PORTABLE DOSIMETER WITH MOSFET SENSOR FOR

RADIOTHERAPY MONITORING

M. A. Carvajal, J. Banqueri, A. J. Palma

Departament of Electronics and Computer Technology, University of Granada, ETSIIT, CITIC-UGR, Granada, Spain

A. M. Lallena

Department of Molecular, Nuclear and Atomic Physics, University of Granda, Granada, Spain

D. Guirado, M. Vilches

Department of Radiophysics, University Hospital San Cecilio, Granada, Spain

Keywords: Radiation sensor, Dosimeter, MOSFET, Radiotherapy, Portable instrument.

Abstract: A portable dosimeter based on unbiased MOSFET sensor is presented. Its main characteristics are an

extended linearity range and a notable thermal drift reduction using a unique p-channel MOSFET (pMOS)

unbiased during the irradiation period, allowing the location of the sensor without wires for patient comfort

and easy-to-use. Both features have been obtained with novel procedures of dose reading and signal

processing applied to low-cost commercial pMOS. In this work, a full description of the electronics of the

dosimetry system and the signal processing techniques are drawn. The system has been tested with photons

from

Co and the complete technical specifications have been obtained. Among them, we can emphasize: i)

dose sensitivity of around 25 mV/Gy; ii) linearity range of more than 50 Gy, with intermediate calibration

each 15 Gy, for each sensor; iii) thermal drift below 3 mGy/ºC; iv) resolution below 1cGy; and v) total

uncertainty of ± 9 mGy in the temperature range from 19 ºC to 36 ºC . We believe that the proposed

dosimeter could be a novel and feasible low-cost alternative to previous commercial dosimetry systems for

radiotherapy monitoring in clinical applications.

1 INTRODUCTION

In the last few decades, MOSFETs (Metal-Oxide-

Semiconductor Field Effect Transistors) have been

used widely as dosimeters due to their small size,

immediate and non-destructive readout, low power

consumption, easy calibration, and reasonable

sensitivity and reproducibility (Holmes-Siedle and

Adams, 1986; Hughes et al., 1988; Soubra et al.,

1994). These desirable characteristics have

promoted the use of this device in different fields

such as spacecraft, radiotherapy, skin dosimetry, and

clinical monitoring (Buehler et al., 1993;

Sarrabayrouse and Siskos, 1998; Benson et al.,

2000; Rosenfeld, 2002; Bloemen et al., 2003; Haran

et al., 2004; Best et al., 2005; Kwan et al., 2008).

Ionizing radiation produces electron-hole pairs in

the MOS structure that are separated by an electric

field established in the gate oxide and near the

silicon-gate oxide interface. During irradiation, the

structure can be configured without an external

voltage (unbiased mode), or the electric field can be

made larger by applying an external voltage between

the gate and the bulk of the device (biased mode).

The charge accumulated in the structure produces

changes in the electrical parameters that can be

measured experimentally, usually after irradiation,

during the so-called readout period. The electrical

parameter of the MOS transistors most commonly

used as a dosimetric parameter (i.e., the parameter

from which the dose is extracted) is the threshold

voltage (V

), which is the gate-source voltage

necessary to induce channel inversion between the

source and drain terminals. Many works show an

approximately linear dependence between the V

shift and the absorbed dose in the oxide for p-

Carvajal M., Banqueri J., Palma A., Lallena A., Vilches M. and Guirado D..

PORTABLE DOSIMETER WITH MOSFET SENSOR FOR RADIOTHERAPY MONITORING.

DOI: 10.5220/0003128100230029

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2011), pages 23-29

ISBN: 978-989-8425-37-9

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

channel MOSFETs (pMOS) (Sarrabayrouse and

Siskos, 1998; Ma and Dressendorfer, 1989; Hughes

et al., 1988).

During readout, the V

determination can be

carried out by extracting the complete current-

voltage characteristic curve of the device (Asensio et

al., 2006) or by using simpler methods based on

constant current measurements. Basically, the

measurement methods consist in recording the drain-

source voltage while the transistor is being polarized

with a constant drain current, before and after

irradiation. Under this configuration, the source-

drain voltage shifts equals, approximately, the

increase in V

(Sarrabayrouse and Siskos, 1998;

Asensio et al., 2006; Carvajal et al., 2010).

