Gamma-Gamma Coincidence Timing with Fast

Scintillation Detectors

Final Year Project

By
Benjamin Thomas Kingswood

Department of Physics
School of Electronics and Physical Sciences
University of Surrey

Abstract

Gamma-ray spectroscopy and the measurement of nuclear lifetimes provide an

excellent tool for obtaining information on nuclear structure. In these experiments
two Barium fluoride (BaF2) detectors were chosen for their fast light output
component, which has a decay time of 0.6ns. The BaF2 detectors were
characterised using 60Co, 137Cs and 152Eu and found to have an energy
resolution of 18.5-19.0 %. This value is similar to other published results.
However, it is slightly higher than would be expected, this maybe due to
fluctuations in the supplied voltage, detector variance in the PMT or the crystal’s
intrinsic energy resolution. The timing experiments were conducted to determine
what levels of sensitivity and resolution the BaF2 detectors possess. These timing
experiments looked at both 60Co and 152Eu. The latter of which is of particular
interest as it can provide information on the 2+ to 0+ state transition in 152Sm. The
timing characteristics of the BaF2 detectors are important if they are to be used
for gated coincidence timing experiments. The average timing resolution
obtained was 10.4±0.1% the smallest resolvable time interval obtained was
0.94±0.05ns. These numbers are consistent with the expected performance of
the experimental apparatus.
 Contents

1. Introduction to Gamma – ray Spectroscopy 2

2. Theory 2
2.1. Gamma Rays and their interactions with matter 3
2.1.1 Photoelectric Effect 3
2.1.2 Compton Scattering 3
2.1.3 Pair Production 5
2.2 Scintillation detectors 5
2.3 Detector Characteristics 6
2.4 Barium Fluoride (BaF2) as a scintillation detector 8

3. Experimental Method 10
3.1 Experimental arrangement 10
3.2 Equipment 11
3.3 Detector Characterisation 13
3.4 Timing Experiments 13

4. Results and Discussion 15

4.1 Detector Characterisation 15
4.2 Timing Experiments 16

5. Conclusions 17

6. Acknowledgements 18

7. Appendices 18

8. References 19

1
1. Introduction to Gamma-ray Spectroscopy
Gamma-ray spectroscopy and the measurement of nuclear lifetimes provide
excellent tools for obtaining information on nuclear structure. Gamma
spectroscopy also provides excellent detailed information in the energy spectrum
of incident gamma-rays. This information can be used to not only quantify but
also identify any gamma emitter/s within a material. In addition gamma
spectroscopy can be a valuable tool for understanding the internal structure of
atomic nuclei. Lifetime measurements of quadrupole transitions can also provide
information on nuclear deformation [1] [2].

Barium fluoride (BaF2) detectors were chosen for their characteristics that make
them ideal fast-timing scintillation detectors. The light pulse output from de-
excitation of BaF2 crystals has two components. The first is a slow component,
which has a wavelength of 310nm with a decay time of 630ns [2]. The second is
the fast component, with a wave length of 220nm and a decay time of only 0.6ns
[2]. It is this component that makes BaF2 so ideal, as the 220nm emission makes
BaF2 the fastest known scintillator. Unfortunately the fast component only
accounts for 20% of the total light output of BaF2, the remaining 80% is made up
by the slow component [2].

The purpose of using two BaF2 detectors together in a coincidence arrangement

is to allow the measurement of lifetimes down to the nanosecond scale. In order
to do this, the BaF2 detector will have to be characterised in terms of its energy
resolution and efficiency.

2. Theory
The use of scintillation capable materials is one of the oldest known methods of
detecting and measuring ionising radiation. Despite its age, scintillation is still
one of the most useful techniques for carrying out spectroscopy on nearly all
types of radiation. When a photon or particle of ionising radiation enters a
material capable of scintillation (usually a crystal), some or all of its energy is
transferred into the material and is re-emitted as electromagnetic energy. This
electromagnetic energy is usually within the visible light spectrum and may
sometimes extend into the ultraviolet. In a detector the emitted photons of light
will then be transferred via means of a ‘light guide’ to a photomultiplier tube
(PMT). Within the PMT a series of cascade reactions occur increasing the
intensity of the light pulse, which is then finally converted into an electrical pulse
[1].

