Download The UK primary standard calorimeter for photon

Phys. Med. Biol. 41 (1996) 137–151. Printed in the UK
The UK primary standard calorimeter for photon-beam
absorbed dose measurement
A R DuSautoy
National Physical Laboratory, Teddington, Middlesex TW11 OLW, UK
Received 30 June 1995
Abstract. A description is given of the UK primary standard graphite calorimeter system. The
calorimeter measures absorbed dose to graphite for photon radiations from 60 Co to 19 MV x-rays,
and is the basis of the NPL therapy-level absorbed dose to water calibration service. Absorbed
dose to graphite from the photon calorimeter has been compared with three other standards: an
ionization chamber and cavity theory, for 60 Co gamma radiation; the NPL electron calorimeter,
for 12–14 MeV electron beams; and the BIPM 60 Co absorbed dose standard. The three standards
agreed within 0.5% which is similar to the measurement uncertainties.
1. Introduction
The National Physical Laboratory (NPL) uses a calorimeter as the UK primary standard
of absorbed dose. The advantage of calorimetry over exposure standards is that the
radiation energy is absorbed in the calorimetric medium and absorbed dose is directly
related to the temperature rise. Ideally the calorimetric medium would be water, but
the chemical reactions in irradiated water may evolve or absorb heat and so interfere
with the measurement of absorbed dose. Graphite is similar to water in its radiation
absorption and scattering properties, but without any chemical reactions. It is therefore the
calorimetric material recommended by the International Commission on Radiation Units and
Measurements (ICRU) Report 14 (ICRU 1969). The calorimetric measurement of absorbed
dose to graphite still requires conversion to the quantity of interest i.e. absorbed dose to
water as described by Burns (1994).
The calorimeter described here has been called the photon-beam calorimeter to
distinguish it from the electron-beam calorimeter described in section 9.2. It can be used to
measure absorbed dose to graphite in a 60 Co gamma-ray beam, in x-rays of nominal energy
2 to 19 MV, and in electron beams of nominal energy 4 to 19 MeV. This calorimeter has
been the primary standard for the NPL calibrations in absorbed dose to water (Burns et al
1988, Rosser et al 1994) and the Chemical Dosimetry services since 1988.
The mean absorbed dose to carbon (graphite) in the core of the calorimeter Dc is
determined by measuring the energy deposited by ionizing radiation ER in the core, of mass
m, according to the definition of absorbed dose given by the ICRU in Report 33 (1980):
ER
Dc =
.
(1)
m
For a dose of one gray, the energy absorbed ER raises the temperature of the core by
about a millikelvin. This temperature rise is measured with a resolution of a few tens of
microkelvin by a thermistor embedded in the core. The decrease in thermistor resistance is
c 1996 British Crown Copyright
137
138
A R DuSautoy
detected by an AC bridge, the voltage output of which is read by a computer which also
controls the measurement sequence and processes the data.
The calorimeter is calibrated by dissipating a measured electrical energy in the core,
substituting the electrical energy for radiation energy. The total uncertainty in the calibration
of an ion chamber in terms of absorbed dose to graphite in the NPL 60 Co beam is 0.24%.
2. Calorimeter design
The calorimeter was built to a design by Domen and Lamperti (1974) with modifications to
the vacuum housing, electrical connector and one of the graphite jackets, as described by
Cross (1988). The calorimeter consists of a small disc-shaped core of graphite surrounded by
three graphite jackets and an extensive graphite phantom. The jackets and phantom present
a nearly homogeneous graphite medium for the absorption and scattering of the beam and
thermally insulate the core. Because of the small temperature rise in the calorimeter, the
heat transfer between the core and its surroundings must be minimized. The conversion
of absorbed dose to graphite into dose to water require that the dose be measured in
homogeneous graphite, as described by Burns and Dale (1990) and Burns (1994). Therefore
great care was taken to minimize the use of materials other than graphite. A premium
quality, fine-grain, flaw free, high-strength graphite (ATJ grade supplied by UCAR Carbon
Ltd) was used in the construction of the calorimeter. Its mean density was measured as
1.737 g cm−3 .
3. Construction
The optimum size and shape for the core is discussed in ICRU Report 14 (ICRU 1969).
Figure 1 shows the vertical cross-section through the axis of the assembled calorimeter and
DuSautoy (1991) gives the dimensions in more detail. The core is placed near the front so
that electron-beam measurements can be made. The minimum measurement depth is about
1 g cm−2 . Two thermistors were included in the core; one is used as a sensor and the other
as a heater for the electrical calibration.