Most commercial dosimetric systems based on

MOSFETs, measure increases in drain-source

voltage at constant drain current

(Thomson&Nielsen, 1991; Halvorsen, 2005).

Usually, in order to minimize thermal drift, the drain

current selected is the Zero Temperature Coefficient,

ZTC

, because the drain-source voltage is invariant

when temperature increases. The dosimetric

parameter used by one of the commercial dosimetric

systems is the difference of the increase in drain-

source voltages of two transistors under different

polarizations. With this arrangement, a wide linear

range and thermal compensation are achieved in the

biased mode (Thomson&Nielsen, 1991). There is

another commercial system based on transistors in

the unbiased mode, but the linear range is up to 5 Gy

and its use is limited to only one irradiation session

(Halvorsen 2005; Sicel Technologies, 2005, Best et

al., 2005).

In previous works (Asensio et al. 2006, Carvajal

et al., 2010; Carvajal et al., 2010b), we showed the

feasibility of using a general-purpose low-cost

pMOS as dosimeter, irradiated in unbiased mode,

with a significant increase in linear range compared

with similar systems and reduced thermal drift.

These improvements are based on gate-source

voltage measurements at three different drain

currents during readout instead of the usual method

with a single polarization current.

Although this work has focussed on the pMOS

dosimeter, it is important to point out that the

proposed methodology could be applied to other

sensors based on MOSFET, such as CHEMFET or

ISFET, where the magnitude under study is

extracted from the threshold voltage. In these cases

(Fraden, 1996), the threshold voltage is also

measured at constant drain current.

In the following section, a complete

measurement system description is carried out,

including the measurement algoritmh and an

explanation of the electronic system. After this, the

experimental setup used in our experiments is

described. We report experimental results of the

dosimetric characterization of the dosimeter

presented here and its technical specifications are

shown. Finally, the main conclusions are drawn.

2 DOSIMETER DESCRIPTION

In this section, a detailed description of the

measurement algorithm and the electronic design of

the dosimeter will be carried out.

2.1 Measurement Algorithm

First, the theoretical background of the measurement

process, developed in previous works, will be

resumed (Carvajal et al., 2010; Carvajal et al.,

2010b). This is the basis on which the electronic

design relies on. The operation method of the dose

measuring of the presented dosimeter is the same of

most previous devices based on MOSFET sensors.

Before the sensor irradiation, an electronic reader

unit measures the electrical characteristics of the

sensor; this is usually named as sensor zeroing (first

readout period). Then, sensor is irradiated

(irradiation period) and, finally, another readout

process is carried out to measure the MOSFET

electrical characteristics shifts caused by the

irradiation. Here during the irradiation period, we

have configured the sensor in the unbiased mode, i.e.

short-circuiting all its terminals and without

polarization, obtaining a sensor without wires in this

period.

During the dose readout under constant drain

current bias, the gate and the drain terminals of the

MOSFET are usually short-circuited and grounded,

and the bulk and the source are also inter-connected

(Sarrabayrouse and Siskos, 1998; Best et al., 2005).

In this configuration, the transistor operates in the

saturation region, where

VI − can be modelled

for pMOS (Sze, 1981):

()

TSD

I −=

(1)

where the drain current,

I , depends on the source

voltage,

V , the threshold voltage,

V , and the

parameter,

oxp

μβ

, where

is the carrier

mobility in the channel,

C is the oxide

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

capacitance, and W and L are the channel width and

length respectively.

From Equation (1), the threshold voltage can be

calculated before (pre) and after (post) irradiation:

pre

IVV

2−=

(2a)

post

IVV

2−=

(2b)

Therefore, the voltage threshold shift can be

calculated as follows:

⎟

⎠

⎞

⎜

⎝

⎛

−−Δ=Δ

prepost

DSt

IVV

ββ

(3)

If the increase in the source voltage is measured

at two different currents, we can write:

⎟

⎠

⎞

⎜

⎝

⎛

−−Δ=Δ

prepost

DSt

IVV

ββ

(4a)

and

⎟

⎠

⎞

⎜

⎝

⎛

−−Δ=Δ

prepost

DSt

IVV

ββ

(4b)

where

VΔ is the source voltage shift measured at

constant current

I . Using (4a) and (4b), the

threshold voltage shift can be calculated as (Carvajal

et al, 2010):

−

Δ−Δ

+Δ=Δ

(5)

With Equation (5), the threshold voltage shifts can

be calculated exactly from the source voltage shifts

instead of approximated to the source voltage, as has

been done until now.

