R15-2
2008 IEEE Nuclear Science Symposium Conference Record
Micro-Structured High-Efficiency Semiconductor Neutron Detectors D.S. McGregor, Member, IEEE, S.L. Bellinger, W.J. McNeil, Abstract- Perforated semiconductor diode detectors have been under development for several years at Kansas State University for a variety of neutron detection applications. The detectors are fabricated from high purity n-type Si. Sinusoidal trenches are etched into the substrate, into which shallow p-type junctions are diffused. The trenches are then backfilled with 6LiF powder to make the device sensitive to neutrons. Thermal neutron measurements from a 0.0253 eV diffracted neutron beam yielded 170/0 intrinsic detection efficiency for devices with 50 micron deep trenches and 29% intrinsic detection efficiency for devices with 100 micron deep trenches.
I.
INTRODUCTION
icrostructured, or perforated, semiconductor neutron detectors have been under investigation at Kansas State University for several years [1-11]. The devices are constructed by etching features into a semiconductor substrate and subsequently backfilling those trenches with neutron reactive material. The device structure has the potential of achieving intrinsic thermal neutron detection efficiencies, Em, greater than 35% [4, 7, 11], much greater than common coated diode detectors which are restricted to ttn values no greater than 5% [12]. A clear advantage of the perforated structure is the high efficiency achieved with a single device. Although aligning common coated diode detectors in a stack can also achieve high efficiency, it would take more than 20 coated diode detectors aligned in a row to match the efficiency of a single perforated detector [12]. In the present work, the neutron reactive material is 6LiF, which relies on the 6Li(n,ttHe reaction. When thermal neutrons are absorbed in 6Li, a 2.73 MeV triton and a 2.05 MeV alpha particle are ejected in opposite directions. The reaction products from the 6Li(n,t)4He reaction are more energetic than those of the IOB(n,a)7Li or 157Gd(n;y) 158Gd reactions and, hence, are much easier to detect and discriminate from background radiations. 6Li has a relatively large microscopic thermal neutron absorption cross section of 940 b, although it is less than those for 157Gd and lOB. Further, Li metal is highly reactive and suffers decomposition outside of inert environments. Hence, the stable compound 6LiF is used instead. The use of inductively-coupled-plasma reactive-ion-etching (ICP-RIE) high-aspect ratio deep etching (HARDE)
M
Manuscript received Novemberl2, 2008. This work is supported in part by the Defense Threat Reduction Agency, contract DTRA-OI-03-C-0051, National Science Foundation, Grant 0412208, and the US Department of Energy, NEER Program grant No. DE-FG07-04IDI4599. D.S. McGregor, (telephone: 785-532-5284, e-mail:
[email protected]) S.L. Bellinger, S.A. Cowley, W.J. McNeil, E. Patterson, I.C.Unruh are with the SMART Laboratory of Kansas State Univ., Manhattan, KS USA 66506.
978-1-4244-2715-4/08/$25.00 ©2008 IEEE
T.e. Unruh
techniques have allowed for unique perforated neutron detector structures to be realized [5-7, 9, 10]. Previous test devices utilized isolated diffusions around the perforated regions and oxide passivation within the perforations to reduce leakage current [5, 6, 7]. A similar approach has been taken by other groups now duplicating this work [13, 14]. However, another design that reduces the leakage current offers performance with less process difficulty. The design incorporates dopants inside the perforations as well as the top surface, thereby making a pn junction within the perforations [3]. The diffusion process consumes surface damage caused by the reactive ion etch process, thereby eliminating the need for the oxide passivation, while reducing the reverse bias leakage current. As a result, pulse height spectra are more distinct and exhibit the expected shapes predicted elsewhere [11 ].
Fig. 1. Streaming effects are eliminated by using sinusoidal perforation patterns. The pattern increases the intrinsic thermal neutron detection efficiency while producing a flat detection response. In inset shows the pattern detail.
II. DETECTOR FABRICATION The perforations are etched with ICP-RIE into 10 k.Q n-type Si. The detector active area is 6 mm in diameter, which are batch processed on 3 inch diameter wafers. The devices reported in the present work have sinusoidal etched trenches ranging between 22 - 26 microns wide and 100 microns deep (Fig. I ). The sinusoidal design eliminates the angular dependence on sensitivity resulting from neutron streaming [5, 7-9], thereby increasing the intrinsic thermal neutron detection efficiency while producing a flat detection response throughout a wide solid-angle around normal incidence [5, 79]. After the etch process, the surface is patterned with photoresist and p-type regions are diffused uniformly into the
446
device structure, thereby forming pn junctions within the trenches. A Ti-AI metal contact ring is fabricated around the detector perimeter to make electrical contact to the p-type region. The back surface is coated with a metal contact to complete the diode structure. Finally, 6 LiF powder is packed into the perforations to perform as the neutron absorbing converter layer material (Fig. 2).
