Proportional counter report

Tomás Fernández Bouvier

March 29, 2019

Abstract
An economic radiation gaseous detector is built from an aluminium can and other relatevely
cheap materials. The device is tested with two radiatiove sources (Fe-55 and Am-241) to gure
out the optimal conditions of operation and measurement as well as the energy resolution.
Finally the detector is used to measure the spectrum and analyze it obtaining as a result the
photo and escape peaks of both sources.

Student number: 014960738

Teachers: Jens Brücken, Erik; Tiina Naaranoja
Helsingin Yliopisto

1
Contents

1 Theory 3
1.1 The physics behind gas-lled detectors . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 ion migration phenomena and eects . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Proportional counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Building process of a proportional counter 6

3 Experimental procedure 12

4 Results discussion and analysis 15

4.1 Spectrum analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5 Conclusion 20

6 Appendix (result tables) 21

6.1 Fe-55 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
6.2 Am-241 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2
1 Theory

Note All the contents expressed in the theory are nothing but a recap of contents from Glenn F.
Knoll-"Radiation and measurement" (chapters 2,4,5 and 6) [1].

1.1 The physics behind gas-lled detectors

When a phenomenom happens in a nucleous, this emits radiation. Radiation is dened as the
transmision among the space and materials of energy [8]. This can be electromagnetic waves or
particles . In other words, radiation is an energy carrier. There are dierent types of radiation
depending on the type of source and carrier. α, β and γ are the most common ones.
Consequently radiation interacts with matter and has an eect on it, it modies their structure
transmit energy. Detecting radiation is usually complicate and tidious. However by measuring the
eect that it produces on materials we can measure nearly all about it. The detector we built in
the laboratory course essentially detects and measures currents produced in a gas and, starting from
their amplitude, measures radiation. This is one of the oldest types of detectors and was usefull in
low energy rst attempts of spectroscopy.
The principle is simply: when radiation interacts with the gas, the energy can produce excited
molecules or ionization. The number of ions is theoretically
n0 = E/W (1)
(where E is the energy of the incident radiation and W is the energy an ion needs to be created).
If an ionization occurs, both free charges will move through the gas creating a current that will
be collected. The amplitude of the collected current will be measured and from this result we can
calculate the number of initial ion pairs and hence their energy. Since sometimes the energy deposited
is not high enough to create ions but either excited molecules, the energy value measured will be
always less than the real one. However the aproximation is still good.

1.2 ion migration phenomena and eects

Normally, the speed at which electrons move is only thermal and described by molecular dynamics.
However, if we apply an external electric eld the speed is superposed and the latter term can even
be dominant. It is given by the formula:
µE
v= (2)
p

where v is the speed, mu is the mobility constant which depends on the ion type and p is presure and
E the magnitude of the electric eld. In the process of migration of the ions to the cathode (diusion)
several phenomena can happen which should be taken in account when analyzing our results:

ˆ charge transfer: when a charged particle inpacts in a neutral one, charge can be transferred
inverting the particles their charges.
ˆ electron attachement: a free electron can inpact in a neutral molecule or atom and be absorbed
by it. Consequently the new molecule is negatively charged.

3
 ˆ recombination: if positive ions caused in the ionization inpact with negative ions, recombination
can happen. This means that charges from negative ion will be given to positive one creating
at the end two neutral particles. Consequently their contribution to the current will be avoided
and hence they won't be detected. The idea is to separate positive and negative created ions
in order to avoid them to recombinate. This can be done if we make them travel fast enough,
result that can be obtained from 2 increasing the electric eld.
ˆ avalanche formation: while travelling towards the anode, a suciently energetic electron (cre-
ated with the ion in the primary ionization) is likely to produce a secondary ionization in a
molecule of the gas. The electric eld will accelerate this electron which is also likely to produce
again secondary ionization creating more and more electrons. This will result in an avalanche
of electrons and therefore a higher current than the original one. The avalanche formation
and intensity depends mainly on the electric eld intensity dening dierent regions. We are
specially interested in the proportional region.

Avalanche is the most important phenomenom since it amplies current in order to be detected. The
avalanche depends on the electric eld (applied voltage), the type of gas and the geometry of the
detector. We can then dier ve essential regions depending on their avalanche rate:

ˆ very low values:In that region the electric eld cannot avoid recombination so the detector
doesn't measure properly the energy of the incident radiation.
ˆ ion saturation region: The recombination doesn't happen anymore but there is not multiplica-
tion
ˆ Proportional and limited proportional region: here the collected charge in the detector is
linearly proportional to the number of primary ions since there is an avalanche and hence a
multipliation. At higher voltages we nd non-linearities since positive ions create a sort of
electric eld which cancels with the original one. Here the equation for the collected charge
would be
Q = n0 eM (3)
(where e is the electron charge and M the average multiplication factor)
ˆ Geiger-Müller region: The collected charge is no longer proportional to the number of primary
ions and hence the energy cannot be measured that way. The detector type which can prevent
the problem are called Geiger-Mueller detectors but we won't study them here.