2
2.1. Gamma Rays and their interactions with matter
Due to their high energy gamma-ray photons induce little direct ionisation.
However, they can transfer their energy to a material by interacting with charged
particles within the material and thus produce scintillation events. The three
principle means by which gamma-ray photons can interact with a material are the
photoelectric effect, Compton scattering and pair production [1].

2.1.1 Photoelectric Effect

The photoelectric effect is the process by which an incident gamma photon
interacts with an ‘absorber atom’ transferring all its energy to a bound electron
within the atom. This energetic ‘photoelectron’ is no longer bound and therefore
expelled from the atom. This process cannot take place with free electrons. The
energy of these emitted photoelectrons (Ee) is given by

Ee = hv − Eb (1)

Where Eb is the binding energy of the electron, prior to interaction with the
gamma photon. For Gamma photons incident energies of several hundred keV
are typical.

The excited atom now has a vacancy in one of its electron shells this can be filled
in a combination of one or two ways, either the atom absorbs a free electron or
the remaining bound electrons reorganise themselves within the bound shells. It
is possible for an electron in a higher shell to drop down into the vacancy in a
lower shell. This results in the emission on a characteristic X-ray photon. This is
potentially a problem as these X-ray photons can cause a count to be recorded in
any detector being used. The photoelectric effect is the most common
mechanism by which lower energy gamma photons may interact with a material
[1].

2.1.2 Compton Scattering

Compton scattering describes the deflection of a gamma-ray photon by an
electron within a target material. This interaction produces a lower energy photon
and a ‘recoil electron’. The amount of energy transferred to the recoil electron is
dependant on the angle through which the photon is deflected (θ).

3
 Recoil Electron

Φ
Incident Photon

energy = hv
θ

Scattered Photon
energy = hv’
Figure1 Diagram of the Compton scattering process [3].

By considering the conservation of energy and momentum it is possible, from

figure 1, to derive an expression relating incident photon energy (hv) to scattered
energy (hv’). i.e.

hv
hv' = (2)
hv
1+ (1 − cos θ )
m0 c 2

The rest mass of the electron is m0c2 (0.511 MeV). From equation 2 we can see
that, since θ is continuous so must be the energy of the scattered photon (hv’).
As the scattering angle θ is continuous so is the amount of energy transferred to
the recoil electron. This can be seen in an acquired spectrum, see example
below.
θ =0 θ =π Eγ

dN
dE

Compton continuum Energy

Compton Edge
Figure 2 The Compton Continuum

4
2.1.3 Pair Production
Pair production can only take place within the Coulomb field of a nucleus and is
limited to gamma-ray photons, with energy greater than twice that of the rest
mass of an electron (1.02 MeV). The process entirely converts the photon and all
its energy into an electron-positron pair. If the photon has energy of exactly 1.02
MeV pair product can still occur, however the particle anti-particle pair will
immediately annihilate upon creation. All excess energy above 1.02 MeV is
converted into kinetic energy of the electron-positron pair [1].

2.2 Scintillation detectors

Scintillation materials produce three different types of emission.

• Fluorescence describes the prompt emission of a photon in the visible

spectrum from an excited subatomic state. This is normally the most
desirable.
• Phosphorescence is again normally within the visible spectrum but has a
longer wavelength and characteristic decay time.
• Delayed fluorescence as the name implies, results in emissions much like
fluorescence but with a longer characteristic decay time.

For a material to be considered a good scintillator it must convert a large fraction

of incident radiation into prompt fluorescence [3].

In a scintillation detector the photons emitted by the scintillation crystals have to

be converted into an electrical pulse for counting. This is done by means of a
photomultiplier tube. The emitted photon pluses are weak and are directed from
the crystal to the photocathode of the PMT by means of a light guide or optical
window. Once the pulse reaches the photocathode a proportional number of
electrons are emitted. The emitted electron pulse is then amplified (normally 10-
12 times) by a series of dynodes within the tube. Therefore, for each electron
emitted at the photocathode, 106-108 electrons arrive at the anode located at the
tubes base [1].

