ABSTRACT Title of dissertation: NEUTRON DETECTION BY SCINTILLATION OF NOBLE-GAS EXCIMERS Jacob Collin McComb, Doctor of Philosophy, 2012 Dissertation directed by: Professor Mohamad al-Sheikhly Nuclear Engineering Program Neutron detection is a technique essential to homeland security, nuclear reactor in- strumentation, neutron di raction science, oil-well logging, particle physics and ra- diation safety. The current shortage of helium-3, the neutron absorber used in most gas- lled proportional counters, has created a strong incentive to develop alternate methods of neutron detection. Excimer-based neutron detection (END) provides an alternative with many attractive properties. Like proportional counters, END relies on the conversion of a neutron into energetic charged particles, through an exother- mic capture reaction with a neutron absorbing nucleus (10B, 6Li, 3He). As charged particles from these reactions lose energy in a surrounding gas, they cause electron excitation and ionization. Whereas most gas- lled detectors collect ionized charge to form a signal, END depends on the formation of diatomic noble-gas excimers (Ar 2, Kr 2, Xe 2). Upon decaying, excimers emit far-ultraviolet (FUV) photons, which may be collected by a photomultiplier tube or other photon detector. This phenomenon provides a means of neutron detection with a number of advantages over traditional methods. This thesis investigates excimer scintillation yield from the heavy noble gases fol- lowing the boron-neutron capture reaction in 10B thin- lm targets. Additionally, the thesis examines noble-gas excimer lifetimes with relationship to gas type and gas pressure. Experimental data were collected both at the National Institute of Standards and Technology (NIST) Center for Neutron Research, and on a newly developed neutron beamline at the Maryland University Training Reactor. The com- ponents of the experiment were calibrated at NIST and the University of Maryland, using FUV synchrotron radiation, neutron imaging, and foil activation techniques, among others. Computer modeling was employed to simulate charged-particle trans- port and excimer photon emission within the experimental apparatus. The observed excimer scintillation yields from the 10B(n; )7Li reaction are compa- rable to the yields of many liquid and solid neutron scintillators. Additionally, the observed slow triplet-state decay of neutron-capture-induced excimers may be used in a practical detector to discriminate neutron interactions from gamma-ray inter- actions. The results of these measurements and simulations will contribute to the development and optimization of a deployable neutron detector based on noble-gas excimer scintillation. NEUTRON DETECTION BY SCINTILLATION OF NOBLE-GAS EXCIMERS by Jacob Collin McComb Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park in partial ful llment of the requirements for the degree of Doctor of Philosophy 2012 Advisory Committee: Professor Mohamad al-Sheikhly, Chair/Advisor Professor Michael A. Coplan, Dean?s Representative Professor Charles W. Clark Professor Mohammad Modarres Professor Gary Pertmer ? Copyright by Jacob Collin McComb 2012 To my grandfathers, Robert Lewis Chapman and Dr. Robert Bradbury McComb. ii Acknowledgments The hard work, expertise, and patience of many contributed to this research. Their support has been fundamental to my success. I must rst thank my advisor, Professor Mohamad al-Sheikhly, for his mentorship during the past four years. His devotion to his students, myself included, is an inspiration. Secondly, I must thank Professor Michael Coplan for his unparalleled enthusiasm, and attention to detail. It has been a privilege