1.3 Proportional counters

Proportional counters operate through the proportional and limited proportional region where the
multiplication happens. In our comparisson between the experimental and theoretical results we will
see how does multiplication happen and where are the limits between proportional and other regions.
But, which is the theoretical prediction for the values of M?. Since M is a consequence from the
avalanche we rstly dene an avalanche rate α through the "intuitive" equation dn n
= α(r)dr (the
ratio amount of created ions is proportional to the radius where it is measured 1 ). The solution for
1 asuming that our detector is cylindrical and its eld radial (we will later explain in more detail why we took that
shape instead of other.

4
this dierential equation is
n(r) = n0 eαr

M is an average of a continuos distribution of the amplication (n/n0 and hence from previous
equation it can be calculated as it follows
Z b
ln M = α(r)dr
a

although it is usefull to write the equation depending on electric eld in order to use assumptions:
Z b
∂r
M= α() d
a ∂

For a cilinder with an applied voltage between a and b we have

V ∂r r
(r) = → =−
r ln(b/a) ∂ 

Inserting that in previous equation we obtain

b a
−r
Z Z
V α()
ln M = α() d = d
a  ln(b/a) b 2

Since we are in a linear region we supose that α is linearly dependent on  (α= t  where t is a
constant). Then:
Z a
V 1 V V V
ln M = t d = t(ln − ln
ln(b/a) b  ln(b/a) a ln(b/a) b ln(b/a)

This is all we can do. The determination of the constant t is attributed to Diethorn who gave his
name to the following equation:

V ln(2) V
ln(M ) = · (ln − ln K) (4)
ln(b/a) ∆V pa ln(b/a)

where p is the preasure of the gas which in our case can be aproximated to the ambient preasure.
∆V is the potential dierence between two ionization events, in our case 23,6 V and K is a threshold
below multiplication won't happen, in our case 48000 atm·cm
V
.

About calculation of uncertainties

For the calculation of uncertainties we will mainly recall the typical propagation of uncertainties
formula: s
X ∂y
sc (y) = ( s(xi ))2 (5)
i
∂xi

where y is a function of xi variables.

5
2 Building process of a proportional counter

For the proportional counter we used an empty can of soda that we cleaned and polished inside in
order to remove the non-conducting coating inside. We also removed the top and we made two little
holes at the bottom. We then operated on two plexiglas coverages where we did also the same holes
as in the bottom of the can.
We used also a thin wire with the length of the can as the anode of our detector and one conector
which relates to the preamplier and the experimental setup 2 . For the cathode we used a long
narrow copper sheet xed externally on the can and in contact with the connector.
Starting of a cilinder, we could have built our detector in a lineal conguration. The reason why
we chose the radial conguration rather than the other was because otherwise the voltage needed to
create a strong enough eld which could avoid recombination would be greater than 20000 V that is
very hard to obtain in a laboratory. With radial conguration, if our anode is very thin (around 50
µm) we can obtain strong electric elds with weaker voltages.
Since a soda-can is not an uniform cilinder we had to remove this unregularities that aect the
electric eld and hence the results. We used two thin brass-tubes inserted in both top and bottom
through the plexiglass coverages. Then, in those regions there won't be electric eld at "non uniform
zones".
This are all the parameters for our proportional counter:

Parameter Size Uncertainty

anode radius (estimation) 0,00375 cm -
cathode radius 3,34 cm 0,50 cm
Cylinder Length 12,50 cm 0,50 cm
Big brass tube length 3,00 cm 0,20 cm
Small brass tube length 2,00 cm 0,20 cm

We placed all materials in the ultrasonic bath to clean impurities.

Next, we inserted the thin wire into one of the brass tubes which goes into the bottom coverage and
made it cross all the chamber. We took it from the other side and soldered it on the conector. At
the end we had a concentric cylindrical anode (wire) and cathode (soda-can) with in its ends two
coverages and brass tubes and conectors.
Once we did that we glued both coverages to the can carefully since the anode wire is fragile. At the
end our detector must be completely sealed to avoid any leackage.
Finally we cut the exceding anode and xed the cathode as mentioned before.
This is our result:
2 see section 3

6
 (a) face where we can see input and interior of our (b) back where we can see the output
device as well as the connector which goes to readout
electronics

Preamplier
Schematics and operating mode

Our preamplier has in short 2 parts:

ˆ Voltage divider/lter: As the name says, these device divides voltage throwing through it
in order to obtain a smaller one as a suply to our circuit. It is clasically built with two
series resistors. We added additionally two capacitors that act as a protection against current
surcharges but this is optional.

