)]. The effect of

the encapsulation, denoted by , can be determined by

comparing the PTW/diamond response in a water phan-

tom to the response of a nonencapsulated diamond detec-

tor of identical sensitive volume (identical dimensions)

irradiated at the same depth with the same incident elec-

tron energies. [11]

TABLE 8.2

Monte Carlo–Determined Variation with Energy of the Average

Absorbed Dose Ratio of Water to the Sensitive Volume of the

PTW/Diamond Detector in Monoenergetic Electron Beams

(MeV) Depth (cm) E

(MeV)

5 1.2 2.2 1.145  0.002

10 2.7 4.1 1.146  0.002

15 3.6 6.7 1.149  0.002

20 5.4 7.0 1.148  0.002

The electron beam size is a square of dimension 10 cm at the phantom surface, and

the incident electron beam is assumed to be parallel and nondivergent. The mean

energy at depth (E

) was calculated following the IAEA method.

From Reference [11]. With permission.

FIGURE 8.16 Composition of the PTW/diamond detector (the simulated geometry).(From Reference [11]. With permission.)

PTW

D

PTW

w,dia

cav

PTW

 P

wall

PTW



D

PTW

()



PTW

med,cav

med

cav



med,cav

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



med,cav

med

D

cav

()

med,cav

wall

PTW

Ch-08.fm Page 371 Friday, November 10, 2000 12:03 PM

372 Radiation Dosimetry: Instrumentation and Methods

II. MOSFET RADIATION DOSIMETER

MOSFET is a sandwich-type device consisting of a p-type

silicon semiconductor substrate separated from a metal

gate by an insulating oxide layer. The advantages of MOS-

FET devices include being a direct reading detector with

thin active area (25



m) and having a small size. The

signal can be permanently stored and is dose-rate inde-

pendent. Dose enhancement was observed for photon

energies below 40 keV, where the photoelectric effect is

dominant. With adequate ﬁltration, uniform response was

reported for photon energies above 80 keV. The sensitivity

and linearity of MOSFET devices is greatly inﬂuenced by

the fabrication process and the voltage-controlled opera-

tion during and after irradiation.

When a metal-oxide-semiconductor (MOS) device is

irradiated, three mechanisms within the silicon dioxide

layer predominate: the build-up of trapped charge in the

oxide; the increase in the number of interface traps; and

the increase in the number of bulk oxide traps. Electron-

hole pairs are generated within the silicon dioxide by the

incident-ionizing radiation. Electrons, whose mobility in

SiO

at room temperature is about 4 orders of magnitude

greater than holes, quickly move toward the positively

biased contacts. Depending on the applied ﬁeld and the

energy and kind of incident particle, some fraction of

electrons and holes will recombine. The holes that escape

initial recombination are relatively immobile and remain

behind, near their point of generation.

A negative bias applied to the FET gate causes a

positive charge to build-up in the silicon substrate. This

build-up of charge allows current to pass through the silicon

substrate from the source to the drain terminals. The gate

voltage necessary to allow conduction through the MOS-

FET is known as the threshold voltage.

When the MOSFET is exposed to ionizing, radiation,

electron hole pairs are formed in the oxide insulation layer.

The junction potential between the device layers, or an

applied positive potential to the gate, causes the electrons

to travel to the gate while the holes migrate to the oxide

silicon interface, where they are trapped. These trapped

positive charges cause a shift in the threshold voltage,

since a larger negative voltage must be applied to the gate

to overcome the electric ﬁeld of the trapped charges to

achieve conduction. The threshold voltage shift is propor-

tional to the radiation dose deposited in the oxide layer,

and this relationship is the basis for using MOSFETs as

dosimeters. Irradiated MOSFETs have been stored under

zero bias condition for over ten years with approximately

1% loss of signal.

