Unit 2
General Properties of Radiation Detectors
1 Nuclear Instrumentation and Measurement
MANE-4350
Chapter 4: General Properties of Radiation Detectors
Simplified Detector Model
• A single particle interacts with detector chamber
• Instantaneous energy transfer and deposition
• Electric charge Q appears within detector at t = 0
• Collection of charges through electric field within detector
• Collection time ms ~ ns (dependent upon mobility and distance)
• Time dependence of the detector current
In the real world, there are multiple particles with randomly distributed time intervals.
Detector Operation
Three distinct modes of operation
(1) Pulse Mode – for radiation spectroscopy
(2) Current Mode – high event rate
(3) Mean Square Voltage (MSV) – high event rate
2 Nuclear Instrumentation and Measurement
MANE-4350
A. Current Mode
t +T
I (t ) = 1 ∫ i(t ' )dt ' = average of individual current bursts
T t
The average current = I = rQ = r E q
0 w
(Eq. 4-1 and Eq. 4-2)
where r = event rate
Q = charge from each event
E = average energy deposition by each event
w = average energy required to produce an
electron-ion pair
q = 1.6 x 10-19 C
Recall from Poisson statistics that σ n = n = rT (Eq. 4-5)
Define σ T (t ) = the time average of standard deviation in the measured current.
The fractional standard deviation is the measured signal due to random fluctuations in pulse
arrival time is given by:
3 Nuclear Instrumentation and Measurement
MANE-4350
σ I (t ) σn rT 1
= = = (Eq. 4-7)
I0 n rT rT
• Small T → large uncertainty at any time t, but fast response
• Large T → small uncertainty at any time t, but slow response
B. Mean Square Voltage (MSV) Mode
The result corresponds to the time average of the standard deviation in the measured current
σ I (t )
σ I (t ) 1
From = and I 0 = rQ
I0 rT
r rQ 2
⇒ σ I (t ) = Q ⇒ σ I (t ) =
2
~ output
T T
Most useful for making measurements in mixed radiation fields (neutron and gamma).
4 Nuclear Instrumentation and Measurement
MANE-4350
C. Pulse Mode
Time constant of the measuring circuit τ = RC
Case #1: Small τ (τ = RC << tc)
• V(t) = i(t) in shape
• Used for high event rates or timing info is important
Case #2: Large τ (τ = RC >> tc)
• Detector current is integrated on the capacitance
• Used for low event rates of timing info is not important
• Most common means of pulse-type mode
Observations
(1) The leading edge is detector dependent only (not external circuit)
(2) The trailing edge is circuit dependent
Q
(3) The amplitude of the signal Vmax = ~ Q ~ Energy of particle
C
(4) Rate of measured pulses → rate of radiation interaction
5 Nuclear Instrumentation and Measurement
MANE-4350
Pulse Height Spectra
One interaction → pulse height Vmax ~ Q ~ energy
Many interactions → pulse height fluctuates due to
(1) fluctuation of energy
(2) fluctuation of detector response
(3) fluctuation of C
6 Nuclear Instrumentation and Measurement
MANE-4350
Most common way of displaying pulse amplitude is through the “differential pulse height
distribution.”
7 Nuclear Instrumentation and Measurement
MANE-4350
The number of pulse whose amplitude lies between two values H1 & H2 can be obtained by
integrating the area under the distribution:
H2
dN
# of pulses ( H1 → H 2 ) = ∫
H1
dH
dH (Eq. 4-10)
∞
dN
Total # N 0 = ∫ dH (Eq. 4-11)
0
dH
The integral pulse height distribution (fig 4-2b)
∞
dN
N (H ) = ∫ dH
H
dH
8 Nuclear Instrumentation and Measurement
MANE-4350
Counting Curves and Plateaus
To experimentally determine N0(H), a discriminative level (Hd) can be used in the counting
device so that only pulse exceeding Hd will be counted.
Vary Hd (by adjusting Gain) and plot N0(H) vs. Gain
Look at the plateau (called counting plateau)
Variation in Gain (Hd) results minimal change in N0(H)
⇒ Maximum counting stability!
9 Nuclear Instrumentation and Measurement
MANE-4350
Energy Resolution
FWHM
resolution R = (Eq. 4-13)
H0
FWHM = Full Width at Half Maximum
(Background on which the peak may be superimposed is subtracted)
Causes of imperfect energy resolution:
10 Nuclear Instrumentation and Measurement
MANE-4350
• Drift of detector operating condition (ex. Power supply)
• Random noise within the detector
• Statistical fluctuation from the signal itself → dominant source of fluctuation
Consider statistical fluctuation only:
Assuming the formation of each charge carrier is a Poisson process, a standard deviation of
N is expected for a total of N charge carriers. If this is the only source of fluctuation, the
FWHM figure above is a Gaussian shape.
