Soil and Environmental Analysis: Modern Instrumental Techniques - Chapter 8 ppsx

It is important to recognize that, because there may be a variety of routes from parent ground state to product ground state, the decay of a particular isotope may be accompanied by the

Measurement of Radioisotopes

and Ionizing Radiation

Olivia J Marsden and Francis R Livens

The University of Manchester, Manchester, England

Approximately 1700 different isotopes are known, of which around 275 are stable The remainder are radioactive; that is, their nuclear configurations are unstable and can change to more stable forms by nuclear transformations that are collectively known as radioactive decay These radioactive decay processes are accompanied by the emission of particles and/or photons from the nucleus Isotopes (or nuclides) are distinguished by the number of protons and neutrons (collectively known as nucleons) they contain and are commonly designated using mass number (A: number of protons þ neutrons) and atomic number (Z: number of protons) For example, 14

6 C is

an isotope of carbon in which the nucleus contains 14 nucleons, of which six are protons The proton number defines the chemical identity of the atom, since the proton charge must be balanced by the appropriate number of electrons, but it also duplicates the information provided by the chemical symbol and, in practice, is often omitted, hence 14 C Differences in the neutron number may control the stability, or otherwise, of a nucleus but have only subtle effects on chemistry, although these can be exploited in studies of stable isotope fractionation in natural systems, for example 2 H/ 1 H, 13 C/ 12 C,

15

N/ 14 N, 17 O/ 16 O, 34 S/ 32 S (see Chap 9).

Only a minority of the unstable isotopes are formed in nature Most are man-made, and the majority of these are available only in such small amounts, or are so short-lived or both, that they are unlikely to be

encountered in the environment or to be of any use as radiotracers The naturally occurring radioisotopes fall into three groups:

1 Primordial isotopes that have existed since the formation of the Earth about 4.5  10 9 years ago These have half-lives (see later) comparable

to the age of the Earth, that is, in the range 10 8 to 10 12 yr Examples include 235 U (t ½ 7.5  10 8 yr) and 138 La (t ½ 1.35 10 11 yr).

2 Short-lived isotopes formed by the decay of long-lived parents These have widely varying half-lives, ranging from 3  10
7

s ( 212 Po) to 2.4  10 5 yr ( 234 Pa) and are constantly being formed by decay of the parent isotope and removed by their own decay In many cases, these isotopes form part of long decay series, in which multiple transforma- tions occur before a stable nucleus is reached An example is the decay of

7

Be (t ½ 53.4 d), and 14 C (t ½ 5736 yr).

The exploitation of nuclear reactions in nuclear weapons and nuclear power, and the use of radioisotopes in industrial and medical applications, has led to the global dispersion of radioisotopes Some of these are also formed to a significant extent in nature (for example, 3 H and 14 C) but many are not (examples include 239 Pu, 237 Np, 137 Cs, 125 I, 129 I, 35 S).

Figure 1 Isotopes produced by the decay of 238 U.

II THEORY OF RADIOACTIVITY

The phenomenon of radioactivity originates in the instability of some nuclear configurations Very simply, for a nucleus to remain intact, the Coulomb repulsion arising from the interaction of the protons must be less than the ‘‘strong’’ attractive force between all nucleons Thus the forces arising from the presence of protons in the nucleus have both attractive and repulsive components, whereas those arising from the presence of neutrons are predominantly attractive The balance between attractive and repulsive forces controls nuclear stability and is dependent on the neutron/proton ratio in the nucleus In the lightest nuclei, up to about A ¼ 40, a neutron/ proton ratio of about 1.0 is adequate for stability, but beyond this point, the Coulomb repulsion for each extra proton rises faster than the attractive force for each extra neutron, and so the neutron/proton ratio needed for stability increases progressively up to a value of about 1.5 in 209 Bi, the heaviest stable nucleus Nuclei with an unstable nuclear configuration (i.e., with an imbalance of protons or neutrons) can move toward stability

by changing the neutron/proton ratio.