5
 Scintillation
crystal Multi-channel analyser
(MCA)
Photocathode

Amplifier
PMT,
Dynodes HV
supply
Preamplifier

Anode

Figure 3 Diagram of the general experimental arrangement for gamma-ray

spectroscopy [4].

As shown in figure 3, once the electrical pulse arrives at the anode it is fed into a
pair of amplifiers to shape the pulse and typically boost it from a few mV to the
order of a few volts. The signal can then be processed by a multi-channel
analyser. The PMT is comprised of two components. The first is the
photocathode, which is a layer of photosensitive material that will emit electrons
when exposed to the optical pulses from the scintillation crystal. The second is
an arrangement of positively charged dynodes required to increase the number
of electrons in the pulse produced by the photocathode. This is necessary as the
initial pulse of electrons is fairly small and cannot easily be measured. The size
of pulse received at the anode is typically about 1010 electrons [1].

2.3 Detector Characteristics

The three most important properties of a scintillator are efficiency, energy

resolution and decay time.

Energy resolution

A detector’s ability to resolve the small differences in the energy of incident

photons is referred to as its energy resolution (R). The smaller the energy
resolution the better a detector is able to distinguish between two photons of
similar energies. The peaks on a spectrum all have a finite width. From this the
full width half maximum (FWHM) can be determined.

6
 dN
dH

Y/2
FWHM

H0 H

Figure 4. Definition of detector resolution [3].

Energy resolution as a percentage can therefore be found by

FWHM
R= × 100 (3)
H0

Where FWHM and H0 are both expressed in terms of channel number. It is

important to note that energy resolution not only varies for different detectors but
also at different incident radiation energies. Energy resolution can also be
affected by such things as electrical noise, drift and fluctuations in the baseline
voltage supplied to the detector [1].

Detector efficiency

In principle all radiation detectors will produce an output pulse for every quantum
of radiation that interacts with the detector’s volume. However, uncharged
gamma-rays can travel a significant distance through a material before
interacting. Therefore, due to the large distance between interactions it is normal
for gamma-ray detectors to be less than 100% efficient. A precise figure for
detector efficiency can be calculated in two ways. The first is an absolute value
defined as

Eabs = ___ number of pulses recorded________ (4)

Number of radiation quanta emitted by source

7
The second is the intrinsic value defined as

Eint = _______number of pulses recorded_________ (5)

Number of radiation quanta incident on detector

[3]

Decay time

The decay time characterises the prompt scintillation yield of a material and its
time of recovery before the material returns to its unexcited state and can
undergo another scintillation event [3].

2.4 Barium Fluoride (BaF2) as a scintillation detector

Barium Fluoride is an inorganic scintillator with a high atomic number (Z=56) and
a density of 4.88 gcm-3. This means BaF2 has high detection efficiency with a fast
response time. The scintillation light from BaF2 has two components, one ‘fast’
the other ‘slow’. The decay time for the ‘fast component’ emitted in the short
wavelength region of the spectrum is only 0.6-0.8 ns. However the ‘slow
component’ in the longer wavelengths has a much longer decay time of 630ns.

The ‘fast component’ is within the ultraviolet region of the spectrum and is
therefore difficult to detect with most PMTs. In order to detect the short
wavelength the PMT requires the use of a quartz end window tube or some other
non-standard light detector [1] [3] [4].

8
Figure 4. The attenuation coefficients for BaF2 as a function of gamma photon
energy [5].

9
The main advantage of using BaF2 is its incredibly short decay time. At about
0.6ns it is a full order of magnitude shorter than any other scintillation crystal in
common use (see Table 1). This makes BaF2 an extremely useful detector for
certain investigations.

Inorganic Scintillation Crystal Decay time (ns)

BaF2 0.6-0.8
CsI 2, 20
CeF3 5, 27
LuAP 17

Table 1. Decay time for some inorganic Scintillators (note. Both BaF2 and CsI
have ‘slow’ components) [1]

3. Experimental Method

3.1 Experimental arrangement

The experiment made use of two identical BaF2 detectors. Each consisted of a
BaF2 crystal optically coupled to a PMT and bound in black tape to prevent
ambient light from triggering scintillation events within the crystal. Both detectors
had three connectors on their rear, high-voltage (HV), anode and dynode.

