Key comparison BIPM.RI(I)-K6 of the standards for absorbed

dose to water at 10 g cm–2

of the NMIJ, Japan and the BIPM in

accelerator photon beams

S. Picard1, D. T. Burns

1, P. Roger

1

M. Shimizu2, Y. Morishita

2, M. Kato

2, T. Tanaka

2, T. Kurosawa

2 and N. Saito

2

Final report, Version 0.1

1Bureau International de Poids et Mesures, Pavillon de Breteuil, F-92312 Sèvres cedex (France)

2National Metrology Institute of Japan, AIST, 1-1-1 Umezono, Tsukuba, Ibaraki 305-8568 Japan

___________________________________________________________________

ABSTRACT

A comparison of the dosimetry for accelerator photon beams was carried out between

the National Metrology Institute of Japan (NMIJ) and the Bureau International des Poids

et Mesures (BIPM) from 9 to 23 April 2015. The comparison was based on the

determination of absorbed dose to water at 10 g cm–2

for three radiation qualities at

the NMIJ. The results, reported as ratios of the NMIJ and the BIPM evaluations (and

with the combined standard uncertainties given in parentheses), are 0.9966 (47) at

6 MV, 0.9965 (60) at 10 MV and 0.9953 (50) at 15 MV. This result is the eighth in

the on-going BIPM.RI(I)-K6 series of comparisons.

2/25

1. INTRODUCTION

To compare the determination of absorbed dose to water of the National Metrology

Institutes (NMIs) for accelerator photon beams, the Bureau International des Poids et

Mesures (BIPM) has developed a transportable standard for absorbed dose to water

based on a graphite calorimeter [Picard et al. 2009, 2011a]. A comparison programme

was adopted in 2008, currently proposed for fifteen NMIs, and is registered in the

BIPM key comparison database [KCDB] as BIPM.RI(I)-K6; the comparison protocol

ADWG(I) [2009] is also available in the KCDB. The CCRI decided in 2013 to adopt

the value of absorbed dose to water as determined by the BIPM on-site at the NMI as

the key comparison reference value [CCRI 2013, Picard et al. 2013a].

The comparison between an NMI primary standard and that of the BIPM is

established by the determination of absorbed dose to water by each standard at several

accelerator radiation qualities. The BIPM absorbed-dose determination, Dw,BIPM, is

made directly at the NMI using the transportable BIPM standard. The NMI

determination, Dw,NMI, is realized during the comparison using one or more NMI

reference ionization chambers calibrated in advance. The comparison result for each

quality is the ratio

BIPMw,

NMIw,

D

DR

(1)

and its associated uncertainty uc(R) 1

.

Within this framework, a comparison has been made between the National Metrology

Institute of Japan (NMIJ) and the BIPM. The measurements were carried out in the

accelerator laboratory of the NMIJ in Tsukuba during the period 9 to 23 April 2015.

The BIPM calorimeter, water phantom and electronics were shipped in advance,

while the transfer ionization chambers were hand-carried.

2. DESCRIPTION OF STANDARDS AND MEASUREMENTS

2.1. The NMIJ Determination of Absorbed Dose to Water

The NMIJ determines the absorbed dose to water in 60

Co and high-energy photon

beams using a graphite calorimeter (see Figure 1). The Domen-type graphite

calorimeter has been described in detail by Morishita et al. [2013]. It consists of a

graphite disk (core), enclosed by a graphite jacket and shield. Two thermistor

sensors and four thermistor heaters are attached to the core to measure and control

1 Certain commercial equipment, instruments, or materials are identified in this report in order to

specify the experimental procedure adequately. Such identification does not imply recommendation or

endorsement, nor that the materials or equipment identified are necessarily the best available for the

purpose.

3/25

its temperature. To determine the absorbed dose to water in high-energy photon

beams, the calorimeter is operated in a quasi-adiabatic mode in which the jacket

temperature is maintained at the same temperature as the core using thermistor

sensors and heaters attached to the jacket.

The absorbed dose to the graphite core, Dc,NMIJ, can be expressed as follows:

𝐷c,NMIJ =𝐶p,c∆𝑇

𝑚c,NMIJ (2-a)

where mc,NMIJ and Cp,c represent the mass and heat capacity of the graphite core,

respectively, and ∆T is its associated temperature rise. The heat capacity Cp,c is

determined from the temperature rise of the core ∆TR when induced by a known

amount of electrical energy UR generated by the thermistor heaters of the core:

𝐶p,c =𝑈R

∆𝑇R . (2-b)

The heat capacity is measured before and after each irradiation.

The ratio between Dc,NMIJ and absorbed dose to water Dw,NMIJ is determined from

charge measurements using a core-sized graphite cavity chamber when exposed to

the same photon beam. The absorbed dose to water Dw,NMIJ is expressed as

𝐷w,NMIJ = 𝐷c,NMIJ

𝑄w,NMIJ

𝑄c,NMIJ(

𝐷w,NMIJ/𝐷c,NMIJ

𝐷cav,w,NMIJ/𝐷cav,c,NMIJ)

MC

𝑘def ∙ 𝑘imp , (3)

Dcav,w,NMIJ and Dcav,c,NMIJ denoting the calculated dose of the graphite cavity

chamber placed in a water phantom and a graphite calorimeter phantom at the

respective reference points. The quantities in parenthesis are calculated using a

Monte-Carlo code (EGS5) [Hirayama 2005, Shimizu et al. 2014]. The factors kdef

and kimp represent the correction for the heat defect and impurity of the core,

respectively.

The graphite calorimeter has been used to determine the absorbed-dose calibration

coefficients ND,w for the transfer ionization chambers. The calibration coefficient

can be expressed as

𝑁𝐷,w =𝐷w,NMIJ

𝑄w,NMIJ𝑘s𝑘𝑇𝑃𝑘H𝑘rad𝑘axl (4)

where k is the correction factor for ion recombination (s), temperature and

pressure (TP), humidity (H), radial non-uniformity (rad) and axial non-uniformity

(axl).