The dimensions of the jackets are governed by those of the core. Discontinuities in the
graphite have to be kept as small as possible while maintaining thermal insulation of the
core. A sensor and heating thermistor are also included in the first jacket. The mass of
the first jacket is similar to that of the core so the calorimeter can operate in the ‘heat loss
compensated’ mode described by Domen and Lamperti (1974).
The sensing thermistor in the second jacket is placed near the front. A carbon resistor
at the back provides heating. Aluminized Mylar film is used to reduce thermal radiation
between components. It covers the inside and outside of the second jacket and the inside of
the third jacket. Heating wires are wound around a channel in the back of the third jacket
to provide temperature control.
A vacuum housing made of Perspex (polymethyl methacrylate) was shaped to fit closely
around the sides and back of the third jacket. The vacuum housing was completed by a sheet
of aluminized Mylar film over the front face. Each of the calorimeter components inside
the vacuum housing was electrically earthed to prevent charge build-up. All electrical
connections were fed through a Perspex tube to an air-tight electrical connector. A
connection was also made from this tube to an air cooled turbomolecular pump. This
pump evacuates the calorimeter to a pressure of 10−3 torr, which eliminates convection and
greatly reduces thermal conduction.
The UK primary standard calorimeter
139
Figure 1. Schematic diagram of the NPL calorimeter. Side view cross-section. Not to scale.
Dimensions in mm.
The phantom consists of a graphite annulus, shaped to fit closely around the outside of
the vacuum housing. Build-up plates of appropriate thicknesses can be added to the front
of the calorimeter to give the required depth in graphite, and to the back to give adequate
scatter.
To derive absorbed dose from the electrical calibrations it is necessary to weigh the core.
140
A R DuSautoy
The distance from the midpoint of the core to external reference points were measured, as
was the depth of material in front of the midpoint of the core. The components of the
calorimeter were weighed and measured at each stage of assembly and the effective mass
of the core m was found to be 1.5043 g ± 0.05%. The distance from the midplane of the
core to the front of the vacuum housing was 7.83 ± 0.1 mm and the actual mass depth was
0.9327 ± 0.006 g cm−2 of graphite.
4. Method of operation
The core sensor thermistor is one arm of an AC bridge, which is operated like a DC
Wheatstone bridge. The first and second jackets employ bridges of a similar design.
The core resistance measurement and heating circuits are shown in figure 2. Initially
the second jacket heating is regulated to produce a constant second jacket temperature, and
the core and first jackets are allowed to reach thermal equilibrium. The core, first-jacket
and second-jacket bridges are then balanced, by adjusting the balancing resistor (e.g. RM in
the core circuit). Before a measurement, the heating power in the second jacket is fixed at a
level which gives a steady increase in temperature at a rate of approximately 12 µK min−1 .
Figure 2. Schematic diagram of electrical circuits.
Figure 3 shows a typical measurement sequence. The bridge amplifier output voltages
of the core and two jackets are recorded for about 300 seconds (A to B). The calorimeter
is then irradiated (B to E), reducing the resistance of the sensing thermistors embedded in
the core and two jackets and increasing their bridge output voltages.
Irradiations last for about 80 seconds. Midway through the irradiation time (at C), the
switch, S1 , is closed, reducing the resistance of that arm of the core bridge by a small fixed
amount (i.e. 0.1 ). Simultaneously the balancing resistances of the first- and second-jacket
bridges are reduced to offset their bridge output voltages below their balance points. The
The UK primary standard calorimeter
141
irradiation continues to heat the calorimeter, increasing the bridge output voltages. The
irradiation is stopped (at E), when the core bridge output voltage returns to the pre-heating
balance voltage (at B). The output voltages of all the bridges then level off and are recorded
for a further 300 s (E to F) after the heating.
Figure 3. Plot of bridge output voltage against time.
Straight lines are fitted to the data collected before and after the heating period using
the method of least squares. Heat gains or losses by the core (other than those due to the
irradiation) are corrected for, by extrapolating these lines to half way through the heating
period (D). The difference in voltage found by extrapolating these lines gives the bridge
amplifier output voltage difference, called Vdiff . This follows the method of Laughlin and
Genna (1966). The thermal insulation of the core and the control of the second jacket
temperature removes the need for more complicated fitting procedures.
Electrical calibration measurements are similar to the radiation measurements except
that electrical heating is used instead of radiation heating. Electrical heating is applied to
the first and second jackets to minimize heat losses from the core. This is achieved by
minimizing the change in temperature gradient between the pre- and post-heating periods
for the core, and also for the first jacket.