The most common route toward stability is the transformation of a neutron into a proton (for a neutron-rich nucleus) or of a proton into a neutron (for a proton-rich nucleus) There are three ways in which this can occur and these are collectively known as beta decay processes In the neutron-rich nuclei, the reaction:

The positively charged particle (positron) emitted in this case is identical to

an electron in all respects except its charge, which is equal in magnitude to that of the electron but positive in sign.

In general, in positron emission,

a-particle) is ejected from the parent nucleus:

B Gamma Ray Emission

Both a- and b-decay processes may proceed via one or more excited states

of the product nucleus, although transitions directly to the product ground state do occur and indeed are the dominant transitions in a number of important isotopes (e.g., 3 H, 14 C, 32 P, 35 S, 90 Sr, 241 Pu) Where the first decay product is an excited state, this can de-excite to the ground state

by a number of mechanisms The most important of these is the emission of one or more photons of electromagnetic radiation, which are known as

g-rays These photons are characterized by high energies (typically 40 to

2000 keV), and the energy with which they are emitted is that separating the nuclear states between which the g-transition has occurred In most cases, the transition from excited to ground state occurs unmeasurably fast and is effectively coincident with the accompanying a- or b-decay event However,

in some cases, the excited states have measurable lifetimes, in which case they are known as ‘‘nuclear isomers’’ or ‘‘metastable states.’’ One of the best known of these is 110m Ag (t ½ 249.9 d), which was released in significant quantities during the Chernobyl reactor accident.

It is important to recognize that, because there may be a variety of routes from parent ground state to product ground state, the decay of a particular isotope may be accompanied by the emission of photons with a number of different energies and also that not all decay events lead to the emission of a g-photon of a particular energy (or indeed any g-photons at all) The proportion of decays that lead to a particular g-transition is known

as the ‘‘abundance’’ and is usually expressed as a percentage Thus the decay of 241 Am proceeds directly to the 237 Np ground state in 65% of cases and via an excited state at 59.5 keV in the remaining 35% De-excitation takes place in one step, so 241 Am is described as an a-emitter giving a 59.5 keV g-ray with 35% abundance.

Since all radioactive decay events arise as an unstable nucleus moves toward

a more stable (i.e., lower energy) state, the energy difference between the initial and final states has to be dissipated during the decay process Both initial and final nuclear states are of well defined energies, and it is easy to calculate the total expected decay energy Where the product nucleus is formed in a nuclear excited state, which subsequently de-excites to the ground state (see above), the decay energy corresponds to the gap between initial and excited product states.

In a-decay processes, the energy associated with the transition from parent to product states is almost all dissipated as kinetic energy associated

with the a-particle The product nucleus has some recoil energy, but conservation of momentum and the usually much larger mass of the product nucleus leave the a-particle with the majority of the kinetic energy For example,

238

92 U ! 234 90 Th þ 4 2 He 2þ

Since momentum before decay is approximately 0, momentum after decay

is approximately 0 as well Thus

a spectrum of energies with their supposed origin in transitions between well-defined, quantized energy levels When the existence of the neutrino was proposed in 1927, this paradox was resolved, although the neutrino was not detected experimentally until 1953 The neutrino is a particle with zero rest mass, which is also formed during a positron or negatron emission and which carries away a variable proportion of the total decay energy Thus the decay energy is the sum of neutrino and b-particle energies Where the b-particle energy is zero, all the decay energy is removed by the neutrino, and conversely when the b-particle energy is equal to E max the neutrino energy is zero.

D Quantitative Treatment of Radioactive Phenomena

For any given isotope, the rate of radioactive decay (the activity, denoted A)

is proportional only to the number of atoms present in the source, denoted

N Thus if one compares the a-count rate from a 1 ng source of 241 Am with that from a 2 ng source under identical conditions, the first count rate will be half the second The activity is related to the number of atoms present by the decay constant, l, which has units of time
1 i.e.

to fall to 50% of the initial value) This time is known as the half-life and

is denoted t 1/2 It is straightforward to demonstrate the relationship between

l and t 1/2 After one half-life has elapsed, a source’s activity will have fallen from A 0 to A 0 /2, thus

and for our 241 Am example, t 1/2¼433 years.

