Lowenthal G., Airey P. Practical Applications of Radioactivity and Nuclear Radiations
Подождите немного. Документ загружается.
(Section 5.4.2). These effects cause the ef®ciencies of LS counting to fall well
short of 100% notwithstanding the absence of source self-absorption. Losses
in counting ef®ciency are larger the lower the average energy of the emitted
radiation.
4.3 Gamma ray applications
4.3.1 The role of electronic instruments
When ionising radiations interact in a detector their energies are converted
into signals which operate ratemeters or scalers or multichannel analysers.
Signals are also stored for subsequent computer processing. This section
offers a brief introduction to the principal functions of nuclear instrumenta-
tion normally used for routine applications of g ray emitters detected with
NaI(Tl) crystals. Figure 4.1 is a schematic diagram of a NaI(Tl) detector
which illustrates the three g ray interactions (g
1
, g
2
, g
3
) described in Section
3.4.3 and shown in Figure 3.7(a). Figure 4.2 shows a block diagram of the
associated signal processing equipment.
The cathode ray oscilloscope (CRO) serves to monitor pulses at any of the
terminals. The very short negative or positive going parts of the tail pulses
shown in the ®gure are due to the rapid collection (within a fraction of a
microsecond) of the very light and mobile photo-electrons emitted from the
photocathode of the PM tube and multiplied in the strong electric ®eld across
its dynode chain. The very much heavier positive ions, the molecules which
lost an electron, take some 10
4
times longer to be collected. This is illustrated
by the long tails which should be very much longer than could be shown in
the ®gure.
The preampli®er is built into the base of the PM tube adjacent to the
output terminal of the detector. It serves to process and transmit the range of
pulses (down to submillivolt heights) generated by ionising radiations in the
detector and this without signi®cant pulse degradation or deformation which
would occur in long cables (exceeding about one metre) if pulses which were
not pre-processed were allowed to reach the ampli®er.
The preampli®er also reduces the length of the tail pulses (clipping), so
reducing pulse pileup. Moreover, it provides an input for test pulses (Figure
4.2) from a precision pulser, the output of which serves to monitor the proper
functioning of all components of the system and to determine corrections for
random pulse summing (Section 4.5.3).
The ampli®er is the core of the pulse processing equipment. Spectroscopy
ampli®ers provide highly stable and linear ampli®cation, but the ampli®ed
4.3 Gamma ray applications 101
Nuclear radiations: a user's perspective102
Figure 4.2. A block diagram of equipment used for the signal processing of pulses
due to g rays detect ed in a NaI(Tl) crystal (see text, Section 4.3.1 for explanations).
The pulser and dead time gate (DT) are rarely required for routine work.
Figure 4.1. Gamma ray interactions in a NaI(Tl) detector shown with the asso-
ciated electronics. The photoelectric effect (g
1
), Compton scatter (g
2
) and pair
production (g
3
) are illustrated.
pulse at the output should not exceed 10 V to avoid ampli®er overload.
Another function of spectroscopy ampli®ers is to shape the tail pulses from the
preampli®er into Gaussian-type pulses (Figure 6.11). Tail pulses are not always
readily accepted by the pulse analysers that follow the ampli®er (Figure 4.2).
Electronic instrumentation gives rise to electronic noise pulses which have
to be kept well below pulse heights due to signals generated by ionising
radiations in the detector. Higher noise levels picked up from external sources
can be greatly reduced by connecting all parts of the equipment to a common
earth terminal with thick copper leads which should be as short as possible.
The single channel analyser (SCA) consists of a lower level discriminator
(LLD) topped by an adjustable gate or window (W) or independently
adjustable lower and upper level discriminators (LLD, ULD). In Figure 4.3,
the LLD and ULD were set to obtain the countrate in the 1173 keV peak of a
60
Co spectrum (Section 4.4.3).
4.3.2 NIM bin and portable equipment
Signal processing instrumentation for signals from nuclear radiation detec-
tors is normally used as NIM bin equipment (Nuclear Instrument Modules),
4.3 Gamma ray applications 103
Figure 4.3. A measurement of the countrate in the 1173 keV full energy peak of a
spectrum due to
60
Co, using a single channel analys er (SCA), to de®ne the window
V
1
to V
2
(Mann et al., 1980, Figure 7.13).
an internationally adopted system. The frame for the modules ®ts up to
twelve single width units and comes with or without power supply, with
frame and modules fabricated in accordance with strictly de®ned speci®ca-
tions. Modules are available for each of the many different functions required
for signal processing: power and high-voltage supplies, ampli®ers, discrimina-
tors, SCAs, timers, scalers, pulsers and other more specialised instruments
described in the catalogues of manufacturers.
