Chapter 16

Ionisation Chambers

Bryan Muir and Alan Nahum*

CONTENTS
16.1 Introduction........................................................................................................................................................... 321
16.2 Physical and Operational Principles......................................................................................................................... 322
16.2.1 Influence of the Applied Voltage – the Ionisation Region............................................................................ 322
16.2.2 Chamber Design......................................................................................................................................... 323
16.2.3 Insulators and Guard Rings......................................................................................................................... 324
16.2.4 Electrometers.............................................................................................................................................. 324
16.2.5 Cables and Connectors................................................................................................................................ 325
16.3 Types of Chambers................................................................................................................................................. 326
16.3.1 Free-Air Chambers...................................................................................................................................... 326
16.3.2 Cavity Chambers......................................................................................................................................... 327
[Link] Desirable Properties of Cavity Chambers – Perturbation Corrections............................................ 328
[Link] Cylindrical Chambers.................................................................................................................... 329
[Link] Plane-Parallel Chambers................................................................................................................ 329
16.3.3 Transmission Monitor Chambers................................................................................................................. 332
16.4 Determination of the Charge Produced.................................................................................................................. 333
16.4.1 Recombination............................................................................................................................................ 333
[Link] Initial and General Recombination................................................................................................ 333
[Link] Collection Efficiency: Boag’s Theory............................................................................................. 333
[Link] Collection Efficiency: the Two-Voltage Method............................................................................. 335
16.4.2 Polarity Effect............................................................................................................................................. 336
16.4.3 Stabilisation and Pre-Irradiation.................................................................................................................. 336
16.4.4 Leakage and Stem Effect............................................................................................................................. 337
16.4.5 Correction for the Changes in the Mass of Air within the Cavity................................................................. 337

16.1 Introduction standard 60731 (IEC 2011) distinguishes between three dif-

ferent classes of measuring instruments based on the use of
Ionisation is the creation of ion pairs. This results from inter- ionisation chambers†:
actions that transfer energy from the radiation to the atomic
system. For energy transfers greater than the electron bind- • Reference-class instruments
ing energies, electrons are ejected (see Section 1.4.2), leaving • Field-class instruments
behind a positively charged ion. The ionisation of a gas was
• Scanning-class instruments.
one of the first properties of ionising radiation to be observed;
it is the defining property. The measurement of ionisation in a The differences among these three classes are related to their
gas was also the basis for the first radiation detectors. performance requirements and recommended applications.
Ionisation (or ion) chambers have remained central not Reference-class instruments have the most stringent require-
only to the detection of radiation but also to the precise ments; they are appropriate for the measurement of reference
determination of absorbed dose. More is known about the (or absolute) dose, also referred to as calibration of a radia-
response of ionisation chambers than about any other radi- tion beam (see Chapter 19). They can also be used for cross-
ation-detecting instrument. They are the principal instru- calibrating field-class instruments. Field-class instruments are
ments in the radiotherapy clinic for both reference and used for the measurement of air kerma or absorbed dose in
relative dose measurements (see Chapter 19 and Chapter 20, a clinical radiation beam as part of routine measurements
respectively); the International Electrotechnical Commission
† Since the ionisation chambers are the main components of these
* The contributions of Pedro Andreo and David Thwaites, co-authors instruments, similar classification can also apply to the chambers as
of this chapter in the 1st Edition, are acknowledged. such (McEwen 2010).

321
322 Part D: Dose Measurement

for quality assurance. Scanning-class instruments are typi- • The chamber cavity volume, V
cally used for relative dose distribution measurements with a • The ratio of the (mass electronic) stopping power
scanning system (see Section [Link]). in water to that in air, swater,air, which is principally
Because of the key role that ionisation chambers play in dependent on the radiation quality†
radiotherapy dosimetry, it is vital that practising medical • The mean energy required to produce an ion pair,
physicists and others involved in radiation measurements in divided by the electronic charge, (Wair/e), which
radiotherapy beams have a thorough knowledge and under- for photon or electron radiation has the value of
standing of their characteristics. As will be made clear in 33.97 eV per ion pair (or joules per coulomb) (see
what follows, although widely used and well understood, Section 19.4.3) and is independent of radiation
ionisation chambers are far from simple in their response or quality except at energies well below the range of
in their mode of operation. interest in radiotherapy (ICRU 1979, 2016).
For the clinical use of radiation, we are interested in the
dose to biological tissue, which consists mostly of water. In For relative dose measurements, the dose at a given location
an ionisation chamber, the sensitive medium is a gas, usually in given conditions is compared with the dose at a different
air at atmospheric pressure, with a density around 800 times location and/or in different conditions. For such measure-
less than that of water. Hence, the signal strength from an ments, since rair, V and (Wair/e) cancel out, the dose ratio is
ionisation chamber will always be relatively low compared equal to the ratio of the ionisation chamber responses, cor-
with so-called solid-state detectors (see Chapters 17 and 18). rected, if necessary, by any modification of s water, air or any dif-
The quantity measured, electrical charge, has first to be ference between the produced and the collected charge.
related to dose to the gas, Dg (or Dair in practice); cavity the- The collected charge Q can usually be determined with
ory (Section 5.7) is then employed to derive the dose to the high accuracy, but a correction for recombination losses may
medium, Dmed. be required to determine the charge liberated (see Section
According to Equation 5.49, the dose in water is related to 16.4.1). For absolute dosimetry, a great deal of attention has
the dose in the air of a (small) cavity by the expression: been paid to the accurate computation of stopping-power
D water = Dair s water, air (16.1) ratios (see Section 19.4.1). The determination of the cham-
ber volume with sufficient accuracy is the major practical dif-
where s water, air is the water-to-air mass stopping-power ratio. ficulty; this is the principal reason why ionisation chambers
As explained in Section 5.7.6, Equation 16.1 implies that used for reference dose determination in radiotherapy, where
the medium – water – is undisturbed by the presence of the accuracy of the order of 1% or better is required, are provided
cavity. Since this is not completely correct, a correction or with an instrument- and radiation-quality-specific calibra-
perturbation factor must be applied. This is further addressed tion factor, which automatically accounts for the chamber
in Section [Link]. volume. For a calibration performed in terms of absorbed dose
The mean absorbed dose in the air of the cavity is related to water per measured charge, other factors are also included
to the charge created, Q, by the expression: (see Section [Link]).
This chapter begins with an introduction to the physical
Dair =
(Q /e )Wair (16.2)
and operational principles behind gas-filled ionisation cham-
mair bers (Section 16.2). The different types of ionisation cham-
bers and their applications are described in Section 16.3.
where
Finally, Section 16.4 describes how the charge produced in
Q /e is the number of ions created in the air cavity (e being
the ionisation chamber is determined and the corrections
the electron charge)
required to accurately derive the absorbed dose from the
Wair is the mean energy required to produce an ion pair in
reading of the ionisation chamber dosimetry system.
dry air
mair is the mass of air:

Combining Equations 16.1 and 16.2, and replacing the

mass of air by the product of the cavity volume V and the air
16.2 Physical and Operational Principles
mass density rair, we obtain; 16.2.1 Influence of the Applied Voltage – the
Q Ionisation Region
D water = (Wair e ) s water, air (16.3)
rairV Ionisation chambers are, in principle, the simplest of all gas-
filled detectors. They consist of a gas-filled cavity with two
This demonstrates that an absolute determination of the dose to
electrodes between which a voltage is applied. As the gas is
water (in Gy) with an ionisation chamber is achievable provided
theoretically a perfect insulator, no electric current will flow
that each of the quantities in Equation 16.2 is known accurately:
between the electrodes. The interactions of the radiation
• The charge Q liberated during the radiation exposure
• The density of the air* in the chamber cavity, rair † The quantity s water,air exhibits a weak dependence on the chamber
properties through the size of the energy cutoff parameter Δ, which
* This derivation is strictly valid only for dry air; in practice, a small is related to the dimensions of the air cavity in the Spencer–Attix for-
correction for relative humidity is required (see Section 16.4.5). mulation of the stopping-power ratio (see Section 5.7.5).
Chapter 16: Ionisation Chambers 323

with the gas molecules give rise to (negative) electrons and This typical behaviour, in addition to the possible elec-
ions with a positive charge, i.e. the gas is ionised; the electric trical breakdown of insulators or distortion of the detec-
field (due to the applied voltage) causes the electrons and tor volume, results in a practical upper limit of the applied
positive ions to move towards the positive and negative elec- voltage, corresponding to the position of the first left-hand
trode, respectively, thereby creating a current that is mea- vertical dashed line in Figure 16.1. A slightly lower limit, at
sured with an electrometer. which the chamber can be safely used, is generally specified
At low voltages, the charge collected from the gas is rela- in the manufacturer’s recommendations. Typical voltage val-
tively small, because the electrons and the positive ions travel ues for ionisation chambers used in radiotherapy dosimetry
relatively slowly towards the electrodes; therefore, their are in the range 200 V to 400 V, although some parallel-
mutual attraction gives rise to considerable ion recombina- plate chambers with a very small distance between electrodes
tion. As the voltage is increased, the signal increases because can require values lower than this (see, for instance, IPEM
the velocity of the ions increases, and consequently, a greater 2003); this also applies to certain small-diameter cylindrical
proportion of the charge is collected before it can recom- chambers.
bine. Finally, the signal reaches a saturation value, where
100% of the ion pairs created by the fast charged particles
16.2.2 Chamber Design
are collected;* the detector is then said to operate in the
ionisation chamber region. Increasing the voltage to values Multiple designs of ionisation chambers have been devel-
beyond this region causes the electrons produced in the gas oped for use in different situations. These are discussed in
to be accelerated to such an extent that they give rise to addi- Section 16.3, where it is shown that the ionisation chamber
tional ionisation of the gas, thereby increasing the charge is a highly versatile radiation detector, capable of operating
collected beyond that corresponding to Isat (see Figure 16.1). over a wide range of radiation intensity (or fluence rate). For
Successive increases in the applied voltage cause the detector the determination of absorbed dose to water in radiother-
to operate successively in the proportional region, limited pro- apy beams, the most common types are cylindrical (thim-
portionality region, Geiger region and continuous discharge ble) and plane-parallel (parallel-plate or coin) chambers,
region (Figure 16.1). with the gas almost always being air (see Section 16.3.2).
The international standard 60731 from the International
Electrotechnical Commission on dosimeters with ionisation
chambers as used in radiotherapy and its amendment 1 (IEC
Continuous

