Resistive Plate Chambers with Gdcoated electrodes as thermal neutron detectors

1
Resistive Plate Chambers with Gd-coated
electrodes as thermal neutron detectors

• R. Trentadue for
• DIAMINE Collaboration WP-2 BARI
• M. Abbrescia, G. Iaselli, T. Mongelli,
• R. Trentadue, A. Ranieri, V. Paticchio

2
Total cross section for neutrons
3
Natural Gd
Natural Gd has the following isotopic composition
interesting isotopes are about 30
As a consequence of the capture process of a
thermal neutron, Gd produces, in the 60 of
cases, an electron from internal conversion
complex energy spectrum
4
Gd as a converter
Gd is a metal, weakly reacting in humid air,
where it oxidises. It is cheap, except when
required in very thin layers (order of ?m).
It is difficult and expensive to obtain Gd
enriched in 157Gd (material of strategic interest)
Gadolinium Oxide Gd2O3 (vulg. Gadolina) is a
white inert powder (easy to handle), with granuli
of 1-3 ?m in diameter, very cheap.
5
The layer of converter
It is constituted by Gd2O3 mixed with linseed
oil the mixture is sprayed on the bakelite
electrodes, which are used to build standard RPCs.
Linseed oil is standardly used on the inner
surfaces of RPCs built with bakelite (but it is
deposed in a different way). It is used by the
future LHC experiments, by ARGO, OPERA, etc.
(also by BABAR)
Mirror surfaces
6
The chambers
3 RPCs 10x10 cm2 in dimensions
1 without Gd2O3, used as a reference
2 with a different concentration of the oil-Gd2O3
mixture
High Voltage
Gas
Signal readout
7
Operating conditions
Signal readout copper pad Signal input to NIM
discriminator, Vthr30 mV
Gas mixture in Bari (with cosmics) Ar-Isobutane-
C2H2F4 70/27/3 C2H2F4-Isobutane 97/3
At Geel, kindly offered by out colleagues from
Torino, that we thank very much
C2H2F4-Isobutane-SF6 97/2.5/0.5
Operating voltage 10-11 kV (streamer mode)
8
How the system works
e-
U
RPC
CI
t0 start DAQ
TDC1
TDC2
tn stop to a multihit TDC
9
The chambers at Geel
Flightpath15 m (CI 13.5 m)
Frame in plastic material (the RPCs are in
plastic material too)
Backward configuration
10
Layer thickness is not important!
Home made simulations checked with the ones
already present in literature, made with MCNP
backward e-
Since neutron intensity, in Gd, decreases
exponentially, just the first layer takes
part to the conversion process
Backward e- have always the same thickness to
cross
11
Some raw data
Ionisation Chamber
These are events whose time with respect to t0
is known
Spectra acquired at the same time for RPC and CI
RPC HighConc Gd
comparison between RPC and CI

• Two regions
• thermal n
• resonances

12
Some raw data 2
Spectra in the resonance zone (few eV)
Ionisation Chamber
These are resonances due to the presence of
filters on the beam W, Na, Ag, S, Co
RPC HighConc Gd

• They are not coincident because
• the flightpath is different between RPC and CI
• the delays are different

13
How you pass from time to energy
Particles with known energy are taken as a
reference
T.O.F. known
delays of the system
energy of the other particles
A resonance can be used as a reference
For the other energies
14
The results
Now the two spectra are calibrated energy of
the resonances are coincident
Ag
Some peaks are present in the spectrum peaks of
the Gd cross section
W
Energy resolution for RPC worse than for CI
15
The thermal neutron region
Relative efficiency
Roughly ? 2.5-3
Conversion efficiency of 10B well known
16
The background of the measure
A way use a Cd filter opaque to neutrons
with Ekinlt 0.5 eV (Cd cutoff)

• Advantages
• data coming from the same chamber (also for CI)
• run in the same conditions

Noise distributed uniformly in time and not in
energy
17
Efficiency
Subtracting the background
Integral efficiency
Differential efficiency
18
Simulation
We have simulated the behaviour of our detector,
by means of a simple model, with the following
hypotheses

• Input
• spectrum deconvolution of Ionising Chamber
  spectrum with respect to 10B cross section
• cross section of natural Gd computed as the
  weighted average of the most important isotopes

• Backward configuration
• emission probability for electrons after neutron
• capture by Gd 60
• Range for electrons in Gd ? 25 mm

Conditions
19
Simulation result
This is the spectra foreseen using as input a
Maxwellian spectrum and Gd cross section

• The agreement between the two curves
  demonstrates
• all sources of background have been correctly
  taken into account and subtracted
• we are observing only thermal neutrons

20
Conclusions

• We have developed a simple method (useful for
  practical and industrial applications) but very
  effective, to render RPCs detectors for thermal
  neutrons

• These are detectors easy to build, robust,
  light-weighted, cheap, adapt for industrial
  applications, etc.

• Detector has an efficiency gt 2.5 eff. CI ? 6

• RPC-Gd exp. eff. is gt 10B theoretical maximum
  eff. gtgt 10B-RPC experimental eff.