According to equation (5), threshold voltage

shifts can be found from voltage increases measured

with two different currents, separating the

contribution to the source-voltage shifts due to

and

VΔ and resulting in an extended linear range

for the sensor (Carvajal et al, 2010). Moreover, if a

third current is applied, a thermal-compensation

method and the extension of the linear range can be

used simultaneously. Two currents are needed to

evaluate equation (5),

I and

I , and an

additional current,

I , is required for the thermal

compensation of the source voltage shifts measured

I and

I (Carvajal et al., 2010b):

()

ZTCD

SSCSS

ZTCD

SSCSS

VVVV

−

Δ−Δ+Δ=Δ

−

Δ−Δ+Δ=Δ

(6)

where I

ZTC

is the current with minimal thermal

coefficient of the threshold voltage. According to

equation (5), we have:

−

Δ−Δ

+Δ=Δ

(7)

Evaluating equations (6) and (7) sequentially

allows both the thermal compensation and the

linearity improvement methods to be taken into

account, and provides the thermally compensated

threshold voltage shifts from the source-voltage

increments measured at three different currents. One

need only know the value of

ZTC

I and the intensities

used to polarize the transistor during readout. If one

of the currents,

I or

I , is equal to

ZTC

I ,

equation (7) is simplified. For example,

ZTCD

, the source voltage shifts at

I are

thermally compensated, and all that is left is to

evaluate equation (7) for

and

VΔ in order to

find

VΔ

Moreover, polarizing with pulsed instead of

constant currents, most of the 1/f noise is avoided,

obtaining a higher signal-to noise ratio of the

measurement (Carvajal et al, 2010).

2.2 Measurement Electronic System

The dosimetric system consists of a read-out unit

and a set of sensor modules.

The sensor module is based on the commercial

general-purpose pMOS 3N163, which is located on

a printed circuit board (PCB). During irradiation and

storage periods, all the terminal of the MOSFET

must be short-circuited together. In previous works,

we used a power supply jack for this purpose, but

resulted in a sensor module too thick that could

produce radiation shadow areas. For minimizing this

effect in the radiation field, the jack was replaced by

an n-channel JFET (see Figure 1), with SMD case

(surface-mounted-device). We used the MMBF4391

of NXP Semiconductor (Netherland), with a cut off

PORTABLE DOSIMETER WITH MOSFET SENSOR FOR RADIOTHERAPY MONITORING

voltage of -10V and resistance of 30 Ω, and it was

welded in the bottom layer of the PCB. This

transistor is normally on and maintains connected

the source and bulk terminals with the drain and

gate. This connection must be opened for zeroing

and voltages shifts measurements. Therefore the

read-out unit has to provide the adequate source-gate

voltage for JFET cut off during the readout process.

When the sensor module is removed from the read-

out unit, the gate capacitance of the JFET is

discharge through a resistor, R

. Therefore, all the

MOSFET terminals are connected again.

pMOSFET

JFET

Figure 1: Sensor module schematic.

The readout unit is an electronic system

controlled by a microcontroller (Figure 3). We used

the PIC-16F877 of the mid-range family of

Microchip (USA) due mainly to its low power

consumption, its 10 bits analog-to-digital converter

(ADC), four independent input/outputs ports and

I2C and UART ports. An EEPROM memory is

included in order to storage different parameters of

each sensor module. One reader unit can support up

to 256 different sensor modules. Each sensor module

must be identified for zeroing and for dose

measurements.

Figure 2: Reader and sensor module.