The efﬁciency of charge trapping or effective sensi-

tivity of the MOSFET depends on the thickness of the

oxide layer and the bias potential applied across the

layer during irradiation. For a given oxide layer thick-

ness, the sensitivity can be controlled by the bias potential

applied between the gate and substrate during irradiation,

since large electric ﬁelds applied across the oxide layer

result in more rapid separation of the electrons from the

holes.

Positive-ion diffusion toward the oxide-silicon inter-

face results in a negative shift of the threshold voltage,

while phonon-induced (thermal) release of trapped

charges, as well as charge annihilation by electrons tun-

neling from the silicon to the oxide layer (nonthermal),

results in annealing or a positive shift of threshold voltage.

These signal drifts become particularly apparent during

low dose rate, long-time irradiations. In order to obtain

meaningful dosimetric information from a MOSFET

dosimeter under low dose rate conditions, the drift effects

must be characterized and deconvoluted from the thresh-

old voltage data.

Prototype miniature dosimeter probes have been de-

signed, built, and characterized by Gladstone et al. [12],

employing a small, radiation-sensitive metal oxide semicon-

ductor ﬁeld effect transistor (MOSFET) chip. It was used to

measure, in vivo, the total accumulated dose and dose rate

as a function of time after internal administration of long-

range beta particle radiolabeled antibodies and in external

high-energy photon and electron beams. The MOSFET

detector is mounted on a long narrow alumina substrate

to facilitate electrical connection. The basic MOSFET

probe design is shown in Figure. 8.17.

A plot of the sensitivity vs. bias voltage is shown in

Figure 8.18. A line is drawn through the data points, show-

ing that the sensitivity of the MOSFET detector varies

logarithmically with bias potential between 1.5 and 9 V.

The sensitivity does not change as a function of total

absorbed dose from 1 to 1000 cGy or dose rate from

0.195 cGy/h to 400 cGy/min.

Figure 8.19 is a plot of the threshold voltage vs. tem-

perature for our MOSFET probe. The threshold voltage

for this MOSFET type is linear with temperature, with a

slope of 7.4 mV/°C. This linear temperature shift is easily

characterized and corrected for by adding an offset to

account for the temperature of the MOSFET at the time the

threshold voltage is measured. If pre- and post-irradiation

measurements are made at the same temperature, no cor-

rection is needed.

The drift rate of the threshold voltage is observed

to be linear with ln (

t/t

) (t is time after irradiation and

is normalization time of 1 min) at times greater than

150 min, with the rate of drift increasing with increasing

temperature.

A direct-reading semiconductor dosimeter has been

investigated by Soubra et al. [13] as a radiation detector for

photon and electron therapy beams of various energies. The

operation of this device is based on the measurement of the

threshold voltage shift in a custom-built metal oxide–silicon

semiconductor ﬁeld effect transistor (MOSFET). This volt-

age is a linear function of absorbed dose. The extent of the

Ch-08.fm Page 372 Friday, November 10, 2000 12:03 PM

Solid-State Dosimeters 373

linearity region is dependent on the voltage-controlled oper-

ation during irradiation. Operating two MOSFETS at two

different biases simultaneously during irradiation will result

in sensitivity (V/Gy) reproducibility better than 3% over

a range in dose of 100 Gy and at a dose per fraction greater

than 20  10



Gy. The modes of operation give this device

many advantages, such as continuous monitoring during

irradiation, immediate reading, and permanent storage of

total dose after irradiation.

The basic MOSFET structure is depicted in Figure 8.20.

The type shown is known as a p channel enhancement

MOSFET, which is built on a negatively doped (n-type)

silicon substrate. Two terminals of the MOSFET, called

the source (s) and the drain (d) are situated on top of a

positively doped (p-type) silicon region. The third termi-

nal shown is the gate (g). Underneath the gate is an insu-

lating silicon dioxide layer, and underneath this oxide

layer is the n silicon substrate. The region of the substrate

immediately below the oxide layer is known as the channel

region. When a sufﬁciently negative V

(gate voltage) is

applied (with reference to the substrate), a signiﬁcant num-

ber of minority carriers (holes in this case) will be attracted

to the oxide-silicon surface from both the bulk of the silicon

and the source and drain regions. Once a sufﬁcient number

of holes have accumulated there, a conduction channel is

formed, allowing an appreciable amount of current to ﬂow

between the source and the drain (I

). Figure 8.21 shows

how the magnitude of current varies with gate voltage.