The Gaussian function can be written as:
(H − H 0 )2
−
A 2σ 2
G(H ) = e
σ 2π (Eq. 4-12)
σ determines the FWHM by FWHM = 2.35σ
Because the response of detectors is linear
⇒ H0 = KN
where K = a proportionality constant
N = charge carriers
The standard deviation σ of the peak is the pulse height spectrum ⇒
σ H = K N ⇒ FWHM = 2.35K N
0
(Because σ N = N and H 0 = KN )
Resolution due to statistical fluctuation
11 Nuclear Instrumentation and Measurement
MANE-4350
FWHM 2.35 K N 2.35
R= = = (Eq. 4-13)
H0 KN N
R improves as N increases
R ≤ 1% when N ≥ 55,000
(Implication to detector design Semiconductor → large # of N)
2.35
a) Robserved =
N
For some detectors, Poisson statistics is not valid.
Fano factor is used to quantify the departure of the observed statistical fluctuation in the number
of charge carriers from pure Poisson statistics.
Observed Variance in N ( N obs )
F= 0 ≤ F ≤ 1.0 (Eq. 4-14)
Poisson predicted Variance ( N )
⇒ N obs = NF
FWHM 2.35K NF F
⇒ Rabs = = = 2.35 (Eq. 4-15)
H0 KN N
2.35
b) Robserved ≥
N
(FWHM )overall
2
= (FWHM )stat + (FWHM )noise + (FWHM )drift + ....
2 2 2
12 Nuclear Instrumentation and Measurement
MANE-4350
Detection Efficiency
number of pulses recorded
Absolute (∈abs ) = (Eq. 4-16)
# of particles emitted by source
number of pulses recorded (Eq. 4-17)
Intrinsic (∈int ) =
# of particles incident on detector
4π
∈int = ∈abs
Ω
Ω = solid angle of the detector seen from the source
cos α
Ω=∫ 2
dA
A r
d
= 2π 1 − (Eq. 4-21)
d 2 + a2
πa 2
≈ (if d >> a )
d2
13 Nuclear Instrumentation and Measurement
MANE-4350
∈ peak
Peak-to-total ratio γ =
∈total
14 Nuclear Instrumentation and Measurement
MANE-4350
Dead Time
In all detectors, there will be a minimum amount of time that must separate two events in order
that they be recorded as two separate pulses.
This minimum time separation = dead time
A true event that is lost because it occurs too quickly following a preceding event = a dead time
loss
Two models of dead time behavior:
(1) Paralyzable
(2) Nnonparalyzable
(1)
(2)
• The two models differ only when true event rates are high.
• Represent two extremes of idealized behavior; real detectors often display a behavior that
is in between.
15 Nuclear Instrumentation and Measurement
MANE-4350
Let
n = true interaction rate
recorded counts
m = recorded count rate =
total time
τ = system dead time
a) For the nonparalyzable case:
The fraction of time that the detector is dead = mτ
Rate of true event lost: nmτ = n – m
m
n=
1 − mτ (Eq. 4-24)
Example:
Suppose τ = 200 µs and m = 30,000 counts/min
Find the fraction of counts due to dead time and the true counting rate.
Solution:
30000
m 60 counts
True counting rate n = = = 555.5
1 − mτ 1 − 30000 200 × 10 −6 s
60
555 − 500 30000
Fraction of dead time loss = = 10% of counts = mτ = 200 × 10 −6
555 60
b) For paralyzable case:
Dead periods are not always of fixed lengths
m = ne − nτ (Must solve by iteration) (Eq. 4-27)
16 Nuclear Instrumentation and Measurement
MANE-4350
Dead Time Measurement:
Two-source method:
-Count source 1 and source 2 separately
(counting rates m1 and m2)
-Count source 1 and source 2 together
(counting rates m12)
• Let n1, n2, and n12 be the true counting rates (sample + background) with source 1, source 2, and
the combined sources respectively.
• Let nb and mb be the true and measured background rates with both sources removed.
(Eq. 4-30)
n12 − nb = ( n1 − nb ) + ( n2 − nb )
using nonparalyzable model
m12 mb m1 m2
+ = + (Eq. 4-31)
1 − m12τ 1 − mbτ 1 − m1τ 1 − m2τ
Solve fro τ (assuming m b = 0) ⇒
m m − [m1m2 (m12 − m1 )(m12 − m2 )]2
1
τ= 1 2 (Eq. 4-33)
m1m2 m12
17 Nuclear Instrumentation and Measurement
MANE-4350