The electrical calibration data are fitted in the same way as for the radiation heating and
information from each measurement (each component’s balancing resistor, the coefficients
of the fits, the electrical power applied to each component, the duration of the periods
before, after and during the heating, etc) is printed out and stored on disk for later analysis.
By employing the offset resistance and restoring the bridge to balance after the heating,
the amplifier is used as a null detector. Therefore the gain of the amplifier must be high,
to achieve sensitivity, but does not need to be accurately known.
The electrical calibration measures the electrical heating energy corresponding to the
core bridge offset. The radiation measurement uses the same offset (figure 3), so if the Vdiff
142
A R DuSautoy
is zero, then the energy deposited during the irradiation equals that in the electrical heating.
Since the core mass is known (sections 3 and 8.1), the absorbed dose can be calculated
from equation (1).
5. Resistance measurement circuits
The resistance measurement bridge used to compare the electrical and radiation heatings
in terms of the offset resistance is described in detail by DuSautoy (1985) and Stoker and
DuSautoy (1991). Other workers have used DC circuits, but these are electrically noisy,
partly due to thermal EMFs; instead low-frequency double Kelvin AC bridge circuits are
used in this work. Some of the comparisons described in section 9 used an earlier system,
but where necessary corrections have been applied to the results to allow for this change.
All thermistors used in the calorimeter are low-noise miniature beads encapsulated in glass.
At the operating temperature of 24.4 ◦ C the core thermistor resistance is 2058  and the
resistance change with temperature is about –3.5% K−1 .
A smooth sine wave produced by the phase sensitive detector drives the bridges and
provides a reference input to the phase sensitive detector. The following frequencies are
used: core 173.3 Hz, first jacket 133.4 Hz and second jacket 104.2 Hz. These were chosen
so that they were not harmonics of each other or of the mains electricity supply (48 to
52 Hz). The phase of the reference wave relative to the bridge drive is under computer
control. This system resolves 8 µK or 7 mGy, while dissipating 2 µW in the sensing
thermistor. For comparison, the power input to the heating thermistor during electrical
calibration is 35 µW.
6. Electrical heating measurements
The electrical energy dissipated by the heating thermistor in the core is found by measuring
the potential across the thermistor, the current flowing through it and the duration of the
heating t. The electrical energy is only measured precisely in the core circuit.
Measurements of the potential drop across the thermistor V1 and the potential drop V2
across a standard resistor (of value R = 5 k) in series with it give the power applied
to the thermistor, shown in figure 2. The potential V2 is sampled half way through the
heating period, so although the power in the heater changes during the heating, the sample
approximates the average power.
The duration of the electrical heating t is found by counting the cycles of a standard
frequency from a quartz clock. Thus
V2
(2)
EE = V1 t.
R
The heating of the core and first two jackets is controlled from a digital-to-analogue converter
and relay.
7. Analysis of radiation and electrical heating measurements
In radiation measurements it is not possible to stop the irradiation so that the bridge balances
exactly, there is normally a small undershoot or overshoot giving a bridge amplifier output
voltage difference Vdiff . Therefore the information required from the electrical calibration
is not only the electrical energy required to regain balance, but also the relationship between
the energy dissipated in the core and Vdiff .
The UK primary standard calorimeter
143
The electrical energy required to regain balance of the core bridge after the fixed
resistance offset is also a function of the core temperature. Therefore the information
from a series of electrical calibration measurements is used to evaluate the coefficients in
the following empirical equation:
EE = a1 + a2 (Vdiff ) + a3 (RM )
(3)
where EE is the electrical energy dissipated in the core, Vdiff is the measured core bridge
amplifier output voltage difference and RM is the core balancing resistance. RM is chosen
as an indicator of the temperature. The current estimates of the coefficients (equation (4))
differ significantly from those given in the report by DuSautoy (1991), because the stray
resistance correction was reassessed in 1992.
The present estimate of the coefficients is based on about 290 calibrations. To determine
the coefficients accurately the electrical calibrations were carried out over a range of Vdiff
and resistance values. The duration of the heating was adjusted to obtain a spread of Vdiff
from −4 to +4 volts. The temperature of the temperature-controlled cabinet where most
electrical calibrations are carried out was varied to produce a spread of core balancing
resistor values of about 40 ohms.