1000 -C
CV
-C
•
o Volts
• • •
0 Volts, Cd Shutter Closed -10 Volts -10 Volts, Cd Shutter Closed
Sinusoid Depth =50 microns Efficiency = 170~
~
0
CJ
CV
a::
100
en ...., c ~
0
0
10 0
500
1000
1500
2000
2500
3000
Channel Number Fig. 3. Pulse height spectra for 50 flIll deep sinusoidal patterns backfilled with 6LiF. At LLD = 100 keY, the measured thermal neutron detection
efficiency was 17% (14.82% at LLD = 300 keY, 13.8% at LLD = 500 keY).
1000 - - - - - - - - . . . , - - - - - - - - - - - ,
• o Volts
• • •
-c
CV
:
-C
I
0 Volts. Cd Shutter Closed -10 Volts -10 Volts, Cd Shutter Closed
~
Fig. 2. 6LiF powder is backfilled into the sinusoidal perforations.
o
III. NEUTRON MEASUREMENTS
a::CV
CJ
Neutron counting efficiency was measured with a 0.0253 eV diffracted neutron beam from the Kansas State University TRIGA Mark II nuclear reactor. The neutron flux was calibrated with a Reuter-Stokes 3He gas-filled proportional detector and found to be 1.2 ± 0.02 x 104 n cm-2 s-1. Pulse height spectra were collected from a sample perforated detector in the diffracted neutron beam with and without a Cd shutter, thereby allowing the collection of responses without and with neutrons. Prompt gamma-rays emitted from the thin Cd shutter appear in the spectrum as numerous pulses at low energy very near the noise level of the detector system (Fig. 3). The neutron counting efficiency was calculated by dividing the neutron counts collected from the detector by the flux determined with the 3He detector. Fig. 3 shows pulse height spectra from a sinusoidal device with 50 micron deep etched features. The device works well as a self-biased detector [15], operating solely on the built-in potential (Vb;) of the pn junction. However, they also operate well with approximately 10 volts reverse bias, as also shown in Fig. 3. With the shutter closed, the gamma-ray component (enhanced by the Cd shutter) was negligible at lower level discriminator (LLD) settings above 300 keV equivalent. The efficiency was determined by subtracting the gamma ray component from the total neutron spectrum, which actually underestimates the value of Etn • At a reverse bias of 10 volts, and with the LLD set to 100 keV, Etn was found to be 17%. Raising the LLD to 300 keY yielded Etn = 14.8%, and raising the LLD to 500 keY yielded Etn = 13.8%.
100
en ...., c
:::s
o o
500
1000
1500
2000
2500
3000
Channel Number Fig. 4. Pulse height spectra for 100 flIll deep sinusoidal patterns backfilled with 6LiF. At LLD = 100 keY, the measured thermal neutron detection
efficiency was 29% (27% at LLD = 300 keY, 25.1 % at LLD = 500 keY).
Fig. 4 shows pulse height spectra from a sinusoidal device with 100 micron deep etched features. The device also operates well as a self-biased detector [15], yet performs better when operated with approximately 10 volts reverse bias. With the shutter closed, the gamma ray component (enhanced by the Cd shutter) was negligible at LLD settings above 400 keY equivalent. The efficiency was determined by subtracting the gamma ray component from the total neutron spectrum, which again underestimates the value of Etn, since the Cd shutter actually increases the gamma-ray background. At a reverse bias of 10 volts, and with the LLD set to 100 keY, Etn was found to be 29%. Raising the LLD to 300 keY yielded Etn = 27%, and raising the LLD to 500 keY yielded Etn = 25.1 %. The actual gamma-ray flux with the Cd shutter in place is not known, however the n/y ratio for the detector was found to be 1.8 times larger than a 5 cm diameter, 10 cm long 3He detector operated under identical measurement conditions. Finally, the spectral shapes of the spectra in Figs. 3 and 4 correspond well with the expected results calculated elsewhere, in which a dip in the spectrum appear in the region between 100 keV and 800
447
Detector Filled with l~," Appl. Phys. Lett., 93, pp. 133502.1 133502.3,2008. [15] D.S. McGregor, S.M. Vernon, H.K. Gersch, S.M. Markham, SJ. Wojtczuk and D.K. Wehe, "Self-Biased Boron-l0 Coated High Purity Epitaxial GaAs Thermal Neutron Detectors," IEEE Trans. Nucl. Sci., 47 pp. 1364-1370,2000.
keV [11]. It is this same reduction in counts that allows for the sinusoidal design to be operated with an LLD setting of 500 keV without severely reducing the detector efficiency. IV.