With electronic equipment continuously being reduced in size, modules are
now of reduced height (`slimline'), but existing frames and power supplies
remain serviceable.
Numerous applications of nuclear radiations include investigations `out
of doors', requiring portable and rugged equipment which may have to be
operated in a dif®cult environment. Field work employs robust, battery-
powered stand-alone units incorporating the preampli®er, ampli®er, LLD
and scaler±timer and sometimes a single-channel analyser. Signal proces-
sing units powered from the mains are also used in industry and in school
and undergraduate laboratories, as are many components of NIM equip-
ment.
4.3.3 Comments on instrumentation and its supply
The range of nuclear instrumentation available for industrial and other
purposes is much larger than is often realised. The following are just a few
comments on the characteristics of radiation detectors and in particular
scintillation detectors. The NaI(Tl) crystal detector is well known, ef®cient
and stable, but it is only one scintillator among numerous others which have
been developed over the past two decades. Several of the more recently
designed detectors exceed NaI(Tl) in density and luminosity and are non-
hygroscopic. Readers should obtain full information on these detectors from
manufacturers, several of whom are listed below. Others can be identi®ed
from advertisements in the technical literature.
Thallium-activated NaI (0.2% Tl by weight, r = 3.67 g/cm
3
) is probably
still the most widely used material for g ray detection. It is, however, strongly
hygroscopic and so has to be protected from atmospheric moisture, as well as
from mechanical damage. This has to be done while minimising any reduction
in the ef®ciency with which the detector responds to incoming radiation
(Section 4.4.1).
Silver-activated ZnS crystals serve to detect a particles, being too small to
be suf®ciently sensitive for g radiations. Since the background is nearly 100%
due to g rays (Section 4.6.4), these small crystals can be used unshielded,
Nuclear radiations: a user's perspective104
assuming that there are no other radiations likely to cause interference with
the signals of interest.
CsI(Tl) crystals are only slightly hygroscopic but, other things being equal,
they have barely half the light output of NaI(Tl) crystals. Scintillators are
also made from organic crystals dissolved in a suitable solvent, notably
anthracene and trans-stilbene, which are useful b particle detectors of very
short dead time (<0.1 ms) but of too low a density (1.1 to 1.3 g/cm
3
) for
routine g ray detection, though there are the scintillation detectors referred to
earlier. Plastic scintillators loaded with a few per cent of
6
Li or
10
B atoms are
useful detectors for neutrons.
Manufacturers developed models of NaI(Tl) detectors which can be used
under water or can tolerate mechanical vibrations or shocks or temperatures
up to 200 8C. Progress has been made with high-density scintillators, which
are ef®cient detectors for high-energy g rays and which have come into
widespread use over the past few years. In particular, BGO, a compound of
oxides of bismuth and germanium, is commonly used.
The high Z number of bismuth (Z = 83) and the high density of bismuth
germinate crystals (7.4 g/cm
3
, twice the density of NaI) make this crystal a
much more ef®cient detector of higher energy g rays than is NaI(Tl). It is fully
transparent to its own radiation, non-hygroscopic (requiring no airtight seal),
and easily machined, but it is expensive and its light output is only about 15%
of that obtained from NaI(Tl) crystals in similar conditions. BGO is widely
employed for the tomographic imaging of 511 keV annihilation g rays
(Section 1.4.4). Other high-density phosphors are being developed to obtain a
higher light output than is available from BGO crystals. High-resolution
detectors will be discussed in Section 5.5.
Information about developments in instrument technology is regularly
published by suppliers of such equipment in their highly informative cata-
logues. These ®rms advertise extensively in the scienti®c literature and many
of them maintain service departments all around the globe. Here it is only
possible to offer a very small selection without prejudice. For example:
. Canberra Industries Inc., and its subsidiary, Packard Instrument Company Inc.,
Meriden, CT 06450, USA.
. PerkinElmer Instruments, Oak Ridge, TN 37831-0895, USA (formerly EG&G
Ortec Nuclear Instruments).
Other manufacturers specialise in radiation detectors and pressurised
ionisation chambers.
. Bicron Scintillation Products, 12345 Kinsman Rd, Newbury, OH 44065, USA.
. NE Technology Ltd, Reading, Berkshire RG7 5PR, England.
4.3 Gamma ray applications 105
. Centronic Ltd, New Addington, Croydon CR9 0BG, Surrey, England
. Nuclear Associates, PO Box 349, Carle Place, NY 11514-9895, USA
. Capintec Inc, Arrow Rd, Ramsay, NJ 07446, USA.