Figure 1: voltage divider/lter

7
 ˆ Preamplier: We built a charge sensitive preamplier. It works as an integrator that produces
voltage proportionally to the input current. Consequently, it allows us to measure with high
accuracy the incoming pulses. Also it is useful since it can remove noise from the signal by
setting a threshold to it.
Our circuit consists on a printed circuit where we have solded an operation amplier, "OP amp
657"("texas instruments"), where the amplication itself occurs, resistors to control the gain
and avoid surcharges, two capacitors that allow us to control the size/amount of amplication
obtained and an input and output plug which will be connected to our detector and our display
respectively. Also we added in parallel to the op-amp a big value resistor. This is intended
to reset the system after each amplication (otherwise the output signal would increase until
maximum output is reached). We had the possibility to build our preamplier up in non-
inverting or inverting conguration. We chose the latter because the signal incoming from the
detector is inverted (electrons travelling from + to -).

(a) inverting-amplier circuit (b) op-amp chip where

schematics the amplication really
occurs

Our preamplier

Our preamplier was made following the schematics of the example one:

Figure 2: voltage divider

8
 (a) face (b) back

Figure 3: printed circuit

Below are the dierent component's values for our preamplier:

Param. Value
R4 0Ω
R1 470 Ω
R2 47 kΩ
R3 22 MΩ
C0 2,2 µF

The theoretical gain obtained with our preamplier is G = − R1

R2
= 47kΩ
470Ω
= 100
In the reality, the preamplier is not ideal. In the lab we measured the dependance between the
input and ouput voltage with the help of an oscilloscope:

Vin Vout s(Vout ) Vin Vout s(Vout )

10 1.0700 0.0014 65 6.3900 0.0035
15 1.4200 0.0014 70 6.8400 0.0032
20 1.9400 0.0014 75 7.1600 0.0037
25 2.3900 0.0014 80 7.3600 0.0025
30 2.8300 0.0016 85 7.4100 0.0037
35 3.4400 0.0019 90 7.4500 0.0038
40 3.8800 0.0022 95 7.4800 0.0030
45 4.3400 0.0022 100 7.4900 0.0022
50 4.7800 0.0021 105 7.5000 0.0027
55 5.2300 0.0020 110 7.5100 0.0030
60 5.6600 0.0019 115 7.5100 0.0034
120 7.5300 0.0040

9
(s(Vin )=0,1 mV) Plotting those results, we see that we have two regions, one linear region where the
amplication really happens and one saturation region where the output voltage doesn't increase
despite of growing the input voltage.

Figure 4: Amplier input-output voltage dependance

For the linear region Vout (V ) = Vin (mV )·(9, 498±0, 096)·10−2 +(0, 043±0, 036) The slope corresponds
to the real amplication, G = (9, 498 ± 0, 096) · 10−2 mV V
≡ 94, 98 ± 0, 96 which matches what we
expected.
For the saturation region Vout (V ) = Vin (mV ) · (4, 160 ± 0, 082) · 10−3 + (7, 0562 ± 0, 0082)
The intersection of these two curves gives us the saturation point which is: Vins = 7,0562−0,042
9,498−4,160
· 102 =
77, 2 ± 3, 2 mV
Beyond this point amplication won't really happen so we might avoid this region in our calculus.
Here is an example of the response outcoming the preamplier for an incoming pulse:

10
Note

After we rst ended building our preamplier, we noticed that we had a lot of noise during our
measurements. We rstly suposed that this could be due to a bad soldering somewhere but we
didn't nd any bad conection. It was only with the help of a teacher that we issued that the probem
was in the operational amplier itself which was damaged because of an overload. Despite changing
it there was still a problem in one of the legs of the voltage input which was solved.