Having evaluated the coefficients from the electrical calibration, the energy absorbed
during each of the radiation measurements ER in joules was calculated from equation (3)
using the measured Vdiff (in volts) and the core balancing resistance (in ohms).
ER = 1.620 64 × 10−3 + 1.2797 × 10−5 (Vdiff ) − 1.059 × 10−6 (RM ).
(4)
The calorimeter has been operated in the ‘heat compensated mode’ of Domen and
Lamperti (1974) and the ‘quasi-isothermal’ mode of Witzani et al (1984). The agreement
found between different modes of operation gave us confidence that different temperature
distributions have little affect on the final results. Described here is the quasi-adiabatic
mode; the other modes are described by DuSautoy (1985).
8. Radiation independent corrections and uncertainties
Uncertainties are treated following NAMAS (1990). In this paper the total uncertainties
are quoted, unless further information is given. Measurements are a sample from a parent
distribution which is assumed to be Gaussian. Uncertainties indicate the spread of the parent
distribution at one standard deviation. The total uncertainty is found by adding the random
and nonrandom components in quadrature. The random component is found by statistical
means from repeated measurements and is typically a standard deviation of the mean. The
non-random uncertainties are assumed to be Gaussian. The non-random uncertainty in a set
of measurements is found from a calibration certificate or estimates of the likely variation
of its components.
8.1. Core mass
The estimated non-random uncertainty in absorbed dose arising from the measurement of the
mass of graphite in the core is ±0.01% and that for the mass of the non-graphite components
(e.g. thermistors, adhesive) is ±0.04%. The measurement has a random uncertainty of less
than ±0.01%.
144
A R DuSautoy
The dose the core would have received if it had been entirely graphite Dc can be
calculated from
n
X
Di
mi
mg +
(5)
Dc = ER
Dg
i=1
where mg is the mass of graphite, mi is the mass of the non-graphite impurity identified
by the subscript i, n is the number of non-graphite impurities, Dg is the average absorbed
dose in graphite and Di is the average absorbed dose in the impurity labelled i.
The ratio of the doses Di /Dg is sufficiently near unity and the impurities are so small
for this calorimeter thatPthe effective mass of the core m was taken (for all beams and
energies) as m = mg + ni=1 mi , i.e. 1.5043 g ± 0.05%.
8.2. Electrical calibration
The electrical calibration measurements used to evaluate the coefficients in equation (3)
relating the electrical energy dissipated in the core to the output voltage difference Vdiff
and the initial core balancing resistance RM were described in section 7.
The random uncertainty derived from electrical calibrations is ±0.11%. The percentage
corrections ki (giving rise to non-random uncertainties in the measurement of energy) can
be written as follows
EE = V1
V2 Y
((ki /100) + 1).
t
R i
(6)
The first correction accounts for the heat lost in the leads delivering power to the core.
The 0.018 mm platinum leads from the thermistors were soldered to 0.03 mm diameter
enamelled copper leads. The copper leads return to the printed circuit board (PCB) at the
back of the second jacket. The correction was calculated from the resistances of the circuit
components by attributing half the energy dissipated in the leads between the core and the
first jacket to the core.
One core heater lead is not electrically insulated from the core earthing wire. A
correction for this stray resistance was calculated from the resistances of the circuit
components and the power in each individual heating measurement. The input impedance
of the DVM (value 10 G) requires a negligible correction.
The following factors contribute to the uncertainty in electrical energy, but no correction
is made for their effects; DVM calibration, timer calibration and calibration of the 5 k
resistor in the power circuit. The lack of linearity of energy with the phase sensitive detector
output voltage and with the core balancing resistor give rise to small uncertainties.
Typical values of the corrections and their associated uncertainties are given in table 1.
The total uncertainty in the electrical energy appears as a non-random uncertainty in the
absorbed dose measurements.
9. Intercomparisons
These comparisons have been described elsewhere so will only be briefly summarized
here. The comparisons described in the following sections were all in graphite, in terms of
absorbed dose to graphite.
The UK primary standard calorimeter
145
Table 1. Typical corrections and uncertainties for electrical calibration.
Effect
% Correction
Random effects
Not applicable
% Uncertainty
in electrical
energy (±1σ )
0.11
Non-random effects
Lead loss
−0.45
Stray resistance
+0.11
DVM impedance
−< 0.01
DVM calibration
0
Timer calibration
0
5 k resistor in power circuit
0
Vdiff linearity
0
RM linearity
0
0.14
0.03
< 0.01
< 0.01
< 0.01
< 0.01
< 0.01
< 0.01
−0.34
0.18
Total
9.1. Comparison with NPL cavity ionization standards
A comparison was made between absorbed dose to graphite determined by the calorimeter
described in this paper and that derived from an ion chamber and cavity theory. The
information for this section was mainly provided by Burns (1991), Burns and Dale (1990)
and DuSautoy (1991).