Figure 5. Photograph of PMT and connections.

The detectors were securely mounted to the table, facing one another and with
the source held within a stand positioned centrally between the two detectors
10cm from either face. Lead shielding was used to protect the operator when
accessing the nearby computer.

10
 Figure 6. Photograph of the two BaF2 detectors in place within the experimental
arrangement.

Both detectors had their own independent HV supply, capable of supplying

between 0 and 3000 volts. The HV supply, amplifier, Time to amplitude converter
(TAC), Constant fraction discriminator (CFD) and Delay modules were all
mounted in the NIM.

Figure 7. Electronics in NIM module

3.2 Equipment
High voltage (HV) supply

It was essential that the HV supplied at least 2000 volts, while at the same time
keeping the current flow to a few milliamps if it is to be used with the BaF2
detectors. Voltages of less than 2000 volts will fail to adequately activate the
BaF2 detectors. The supply also had to be well regulated to prevent drifts in the
supplied voltage from affecting the experiments. The HV supply is operated by 3
controls; the first could be adjusted from 0V to 100V in 1V divisions, the second
from 0-400V in 100V divisions and the third from 0-2500V in 500V divisions.

11
Preamplifier (preamp)

The purpose of the preamp is to avoid changes that can occur to the time
constant of a pulse if the PMT is directly connected to the main/linear amplifier.
The preamp has several ways of maximising its signal to noise ratio. The first is
simply to attach it as closely as possible to the detector itself. The second is to
terminate its capacitance rapidly. However, the preamp does not shape the
received pulse and has a short output time with a linear tail.

Amplifier

The main/linear amplifier is used for two reasons. The first is to produce a gain in
amplitude; the second is to shape the tail pulse from the preamp into a linear
pulse. The amplification factor can be modified using both fine and coarse gain
controls. In addition, the shaping time can also be adjusted and the affect of this
will vary with differing rates [1].

Constant fraction discriminator (CFD)

The CFD function is to improve the timing performance of the experimental

setup. It does this by selecting a point on the received pulse that is independent
of amplitude. Ideally this would be the leading edge of the pulse for the greatest
increase in performance. The CFD achieves this by first inverting and delaying
the inputted signal and adding this signal to attenuated prompt signal. This
results in a bipolar signal with zero crossing.

Delay Box

The delay box can be used to introduce a delay in the signal pulse of up to 63ns.
This is achieved easily using cables of differing length.

Time to amplitude converter (TAC)

The TAC is required to make the conversion from energy to time intervals. It
must work in conjunction with the Multi-channel analyser (MCA). The TAC
coverts the duration of a received pulse into a new pulse of proportional
amplitude, which can in turn be analysed by the MCA. Therefore, each channel
on the MCA no longer represents an energy window, but instead an interval of
time. The TAC used is of the start-stop type. The start pulse initiates circuit action
by providing a constant current source. This current is switched off by the stop
pulse. Two coincidence pulses are required to both start and stop the TAC. Both
are fast and negative. The TAC’s output is slow and positive. The stop pulse will
have a larger decay due to the decay box. The output pulse of the TAC is
typically measured in milliseconds [1].

12
Multi channel analyser (MCA)

The PCA –II and PCA –III MCAs were used during these experiments. While
both packages are very similar version III is required for the more advanced time
acquisitions.

The data produced by the MCA is produced in graph format i.e. a one
dimensional array. The MCA when used for the characterisation of BaF2, displays
the number of counts recorded per channel. The MCA was then calibrated to
show the number of counts per energy. The PCA –III was set to coincidence
mode and so would only display counts within a gate, as defined by the input
channel, or out side the gate in anti-coincidence mode [1].