2011) provide recommended limits for the performance

discharge
Geiger region

characteristics of these devices. This IEC standard defines

an ionisation chamber as ‘a detector consisting of a cham-
Limited proportionality

ber filled with air, in which an electric field insufficient to

produce gas multiplication is provided for the collection at
the electrodes of charges associated with the ions and the
electrons produced in the measuring volume of the detec-
Proportional region

tor by ionising radiation’. The term dosimeter describes the

Ionisation chamber region

entire charge measurement apparatus, including chamber,

electrometer and cables – although some manufacturers use
Ion collection

this term solely to refer to the electrometer. IEC 60731 stan-

dard includes in the term radiotherapy dosimeter not only the
chamber and measuring assemblies but also stability check
sources and phantoms.
The most frequently used ionisation chambers are cylin-
drical. They consist of the measuring volume, the central
collecting electrode and the outer electrode (which con-
sists of the chamber wall and possibly a conductive coat-
Isat ing). A build-up cap or waterproof housing may be added
if needed. It is clear from Equation 16.3 that for a given
dose, the charge (in integrating mode) and therefore the cur-
rent intensity (in rate mode) will be directly proportional to
Voltage the chamber volume. From this equation, the sensitivity of
ionisation chambers can be calculated; for the photon and
FIGURE 16.1 Variation of the charge collected (or current in rate
electron energies currently used in radiotherapy, it is around
mode), as a function of the applied voltage, in a gas detector. The dif-
ferent operating regions of the detector are indicated; the plateau cor- 40 nC Gy−1 cm−3. The collecting volumes of cavity chambers
responds to the so-called ionisation chamber region, where a saturation used for radiotherapy dosimetry are typically in the range
current Isat is achieved. 0.03 cm3 to 1 cm3, exceptionally down to 0.003 cm3 (see
Section 16.3.2). For dose rates that can vary from 0.1 Gy min−1
* The deviations from perfect charge collection are discussed in to 10 Gy min−1, the typical currents created from the ionisa-
Section 16.4.1. tion of the air molecules in the chamber cavity are in the
324 Part D: Dose Measurement

range 6 pA to 2 nA. Such extremely small currents are dif- current adds to the signal due to the radiation. This leakage is
ficult to measure and require high-quality connectors and increased by the presence of moisture or small dirt particles;
amplifiers. Therefore, since ionisation chambers are intended this is the reason why insulators should not be breathed on
to determine the absorbed dose ‘at a point’ and may be used or touched with the fingers. To minimise the voltage across
in high-dose-gradient regions, depending on the practical the insulator and therefore, the leakage current, the insulator
application, a compromise must be found between spatial is usually segmented into two parts with a conducting guard
resolution and sensitivity. In radiation protection, where the ring; one part separates the ring from the negative electrode,
dose rates are much lower and where high spatial resolution and another separates it from the positive electrode (see
is not required, the volume of the type of ionisation cham- Figure 16.2a). The guard ring extends through the insula-
ber employed is typically several hundred cubic centimetres; tor assembly in the chamber stem and connects to the inner
even with such large volumes, dose rates need to be at least braid in the cable. The guard and the collector are held at the
10 μGy h−1 to be determined with sufficient accuracy. same potential, so any charge produced in the cable is col-
With the development of MR-guided linear accelerators lected between outer and inner braids and does not affect the
(see Section 14.4), where dose measurements are performed measurement. Different guard ring designs are implemented
in the presence of a magnetic field, specific MR-compatible in plane-parallel chambers, where they are used to avoid any
ionisation chambers constructed without ferrous material are field distortion and therefore, define the effective collection
becoming commercially available. However, the magnetic volume between the collecting electrodes where the electric
field also influences the electron tracks within the chamber field is uniform (see Figure 16.2b). It must be stressed that
cavity, and this must be accounted for in the dose determina- because of the influence of the electric field, even for cylin-
tion with an ionisation chamber (see Section 19.7). drical ionisation chambers, the effective collection volume dif-
fers from the mechanical volume, which could be obtained
accurately, for instance, from micro-computed tomography
16.2.3 Insulators and Guard Rings
(CT) imaging (McNiven et al. 2008; Ross 2009).
The charge (or current) created in the air cavity is conducted
to the measuring device (the electrometer) through leads
16.2.4 Electrometers
connected to the electrodes. These leads pass through the
chamber stem, which is attached to the chamber cavity and The electronics required to measure the charge (or current)
designed to hold the whole assembly. Because the charge is provided by the electrometer, which in most cases, includes
circulating between the electrodes should be restricted to a voltage supply and a display unit. Most modern electrom-
that created within the air cavity, there must be no other eters also have the ability to interface with a computer for
conducting path between the electrodes except that of the both parameter selection and data acquisition. For further
ionised air, and therefore, an insulator must be provided details on electrometers, readers are referred to Andreo et al.
between them. Although, in the past, various naturally (2017). In the early days of radiation dosimetry, physicists
occurring substances were used as insulators (amber, quartz used a gold-leaf electroscope equipped with a quantitative
and ceresin wax), today, synthetic plastics can be used. These scale as an electrometer. Today, most electrometers are based
are polymethyl methacrylate (PMMA), polystyrene, polyeth- on negative-feedback operational amplifiers.
ylene, Nylon, polytetrafluoroethylene (PTFE, Teflon), Mylar An operational amplifier, represented by the triangle in
film and polytrifluoro-monochloroethylene (Kel-F) (Boag Figure 16.3, has a high open-loop gain (>10 4) and high input
1966). Surface leakage because of humidity is resisted well impedance (>1012 ohms), and consequently, a low potential
by Teflon; however, Teflon is more prone to radiation dam- (<100 mV) is maintained between its positive and nega-
age than, for example, polystyrene. tive inputs. Negative feedback means that some or all of the
As a result of the electric field between the electrodes, a so- amplifier’s output is fed back to its negative input. Total feed-
called leakage current can be created in the insulator, and this back results in a gain of unity. When a capacitor or a resistor

Stem Guard ring

Outer electrode Guard ring

Collecting electrode
Collecting electrode

Outer insulator
Inner insulator
V
V

(a) (b)

FIGURE 16.2 Set-up of the guard ring in a cylindrical ionisation chamber (a) to avoid leakage current in the insulator. In some chamber designs
the guard ring does not protrude into the air cavity (see Figure 16.6). In a plane-parallel chamber (b), the guard ring defines the effective collec-
tion volume as described in the text.
Chapter 16: Ionisation Chambers 325

Operational
− amplifier

+
Ionisation chamber
Readout
V V

Electrometer

FIGURE 16.3 The principle of a negative-feedback operational-amplifier electrometer. The capacitor, C, or the resistor, R, is the feedback element.
The use of the capacitor or the resistor corresponds, respectively, to an electrometer operating in integrating or rate mode.