11
3 Experimental procedure

The day before the measurements we veried that there was not any leackage in our detector. Then
we left the detector connected to gas ush in order to only have gas inside the ion chamber. We
left this conguration over one night under some minimal gas ow to avoid any oxygen and water
coming in.
Note: The used gas was P-10 which is the most common one in such experiments. Its composition is
90 % Ar and 10 % CH4 and its ionization energy is around 26 eV/ion. Actually a better understanding
of the absorption can be attempted by looking at the graph below which shows absorption of radiation
from the detector (gas+aluminium recipient) depending of the energy carried:

Figure 5: Relative intensity vs radiation's energy extracted from [2]

We can see that the resolution in terms of gas+can response decreases after 10 keV. So all the counts
distribution registered in our detector won't be realistic. For xing the problem we would have to
deconvolute the two functions and apply the correction to our measurements but this is tidious and
it isn't the objective of our experiment.
The day of the measurements, we connected the detector to a permanent gas ush of 5,25 ml in
order to avoid possible leackages. We also plugged the detector's connector into the preamplier that
amplies outcoming signals and splits the signal into two. One goes to the oscilloscope. From here
we can measure if there is a signal outcoming and it's shape. It is used in also in the calibration of
our data adquisition. The other connector goes into the multichannel analyzer in order to measure
the pulse height and register counts which will be described later. All this constitutes the readout
electronics. The electric eld is generated by the voltage regulator.
The user is protected by a radiation absorbing metal sheet recovering the device that screens the
source. Electromagnetic interferences are avoided with an aluminium foil that operates like an anecoic
chamber.

12
 Figure 6: Experimental setup schematics

Figure 7: Experimental setup

Once we were ready, the rst thing we did before starting to do measurements is calibrating our
electronic readout chain. For that, we measured dierent voltages coming from a pulse generator in
the oscilloscope and in the MCA to get an equivalence between them. We used a coursegain of 10.
We performed thirteen measurements which are shown in table 6.
After that we put a radioactive source above our proportional counter. We started by iron 55 (Fe-55).
Then we raised the input voltage until we saw a clearly distinguishable peak on the MCA and a good
signal in the oscilloscope. The latter represents the amplied pulses arriving from the anode to the
cathode because of the electric eld. For Fe-55 our threshold was 1689 V. Since the signals are very

13
weak, they were initially amplied by 100. We then initizalized the MCA for 1 minute (uctuations
on time are not important) which starts to count pulses and distribute them on channels depending
on their height. This will yield a spectrum with a quite well dened peak. We measure the centroid of
the peak with its uncertainty and full width at half maximum (FWHM)3 . After that we raised by 50
volts the input voltage. This would result in a displacement of the peak to the right (higher energy).
We repeated the measurement and procedure of increasing voltage until the peak reached the limit
of our MCA. At this point, we reduced the amplication to 40 and did again the described procedure
also for an amplication of 20 10 4 and 2. All those measurements appear in table 6.1. Finally we did
a long measurement in optimal conditions (V=2236 and coursegain=10) for 8,58 minutes in order to
get the best dened spectrum as we could (i.e. well denide caracteristic peaks which will be used
for calibrating in the later analysis).
We then substituted our source by Am-241 and repeat the procedure. The dierence is that here
our threshold input voltage would be quite lower. This can be explained since Am-241's radiation is
more energetic than Fe-55's one and it is then easier to see peaks for less voltage. We did the big
measurement of the spectrum for 15,03 minutes with the same conditions (voltage, amplication,
etc.).
Last thing we did was to measure the background spectrum for 34,92 minutes and same settings as
the other two big measurements in order to improve the latter by substraction. We also measured
the preasure of the room p = 1005 ± 5HP a, and its temperature T = 295 ± 1K .

About the used sources

Fe-55 [4] is an isotope of iron which nucleus has 29 protons and 26 electrons. It decays to Mn-55 via
electron capture. This means that after a electronic readjustement, an electron from a shell lls a
lower-energy one. The resulting energy dierence is released as two types of radiation:

ˆ auger eect: the energy is absorbed by a higher level electron and hence it is released. This
cannot be detected by our proportional counter. It covers 60% of the decays.
ˆ x-rays radiation: the energy is released as photons, 2 K-alpha with 5,9 KeV (28 %) and 2
K-beta 6,49 KeV (2,85 %). This depends on the initial energy-level of the "decaying" electron.

Am-241 [5] is an isotope of ameritium which nucleus has 95 protons and 146 neutrons. It decays
mainly to Np-237 by emiting α-particles and γ -rays:

ˆ α-decay: The nucleus emits very high energy α particles (around 5,5 MeV). Since they do not
produce any ionization, those particles won't be detected in our proportional counter.
ˆ γ -decay: The energy of gamma rays will be mostly 59,54 KeV but some other peaks might be
detected around 13,9 KeV, 17,8 KeV, 26,4 KeV (escape peaks and characteristic peaks from
Np)

Am-241 can also decay by emiting nanoclusters of Si or spontaneous ssion emiting 3 n0 . Samely
as α-decay, they won't theoretically be detected but all of them could produce some noise at lowest
energies 4
3 Here we are only interesed by the position of the peak, not its size or the rest of the spectrum. That's why we do
not pay special attention to the rest of the spectrum
4 see section 4