4/25

Figure 1. Absorbed dose to water standard for high-energy photon beams at the NMIJ

where the NMIJ graphite calorimeter in placed in the beam (to the right) and a water

phantom is seen in the foreground (to the left).

2.2. The BIPM Determination of Absorbed Dose to Water

2.2.1. Description of the calorimeter system

The BIPM absorbed-dose graphite calorimeter is described in Picard et al. [2009,

2011a]. No electrical heating is employed, but rather the specific heat capacity of

the graphite core, cp,c, has been determined previously in a separate experiment

[Picard et al. 2007]. Quasi-adiabatic conditions are achieved by irradiating the

core in a graphite jacket that is smaller than the radiation field, resulting in a

relatively uniform dose distribution in the jacket. This arrangement is mounted in

a PMMA2 vacuum container with graphite build-up plates to centre the core at the

reference depth of 10 g cm–2

. The mean absorbed dose, Dc, in the graphite core is

determined using

impc,c )( kTTcD p .

(

(5)

2 Polymethylmethacrylate

5/25

where T is the temperature rise in the core and kimp corrects for non-graphite

materials in the core.

Two nominally identical parallel-plate ionization chambers with graphite walls

and collector, similar in design to the existing BIPM standards for air kerma and

absorbed dose to water, were fabricated for the determination of the absorbed

dose to water from the measured absorbed dose to the graphite core. The first

chamber is housed in a graphite jacket, nominally identical to the calorimeter

jacket, and is irradiated in the same PMMA support and phantom arrangement,

but without evacuating the phantom.

The second chamber is mounted in a waterproof sleeve and irradiated at the

reference depth in water. These measurement arrangements are represented

schematically in Figure 2.

The method adopted by the BIPM combining calorimetric and ionometric

measurements with Monte Carlo simulations to determine the absorbed dose to

water is described in detail in Burns [2014] and has been applied in a number of

previous comparisons [Picard et al. 2010, 2011b, 2013(b,c), 2014, 2015(a,b,c)].

The absorbed dose to water at the reference point, Dw,BIPM, is evaluated as

BIPMrn,

MC

wcav,

ccav,

MC

c

w

c

w

cBIPMw, kD

D

D

D

Q

QDD

(6)

where

Dc - measured absorbed dose to the graphite core;

Qc - ionization charge measured when the transfer chamber is

positioned in the graphite jacket, replacing the core;

Qw - ionization charge measured when the transfer chamber is

positioned in water;

MC

c

w

D

D

- calculated ratio of absorbed dose to water and to the graphite

core using Monte Carlo simulations;

MC

wcav,

ccav,

D

D

- calculated ratio of cavity doses in graphite and in water using

Monte Carlo simulations;

kr n,BIPM - measured correction for radial non-uniformity in water.

In abbreviated form, Dw,BIPM can be expressed as

BIPMrn,c,wc

wcBIPMw, kC

Q

QDD , (7)

where Cw,c represents the total Monte Carlo conversion factor.

6/25

Figure 2. Schematic representation of the three different measurement situations (a, b

and c) and four Monte Carlo models (a, b, c and d). All measurements are made in a

cubic PMMA phantom, here represented by the dark blue square. a) The calorimeter is

used in vacuum and Dc is both measured and calculated. b) The graphite core is replaced

by the transfer ionization chamber at atmospheric pressure. The ionization charge in

graphite, Qc, is measured and the corresponding cavity dose, Dcav,c, calculated. c) The

ionization chamber is placed in a waterproof sleeve inside a similar phantom filled with

water. The ionization charge in water, Qw, is measured and the corresponding cavity dose,

Dcav,w, calculated. For calculation d), the mean absorbed dose to water, Dw,BIPM, in the

absence of the chamber and sleeve is calculated for a water detector with the same

dimensions as the cavity. It follows that a correction factor kr n,BIPM is required for the

radial non-uniformity of the radiation field over this dimension, measured for a

homogeneous water phantom.

2.2.2. Monte Carlo simulations

The Monte Carlo calculations are described in detail in Burns [2014] and make

use of the PENELOPE code [Salvat et al. 2009]. As noted in the preceding

section, four geometries are simulated and the accuracy of the method relies on

the symmetry of the geometries. A novel aspect of this is the use of a disc-shaped

transfer chamber whose total graphite thickness on-axis is the same as that of the

calorimeter core. Very few of the geometrical bodies appear in only one of the

four simulations so that the fine details should not need to be simulated.

Nevertheless, a very detailed geometrical model was constructed. Similarly,

although detailed electron transport should not be essential for the same reasons,

sufficient detail was used to permit the cavity dose to be calculated in a way that

gives the same results as a full calculation using event-by-event electron transport

(as demonstrated in an earlier work [Burns 2006]). Reference Burns [2014]

includes a detailed uncertainty analysis for the calculation of the conversion factor

Cw,c.

The phase-space files of incident photons at 80 cm from the source were supplied

to the BIPM by the NMIJ, calculated using the EGS5 (ver. 1.0.4) [Hirayama 2005,

Shimizu et al. 2014]. In total, 2.4 × 107 independent photons were used,

distributed for convenience in 24 files in the format specified in Appendix VI of

the comparison protocol [ADWG(I) 2009]. The multiple use of each photon until

the statistical uncertainty is optimized is also discussed in the protocol.

7/25

The phase-space files are used to calculate Cw,c and the corresponding TPR20,10

(calculated for a detector of radius 3 mm). To test the consistency of these data

and account for any difference between the calculated TPR20,10 and the measured

value, the following procedure is adopted. The calculated values for Cw,c

accumulated so far for the BIPM.RI(I)-K6 comparison series (17 beam qualities in

total, including those for the NMIJ) are plotted as a function of the corresponding

calculated TPR20,10 (see Figure 6). A quadratic fit to these data shows an r.m.s.

deviation within the standard uncertainty of each calculation (typically 0.06 %).

Using this quadratic fit, the value for Cw,c corresponding to the measured TPR20,10

is extracted and used for the BIPM determination of absorbed dose in this beam.

