IM.3.

The Geiger-Müller Counter

1. Purpose: Do some measurements in nuclear decay, notions of statistics

2. Apparatus: Scaler-Timer (The Nucleus model 550),

Geiger-Müller tube,
oscilloscope,
radioactive sources.

3. Introduction:
A typical Geiger-Müller (GM) Counter consists of a GM tube having a thin end-window (e.g.
made of mica), a high voltage supply for the tube, a scaler to record the number of particles
detected by the tube, and a timer which will stop the action of the scaler at the end of a preset
interval.

The sensitivity of the GM tube is such that any particle capable of ionizing a single
atom of the filling gas of the tube will initiate an avalanche of electrons and ions in the tube.
The collection of the charge thus produced results in the formation of a pulse of voltage at the
output of the tube. The amplitude of this pulse, on the order of a volt or so, is sufficient to
operate the scaler circuit with little or no further amplification. The pulse amplitude is largely
independent of the properties of the particle detected, and gives therefore little information
as to the nature of the particle. Even so, the GM Counter is a versatile device which may be
used for counting alpha particles, beta particles, and gamma rays, albeit with varying degrees
of efficiency.

4. Set-up:
Set up equipment as shown in Fig. 1.
Scaler, Timer, and High Voltage Supply may well be contained in one package.

1
5. Measurements to be performed:
5.1 Characteristics of the GM counter:
Put a radioactive source below the GM tube. Put the counter in a counting mode and
raise the voltage until counts are observed. Note the shape of the pulse and what happens
as the voltage on the GM tube is increased. Do not exceed 900 volts at this time. What is
the minimum voltage pulse necessary to activate the counter? After properly triggering the
scope, sketch a picture of the pulse shape. How would you describe it?

Every GM tube has a characteristic response of counting rate versus voltage applied to
the tube. A curve representing the variation of counting rate with voltage is called a plateau
curve because of its appearance. The plateau curve of every tube that is to be used for the
first time should be drawn in order to determine the optimum operating voltage. Find the
plateau curve for your tube using the procedure outlined below.
(a) Check to see that the high voltage as indicated by the meter on the instrument
is at its minimum value.
(b) Insert a radioactive source into one of the shelves of the counting chamber.
Turn on the count switch and slowly increase the high voltage until counts just
begin to be recorded by the scaler. The voltage at which counts just begin is called
the "starting voltage" of the tube. Using the oscilloscope, measure the minimum
signal size necessary to trigger the scaler.
(c) Beginning at the nearest 20 volt mark above the starting voltage, take counts
every 40 volts. Choose shelf and counting time such that you have at least 1000
counts when the high voltage is about 50V above the starting voltage. Continue
taking counts until a voltage is reached where a rapid increase in counts is
observed. Reset scaler to zero before each count. Tabulate counts versus
voltage.
(d) Plot the data of (c). Identify the “plateau” (i.e. the flattest part of the curve
between the knee and before the onset of the fast rise).

The slope S of the plateau of a GM tube serves as a figure of merit for the tube. The slope is
defined to be the percent change in count rate per 100 volts change in applied voltage in the
plateau region. A slope of greater than 10% indicates that the tube should no longer be used
for accurate work.

The slope may be computed using

2 (R2 - R1) 104

S (% per 100 V) = ------------------------ (1)
(R2 + R1) (V2 - V1)

where V2 is the voltage at the high end of plateau, R2 is the count rate at this voltage, V1 is the
voltage at the low end of the plateau (just above the “knee”), and R1 the corresponding rate.
Do you understand this equation and can you explain it?
In order to compare, obtain a similar plateau curve for an old tube (if available).

2
 The optimum operating voltage will be about the middle of the plateau, usually some
150 to 200 volts above the knee of the curve. Set the high voltage to this point and record
the value. All the subsequent measurements are to be done with the HV set to this optimum
operating voltage.

5.2 Dead-time of the GM counter

There is an interval of time following the production of a pulse in the GM tube during
which no other pulse can be recorded. This interval is called the dead-time of the system.
If this time is known it can be used to make a correction to the observed count rate to yield
the true count rate. The procedure below can give a good estimate of the dead-time.

(a) Obtain a dead-time source (a “split source”) from the instructor.

This source is split into two parts. Remove one half of the source and set it aside.
(b) Place the carrier containing one part on the second shelf of the counting chamber
and make a trial count of 1 minute duration. Get the maximum count rate you can.
This should be more than 20,000 counts per minute, but if not use what you can
get.
(c) Make a 2-minute count and record the counts, N1.
(d) Put the two parts of the source back together, taking care not to disturb the position
of the first part. Make a 2-minute count of the combined parts and record as Nc.
(e) Remove the part initially counted and make a 2 minute count on the second part --
record counts as N2.
(f) Calculate the dead-time of the GM counter using the relation:

N1 + N2 - Nc
τ= -------------------- T (2)
2N1N2

Convert the time thus found to microseconds (note that T is the duration of your
counting -- careful – use the correct time) and record. To understand the origin
of the equation, see refs [1 - 3].