Figure 3 shows the block diagram of the analog

circuitry and the microcontroller. The analog

subsystem consists mainly, in a programmable

current source, an instrumental amplifier and a

digital to analog converter (DAC). During the

measurements process of the sensor module, the

JFET is cut off, and the programmable source

polarizes the MOSFET at different currents, being

measured the source voltages by the read-out unit.

For zeroing, the microcontroller calculates the DAC

output voltage necessary to reduce the instrumental

amplifier output down to tens of milivolts. The

digitalized source voltages, and the DAC words are

storage in the EEPROM, mapped in EEPROM

depends on the sensor module identification number,

ID. At least, two minutes after the irradiation ends,

the dose measurement can be carried out. The sensor

ID is introduced in the keypad, and DAC words are

restored for each polarization currents. Then, the

source voltage source shift can be calculated. The

calibration parameters are read from the EEPROM,

and the dose are calculated. Finally the results are

shown on the display or can be downloaded to the

computer.

Programable

current

source

Timer

555

pMOSFET

BJT

Microcontroller

aconditionating

circuit

A.I. A/D

DAC

A/D

Figure 3: Block diagram of the analog circuitry.

In addition, during the readout process, the

pMOS current is switched in order to reduce the 1/f

noise and to increase the linearity as Carvajal et al.

2010 shown in a laboratory setup. The collector of a

bipolar transistor, the BC547 of NXP

Semiconductors (Netherland), is connected to the

source of the MOSFET and the emitter is grounded.

When the bipolar transistor is activated, the current

drains by the BJT, forcing the source voltage of

pMOS to zero. But, when the BJT is deactivated, the

current crosses the pMOS transistor, and then the

source measurement are carried out. The pulse

signal was generated with a timer 555. The PWM

module of the microcontroller was not used because

the PIC was in sleep mode during A/D

measurements to reduce the noise produced by the

crystal oscillator of the PIC.

The readout unit can be configured by a 4x4

keypad or by computer via the USB port. The results

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

can be shown on the LCD screen and/or be sent to

the PC (Figure 2).

3 EXPERIMENTAL RESULTS

3.1 Experimental Setup

For the evaluation of the presented dosimetry

system, a total of eight sensor modules were

irradiated with an AECL Theratron 780 located at

the University Hospital of San Cecilio in Granada

(Spain). This is a teletherapy unit with a

Co source.

All the irradiations were carried out with all four

transistor terminals connected to each other, hence

without bias voltage, in the ‘unbiased’ mode. The

transistors were irradiated with a 40 x 40 cm

field

and were placed at the isocentre, at 80 cm from the

source. Five transistors were used to study the

linearity and measurement range, two for the long-

term fading and the last one for short-term fading.

According to a previous work, the three

polarization currents included in the dosimeter setup

and applied to the sensor during the readout periods

were I

= 30 µA, I

= I

ZTC

= 225 µA and I

= 120

µA

. With these current values a thermal drift below

3 mGy/ºC has been obtained (Carvajal et al, 2010b).

Moreover, for sensor equalization, an initial pre-

irradiation of 30 Gy was necessary.

3.2 Dosimetric Evaluation

In order to evaluate the dosimetric behaviour of the

dosimeter, the post-irradiation fading and the

linearity range. All these results have been obtained

with equations 6 and 7 included in the signal

processing.

3.2.1 Post-irradiation Fading

The reordering of the MOSFET charge after the

irradiation produces shifts in the dosimetric

parameter, i.e. the voltage threshold, which could

false the dose reading. This effect is inherent to the

transduction principle (measurement of the dose by

the charge creation) and very difficulty avoidable.

However, what you can do is to characterize it to

keep it under control. This has been done with our

dosimeter, measuring the short-term fading and the

log-term fading to give a time period when both are

negligible.

In Figure 4, the short-term post-irradiation

fading is shown for several accumulated dose. In this

figure, the threshold voltage transient (in arbitrary

units) is measured as a function of time.

260

280

300

320

340

360

380

400

0 50 100 150 200 250 300

t(s)

n_ADC

3Gy

9Gy

12Gy

15Gy

18Gy

21Gy

24Gy

27Gy

30Gy

Figure 4: Short-term post-irradiation fading.