The simplest example of a radioactive decay is a one-step transformation

of a radioactive parent into a stable product nucleus, for example 14 C:

14

6 C ! 14 7 N þ b

However, in many cases e.g., 238 U, 235 U, 232 Th, there are multiple decays before a stable product nucleus is reached The decay scheme of 238 U has been illustrated in Fig 1 as an example In this, and in the corresponding schemes of 235 U and 232 Th, the decay of the long-lived parent isotope generates shorter-lived isotopes These short-lived isotopes may be radiologically significant (e.g., 220 Rn and 222 Rn), or they may be exploitable

as tracers (e.g., 210 Pb, 234 U, 234 Pa) in natural systems.

The number of atoms of any of these short-lived isotopes which is formed depends on the balance between the production rate (i.e., the decay of the parent) and the removal rate (i.e., the decay of the product) The mathematical relationships between the activities of different members

of the natural decay series were first derived by Bateman (1910) and are discussed in Choppin et al (1995) Three cases can be identified, depending on the half-lives of the parent and product isotopes.

The most important in the environmental context, and the only one discussed here, is that where a long-lived parent decays to a short-lived product (‘‘secular equilibrium’’) If we isolate an isotopically pure sample of the parent isotope, its decay starts to form the product isotope immediately Initially, the production rate (¼ N parentlparent ) is greater than the removal rate (¼ N productlproduct ), so that the product isotope activity increases with time Eventually, the situation is reached where production is balanced by

removal, at which point the product isotope activity is said to be in a

‘‘steady state’’ or ‘‘secular equilibrium’’ An example is provided by the production of 234 Th (t 1/2 24.1 days) from the a-decay of 238 U The change

in the 234 Th activity with time in an initially isotopically pure source of 238 U

is illustrated in Fig 2.

The particles and photons emitted during radioactive decay processes and nuclear reactions are, in general, highly energetic, with energies in the range keV to MeV These are much greater than the energies typically associated with chemical reactions, where only a few eV are needed to make or break a chemical bond As a result, the particles and photons are capable of breaking large numbers of chemical bonds or generating a large number of atoms in excited states during their passage through matter, often forming ions and/or free radical species These processes allow ionizing radiations

to be detected very efficiently.

The rate of dissipation of energy (linear energy transfer or LET) is an indicator of the efficiency with which a particle or photon loses its kinetic energy during passage through a medium Different particles and photons have very different LET values; for example, the LET for a-particles in air is

of the order of 1 MeV cm
1 , while that for g-photons of moderate energy (about 500 keV) in air is around 1 keV cm
1 These very different characteristics lead to the use of different types of detectors to measure the different types of radiation even though the vast majority of detectors

Figure 2 Ingrowth of 234 Th in 238 U to reach secular equilibrium.

measure the ionization or excitation induced by passage of a particle or photon.

The majority of radiation detectors rely on the interaction of the radiation with the detector medium, which may be a suitable solid, liquid, or gas, and the generation of a measurable electrical signal This may be generated indirectly, for example through the stimulation of light (visible or UV) emissions, which are converted to electrical impulses, or it may be stimulated directly Most measurement systems also include amplifiers to boost the often small output from the detectors; analog-to-digital converters (ADCs) to convert the amplifier outputs into digital signals; and data acquisition devices These may be multichannel analyzers (MCAs, also known as pulse height analyzers, PHAs), where appropriate, which display the output data as a histogram of channel number on the x-axis (proportional to photon or particle energy) against number of counts (proportional to emission intensity), or they may be simpler scalers that just measure the total number of signals A key parameter in the measurement

of detector performance is the energy resolution, that is, the sharpness of the signals generated, which is usually expressed in terms of the full width at half maximum (FWHM), although full width at tenth maximum values are sometimes quoted as well The other parameter of general importance

is the detection efficiency, that is, the probability of a particle or photon being detected Other parameters can also be defined, for example peak : Compton intensity ratios in g-spectroscopy, or figures of merit in liquid scintillation counting.