• Coupling two of these detectors together
  efficiency reaches about 3.5-4 eff. CI (analysis
  in progress)

We are still far from max. th. eff. for Gd-RPC
... Possible improvements Gd concentration
optmisation, linseed oil polimerisation
procedure, more layers, ...
21
RPC FOR THERMAL NEUTRONS WITH BORON CONVERTER

• Converter material
• ? B or B4C with B enriched of
  10B ( gt 97)
• Charged particles to be detected
• ? a, 7Li through
  the reaction 10B(n,a)7Li
• cross section for thermal
  neutrons s ? 4 Kbar
• particles energies Ta ?
  1.8 MeV T7Li ? 1.0 MeV ( 6 of events)
• 
  Ta ? 1.5 MeV T7Li ? 0.8 MeV ( 94 of
  events)
• Conversion of thermal neutrons into charged
  particles inside the active volume
• ? converter coating on the
  electrode surface facing the gas gap
• It is not possible to use linseed oil on the
  converter
• ? the surface of the converter
  layer must be very smooth
• ? realization of a prototype
  with glass electrodes

Pick-up x strips 2 cm wide
HV
Insulating film
glass 2mm
Converter coating
Gas gap 2mm
glass 2mm
Insulating film
Pick-up y strips 2cm wide
GND
22
RPC PROTOTYPE TEST
RPC prototype with 0.35 ?m B4C coating has been
tested in avalanche mode with thermal neutrons
at the European research centre JRMM
operation voltage for neutron detection
9500?9750 V efficiency for cosmic rays at 9500 V
? 20 single count rate due to detector
noise at 9500 V 0.04 Hz/cm2
Example of time of flight spectrum for neutrons
at 9750 V
W 18 eV
W 4 eV
Ag 5 eV
23
RPC PROTOTYPE TEST WITH THERMAL NEUTRONS
?E? ? 1000 ?Emip
? possibility to operate the RPC at low voltage
where the RPC is not fully efficient for cosmic
rays with a consequent noise reduction
HV ? 9500 ? ? ? constant
efficiency plateau for neutrons is
shifted at lower operation voltage than the
efficiency plateau for m.i.p.
24
How the system works 2
t0 arriving time of the e- bunch on Uranium
delays start signal for two multihit TDC (0.5
ns/bin)
RPC Ionisation Chamber
2 multihit TDC, each with 1 channel
DAQ separated for CI and RPC
CI two layers of 10B of 0.35 ??m each RPC in
order RPC-GdB, RPC-GdNB, RPC-Oil,
OR(RPC-GdBRPC-GdNB)
25
How the system works 4
Neutron energy is computed by starting from
tn-to Time Of Flight of neutrons
( every delay due to cables, electronics, etc.)
burst 14 ns long
limit on energy resolution
26
How the system works 3
Operating conditions an e- burst every 10 ms
(100 Hz)
Maximum acquisition time/burst 10 ms
Maximum time tmax between the first ? flash
and the last particle slow
n
10 ms
tmax 10 ms (100 Hz) Emin11.7 meV tmax
1.25 ms (800 Hz) Emin 748 meV


27
How the story goes on
The chambers have been brought to Geel, where we
could use
GELINA Geel Electron Linear Accelerator
An e- beam on an Uranium target produces, for
Bremsstrahlung, ? which, in turn, produce, via
photonuclear emission, neutrons
Energy from a few meV to 20 Mev 12 flightpaths
from 8 to 200 m
Peak Yield 4.5x1019 n/s Average Yield 3.4x1011
n/s
28
RPC energy resolution is worse because

• Gd conversion efficiency is smaller for E?eV
  (smaller cross section) less events, therefore
  poorer statistics
• Measured background about 6-9 greater near RPCs
  with respect to the CI zone about 1-1.5 ?Sv for
  CI, about 9 ?Sv for RPC (neutron dose). These
  are out of time neutrons, beacuse they do not
  come directly from the beam, but from other
  zones (B layer downstream the chambers, other
  scattering zones, etc.)
• The frame and the chambers are made from plastic
  materials again another source of out of time
  neutrons.

Are we interested (for these applications) to
resolve well resonances ???
29
The advantages of the method

• It is possible to obtain extremely uniform
  layers, with very constant thickness and density

• The electric properties (surface resistivity) of
  bakelite electrodes are not altered

• It is a method easily appliable to surfaces
  having large dimensions

• It can be used for industrial-scale applications
  (as required for practical uses), and factories
  have a great experience about it it is the very
  same method used to paint cars

30
The background of the measure

• An upper limit to the background for our measure
  can be obtained
• by examining the run acquired using the chamber
  without Gd
• In this case events in the thermal neutron region
  are due to
• Noise of the chamber
• Neutrons converting somehow in bakelite or gas
  (conversion prob. ?10-3)
• Out of time neutrons (from the background)

31
that is calibration
A filter is inserted and taken off, and the peak
which desappears is observed
resonance energy
For the other energies
It is better to use a resonance as a reference,
because operating conditions are more similar to
the ones of the thermal neutron region
This procedure is made separatly for CI and RPC
32
10B
Solid converters with the largest cross section
for thermal neutrons are
157Gd 250 kbarn 155Gd 75 kbarn 10B 3.8 kbarn
(it is obvious to use these materials)
Natural B is composed for 19.7 of 10B, but
obtaining enriched Boron (gt90) with 10B is
relativly simple and cheap
73Li 42?
E? 1.8 MeV (ground state) ? 6.3
105B 10n
73Li 42?
E? 1.5 MeV (excited state) ? 93.7
Produced ? have a range in 10B of about 3-4 ??m
Maximum useful thickness