is connected in the feedback loop of an operational amplifier 16.2.5 Cables and Connectors

whose negative input is connected to the chamber collection
The cables connecting the chamber and the electrometer are
electrode and the positive input to the polarising voltage or
an important component of the dosimetry system (Humphries
to the case ground, the voltage measured at the output of the
and Slowey 1986). High-quality low-noise triaxial cables
amplifier is proportional to the charge collected and is equal
having three separated conductive regions are the obvious
to the voltage across the capacitor or resistor. If a capacitor
choice, because this allows a single cable to be used to trans-
C is placed in the feedback loop, and the voltage measured
port the signal from the collecting electrode and to provide
across the capacitor is V, then the charge collected, Q, is equal
chamber grounding as well as high voltage. The microphonic
to CV. This is known as capacitive feedback circuit mode, and
noise and dielectric-response time in cables provided by dif-
unity gain ensures that exactly the amount of charge col-
ferent manufacturers can vary significantly These usually
lected by the chamber collection electrode is deposited in
have a minor influence on the response of the system, but a
the feedback capacitor. Similarly, if instead of the capacitor a
more important aspect is the radiation-induced signal, which
resistor R is placed in the loop, the voltage V measured across
varies with the beam energy and, as shown by Humphries
the resistor allows the current I (related to the charge rate) to
and Slowey (1986), is often rather different between high-
be determined as V/R. The reproducibility of a measurement
energy photons and electrons and from one manufacturer to
depends almost entirely on the quality and stability of the
another. The desired cable characteristics are the following:
capacitor or the resistor. Typical desired characteristics of the
short equilibration time (less than 1 minute) when the high
chamber–electrometer assembly are the following:
voltage is changed; low radiation-induced signal; low micro-
• Fast warm-up (5 min–10 min); phonic noise; low leakage (less than 10−14 A; see also Section
16.4.4 for a discussion of leakage currents); low capacitance;
• Short equilibration time after changing the high
pliability; and easy and robust connector installation.
voltage across the electrodes;
The connectors between the chamber lead and the elec-
• Negligible (<10 −13 A) pre- and post-irradiation leak- trometer have not been given much attention in the dosim-
age currents; etry literature. However they can cause numerous problems
• Sensitivity independent of ambient conditions, time for users, such as the effects of dirt, moisture or small metal
or scale position; fragments, bad contacts, misaligned pins, breaks in connec-
• Ability to change the bias voltage in magnitude and tion of the cable leads to the connector contacts, mismatches
polarity. of connectors, slightly different sizes of the same nominal
connector, and strain of the cable where it meets the connec-
More detailed specifications can be found in the IEC standard tor. Simple precautions should be taken, such as avoiding too
60731 (IEC 2011), the requirements being more stringent severe bending (special cable reels are commercially available),
for reference-class instruments than for field-class or scanning- systematically replacing dust caps on the plugs when unused,
class instruments (McEwen 2010). Some of them are based storing the cables and connectors in a dry and safe environ-
on the concept of effective range, which defines the range of ment, etc. Since the distance between the treatment machine
indicated values for which the instrument complies with a and the control console where the electrometer sits could be
stated performance. For instance, for a field-class instrument, quite large, extension cables are often required, which can
the resolution of the display (or data output terminal) must be cause additional problems, and permanently installing cables
better than 0.5% of the minimum effective reading. with wall-mounted connectors is an interesting option. If
326 Part D: Dose Measurement

A B C D
BNC triax M/F BNC triax F/M TNC triax F/M TNC triax M/F

E F G H
BNC coax M/F Banana M BNC coax F/M Banana F

J K L M
MHV coax M/F SHV coax M/F Type "M" M/F Type "M" F/M

FIGURE 16.4 Different types of connectors between ionisation chambers and electrometers (courtesy of Standard Imaging Inc., WI). BNC coax-
ial (PTW, Exradin); BNC triaxial (Keithley); Miniquick coaxial (Nuclear Enterprise); TNC triaxial (Nuclear Enterprise, Capintec); PTW biaxial
(PTW, type ‘M’). The connector ‘gender’ is also indicated (e.g. M/F means that it is a male connector designed to be connected to a female one).

the connectors between two devices do not match, an appro- type of chamber of great relevance to accurate dosimetry in
priate adapter must be used. In addition male and female radiotherapy; namely, the free-air ionisation chamber that is
connectors must be considered. There is a range of connec- used at Standards Dosimetry Laboratories (SDLs)† to realise
tor types for commercially available chambers and electrom- air kerma standards, which are the basis for the majority of
eters (Nette and Czap 1994), some of which are shown in calibrations of cavity chambers in kilovoltage x-ray beams.
Figure 16.4.* Not all combinations are possible, and some of
them can be dangerous. One reason is, for instance, that the
16.3.1 Free-Air Chambers
outer shielding can be used either for grounding or for high-
voltage supply. It could result in an improper grounding of The so-called free-air ionisation chamber is a specialist
the chamber and sometimes, in risk of electric shock when instrument usually only to be found at Primary Standard
accessible parts of the chamber conduct high voltage. Dosimetry Laboratories (PSDLs). It is called a free-air
chamber because in principle, the walls of the chamber do
not play any role in its response. The basic idea underly-
ing the design is to produce perfect charged particle equilib-
16.3 Types of Chambers
rium (see Section 5.5) in the defined collecting volume for a
In what follows, those types of ionisation chambers that are particular photon beam and hence, to measure exposure (or
used in radiotherapy for measuring the dose in the treat- air kerma) directly.‡ This necessitates a consideration of the
ment beam will be further discussed. The discussion will not maximum range in air of the secondary electrons for various
include chambers designed for radiation protection purposes photon-beam qualities, as the dimensions of the chamber in
(see Section 16.2.2). In the radiotherapy clinic, essentially
only one type of chamber is employed. This is known as a † The Primary and Secondary Standards Dosimetry Laboratories are
cavity chamber (see Section 16.3.2), as its air volume is encap- responsible for the calibration of the dosimeters at the international,
sulated and delimited by walls. However, there is another regional or national level (see Section 19.1).
‡ The quantity exposure is employed in this section, since its original
definition was closely related to the measurements that could be
* For most supplied ionisation chambers, the most common connec- performed with a free-air ionisation chamber, although these days,
tors are BNC triax and TNC. Many ion chamber manufacturers also air kerma is preferred. Air kerma can be considered as the modern
provide several types of extension cables. replacement quantity for exposure (see Section 5.3.2).
Chapter 16: Ionisation Chambers 327

TABLE 16.1
Approximate Continuous Slowing Down Approximation
(CSDA) Ranges in Air for the Mean Energies of the Secondary
Electrons for X-ray Qualities Ranging from 30 kV to 300 kV
and for 60Co γ-Rays
Radiation Average Initial
Quality Secondary Electron CSDA Range
(kVp) Kinetic Energy (keV)a in Air (cm)b
30 10   0.24
50 13   0.36
80 12   0.33
100 11   0.25
200 11   0.25
300 18   0.64
60Co γ-rays 590 210 FIGURE 16.5 Schematic cross-sectional view of a typical standard
free-air ionisation chamber. The ionisation desired for an exposure mea-
Sources: From Johns, H. E. and Cunningham, J. R., The Physics of surement is that produced by the electrons, such as e1 originating in
Radiology, 4th Ed., Charles C. Thomas ed., Springfield, Illinois, interaction volume V (volume defined by the intersection of the beam
1983; ICRU Report 37, ICRU, Bethesda, Maryland, 1984b. and the collecting volume V′, in turn defined by the collector electrode).
a The average initial secondary electron kinetic energy was estimated
The measured ionisation is that collected from volume V′. Charged par-
from Table A-3a in Johns and Cunningham (1983) on the basis of the ticle equilibrium ensures that if a particle e2 escapes from the interaction
mean x-ray energy in the spectrum approximated as 0.37 of the maxi- volume V, it is compensated by a particle such as e3 collected in V′. The
mum energy (i.e. kVp). collector length, l, is typically between 4 cm and 10 cm, and the elec-
b The CSDA ranges (see Section 3.5.3) were taken from ICRU Report
trode spacing, h, is typically between 6 cm and 26 cm, depending on
37 (1984b). the beam x-ray quality. This chamber is used to cross-calibrate a cavity
chamber located at point P where the exposure (or air kerma) in free air
is easily derived (distance inverse square correction) from the ionisa-
all directions from the interaction volume must be at least tion referred to P¢, the effective point of measurement for this chamber.
as large as this for the energy of the particular beam under (Adapted from Attix, F. H., Introduction to Radiological Physics and
consideration. Table 16.1 contains estimates of this quan- Radiation Dosimetry, Wiley, New York, 1986. With permission.)
tity. It can be seen that for 60Co γ-rays (and for megavolt-
age qualities), it becomes totally impractical to build such where
instruments. g is the fraction of the initial secondary electron
Figure 16.5 is a schematic illustration of a free-air cham- kinetic energy re-radiated as bremsstrahlung in
ber. The ionisation to be measured is that produced by sec- air (see Section 5.3.2)
ondary electrons created in the interaction volume V. Note (Wair/e) is the mean energy required to produce an ion
that the collector, length l, is guarded by plates G. The two pair in air, divided by the electron charge (and
illustrative electron tracks e3 and e2 are shown balancing each has units of J C −1).
other to indicate that there is charged particle equilibrium
Eliminating exposure between Equations 16.4 and 16.5
over the complete inter-electrode collecting volume.
yields the following expression for air kerma:
The mass of air for which the ionisation is collected is
derived from the product of the cross-sectional area A of the
beam aperture and the length l of the collecting electrodes in K air =
Q
mair
(Wair e ) (1 −1 g ) (16.6)
the beam direction (V = Al) multiplied by the density r of the
air. The exposure rate ∆X/∆t is then given by: It can be seen that the determination of air kerma relies
directly on the value of (W/e) for air; this key quantity is dis-
cussed in Section 19.4.3, in ICRU Report 31 (ICRU 1979)
DX Q Q
= = (16.4) and in ICRU Report 90 (2016).
Dt mair A l r Free-air chambers of various dimensions are used at SDLs
to determine air kerma in beams of x-rays with qualities
where Q is the charge collected in a specified time interval between 10 kV and 300 kV. At higher energies, the much
∆t. Small corrections must be made for the influence of, for longer secondary electron ranges in air make it impractical to
example, the attenuation by the air and photons scattered build such a chamber. At photon qualities of 60Co γ-rays and
in the chamber. Details of these can be found in Andreo above, air kerma is determined through the use of so-called
et al. (2017). cavity ionisation chambers.
The quantity air kerma, K air, has replaced exposure; these
quantities are related through:
16.3.2 Cavity Chambers
X
K air =
(1 − g )
(Wair e ) (16.5) Cavity chambers are small chambers with walls enclosing the
air volume. They are used both in SDLs and in radiotherapy
328 Part D: Dose Measurement