The use of MCAs allows the accumulation of energy-resolved spectra, which is particularly useful in a- and g-spectroscopies and in some types of

b-counting Accurate timing is, obviously, essential in measuring count

rates, and much modern instrumentation is based around personal computers, which are readily adapted to data acquisition and can run the computer programs needed to collect and analyze what can sometimes be complex data.

1 Gas-Filled Counters

Gas-filled counters detect the ionization of a gas contained within the detector and are best used for quantitative measurement of b-particles, although portable instruments can be used to measure combined b,g- radiation fields Such detectors have relatively limited use in quantitative

g-ray measurement, since they have small volumes, and the probability of

complete deposition of -photon energy in the sensitive volume of the detector is relatively low The gas-filled counters are among the simplest detection devices and have now been used for over 80 years, The Geiger– Mu¨ller and proportional counters are technically very similar, consisting of

a gas-filled tube with a central axial anode A potential difference is maintained between the anode and a cathode, and the detector can operate

in either ‘‘proportional’’ or ‘‘Geiger–Mu¨ller’’ mode, depending on the applied voltage The ionizing radiations can enter the tube through a thin mica window on the end of the tube or through a thin glass side wall of the tube, although both attenuate the particle or photon energies Low-energy (< 200 keV) b-particles can be measured using windowless gas-flow detectors, in which the fill gas flows through and is constantly replenished.

At relatively low applied voltages (200 to 600 V), the incident radiation ionizes the detector fill gas and the electrons formed are collected at the central anode The number of electrons released is directly proportional to the energy dissipated in the detector, so the charge collected at the anode is proportional to the b-particle energy, hence the term ‘‘proportional’’

Table 1 Summary of Principal Radiation Types and Detection Methods

Radiation type

Gas counters possible but

unusual

common possible, esp.

portable Geiger detectors Surface barrier and

common common Liquid scintillation

counters

possible, esp.

using PERALS

common

counting Data from proportional counters are usually collected using some form of PHA equipment in order to resolve the different emission energies.

At higher applied voltages (800 to 1000 V), the electrons released in the primary ionization of the fill gas are accelerated so rapidly by the applied potential that they are capable of inducing secondary ionizations, which

in turn cause further ionizations and so on This leads to a signal of the same size, more or less regardless of the energy of the incident particle or photon The large size of the signal means that the ancillary signal processing equipment need only be very simple, and the loss of proportionality to particle energy means that there is no need for energy resolution This makes the G–M tube ideal for use in portable monitoring equipment and in measuring the activity of chemically separated b-emitters.

Scintillation devices are among the oldest established radiation detectors and were used in much of the pioneering radioactivity research in the early 20th century Different scintillators have been developed for the efficient measurement of all the different radiation types.

Liquid scintillation counting (LSC) was developed in the 1950s as a technique for measuring b-particles, particularly the low-energy emissions from isotopes such as 3 H or 14 C In LSC, the radioisotope is intimately mixed with a solution containing one or more scintillants, compounds that can be efficiently promoted into an excited state by energy transfer from

a b-particle On de-excitation, the excited molecules emit light in the visible/near-UV region, which can be detected by a photomultiplier tube The wavelength of the light emissions from those molecules which are most efficiently excited by b-particles does not match the most sensitive spectral region for detection using a PM tube Modern scintillants are therefore

‘‘cocktails’’ containing a ‘‘primary’’ scintillant, which is an efficient absorber

of b-particle energy, and one or more ‘‘wavelength shifters’’ or ‘‘secondary’’ scintillants, which reabsorb the light from the primary scintillants and re- emit at a wavelength better suited to the PM tubes The scintillants are usually complex organic compounds containing extensive delocalized aromatic structures, e.g., 2,5-diphenyloxazole (PPO), used as a primary scintillant, and 1,4-bis-[2-(5-phenyloxazolyl)]-benzene (POPOP), used as a secondary scintillant The early scintillants were soluble only in nonpolar organic solvents such as benzene and dioxan, which severely limited their use with aqueous solutions Much effort has been devoted to the development of scintillation cocktails that can accommodate significant quantities of aqueous material, often as an emulsion, and it is now possible

to dissolve or disperse up to 50 % by volume of aqueous solution into some

modern scintillants Even so, great care has to be taken to avoid phase separation, and there are severe restrictions on the pH of the samples and on the salt loading that can be tolerated Precipitates can be dispersed in gel scintillants for counting (e.g., Anon., 1989).