Radiochemical suppliers include:
. Amersham International plc, Li ttle Chalfont, Buckinghamshire HP7 9NA ,
England
. Mallinkrodt International, 675 McDonnell Blvd, St Louis, MO 63134, USA
. Nordian International Inc, March Rd, Kanata, Ontario, Canada K2K 1X8.
4.4 Gamma ray counting with NaI(Tl) detectors
4.4.1 Further comments on NaI(Tl) detectors
NaI(Tl), an inorganic scintillation detector
When ®rst employed, scintillation detectors were commonly organic crystals
of low density, e.g. naphthalene, r = 1.14 g/cm
3
, suitable mainly for the
detection of b particles and lower energy X rays, up to 10 keV or even less.
During the 1940s it was discovered that scintillations triggered in crystals of
NaI could make them useful g ray detectors because their density was
adequately high (r = 3.67 g/cm
3
). Also, the addition of 0.2% by weight of
thallium (Z = 81) made even very large NaI crystals (Figure 4.4) fully
transparent to the light quanta generated by g rays throughout their volume.
However, the crystals have to be kept free from moisture and there are other
fabrication problems to be overcome.
Selected characteristics of integral assemblies
Routinely used NaI(Tl) detection systems are often known as integral line
assemblies or just integral assemblies, combining the cylindrical detector, the
PM tube and the preampli®er in a metallic shield. A frequently used design is
to protect the cylindrical surface of the crystal (Figure 4.1) by a 2±3 mm
thick aluminium sheet the inner surface of which is pre-treated to maximise
its re¯ectivity. A 3 mm thick Pyrex or quartz window, selected to optimise the
transmission and focussing of light pulses, connects the NaI crystal to the PM
tube, the latter being protected by an outer Mumetal shield which extends
around the preampli®er.
The entrance window is made as thin as practical to transmit photons with
minimum attenuation. For laboratory applications the thinnest and rather
fragile window is 0.25 mm beryllium, which transmits 5 keV X rays with
close to 80% ef®ciency. Windows for detectors used in the ®eld are made
Nuclear radiations: a user's perspective106
from 0.25±0.5 mm Al, the former transmitting 20 keV photons with 80%
ef®ciency. The corresponding ef®ciency of a 0.5 mm thick window is 60%
(Hubbell 1982, see also Bicron catalogues). All assemblies are engineered to
optimise light collection while minimising light losses.
On average it is necessary to absorb about 0.3 keV of g ray energy in the
NaI(Tl) crystals to produce a countable voltage pulse at the anode (Figure
4.1). The light output from NaI(Tl) crystals peaks at a wavelength near
430 nm (at the blue end of the optical spectrum), the wavelengths at which
the PM tube is at its most sensitive (Neiler and Bell, 1966).
Total ef®ciencies and peak-to-total ratios
The ef®ciency with which NaI(Tl) detectors respond to g rays increases with
their volume, as shown in Figure 4.4. The ef®cient detection of high-energy
radiations (>1 MeV) is favoured by larger detectors (75 mm675 mm or
larger). Lower energy radiations (<200 keV) can be ef®ciently detected by
25 mm625 mm or smaller detectors with a compensating advantage that the
smaller the crystal the lower the countrate due to the background.
To ensure as high a g ray detection ef®ciency as practical for applications
in the ®eld, it is normal practice to include the entire pulse height spectrum
(above a low LLD setting) in the count, a procedure known as integral
4.4 Gamma ray counting 107
Figure 4.4. Total detection ef®ciencies of NaI(Tl) detectors for collimated g rays as
functions of the dimensions of the crystals in millimetres (shown on the curves) and
the en ergies of the g rays (from catalogues of manufacturers).
counting (Section 4.4.2). Integral counting normally requires shielding of the
detector, preferably behind 50 mm thick lead to reduce the background
count which could exceed 400 cps in unshielded 50 mm650 mm detectors.
Procedures to realise spectra with optimum peak-to-total ratios (Figure 4.5)
will be referred to in Sections 4.4.3 and 4.4.4.
4.4.2 Integral counting
Gamma rays are emitted with a precisely de®ned energy (Section 3.5.3), but
their detection results in a spectrum from near zero to the upper edge of the
Nuclear radiations: a user's perspective108
Figure 4.5. Peak-to-total ratios for NaI(Tl) detectors as functions of the dimensions
of the detectors and energies of the g rays. These ratios apply for narrow beam
geometry (Figures 4.6(a)) but depend only weakly on the source±detector distance
(Neiler and Bell, 1966).
highest full energy peak (Figures 3.9(b), (c)), with many spectra due to
multiple g ray emitters and so including large numbers of peaks (Section
3.7.3). When a NaI(Tl) detector is required to detect the g rays from a
radionuclide with maximum ef®ciency, the ampli®er gain and the discrimi-
nator levels should be adjusted to ensure that countrates include the entire
pulse height spectrum generated in the detector (Section 3.5.3), assuming
there is no overlap with other spectra.