From Figure 4, we can observe a measurable

transient in the first 60 s and that for long times it is

negligible even at higher accumulated dose. The

long-term post-irradiation fading was measured

every 12 hours, for two sensors with accumulated

dose of 27 Gy in 3 Gy sessions. The threshold

voltage long-term transients are shown in Figure 5.

y = 9.33E+01e

-1.62E-06x

= 9.97E-01

y = 8.90E+01e

-1.77E-06x

= 9.97E-01

100

0.0E+00 2.0E+05 4.0E+05 6.0E+05

t (s)

Vs (mV)

Figure 5: Long-term post-irradiation fading.

From Figure 5, a time constant of around one

week was obtained with an exponential fit, in

agreement with reported results (Ma and

Dressendorfer, 1989). Thus, this dosimeter obtains

dose measurements without fading interference at

times longer than 60 s and shorter than a few hours

after the sensor irradiation. Therefore, we have

stated a delay for the measurement between 2 and 3

minutes.

3.2.2 Linearity

One of the disadvantages of using MOSFET sensors

in the unbiased mode is the reduced linearity range

(Sarrabayrouse and Siskos, 1998). With the readout

procedure and the developed instrumentation, this

characteristic has been notably improved. Session

sensitivities (normalized voltage threshold shift,

mV/Gy, per irradiation) are depicted in Figure 6. In

Figure 7, the accumulate threshold voltage shifts as a

function of the dose up to 52 Gy are shown for five

PORTABLE DOSIMETER WITH MOSFET SENSOR FOR RADIOTHERAPY MONITORING

sensors, where the first 30 Gy have been subtracted

and used as pre-irradiation dose for sensor

equalization and to improve the sensitivity stability

0 20406080

D (Gy)

Sen (mV/Gy)

irradiations

Dose

measurements

Figure 6: Irradiation session sensitivities versus

accumulated dose.

y = 0.0233x

= 1.0000

y = 0.0240x

= 1.0000

y = 0.0240x

= 1.0000

y = 0.0243x

= 1.0000

y = 0.0232x

= 1.0000

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0 102030405060

D-Dpre (Gy)

(V)

Li l (5)

Figure 7: Accumulated threshold voltages shifts versus

accumulated dose.

From figures 6 and 7, several remarks can be

extracted:

i. A pre-irradiation of 30 Gy is necessary for

equalization.

ii. After pre-irradiation, a good linearity is observed

iii. Intervals of accumulated dose of 15 Gy produce

linearity errors below 5%

iv. User recalibration every 15 Gy are needed to sure

good accuracy for extended linear range up to 52 Gy

with a unique sensor.

3.2.3 Technical Specification

In Table 1, the main technical specifications of the

presented dosimeter are resumed:

Table 1: Dosimeter technical specifications.

Temperature range 10 – 40 ºC

Resolution 1 cGy

Accuracy ± 3 %

Linear range

15 Gy

> 80 Gy

Thermal drift < 3mGy/ºC

Delay after irradiation 2 – 3 minutes

* with recalibrations every 15 Gy.

4 CONCLUSIONS

A novel dosimeter based on MOSFET sensor is

presented, suitable for radiotherapy monitoring. In

the same instrument, main advantages of previous

commercial equipment have been joined such us low

thermal and fading interference, enough resolution

and accuracy for clinical usage, extended linear

range and portable reader with wireless sensors.

Moreover, these features have been obtained with a

low-cost commercial MOSFET instead of expensive

and dedicated radiation sensor based on MOS

technology.

Regarding with future works, preliminary results

with LINAC irradiations (6 and 18 MV) are quite

promising. Build-up layer needs to be increased as

energy photons increases but general performance is

similar to the presented results for Cobalt source.

ACKNOWLEDGEMENTS

This work has been funded by the Ministerio de

Educación y Ciencia, Dirección General de

Enseñanza Superior (Spain) (Projects CTQ2009-

14428-CO2-01, CTQ2009-14428-CO2-02), from the

Junta de Andalucía (Projects P09-FQM-5341, P08-

FQM-3535) and from the Ministerio de Ciencia e

Innovación (Project PS-300000-2009-5).

REFERENCES

Asensio L J, Carvajal M A , López-Villanueva J A, M.