The comparison was made in three stages: the NPL transfer chamber 260 was calibrated
by the calorimeter (in a graphite phantom, in a 60 Co beam) giving Ncal as described in
section 9.1.1; the chamber 260 was compared with the standard chamber SCC3 (in air, in
a 60 Co beam) from which the volume V of chamber 260 could be found as described in
section 9.1.2; knowing the volume of chamber 260, cavity theory was used to calculate its
calibration factor in absorbed dose to graphite Ncav , in a graphite phantom as described in
section 9.1.3. The ratio of calibration factors Ncal /Ncav gives the ratio of doses based on
each standard.
The transfer chamber was an NPL secondary-standard chamber, manufactured by NE
Technology Ltd, type number NE2561 and serial number 260 (volume approximately
0.3 cm3 ). The standard chamber SCC3 (volume approximately 1.5 cm3 ) was one of the
three graphite cavity chambers designed to be the new 2 MV primary standard of air kerma.
Its volume has been accurately measured. This procedure was adopted to avoid having
to calculate a large and uncertain displacement correction for the standard chamber SCC3
which would have been necessary if it had been used directly in a graphite phantom.
9.1.1. Calibration of chamber 260 by calorimeter. The absorbed dose calibration factor
Nc for chamber 260 is Dc /Q, where Dc is the absorbed dose to graphite, in homogeneous
graphite at the centre of the chamber with the chamber replaced by graphite, and Q is the
corresponding charge collected from ion chamber number 260.
First, absorbed dose in the calorimeter was found from equation (1). The energy
deposited was calculated from equation (4) having measured the output voltage difference
and balancing resistance as described in section 4. The measurement of the core mass is
described in sections 3 and 8.1.
Since 1991 a monitor chamber has been used to link the ion chamber and calorimeter
measurements, in 60 Co. This avoided the need to correct the calorimeter measurements for
146
A R DuSautoy
the transit time of the 60 Co source. If the monitor reading when the calorimeter is in the
beam is MUcal and that when ion chamber 260 is in the beam is MUcham then
Dc /MUcal
.
(7)
Nc =
Q/MUcham
The irradiations took place in a Mobaltron 60 Co beam, with source to detector distance of
70 cm and field size 10 × 10 cm2 . All field sizes are quoted at the measurement plane.
The measurement depth in the calorimeter was 5.5745 g cm−2 and this was corrected to the
calibration depth 5.5456 g cm−2 which corresponded to a scaled depth of 5 cm in water as
described by Burns and Dale (1990).
The random uncertainty of the mean of 73 measurements taken in November 1991 and
August 1992 was ±0.06%.
The percentage corrections ki (giving rise to non-random uncertainties in the
measurement of absorbed dose) can be written as follows:
ER Y
Dc =
((ki /100) + 1).
(8)
m i
Typical values of the corrections and their associated uncertainties are given by table 2. The
correction for vacuum and air gaps is given by Owen and DuSautoy (1990).
Table 2. Corrections and uncertainties for the calorimeter in
60 Co
radiation.
Effect
% Correction
% Uncertainty
in dose (±1σ )
Random effects
Not applicable
0.06
Non-random effects
Electrical calibration
Mass
Vacuum and air gaps around core
Beam radial non-uniformity
Depth in graphite
Distance from source
Difference in phantom size
Electrometer calibration etc
Not applicable
Not applicable
+0.50
+0.14
+0.09
0.00
0.00
Not applicable
0.18
0.05
0.13
0.01
0.17
0.10
0.02
0.05
Total
+0.73
0.31
The ion chamber phantom was a 17 cm cube. The centre of the ion chamber was placed
at a depth of 5.6031 g cm−2 in graphite of average density 1.76 g cm−3 .
The second part of the calibration was the determination of the charge from chamber 260
in a graphite phantom under the same measurement conditions as for the calorimeter. The
chamber was connected to a NE Technology Ltd electrometer type number NE2560 (serial
number E3) with a charge calibration of 9.990 03 nC V−1 which supplied the polarizing
potential of −230 V.
The mean ion chamber reading was corrected for the following effects: depth in graphite,
ion recombination and distance from the source (which sum to 0.29 ± 0.05%). Also each
ion chamber measurement was individually corrected for air density and humidity.
