|Fig. 1: Device design and fundamental operation for a planar conversion layer p-n junction thermal neutron detector.|
Since neutrons do not interact with matter through Coulomb forces, charge-based neutron detectors detect the presence of thermal neutrons (0.026 eV) through nuclear reaction products generated from neutron absorption events with nuclei of the detection media. When a thermal neutron collides with and is absorbed by a target nucleus, one or more primary nuclear reaction products such as alpha particles, gamma rays, nucleons, and heavy and light ions can be released.  As these high energy reaction products travel through the detection media, they can generate secondary reaction products by ionizing and exciting atoms they encounter along their path. Both primary and secondary reaction products can then contribute to an electrical signal which indicates the initial thermal neutron interaction.
Since neutron detection relies on the transduction of energy though nuclear reactions, a primary consideration for thermal neutron detector designs is how strongly the constituent materials interact with slow neutrons. The probability that a neutron scattering or absorption event will occur depends on the energy-dependent microscopic cross-section and the atomic density of the target.  3He, with its large thermal neutron cross-section (5337 barns) and insensitivity to background gamma radiation, is routinely used as a converter gas for high efficiency neutron detection.  However, the recent major increase in demand for border security devices capable of detecting the transport of fissile materials in cargo and the dwindling U.S. stockpile of 3He gas, have highlighted the crucial need for alternative neutron detector technologies. [3-5] Semiconductor-based neutron detectors are one such alternative.
Most conventional semiconductor materials have a very low probability of interacting with free neutrons. For example, the microscopic thermal neutron cross-section for naturally occurring silicon is ~ 2.24 barns, which means a thermal neutron would have to travel an average distance of ~ 8.6 cm (mean free path) before a scattering or capture event would occur.  The large mean free path for neutrons in Si makes neutron detection using only modern Si devices completely impractical. However, solid state semiconductor thermal neutron detectors are still highly desirable because when compared with gas-filled detectors, they have the potential to be more compact, operate at lower voltages, and be more robust against vibration induced noise.
The most straight forward way to overcome the low probability of neutron interaction with conventional semiconductor materials and improve detection efficiency is to integrate a layer of neutron reactive material into the semiconductor device architecture. These designs, commonly referred to as conversion layer devices, typically use Si or GaAs p-n junction or Schottky diodes to separate electron-hole pairs generated from interactions with primary nuclear reaction products. The device design and fundamental operation of a planar conversion layer p-n junction thermal neutron detector is shown in Fig. 1.
A typical planar thermal neutron conversion layer detector consists of a thin film of neutron reactive material deposited on the surface of a semiconductor diode. 10B and 6LiF are frequently used as conversion layer materials because of their stability, large thermal neutron cross-sections, and primary reaction products.  In the case of 10B, the most common primary reaction products generated by neutron absorption are alpha particles (1.47 MeV) and 7Li ions (840 keV).  When a nuclear reaction occurs in the 10B conversion layer, these charged particles will travel into the semiconductor material where they can create electron-hole (e-h) pairs. Since the detector is typically operated under reverse bias, electron-hole pairs generated inside of or within a diffusion length of the space charge region are separated by the externally applied electric field and collected at the contacts. To enhance the charge collection efficiency and response time of the planar conversion layer detector, a p-i-n diode can be used in place of the p-n junction.
A critical design parameter for planar conversion layer devices, which is ultimately responsible for limiting the maximum thermal neutron detection efficiency, is the thickness of the conversion layer material. The probability that a neutron will be absorbed in the conversion layer increases with increasing layer thickness. For 10B, ~ 90 % of incident neutron flux will be absorbed in a ~ 50 μm thick layer.  However, there is a trade off since the primary nuclear reaction products are released inside of the conversion layer and must travel a finite distance before they can escape into the semiconductor material. As the charged particles pass through the conversion layer, their intensity rapidly drops off as some of their energy is absorbed by or transferred to the conversion material.  If the conversion layer is too thick, the reaction products may be absorbed before reaching the semiconductor. It is also possible that if the layer is too thick, by the time reaction products reach the semiconductor they may not have enough energy for adequate electron-hole pair generation. Average travel ranges for the 1.47 MeV alpha particle and 840 keV 7Li ion in 10B are 3.6 μm and 1.6μm, respectively.  Conversion layer thicknesses for planar conversion layer devices are limited by the average travel ranges of the reaction products.
|Fig. 2: Device design and fundamental operation for a pillar structured thermal neutron detector.|
In addition to the efficiency limiting conversion layer thickness, the planar geometry of the device restricts the primary reaction products available to contribute to electron-hole pair generation to 50%. For most isotopes used for conversion layer materials, the two charged particles released from nuclear reactions are emitted at 180 degrees relative to one another which means that only one reaction product has the ability to reach the semiconductor and create electron-hole pairs. 