FIGURE 8.17 Miniature MOSFET radiation detector probe design. The MOSFET chip is mounted on an alumina substrate. Gold

wire is used to electronically connect the MOSFET to the leads of the substrate. A cable is connected to the end of the substrate not

shown. The probe is inserted into a plastic ﬂexineedle for protection before insertion into tissue. (From Reference [12]. With

permission.)

FIGURE 8.18 Sensitivity vs. bias voltage. The sensitivity

increases with increasing bias voltage.(From Reference [12]. With

permission.)

FIGURE 8.19 Threshold voltage vs. temperature. The magni-

tude of the threshold voltage decreases with increasing temper-

ature. (From Reference. [12] With permission.)

Ch-08.fm Page 373 Friday, November 10, 2000 12:03 PM

374 Radiation Dosimetry: Instrumentation and Methods

Threshold voltage (V

) is deﬁned as the gate voltage

needed to attain a predetermined current ﬂow (I

Typically commercial MOS transistors have 200- to

800-Å-thick gate oxides. To maximize the MOSFET sen-

sitivity to ionizing radiation, a relatively thick gate oxide

is preferred. However, there is a limitation to how thick a

gate oxide can be grown. Growth of extremely thick oxide

layers (l



m) requires growth times of more than 100 h

at temperatures of 1000°C. These oxidation times are not

feasible and, in addition, thick oxide will tend to be

highly stressed, leading to dislocations at the Si-SiO

interface, resulting in fast surface states and threshold

instability. [13]

The response of the single MOSFET detector (V

) as

a function of accumulated dose will exhibit a nonlinear

region at high dose levels. The main reason for this phenom-

enon is the accumulation of the radiation-generated positive

charge (holes) in the oxide traps, which effectively reduces

the magnitude of the electric ﬁeld produced by the positive

gate bias between the positive-charge region and the gate.

The MOSFET is approximately 200



m in diameter

and consists of a 0.5-



m Al electrode on top of a l-



SiO

and 300-



m Si substrate. Results for percentage

surface dose measured by Butson et al. [14] were within

2% compared to the Attix chamber and within 3% of

TLD extrapolation results for normally incident beams.

Percentage surface dose for 10  10-cm and 40  40-cm

ﬁeld size for 6-MV x-ray at 100-cm SSD using the MOS-

FET was 16% and 42% of maximum, respectively. Factors

such as its small size, immediate retrieval of results, high

accuracy attainable from low applied doses, and recording

of its dose history make the MOSFET a suitable in vivo

dosimeter where surface and skin doses need to be deter-

mined.

Dose in the build-up region varies considerably with

changes in parameters such as beam energy, SSD, patient

geometry, angle of incidence, ﬁeld size, and the use of

blocks, block trays, and wedges and boluses. Dose depos-

ited at the surface and in the build-up region has two main

contributors:

1. In-phantom scatter, where x-rays are incident on

the phantom and interact with electrons via

mainly the Compton effect and, to a lesser extent,

photoelectric and pair production at 6-MV x-ray

energy, which, in turn, deposit their energy along

their path; and

2. Electron contamination from electrons pro-

duced outside the phantom and incident on the

phantom to deposit their dose. These electrons

are produced by x-rays interacting with material

in and on the machine head, such as ﬂattening

ﬁlters, ion chambers, collimators, block trays

and wedges, plus the air column between the

source and phantom. These electrons will be

referred to as electron contamination.