Therefore the chamber calibration factor in the specified irradiation conditions in the
ion chamber phantom, based on the calorimeter, was
Ncal = 90.94 mGy nC−1 ± 0.24%
corrected to dry air at 20 ◦ C, 1013.25 mbar, and to zero ion recombination.
The UK primary standard calorimeter
147
9.1.2. Determination of the volume of chamber 260. The chamber 260, fitted with
a graphite build-up cap, was compared in air with standard chamber SCC3 (volume
1.4555 cm3 ). The irradiations took place in a 60 Co beam, with a source–detector distance
of 80 cm and field size of 12 × 12 cm.
The percentage corrections ki (giving rise to non-random uncertainties in the
measurement of the volume V of chamber 260) can be written as follows
Y
I260
volume of SCC3 ((ki /100) + 1)
(9)
V =
ISCC3
i
where I260 is the current collected from chamber 260 and similarly for ISCC3 . Corrections
were made for the following effects: air density and humidity, ion recombination in chamber
260 and chamber SCC3, effects of the stem in both chambers and their difference in wall
thickness. The total correction was typically 0.19 ± 0.06%. The volume V of chamber 260
was found to be 0.30641 cm3 ± 0.06%.
9.1.3. Calculation of Ncav .
by
Ncav =
The calibration factor Ncav for a chamber in graphite was given
(W/e)a
(L/ρ)ca pc sc
Vρa
(10)
where the symbols have the following meanings: (W/e)a is the mean energy required to
produce an ion pair in dry air, per unit charge (33.97 J C−1 ); V is the effective volume of the
NE2561 chamber 260 (3.0641 × 10−7 m3 ); ρa is the density of dry air at 20 ◦ C, 1013.3 mbar
(1.2044 kg m−3 ); (L/ρ)ca is the ratio of mean restricted mass stopping powers of carbon
to air (1 = 10 keV), 1.0031 (interpolated quality index 0.565 from values in table A.14.2
of Burns and Dale (1990) graphically smoothed); pc is the correction for the effect on the
response of displacing graphite by the air cavity from AAPM (1983) (−1.7 ± 0.33%) and
sc is the experimentally determined correction for stem scatter in graphite (1.000). This
gave a calibration factor in absorbed dose to graphite for chamber 260 of
Ncav = 90.77 mGy nC−1 ± 0.36%.
9.1.4. Result. The ratio of absorbed dose to graphite measured by the calorimeter to that
measured by an ion chamber and cavity theory was therefore
Ncal
Dcal
=
= 1.0019 ± 0.47%.
Dcav
Ncav
(11)
9.2. Comparison with NPL electron-beam calorimeter
NPL has developed a graphite electron calorimeter specifically for measuring absorbed dose
in electron beams as described by Burns and Morris (1991) and Burns et al (1994). The
calorimeter consists of a small disc-shaped graphite core 50 mm in diameter and 7 mm in
thickness separated by an air gap of about 0.5 mm from a graphite phantom. Absorbed
dose to graphite is calculated from the product of the temperature rise in the core due
to irradiation and the specific heat capacity of the core. The specific heat capacity was
measured independently by Williams et al (1993). The temperature rise is measured by
means of a calibrated thermistor in the core. The thermistor forms one arm of a DC
Wheatstone bridge and extrapolation of the pre- and post-irradiation periods accounts for
heat losses.
148
A R DuSautoy
Table 3. Corrections and uncertainties for electron calorimeter in electron beams.
Effect
% Correction
% Uncertainty
in dose (±1σ )
Random effects
Not applicable
0.08
Non-random effects
Specific heat capacity of core
Thermistor calibration
Beam radial non-uniformity
Temperature rise
Gap effect
Not applicable
Not applicable
+0.10
0
0
0.08
0.05
0.03
0.10
0.15
Total
+0.10
0.22
The photon and electron calorimeters have been compared in electron beams, in terms
of absorbed dose to graphite. The comparison was based on five runs at 14.6 MeV and
one at 12.6 MeV. Details are given by Burns and Morris (1991). Dose rates from 15 to
60 Gy min−1 were used, at a beam pulse repetition frequency of 240 Hz and a pulse width
of 3 µs. All measurements were made with the beam scattered by 2 mm aluminium at a
distance of 2 m ± 0.3 mm from the calorimeter core centres. The centres of the calorimeter
cores were placed at a depth of 2.034 ± 0.17 g cm−2 in graphite. The uncertainty here
has little effect on the absorbed dose, as the dose varies slowly with depth under these
conditions. A transmission ion chamber was used to monitor the beam during runs with
each calorimeter. At present the effects of gaps around the cores of each calorimeter are not
known, but are thought to be small and similar, so an uncertainty was attached to the electron
calorimeter measurements only. The percentage uncertainties in absorbed dose measurement
for the electron calorimeter are given in table 3 and those for the photon calorimeter are
given in table 4. In each run, each calorimeter was given up to 25 irradiations. The mean
ratio of absorbed dose to graphite (weighted by the random uncertainty in each comparison)
was photon calorimeter/electron calorimeter = 1.0017 ± 0.34%.