clinics. The key distinction is that the design of the cavity later, are discussed in more detail in Section 19.4.2. A brief
chambers employed at SDLs (reference-class dosimeters) must description of the various sources of perturbation is, how-
enable highly precise determinations to be made of the sen- ever, useful here to appreciate their importance for the design
sitive air volume, and that the materials used in the walls of ionisation chambers.
and central electrode must be as close to homogeneous as Referring to Conditions 2 and 3 in the preceding list, the
possible for the correction factors to be known as precisely chambers wall(s) should ideally either have negligible thick-
as possible. ness or be continuous with, or equivalent to, the surrounding
There are basically two types of cavity chambers: cylindri- medium so that all the secondary electrons (assuming photons
cal (sometimes known as thimble) chambers and plane-parallel are the primary radiation) in the cavity originate either in the
(also known as parallel-plate, flat or coin) chambers. Both medium or in medium-equivalent material. However, there
of these types are designed to behave as Bragg–Gray cavi- are no perfectly water-equivalent materials. Furthermore, for
ties in megavoltage photon and electron beams (see Section the minimum wall thickness compatible with robustness (i.e.
5.7.4), and their design features can best be understood in around 0.5 mm), some 50% of the ionisation measured with
the light of this. Somewhat different considerations apply to a typical cylindrical chamber in a cobalt-60 γ-ray beam is due
chambers used in kilovoltage x-ray beams (see Section 19.8). to electrons generated in the wall, falling to no lower than
In all cases, the gas in the cavity is air, and the cavity should around 20% at the highest photon energies used in radio-
not be sealed so that the air pressure can equilibrate with the therapy (Andreo et al. 1986; IAEA 1997a). Consequently,
ambient pressure (see Section 16.4.5). a correction for the effect of the wall will generally be nec-
essary, and this is expressed through the wall perturbation
correction factor pwall; generally, this is within 2% of unity
[Link] Desirable Properties of Cavity Chambers –
for the wall materials used in practice. The wall perturba-
Perturbation Corrections
tion correction factor is also thought to include the effect
For an ionisation chamber to be a useful dosimeter in a of a waterproofing sleeve,* if required, and the influence of
phantom irradiated by a high-energy photon or electron the chamber stem. Note that although this effect is generally
radiation under reference conditions the chamber and the only considered in photon beams, an approximate theory for
radiation field, taken as a unit, should have the following wall effects in electron beams (Nahum 1988) predicts effects
properties: of the order of 0.5%. More recent Monte Carlo simulations
indicate that the effect can be as large as 2% at the reference
1. The sensitive volume of the air cavity should be well depth in low-energy electron beams and even larger with
defined in space and ideally, small in relation to dose increasing depth (Buckley and Rogers 2006).
gradients in any direction. In practice, a potential The second important departure from Bragg–Gray behav-
difference of some 300 V should be maintained iour is due to the finite size of the air cavity (Condition 1).
between the electrodes and the sensitivity (which is This is generally dealt with by assigning to the chamber an
proportional to the air volume) should not be too effective point of measurement,† denoted by Peff or EPOM; in
low to preclude the lengths of any dimension being general, this will lie upstream of the centre C of the cavity;
less than around 2 mm and the cavity volume being the magnitude of the (Peff–C) shift is non-negligible in most
much less than 0.03 cm3; in very small fields, the air practical chamber designs. More details on Peff are given in
volume would need to be much less than this for Section [Link]. The effect is similar in photon and in elec-
the chamber to fulfil Bragg–Gray conditions (see tron beams.
Section [Link] and Kumar et al. 2015). There is a further effect due to the finite cavity size con-
2. The wall (and central electrode in the case of thim- fined to measurements in electron beams. The large differ-
ble chambers) should be made out of materials that ence in density between the gas and the wall, or medium,
are as homogeneous as possible and also as water- causes more electrons to be scattered into the cavity than are
(or medium-)equivalent as possible; in practice, this scattered out. For certain cavity shapes, especially in relatively
is often unachievable. low-energy electron beams, there is an appreciable increase in
3. The walls should be as thin as possible, especially the electron fluence in the cavity compared with that in the
in the beam direction on the upstream side of the (undisturbed) medium at the equivalent position (Harder
cavity, consistently with the demands of robustness 1968; ICRU 1984a; Nahum 1996; IAEA 1997b). This flu-
(although in principle, this would not matter if the ence and signal increase is corrected for by introducing a per-
walls were perfectly medium-equivalent and no turbation correction factor less than unity, usually denoted
measurements were required close to the surface of by pcav. There is no corresponding effect in photon beams; i.e.
the phantom). pcav = 1 in Equation 16.7; the reason was explained by Harder

Departures from Bragg–Gray conditions, i.e. that the

charged-particle fluence in the air cavity is identical to that * Many recent ionisation chambers are waterproof, and such sleeves are
in the undisturbed medium at the same position (see Section only rarely used nowadays.
† An alternative approach, presented in Section [Link], is to add a cor-
5.7.4), require perturbation correction factors (see Section rection factor p dis to Equation 16.7 (to account for the displacement of
5.7.6) for accurate determination of the dose to the medium. the medium by the air cavity), enabling Peff to be replaced by C, the
These factors, which are included in the Equation 16.7 given geometrical centre of the cavity.
Chapter 16: Ionisation Chambers 329

(1974) – (partial) charged particle equilibrium means that also models where the central electrode is made of graphite,
the Fano theorem can be invoked (see Section 5.7.2), and A-150 or C-552 instead of aluminium.
therefore, the charged-particle fluence must be the same in a Other chamber designs may be smaller in diameter and
low-density gas cavity (strictly with identical mass interaction length to achieve a higher spatial resolution (particularly
coefficients) as in the high-density medium. important in situations where the dose is changing rapidly,
The expression to obtain the dose to the medium, Dmed, e.g. in the penumbra region, and in so-called small fields)
from the dose to the air measured by the ionisation chamber, but at the cost of a lower signal per unit dose. A represen-
Dair, is modified for departures from Bragg–Gray behaviour tative selection of cylindrical chamber designs is given in
(Equation 16.1) to read: Table 16.2.
Most cylindrical chambers are supplied with a build-up cap.
D med (Peff ) = D med s med, air p wall pcav pcel (16.7) This is placed around the chamber wall to achieve charged
particle equilibrium (i.e. full build-up in the wall) when the
where the measurement is assigned to position Peff in the chamber is irradiated free-in-air in a 60Co γ-ray beam, as is
medium; the correction, pcel, for the effect of the central elec- done when providing the chamber with an air kerma calibra-
trode material has been added for completeness. tion factor, NK , at an SDL (see Section 19.3.2). It should be
Some designs of ionisation chamber for use in radiotherapy noted that the build-up cap is not used when the chamber is
are not intended to fulfil the Bragg–Gray criteria; in kilo- irradiated in a phantom at the SDL for a calibration in terms
voltage x-ray beams (see Chapter 22), practical chambers do of dose to water ND,w (see Section 19.2) or irradiated in a
not behave as Bragg–Gray cavities due to the vastly reduced phantom in the clinic. However, if the chamber is not water-
ranges of the secondary electrons (see Section [Link]). Here, proof, it is necessary to place a thin waterproof sheath around
the primary concern is that the air kerma calibration factor, it before using it in a water phantom.‡
NK (see Sections 19.3 and 19.8), should show as little varia- Cylindrical chambers such as those given in Table 16.2 can
tion as possible over a wide range of radiation quality, from be used for either reference or relative dose measurements. For
say 50 kV (half value layer* ≈ 1 mm Al) up to 300 kV (half reference dose determination, they must be calibrated, trace-
value layer ≈ 20 mm Al) and sometimes even extending to able to national standards, and used with a calibration factor
cobalt-60 energy (with a build-up cap), as it may be desirable (ND,w or NK) to determine the dose in absolute units (Gy) at
that the same chamber, e.g. a graphite-walled Farmer (see the reference point in water; it is emphasised that this must be
Section [Link]), be provided with NK factors for the com- done by following the Code of Practice currently applicable in
plete range of qualities existing in the radiotherapy clinic. It the country or region in question. They must also be shown
was found experimentally, for example, that a graphite wall to behave reasonably well for reference dose measurements.
in combination with an aluminium central electrode resulted The addendum to the TG-51 protocol gives specifications for
in a variation in NK of no more than around 5% (Humphries a reference-class chamber (McEwen 2010; AAPM 2014). For
and Slowey 1986). relative dose measurements, other types of cylindrical cham-
The extent to which these various requirements are ful- bers (field-class or scanning-class chambers) may be used in
filled is now discussed in relation to the two main chamber conjunction with computer-controlled beam-data acquisition
types, cylindrical and plane-parallel. systems. These are preferably considerably smaller so as to
improve the spatial resolution (see Section [Link]).
[Link] Cylindrical Chambers
[Link] Plane-Parallel Chambers
The classic Farmer chamber (Farmer 1955; Aird and Farmer
1972), a very common design of cylindrical chamber for The plane-parallel or parallel-plate chamber design appears
radiotherapy, is shown in Figure 16.6. The cylindrical outer to be rather different from that of the cylindrical chamber,
wall acts as one of the electrodes and is made of graphite. The although both have cylindrical air volumes. The cylindrical
other electrode is a thin aluminium central rod. The nominal (or thimble) chamber is generally irradiated with its central
cavity volume is 0.6 cm3. The cavity length is 24 mm with axis perpendicular to the beam direction, whereas the plane-
an inner diameter of 6.25 mm. The graphite outer wall is parallel chamber is generally irradiated with its axis parallel
0.5 mm in thickness. to the beam direction.
Several other types of Farmer chambers are commercially The designs discussed up to this point are intended to
available, with walls made variously of graphite, A-150 (mus- be used principally in photon or electron beams from lin-
cle-tissue-equivalent conducting plastic developed by Francis ear accelerators. For kilovoltage x-rays, the plane-parallel
Shonka), C-552 (air-equivalent conducting plastic), Delrin, chambers currently used have a very thin entrance wall, e.g.
Nylon 66 and PMMA.† The closest of these to water equiva- 0.004 mm of conducting Mylar. Their wall and body materi-
lence in terms of atomic composition is PMMA. There are als are chosen with care to avoid significant energy depen-
dence of the response for low-energy x-rays. These chambers
are to be used with the window flush with the surface of
* For the definition of half value layer, see Section 22.2.2.
† When PMMA is used, it is coated on the inside with a conduct- ‡ The absorbed dose to water calibration of the NPL secondary stan-
ing layer, usually dag, which is basically a suspension of graphite in dard chamber NE2561 is carried out in its waterproof PMMA sheath
alcohol. at the SDL.
 330