In practice, liquid scintillation measurements are made by dissolving

or dispersing the sample in around 15 mL of scintillant in a polythene or glass vial The liquid scintillation counter transfers the vial into a light-tight counting chamber, and the light emissions are counted by two PM tubes

180 apart in coincidence mode (in other words, only signals detected by both PM tubes simultaneously are treated as real; those detected only by one, for example stray light, are rejected) Instrument background can be further reduced by heavy shielding of the counting chamber and/or by pulse shape analysis which allows signals from different sources to be distinguished There are a number of potential interferences in liquid scintillation The most important is ‘‘quenching’’, which arises from a number of sources, but commonly from absorption of the light photons within a colored sample (color quench) or else from interference with the energy transfer from

b-particle to scintillant (chemical quench) There are several techniques

available to correct for quench, and it is essential that adequate corrections are made if reliable results are to be obtained Not all these techniques are straightforward and, while most modern counters can be programmed to carry out automatic quench corrections, it is important to ensure that the corrections they perform are satisfactory Quench correction is probably the most difficult aspect of liquid scintillation counting, whereas the measure- ment of background and of counting efficiency is relatively straightforward The energy of an emitted b-particle is rapidly, and often quantita- tively, transferred to the scintillator Thus the number of excited states produced, and the intensity of the light emissions stimulated, are proportional to b-particle energy Early scintillation counters had relatively simple energy-resolving equipment, consisting of one or more variable

‘‘discriminators,’’ which defined the lower level below which signals were rejected, and ‘‘windows,’’ within which signals were counted These instruments provided sufficient energy resolution to allow simultaneous analysis of, for example, 3 H (E max 18.6 keV) and 14 C (E max 156 keV) More modern electronics and PHAs allow more sophisticated data acquisition and display, although the analysis of isotope mixtures will always be difficult owing to the b-particles’ continuous energy distributions (Fig 3).

Cerenkov counting is a variation on basic LSC It relies on the emission of light (Cerenkov radiation) in the visible/UV region of the spectrum when a particle passes through a medium with a velocity greater than that of light in that medium The two great attractions of Cerenkov radiation are its automatic discrimination against low-energy b-particles

(the minimum energy needed to stimulate Cerenkov radiation in aqueous solution is about 500 keV) and the avoidance of any scintillators, thus allowing direct counting of aqueous solutions in a conventional LSC instrument The wavelength of Cerenkov light (blue and near-UV) is such that the signals appear in the same spectral regions as scintillations stimulated by very-low-energy b-emitters such as 3 H Although the absence

of scintillators means that chemical quench is not a problem, color quench can be Efficiency calibration and background subtraction are carried out exactly as in normal LSC Among the isotopes measured by Cerenkov counting are 32 P in radiotracer experiments (e.g., Harrison et al., 1991) and 90 Y (decay product of 90 Sr) in environmental samples (e.g., Zhu et al., 1990; Scarpitta et al., 1999; Vaca et al., 1999).

Historically, a-particles were measured using solid scintillators [ZnS(Ag) screens], which functioned in the same way as those used for

b-particle measurement However, given the limited range and penetrating

ability of a-particles, a-scintillators need only to be arranged as very thin films of scintillator, coupled to a photomultiplier Such detectors are still in use today, primarily in simple monitors since they have no energy-resolving capability but a high sensitivity.

