EXPERIMENTS

WITH
G.M. COUNTER
 Experiments #3: Dead time and nuclear counting statistics

Objective:

1. Measurement of dead time

2. To investigate the statistics related to measurements with a Geiger counter:
counter Poisson and Gaussian distribution
distr

Apparatus:
• Set-up for ST-350 Counter

• GM Tube and stand

• Shelf stand, serial cable, and a source holder

• Radioactive Source (e.g., Cs-137, Sr-90,

90, or Co-60)
Co

Fig. 1 Experimental set up for studying dead time and counting statistics

Dead Time

In nearly all detector systems, there will be a minimum amount of time that separates two events in order
that theyy may be recorded as two separate pulses. In some cases the limiting time may be set by processes
in the detector itself, while in other cases the limit may arise due to the delays associated with the
electronics. This minimum time separation is usually called
ca the dead time of the counting system.
Because of the random nature of radioactive decay, there is always some probability that a true event will
be lost because it occurs too quickly following a preceding event. Two models of dead time are in
common use,
se, categorized on the basis of paralyzable and nonparalyzable response of the detector. The
fundamental assumptions of the two models are illustrated in Fig. 3. At the centre of the figure, a time
scale is shown on which six randomly spaced events in the detector are indicated. At the bottom of the
figure is the corresponding dead time behaviour of a detector to be nonparalyzable. A fixed time τ is
assumed to follow each true event that occurs during the “live period” of the detector. True events that
occurr during the dead period are lost and assumed to have no effect whatsoever on the behaviour of the
detector. In the example shown the nonparalyzable detector would record four counts from the six true
events. In contrast, the behaviour of a paralyzable detector
detector is shown along the top line of Fig. 3. The same
dead time τ is assumed to follow each true interaction that occurs during the live period of the detector.
True events that occur during the dead period are not recorded but they extend
xtend the dead time by another
period τ following the lost event. In the example shown, only three counts are recorded for the six true
events. The two models predict the same first-order
first order losses and differ only when true event rates are high.
They are in some sense two extremes of idealized system behaviour,, and real counting system will often
display a behaviour that is intermediate between these extremes. The detailed behaviour of a specific
counting system may depend on the physical processes taking place in the detector itself
itse or on delays
introduced by the pulse processing and recording electronics.

If the system dead time is τ,, and the measured count rate is m , then the true count rate n

predicted by the two models can be expressed as


Nonparalyzable Model:
  (1)

Paralyzable Model:  

 (2)

The derivations of the above results are given in Ref. 1. You may show
show that for low counting

rates (
<< 1/τ ) both models give the same expression for n . In the present experiment we will

utilize the nonparalyzable model in the calculation of true counts and the dead time τ.

A commonly used method for dead time measurements is known as two source method. The method is
based on observing the counting rate from two sources individually and in combination. Because the
counting losses are nonlinear, the observed rate due to the combined
combined sources will be less than the sum
of the rates due to the two sources counted individually, and the dead time can be calculated from the
discrepancy.
Procedure

(i) To find the dead time we have to use two γ sources say S1 (137Cs) and S2 (60Co). While performing
the experiment as per the steps given below, care must be exercised not to move the source already in
place and consideration must be given to the possibility that the presence of a second source will
scatter radiation into the detector which would
would not ordinarily be counted from the first source alone.
In order to keep the scattering unchanged, a dummy second source without activity is normally put in
place when the sources are counted individually.
(ii) Keep source S1 in one of the pits in the source
source holder made for this purpose. Keep a dummy
source in the second pit. Record the counts for a preset time (say 300 s).
(iii) Without removing source S1 remove the dummy source from the second pit and keep the source S2
in its place. Record the number of counts for the combined sources S1 and S2 for the same preset
time as in (ii).
(iv) Remove source S1 and measure the counts due to source S2 alone, for the same preset time as in
(ii).
(v) Remove source S2 as well and record the background counts for the same
same period. Calculate the
count rates in all the cases. Let
 ,
 and
 be the true counts (sample plus background), with
sources S1, S2 and (S1 +S2), respectively. Let ,  and  represent the corresponding observed
rates. Also let
 and  be the true and measured background rates with both the sources
removed. Assuming the nonparalyzable model, the dead time τ is given by (see Ref. 1 for details)

Table: Data for resolving time

Counting statistics:

Radioactive decay is a random process. Consequently, any measurement based on observing the radiation
emitted in a nuclear decay is subject to some degree of statistical fluctuations. These inherent fluctuations are
unavoidable in all nuclear measurements.
measurements. The term counting statistics includes the framework of statistical
analysis required to process the results of nuclear counting experiments and to make predictions about the
expected precision of quantities derived from these measurements.