Any significant difference between the measured and calculated values for the

TPR20,10 is included as an uncertainty component. Likewise, any significant

deviation from the fit of the Cw,c values calculated for a particular NMI is also

included as an uncertainty.

We note a minor modification to the above procedure. Until 2014, all calculations

were made using a model for the water transfer chamber calo-6, including its

waterproof sleeve. This chamber failed in 2014 and was replaced by a nominally

identical chamber (calo-5) which, for practical reasons, has a different sleeve

design. For the present comparison, calculations were made for both sleeve

designs and show Cw,c to be reduced by around 0.1 % (less at higher energies).

The results presented graphically in Figure 6 are those for the original sleeve

design, while the values used for the present comparison (given below) are those

for the new sleeve, modified as above for any difference between the measured

and calculated values for the TPR20,10.

The results for Cw,c so-obtained are listed in Table 1 along with the measured and

calculated TPR20,10. The figures in parentheses for the calculated parameters

represent the combined standard uncertainty in the trailing digits based on the

analysis presented in Burns [2014], including components arising from the

simulation geometries, input spectra, radiation transport mechanisms and cross-

section data used. A component of 6·10–4

, 7·10–4

and 8·10–4

has been included for

the beam energies 6 MV, 10 MV and 15 MV, respectively, arising from the

uncertainty in TPR20,10.”The statistical standard uncertainty for each value for Cw,c

is around 0.05 %.

Table 1. Results of the calculations of the conversion factor Cw,c for the BIPM

calorimeter using the Monte Carlo code PENELOPE [Salvat et al. 2009], calculated using

the phase-space files supplied by the NMIJ. Also given are the measured and calculated

values for the TPR20,10. The values in parenthesis represent the combined standard

uncertainty of the calculations based on Burns [2014].

Beam quality Measured TPR20,10 Calculated TPR20,10 Cw,ca

6 MV 0.679(4) 0.684(1) 1.1227 (25)

10 MV 0.729(4) 0.736(1) 1.1309 (29)

15 MV 0.758(5) 0.761(1) 1.1359 (30) a These values for Cw,c include the use of chamber calo-5 in water, which has a different design of

PMMA sleeve from chamber calo-6 used for previous comparisons in the series. Consequently,

the values are lower than those calculated for previous comparisons by 0.12 % at 6 MV, 0.09% at

10 MV and 0.08 % at 15 MV.

8/25

2.3. Measurements in the High-Energy Photon Irradiation Facility at the NMIJ

2.3.1. The NMIJ accelerator and calibration of reference instruments

The comparison was carried out at the NMIJ accelerator facility, housing an

Elekta Precise linear accelerator. All measurements for this comparison were

made with the gantry fixed for horizontal irradiation. The source-to-surface

distance (SSD) was set using a graduated rod from a fixed reference plane to the

centre of the front face of the calorimeter core; the ionization chambers were

aligned with the centre of the beam using two laser systems. The accelerator head

is equipped by a multi-leaf collimator.

One PTW 30013, serial number 6740, was chosen as the NMIJ reference standard

for the realization of the absorbed dose to water. In 2015, the NMIJ and the

ARPANSA carried out an indirect comparison of absorbed dose to water for 60

Co

and high-energy photon beams by calibrating three Farmer-type ionization

chambers as transfer standards [Shimizu et al. 2015]. The selected chamber has

shown good stability and is one of these three chambers.

2.3.2. Experimental set up for the key comparison

The comparison measurements were made in three of the seven photon beams

available on the accelerator: 6 MV, 10 MV and 15 MV, whose measured tissue-

phantom ratios TPR20,10 are given in Table 1. The absorbed dose rate was chosen

around 2 Gy min–1

, the typical dose rate used for calibrations carried out by the

NMIJ, at the depth of maximum dose and the pulse repetition frequency was 200

Hz at 6 MV and 100 Hz at 10 MV and 15 MV.

While the 10 MV and 15 MV qualities are both important for Japanese traceability

in dosimetry, they correspond to the same TPR20,10 interval defined for the

evaluation of Degrees of Equivalence. It was therefore decided before the start of

the comparison that the Degrees of Equivalence for the interval in question should

be determined from the average result at 10 MV and 15 MV.

All measurements were made at 10 g cm–2

in accordance with the Japanese code

of practice for high-energy photons [JSMP 2012]. For the measurements in water,

a cubic PMMA phantom of side length 30 cm was used, positioned at a source-to-

surface distance of 90 cm. The field size in the detector plane was 10 cm 10 cm.

The reference point for each ionization chamber was positioned at a distance of

100 cm from the source and at a depth of 10 g cm–2

; in evaluating this depth, the

density of the 3.81 mm PMMA entrance window and of the water were taken into

account.

The ionization chamber readings were normalized to the reference temperature of

20 °C and pressure of 101.325 kPa chosen for the comparison. No correction was

made for air humidity; the relative humidity in the accelerator laboratory remained

within the interval 20 % to 80 % throughout the comparison measurements. An

irradiation time of 60 s was used for all measurements with the BIPM calorimeter,

the BIPM transfer chambers and the NMIJ reference chamber.

A mechanical table was used as support for the calorimeter phantom and the water

phantom in turn, adjustable in three dimensions with a reproducibility better that

0.2 mm.

9/25

2.3.3. Beam monitoring and measurement system

The indicated dose from a clinical accelerator is not normally sufficiently stable in

time for comparisons of primary standards when used with its internal

transmission monitor alone. For this reason the BIPM used a commercial parallel-

plate transmission chamber to serve as a monitor during irradiation and a

reference class thimble chamber (type NE 2571 serial number 2106) to determine

the stability of the transmission monitor. All ionization chamber current

measurements are ultimately normalized to the reading of the thimble chamber.