The low-energy limit for photons to trigger a NaI(Tl) detector is ®xed by
the thickness of the entrance window, but low-energy pulses are also due to
interactions within the detector material. They are consequences of Compton
scatter and photoelectric effects, which account for the low-energy section of
g ray spectra, e.g. the
137
Cs±
137m
Ba spectrum shown in Figure 3.4(d),
together, in this case, with ¯uorescent X rays (Section 3.9).
Integral counting offers two useful advantages ± a high counting ef®ciency
and the use of the relatively simple electronics shown in Figure 4.2 which are
available in NIM bins or as a compact, easily portable unit. Before applying a
high voltage (HV) across the PM tube the correct polarity must be ensured. It
is helpful, if not essential, to have a CRO available to monitor the output.
The operating voltage range (the plateau) for these detectors is normally at
least 200 V, beginning between 800 and 1200 V. If it is known to begin near
800 V, the HV is set initially to 400 or 500 V and then increased, taking
counts every 50 V. The countrate increases sharply to begin with, but as the
operating voltage plateau is reached the rate of increase should be down to
about 1% per 100 V, when all pulses due to g rays exceed the LLD and are
counted. The g ray source used for the setting up procedure should be placed
a few centimetres in front of the detector window and cause a countrate
between 400 and 800 cps at the operating voltage.
The growth of pulses can be monitored using the CRO. On reaching the
operating voltage (the plateau), the pulses should be of nearly identical
height, allowing for random variations. Noise pulses should be preferably at
least an order of magnitude smaller. The plateau ends at an HV value when
noise pulses are suf®ciently ampli®ed to pass the lower discriminator, so
causing a rapidly increasing countrate.
Having identi®ed the position of the voltage plateau, the operating voltage
should be selected within the plateau, say half way along, bearing in mind
that pulse heights at the ampli®er output must remain below 10 V. For
example, to count a
60
Co spectrum (Figure 4.3) the pulse heights should not
exceed 0.6 V per 100 keV of g ray energy, since the upper edge of the 1332
keV peak would then be close to about 9.0 V, rather close to the 10 V limit.
If an upper-level discriminator is used, the cut-off energy should be high
4.4 Gamma ray counting 109
enough not to interfere with the spectrum. It is also important to bear in
mind that pulse heights from a NaI(Tl) detector are only an approximately
linear function of the g ray energy (Section 6.4.4).
4.4.3 Peak counting
Peak counting is the term used when only the g rays that generate pulses
within one or more selected full energy peaks are counted.
For routine applications peak counting normally employs a NaI(Tl)
detector and an SCA. Adding a CRO and a precision pulser is likely to lead
to more satisfactory results (Section 4.4.2). If a semiconducting detector is
used, a multichannel analyser (MCA) and computerised signal processing
equipment are required to cope with the large amount of data. If at all
possible, peak counting should be done in narrow beam geometry, since
pulses due to scattered g rays are lost from the peak. If scattered g rays are
unavoidable, the measurements should be taken in conditions where con-
tributions from scattered g rays can be reproduced as closely as possible
(Section 4.4.4).
Assuming that the 1173 keV peak from a
60
Co source is to be counted
(Figure 4.3), the ampli®er is used to place the centre of this peak close to 5 V,
so keeping the spectrum well below 10 V. Next, the SCA is set for a 0.10 V
window (about 0.23keV). The window is moved across the peak, taking short
counts every 0.2 V and is then set to the position where the short count was a
maximum which is the centre of the peak. Next, the window is opened to
encompass the range V
1
to V
2
, when peak counting can be effected between
these limits. For the spectrum shown in Figure 4.3 the backscatter peak is all
but absent, indicating that pulses due to scattered g rays had been minimised
(see also Section 4.4.4).
Given the relatively low resolution of NaI(Tl) detectors, the positions of the
low- and high-energy edges of full energy peaks (and therefore the positions of
V
1
and V
2
) are subject to uncertainties that could reach several per cent but
could be greatly reduced when working with the well de®ned peaks due to
high-resolution germanium detectors (Figure 3.9(b)). However,
60
Co peaks
can often be counted without shielding, even in a laboratory, since at these
high g ray energies the background rate is almost negligible (Section 4.6.4).
4.4.4 Precautions to avoid errors due to Compton scatter
Integral and, in particular, peak counting yield the most accurate results if
carried out in near scatter-free conditions, i.e. subject to narrow beam
Nuclear radiations: a user's perspective110