TABLE 16.2
Common Cylindrical Ionisation Chambers
Cavity Cavity Cavity Wall Build-up Build-up Cap Central
Volume Length Radius Wall Thickness Cap Thicknessb Electrode
Ionisation Chamber Typea (cm3) (mm) (mm) Material (g/cm 2) Materialb (g/cm 2) Material Waterproof
Capintec PR-05 mini 0.14 11.5 2.0 C-552 0.220 Polystyrene 0.568 C-552 N
Capintec PR-06C/G Farmer 0.65 22.0 3.2 C-552 0.050 PMMAc 0.547 C-552 N
Exradin T2 Spokas (4 mm cap) 0.53 11.4 4.8 A-150d 0.113 A-150 0.451 A-150 Y
Exradin T1 mini Shonka (4 mm cap) 0.05 5.7 2.0 A-150d 0.113 A-150 0.451 A-150 Y
Exradin A12 Farmer 0.65 24.2 3.1 C-552 0.088 C-552 0.493 C-552 Y
Exradin A19 Farmer 0.63 25.0 3.1 C-552 0.088 C-552 0.493 C-552 Y
Exradin A1SL scanning 0.05 6.0 2.0 C-552 0.194 C-552 Y
Exradin A16 micro 0.007 2.4 1.2 C-552 0.088 SPCe Y
NE 2571 Farmer 0.6 24.0 3.2 Graphite 0.065 Delrin 0.551 Al N
NE 2581 Farmer (PMMA cap) 0.6 24.0 3.2 A-150d 0.041 PMMA 0.584 A-150 N
NE 2561/2611 Sec. Std 0.33 9.2 3.7 Graphite 0.090 Delrin 0.600 Al (hollow) N
PTW 30010 Farmer 0.6 23.0 3.1 PMMAf 0.057 PMMA 0.541 Al N
PTW 30013 Farmer 0.6 23.0 3.1 PMMAf 0.057 PMMA 0.541 Al Y
PTW 31013 Semiflex 0.3 16.3 2.8 PMMAf 0.078 PMMA 0.357 Al Y
PTW 31010 Semiflex 0.125 6.5 2.75 PMMAf 0.078 PMMA 0.357 Al
PTW 31014 PinPoint 0.015 5.0 1.0 PMMAf 0.085 PMMA 0.357 Al Y
Scanditronix-Wellhöfer IC-25/IBA 0.25 10.0 3.0 C-552 0.068 C-552 Y
CC25
Scanditronix-Wellhöfer IC-70 0.65 23.1 3.1 Graphite 0.073 POMg 0.560 Al Y
Farmer/IBA FC65G
IBA FC65P 0.65 23.1 3.1 POMg 0.057 Al Y
IBA CC13 0.13 5.8 3.0 C-552 0.070 C-552 Y
IBA RAZOR 0.01 3.6 1.0 C-552 0.088 Graphite Y
IBA RAZOR NANO 0.003 – 1.0 C-552 0.088 Graphite Y
Source: Specifications are taken from International Atomic Energy Agency, Technical Report Series no. 483, Tables 4 and 5, IAEA, Vienna, 2017.
a As obtained from the manufacturers.
b Blanks correspond to either no build-up cap offered or no information available.

c Polymethyl methacrylate (C H O ), also known as acrylic. Trade names are Lucite, Plexiglass and Perspex.
5 8 2
d Mijnheer (1985) showed that the chamber response for chambers that use A-150 walls can vary due to their hygroscopic nature, and they are therefore no longer recom-

mended for reference dosimetry (AAPM 2014).

e SPC is an abbreviation for silver-plated copper.

f Like most chamber types with non-conductive plastic walls, the chamber wall has an inner conductive layer made of graphite. For this chamber type, the thickness and

density of the graphite layer is supplied in the chamber specifications.

g Poly Oxy Methylene (CH O). A trade name is Delrin.
2
Part D: Dose Measurement
Chapter 16: Ionisation Chambers 331

Graphite PTCFE
Aluminium Dural

7 6.25
1

21.4
24.0
36.8

FIGURE 16.6 A Farmer-type ionisation chamber (redrawn from Aird and Farmer 1972), dimensions in millimetres.

Open to Guard ring

atmosphere Collecting electrode

Insulator

FIGURE 16.7 The NACP plane-parallel design of ionisation chamber. (From Nordic Association of Clinical Physics, Acta Radiol. Oncol., 20,
401–415, 1981. With permission.)

a phantom (skin dose). In a so-called Grenz-ray chamber, TABLE 16.3

the phantom forms part of the chamber body.* For electron Desirable Properties for a Plane-Parallel Chamber in
or megavoltage-photon beams, the entrance wall may be Electron Beams
thicker, as shown in Figure 16.7.
Plane-parallel chambers are preferred to cylindrical cham- Chamber dimensions
bers in situations where the dose gradient is steep in the Front window thickness ≤1 mm
direction of the beam, for example in the build-up region Collecting electrode diameter ≤20 mm
in a high-energy photon beam (see Section 23.3.1) or in the Ratio of guard ring width to cavity height ≥1.5
Bragg-peak region of a proton beam (see Chapter 25). Cavity height ≤2 mm
Historically, their use has been recommended for absolute In-scattering perturbation effect, pcav <1%
dose determination in electron beams of energy typically Wall and backscattering perturbation effects, pwall <1%
below 10 MeV†, although this upper limit varies in different Polarity effect <1%
Codes of Practice. This recommendation was mainly because Leakage current <10−14 A
the in-scattering perturbation effect (see Section [Link]) Long-term stability <0.3%a
was thought be eliminated with a sufficiently wide guard Source: From International Atomic Energy Agency, Technical Report
ring (IAEA 1997b; Nahum 1996), so that the plane-parallel Series no. 381, IAEA, Vienna, 1997a. With permission.
chamber samples the electron fluence incident only through a Response variation per year for any radiation quality in the rated range.

the front window. The in-scattering or fluence perturbation

correction could therefore be assumed to be unity, and the
position of the effective point of measurement is well defined suggests that cylindrical ionisation chambers can be used
at the centre of the inside surface of the front window (see in electron beams with nominal energies lower than previ-
Section [Link]). Furthermore, the in-scattering effect or ously thought, because the fluence perturbation can now be
fluence perturbation for cylindrical chambers was thought accurately accounted for. Problems associated with the long-
to be large and variable. It is now known (Zink et al. 2014) term stability of plane-parallel chambers noted by Muir and
that the in-scattering perturbation for plane-parallel cham- McEwen (2017) and references therein suggest that it might
bers is not completely eliminated no matter how wide the be appropriate to use well-behaved cylindrical chambers for
guard ring. More recent evidence (Muir and McEwen 2017) the reference dosimetry of all clinical electron beams. This is
discussed in more detail in Section [Link].
* Grenz rays are x-rays at or below 20 kVp.
The Nordic Association of Clinical Physics (NACP) cham-
† See table 16.3 for desirable properties of plane-parallel chambers in ber (Figure 16.7) is a well-established flat chamber design.
such beams. It has a 0.6 mm thick front window made of Mylar foil and
332 Part D: Dose Measurement

graphite, enabling measurements to be made at small depths. 16.3.3 Transmission Monitor Chambers

The back wall is made of graphite and the side walls of rexo-
A further type of ionisation chamber, usually referred to as
lite. The collecting electrode is a very thin layer (≈0.1 mm)
a monitor chamber, can be found in the treatment heads of
of graphited rexolite mounted on a thin insulating layer to
linear accelerators. Such monitor chambers are essentially
minimise the polarity effect (see Section 16.4.2). There is
used to control the total dose, the dose rate and the beam
a 3 mm guard ring to minimise the signal from electrons
position, as described in Section 11.4.2. Their principle and
scattered-in from the side walls. The distance between the
design have some similarities to plane-parallel chambers, but
electrodes is 2 mm.
since they are used as transmission chambers, their diameter
Table 16.4 lists the characteristics of plane-parallel cham-
is much larger, and their electrodes (i.e. entrance and exit
ber types in common use. It should be noted that relatively
windows) are very thin to prevent significant perturbation
large values of the polarity effect have been observed with
of the beam characteristics. Some monitor chambers (e.g.
certain plane-parallel chamber types, and particular care
for Varian accelerators) are sealed to keep the enclosed mass
should be taken to investigate this property when choosing
of air constant, independently of the pressure and (machine
a chamber.