The regular use of the thimble monitor chamber to calibrate the transmission

monitor, as well as the need to pre-irradiate all chambers, presents a problem

when the BIPM calorimeter is in place because the calorimeter must be shielded

during these measurements to avoid unnecessary heating of the core. This was

achieved for previous comparisons by manually positioning a tungsten block,

3 cm in thickness, between the calorimeter and the external transmission monitor

when calibrating the transmission monitor or pre-irradiating the chambers. This

block produces significant backscatter into the external monitor, increasing its

response by up to 40 % depending on how close it is to the transmission chamber.

Therefore, to maintain the consistency of the measurements and correctly

determine the ratio Dc/Qc, this shielding block is also positioned during the

corresponding monitoring for the measurement of Qc.

The BIPM transmission and thimble monitor chambers, and since 2014 the

tungsten block, are mounted on a motorized and remote-controlled support, fixed

to a shadow tray. This device allows the thimble chamber and tungsten block to be

moved into the beam when the transmission monitor is to be calibrated, and

moved out again for calorimetry or ionometric measurements. This system allows

for a high reproducibility of positioning and saves time by avoiding repeated entry

into the irradiation area. The support incorporates a probe to measure the

temperature close to the chambers.

The monitoring procedure was as follows. Before, and in some cases after, each

series of measurements made by the BIPM or the NMIJ, the thimble chamber,

with its POM3 build-up cap, was positioned in the beam about 1 cm downstream

from the external transmission monitor and used to calibrate the latter. In effect,

the transmission chamber serves as monitor during each series of measurements

while the thimble chamber is used to transfer the monitoring between series of

measurements. In this way the uncertainty associated with the monitoring is

included in the reproducibility of the repeat measurements for each device, that is,

in the statistical standard uncertainties (see Tables 4, 5 and 6).

It follows that it is the thimble chamber that is used to link the BIPM and

NMIJ dose measurements. That is, over the course of the comparison, this

chamber was calibrated using both the BIPM graphite calorimeter and the NMIJ

transfer standards, which had previously been calibrated against the NMIJ

graphite calorimeter (Section 1.1). The thimble chamber was also used for a

robust determination of the charge ratio Qw/Qc for the BIPM transfer chambers.

3 Polyoxymethylene, commonly referred to as Delrin

10/25

All measurements using the NMIJ transfer standard were made using the NMIJ

measurement system, including the NMIJ air temperature and pressure

measurements. The NMIJ ionization measurements were synchronized with the

BIPM external transmission chamber measurements. For the BIPM transfer

chambers, the BIPM data acquisition system was used, air pressure and

temperature being measured using the BIPM detectors.

2.3.4. Beam Profile

The BIPM calorimeter core is 45 mm in diameter and, depending on the beam

profile, the correction factor for radial non-uniformity kr n,BIPM can be significant.

This is particularly true for clinical accelerators where the uniformity over a

10 cm by 10 cm field can be compromised somewhat so that uniformity

specifications are met for all field sizes. The NMIJ measured the beam profile in

the horizontal and vertical directions for each beam quality. The vertical profiles

of the three beams are shown in Figure 3. The observed non-uniformity of the

10 MV profile is not unusual for clinical accelerator beams.

The beam profiles measured by the NMIJ were used to determine the values for

kr n,BIPM used for this comparison. A cubic spline was fitted to each of the NMIJ

horizontal and vertical profiles. The correction factor kr n,BIPM for a given curve is

determined by dividing the calculated (fitted) response at the centre by the mean

response (weighted by radius) over a diameter of 45 mm. Each value for kr n,BIPM

listed in Table 2 is the mean of the evaluations in the horizontal and vertical

directions. The stated standard uncertainty is an estimate including the difference

between the horizontal and vertical scans.

Table 2. Correction factors for radial non-uniformity, kr n,BIPM, measured for a diameter of

45 mm at a depth of 10 g cm–2

and for a source-to-detector distance of 100 cm. The

standard uncertainty in the last digits is indicated in parenthesis.

Nominal accelerating voltage / MV kr n ,BIPM

6 0.999 9 (10)

10 0.992 0 (35)

15 1.000 4 (13)

11/25

Figure 3. Beam profile in the horizontal (cross-line) measured by the NMIJ at 6 MV,

10 MV and 15 MV. The measured current I is normalized to the dose at the centre.

3. RESULTS AND UNCERTAINTIES

Typically, after setting up a given BIPM device (calorimeter, transfer chamber in

graphite or water) and selecting the radiation quality, the monitors were pre-

irradiated. Then the BIPM thimble monitor was used to calibrate the transmission

monitor, the BIPM device was measured and finally the transmission monitor re-

calibrated. Each measurement series normally involved between seven and ten

measurements. For a given BIPM device, the radiation quality was interchanged

between 6 MV, 10 MV and 15 MV and this cycle was repeated three times before

switching to a new BIPM device. This entire procedure, including setting up, was

repeated for each device at a later date to determine the degree of reproducibility.

Thus in total each quality was measured typically five times for each device. For

calorimetric measurements, when possible, the calorimeter was set up on a Friday

evening to benefit from the weekend for temperature stabilization and to obtain an

adequate vacuum.

12/25

3.1. Estimation of Uncertainties

3.1.1. Uncertainty in the determination of Dw,NMIJ

The relative standard uncertainties for the NMIJ determination of the calibration

coefficient ND,w,NMIJ of one selected reference ionization chamber used as a

transfer standard, based on measurements in water using the NMIJ primary

standard graphite calorimeter, are outlined in Tables 3-a, 3-b and 3-c. The

combined standard uncertainty of the determination of Dw,NMIJ, normalized to the

thimble monitor charge Qth, is given in Table 4. The Type A uncertainty,

attributed to reproducibility, is the standard deviation of the mean of 6 runs for

each chamber at each beam quality. Each run consisted of 10 exposures of 60 s

duration.

3.1.2. Uncertainties in the determination of Dw,BIPM

The uncertainties for the determination of Dc, normalized to the thimble monitor

charge Qth, are listed in Table 5. The type A standard uncertainty of Dc/Qth

depends on the time devoted to measurements at the NMI, the temperature

stability of the NMI laboratory and the stability of the beam and beam monitoring.