CONCLUSIONS
Perforated Si detectors with sinusoidal trenches backfilled with 6LiF powder have been tested in a characterized diffracted 0.0253 eV thermal neutron beam from a TRIGA Mark II nuclear reactor. Detectors with 50 micron deep trenches achieved 17% ttn and detectors with 100 micron deep trenches achieved 29% ttn. The design with pn junctions diffused within the perforations has lower leakage current than previous passivated designs. The devices have a higher n/y rejection ratio than a 5cm x 10cm 3He detector. Spectra from the detectors match well to predicted results. REFERENCES [1]
[2] [3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
D.S. McGregor, R.T. Klann, H.K. Gersch, E. Ariesanti, J.D. Sanders, and B. Van Der Elzen, "New Surface Morphology for Low Stress ThinFilm-Coated Thermal-Neutron Detectors," IEEE Trans. Nucl. Sci., 49, pp. 1999-2004, 2002. D.S. McGregor and R.T. Klann, "Pocked Surface Neutron Detector," patent US-6545281; allowed April 8, 2003. D.S. McGregor, R.T. Klann, "High-Efficiency Neutron Detectors and Methods of Making the Same," patent US-7164138; allowed January 16, 2007. 1.K. Shultis and D.S. McGregor, "Efficiencies of Coated and Perforated Semiconductor Neutron Detectors," IEEE Trans. Nucl. Sci., NS-53, pp. 1659-1665,2006. WJ. McNeil, S. L. Bellinger, T. C. Unruh, E. L. Patterson, 1. K. Shultis, and D. S. McGregor, "Development of Perforated Si Diodes for Neutron Detection," IEEE Nucl. Sci. Symp. Con! Rec., San Diego, CA, Oct. 29Nov. 3,2006. WJ. McNeil, S.L. Bellinger, BJ. Blalock, C.L. Britton Jr., J.L. Britton, S.C. Bunch, W.L. Dunn, C.M. Henderson, TJ. Sobering, R.D. Taylor, T.C. Unruh, D. S. McGregor, "Preliminary Tests of a High Efficiency 1D Silicon Pixel Array for Small Angle Neutron Scattering," IEEE Nucl. Sci. Symp. Con! Rec, Waikiki, Hawaii, Oct. 28-Nov. 3,2007. S.L. Bellinger, WJ. McNeil, T.C. Unruh, D.S. McGregor, "Angular Response of Perforated Silicon Diode High Efficiency Neutron Detectors," IEEE Nucl. Sci. Symp. Con! Rec, Waikiki, Hawaii, Oct. 28Nov. 3,2007. CJ. Solomon, J.K. Shultis, WJ. McNeil, T.C. Unruh, B.B. Rice, and D.S. McGregor, "A Hybrid Method for Coupled Neutron-Ion Transport Calculations for lOB and 6LiF Coated and Perforated Detector Efficiencies," Nucl. Instrum. and Meth., A580, pp. 326-330,2007. D.S. McGregor, S.L. Bellinger, D. Bruno, WJ. McNeil, E. Patterson, 1.K. Shultis, CJ. Solomon, T. Unruh, "Perforated Semiconductor Neutron Detectors for Battery Operated Portable Modules," Proc. SPIE, 6706, pp. ONl- ONI2, 2007. D.S. McGregor S. Bellinger, D. Bruno, W.J. McNeil, E. Patterson, B.B. Rice, "Perforated Semiconductor Neutron Detector Modules," Proc. of 32nd Annual GOMACTech Conf, Lake Buena Vista, FL, March 19-22, 2007. D.S. McGregor and J.K. Shultis, "Designs for Micro-Structured Semiconductor Neutron Detectors," Proc. SPIE, 7079, pp. 0601 - 0615, 2008. D.S. McGregor, M.D. Hammig, H.K. Gersch, Y-H. Yang, and R.T. Klann, "Design Considerations for Thin Film Coated Semiconductor Thermal Neutron Detectors, Part I: Basics Regarding Alpha Particle Emitting Neutron Reactive Films," Nucl. Instrum. and Meth., A500, pp. 272-308, 2003. 1. Uher, F. Frojdh, 1. Jakubek, C. Kenney, Z. Kohout, V. Linhart, S. Parker, S. Petersson, S. Pospisil and G. Thungstrom, Characterization of 3D Thermal Neutron Semiconductor Detectors," Nucl. Instrum. and Meth., A576, pp. 32 - 37,2007. RJ. Nikolic, A.M. Conway, C.E. Reinhardt, R.T. Graff, T.F. Wang, N. Deo and C.L. Cheung, "6:1 Aspect Ratio Pillar Based Thermal Neutron
448