Skin dose measurements represent a greater technical chal-

lenge because of the small distances to superﬁcial layers;

hence, any suitable detector should have minimal build-

up due to encapsulation.

The experimental conﬁguration for the MOSFET con-

sisted of a scientiﬁc grade (stock item) n-type MOSFET

device from which Butson et al. have removed the nickel

casing, leaving the bare crystal intact and exposed, as

shown in Figure 8.22. This was easily achieved using

pliers to ease off the cap that was connected at three

positions around its base to the device’s main body. We

FIGURE 8.20 Schematic cross section of a p-channel MOSFET

showing the oxide (SiO

), the substrate (Si), the source, the gate,

and the drain. (From Reference [13]. With permission.)

FIGURE 8.21 Typical MOSFET source to drain current I

vs.

gate voltage

. The difference in the gate voltages V

needed

to attain a predetermined current ﬂow

, before and after radi-

ation, is equal to the threshold voltage shift (

V

). (From Ref-

erence [13]. With permission.)

Ch-08.fm Page 374 Friday, November 10, 2000 12:03 PM

Solid-State Dosimeters 375

refer to this state as the unencapsulated MOSFET. During

irradiation, the MOSFET was connected to a 6-V bias

voltage supplied by a DC battery with a circuit, as shown

in Figure 8.23.

The unencapsulated MOSFET has a few advantages over

the TLD:

1. Immediate retrieval. Dose can be ascertained

accurately within minutes after irradiation with

as low as 10 cGy of applied dose.

2. MOSFETs can be used again immediately.

They do not require an anneal process after

irradiation; TLDs do require such a process.

3. Inexpensive support equipment. MOSFETs

require only a small power source and voltmeter

for data retrieval.

4. MOSFETs retain their dose history. Informa-

tion is not lost during the retrieval process.

All of the above properties and their accuracy make the

unencapsulated MOSFET an ideal in vivo dosimeter for

surface and skin dose measurements.

Rosenfeld et al. [15] was concerned with the role of the

package of MOSFETs used in measurements of gamma

dose in mixed gamma-neutron ﬁelds, in high energy

bremsstrahlung, and in soft x-ray ﬁelds. It is shown that

a kovar cap should be avoided for dosimetry applications

in the presence of strong thermal neutron ﬂuences. In

regions of strong electronic disequilibrium, the ‘‘bare’’ or

unencapsulated MOSFET is a unique tool for surface dose

measurements and Monte Carlo model validation. For

depths where electronic equilibrium exists (i.e.,

x  depth

of D

max

), the MOSFET package is not critical. For low-

energy X-ray ﬁelds, the energy dependence of the dose

enhancement factor (DEF) will differ, depending upon

whether the irradiation is performed in free-air geometry

or on the surface of a phantom.

The study of the neutron response of MOSFET

dosimeters was initiated primarily for the dosimetry of

mixed gamma-neutron ﬁelds. It has previously been

shown that such MOSFETs are much less sensitive

(~100 times) to fast neutron tissue dose than to gamma

tissue dose. The investigations have been carried out in

TO-5 and TO-18 kovar packages exposed to 3-MeV

and 15-MeV fast neutron ﬁelds and cadmium ﬁltered

FIGURE 8.22 Physical design of unencapsulated MOSFET. (From Reference [14]. With permission.)

FIGURE 8.23 MOSFET circuit diagram. (From Reference [14].

With permission.)

Ch-08.fm Page 375 Friday, November 10, 2000 12:03 PM

376 Radiation Dosimetry: Instrumentation and Methods

reactor beam spectra with an average neutron energy

of 1 MeV.

The MOSFET dosimeter is unique due to the small

size of its active dosimetric volume of SiO

. This makes

the MOSFET very attractive for dosimetry in regions of

electronic disequilibrium. Such ﬁelds occur in radiation

hardness testing on pulsed bremsstrahlung facilities (~10

Gy(Si)/s) and also on medical linear accelerators having

end energies in the range 6–25 MeV. Different approaches

are required for tissue dosimetry in soft x-ray spectra

compared to high-energy bremsstrahlung ﬁelds. The

appropriate MOSFET packaging will differ between these

cases.