The uncertainties associated with the determination of the ratio Qw/Qc are listed in

Table 6. The chamber orientation corrections noted in Table 6 were determined in

the BIPM 60

Co reference beam. The polarity effect for the parallel plate chambers

was determined in a previous work where the correction factor for recombination

losses in pulsed beams was determined for the BIPM transfer chambers [Picard et

al. 2011c].

Table 3-a. Uncertainties associated with the determination of Dc,NMIJ using the

NMIJ graphite calorimeter.

Uncertainty component uA(y)/y / 10–3

uB(y)/y / 10–3

Measured Dc,NMIJ 0.84 0.51

Mass of graphite core 0.003

Heat defect 1.0

Impurity effect 0.5

SDD, graphite calorimeter 0.12

Graphite depth 0.49

Beam monitor stability 1.0

Temperature and pressure correction 0.22

Humidity correction 0.2

SDD, ionization chamber 0.12

Water depth 0.29

Radial and axial non-uniformity correction 1.0

Total uncertainty 0.84 0.20

Combined relative standard uncertainty 2.2

13/25

Table 3-b. Uncertainties associated with the determination of Qw,NMIJ/Qc,NMIJ.

Uncertainty component uA(y)/y / 10–3

uB(y)/y / 10–3

Measured Qw,NMIJ/Qc,NMIJ 1.0 0

SDD in graphite phantom 0.12

Graphite depth 0.49

SDD in water phantom 0.12

Water depth 0.49

Beam monitor stability 1.0

Monte Carlo calculation of the conversion

factor 2.0

Total uncertainty 1.0 2.3

Combined relative standard uncertainty 2.6

Table 3-c. Uncertainties associated with the determination of calibration

coefficient ND,w of the NMIJ transfer standards.

Uncertainty component uA(y)/y / 10–3

uB(y)/y / 10–3

Dc,NMIJ 0.84 2.0

Qw,NMIJ/Qc,NMIJ 1.0 2.3

Total uncertainty 1.3 3.1

Combined relative standard uncertainty 3.4

Table 4. Standard uncertainty components for the determination of Dw,NMIJ in the 6 MV,

10 MV and 15 MV accelerator photon beams. The data are derived using the external

transmission monitor.

Type A relative standard uncertainty component uA(y)/y / 10–3

6 MV 10 MV 15 MV

Qw,NMIJ (n = 4) 0.2 0.2 0.3

ND,w (Table 3-c) 1.3 1.3 1.3

Type B relative standard uncertainty component uB(y)/y / 10–3

6 MV 10 MV 15 MV

ND,w (Table 3-c) 3.1 3.1 3.1

positioning 0.5 0.5 0.5

temperature and pressure correction 0.4 0.4 0.4

Combined relative standard uncertainty [uc(y)/y] / 10–3

: 3.4 3.4 3.4

14/25

Table 5. Standard uncertainty components for the determination of Dc/Qth at 6 MV,

10 MV and 15 MV. The statistical uncertainty is based on n = 5 determinations for each

radiation quality, each one the result of 10 irradiations of 60 s. The data are derived using

the external transmission monitor.

Type A relative standard uncertainty component uA(y)/y / 10–3

6 MV 10 MV 15 MV

typical standard uncertainty of the mean (n = 5)

including the transfer using the external transmission monitor

0.6 0.6 0.6

Type B relative standard uncertainty component uB(y)/y / 10–3

specific heat capacity of graphite [Picard et al. 2007] 0.9

impurity correction 0.2

temperature calibration 0.5

linear model for temperature extrapolation 0.7

axial position of calorimeter 0.5

Combined relative standard uncertainty [uc(y)/y] / 10–3

: 1.5 1.5 1.5

Table 6. Standard uncertainty components for the determination of Qw/Qc at 6 MV,

10 MV and 15 MV. The statistical uncertainty is based on n = 5 determinations for each

radiation quality, each one the result of 7 irradiations of 60 s. The data are derived using

the external transmission monitor and the thimble monitor.

Type A relative standard uncertainty component uA(y)/y / 10–3

6 MV 10 MV 15 MV

typical standard uncertainty of the mean (n = 5)

including the transfer using the external transmission monitor

0.3 0.3 0.3

Type B relative standard uncertainty component uB(y)/y / 10–3

difference in graphite jackets of core and transfer chamber 0.1

chamber orientation for Qc 0.1

chamber orientation for Qw 0.4

volume correction kV for the BIPM chambers 0.3

temperature and pressure correction 0.3

axial position of chamber 0.7

Combined relative standard uncertainty [uc(y)/y] / 10–3

: 1.0 1.0 1.0

15/25

3.2. Combined Uncertainty

The significant uncertainties in the determination of Dw,NMIJ/Qth and Dw,BIPM/Qth in

high-energy photon beams are listed in Tables 1, 2, 4, 5 and 6. The combined

uncertainties are listed in Table 7.

Table 7. Standard uncertainty components in the determination of the comparison ratio

Dw,NMIJ/Dw,BIPM.

Relative standard uncertainty component u(y)/y / 10–3

6 MV 10 MV 15 MV

calibration of thimble monitor in terms of Dw,NMIJ (Table 4) 3.4 3.4 3.4

calibration of thimble monitor in terms of Dc,BIPM (Table 5) 1.5 1.5 1.5

Qw/Qc for BIPM chambers in NMIJ beams (Table 6) 1.0 1.0 1.0

Cw,c for BIPM standard in NMIJ beams (Table 1) 2.5 2.9 3.0

kr n,BIPM for BIPM standard in NMIJ beams (Table 2) 1.0 3.5 1.3

Combined relative standard uncertainty [uc(y)/y] / 10–3

: 4.7 6.0 5.0

3.3. Determination of Dc, Qw, and Qc

The absorbed dose to graphite Dc was obtained by taking the mean of the

temperature rises detected by two thermistor bridges (the third bridge was not in

operation at the time of the comparison) and applying equation (5) with its correction

for impurities, kimp = 1.0004.

