Detection of Fission Neutrons Using Semi-Insulating Silicon Carbide Detectors*

1
Detection of Fission Neutrons Using
Semi-Insulating Silicon Carbide Detectors

• Frank H. Ruddy, Robert W. Flammang, John G.
  Seidel, Scott M. Watson, and James T. Johnson
• Science Technology Dept., Westinghouse Electric
  Co.
• Pittsburgh, Pennsylvania, USA
• Idaho National Laboratory, Nuclear
  Nonproliferation Department, Idaho Falls, Idaho,
  USA

This work was supported by the US Department of
Homeland Security Domestic Nuclear Detection
Office under Contract HSHQ-07-C-00041
2
Detection of Fission Neutrons Using
Semi-Insulating Silicon Carbide Detectors

• Silicon Carbide (SiC) semiconductor
  fission-neutron detectors have been developed for
  US Department of Homeland Security applications
• SiC detectors have been shown to be capable of
  detecting fission neutrons from concealed Special
  Nuclear Materials in pulsed neutron and photon
  interrogation environments.

3
Advantages of SiC Detectors

• SiC detectors can count fission neutrons from SNM
  immediately following intense neutron bursts from
  an interrogating source
• Thermal and epithermal neutron-induced fission
  neutrons will be missed by alternative detectors,
  which require time to recover from the neutron
  burst
• Highly selective - in the example shown, 2336
  fission neutrons were detected with a 235U sample
  present and zero counts when the sample was
  removed

(Results reported at the IEEE2007 NSS)
4
Advantages of SiC Detectors

• SiC detectors are insensitive to thermal and
  epithermal neutrons and gamma rays and provide
  excellent g/n discrimination
• Fast-neutron pulses have been observed during
  intense gamma bursts at the INL Varitron 10-MV
  electron accelerator
• An unshielded SiC detector was located 1 m from
  the converter in a 9000-12,000 R/hr gamma field

(Results reported at the IEEE2007 NSS)
5
Development of Radiation Detectors Based on
Semi-Insulating Silicon Carbide

• Silicon Carbide (SiC) radiation detectors based
  on diode devices have been developed by many
  groups for a variety of applications
• SiC Schottky diode and p-i-n diode detectors use
  lightly doped epitaxial SiC layers on conducting
  SiC substrates
• The thickness of the detector active volume
  corresponds to the depletion thickness of the
  epitaxial layer
• Although SiC epitaxial layers thicker than 200 µm
  have been produced, only layers up to about 120
  µm can be produced routinely
• Furthermore, to fully deplete the epitaxial layer
  at modest voltages (lt1000 V), the unintentional
  doping concentration must be limited to lt1014
  atoms/cm3 (Nitrogen is typically used)

6
SiC Fast-Neutron Detectors

• A SiC Schottky diode is shown
• The n- epitaxial layer, when depleted by a
  reverse bias, is the active volume for the
  detector
• Ionization in the
• active volume is
• collected under
• the influence of
• the applied bias
• Detector response
• is due to fast-
• neutron reactions
• no internal detector die-away time

7
SiC Fast-Neutron Detectors

• The detector response is based on ionization
  produced by neutron-induced reaction products in
  the detector active volume
• Carbon and Silicon neutron-reaction response is
  proportional to active volume

8
Development of Radiation Detectors Based on
Semi-Insulating Silicon Carbide

• An alternative is to use semi-insulating SiC
• Semi-insulating SiC detectors have been
  demonstrated
• M. Rogalla, et al., Nuclear Physics B 78, 516-520
  (1999)
• W. Cunningham, et al., Nuclear Instruments
  Methods A 509, 127-131 (2003)
• However, the semi-insulating SiC used was heavily
  doped with Vanadium, which produces deep-level
  charge traps and causes low charge-collection
  efficiency

9
Development of Radiation Detectors Based on
Semi-Insulating Silicon Carbide

• Recently, high-quality un-doped semi-insulating
  SiC wafer material has become available from
  Cree, Inc.
• These wafers contain low residual impurities
  corresponding to background net doping
  concentrations lt105 cm-3
• 360-µm thick wafers are readily available, and
  thicknesses up to 2-mm are potentially available
• SiC fast-neutron detectors using these
  semi-insulating materials have been manufactured
  and tested

10
Diode Detectors vs. Semi-Insulating SiC
SiC Schottky Diode
Semi-Insulating SiC

• SiC diode detectors thickness of epitaxial
  layer determines active volume for the detector
  (limited to 120 µm)
• Semi-insulating (S-I) detectors entire wafer
  thickness comprises the active volume
• 150-µm, 160- µm, and 360- µm S-I SiC detectors
  have been tested and up to 2000-µm wafers are
  available

11
SiC Detector Fabrication

• 0.78-mm2, 7.07-mm2, and 28.3-mm2 detectors were
  made on 150-µm, 160-µm and 360-µm SiC wafers by
  GeneSiC Semiconductor, Inc. of Dulles, VA USA