The Ukrainian n-MOSFET with a 1-



m-thick oxide

layer and aluminum gate and encapsulated in a standard

kovar package has been investigated in the MOATA

research reactor thermal (TC-10) facility.

To study the role of different thermal neutron ﬂuences

on the threshold shift while maintaining the gamma and

fast neutron ﬂuences approximately constant, Rosenfeld

et al. surrounded the TO-18 packaged MOSFETs with

additional

LiF/epoxy encapsulation. Each of these lithi-

ated covers was cylindrical in shape and of uniform thick-

ness. To maintain constant gamma irradiation conditions,

MOSFETs were exposed at the same point in the irradi-

ation channel and for the same period of time but with

different thicknesses of

Li encapsulation. The change in

threshold voltage of the MOSFETs as a function of ther-

mal neutron ﬂuence is shown in Figure 8.24.

The model of the MOSFET follows the geometry

shown in Figure 8.25. The gold-plated leads and kovar base

were represented as a homogeneous mixture of gold and

nickel. Similarly, the silicon oxide layer and aluminum gate

electrode were considered a homogeneous mixture of sil-

icon oxide and aluminum. These approximations are standard

techniques in MCNP modeling. The thickness of the SiO

layer was increased from its real value of 1



m to 450



in the model to get better statistics.

Figure 8.26 shows the role of MOSFET packaging on

the relative dose depth distribution in the phantom build-

up region for 6-MeV bremsstrahlung. The MOSFET

encapsulated in a TO-18 package showed a strong dose

enhancement phenomenon in the region of strong elec-

tronic disequilibrium. At the interface between the air and

the solid water, the encapsulated MOSFET yielded an

overestimation of the dose by four times relative to Attix

chamber measurements.

FIGURE 8.24 Change in MOSFET threshold voltage as a function of thermal neutron ﬂuence (V

 0 V, passive mode). (From

Reference [15]. With permission.)

FIGURE 8.25 Schematic diagram of the geometry used in the

Monte Carlo model of the MOSFET. (From Reference [15]. With

permission.)

Ch-08.fm Page 376 Friday, November 10, 2000 12:03 PM

Solid-State Dosimeters 377

The semi-encapsulated MOSFET showed a 2.5 times

overestimation of dose for the same comparison. The ‘‘bare’’

MOSFET, however, was in agreement with the Attix chamber

to more than 1% for the entire 15-mm build-up region.

Figure 8.27 illustrates the DEF for a thick-oxide

n-MOSFET in three different packages. The energy axis

shows the end energy of the x-ray spectrum, and the DEF

is normalized to the sensitivity observed for an end energy

of 6 MeV. All irradiations were performed with the beam

normally incident on the top surface of the MOSFET can

or die.

MOSFET integral dosimetry based on charge build-up

in the SiO

layer, when exposed to radiation was studied

by Rosenfeld et al. [16]. Microdosimetry using silicon

micro-volumes based on spectroscopy of charge in a p-n

junction, is a further step in the characterization of mixed

radiation environments. For that characterization LET

must be measured. That is the most serious problem

encountered in the data interpretation, i.e., the determina-

tion of the average chord length of the sensitive volume.

This quantity is required for the conversion of pulse height

spectrum to LET spectrum.

The n-MOSFET chip (p-Si substrate, doping level

1.35  10

3

, drain surrounded by 60-



m length gate,

drain-active area 1.5  10

2

) originally was used in

this study in a TO-18 package from which the cover was

removed. Either

210

Po or

241

Am α source was placed inside

the vacuum chamber, with the MOSFET chip located

immediately above it. The measurements were made in two

modes. In the ﬁrst mode (Figure 8.28a), the source and gate

FIGURE 8.26 Depth dose distributions measured using a MOSFET in the build-up region of a solid water phantom irradiated with

a 10

 10-cm

ﬁeld from a 6-MeV bremsstrahlung source. (From Reference [15]. With permission.)