The mean values of Dc, Qc and Qw, each normalized to the thimble monitor charge

Qth, are listed for each beam quality in Table 8 along with the statistical standard

uncertainties in parentheses. For practical reasons, the BIPM method to convert to

absorbed dose to water uses different transfer chambers in water and in graphite.

While nominally identical, the two chambers have slightly different volumes and an

appropriate correction is made to the measured charge ratio. The stated values for Qc

and Qw in Table 8 are normalized to the reference air density and Qw is corrected for

the difference in volume (kV = 0.9955).

16/25

Table 8. Experimental results obtained for the BIPM calorimeter and transfer chambers,

normalized to the thimble monitor charge, Qth. Values in parentheses represent the standard

uncertainty of the mean.

Nominal accelerating voltage / MV Dc/Qth

[Gy µC –1

]

Qc/Qth

Qw/Qth

6 16.630 (9) 4.104 2 (11) 3.978 6 (3)

10 18.377 (11) 4.563 9 (13) 4.420 5 (2)

15 19.526 (12) 4.887 2 (12) 4.736 9 (7)

Table 9. Absorbed dose to water determined by the BIPM, Dw,BIPM. The dose values are

given relative to the charge, Qth, measured by the thimble monitor using the external

transmission monitor. Also given are the determined values for Dc/Qc, Qw/Qth and kr n,BIPM as

well as the total BIPM Monte Carlo conversion factors Cw,c.

Nominal accelerating

voltage / MV

Dc/Qc

[Gy µC–1

]

Qw /Qth

kr n,BIPM Cw,c

Dw,BIPM/Qth

[Gy µC–1

]

6 4.051 9 3.978 6 0.9999 1.1227 18.097

10 4.026 6 4.420 5 0.9920 1.1309 19.968

15 3.995 3 4.736 9 1.0004 1.1359 21.507

3.4. Comparison Results

The NMIJ determination of the absorbed dose to water was transferred to one

transfer chamber of the NMIJ in the form of a calibration coefficient for each beam

quality, determined under the reference conditions for the comparison. The use of

this chamber in the BIPM water phantom under the same reference conditions

resulted in the determination of Dw,NMIJ used for the comparison result. Tables 10

and 11 summarize the measurement results for the BIPM and the NMIJ.

Table 10. Absorbed dose to water determined by the NMIJ, Dw,NMIJ and the BIPM, Dw,BIPM,

and the quotient Dw,NMIJ/Dw,BIPM. The dose values are given relative to the charge, Qth,

measured by the thimble monitor using the external transmission monitor. The comparsion

result R and associated uncertainty is also listed for each radiation quality.

Nominal

accelerating

voltage /MV

Measured

TPR20,10

Dw,NMIJ/Qth

[Gy µC–1

]

Dw,BIPM/Qth

[Gy µC–1

]

Dw,NMIJ/ Dw,BIPM

= R

uc(R)/R

6 0.679 18.035 18.093 0.9966 0.0047

10 0.729 19.898 19.963 0.9965 0.0060

15 0.758 21.404 21.503 0.9953 0.0050

17/25

Figure 4. Results of the 2015 comparison at high energies of the calorimetric

absorbed dose to water standards of the NMIJ and the BIPM, shown as a function

of the measured TPR20,10. The uncertainty bars represent the combined standard

uncertainty of each comparison result.

4. DISCUSSION

All measurements were carried out under the reference conditions normally used by

the NMIJ [JSMP 2012]. The results for all three beam qualities are within one

standard uncertainty. The comparison reported here presented no technical problems.

The calculated horizontal and vertical values for the non-uniformity correction were

similar at all three beam qualities and give rise to relatively small uncertainties except

at 10 MV. However, given the limited data available to characterize beam uniformity

in two dimensions and its potential variation over the duration of the comparison, the

stated uncertainties for kr n might be under estimated.

The correction for recombination losses in pulsed beams was determined for the

BIPM transfer chambers in a separate series of measurements [Picard et al. 2011c].

While this effect will cancel for the measured ratio Qc/Qw, the correction of Qc for

recombination allows the calibration coefficient ND,c measured at different NMIs

using the same chamber (calo-3) to be directly compared. These results, shown in

Figure 5, serve as an additional check on the BIPM measurements during a

comparison and include the National Research Council (NRC) Canada, the

Physikalisch-Technische Bundesanstalt (PTB) Germany, the National Institute of

Standards and Technology (NIST) USA, the Laboratoire National de Métrologie et

d’Essais – Laboratoire National Henri Becquerel (LNE-LNHB) France, the

Australian Radiation Protection and Nuclear Safety Agency (ARPANSA) Australia,

the National Physical Laboratory (NPL) UK and the NMIJ (present work), all

18/25

determined at a depth of 10 g cm–2

[Picard et al. 2010, 2011b, 2013b, 2013c, 2014,

2015b].

Monte Carlo calculations of the conversion factor Cw,c for different beam qualities

have now been made for the NRC, PTB, NIST, LNHB, ARPANSA, NPL and the

NMIJ (present work), all determined at 10 g cm–2

[Picard et al. 2010, 2011b, 2013b,

2013c, 2014, 2015b]. The results are shown in Figure 6 along with a weighted

quadratic fit to the data. The deviations about this line are consistent with the typical

statistical standard uncertainty of 5 parts in 104. As noted in Section 2.2.2, these

results are for calo-6 and its waterproof sleeve, while the values used for the present

comparison (and the previous comparison with the NPL) are for calo-5 and its sleeve.

Figure 5. The measured calibration coefficient ND,c = Dc/Qc, corrected for

recombination, as a function of measured TPR20,10. The solid line traces a weighted

quadratic fit to the data.

19/25

Figure 6. The dose conversion factor Cw,c for the BIPM standard, calculated using the

phase-space files supplied by the participating NMIs. The line is a weighted quadratic fit to

the data; the deviations about this line are consistent with the typical statistical uncertainty

of 5 parts in 104. Note that these are the results using a model for the original water transfer

chamber (calo-6) and its waterproof sleeve. For the present comparison, the values given in

Table 1 were calculated using a model for the new chamber (calo-5), the main difference

being the sleeve design.