TABLE 16.4
Characteristics of Plane-Parallel Chambers in Common Use
Collecting Guard
Ionisation Electrode Electrode Ring Recommended
Chamber Typea Materials Window Thickness Spacing Diameter Width Phantom Material
NACP-01 Graphite window 90 mg cm–2, 0.5 mm 2 mm 10 mm 3 mm Polystyrene, graphite,
(Scanditronix) Graphited rexolite electrode water (with
Graphite body (back wall) waterproof housing)
Rexolite housing
IBA/Scanditronix Mylar foil and graphite 104 mg cm–2, 0.6 mm 2 mm 10 mm 3 mm Water, PMMA
NACP-02 window
Graphited rexolite electrode
Graphite body (back wall)
Rexolite housing
PTW 23343 Graphited polyethylene foil 106 mg cm–2, 0.9 mm 2 mm 5.3 mm 0.2 mm Water, PMMA
Markus chamber window (incl. cap)
Graphited polystyrene
collector
PMMA body
PMMA cap
IBA/Scanditronix Window and body C-552 176 mg cm–2, 1 mm 0.5 mm 10 mm 3.5 mm Water
PPC 05 Graphited (PEEKb) electrode
Exradin 11 Conducting plastic wall and PI 1: 104 mg cm–2, 1 mm 2 mm 20 mm 5.1 mm P11: Polystyrene,
electrodes, water
Model P11: polystyrene
equivalent
Model A11: C-552, air
equivalent
Model T11: A-150, tissue
equivalent
PTW Roos chamber PMMA, graphited 118 mg cm–2, 1 mm 2 mm 16 mm 4 mm Water, PMMA
electrodes
PTW 34045 Graphited polyethylene foil 106 mg cm–2, 0.9 mm 1 mm 5.0 mm 2 mm Water, PMMA
Advanced Markus window (incl. cap)
chamber Graphited polystyrene
collector
PMMA body
PMMA cap
PTW 34070 Bragg PMMA, graphited 118 mg cm–2, 3.45 mm 2 mm 81.6 mm 1.1 mm Water
Peak chamber electrodes
Source: Taken from International Atomic Energy Agency, Technical Report Series no. 398, Table 4, IAEA, Vienna, 2000. With permission.
a Some of the chambers listed in this table fail to meet the requirements described in Table 16.3. They have been included here because they are used
in certain clinics.
b Polyetheretherketone (C H O ) 1.265 g cm–3.
19 18 3
Chapter 16: Ionisation Chambers 333

head) temperature. However, keeping the mass of air accu- tracks with a high ionisation density, i.e. high-linear energy
rately constant is difficult to achieve in practice, and the transfer (LET) particles such as neutrons or heavy ions (see
monitor response must be checked frequently to prevent any Section 6.11.5).
drift. Alternatively, some monitor chambers (e.g. in Elekta General recombination proceeds once diffusion and ion
machines) are unsealed, and their response is automati- drift have modified the initial track structure, and posi-
cally corrected for changes in pressure and temperature (see tive and negative ions of different tracks recombine as they
Section 16.4.5) directly measured in the machine head. This drift toward the chamber electrodes. The process is also
does not, however, remove the need for frequent checks (see known as volume recombination; it increases as the density
Section 46.8.1). of ions increases, being dependent on dose rate. This is the
There are also commercially available transmission moni- most significant process in the majority of radiotherapy
tor chambers that are mounted externally to the linac and beams.
are used for dose monitoring (i.e. to account for variations Approximate empirical relationships have been proposed
in linac output) when making relative measurements or for to describe initial and general recombination; so-called Jaffé
providing patient dose quality assurance, taking into account plots (1/q as a function of 1/V or 1/V 2, where q is the mea-
the fluence modulation achieved with the multi-leaf collima- sured charge and V the polarising voltage) can be created to
tor in intensity-modulated radiation therapy (IMRT). For represent the complete saturation curve; Figure 16.8 shows
this latter purpose, pixel-segmented ionisation chambers an example of a Jaffé plot.
have been developed (see e.g. Thoelking et al. 2016 and Initial recombination in a continuous beam (e.g. cobalt-60)
Section 18.6.6). is described by:

1/q = 1/q 0 + constant/V (16.8)

16.4 Determination of the Charge Produced

whereas general recombination in a continuous beam (e.g.
Ionisation chambers must be used in the voltage range cobalt-60) follows the relationship:
where the collected charge reaches, or asymptotically
approaches, a saturation value (see Figure 16.1). However, 1/q = 1/q 0 + constant/V 2 (16.9)
because of the physics of ion transport and chamber elec-
trical design, the charge collected can be different from For pulsed (linac) beams, both initial and general recom-
the charge produced in the chamber gas by the passage of bination are described by Equation 16.8. Jaffé plots can be
radiation. The most important phenomena affecting the used to extract an accurate measure of the ion recombina-
charge collected are ion recombination, polarity effects, tion correction through extrapolation to infinite voltage (i.e.
and leakage and stem effects. Stabilisation and pre-irra- 1/V = 0) using a fit to the linear region of the curve. These
diation are also required for reproducible measurements. include effects from both initial and general recombination
An additional important issue arises at the time of deter- at a given average dose-per-pulse, d. They are also useful for
mining the charge per unit mass of air in the chamber determining the maximum ‘safe’ operating voltage, Vmax,
cavity – a correction must be made for the local pressure before effects from processes such as charge multiplication
and temperature. produce the non-linear behaviour shown in Figure 16.8 (see
the proportional region in Figure 16.1). This behaviour can
differ between chambers of the same type, and the maxi-
16.4.1 Recombination mum operating voltage may also differ from the value speci-
fied by the manufacturer.
[Link] Initial and General Recombination
The recombination of positive and negative ions (i.e. elec-
[Link] Collection Efficiency: Boag’s Theory
trons) within the air cavity reduces the amount of charge
collected. Two main mechanisms for the recombination of For general recombination, Boag (1966, 1987) defined the
ions, related to the structure of radiation tracks and their collection efficiency f of an ionisation chamber as the ratio
distribution, have been described (Boag 1966, 1987); these of the collected (i.e. measured) charge qm to the ideal satura-
are termed initial and general recombination. There are also tion charge qi that would be produced without recombina-
effects from ion diffusion, which are indistinguishable from tion (i.e. f < 1). It should be noted that the recombination
initial recombination. correction factor, ks, commonly used in dosimetry protocols,
In the process of initial recombination, positive and is the inverse of the collection efficiency. In his comprehen-
negative ions in the track of a single particle recombine; sive review, Boag developed step by step the theory for the
this process is also called columnar recombination. Because collection efficiency in continuous radiation, pulsed beams
only one single particle track is involved for a given elec- and pulsed scanning beams for different types of ionisa-
tric field strength, this intra-track process is independent tion chambers. Here, a summary is given for the case of the
of the number of tracks formed per time interval (i.e. the pulsed beams commonly found in clinical accelerators; for
dose rate). In practice, this effect is only significant for more details, readers should consult Boag (1987). A more
334 Part D: Dose Measurement

100 V

Is/I

NACP plane-parallel chamber

1/V (V–1)

FIGURE 16.8 Example of a Jaffé plot. The inverse of the chamber current I as a function of the inverse of the polarising voltage V for a particular
plane-parallel chamber at a dose per pulse of 0.14 mGy to air. The chamber current is normalised by the saturation current Is as determined by a
full Boag analysis of the data below 100 V. The solid line is a linear fit to the same data. (Reproduced from Burns D.T. and McEwen M.R., Phys.
Med. Biol., 33, 2033–2045, 1998. With permission.)

condensed, practical approach is given in the report 34 of the collected charge density qm is equal to fqi, and Boag intro-
ICRU (1982). duced an auxiliary variable v = fu = ln(1 + u) with the result
In pulsed beams, the collection efficiency is given by: that:

1 v
f = ln(1 + u) (16.10) f = (16.13)
u ev - 1

where the variable u is defined as: This equation allows f to be determined from the measured
charge, taking v = [m AI V ]qm from Equation 16.11 where
qi u = v (qi /qm ).
u=m AI (16.11) A simpler approximate equation, essentially a fit to calcu-
V
lated values from the complete expressions, has been given in
and where
IAEA (1997b); for plane-parallel chambers, the recombina-
μ is a constant that depends only on properties of the
tion correction factor k s is given by:
ions and their mobility (3.02 × 1010 V m C −1)
qi is the initial charge density (e.g. C m−3) s2
V is the polarising voltage (e.g. V) ks = 1 + 0.54d (16.14)
V
A I is the solution of a double integral that in the case of
a plane-parallel chamber, is equal to the square of the where
distance between the electrodes (e.g. m 2). d is the average dose per pulse in mGy
s is the electrode spacing in millimetres
In the case of a cylindrical chamber, A I is given by: V is the polarising voltage in volts.