Device reticule is repeated over entire 7.62-cm
wafer area
12
Present Work

• Fission-neutron response characteristics were
  compared for a 28.3-mm2 x 100-µm Schottky diode
  and a 28.3-mm2 x 160- µm S-I SiC detector
• Both detectors were equipped with 100- µm
  polyethylene proton recoil converter foils
• Detection of thermal-neutron induced fission
  neutrons was demonstrated in a pulsed neutron
  interrogation environment

13
Detection of 252Cf Fission Neutrons

• Detectors were placed within an RF shielding box
  at a distance of 10 mm from a 2.11 x 107 s-1
  252Cf neutron source
• A fast Ortec VT-120 amplifier was used with an
  Acqiris digitizer to record response pulses

• Pulse amplitude spectra were recorded for the
  Schottky diode and for the S-I SiC detector

14
SiC Schottky Diode 252Cf Fission-Neutron Results

• Pulse amplitude spectra were recorded as a
  function of voltage at two different triggering
  levels
• Pulse amplitude increases with voltage
• Count rate increases with voltage

Triggering Level 1
Triggering Level 2
15
SiC Schottky Diode 252Cf Fission-Neutron Results

• The count rate increases are due to increasing
  depletion depth with voltage
• The increases in pulse amplitude are a result of
  decreasing capacitance with voltage
  (C?1/thickness)
• The maximum pulse height is VmaxQ/C , where Q is
  the charge collected and C is the capacitance
• Since the neutron-induced reaction products have
  ranges that are much less than the active volume
  dimensions, Q is nearly constant. Therefore the
  increase in pulse amplitude with voltage is a
  result of increasing depletion depth and
  decreasing C.

16
SiC Schottky Diode 252Cf Fission-Neutron Results

• The integral counts for each spectrum are shown
  as a function of voltage
• The depletion depth for each voltage can be
  calculated
• The number of counts recorded is directly
  proportional to depletion depth or active volume
  as expected

17
S-I SiC Detector 252Cf Fission-Neutron Results

• Pulse amplitude spectra were recorded as a
  function of voltage for both positive and
  negative biases
• Pulse heights increase with increasing voltage
• Since C is constant, increase must be due to Q
  more charge collected at higher voltages

Positive Bias
Negative Bias
18
S-I SiC Detector 252Cf Fission-Neutron Results

• Pulse height spectra are similar at the same
  absolute voltage
• More counts are observed when bias is negative
• Both show more counts at higher voltages
  positive-bias case is nearly linear
• Many fewer counts observed for S-I SiC than for
  Schottky

19
S-I SiC Detector 252Cf Fission-Neutron Results

• Pulse amplitudes are also much less for S-I SiC
  than for the Schottky diode
• S-I result at 800 V is similar to Schottky result
  at 200 V the 200-V Schottky depletion depth is
  only 39 µm
• Both reduced pulse heights and fewer counts in
  S-I SiC are indicative of incomplete charge
  collection
• Although increasing counts were observed with
  increasing voltages for S-I SiC, high-voltage
  testing was limited by detector packaging issues

20
S-I SiC Detector Packaging Issues

• On-chip I-V measurements yielded excellent
  results
• Detectors showed resistive behavior with leakage
  currents generally less than 1 nanoampere
• Packaged devices had high leakage currents and
  breakdown voltages 800 V

(Measurements made by GeneSiC Semiconductor, Inc.)
21
S-I SiC Neutron Efficiencies

• S-I SiC results are both disappointing and
  encouraging
• Fast-Neutron detection efficiencies are lower
  than expected
• Results indicate that higher efficiencies can be
  obtained than for thickest Schottky diodes
  available
• Higher efficiencies can be expected with better
  packaging and higher voltages

22
Improvement of Fast-Neutron Efficiency for
Semi-Insulating SiC Detectors

• In-addition to higher voltage packaging, 3D
  electrode designs can be explored
• Proposed by Parker, et al. (NIMA 395, 328-343
  1997), 3D configurations are being explored for
  SiC as well as Si and other semiconductor
  detectors for high-energy physics experiments

23
Summary and Conclusions

• Semi-insulating SiC neutron detectors are a
  promising alternative to diode detectors
• Semi-insulating SiC detectors are less expensive
  than epitaxial SiC diode detectors
• Detectors with 360-µm thick active layers are
  readily available, and wafers up to 2 mm thick
  can be obtained
• However, charge collection problems with
  semi-insulating detectors must be overcome

24
(No Transcript)
25
Advantages of SiC Detectors
(Results reported at the IEEE2006 NSS)

• Temperature insensitivity
• No drift in response observed in the range from
  from 18 0C to 304 0C
• Resistance to Radiation Effects
• The response of SiC detectors is unaffected by
  large radiation exposures
• Therefore, SiC detectors can operate in extreme
  ambient temperature and radiation environments
  for extended time periods

Detector still works after massive gamma-ray
exposure
26
Advantages of SiC Detectors

• Unprecedented signal-to-background has been
  achieved in high-energy neutron interrogation
  environments
• 2336 fission neutrons were observed between
  100-ms source bursts with zero background when
  the SNM was removed

(Results reported at the IEEE2007 NSS)