Although eachh measurement (number of decays in a given interval) for a radioactive sample is
independent of all previous measurements (due to randomness of the process), for a large number of
individual measurements the deviation of the individual count rates from the average count rate behaves in a
predictable manner. Small deviations from the average are much more likely than large deviations. These
statistical fluctuations in the nuclear decay can be understood from the statistical models utilizing Poisson
distribution
ion or Gaussian (Normal) distribution. If we observe a given radioactive nucleus for a time t and
define the success as “the nucleus decays during the process” then the probability of success “p” is given by (1
− e −λt ). The Poisson distribution applies when
when the success probability p is small and the number successes
(i.e. number of counts measured) is also small (say <30). In practical terms, this condition implies that we
have chosen an observation time that is small compared with the half-life
half life of the source.
so When the average
number of successes becomes relatively large (say > 30) we can utilize the Gaussian model of distribution.
Since in most of the cases the count rates are reasonably large (few tens of counts per second) the Gaussian
model has become widely applicable to many problems in counting statistics. On the other hand the Poisson
distribution is applicable in the case of background counts. The details of experimental, Poisson and Gaussian
distributions are given below.

n function
Experimental distribution

We assume that we have a collection of N independent measurements of the same physical quantity. In this
particular case the quantity is the number of counts recorded by the detector in a specific time interval. We
denote the result of these N measurements as

y1, y2, y3 …………….......yi…...........….yN .

The experimental mean is given by

(7)

The data set is conveniently represented by a frequency distribution function F(y). The value of F(y) is the
relative frequency with which the number appears
appears in the collection of data. By definition

  ℎ 
  ℎ    ≡
  (8)



  
  !

A plot of F(y) versus y gives the frequency distribution

distribution of the data (The number of occurrences can also be
calculated by choosing a suitable interval for the values of y).
y). The standard deviation of the distribution is
given by

(9)
 #
Notes regarding σ exp and "

Remember that Eq. (9)) is applicable to the quantities directly measured in the experiment and not to the
derived quantities. To illustrate, in the present experiment if you measure the number of counts for a preset
time interval (say 30 s) and call it yi. Then Eq. (9)) is applicable to these counts only and not to the counting
rates calculated using these values. To determine the deviations for the derived quantities proper error
propagation methods should be used.

To be precise, $ is the true mean value determined from a set having infinitely large number of
measurements and cannot be determined experimentally as such. However for a reasonably large set of
measurements the value of $ can be set equal to $ (Eq. (7).

The Poisson distribution

As mentioned above it is applicable when p <<1 and the number of successes are very few.

(10)

In this case the standard deviation is given by

%&  '$ (11)

The Normal or Gaussian distribution

When p << 1 and the successes are large one can model the experimental data using the Normal distribution
whichh is also called Gaussian distribution (as per R.D. Evans it is erroneous to call this as Gaussian because
its derivation by Gauss (1809) was antedated by those of Laplace (1774) and DeMoivre (1735)). This is given
by

(12)

The standard deviation in this case

se is the same as that for the Poisson distribution

(13)

We will denote both σP and σG as σth.

Applications of statistical models in nuclear physics

There are two major applications of counting statistics in nuclear measurements. The first application
involves the use of statistical analysis
analysis to determine whether a set of multiple measurements of the same
physical quantity shows an amount of internal fluctuation that is consistent with statistical predictions. In this
case the motivation is to determine whether a particular counting system is
is functioning normally. The second
application is more important in which we examine these methods to make a prediction about the uncertainty
one should associate with a single measurement. The following procedure and analysis will give you a feel as
to how
w an experimental distribution in a nuclear counting experiment looks like and how does it compare with
theoretical distributions.
Procedure

(i) Set the operating voltage of the Geiger counter at its proper value.