(a + b) ln(a b) The average dose per pulse refers to the absorbed dose
AI = (a − b)2 (16.12) to air. A similar equation can be written for cylindrical
(a − b) 2
chambers, replacing s2 by the corresponding value A I from
where a and b are the radii of the external and internal elec- Equation 16.12.
trodes, respectively. Some analyses of practical cylindrical chambers, such as
A problem with Equation 16.10 is that it depends on qi, the National Physical Laboratory (NPL)-designed stan-
the initial charge density, which is not known until the dard chamber (Kemp 1972; Burns and Rosser 1990) or the
amount of recombination has been determined. However, Farmer chamber (Havercroft and Klevenhagen 1993), have
Chapter 16: Ionisation Chambers 335

used Boag’s approach and taken into account both the cylin- When flattening filter free (FFF) linear accelerators are
drical body and the end of the chamber. Expressions for both used (see Section 23.7), the increase in dose rate results in
these chamber designs can be expressed as: an increase in dose per pulse, d, and can potentially result in
much larger ion recombination corrections (up to 5%); these
d larger corrections must be taken into account for accurate
ks = 1 + n (16.15)
V reference dose determination (Lang et al. 2012; AAPM
where 2014).
d is the dose per pulse in mGy
V is the polarising voltage in volts [Link] Collection Efficiency: the
n is a constant depending on the chamber design. Two-Voltage Method
The values of n are 4.6 for the NPL-designed chamber Situations can be encountered when an ionisation chamber
and 5.75 for the Farmer chamber. Both the simpler equa- does not reach saturation at the highest polarisation volt-
tions given in terms of dose per pulse (Equation 16.14 for age that can safely be used.* However, a method based on
plane-parallel chambers and Equation 16.15 for the Farmer making a second measurement at a much lower voltage has
design) are in close agreement with the illustrative values been described by Boag and Currant (1980). If the collected
given in UK Codes of Practice for electron beams (IPEMB charges Q 1 and Q 2 have been obtained at the voltages V 1 and
1996b; IPEM 2003) and also with experimental measure- V2, and Q i is the desired saturation value, then:
ments using a range of chambers recommended in those
codes (Nisbet and Thwaites 1998a). 1
Q 1 = Q i f1 = Q i ln(1 + u1)
Starting from the Boag (1950, 1966) theory of ion recom- u1
bination, Burns and McEwen (1998) derived a new formal- (16.17)
1
ism for ion recombination. A minor modification of this new Q 2 = Q i f2 = Q i ln(1 + u 2)
u2
formalism was introduced in the addendum to the TG-51
protocol for reference dosimetry of megavoltage photon and their ratio becomes:
beams (AAPM 2014) as:
Q 1 u 2 ln(1 + u1) V1 ln(1 + u1)
= = (16.18)
ks = 1 + C init + C gend (16.16) Q 2 u1 ln(1 + u 2) V 2 ln(1 + u 2)
as u1/u 2 = V2/V 1 from Equation 16.11.
where Cinit and Cgen are the coefficients of initial and general
Equation 16.18 can be solved for u1 (corresponding to the
recombination, respectively. Since the component of initial
highest voltage) numerically or graphically, because both the
recombination is independent of the dose per pulse d, and
ratios Q 1/Q 2 and V 1/V2 are known.
Cgen only includes parameters related to the chamber geom-
Weinhous and Meli (1984) carried out this calculation and
etry and polarising voltage (consider Equations 16.14 and
fitted the results for the recombination correction factor k s to
16.15), ks should vary linearly with d. This enables the coef-
the function:
ficients Cinit and Cgen to be derived by determining ks as a
2
function of d. It is possible to vary d via the linac operating æQ ö æQ ö
parameters, but the easiest way is to change the source-to- ks = a0 + a1 ç 1 ÷ + a 2 ç 1 ÷ (16.19)
Q
è 2ø èQ2 ø
surface distance or depth of measurement. McEwen (2010)
also showed that the relationship between k s and d could be where the coefficients a0, a1 and a 2 were given as a function of
polarity dependent. V 1/V2 for pulsed and for pulsed-scanned beams. These solu-
The addendum to the AAPM TG-51 protocol (2014) notes tions have been implemented in some dosimetry protocols.
that for a reference-class ionisation chamber, Cinit should be Provided that the correction required is small (typically
small (≤0.002 – that is, the component of initial recombina- ks < 1.05), the quadratic term in Equation 16.19 can be
tion should be less than 0.2%). A large value of Cinit could ignored, and in a first approximation, k s can be written as
indicate non-standard behaviour. Of practical importance (IPEM 2003):
here is that some primary and secondary standards laborato-
ries do not normally correct for ion recombination in ionisa- Q1 -1
Q2
tion chamber calibrations in cobalt-60 (d = 0), and so a large ks = 1 + (16.20)
V1 -1
value of Cinit would introduce errors in reference dosimetry V2
measurements. They do, however, correct for ion recombina-
tion when calibrations are performed in linac beams because According to Equation 16.20, in the simplest case, if a volt-
of the potentially large general recombination correction. age ratio of 2 is selected, the percentage recombination cor-
Muir et al. (2012) showed that consistent values of Cinit are rection is equal to the percentage change in reading. This
obtained when ion recombination measurements are made in
cobalt-60, where the dose per pulse is zero, and using a fit to * This value is the threshold voltage, recommended by the ion chamber
ks as a function of dose per pulse in pulsed linear accelerator manufacturer, above which processes such as multiplication begin to
beams. occur (i.e. in the ‘proportional region’ of Figure 16.1).
336 Part D: Dose Measurement

method does, however, rely on the linearity of the Jaffé plot, where
which demonstrates behaviour consistent with the theory; Q + and Q − are the electrometer readings obtained at pos-
therefore, it is wise to verify this by plotting the full curve itive and negative polarity, respectively
for each chamber at least once. Typical voltage values for Q is the electrometer reading obtained with the
cylindrical chambers may be V 1 = 300 V and V2 = 150 V. polarity used routinely (positive or negative).
These voltages are generally in the linear portion of the Jaffé
plot. For plane-parallel chambers, the deviation from linear- The readings Q + and Q − should be made after ensuring
ity may occur at a smaller voltage value (e.g. 100 V). This is that the chamber reading is stable following any change in
the case for the example of response of one particular plane- polarity (some chambers can take up to 20 min to stabi-
parallel ionisation chamber shown in Figure 16.8. lise; see Section 16.4.3 for a discussion of stabilisation and
For very high dose rates, in pulsed electron beams, the pre-irradiation).
recombination can be very important. Petersson et al. (2017) When the chamber is sent for calibration, a decision is nor-
found that, for a specific type of plane parallel Markus cham- mally made, either by the user or by the SDL, on the polaris-
ber, in a 6 MeV electron beam, the two-voltage method and ing potential and polarity to be adopted for the routine use
the Jaffe-plot method (Burns and Mc Ewen 1998) were not of the chamber. The calibration should be carried out at this
any longer accurate for doses per pulse larger than 10 –2 Gy same polarising potential (and polarity if only one polarity is
and 10 –1 Gy (independently of the average dose rate), respec- used for the calibration). The SDL may or may not correct
tively. The application of the general Boag theory gave bet- for the polarity effect at the calibration quality. This should
ter results for voltage smaller than about 50 V. For 1 Gy per be stated in the calibration certificate. When the calibration
pulse they found that the collection efficiency dropped to laboratory has already corrected for the polarity effect, then
around 26%, 47% and 66% for polarisation voltages of 50 V, the user must apply the correction factor kpol derived using
150 V and 300 V, respectively. Therefore for such high doses Equation 16.21 to all measurements made using the routine
per pulse, specific investigation of the recombination correc- polarity. When the calibration laboratory has not corrected
tion factor are required. for the polarity effect, the subsequent treatment of the polar-
ity effect depends on the facilities available to the user and
16.4.2 Polarity Effect on what beam qualities must be measured. Detailed proce-
dures on how to perform such corrections can be found, for
Reversing the polarity on an ionisation chamber (e.g. applying a instance, in IAEA (2006).
positive voltage to the collector or central electrode with respect For megavoltage photon beams, the addendum to the
to the wall and then changing it to a negative voltage) may AAPM TG-51 protocol (2014) specifies that for a reference-
yield different readings; this phenomenon is termed the polarity class ionisation chamber, the polarity effect should be less
effect. For most chamber types, the effect is practically negli- than 0.4% (0.996 < kpol < 1.004) and should vary by less than
gible in high-energy photon beams. In fact, a large and/or vari- 0.5% with energy.
able polarity effect in high-energy photon beams could indicate In electron beams, the polarity effect increases as the
non-standard ionisation chamber behaviour. In electron beams, energy decreases. It should be less than 1% for well-designed
notably at low energy, the effect may be significant. Several chambers (Nisbet and Thwaites 1998a). At low energy (typi-
explanations, related to net charge deposition in the chamber, cally less than 6 MeV) this effect could be of several percent
have been given for this phenomenon (e.g. see Chapter 12 in for some plane-parallel ionisation chambers and therefore,
Andreo et al. 2017); the effect is better understood in the case such chambers should not be used for reference dosimetry
of plane-parallel chambers in electron beams (IAEA 1997b), (Pearce et al. 2006). This effect has been shown to be over
where it is considered to be essentially a charge-balance effect 10%, for a 6 MeV beam, very high dose rate and low polaris-
on electrodes with asymmetric design. The magnitude of the ing voltage (Petersson et al. 2017).
effect depends on the energy and angular distribution of the
incident radiation, the measurement depth in a phantom and
also the field size (Aget and Rosenwald 1991; Ramsey et al. 16.4.3 Stabilisation and Pre-Irradiation
1999). By examining the pattern of charge deposition by elec- Dosimetry protocols recommend that ionisation chambers
trons in a phantom, it can also be seen that the sign of the polar- be pre-irradiated until a stable reading is obtained. This
ity effect may change with depth. requires that readings be repeated using the same setting
In situations where there is a measurable polarity effect, the until no drift of the readings is observed and that this must
true chamber reading is taken to be the mean of the absolute be done each time the applied voltage is changed. The length
values of readings taken at each polarity. In the routine use of of time and/or amount of dose from radiation required for
a given ionisation chamber, a single polarising potential and the reading to stabilise is thought to depend on the ionisa-
polarity are normally employed. The effect on the chamber tion chamber geometry, although it can vary for different
reading of using polarising potentials of opposite polarity for chambers of the same type. It is also thought that stabili-
each user beam quality can be accounted for by using a cor- sation behaviour is related to the irradiation history of the
rection factor: chamber. McCaffrey et al. (2005) showed that the stabilisa-
tion behaviour of a given chamber is related to the extent
Q+ + Q− of the guard of the active volume of the air cavity. McEwen
kpol = (16.21)
2Q (2010) pointed out that for ionisation chambers with active
Chapter 16: Ionisation Chambers 337