3.3 Detector Characterisation

In order to characterise the two detectors, three different sources were used,
60
Co, 137Cs and 152Eu. This provided an energy range of 1210.73 keV (from
121.78 keV to 1332.51 keV). The gain on the amplifier was set to ensure the full
spectrum was visible on the MCA display array. A series of experiments was
conducted to determine the best voltage for the detectors.

HV Supply

BaF2 Scintillation
Crystal Amplifier

PMT Preamp
0.1 m
MCA
Source

Figure 8. Diagram of the experimental arrangement for the characterisation of the

BaF2 detectors.

3.4 Timing Experiments

The timing experiments were conducted to determine what the levels of
sensitivity and resolution are for the BaF2 detectors. To this end the equipment
was configured as shown in figure 9. Two sources were used for these

13
experiments 60Co and 152Eu as both have short lifetimes. The gamma-ray
coincidences of 152Eu should be relatively easy to observe. The coincidence
occurs between the 1408 keV and 121 keV gamma-ray photons. These are
emitted from the 2+ state of 152Sm and the 2+ state of 152Eu respectively (see
appendix figures 14 and 15). 60Co has a lifetime of similar length to that of the
decay time of BaF2 detectors it therefore tested the limits of this detector.

High Voltage High Voltage

Supply Source Supply

Detector Detector
1 2

Preamp Preamp

Amp Amp

CFD

Delay
Box

TAC

MCA

Figure 9. Experimental Arrangement for the gamma-gamma coincidence timing

experiments

14
4. Results and Discussion

4.1 Detector Characterisation

Spectrums were acquired for both detectors across a variety of voltages ranging
from 2000 to 2600volts. The gain had to be varied in each case to ensure that
the spectrum for each source fitted neatly into the data array of the PAC
programs. This allowed the best settings for the detectors at varying energies to
be determined. Figure 10 shows an example composite spectrum obtained at
2000 volts from the HV supply with a gain of 750 and a shaping time of 6µs.

BaF2 Resolution test at 2000 volts with Co-60, Cs-137 and Eu-152

1800

1600

1400

1200
Counts

1000

800

600

400

200

0
0 200 400 600 800 1000 1200 1400 1600
Energy in KeV

Figure 10. Composite 60Co, 137Cs and 152Eu spectra used in characterisation

Energy resolution
The graph below shows the variation of detector resolution with applied voltage.
It can be seen that resolution of a single material is far more dependant on the
source and energy of the radiation measured than the voltage supplied to the
dynode of the PMT. Provided the voltage is sufficient to activate the PMT. The
energy resolution for the BaF2 detector was found to be 18.7 ± 0.2 %.

Given modern manufacturing techniques resolution losses due to imperfections

within the crystal are likely to be very minor. However, the rough surface of the
crystal is another matter and could have a significant effect on the resolution of
the system. In addition to this variances induced by the PMT can contribute
significantly to any loss of resolution. These variances include the loss of photon
uniformity at the photocathode and statistical variations in the effectiveness of the
electron multiplier components [1] [2].

15
 Energy Resolution for a BaF2 detector using various sources

25.00%
Eu-152 121.78 keV Eu-152 344.27 keV
Cs-137 661.64 keV Co-60 1172.23 keV
Co-60 1332.51 keV
20.00%
Resolution as a percentage

15.00%

10.00%

5.00%

0.00%
2000 2200 2400 2600
High voltage supply (volts)

Figure 11. Energy resolution for BaF2 detectors

4.2 Timing Experiments

The first timing experiment was conducted using 60Co. It was unknown what level
of resolution would be achieved in this experiment as the decay time for BaF2 is
very close to the lifetime of the gamma-gamma coincidence for 60Co. However,
as figure 12 below shows, 7 distinct Gaussian peaks were visible. The peaks
were separated by a delay of 8ns and have an average FWHM of 0.94ns which
is just above the 0.6ns decay time of BaF2. As with energy, resolution varied
along the x-axis with an average value of 4.54%

Coincidence Timing with Co-60

40

35

30

25
ounts

20
C

15

10

0
0 10 20 30 40 50 60
Delay (ns)

Figure 12.Gamma-gamma coincidence timing with 60Co

16
The second coincidence timing experiment, conducted using 152Eu, required the
peaks to be separated by 16ns in order that they were clearly distinguishable.
This time the peaks were not Gaussian as before but rather they had an
exponential leading edge. The resolution was lower, with an average of 16.3%.
The calculated average FWHM was 3.24ns seconds which fitted with known
facts as it is longer than that of the 60Co.