(ii) Don’t put any source in the lead castle. Also remove all the sources in the vicinity of the castle.

(iii) Take 100 independent readings of the background counts for a preset time of 10 s. (To set Preset time
10 sec. follow step (ii) of initial procedure).
proce

(iv) Save the data by pressing STORE key. While taking 100 independent reading set ITERATION (Step
(iii) of initial procedure) to 1.

(v) Place one of the γ sources (137Cs or 60Co) far enough away from the window of the G Geiger tube so
that approximately 2000 counts are recorded in a time period of 30 s. Take 100independent readings of the
counts for a preset time of 30s.

(vi) Save the data by pressing STORE key. While taking 100 independent reading set ITERATION (Step
(v) of initial procedure) to 1.

(vii) Transfer the data on PC and plot the required function.

Analysis of Background counts (data set (iii) above)

(viii) Determine frequency of occurrence ν(y)

y) which is the number of measurements in which y = 0, 1,2 , 3,
4 ….counts have been observed and plot the experimental distribution ν(y)
y) versus y .

(ix) Calculate the average number of counts $ and the Poisson distribution

(x) Calculate σexp and σth and compare. The comparison gives clue to the reliability of the measuring
equipment. If σexp is larger than σth , it means that additional fluctuations have been introduced by the
apparatus, such as spurious counts due to voltage surges, sparks in the tube or change of the background
during the course of the experiment
eriment which can occur when you handle the sources (move from one place to
another ) while the measurements are going on.

(xi) Determine the actual number of intervals for which the absolute value of the deviation from the
ard deviation %  '$ and the probable error 0.6745σ.
average is larger than the standard 0.6745 Compare with theory.

Analysis of the counts taken with the source data set (v) above

(xii) Carry out the analysis following steps (viii) to (xi) above. However, in this case use Gaussian
distribution. Also, in order to represent the distribution in the best possible manner, frequency of
occurrence may be calculated by choosing equally spaced, non-overlapping,
non overlapping, contiguous intervals for the
counts. The width of the interval can be anywhere from 2 to 10 counts
counts or more depending on the data set.

(xiii) In addition, you may use different methods of testing the "Gaussian" nature of an experimental data
which are illustrated in the book: Measurement systems, Applications and Design (4th edition) by E. O.
Doebelin, pages 44-58.
Table: Data for counting statistics

References:

1. Glenn F. Knoll, "Radiation Detectors and Measurement", Chapters 3 and 7.

2. R.D.. Evans, "The Atomic Nucleus", page 688.

3. R. M. Singru, "Introduction to Experimental Nuclear Physics".

4. Instruction Manual of G.M. Counting System TYPE: ST 360.

6.0 IMPORTANT DEFINITIONS OF RADIATION TERMS

 Absorbed dose: The energy transferred to a material by ionising radiation per unit
mass of the material.

Unit: J kg-1; Name of unit: Gray (see also Rad)

 Absolute Efficiency: The ratio of number of pulses recorded to the number of

radiations emitted by the source.

 Activity: Measurement of quantity of radioactive material. It is the number of nuclear

transformations or isomeric transitions per unit time.

Unit: s-1 Name of unit: Becquerel (see also Curie)

 Alpha decay: Alpha particles consist of two protons and two neutrons bound
together into a particle identical to a helium nucleus. They are generally produced in
the process of alpha decay, but may also be produced in other ways. Alpha particles
are named after the first letter in the Greek alphabet, α.
A radioactive conversion accompanied by the emission of an alpha particle. In alpha
decay the atomic number is reduced by 2 and the mass number by 4. Alpha decay
occurs, with a few exceptions, only for nuclides with a proton number exceeding 82.

 Alpha radiation: Radiation that consists of high energy helium (4He) nuclei emitted
during alpha disintegration of atomic nuclei. Alpha particles possess discrete initial
energies (line spectra) which are characteristic of the emitting nuclide.

 Anode (in electron tubes): An electrode through which a principal stream of

electrons leaves the interelectrode space.

 Attenuation coefficient: The probability that a photon will be removed from the
incident beam per unit thickness of material traversed.