The results of the present comparison are very similar for all three beam qualities.

The accumulated comparison results obtained to date are shown in Figure 7. The

uncertainty bars represent the standard uncertainty of each comparison and are

therefore correlated through the common use of the BIPM standard (as well as any

correlation that might exist between any of the NMI standards).

It is noted that during the comparison a second ionization chamber of the NMIJ was

included in the measurements, but it was considered unstable by the NMIJ and

therefore discarded at an early stage. However, the analysis of the results from that

chamber, although not included in the final analysis, show consistent results with

those presented.

The chamber of the NMIJ (PTW 30013, serial number 6740), used in this

comparison as the NMIJ reference standard for the realization of the absorbed dose

to water, was also used for an indirect comparison of the NMIJ and ARPANSA

[Shimizu et al. 2015]. Table 12 lists the comparison data of NMIJ and ARPANSA at

TPR20,10 ≈ 0.679 (6 MV) and TPR20,10 ≈ 0.729 (10 MV). The absorbed dose to water

determined by the BIPM has been compared to that of the NMIJ [this work] and

20/25

ARPANSA [Picard et al. 2014] at approximately the same values of TPR20,10. The

ratios of the results of these are listed in the third column of Table 12. It can be

observed that the comparison results are very coherent..

Table 12. Comparison data of NMIJ/ARPANSA, and the results

obtained by combining the BIPM.RI(I)-K6 comparison results of NMIJ

and ARPANSA, respectively.

TPR20,10 NMIJ/ARPANSA

[Shimizu et al.2015]

[NMIJ/BIPM]/ [ARPANSA/BIPM]

[Picard et al. 2014, this work]

≈ 0.679 0.9995 1.0001

≈ 0.729 1.0040 1.0041

Figure 7. Results of eight comparisons of accelerator dosimetry to date, reported as a ratio of

the NMI and the BIPM evaluations of absorbed dose to water plotted as a function of the

measured TPR20,10.4 The results for the NMIJ (this work) are indicated as purple diamonds.

The uncertainty bars represent the combined standard uncertainty of each comparison result.

4 The results of the NPL 2013 comparison were at 5 g cm

–2 and 7 g cm

–2, while those of NPL 2014 were at

10 g cm–2

. It should be noted that the NPL results at 10 g cm–2

are not included in the list of Degrees of

Equivalences as the disseminated standard still uses the depths of 5 g cm–2

and 7 g cm–2

as defined in the

existing UK protocol [Code of Practice 1990].

21/25

In relation to the BIPM KCDB, the purpose of the BIPM.RI(I)-K6 comparison is to

provide information on equivalence and traceability for absorbed dose to water in

accelerator beams.

The full range of TPR20,10 values in normal use at the NMIs has been divided into three

ranges, within each of which NMIs may obtain a degree of equivalence with respect to

the key comparison reference value. The 10 MV and 15 MV qualities are both

important for Japanese traceability in dosimetry, but correspond to the same TPR20,10

range. Consequently, the degree of equivalence for this range has been determined

from the weighted average result at 10 MV and 15 MV (cf. Appendix).

5. DEGREES OF EQUIVALENCE

Following a decision of the CCRI, the BIPM calorimetric determination of absorbed

dose to water as determined by the BIPM on the NMI site is taken as the key

comparison reference value [CCRI 2013, Picard et al. 2013a]. It follows that for each

NMIi having a BIPM comparison result Ri, with a combined standard uncertainty ui,

the degree of equivalence with respect to the KCRV is given by a pair of terms Di

and Ui:

Di = Ri – 1 , (8)

Ui = 2 ui . (9)

These terms represent the relative difference and the expanded uncertainty (k = 2) of

this difference.

The degrees of equivalence, Di and Ui, for each of the NMIs having participated until

December 2015 are listed and represented graphically in the Appendix as they will

appear in the KCDB. As noted above, for the NMIJ the 10 MV and 15 MV beam

qualities belong to the same TPR20,10 interval (from 0.71 to 0.77) and the

corresponding degree of equivalence is therefore calculated as the average (cf.

Appendix).

6. CONCLUSION

This comparison is the ninth in the on-going BIPM key comparison BIPM.RI(I)-K6.

The results, reported as ratios of the NMIJ and the BIPM evaluations (and with the

combined standard uncertainties given in parentheses), are 0.9966 (47) at 6 MV,

0.9965 (60) at 10 MV and 0.9953 (50) at 15 MV, and are available in the BIPM key

comparison database [KCDB].

22/25

REFERENCES

Burns D T 2006 A new approach to the determination of air kerma using primary-

standard cavity ionization chambers Phys. Med. Biol. 51 929-942

Burns D T 2015 The dose conversion procedure for the BIPM graphite calorimeter

standard for absorbed dose to water, in preparation

ADWG(I) 2009 Calorimetric Comparison of Absorbed Dose to Water at High

Energies. BIPM Key Comparison BIPM.RI(I)-K6. Protocol 1.65

Accelerator

Dosimetry Working Group of the CCRI(I)

CCRI 2013 Consultative Committee for Ionizing Radiation, Report of the 24th

meeting (2013)

Code of Practice for high-energy photon therapy dosimetry based on the NPL

absorbed dose calibration service 1990 Institute of Physical Sciences in Medicine

Phys. Med. Biol. 35 1355-60

Hirayama H, Namito Y, Bielajew A F, Wilderman S J and Nelson W R, SLAC-R-

730 (2005) and KEK Report 2005-8 (2005)

JSMP 2012: Japan Society of Medical Physics 2012 Standard dosimetry of

absorbed dose to water in external radiotherapy (Standard dosimetry 12)

Kawrakow I, Mainegra-Hing E, Rogers D W O, Tessier F and Walters B R B 2011

The EGSnrc code system: Monte Carlo simulation of electron and photon transport

NRCC Report PIRS-701 National Research Council of Canada, Ottawa

KCDB The BIPM Key Comparison Database is available online at

http://kcdb.bipm.org

Morishita Y, Kato M, Takata N, Kurosawa T, Tanaka T and Saito N 2013 A

standard for absorbed dose rate to water in a 60

Co field using a graphite calorimeter

at the National Metrology Institute of Japan, Radiat. Prot. Dosim. 154 331-9

Picard S, Burns D T and Roger P 2007 Determination of the Specific Heat Capacity

of a Graphite Sample Using Absolute and Differential Methods Metrologia 44 294-

302

Picard S, Burns D T and Roger P 2009 Construction of an Absorbed-Dose Graphite

Calorimeter Rapport BIPM-2009/01 (Sèvres: Bureau International des Poids et

Mesures) 12 pp.