FIGURE 8.27 Dose enhancement factors for four differently packaged MOSFETs in free-air geometry. (From Reference [15]. With

permission.)

Ch-08.fm Page 377 Friday, November 10, 2000 12:03 PM

378 Radiation Dosimetry: Instrumentation and Methods

were grounded and a charge sensitive preampliﬁer (PA)

was connected to the n



-p drain junction reverse biased

at 18 V. In the second mode (Figure 8.28b), both the source

and the drain were connected to the PA and the gate was

grounded.

Dose response of various radiation detectors to syn-

chrotron radiation was studied by Korn et al. [17] Monoen-

ergetic x-rays from synchrotron radiation offer the unique

opportunity to study the dose response variation with pho-

ton energy of radiation detectors without the compounding

effect of the spectral distribution of x-rays from conven-

tional sources. The variation of dose response with photon

energies between 10 and 99.6 keV was studied for two TLD

materials (LiF:Mg,Ti and LiF:Mg,Cu,P), MOSFET semi-

conductors, radiographic ﬁlm, and radiochromic ﬁlm.

MOSFET detectors and the radiographic ﬁlm were found

to over-respond to low-energy x-rays by a factor of up to 7

and 12, respectively, while the radiochromic ﬁlm underes-

timated the dose by approximately a factor of 2 at 24 keV.

The TLDs showed a slight over-response with LiF:Mg,

Cu,P, demonstrating better tissue equivalence than LiF:Mg,Ti

(maximum deviation from water less than 25%).

The increasing recombination effects with decreasing

photon energy result in a reduced sensitivity. Other con-

tributing factors are the predominance of the photoelectric

effect in SiO

in comparison with tissue for photon ener-

gies less than 100 keV, a dose enhancement effect due to

the metal-SiO

interfaces used in the MOSFET devices,

and the packaging of the chip.

The MOSFET dosimeter system used in Kron’s study

was developed by the radiation physics group at the Univer-

sity of Wollongong (Australia). The n-MOSFET detectors,

which were obtained from the Ukraine, feature a 1-



m-thick

silicon oxide layer with an aluminum gate electrode of

0.5-



m thickness. For the experiments, the detector was

placed in a solid water plate of approximately 20  40 mm

and a thickness of 4 mm.

The gate of the MOSFETs was biased during irradi-

ation with 5 V. They were evaluated using a battery-operated

portable reader which was designed for measurements of

the threshold voltage in the range of 10 V to 10 V.

The stability of the threshold voltage assessment was within

2 mV.

Figure 8.29 shows the reduction of dose response in the

Wollongong MOSFET, A, as a function of total gate voltage.

FIGURE 8.28 Connections for pulse-height measurement using the MOSFET (a) with only the drain connected to the PA and

(b) with source and drain both connected to the PA. (From Reference [16]. With permission.)

FIGURE 8.29 Variation of response with total absorbed dose

in MOSFET detector,

A, exposed to 4-MV photons from a med-

ical linear accelerator (Varian Clinac 600 C/4 MV). The change

required in gate voltage to compensate for the charge accumu-

lated due to 0.5 Gy absorbed dose was used as dose response.

It is shown as a function of total absorbed dose and the actual

MOSFET total gate voltage required to maintain a constant

current between source and drain. The bias applied during irra-

diation was 5 V. The broken line shows the best linear ﬁt to the

experimental data (

 0.877). (From Reference [17]. With

permission.)

Ch-08.fm Page 378 Friday, November 10, 2000 12:03 PM

Solid-State Dosimeters 379

The total gate voltage is related to the total absorbed dose

that is also indicated on the horizontal axis. The reduction

of dose response, expressed as gate voltage change per

absorbed dose, is a linear function of the total absorbed

dose. The detector response is reduced by approximately

3% for a gate voltage change of 1 V. This function was

used to correct the readings of the other Wollongong

MOSFET detectors.