Vilches, A. M. Lallena and A. J. Palma, 2006

Evaluation of a low-cost commercial MOSFET as

radiation dosimeter Sensors and Actuators A 125 288-

295,.

Benson C, Joyce M J, O’Connell B and Silvie J 2000

Neutron Detection at the Extremes of Sensitivity in the

Cosmic Environment IEEE Trans Nucl. Sci. 6 2417-

2422.

Best S, Ralson A and Suchowerska N 2005 Clinical

Application of the OneDose

Patient Dosimetry

System for total body irradiation Phys. Med. Biol. 50

5909-5919.

Bloemen E, Bois W, P. Visser P, Bruinvis I, Jalink D,

Hermans J and Lambin P 2003 Clinical dosimetry

with MOSFET dosimeters to determinate the dose

along the field junction in a split beam technique

Radiother. Oncol. 67 (2003) 351-357.

Buehler M G, Blaes B R, Soli G A and Tardio G R 1993

On Chip p-MOSFET Dosimetry IEEE Trans. Nucl.

Sci. 40 1442-1449.

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

Carvajal M A, Vilches M, Guirado D, Lallena A M,

Banqueri J and Palma A J, 2010 Readout techniques

for linearity and resolution improvements in MOSFET

dosimeters Sensors and Actuators A 157 178–184.

Carvajal M A, Martínez-Olmos A, Morales D P,

López-Villanueva J A, Lallena A M and Palma A J,

2010 Thermal Drift Reduction With Multiple Current

Polarization For MOSFET Dosimeters send for

publication to Phys. Med. Biol.

Fraden J, 1996 Handbook of modern sensor. Physics,

design and application, 2

Edition (New York:

Springer Science+Biseness Media).

Halvorsen P. H. 2005. Dosimetric evaluation of a new

design MOSFET in vivo dosimeter, Med. Phys. 32

110-117.

Haran A, Jaksic A, Refaeli N, Eliyahu A, David D and

Barak J 2004 Temperature Effects and Long Term

Fading of Implanted and Unimplanted Gate Oxide

RADFETs, IEEE Trans. Nucl. Sci. 5 2917-2921.

Holmes-Siedle A and Adams L 1986 RADFET: A review

of the use of metal-oxide-silicon devices as integrating

dosimeters Radiat. Phys. Chem. 28 235-244.

Hughes R C, Huffman D, Snelling J V, Zipperian J E,

Ricco A J and Kelsy A 1988 Miniature Radiation

Dosimeter for In-Vivo Radiation Measurements, Int. J.

Radiat. Oncol. Biol. Phys. 14 963-967.

Kwan I S, Rosenfeld A B, Qia Z Y, Wilkinson D, Lerch M

L F, Cutajar D L, Safavi-Naeni M, Butson M, Bucci J

A, Chin Y and Perevertaylo V L 2008 Skin dosimetry

with newMOSFET detectors Radiation Measurements

43 929 – 932.

Ma T P and Dressendorfer P V 1989 Ionizing Radiation

Effects In MOS Devices and Circuits (New York: John

Wiley & Sons).

Rosenfeld A B 2002 MOSFET Dosimetry on Modern

Radiation Oncology Modalities, Radiat. Prot. Dosim.

101 393-398.

Sarrabayrouse G and Siskos S 1998 Radiation Dose

Measurement using MOSFETs, IEEE Instrum. Meas.

Mag. 1 26-34.

Soubra M, Cygler J and Mackay G 1994 Evaluation of a

dual bias metal oxide-silicon semiconductor field

effect transistor detector as radiation dosimeter, Med.

Phys. 21 (1994) 567-572.

Sicel Technologies 2005 OneDose User´s Manual,

Pre-production draft version, rev. P01, (Morrisville:

Sicel Technologies Inc.).

Sze S M 1981 Physics of Semiconductor Devices, 2

edition. (New York: John Wiley & Sons)

Thomson & Nielsen 1991 Direct reading dosimeter

European Patent Office EP 0471957A2, (02/07/1991).

PORTABLE DOSIMETER WITH MOSFET SENSOR FOR RADIOTHERAPY MONITORING