Coincidence timing with Eu-152

50

45

40

35

30
Counts

25

20

15

10

0
0 10 20 30 40 50 60 70
Delay (ns)

Figure 13.Gamma-gamma coincidence timing with 152Eu

5. Conclusions
The BaF2 detectors were characterised using 60Co, 137Cs and 152Eu sources. This
was done by acquiring several spectra at varying voltages from the high-voltage
supply. It was found that varying the voltage had little effect on the resolution of
detectors. However, the varying voltage did have an effect on the position of the
centroid for the acquired spectra. In order to compensate for this and ensure that
the full spectrum was visible, the amplifiers gain also had to be varied to obtain
all energies from 121 keV to 1333 keV within the available channels. It was
observed that, the variation of the gain did not have a linear effect on the
distribution of the spectra obtained. The Barium Fluoride detector was
characterised determined to have an energy resolution of 18.5-19.0 %. This
value is a similar to other published results. However, it is slightly higher than
would be expected, this maybe due to fluctuations in the supplied voltage,
detector variance in the PMT or the crystal’s intrinsic energy resolution.

The detectors were used to evaluate the coincidence timing potential of BaF2
scintillation material. Timing experiments looked at both 60Co and 152Eu. The
latter of which is of particular interest as it can provide information on the 2+ to 0+
state transition in 152Sm. This takes place when a gamma ray populates the 2+ of
152
Sm. This state then depopulates returning to 0+ via the emission of a 121keV
gamma ray. The timing characteristics of the BaF2 detectors are important if they

17
are to be used for gated coincidence timing experiments. The average timing
resolution obtained was 10.4±0.1%. The smallest resolvable time interval
obtained was 0.94±0.05ns.

There are several methods that could be used to improve the results obtained by
these experiments. As all the data/spectra is dependant on the number of
acquired counts then by acquiring more counts we can improve the quality of the
derived curves. The three most effective ways to achieve this are higher activity,
longer acquisition times and more detectors. I would tend to dismiss the use of
high activities due the risks of working with high energy activity sources. Time
and detectors should be ample means of improving the results obtained.

6. Acknowledgements

I would like to thank my supervisor Dr P Regan for his help assistance during this
project. I would also like to thank James Wallbank and Francois kazadi for their
assistance with the practical aspects to this project.

7. Appendices

Figure 14. 152Eu to 152Sm decay Scheme.

18
 Figure 15. 152Eu to 152Gd decay scheme.

8. References
[1] Amy Catherine Nisbet; Characterisation and Lifetime Measurements with
BaF2 Detectors; Department of Physics, School of Electronics & Physical
Sciences, University of Surrey 2006

[2] Aage Bohr, Ben R. Mottelson; Nuclear Structure, volume 1; World

Scientific Publishing Co. Ltd 1998

[3] Glenn f. knoll; ‘Radiation Detection and Measurement’; Third edition, John
Wiley & Sons, Inc 2000

[4] Burcham and Jobes; ‘Nuclear and Particle Physics’; Longman Group
Limited, 1995

[5] M.J. Berger,1 J.H. Hubbell,1 S.M. Seltzer,1 J. Chang,2 J.S. Coursey,2
R. Sukumar,2 and D.S. Zucker2, 1National Institute of Standards and Technology,
Physics Laboratory, Ionizing Radiation Division
2
NIST, Physics Laboratory, Office of Electronic Commerce in Scientific and Engineering
Data, XCOM: Photon Cross Sections Database (version 3.1)

http://physics.nist.gov/PhysRefData/Xcom/Text/XCOM.html (Accessed
February 2007)

[6] Kenneth S. Krane; ‘Introductory Nuclear Physics’; John Wiley and sons,
Inc 1988

[7] S.J. Williams et al., Journal of Physics G31 (2005) S1979

19