to work with him throughout the course of this project. I am also obliged to the remainder of my thesis committee, Dr. Charles Clark, Professor Mohammad Modarres, and Professor Gary Pertmer, for taking time to review this document, and the work it entails. Thank you to the FUND group, of which I am honored to be a part. To Dr. Alan Thompson for his empathy, wit, and endless supply of ingenious solutions. To Dr. Robert Vest for many lessons on FUV detection, vacuum systems, and the tricks of building an apparatus. To Dr. Patrick Hughes for conferring to me a mountain of knowledge on the experiments leading up to my own. Thank you again to Dr. Clark, Professor Coplan, and Professor Al-Sheikhly for bringing me into the group and supporting this project. Thank you to my groupmates past and present, Layla Shahamat, Slavica Grdanovksa, Chanel Tissot, Cameron Goodwin, Ian Gi ord, Dr. Ali Mohamed, Amy Beasten, and Daniel Vega. To Vince Adams for teaching me how to sweat pipes, melt wax, and operate a nuclear reactor. To Dr. Timothy Koeth for breathing life back into the MUTR, and for his assistance with the PMT electronics. To the Radiation Safety O ce and the Reactor Safety Committee for enduring my experimental proposals. Thank you to everyone in the Neutron Interactions Group at the NCNR for their as- sistance, encouragement, and good humor. To Dr. Daniel Hussey, Dr. David Jacobsen, Dr. David Gilliam, Dr. Craig Heimbach, Dr. Je rey Nico, Dr. Pieter Mumm, and Dr. Muhammad Arif. Thank you to the entire SURF sta for being incredibly welcoming and helpful. To Dr. Uwe Arp, Alex Farrell, Dr. Edward Hagley, Dr. Thomas Lucatorto, Dr. Mitchell Furst, Dr. Ping Shaw, and Dr. Steve Grantham. Thank you to Dr. Gerard Henein at the CNST for teaching me the rst thing about boron EBPVD. Additionally, thank you to Dr. John Abrahams and Dr. Thomas Loughran at the UMD FabLab for all of their help in the clean room. Thank you to Dr. June Tveekrem at NASA Goddard Space Flight Center for her assistance with the photon collection e ciency modeling. Thank you to Dr. Lis Nanver at Delft University of Technology for her insight on boron lm fabrication, and for providing us with epitaxial substrates. Thank you to Dr. Karen Gaskell for her assistance with XPS, Rich Suchoski for his as- sistance with XRD, and Charlie Chen for his assistance with SEM. Thank you to Howie Grossenbacher and Edd Cole at the Aerospace and IPST Machine Shops for turning my hasty drawings and vague requests into perfectly functioning components. iv Thank you to everyone who kept the administrative aspects of my graduate life rolling. To Dr. Kathleen Hart, JoAnne Kagle, Mike McNicholas, Ginette Villeneuve, Olivia Noble, Kay Morris, and Martha Neviaser. Thank you to those who encouraged and nurtured my pursuit of higher education, in- cluding my undergraduate advisor, Professor Je rey Dunham, my mentors at ORNL, Dr. Marc Garland and Dr. Saed Mirzadeh, and all of my professors at UMD. Lastly, and most importantly, thank you to my friends and family. You have always been my biggest source of encouragement and inspiration. v Contents List of Tables ix List of Figures xi List of Abbreviations xiv 1 Introduction 1 1.1 Applications of Neutron Detection . . . . . . . . . . . . . . . . . . . . 1 1.2 Demand for Alternative Methods . . . . . . . . . . . . . . . . . . . . 4 1.3 Excimer-Based Neutron Detection . . . . . . . . . . . . . . . . . . . . 6 1.4 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2 Background 11 2.1 The Neutron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.2 Neutron Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.3 Neutron Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.3.1 Gas- lled neutron detectors . . . . . . . . . . . . . . . . . . . 