Despite major advances in semiconductor material growth and processing technologies over the last 60 years, when the first planar conversion layer thermal neutron detector was reported, the maximum detection efficiencies for semiconductors coated in 10B or 6LiF have not exceeded 5%.  When compared with the greater than 70% thermal neutron detection efficiency for 3He gas detectors, planar conversion layer semiconductor detectors are generally considered a non-viable approach to alternative thermal neutron detection. 
However, recent developments in innovative three-dimensional semiconductor detector designs using photolithography and deep reactive ion etching (DRIE) have renewed interest in semiconductor-based technologies for high efficiency thermal neutron detection.  Fig. 2 provides a typical cross-section of a pillar semiconductor detector design. By filling the regions between the semiconductor pillars with a neutron reactive material, these three-dimensional designs aim to overcome the geometric constraints that limit planar converter layer devices. By increasing the converter material thickness, these designs can achieve maximum neutron absorption and while simultaneously providing the reaction products with a reasonable pathway to escape into the semiconductor.
For three-dimensional detectors, both the diode and conversion material dimensions are important design parameters. Optimized Si pillar designs using 10B are theoretically predicted to reach over 75% thermal neutron detection efficiency.  Unfortunately, achieving high aspect ratio etching and conformal converter material deposition have proven to be challenges that have so far limited detector efficiencies to ~ 20%.  However, a significant research effort is currently being made to address these issues and move closer to realizing theoretical detection efficiencies. If these high detection efficiencies can be achieved, semiconductor-based thermal neutron detection devices may one day be a potentially viable alternative to 3He detectors.
© Sara E. Harrison. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.
 T.W. Crane and M. P. Baker, "Neutron Detectors," in Passive Nondestructive Assay of Nuclear materials, ed. by D. Reilly et al., Nuclear Regulatory Commission, NUREG/CR-5550, March 1991, p. 379.
 P. Rinard, "Neutron Interaction with Matter," in Passive Nondestructive Assay of Nuclear materials, ed. by D. Reilly et al., Nuclear Regulatory Commission, NUREG/CR-5550, March 1991, p. 357.
 R. T. Kouzes, "The 3He Supply Problem", Pacific Northwest National Laboratory, PNNL-18388, April 2009.
 D. Kramer, "DOE Begins Rationing Helium-3," Physics Today 63, No. 6, 22 (2010).
 D. Kramer, "For Some, Helium-3 Supply Picture Is Brightening," Physics Today 64, No. 5, 20 (2010).
 D. S. McGregor et al., "Designs for Thin-Film-Coated Semiconductor Thermal Neutron Detectors," Nuclear Science Symposium Conference Record, (IEEE 2001), p. 2454.
 J. K. Shultis and D. S. McGregor, "Efficiencies of Coated and Perforated Semiconductor Neutron Detectors," IEEE Trans. Nucl. Sci. 53, 1659 (2006).
 Q. Shao et al., "Gamma Discrimination in Pillar Structured Thermal Neutron Detectors," Lawrence Livermore National Laboratory, LLNL-PROC-543492, 30 Mar 12.
 R. J. Nikolic, et al., "Roadmap for High Efficiency Solid-State Neutron Detectors," Lawrence Livermore National Laboratory, UCRL-PROC-213583, 14 Jul 05.
 A. N. Caruso, "The Physics of Solid-State Neutron Detector Materials and Geometries," J. Phys. Condes. Matter 22, 443201 (2010).
 D. S. McGregor et al., "Design Considerations for Thin Film Coated Semiconductor Thermal Neutron Tetectors - I: Basics Regarding Alpha Particle Emitting Neutron Reactive Films", Nucl. Instrum. Methods Phys. Res. A 500, 272 (2003).
 L. F. Voss et al., "Development of 10B2O3 Processing For Use as a Neutron Conversion Material," J. Electron. Mater. 39, 263 (2010).