(g) calculate the dead-time uncertainty

The dead-time τ may be used to correct an observed count rate using the expression:

R = r/(1-r τ) (3)

where
r = Observed count rate
R = True count rate

To see the effect of the dead-time directly, use a very high rate source and view the
pulses on the oscilloscope.

3
5.3 Statistical treatment of counting data

The emission of particles by radioactive nuclei is a completely random process. If the

time over which the decays are observed is small compared to the (mean of half-) lifetime of
the radioactive nucleus, the probability for a decay to occur during a given time interval is
constant (the same for every time interval of the same length). When, under identical
conditions, a series of N measurements is made of the number of particles detected per unit
time it will be observed that the individual measurements will vary about some average or
mean value. The true mean, m, can be determined only by averaging an infinite number of
measurements. However, for a finite number of observations (a finite “sample”) the best
approximation of the true mean is simply the “sample mean”, i.e. the arithmetic average n

m≈ n = (1/N) (n1 + n2 + n3 + .... + nN) (4)

The magnitude of the deviations of individual measurements from the true mean is usually
expressed in terms of a “standard deviation” (σ ). The standard deviation is defined to be the
square root of the average value of the squares of the individual deviations (rms = “root-
mean-square”). The number of counts of radioactive decays for a fixed time is a random
variable whose probability distribution is a Poisson distribution; the Standard Deviation for
such a distribution is simply the square root of the true mean:

σ = √m (5)

Given a finite series of measurements, the best approximation (the “unbiased estimator”) of
the standard deviation is given by the square root of the “sample variance”,

σ n=[{(n1- n )2+(n2- n )2+... +(nN- n )2}/(N-1) ]½ (6)

For values of m > 20 the Poisson distribution can be very well approximated by the Gaussian
(or “normal”) distribution for which certain confidence levels have been established in terms
of the standard deviation. These confidence levels are as follows:

About 68% of the number of observations made will fall within the limits of n ± σ n.
About 95% of the number of observations made will fall within the limits of n ± 2σ n.
About 99% of the number of observations made will fall within the limits of n ± 3σ n.

This means that if one additional measurement is made, it should have a 68% chance
of falling within n ± σ n.

When circumstances permit the making of only a single observation the number of counts
obtained, n, is used as an estimator of the true mean m and n as an estimator for its
uncertainty (standard deviation σ n).

The Standard Deviation of a gross counting rate, Rg is:

4
 Rg
σ Rg = n /t = Rgt t = (7), where t is the duration of the counting.
t
This expression is adequate when the sample counting rates is much higher then the
background counting rate. However, when the background counting rate is appreciable
compared to the sample counting rate the net counting rate, R, and its Standard Deviation,
σ r, are determined by
Rg Rb
σ R = [σ 2
Rg + σ Rg
2
]½ = [ + ] ½
(8)
tg tb
where
R = Net counting rate,
Rg = Gross counting rate of sample,
tg = Period used to determine Rg ,
Rb = Background counting rate,
tb = Period used to determine Rb .

To test the statistical nature of nuclear decay, do the following experiment:

(a) Adjust the height of a source in the counting chamber to produce about 2000
counts per minute.
(b) Take a set of 10 counts of 30 seconds duration (“set A”).
(c) Compute the arithmetic mean n .
(d) Compute the standard deviation σ s for a Poisson distribution of mean n .
(e) Compute the individual deviations from the mean (ni - n ) and record in a table.
Do they sum very nearly to zero?
(f) Square the (ni - n ), sum the square and apply Equation (6) to obtain the
standard deviation σ n. Compare σ n with σ s
(g) Count the number of measurements whose values lie within n ± σ n.

Now take a second set of ten measurements (“set B”) and repeat the same analysis.
Compare the two mean values and sigmas. How many measurements of set (B) fall within
the one-sigma interval of set (A)?

5.4 Background Measurements

Extraneous radiation called background radiation is always present. Gamma rays

emitted by certain radioisotopes in the ground, the air, and various building materials as well
as cosmic radiation can all provide counts in a detector in addition to those from a sample
being measured. This background counting rate should always be subtracted from a sample
counting rate in order to obtain the rate from the sample alone. Obtain a background
counting rate using a 5-minute sample time.