9.3. Comparison with BIPM
The International Bureau of Weights and Measures, in Sèvres, France, maintains a 4 cm
diameter graphite chamber, in a graphite phantom, as a reference standard. It has been
compared with calorimeters from other national standards laboratories, and its calibration
normalized to their average.
In 1987 two calibrated NPL secondary standard ionization chambers (NE Technology
Ltd type number 2561) serial numbers 260 and 261 were compared in a graphite phantom
with the ‘reference’ standard of absorbed dose at BIPM. The comparison was reported at the
CCEMRI meeting in 1988 (Burns and Owen 1988). I will deal only with chamber number
260 as the results from chamber 261 were the same to within 0.01%.
This comparison took place in two stages:
(i) NE2561 chamber serial number 260 was calibrated, in 60 Co, in a graphite phantom
by comparison with the calorimeter at NPL giving the calibration factors NNP L as described
in section 9.1.1.
(ii) NPL chambers 260 and the graphite phantom were taken to BIPM and compared,
in 60 Co, with the BIPM standard ionization chamber in its own graphite phantom.
The field size at the surface of the phantom in both laboratories was 10 × 10 cm2 .
The UK primary standard calorimeter
149
Table 4. Corrections and uncertainties for photon calorimeter in electron beams.
% Uncertainty
in dose (±1σ )
Effect
% Correction
Random effects
Not applicable
0.06
Non-random effects
(Old) electrical calibration
Mass
Depth in graphite
Distance from source
Beam radial non-uniformity
Not applicable
Not applicable
0
0
0
0.19
0.05
0.03
0.15
< 0.01
Total
0
0.26
At NPL, the calibration factors, NN P L , were measured as described in section 9.1.1 for
chamber 260, 90.94 mGy nC−1 with an uncertainty of ±0.43%. The NPL uncertainties
were described in section 9.1.1.
The absorbed dose rate to graphite at BIPM at the point of measurement in the BIPM
graphite phantom was derived from a measurement with the BIPM standard chamber:
ḊBI P M =
I
(W/e)a (s/ρ)ca pc krn kdist kref
m
(12)
where the symbols have the following meanings and values: I /m is the current per unit
mass of air in the BIPM chamber (102.92 µA kg−1 ), (W/e)a is as in section 9.1.3, (s/ρ)ca is
the ratio of mean collisional mass stopping powers of carbon to air for secondary electrons
(1.0030), pc is the correction for the effect on the response of displacing graphite by the
air cavity (0.9896), krn is the correction for the radial non-uniformity of the beam (1.0032),
kdist is the correction from 5.0186 g cm−2 to 5.0 g cm−2 (1.000 56), kref is a normalization
correction with respect to the weighted mean of absorbed dose from comparison with several
graphite calorimeters (0.9992) and the correction from 5.000 to 5.6031 g cm−2 depth in the
NPL phantom was 0.9815.
At BIPM, the BIPM measurement of absorbed dose to graphite at 5.6031 g cm−2 depth
and 1 m source distance was 3.4079 mGy s−1 on 1 January 1987. The corresponding BIPM
measured ionization current in the NPL graphite phantom was 37.635 pA. This gave a BIPM
calibration factor NBI P M of 90.55 mGy nC−1 with uncertainties of ±0.06%.
The ratio of absorbed dose to graphite NPL/BIPM is given by
NN P L
DN P L
=
= 1.0043 ± 0.32% for chamber 260.
DBI P M
NBI P M
(13)
10. Conclusions
The ratios of absorbed dose to graphite derived from the NPL photon calorimeter compared
with other methods are summarized as follows:
60
NPL ion chamber and cavity theory 1.0019 ± 0.47%
Co
−
12–14 MeV e NPL electron calorimeter
1.0017 ± 0.34%
60
1.0043 ± 0.32%.
