BORON CARBIDE BASED

1
BORON CARBIDE BASED NEUTRON DETECTORS
Ellen Day Manuel Diaz Andrew Harken Carl
Lundstedt Brian Robertson Shireen
Adenwalla University of Nebraska-Lincoln
2
Funded by the ONR, State of Nebraska and NASA.
3
NEUTRON DETECTORS
Enforce nuclear safeguard agreements Monitor
reactors, nuclear stockpiles Homeland security
applications Neutron scattering science at the
new Spallation Neutron Source Planetary and
space exploration
ALL SOLID STATE NEUTRON DETECTOR Compact Sturd
y Low power operation (even 0V!)
4
PROBLEMS
Most materials have a tiny capture cross section
for neutrons ENTER BORON CARBIDE SEMICONDUCT
ING LARGE capture cross section
5
SEMICONDUCTING FORM OF BORON CARBIDE 1990S PETER
DOWBENS GROUP
Boron Carbide on n-type Si
Nickel doped boron carbide on p-type Si
6
Source Molecule
DEOPSITION USING PLASMA-ENHANCED CHEMICAL VAPOR
DEPOSITION (PECVD)
C2B10H12 (ortho-carborane) Closo-1,2-dicarbadod
ecaborane
7
Film Deposition
Sublimated ortho-carborane molecules (70 ? )
with Ar gas (90 ?)
Chamber Temp 90 C Pressure 200 mTorr Gas
Flow Rate 10 sccm Deposition rate
80 nm/10 min
Ellen Day
8
FACT SHEET
? Capture c.s. for thermal neutrons ? 3850
barns ? ? 1/vE
10B n ? 7Li(0.84 MeV) 4He(1.47 MeV)
?(0.48MeV) 94 ? 7Li (1.02 MeV )
4He ( 1.78 MeV ) 6
Abundance of 10B in naturally occurring B 20
All neutron data at KSU TRIGA tangential beam
port 10.6 n/cm2secW
9
10B Capture Cross Section
? 1/vE
10
EFFICIENCY1-exp(-N?x)
11
IS IT A TRUE SOLID STATE DEVICE? CONVERSION
LAYER 1. B RICH LAYER CAPTURES NEUTRONS 2. CHARGE
IS CAPTURED BY SEMICONDUCTING LAYER (DISTINCT
FROM B RICH LAYER) TRUE SOLID STATE BOTH
NEUTRON CAPTURE AND CHARGE CAPTURE OCCUR IN THE
SAME MATERIAL IN THIS CASE BORON CARBIDE
IMPLICATIONS FOR HIGHEST ATTAINABLE EFFICIENCY
12
Distinction between Different Devices

• Path 1
• Path 2

Conversion layer
Diode device
13
Simplified Model Comparisons with GEANT4 for
Planar Conversion and All-B5C Detectors
(Lundstedt et al, NIMPR, 2005)
14
Initial 1 mm B5C / Si heterojunction diodes
(from 3 wafers with 276nm B5C)
Up to gt 3 x 105 electrons / neutron
15
Irradiation by 1015 n/cm2
16
All BC device
BC on Sapphire Cr/Au contacts Purely ohmic I-V
17
ALL BC RESISTIVE DEVICE
?1.8 X 106 ?-cm
Vapp7.5 V E field 3 X 104 V/cm dBC 250 nm
18
MONTE CARLO SIMULATION FOR A 500 nm ALL BC DEVICE
USING GEANT 4
NO SMEARING
30 KEV 50 KEV SMEARING
19
Calc. efficiencyN?x
EXPERIMENTAL EFFICIENCY 5 X 10-5 CALCULATED
EFFICIENCY 2 X 10-3
1 40