5.5 Decay constant of an unknown radioisotope

The activity (number of disintegrations per unit time) of a radioisotope is expressed as

5
 A(t2) = A(t1) exp(-λ (t2 – t1) ) (10)

where
A(t) = activity at time t
λ = decay constant, characteristic of the radioisotope = probability per unit time
for a radioactive decay to occur;

The half-life, T½ of a radioisotope is defined to be that interval during which the activity
decreases to one-half its value at the beginning of the internal. In terms of half-life, the time
dependence of the activity is

A(t2) = A(t1) exp(-(t2 – t1) ln2/T½) , where T½ = ln2/λ  (11)

The counting rate of a sample of a radioisotope may be considered to be directly

proportional to the activity at the moment of measurement provided that the counting interval
is short compared to the half-life. Reasonably short half-lives can be determined by
measuring activity at regular intervals.
The logarithm of the activity when plotted as a function of elapsed time should yield points
falling on a straight line. Explain why.
Obtain your unknown sample from the instructor, measure the activity as a function of time
and find the decay constant. You should take at least 20 measurements of the activity, for
half-minute intervals, making sure that the time between counting periods is minimized (why
is that important?)

6. Analysis, error estimation:

Dead-time:
Estimate the uncertainty on your dead-time measurement from the uncertainties on the
number of counts (use Poisson uncertainties for these).
Decay constant measurement:
Having N measurements of activity gives you N-1 independent measurements of the
decay constant. Estimate the uncertainty on each individual decay constant measurement
from the uncertainties on the number of counts (remember that they are Poisson-distributed!).
Determine the average of all of these values, and calculate the standard deviation. In
addition, determine also a weighted mean of these values, with weight = (1/σi2 ). Taking a
weighted average like this is the appropriate method If the uncertainties of the individual
measurements are very different from each other. Furthermore, you should also determine
the decay constant from the slope of the straight line fitted to a graph of the logarithm of the
activity versus elapsed time. The uncertainty on this slope gives you another estimate of the
uncertainty on the decay constant. (See the statistics hand-out or the statistical methods
chapter of the textbook on how to determine the uncertainty on the slope). From the decay
constant, derive the mean life and the half-life (with their uncertainties). Compare the values
and uncertainties obtained by the different methods, and discuss their relative merits.
7. Report:
You should treat every step in this experiment as a different measurement, with its data,
analysis and conclusion together in one section.
Your report should have a clear and complete discussion of the principles underlying the
6
functioning of a GM Counter, as well as its characteristics as determined from your
experimental data. In addition, you should have a complete description of the data analysis,
including determination of uncertainties. You need to have a calculation of the uncertainty on
dead-time, decay constant and half-life (for all methods of determining the half-life).

You must also answer the following questions:

(1) what are physical phenomena that are initiated by the passage of an ionizing particle
in the GM tube,?
(2) What makes it possible for the GM tube to also count X-rays and γ -rays (photons)?
(3) what is the size (in V) of the smallest signal from the GM tube that is big enough to
trigger the counter circuit?
(4) what is the smallest time between two successive GM tube output signals that you can
see on the oscilloscope?
(5) Given the dead-time determined by you, what is the counting rate at which the error
due to the finite dead-time is 1% of the rate? Do you need to apply a dead-time
correction to the counts for the decay constant determination? Why or why not?
(Hint: compare the error due to dead-time with other uncertainties).
(6) When taking the counts for the decay constant measurement, why is it important that
the time between two successive measurement periods be minimized? If you had
appreciable delays between successive counting periods (without correcting for this),
would that cause you to over- or underestimate the half-life? (Explain)

8. References:

[1] Glenn F. Knoll: Radiation Detection and measurement,

John Wiley & Sons, New York 1989 (2nd ed.), 1999 (3rd ed.)
[2] William Leo: Techniques for Nuclear and Particle Physics Experiments :
A How-To Approach; Springer Verlag, New York 1994 (2nd ed.)
[3] A. Melissinos: Experiments in Modern Physics, Academic Press, New York 1966
(a reference copy of this book is available in the lab).
[4] Preston, the Art of Experimental Physics (course textbook -- reference copy in the lab)
[5] http://polaris.phys.ualberta.ca/info/Phys29x/Manual/11GM01.pdf
[6] http://www.mathematik.uni-marburg.de/~kronjaeg/hv/radio/geiger/caltech/exp2.htm
[7] http://www.warren-wilson.edu/~sleavitt/formal/geigerwebpage.htm
[8] http://www.home.fh-karlsruhe.de/~mero0001/master/gm_counter.pdf
[9] http://www.home.fh-karlsruhe.de/~mero0001/master/deadtime.pdf
[10] http://www.lbl.gov/abc/wallchart/chapters/12/2.html
[11] http://www.astro.psu.edu/users/niel/astro485/derivations/geiger1.pdf