19 2.3.2 Scintillation-based neutron detectors . . . . . . . . . . . . . . 22 2.3.3 Semiconductor neutron detectors . . . . . . . . . . . . . . . . 25 2.3.4 Other neutron detection methods . . . . . . . . . . . . . . . . 27 2.4 Boron-Neutron Capture . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.5 Charged-Particle Transport . . . . . . . . . . . . . . . . . . . . . . . 30 2.6 Noble-Gas Excimers . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 2.6.1 Excimer scintillation . . . . . . . . . . . . . . . . . . . . . . . 35 2.6.2 Noble-gas-based detectors . . . . . . . . . . . . . . . . . . . . 42 3 Review of Preliminary Experiments 43 3.1 The NIST Center for Neutron Research . . . . . . . . . . . . . . . . . 43 3.2 Lyman-Alpha Experiment at the NCNR . . . . . . . . . . . . . . . . 45 3.3 Excimer-Based Experiments at the NCNR . . . . . . . . . . . . . . . 46 3.3.1 Helium-3 and noble-gas mixtures . . . . . . . . . . . . . . . . 48 3.3.2 Preliminary thin- lm experiment . . . . . . . . . . . . . . . . 50 4 Excimer Scintillation Experiment at Maryland 53 4.1 The Maryland University Training Reactor . . . . . . . . . . . . . . . 54 4.2 Experimental Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.2.1 Scintillation cell . . . . . . . . . . . . . . . . . . . . . . . . . . 56 4.2.2 Gas-handling system . . . . . . . . . . . . . . . . . . . . . . . 60 4.2.3 Neutron beamline . . . . . . . . . . . . . . . . . . . . . . . . . 63 4.2.4 Electronics con guration . . . . . . . . . . . . . . . . . . . . . 68 4.2.5 Electronics calibration . . . . . . . . . . . . . . . . . . . . . . 72 4.3 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 5 Boron-10 Thin-Film Targets 77 vi 5.1 Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 5.2 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.2.1 Neutron imaging . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.2.2 Pro lometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 5.2.3 X-ray photoelectron spectroscopy . . . . . . . . . . . . . . . . 95 5.2.4 X-ray di raction . . . . . . . . . . . . . . . . . . . . . . . . . 102 5.2.5 Scanning electron microscopy . . . . . . . . . . . . . . . . . . 104 6 Photomultiplier Tube Calibration 109 6.1 The Synchrotron Ultraviolet Radiation Facility . . . . . . . . . . . . . 109 6.2 Calibration Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . 113 6.3 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 6.4 Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 7 Neutron Beam Characterization 131 7.1 Neutron Beam Collimator . . . . . . . . . . . . . . . . . . . . . . . . 131 7.2 Fluence Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . 137 7.3 Beam Pro le Measurement . . . . . . . . . . . . . . . . . . . . . . . . 151 7.4 Energy-Distribution Modeling . . . . . . . . . . . . . . . . . . . . . . 156 8 Modeling 163 8.1 Charged-Particle Transport . . . . . . . . . . . . . . . . . . . . . . . 163 8.2 Photon Emission and Re ection . . . . . . . . . . . . . . . . . . . . . 173 9 Results and Discussion 183 9.1 Excimer Scintillation Yield . . . . . . . . . . . . . . . . . . . . . . . . 183 9.2 Excimer Lifetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 9.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 9.4 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 A Scintillation from 241Am -particle Emission 211 B Construction of an FUV Re ectometer 213 Bibliography 217 vii List of Tables 1.1 SNM neutron emission rates . . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Characteristics of various thermal-neutron scintillators . . . . . . . . 