In the last 10–15 years, much effort has been devoted to the development of liquid scintillation counting for a-particles, since the high LET of an a-particle leads to very efficient transfer of energy to the scintillant Instruments that can discriminate between a- and b-particles are now commercially available These use pulse shape discrimination (Pates et al., 1993, 1998) to distinguish between the particle types, and good energy resolution is now achievable Alpha scintillation counting is particularly attractive where the separation chemistry results in the isolation

Figure 3 LSC spectra of 3 H and 14 C.

of the element of interest in an organic solvent, since many of these can be readily incorporated into scintillants Counting of aqueous solutions is a little more difficult, since these often need to be acidic to prevent hydrolysis, and many scintillants have only a limited capacity for such solutions Photon/Electron Rejecting Alpha Counting by Liquid Scintillation (PERALS) is a more recent development in a-scintillation Since much of the complexity in commercial LSC arises from the need to reduce backgrounds in b-measurement, a spectrometer designed specifically for

a-scintillation can be made relatively simple The PERALS spectrometer

fits in a standard NIM crate and incorporates signal discrimination to reject pulses arising from b-decay events and passage of g-photons Some applications are discussed by Cadieux (1990) and Cadieux et al (1994) Coupled with a conventional MCA, it is capable of providing high detection efficiency (100%) with energy resolution that is adequate for many purposes, although peaks that are close in energy (e.g., 241 Am and 243 Am at 5.5 and 5.35 MeV, respectively) are not always cleanly resolved (Dacheux and Aupiais, 1997, 1998).

Scintillation counting can also be used to measure -photons Since these have low LET, the scintillator needs to be dense in order to maximize the probability of interaction, and liquid scintillation counting is therefore not really practical The most commonly used solid scintillator is NaI(TI), which is available as very large single crystals (up to 30 cm diameter) Other materials such as CsI(Tl) are more efficient, much more expensive, and used occasionally The main advantage of g-scintillation counting is the high detection efficiency that can be achieved, and its main disadvantage is the poor energy resolution (about 50 keV FWHM at 660 keV), which can lead

to uninterpretably complex spectra from mixed isotopic sources The poor energy resolution arises from the amount of energy needed to generate a scintillation event This is about 350 eV, compared with around 3.5 eV needed to create an electron/hole pair in a semiconductor (see below) Gamma spectroscopy, whether using scintillation or semiconductor detectors, is not as simple as it perhaps seems Although the penetrating ability of g-photons means that it is possible to measure solid samples without much or any pretreatment and, almost always, without lengthy chemical separations, the accuracy of the final results is greatly affected by the care that is taken in calibrating the detector It is necessary to establish two different relationships: one to relate position in the spectrum to

g-photon energy (which is diagnostic of the presence of a particular isotope)

and one to establish the efficiency of photon detection as a function of energy.

In practice, since NaI(Tl) detectors are largely restricted to the measurement of single isotopes or simple mixtures, calibration is best

achieved by using a standard containing a known activity of the isotope

of interest Since the way the sample is presented to the detector (the ‘‘geometry’’) and the nature of the sample matrix (the capacity for self-absorption) can have a very significant effect on counting efficiency, the simplest form of calibration standard is a blank of the appropriate matrix material spiked with a known activity of the isotope(s) of interest It is then possible to calculate detection efficiency directly for the relevant combina- tion of isotope, matrix, and geometry.

The semiconductors used for general g-ray spectroscopy are large-volume single crystals, usually of germanium Ge of the purity needed was not readily available initially, so early detectors were Li-drifted, that is, Liþions were diffused into the crystal to eliminate the effects of impurities in the Ge These are known as Ge(Li) crystals and need to be maintained permanently

at liquid nitrogen temperatures, both to prevent the Li from becoming mobile in the lattice and aggregating and also to reduce thermal noise Lower energy photons, in the x-ray region, are better detected with lower atomic mass materials, and Si(Li) detectors are used for this purpose In the last 10 to 20 years, ‘‘hyperpure’’ Ge has become available and has supplanted Li-drifted semiconductors to a significant extent So-called HPGe detectors can be made much larger than Ge(Li) crystals could, so that much higher relative efficiencies [> 150% relative to 7.5  7.5 cm NaI(Tl)] are possible These detectors can be stored at room temperature, although they must still be operated at liquid nitrogen temperatures Different detector configurations (‘‘reverse electrode’’ and ‘‘planar’’) are also available and provide improved low-energy performance Both scintillators and semiconductors are available in ‘‘well type’’ configurations, in which the sample tube is inserted into a narrow well in the center of the detector, surrounding the sample with sensitive material This maximizes counting efficiency, but at the cost of severely restricting sample size Large samples of the order of 1.0 liter in volume can be counted with reasonable efficiency in a recessed Marinelli beaker that surrounds the sensitive region of the detector All semiconductor detectors work on the same principle Within the electronic structure of the crystal, there exists an occupied ‘‘valence band’’ of electronic states and, about 3.5 eV higher in energy, there is an empty