14
4 Results discussion and analysis

Note: All the analysis (ttings and calculations) was performed using OrigiPro 8
The rst step is to nd a formula that relates the channels from the MCA with input voltages. For
that we did the measurements in 6. Plotting voltages vs channel number we clearly see a linear
relation so we make a linear t of our data (V = m · N + V0 ). We obtain that

m = (0, 15920 ± 0, 00078) · 10−2 V /unity V0 = (0, 94 ± 0, 35) · 10−2 V r2 = 0, 998

Figure 8: calibration of MCA channels

With that formula we can easily calculate the charges arriving to the cathode by assuming that the
preamplier behaves as a capacitor with C=1pF.

Q=C ·V

We obtain then the table 6.1.

Since we didn't take the value of the radius of our anode we need a way to nd it before making any
calculus. We claim that the photo-peak for the spectrum of the Fe-55 is around 5,9 KeV and for Am-
241 57,83 KeV. Then we calculated from two dierent ways M, experimentally and theorethically, and
we adjusted the parameter a (anode radius)in order to make values corresponde for same voltages.
The rst way we followed was dividing our Q by "e · n0 " according to 1. We obtained the num-
ber of primary ions dividing the total energy by the energy an single ion pair needs to be created.
For the calculation of the uncertainties, we took s(Q) to be the combination of the centroid un-
certaintypwith the dierence between Q measured at the same voltage but dierent coarsegain:
s(Q) = (∆Qcentroid )2 + (∆Qcoarsegain )2 . We then applied the usual error propagation formula 5 to
calculate s(M).
The second way was by using Diethorn prediction: ln(M ) = ln(b/a)
V
· ln(2)
∆V
V
(ln pa ln(b/a) − ln K). With a
cathode radius of (3, 34 ± 2) cm, the issue we reached in order to make the values coherent was a=

15
(0, 00375 ± 0, 000205 ) cm. For the calculation of associated uncertainties we propagated the ones on
which lnM depends: the cathode radius, the preassure, the anode radius and the high voltage with
5 and from s(lnM) we calculated s(M).
Here are the plot of theoretical vs experimental values of M in logaritmic scale:

Figure 9: theoretical (Diethorn) vs experimental values of M

As we can see, at lower values (1500 V) experimental predictions seem not to be in concordance
with diethorn. This happens because below 1500 V we are reaching the ion-saturation plateau where
amplication doesn't happen. Meanwhile, at higher values it seems that the tendeance dicreases.
This means we are near the Geiger-Müller plateau where the linear relation doesn't stand anymore.
In the linear region the results match very well with the theoretical prediction (in the intervals of
uncertainty)
Also we plotted energy resolution calculated by the expression ∆E
E
= F W HM
centroid
.

5 This is a simple estimation for which we didn't use any statistical method.

16
There is clearly a relation between each isotope's data. Indeed the primary ionization (which is what
we measure to know the energy√of the rays) is nothing but a poissonian √
process which consequently
has a standard deviation σ ≡ E . Hence the resolution R = E α E α E behaves as a rational
∆E E √1

function. Also since the emission energy of Am-241 (characteristic peak) is much higher than Fe-55's
one it is normal to nd it under the other.

4.1 Spectrum analysis

Now we proceed to analyze the full spectrum of Fe-55 and Am-241 respectively. We did a long
measurement in order to have a big amount of counts and hence a well dened spectra. With an
input high voltage of 2236 V and an amplication of 10x, this is what we obtained for both sources:

(a) Fe-55 (b) Am-241

We also measured for the same and conditions the spectra of the background in order to substract
it form our results and depure them from noise caused by dierent interferences (cosmic rays for
example). The substraction formula is used for each channel counts: N = Ns − Ntbstb , where N is the
nal numbe of counts N0 the initial one and tb and ts the measurement times of the background and
source spectra respectively.

17
 Figure 10: Background spectrum

As we can see, the spectrum of Fe-55 is esentially a high peak centered around channel 90 6 and
one quite smaller around channel 50 . The latter corresponds to an escape peak, caused when the
incoming radiation from the source doesn't interact completely with the detector and less gamma
ray escape (producing a peak of less energy than the photo-peak). In the case of Am-241, the results
seem to be a little bit more strange. First of all we can see that the number of counts corresponding
to peaks dicreased a lot compared to Fe-55. While we where having a photo-peak of height ≈ 1500,
here our photo-peak is around only 300 counts. This makes it not to be very well dened and we
can see counts among all the spectra. The most strange result is at lower energies where the peak
raises even higher than the photo-peak.
In our spectrums there are 4 peaks for which we know precisely the energy (see "About used sources")
(three from Fe-55 spectrum and one from Am-241). We can use them to peform a second calibration
(x−xc )2
more precise than the one before. For that we make a multi-peak gaussian t (N = √A e−2 w2 )
w π/2
on our "depured" crude results:

(a) Fe-55 (b) Am-241

6 From databases we know that actually channel 90 peak is in fact two peaks that are indistinguishable here because
of the bad resolution of our detector

18
Note that in the case of Am-241 we only tted the peaks and not the weird noise at lower values in
order to have better stathistics. The results are the following with their corresponding uncertainties.
7
:
Peak xc s(xc ) w s(w)
1 10,09 0,45 10,4 1,1
2 92,86 0,82 11,0 1,5
3 99,8 5,8 14,5 3,4
4 266,4 6,1 64 27
5 337,3 7,1 40 13
6 410,3 1,3 37,4 2,8
7 926,04 0,44 59,63 0,93

For our calibration we will use peaks 1,2,3 and 4:

We obtained that E(keV ) = (0, 0631 ± 0, 0019) · M CA + (0, 99 ± 0, 89) with r=0,997.
Pluggin this calibration in our previous table we obtain nally:

Peak xc s(xc ) w s(w)

1 3,2 1,4 1,6 1,4
2 5,9 2,23 1,7 1,4
3 6,49 2,5 1,9 1,5
4 59,54 18 4,8 1,8
5 17,8 5,3 5,0 3,1
6 22,3 6,6 3,5 2,2
7 26,9 7,9 3,3 1,6

Where all measurements are given in terms of KeV.

7 We don't include the height of the peaks since we are not interested in

19
From the database [9] we identied that peak 7 corresponds to γ2,1 decay (E=26,3446 ± 0,0002
KeV). Meanwhile, peaks 5 and 6 correspond rather to XL X-ray emissions (E=11,89 to 22,2 KeV).
According to what we comented above this is what we expected to nd for Am-241.

20
5 Conclusion

What did we do through the report?. Firstly, after a rst calibrating of our display (MCA), we
measured the spectra of two sources under dierent conditions of high voltage and amplication.
This allows us to understand how does our detector work and which are the optimal parameters of
operation. We discovered that the anode radius of our detector might be 37,5 µm which is quite
reasonable and matches with our espectation. We also studied how the energy resolution varies and
although we couldn't do any rigourous analysis of our results and plots we think that the resolution
decreases with a voltage increase.
All this results allowed us to nd a sweet spot for measuring our spectra, so with optimal parameters,
we did a long measurement of both source's spectra and a background one. By plotting them, we
discovered in the Fe-55 one two peaks that we tted succesfully and, acording to databases we
improved the energy-channel number calibration. Finally we found out three other escape peaks in
the interval 15-30 keV for Am-241. The fact that peaks that are not evident to see appear clearly
in our "homemade" detector proves that the latter works good and it is usefull. However the big
plateau corresponding to lower energies remains unexplained. One approach could be that the other
decays happening in the source (clusters and spontaneous sion) produce energy and hence some
noise wich interfers weakly with the detector causing such perturbations. However we do not have
any experimental way of proving it. The other, more realistic and plaussible is related to the eciency
of the device in terms of photon absoption by the argon and the aluminium. The optimal conditions
will be between 10-20 keV and for higher values the absorption will be quite less, distorting the height
of the peaks as it happens in that case.
Despite of that, our results are still quite satisfactory for a homemade detector.

21
6 Appendix (result tables)

Here are the tables of all results and measurements done with the proportional counter except the
ones related to the spectrum measurements.8

Centroid ∆ Centroid FWHM Volt(mV) ∆Volt(mV)

50,83 2,11 1,25 0 2
358,84 2,11 1,94 58 0,5
125,86 2,11 1,23 20,7 0,5
203,62 2,11 1,75 33,3 1
250,95 2,11 1,19 42,92 1
316,7 2,11 1,5 51,7 1
383,41 2,11 1,83 62,2 1
456,58 2,11 1,86 73,3 1
517,05 2,11 1,18 83 1
592,15 2,11 1,24 95 1
667,66 2,11 1,67 107 1
779,33 2,11 1,65 124 1
974,58 2,11 1,85 157 1
Table 1: Calibration of the MCA
8 the
coarse gain in the calibration is 10 so the voltage coming in the MCA is ten times higher than the readout
voltage on the oscilloscope