Figure 8.30 shows the variation of dose response with

photon energy relative to water for MOSFET detectors.

The two Wollongong-type detectors exhibit a maximum

of dose response relative to water around 35 keV. The

difference between the two detectors in the absolute dose

maximum of 7.1 and 6.5, respectively, approaches statis-

tical signiﬁcance.

For comparison, the results obtained in spectral radia-

tion beams with the Thomson and Nielson [17] MOSFET

detectors are also plotted in Figure 8.30. The reduction of

dose response with total absorbed dose was found to be

less than 1% for a gate voltage change of 1 V, which is

considerably smaller than the one observed in the Wol-

longong MOSFETs. The dose response for different radi-

ation energies was corrected for this variation.

As can be seen in Figure 8.30, the Thomson and

Nielson detectors also exhibit a smaller variation of dose

response with photon energy, which is most likely due to

the use of a 2-mm resin on the detector. The over-response

was found to be on the order of a factor of 3, just above

an effective photon energy of 50 keV.

The sensitive element of a MOSFET detector is a

silicon oxide layer underneath the aluminum transistor

gate. The geometry of the sensing regions of the silicon

dioxide is slightly different for the two MOSFET devices

used by Rosenfeld et al. [18] In the case of the TOT500

p-MOSFET device, the gate oxide has a serpentine shape,

and in the case of the n-MOSFET device it was a rectan-

gular ribbon shape with an outer boundary dimension of

about 200  200



and a thickness of about 1



m for

both detectors.

The change in MOSFET threshold voltage, which is

proportional to the absorbed radiation dose, was measured

under constant current conditions. The n-channel MOSFET

chip was mounted inside a kovar package with the lid

removed. RADFET has a different topology than the

n-channel MOSFET and was mounted on a plastic board

under an epoxy bubble. Both MOSFET and RADFET

detectors were irradiated with a gale bias voltage of 5 V

to increase sensitivity and improve linearity of response.

The

n-MOSFET dosimeters were exposed in air, i.e.,

without any build-up material, to the microbeam in ‘‘normal’’

and ‘‘edge-on’’ orientations, and the p-MOSFET dosimeter

was placed in ‘‘edge-on’’ orientation only. The dose deliv-

ered to MOSFET per single irradiation shot in free-air

geometry was 17 monitor units (MU), which corresponds

to a maximum threshold voltage shift in the center of the

beam of about 150 mV. The results (Figure 8.31) prove that

the spatial resolution of a MOSFET dosimeter is indeed

affected by the orientation of the detector in the beam, the

resolution being superior for the ‘‘edge-on’’ mode. The

results demonstrate that the responses of

n-MOSFET and

RADFET in ‘‘edge-on’’ mode are almost independent of

package geometry and material for irradiation in free-air

geometry.

Electron ﬂuence was calculated by Rosenfeld et al.

[19] from the dose measured in the water at a distance of

1 m from the virtual source. The ﬂuence at the point of

irradiation was calculated using the inverse-square law.

The electron ﬂuence and dose in Si was measured inde-

pendently using the Hamamatsu photodiode at the point

of irradiation from the relation

(8.8)

where,

 photodiode current

 3.62 eV

 irradiation time

e  1.6  10

19



 density of Si

 sensitive volume of the photodiode (1  1 

0.3 cm

A LINAC electron beam was used for monitoring of

the Hamamatsu photodiode response. The degradation of

FIGURE 8.30 Variation of the dose response with effective

radiation energy for MOSFET detector. The response is normal-

ized to 6-MV spectral x-rays. (From Reference [17]. With per-

mission.)

DSi()



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



Ch-08.fm Page 379 Friday, November 10, 2000 12:03 PM