Co
BIPM absorbed dose standard
150
A R DuSautoy
The NPL photon calorimeter agrees with three independent assessments of absorbed dose
to graphite within 0.5%. One of the three comparisons is outside the uncertainties at the
one standard deviation level, as one might expect.
Acknowledgments
The author acknowledges the help of Mr C J Brend, Ms K E Rosser and Miss D H Pritchard
in carrying out the calorimeter and ion chamber measurements. Thanks are given to Mrs
C Cross who built the calorimeter and Mr I Stoker who designed and built the electronics.
The following people contributed to the comparisons: Mr J E Burns and Mr J W G Dale
for the ionization chambers and Dr D T Burns for the electron calorimeter. Finally I would
like to thank Mr B Owen for the guidance he gave throughout this project.
References
AAPM 1983 A protocol for the determination of absorbed dose from high-energy photon and electron beams Med.
Phys. 10 741–71
Burns D T, McEwen M R and Williams A J 1994 An NPL absorbed dose calibration service for electron-beam
radiotherapy Proc. Int. Symp. on Measurement Assurance in Dosimetry (Vienna, 1993) (Vienna: IAEA) pp 61–
71
Burns D T and Morris W T 1991 A graphite calorimeter for electron-beam dosimetry Proc. Int. Symp. on High
Dose Dosimetry for Radiation Processing (Vienna, 1990) (Vienna: IAEA) pp 123–36
Burns J E 1991 private communication
——1994 Absorbed-dose calibrations in high-energy photon beams at the National Physical Laboratory: conversion
procedure Phys. Med. Biol. 39 1555–75
Burns J E and Dale J W G 1990 Conversion of absorbed-dose calibration from graphite into water National
Physical Laboratory Report RSA(EXT) 7
Burns J E, Dale J W G, DuSautoy A R, Owen B and Pritchard D H 1988 New calibration service for high energy
X-radiation at NPL Proc. Int. Symp. on Dosimetry in Radiotherapy (Vienna, 1987) vol 2 (Vienna: IAEA)
pp 125–32
Burns J E and Owen B 1988 NPL/BIPM comparison of absorbed dose 1987–88 BIPM Com. Consult. Etalons Mes.
Ray. Ionisants (Section I) 1988 Document 88-14
Cross C 1988 The construction of a graphite microcalorimeter for the measurement of absorbed dose National
Physical Laboratory Report RS(EXT) 96
Domen S R and Lamperti P J 1974 A heat-loss-compensated calorimeter: theory, design and performance
J. Res. NBS 78A 595–610
DuSautoy A R 1985 The measuring assembly for a primary standard absorbed dose calorimeter National Physical
Laboratory Report RS(EXT) 74
——1991 The UK primary standard calorimeter for absorbed dose measurement National Physical Laboratory
Report RSA(EXT) 25
ICRU 1969 Radiation dosimetry: x-rays and gamma rays with maximum photon energies between 0.6 and 50 MeV
ICRU Report 14 (Washington, DC: ICRU)
——1980 Radiation quantities and units ICRU Report 33 (Washington, DC: ICRU)
Laughlin J S and Genna S 1966 Calorimetry Radiation Dosimetry vol II, ed K R Kase et al (New York: Academic)
pp 389–441
National Measurement and Accreditation Service (NAMAS) 1990 The expression of uncertainty in radiological
measurements NAMAS Information Sheet B0825
Rosser K E, Owen B, DuSautoy A R, Pritchard D H, Stoker I and Brend C 1993 The NPL absorbed dose to
water calibration service for high-energy photons Proc. Int. Symp. on Measurement Assurance in Dosimetry
(Vienna, 1993) (IAEA-SM-330/35) (Vienna: IAEA) pp 73–81
Stoker I and DuSautoy A R 1991 The measuring assembly for the primary standard absorbed dose graphite
calorimeter at therapy levels National Physical Laboratory Report RSA(EXT) 23
Owen B and DuSautoy A R 1990 NPL absorbed dose graphite calorimeter correction for the effect of the gaps
around the core National Physical Laboratory Report RSA(EXT) 12
The UK primary standard calorimeter
151
Williams A J, Burns D T and McEwen M R 1993 Measurement of the specific heat capacity of the electron-beam
graphite calorimeter National Physical Laboratory Report RSA(EXT) 40
Witzani J, Duftschmid K E, Strachotinsky Ch and Leitner A 1984 A graphite absorbed-dose calorimeter in the
quasi-isothermal mode of operation Metrologia 20 73–9