Incomplete charge collection-DISMAL FAILURE!
20
BUT Capacitance measurements indicate
effective area only 10of what we thought it
was! Brings down incident neutron flux by
factor of 10 NOW.
EXPERIMENTAL EFFICIENCY 5 X 10-4 CALCULATED
EFFICIENCY 2 X 10-3
0.25 !!
21
BC/Si DIODE
Vapp 0 V
d230 nm
E.E. Day et al J. Phys. D, 39, 2920 (2006)
22
WITH INCREASING Vapp, PEAK POSITIONS CHANGE
23
PEAK POSITION vs. APPLIED BIAS
Increase of 60 in going from 0V to 3.15V
INCREASED charge collection higher E field,
larger depletion region (in Si)
24
EFFICIENCY
10 increase
1. Operation at 0V bias is both feasible AND
efficient 2. Increased bias DOES increase the
charge collection significantly but not the
efficiency. i.e. such a LARGE amount of charge
is liberated per neutron capture event that even
if only small fractions are collected they still
provide a large enough signal above detector
noise
25
SCOPE TRACES AND CHARGE COLLECTION TIMES
26
CONCLUSIONS

• Semiconducting boron carbide is a promising and
  versatile
• material for neutron detection.
• All BC device WORKS! (albeit inefficiently)
• 3. Detected signal varies considerably depending
  on thickness
• of material, device types, device parameters
• 4. In order to understand the neutron generated
  signal,
• comparisons with modeling and/or simulations
  proves
• insightful.
• Future developments
• Higher resistivity material Homojunction diodes
• Measurements of relevant semiconducting
  parameters
• Enriched boron carbide

27
Monte Carlo

• Monte Carlo simulation toolkit used was GEANT
  4.5.2. and GEANT 4.8.0.
• GEANT handles the entire physical simulation.
• Detector construction (Materials and geometry),
  neutron transport, capture cross sections etc
• GEANT tracks the ion energy loss as it travels
  through the material.

28
Simulation Assumptions

• GEANT simulation assumptions
• All the energy deposited is converted into
  electron hole pairs
• All charges generated are collected by the
  electrodes
• Boron carbide layer (B5C) constructed out of 19
  10B and 81 11B natural occurrence
• No electronic noise incorporated

29
Base Device

• Device Construction
• Thickness of Boron Carbide layer varied
• 381µm Si layer.
• Thermal neutron beam is incident normal to the
  Boron Carbide surface.
• Capture cross section for 0.025eV neutron is 3840
  barns.

30
Diode
Conversion Layer
10 Million Neutrons Incident
31
500nm plus moderator

• Device Construction
• 500nm Boron Carbide layer
• 381µm Si layer.
• Moderator
• H2O
• Thickness varied
• Neutron Energy
• 100eV

32
(No Transcript)
33
Calorimeter

• Measures the energy of particles by
  discrimination.
• Neutrons with energy greater than their
  surroundings will lose energy by elastic
  scattering in the moderator.

34
Simplified Calorimeter

• Device construction
• 500nm Boron Carbide layer
• 381µm Si layer
• Moderator
• H2O
• 1.5cm
• 4.0cm
• Neutron energy
• 100eV

35
Calorimeter
36
Calorimeter Modification

• Graphite Reflector added
• 1.0 cm thick

37
Graphite Reflector
Top Detector
Bottom Detector
____ no moderator no reflector ____ graphite
reflector above moderator ____ moderator no
reflector
38
(No Transcript)
39
Conclusions

• Monte Carlo modeling has proven successful in
  determining the expected spectra for novel
  neutron detectors.
• Modeling has demonstrated that using a
  calorimeter incorporating boron carbide detectors
  can be effective as an energy discriminator.

40
ACKNOWLEDGEMENTS
FUNDING Office of Naval Research
N00014-04-1-0605 Nebraska Research Initiative NASA
MIKE WHALEY TROY UNRUH For technical support
in the TRIGA MARK II Nuclear Reactor Facility
at the Kansas State University.
41
BC/Si Diode d650 nm
42
(No Transcript)
43
500 nm Conversion Layer BC with Additional
smearing (eg incomplete charge collection)
44
MONTE CARLO SIMULATION (GEANT)
45
Substrate Cleaning
Acetone, Methanol, DN water
Ar Plasma Etching
Base Pressure lt 10-7 Torr Working Pressure
200 mTorr RF Power 30 Watt Ar Gas Flow Rate 8
sccm Ar Gas Temperature
Ellen Day