22
6.1 Fe-55

HV (V) Centroid s(Centroid) FWHM Coursegain

1689 39,05 0,98 10,98 100
1737 49,5 1,0 11,88 100
1790 65,4 1,0 13,96 100
1840 90,2 3,0 13,64 100
2043 277,64 0,93 49,94 100
2092 380,1 1,1 55,1 100
2141 507,5 1,0 71,73 100
1892 113,62 0,63 22,28 100
1938 147,27 0,93 28,36 100
1988 197,41 0,97 36,03 100
1988 81,24 0,99 16,03 40
2038 109,20 0,90 19,3 40
2090 150,00 0,94 28,17 40
2139 203,51 0,93 34,8 40
2191 281,85 0,95 45,02 40
2243 395,58 0,97 56,87 40
2291 541,8 1,1 74,64 40
2291 265,83 0,94 46,94 20
2340 368,97 0,98 67,52 20
2386 501,2 1,1 78,78 20
2236 95,53 0,65 16,51 10
2383 245,89 0,95 47,24 10
2443 370,17 0,94 67,15 10
2488 504,77 0,94 77,77 10
2490 210,10 0,93 40,03 4
2537 290,50 0,93 47,59 4
2586 408,59 0,93 63,46 4
2638 591,2 1,1 71,46 4
2638 291,26 0,93 49,05 2
2690 420,9 1,0 68,2 2
2738 588,8 1,0 78,02 2
2784 816,1 1,2 95,56 2
Table 2: Fe-55 measurements

For all measurements s(HV)= 1 V

23
HV(C) Q(C) s(Q)(C) M s(M) ∆E/E s(∆E/E )
1689 7,16E-16 3,8E-17 19,7 1,1 0,2812 0,0071
1737 8,82E-16 3,9E-17 24,3 1,1 0,2400 0,0049
1790 1,135E-15 3,9E-17 31,2 1,1 0,2136 0,0033
1840 1,530E-15 5,9E-17 42,1 1,6 0,1513 0,0050
2043 4,514E-15 4,4E-17 124,2 1,2 0,17987 0,00060
2092 6,144E-15 4,9E-17 169,0 1,4 0,14498 0,00043
2141 8,174E-15 5,5E-17 224,8 1,5 0,14133 0,00029
1892 1,903E-15 3,7E-17 52,3 1,0 0,1961 0,0011
1938 2,439E-15 4,0E-17 67,1 1,1 0,1926 0,0012
1988 3,237E-15 4,1E-17 89,0 1,1 0,18251 0,00090
1988 3,468E-15 9,7E-17 95,4 2,7 0,1973 0,0024
2038 4,581E-15 9,7E-17 126,0 2,7 0,1767 0,0015
2090 6,2E-15 9,96E-17 170,7 2,7 0,1878 0,0012
2139 8,3E-15 1,03E-16 229,3 2,8 0,17100 0,00078
2191 1,145E-14 1,1E-16 315,0 3,0 0,15973 0,00054
2243 1,598E-14 1,2E-16 439,6 3,4 0,14376 0,00035
2291 2,180E-14 1,4E-16 599,7 4,0 0,13776 0,00028
2291 2,163E-14 2,2E-16 595,0 6,0 0,17658 0,00062
2340 2,984E-14 2,4E-16 820,8 6,6 0,18300 0,00049
2386 4,037E-14 2,8E-16 1110,4 7,6 0,15718 0,00035
2236 1,615E-14 3,7E-16 444 10 0,1728 0,0012
2383 4,009E-14 4,3E-16 1103 12 0,19212 0,00074
2443 5,987E-14 4,8E-16 1647 13 0,18140 0,00046
2488 8,130E-14 5,5E-16 2236 15 0,15407 0,00029
2490 8,60E-14 1,0E-15 2365 28 0,19053 0,00084
2537 1,180E-13 1,1E-15 3245 30 0,16382 0,00052
2586 1,650E-13 1,2E-15 4538 34 0,15531 0,00035
2638 2,376E-13 1,5E-15 6537 41 0,12088 0,00022
2638 2,365E-13 2,2E-15 6507 61 0,16841 0,00054
2690 3,397E-13 2,5E-15 9346 70 0,16203 0,00038
2738 4,734E-13 3,0E-15 13021 82 0,13251 0,00023
2784 6,543E-13 3,7E-15 18000 103 0,11709 0,00017
Table 3: Fe-55 results