25 2.2 Properties of noble gases . . . . . . . . . . . . . . . . . . . . . . . . . 35 2.3 Wh values of noble gases during -particle irradiation . . . . . . . . 42 4.1 Noble-gas purities and contaminant concentrations . . . . . . . . . . 61 5.1 Tool settings during the boron- lm depositions . . . . . . . . . . . . . 81 5.2 Areal densities of the 10B lms . . . . . . . . . . . . . . . . . . . . . . 88 5.3 Pro lometer scan parameters . . . . . . . . . . . . . . . . . . . . . . 94 5.4 Film thicknesses and roughnesses from pro lometry . . . . . . . . . . 94 5.5 Concentrations of surface contaminants in boron lms . . . . . . . . . 99 5.6 Concentrations of surface contaminants in boron source materials . . 99 6.1 Speci cations of the Hamamatsu R6835 photomultiplier tube . . . . . 113 6.2 Detector e ciencies for Ar, Kr, and Xe emissions . . . . . . . . . . . 126 7.1 Neutron count tting and uncertainties . . . . . . . . . . . . . . . . . 143 7.2 Neutron transmission fractions of beamline materials . . . . . . . . . 146 7.3 Average neutron absorption rates of each 10B target . . . . . . . . . . 146 7.4 Average total neutron uence in the MUTR beam . . . . . . . . . . . 159 8.1 Average charged-particle escape energies . . . . . . . . . . . . . . . . 166 8.2 Collection e ciencies determined by modeling . . . . . . . . . . . . . 180 9.1 Uncertainty budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188 9.2 Average values of Wh from the 10B(n; )7Li reaction . . . . . . . . . 199 viii List of Figures 1.1 Applications of neutron detection . . . . . . . . . . . . . . . . . . . . 4 1.2 Emission spectra of Ar, Kr, and Xe excimers . . . . . . . . . . . . . . 7 2.1 Neutron interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.2 Cross sections of various nuclei . . . . . . . . . . . . . . . . . . . . . 15 2.3 Gas multiplication curves and proportional tube detectors . . . . . . 20 2.4 Scintillation-based neutron detectors . . . . . . . . . . . . . . . . . . 23 2.5 Fluorescence, phosphorescence, and scintillation decay . . . . . . . . . 24 2.6 Semiconductor neutron detectors . . . . . . . . . . . . . . . . . . . . 26 2.7 Boron-10 neutron absorption cross section . . . . . . . . . . . . . . . 28 2.8 Boron-neutron capture reaction . . . . . . . . . . . . . . . . . . . . . 29 2.9 Ranges of charged particles in boron . . . . . . . . . . . . . . . . . . 33 2.10 Excimer emission spectra at 50 and 400 torr . . . . . . . . . . . . . . 37 2.11 Excimer potential curves . . . . . . . . . . . . . . . . . . . . . . . . . 38 2.12 Fast and slow decay components of noble-gas excimers . . . . . . . . 39 2.13 Literature values for excimer radiative decay times . . . . . . . . . . 40 2.14 Absolute excimer photon yields . . . . . . . . . . . . . . . . . . . . . 41 3.1 The NCNR guidehall . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.2 Layout of the NCNR . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.3 Lyman-alpha signal from 3He irradiation . . . . . . . . . . . . . . . . 47 3.4 PMT response and scintillation signals from behind lters . . . . . . 48 3.5 Excimer signal from 3He and noble-gas mixtures . . . . . . . . . . . . 49 4.1 The MUTR core . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.2 Elevation view of the MUTR thermal column . . . . . . . . . . . . . 55 4.3 Diagram of the scintillation cell . . . . . . . . . . . . . . . . . . . . . 58 4.4 The scintillation cell . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.5 Typical scan of residual gases . . . . . . . . . . . . . . . . . . . . . . 61 4.6 Gas-handling system layout . . . . . . . . . . . . . . . . . . . . . . . 62 4.7 Gas-handling system model . . . . . . . . . . . . . . . . . . . . . . . 