volumes less than 0.01 cm3, the initial offset from the stabi- effect arises from scattered radiation arising from interac-
lised reading can be large (>1%), and the length of time for tions in the stem that may differ from those in the phantom
the chamber reading to settle can be very long – over an hour material and contribute differently to the ionisation in the
in a cobalt-60 beam with a dose rate of 0.5 Gy min−1. cavity volume. This effect can be determined using a dummy
The addendum to the AAPM TG-51 protocol (2014) spec- stem. The chamber is irradiated successively with and with-
ifies that for a reference-class ionisation chamber, the differ- out the presence of the dummy stem, and the ratio of the
ence between the initial reading (per monitor unit) obtained readings allows a correction factor to be determined. Stem
when a warmed-up treatment machine is turned on and the leakage arises as a consequence of the irradiation of the stem
stabilised reading should be less than 0.5%.* and the insulators and cables in it. The effect is usually deter-
mined by irradiating a chamber twice with the same narrow
rectangular field, parallel and perpendicular to the long side,
16.4.4 Leakage and Stem Effect
and a correction factor is then derived in the same way as
Electric leakage in the ionisation chamber and electrometer when using a dummy stem.†
may occur as a consequence of the irradiation of insulators
and chamber parts, cables and electronics of the measur-
16.4.5 Correction for the Changes in the
ing equipment. This is termed post-irradiation leakage, an
Mass of Air within the Cavity
effect that continues after the irradiation has ceased and usu-
ally decreases exponentially with time. IEC standard 60731 As discussed in Section 16.1, the mean absorbed dose to the
(2011) recommends that for a field-class dosimeter, within air in the cavity of an ionisation chamber, Dair , is derived
5 s of the end of a 10 min irradiation, the leakage current shall from:
have decreased to less than 1.0% of the ionisation current
produced in the measuring volume during the irradiation. Q
This specification is rather lax, as most modern chambers Dair = (Wair e) (16.22)
mair
produce leakage readings much less than 0.1% of the ionisa-
tion current during irradiation (AAPM 2014). Another kind where
of leakage is the so-called stray radiation leakage produced Q is the charge produced in the air cavity having a
by scattered radiation in those parts of the chamber assem- mass mair
bly and pre-amplifier that are usually outside the radiation (Wair/e) is the mean energy required to produce an ion
beam (but within the radiation room). Mechanical stress on pair in dry air, divided by the electron charge.
cable insulators can also cause a leakage current, and for this
reason, bending or twisting of the cables should be avoided. The mass of air is equal to V × rair, where V is the cavity vol-
In the case of insulators, leakage may occur either through ume and rair the air density inside the chamber. The number
the volume or across the surface of insulators. The mecha- of interactions and consequently, the charge produced are
nisms of radiation effects on insulators are complicated and proportional to the mass of air, which in turn, is proportional
numerous (Attix 1986). The electrons produced during the to its density. As most ionisation chambers are vented to the
ionisation of atoms in the insulators may be displaced a con- ambient atmosphere, rair is a function of the atmospheric
siderable distance from the ionised atom, requiring some pressure, temperature and humidity.
time to recombine. As a result of the applied voltage, the elec- It is common practice to fix the value of rair for certain
tron–hole dipoles become aligned in a common direction, conditions and convert Q to those conditions, and not the
and their relaxation induces charges in the electrodes con- reverse. The SDLs take the reference air density value as
nected to the insulator, forming a leakage current. Another 1.2930 kg m−3, which is the value for dry air at 0°C and
effect in insulators that has received attention (Galbraith et 101.325 kPa. They then measure charge in the reference
al. 1984; Mattsson and Svensson 1984; Thwaites 1984) is beams at 20°C, 101.325 kPa and 50% relative humidity
charge accumulation in non-conductive plastic phantoms in (common reference conditions). They correct to dry air and
electron beams; this can give rise to a very large electric field to 0°C and 101.325 kPa, using the usual density corrections
around the chamber, which then directs the flow of electrons for dry air. The conversion from charge to absorbed dose to
toward the chamber cavity, yielding an increased signal. This air also involves the value for (Wair/e), and for consistency, in
is discussed further in Section 19.10.3 for phantom materi- Equation 16.22, the value to be used is that for dry air, i.e.
als, but similar effects can occur in the insulators of ionisa- 33.97 J C −1.
tion chambers. When calibrating an ionisation chamber at the SDL (IAEA
The irradiation of the chamber stem is also a compli- 2009), the charge measured by a vented chamber depends
cated phenomenon; fortunately, however, any such effects on air temperature and pressure (and also humidity); there-
have been minimised in modern ionisation chambers. Two fore, the calibration factor must be given for stated reference
mechanisms have been described in the IEC standard 60731
(2011); namely stem scatter and stem leakage. The stem scatter † When performing such measurements, it is recommended to turn the
detector through 90° rather than rotating the collimator or exchanging
* According to the AAPM TG-142 report (AAPM 2009a), the short- the X and Y dimensions of the collimator; this prevents the influence of
term stability of the accelerator monitor response (i.e. during a period the collimator settings on the monitor chamber being interpreted as a
of repeated measurements) is typically better than 0.05%. stem-leakage effect (see also Sections 20.2.5 and 26.2.11).
338 Part D: Dose Measurement

values of these parameters. Most SDLs supply calibration fac- SDL, where P and θ (°C) are the cavity air pressure and tem-
tors to the user that have been corrected to θ0 = 20°C* and perature‡ at the time of the measurements, and P0 and θ0
P0 = 101.325 kPa without humidity correction. Instead, the (°C) are the SDL reference values. The air temperature in a
relative humidity during calibration is controlled within the chamber cavity should be taken to be that of the phantom.
range 45–55%, so the calibration factor applies for relative This is not necessarily the same as the temperature of the
humidity around 50% (in fact, in the range from 20 to 70%, surrounding air, and if a significant difference is observed,
no correction is required†). A detailed treatment of the effect the phantom temperature must be checked regularly, or pref-
of humidity is given by Rogers and Ross (1988). erably, by waiting a sufficient time, the phantom should be
In the clinical beam, the correction factor: close to room temperature. In addition, for measurements
in a water phantom, the waterproof chamber (or chamber
(273.15 + q ) P0 equipped with a waterproof sleeve) should be vented to the
kq P = (16.23)
(273.15 + q0) P atmosphere to achieve rapid equilibrium between the ambi-
ent air and the air in the chamber cavity.
should be applied to convert the measured charge to the
reference conditions used for the chamber calibration at the

* In some countries, the reference temperature is 22°C so it is impor-

tant to check the calibration certificate carefully. ‡ This factor can also be written as k TP, with reference to the Kelvin
† In practice, a correction is only required in tropical countries. temperature T = 273.15+θ.