24
6.2 Am-241

HV (V) Centroid s(Centroid) FWHM Coursegain

1243 56,2 3,1 10 100
1303 68,7 3,5 9,46 100
1363 84,8 3,0 9,57 100
1421 102,9 4,0 10,35 100
1483 115,3 6,3 11,19 100
1539 171,0 1,8 12,75 100
1602 239,8 3,3 14,71 100
1661 319,6 2,8 19,31 100
1723 439,9 2,8 24,83 100
1780 604,4 4,3 12,77 100
1780 240,8 4,4 13,61 40
1842 343,6 3,6 17,26 40
1901 472,4 4,2 24,34 40
1960 676,0 2,7 34,44 40
1960 326,0 4,5 14,15 20
2024 489,1 2,6 31,08 20
2084 710,1 3,3 19,39 20
2084 348,9 3,6 19,19 10
2141 503,1 2,8 19,9 10
2199 736,5 4,1 31,68 10
2236 926,6 1,1 57,5 10
2199 295,9 2,9 20,71 4
2261 443,5 2,5 35,35 4
2318 649,4 2,7 21,55 4
2318 315,4 2,8 18,02 2
2386 487,1 4,7 23,41 2
2441 687,4 8,8 9,33 2
2468 875,8 3,4 1,79 2
Table 4: Am-241 measurements

For all measurements s(HV)= 1 V

25
HV(C) Q(C) s(Q)(C) M s(M) ∆E/E s(∆E/E )
1243 9,89E-16 6,1E-17 2,78 0,17 0,1779 0,0099
1303 1,188E-15 6,7E-17 3,33 0,19 0,1377 0,0071
1363 1,444E-15 6,0E-17 4,05 0,17 0,1129 0,0040
1421 1,732E-15 7,3E-17 4,86 0,21 0,1006 0,0039
1483 1,93E-15 1,1E-16 5,41 0,30 0,0971 0,0053
1539 2,816E-15 4,7E-17 7,90 0,13 0,07457 0,00079
1602 3,912E-15 6,5E-17 10,98 0,18 0,06133 0,00083
1661 5,181E-15 6,2E-17 14,54 0,17 0,06043 0,00053
1723 7,097E-15 6,6E-17 19,92 0,19 0,05644 0,00036
1780 9,716E-15 9,0E-17 27,27 0,25 0,02113 0,00015
1780 9,82E-15 2,0E-16 27,56 0,57 0,0565 0,0010
1842 1,391E-14 1,8E-16 39,03 0,51 0,05024 0,00053
1901 1,903E-14 2,1E-16 53,42 0,59 0,05153 0,00046
1960 2,714E-14 1,9E-16 76,16 0,53 0,05095 0,00020
1960 2,642E-14 4,2E-16 74,1 1,2 0,04341 0,00060
2024 3,940E-14 3,3E-16 110,58 0,93 0,06355 0,00034
2084 5,699E-14 4,2E-16 159,9 1,2 0,02731 0,00012
2084 5,648E-14 7,3E-16 158,5 2,0 0,05501 0,00057
2141 8,103E-14 6,9E-16 227,4 1,9 0,03956 0,00022
2199 1,1818E-13 9,4E-16 331,7 2,6 0,04302 0,00024
2236 1,4846E-13 8,2E-16 416,6 2,3 0,06205 0,00007
2199 1,201E-13 1,5E-15 337,1 4,3 0,06998 0,00068
2261 1,789E-13 1,6E-15 502,0 4,4 0,07970 0,00044
2318 2,608E-13 1,9E-15 732,0 5,2 0,03318 0,00014
2318 2,558E-13 3,1E-15 717,8 8,6 0,05713 0,00050
2386 3,924E-13 4,5E-15 1101 13 0,04806 0,00046
2441 5,518E-13 7,7E-15 1549 22 0,01357 0,00017
2468 7,019E-13 4,7E-15 1970 13 0,0020438 0,0000079
Table 5: Am-241 results

26
References

[1] ,
KNOLL, Glenn, F. Radiation detection and measurement , third edition, John Wiley and Sons,
New York, 2004.
[2] WINKLER, Alexander , HILDÉN, Timo , BRÜCKEN, Erik J. A gaseous proportional
counter built from a conventional aluminium beverage can, Helsingin Yliopisto, 2015, arxiv:
1509.02379.
[3] ,
CHONG, C.S ELYAASEERY, I.S , TAJUDDIN, A.A Gamma ray spectrum of Am-241 in a
back scattering geometry using a high purity germanium detector School of Physics, Universiti
Sains, Malaysia
[4] https://en.wikipedia.org/wiki/Iron-55
[5] https://en.wikipedia.org/wiki/Americium-241
[6] https://en.wikipedia.org/wiki/Proportional_counter
[7] Lecture notes from Laboratory course on instrumentation, PAP328,HILDÉN, Timo ,
BRÜCKEN, Erik J. , Helsingin Yliopisto

[8] https://en.wikipedia.org/wiki/Radiation
[9] http://www.nucleide.org/DDEP_WG/Nuclides/Am-241_tables.pdf

27