Picard S, Burns D T, Roger P, Allisy-Roberts P J, McEwan M, Cojocaru C, Ross C

2010 Comparison of the standards for absorbed dose to water of the NRC and the

BIPM for accelerator photon beams Metrologia 47 Tech. Suppl. 06025, 22 pp.

Picard S, Burns D T, Roger P 2011a The BIPM Graphite Calorimeter Standard for

Absorbed Dose to Water, abstract to International Symposium on Standards,

Applications and Quality Assurance in Medical Radiation Dosimetry in Standards,

5 This is the version available at the time of the comparison. An updated version is available on the KCDB

web site.

23/25

Applications and Quality Assurance in Medical Radiation Dosimetry (IDOS),

2010, vol. 1 55-65, Proceedings Series – International Atomic Energy Agency 2011

Picard S, Burns D T, Roger P, Allisy-Roberts P J, Kapsch R P and Krauss A 2011b

Key comparison BIPM.RI(I)-K6 of the standards for absorbed dose to water of the

PTB, Germany and the BIPM in accelerator photon beams Metrologia 48 Tech.

Suppl. 06020, 21 pp.

Picard S, Burns D T and Ostrowsky A 2011c Determination of the correction factor

for recombination losses of a BIPM ionization chamber of standard design in a

pulsed photon beam Rapport BIPM-2011/06 (Sèvres: Bureau International des

Poids et Mesures) 9 pp.

Picard S, Burns D T and Los Arcos JM 2013a Establishment of degrees of

equivalence of national primary standards for absorbed dose to water in accelerator

photon beams Metrologia 50 Tech. Suppl. 060016, 9 pp.

Picard S, Burns D T, Roger P, Bateman F B, Tosh R E and Chen-Mayer H 2013b

Key comparison BIPM.RI(I)-K6 of the standards for absorbed dose to water of the

NIST, USA and the BIPM in accelerator photon beams Metrologia 50 Tech. Suppl.

06004, 22 pp.

Picard S, Burns D T, Roger P, Delaunay F, Le Roy M, Ostrowsky A, Sommier L

and Vermesse D 2013c Key comparison BIPM.RI(I)-K6 of the standards for

absorbed dose to water of the LNE-LNHB, France and the BIPM in accelerator

photon beams Metrologia 50 Tech. Suppl. 06015, 24 pp.

Picard S, Burns D T, Roger P, Harty P D, Ramanathan G, Lye J E, Wright T, Butler

D J, Cole A, Oliver C and Webb D V D 2014 Key comparison BIPM.RI(I)-K6 of

the standards for absorbed dose to water of the ARPANSA, Australia and the BIPM

in accelerator photon beams Metrologia 51 Tech. Suppl. 06006, 29 pp.

Picard S, Burns D T, Roger P, Duane S, Bass G A, Manning J W and Shipley D R

2015a Key comparison BIPM.RI(I)-K6 of the standards for absorbed dose to water

at 5 g cm–2

and 7 g cm–2

of the NPL, United Kingdom and the BIPM in accelerator

photon beams, Metrologia, 52, Tech. Suppl., 06010 27 pp.

Picard S, Burns D T, Roger P, Duane S, Bass G A, Manning J W and Shipley D R

2015b Key comparison BIPM.RI(I)-K6 of the standards for absorbed dose to water

at 10 g cm–2

of the NPL, United Kingdom and the BIPM in accelerator photon

beams, Metrologia, 52, Tech. Suppl., 06021 23 pp.

Picard S, Burns D T, Roger P, de Prez L. and Jansen B 2015c Key comparison

BIPM.RI(I)-K6 of the standards for absorbed dose to water of the VSL,

Netherlands and the BIPM in accelerator photon beams, in preparation.

Rogers D W O, Walters B R B and Kawrakow I 2011 BEAMnrc User Manual

NRCC Report PIRS-0509(A) National Research Council of Canada, Ottawa

Shimizu M, Morishita Y, Kato M, Kurosawa T, Tanaka T, Takata N and Saito N,

2014 Calculation of the characteristics of clinical high-energy photon beams with

EGS5-MPI J. Phys. Conf. series 489 012023

24/25

Shimizu M, Morishita Y, Kato M, Tanaka T, Kurosawa T, Takata N, Saito N,

Ramanathan G, Harty P D, Oliver C, Wright T and Butler D J, 2015

Comparison of the NMIJ and the ARPANSA standards for absorbed dose to

water in high-energy photon beams, Radiat. Prot. Dosimetry 164 181-186

APPENDIX

Graph of degrees of equivalence Di with the KCRV for each of the NMIs having participated until end of 2015

Key comparison BIPM.RI(I)-K6

MEASURAND : Absorbed dose to water

The key comparison reference value is the BIPM determination of absorbed dose to water in an accelerator beam

The degree of equivalence of each laboratory i with respect to the key comparison reference value is given by a pair of terms:

the relative difference D i and U i = 2u i , its expanded uncertainty (k = 2), both expressed in mGy/Gy.

When required, the degree of equivalence between two laboratories i and j can be evaluated by a pair of terms:

D ij = D i - D j and U ij = 2u ij , its expanded uncertainty (k = 2), both expressed in mGy/Gy.

In evaluating u ij , account should be taken of correlation between u i and u j .

25/25

Degrees of equivalence Di