‘‘conduction band.’’ When a g-photon interacts with the detector crystal, the dissipation of its energy causes promotion of electrons from valence to conduction bands, creating vacancies (‘‘holes’’) in the valence band Under the influence of the applied voltage, the electrons in the conduction band and the holes in the valence band migrate to electrodes attached to the

crystal The magnitude of the charge collected at the electrodes is proportional to the number of electron/hole pairs created, which in turn

is proportional to the energy dissipated in the detector crystal by the

g-photon Conventional signal processing and pulse height analysis

equipment allow the collection of data as an energy-resolved spectrum Gamma spectra contain a number of characteristic features, some

of which are visible in Fig 4, a spectrum obtained from a sediment sample contaminated with artificial radionuclides from a fuel reprocessing plant.

In Fig 4, A is the full energy photopeak, the signal arising from the dissipation of all the photon energy in the sensitive volume of the detector.

B is the Compton edge Many photons can pass through the detector without dissipating all their energy in the sensitive volume Some of these remove an electron from an atom in the detector (Compton scattering), reducing the photon energy by an amount equal to the electron binding energy If the photon then dissipates some or all of its remaining energy in the detector, its energy will be measured as the reduced value The Compton edge thus arises from photons that have undergone a Compton scattering interaction and then dissipated all their remaining energy in the detector Feature C is due to photons that have suffered partial energy loss through interaction with other materials before interaction with the detector (backscattering).

Other spectral features may also be observed If a photon has a total energy greater than 1022 keV, a process known as pair production can occur, in which the photon’s energy is converted into a positron/electron pair The mass equivalent of 1022 keV is 2 electron masses If one electron

is lost from the detector, then the resulting signal will appear at 511 keV below the full energy peak Similarly, double escape may also occur.

Figure 4 Gamma spectrum of a sediment sample, obtained with a semiconductor detector.

The peak due to this will appear at 1022 keV below the full energy peak and arises from the escape of both the positron and the electron formed

by pair production from the detector.

Annihilation radiation may also be observed This signal always arises

at 511 keV and is particularly prominent in the spectra of bþ

-emitters.

It results from the interaction of the positron with an electron, which causes annihilation (the inverse of pair production) and the production of two photons, each with 511 keV energy Since these photons do not arise from a specific isotope, it is very difficult to use the 511 keV peak in analysis The calibration of semiconductor g-detectors is relatively complex, since they are so often used for the identification and measurement of a very wide range of isotopes, not all of which are available for the preparation of calibration standards These detectors are usually calibrated using a mixture

of g-emitters, present in known activities and covering the energy range from about 50 to 2000 keV Typically, such mixed standards will give spectra containing about a dozen peaks, and both the energy and the efficiency calibrations can be performed using these materials While the energy calibration should be linear, detection efficiency varies with energy in

a complex fashion (Fig 5) As with NaI(T1) detectors, geometry and matrix effects are of critical importance Most commercial g-spectroscopy systems include software for the calculation of calibration equations from mixed isotopic standards It should be noted that error in efficiency calibration is a major source of error in g-spectroscopy, since actual counting efficiencies are low (often 1% or less for large-volume samples), so small errors in calculated efficiencies lead to large errors in the final results The calibration

of semiconductor detectors at low energies (< 50 keV) can be quite difficult, and single isotopes or simple isotope mixtures are often used Photons above around 2000 keV are rarely encountered in environmental measurements.

Figure 5 Variation of detector efficiency with energy, for a semiconductor detector.