62 4.8 Collimator insert for the MUTR thermal column . . . . . . . . . . . . 64 4.9 Scintillation experiment con guration (close view) . . . . . . . . . . . 66 4.10 Scintillation experiment con guration (wide view) . . . . . . . . . . . 66 4.11 The scintillation cell in the beamline . . . . . . . . . . . . . . . . . . 67 4.12 The scintillation experiment beside the MUTR containment . . . . . 67 4.13 Pulse traces from various components . . . . . . . . . . . . . . . . . . 69 4.14 Photon-counting electronics . . . . . . . . . . . . . . . . . . . . . . . 70 4.15 Visualization of TAC inputs and outputs . . . . . . . . . . . . . . . . 71 4.16 Pulse-timing electronics . . . . . . . . . . . . . . . . . . . . . . . . . 71 4.17 Fission-chamber beam-monitor electronics . . . . . . . . . . . . . . . 72 4.18 Oscilloscope traces from scintillation . . . . . . . . . . . . . . . . . . 73 5.1 Electron-beam evaporator . . . . . . . . . . . . . . . . . . . . . . . . 79 x 5.2 Source material and 10B lms . . . . . . . . . . . . . . . . . . . . . . 82 5.3 Delamination of 2.5 m lms . . . . . . . . . . . . . . . . . . . . . . 83 5.4 Plan view of the NCNR neutron imaging facility . . . . . . . . . . . . 85 5.5 Diagram of the NIF detector station . . . . . . . . . . . . . . . . . . 86 5.6 Boron-10 lm and holder used for neutron imaging . . . . . . . . . . 89 5.7 Thermal-neutron image of a 300-nm 10B lm . . . . . . . . . . . . . . 89 5.8 Thermal-neutron image of a 600-nm 10B lm . . . . . . . . . . . . . . 90 5.9 Thermal-neutron image of a 900-nm 10B lm . . . . . . . . . . . . . . 90 5.10 Thermal-neutron image of a 1200-nm 10B lm . . . . . . . . . . . . . 91 5.11 Thermal-neutron image of silicon substrate . . . . . . . . . . . . . . . 91 5.12 Neutron-wavelength distribution of the NIF beam . . . . . . . . . . . 92 5.13 Pro lometer scans of 10B thin- lms . . . . . . . . . . . . . . . . . . . 94 5.14 X-ray photoelectron spectrometry . . . . . . . . . . . . . . . . . . . . 96 5.15 Typical electron binding energies of boron compounds . . . . . . . . . 98 5.16 EBE distribution for a boron lm stored in atmosphere . . . . . . . . 100 5.17 EBE distribution for a boron lm stored in N2 glovebox . . . . . . . . 100 5.18 Boron (1s) electron peaks from XPS . . . . . . . . . . . . . . . . . . . 101 5.19 Contaminant concentrations during sputtering . . . . . . . . . . . . . 101 5.20 Bragg scattering from a crystal lattice . . . . . . . . . . . . . . . . . 103 5.21 X-ray di raction spectrum from a boron lm . . . . . . . . . . . . . . 103 5.22 Diagram of a scanning electron microscope . . . . . . . . . . . . . . . 105 5.23 SEM images of a 600-nm boron lm . . . . . . . . . . . . . . . . . . . 106 5.24 SEM images of a 900-nm boron lm . . . . . . . . . . . . . . . . . . . 107 6.1 Synchrotron radiation . . . . . . . . . . . . . . . . . . . . . . . . . . 110 6.2 SURF III radiant power . . . . . . . . . . . . . . . . . . . . . . . . . 110 6.3 The Synchrotron Ultraviolet Radiation Facility at NIST . . . . . . . . 112 6.4 Diagram of BL-4 at SURF . . . . . . . . . . . . . . . . . . . . . . . . 112 6.5 The PMT calibration xture . . . . . . . . . . . . . . . . . . . . . . . 114 6.6 Schematic of an np-type VUV photodiode . . . . . . . . . . . . . . . 115 6.7 PMT counts during the SURF calibration . . . . . . . . . . . . . . . 119 6.8 Statistical uncertainties in PMT counts . . . . . . . . . . . . . . . . . 119 6.9 Responsivity of the silicon photodiode . . . . . . . . . . . . . . . . . 122 6.10 Photon ux coe cients of BL-4 . . . . . . . . . . . . . . . . . . . . . 123 6.11 Relative uncertainties in k( ). . . . . . . . . . . . . . . . . . . . . . . 123 6.12 Intrinsic e ciency of the PMT detector package . . . . . . . . . . . . 