 Background counts (radiation counters): Counts caused by ionizing radiation

coming from sources other than that be to measured.

 Becquerel (Bq): Name of the derived SI unit of activity. Number of radioactive

transformations or isometric transitions per seconds s-1 = Bq.

81
 1 Bq = 27 x 10-12 = 27 pCi

1 kBq = 27 x 10-9 = 27 nCi

1 MBq = 27 x 10-6 = 27 mCi

1 GBq = 27 x 10-3 = 27 mCi

1 TBq = 27 Ci = 27 Ci
 Beta decay: Radioactive conversion accompanied by the emission of a beta particle,
i.e. a negatively charged electron (b- decay) or a positively charged electron (b+
decay). When a negatively charged electron is emitted, a neutron in the atomic
nucleus is converted to a proton with the simultaneous emission of an antineutrino, so
that the proton number Z is increased by 1. When a positively charged electron
(positron) is emitted, a proton in the nucleus is converted to a neutron with
simultaneous emission of a neutrino, so that the proton number Z is decreased by 1.

 Beta Radiation: Radiation that consists of negative or positive electrons which are
emitted from nuclei undergoing decay. Since the decay energy (or, if it is followed
by gamma radiation, the decay energy less that photons energy) is statistically divided
between beta particles and neutrinos (or antineutrinos), the energy spectrum of beta
radiation is continuous, extending from zero to a maximum value characteristic of the
nuclide concerned. The maximum beta energy is generally termed the “beta end-point
energy of the nuclide”.
 Bremsstrahlung: Radiation that results from the acceleration/deceleration of
charged particles in the Coulomb field of atoms.

 Curie (Ci): Name for derived unit of activity. One Curie corresponds to 3.7 x 1010

nuclear disintegrations or isomeric transitions per second 1 Ci = 3.7 x 1010 s-1.

1 Ci = 37 GBq
1 mCi = 37 MB1
1 Ci = 37 kBq
1 nCi = 37 Bq
1 pCi = 37 mBq
 Dose: See absorbed dose, exposure value, and dose equivalent

 Dose equivalent: A term used in radiation protection for the radiation dose. It is the
product of absorbed dose times the quality factor.

Unit: J kg-1; Name of unit: Sievert (see also Rem)

82
 Dose rate: Dose absorbed per unit time

 Dynode: An electrode which performs a useful function, such as current

amplification, by means of secondary emission.

 Electron radiation: Particle emission consisting of negatively or positively charged

electrons.

 Exposure dose: The ratio of the amount of electric charge of the ions of one polarity
that are formed in air by Ionizing radiation and the mass of the air.

Unit: C. kg-1 (see also Roentgen)

 Full width at half maximum (FWHM): The full width of a distribution measured at
half the maximum ordinate.

 Gamma radiation: Gamma radiation, also known as gamma rays, and denoted by
the Greek letter γ, refers to electromagnetic radiation of extremely high frequency and
therefore high energy per photon. Gamma rays are ionizing radiation, and are thus
biologically hazardous. They are classically produced by the decay from high energy
states of atomic nuclei (gamma decay), but are also created by other processes. Paul
Villard, a French chemist and physicist, discovered gamma radiation in 1900, while
studying radiation emitted from radium. Villard's radiation was named "gamma rays"
by Ernest Rutherford in 1903.
Natural sources of gamma rays on Earth include gamma decay from naturally
occurring radioisotopes, and secondary radiation from atmospheric interactions with
cosmic ray particles. Rare terrestrial natural sources produce gamma rays that are not
of a nuclear origin, such as lightning strikes and terrestrial gamma-ray flashes.
Additionally, gamma rays are also produced by a number of astronomical processes in
which very high-energy electrons are produced, that in turn cause secondary gamma
rays via bremsstrahlung, inverse Compton scattering and synchrotron radiation.
However, a large fraction of such astronomical gamma rays are screened by Earth's
atmosphere and can only be detected by spacecraft.

 Gray: The SI unit of absorbed radiation dose. 1 Gray of absorbed dose corresponds to
1 joule of energy per kilogram of mass.
1 Gray = 100 rad

 Half-value thickness (T1/2): The thickness of material layer that reduces the initial
intensity of radiation by a factor of two.

83