124 6.13 Spectral response of the PMT supplied by the manufacturer . . . . . 125 6.14 Relative uncertainties in pmt( ). . . . . . . . . . . . . . . . . . . . . 125 6.15 Spatial response of PMT detector package (3D) . . . . . . . . . . . . 128 6.16 Spatial response of PMT detector package (2D) . . . . . . . . . . . . 128 6.17 PMT response as a function of operating voltage . . . . . . . . . . . . 129 7.1 Collimator insert design . . . . . . . . . . . . . . . . . . . . . . . . . 132 7.2 Collimator insert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 7.3 Collimator insert frame . . . . . . . . . . . . . . . . . . . . . . . . . . 134 xi 7.4 Sample of the collimator ll mixture . . . . . . . . . . . . . . . . . . 135 7.5 Collimator insert lled with shielding mixture . . . . . . . . . . . . . 135 7.6 Collimator shutter assembly . . . . . . . . . . . . . . . . . . . . . . . 136 7.7 The ssion-chamber beam monitor . . . . . . . . . . . . . . . . . . . 137 7.8 Pulse-height distributions from the ssion chamber . . . . . . . . . . 147 7.9 Neutron counts on 3.09.12 . . . . . . . . . . . . . . . . . . . . . . . . 148 7.10 Neutron counts on 3.14.12 . . . . . . . . . . . . . . . . . . . . . . . . 148 7.11 Neutron counts on 3.16.12 . . . . . . . . . . . . . . . . . . . . . . . . 149 7.12 Neutron counts on 3.23.12 . . . . . . . . . . . . . . . . . . . . . . . . 149 7.13 Thermal cross sections of 10B and 235U . . . . . . . . . . . . . . . . . 150 7.14 Value of over the thermal energy region . . . . . . . . . . . . . . . . 150 7.15 Beam pro le immediately behind the aperture . . . . . . . . . . . . . 154 7.16 Beam pro le 12 cm behind the aperture . . . . . . . . . . . . . . . . 155 7.17 Geometry of the MUTR MCNP model . . . . . . . . . . . . . . . . . 158 7.18 Neutron energy distributions in the MUTR thermal column . . . . . 160 7.19 Neutron ux ratios across the MUTR thermal column . . . . . . . . . 160 8.1 TRIM particle trajectory plots . . . . . . . . . . . . . . . . . . . . . . 165 8.2 Charged-particle escape energy as a function of lm thickness . . . . 166 8.3 Charged-particle range and stopping time . . . . . . . . . . . . . . . 169 8.4 Ionization density in argon . . . . . . . . . . . . . . . . . . . . . . . . 170 8.5 Ionization density in krypton . . . . . . . . . . . . . . . . . . . . . . . 171 8.6 Ionization density in xenon . . . . . . . . . . . . . . . . . . . . . . . . 172 8.7 Photon-emission model geometry . . . . . . . . . . . . . . . . . . . . 174 8.8 Scoring regions and the spatial response of PMT . . . . . . . . . . . . 176 8.9 Boron re ectivity over the FUV region . . . . . . . . . . . . . . . . . 177 8.10 Copper-black BRDF data at 121.6 nm . . . . . . . . . . . . . . . . . 178 8.11 Collection e ciency as a function of %B and h . . . . . . . . . . . . . 181 8.12 Collection e ciency as a function of %B and %Cu . . . . . . . . . . . . 181 9.1 Counts from scintillation of argon . . . . . . . . . . . . . . . . . . . . 189 9.2 Counts from scintillation of krypton . . . . . . . . . . . . . . . . . . . 190 9.3 Counts from scintillation of xenon . . . . . . . . . . . . . . . . . . . . 191 9.4 Relative uncertainties in nh . . . . . . . . . . . . . . . . . . . . . . . 192 9.5 Contributions to nh total . . . . . . . . . . . . . . . . . . . . . . . . . 193 9.6 Pulse-charge distributions from excimer scintillation . . . . . . . . . . 194 9.7 Scintillation yield from 10B(n; )7Li in argon . . . . . . . . . . . . . . 195 9.8 Scintillation yield from 10B(n; )7Li in krypton . . . . . . . . . . . . . 196 9.9 Scintillation yield from 10B(n; )7Li in xenon . . . . . . . . . . . . . . 197 9.10 Wh values as a function of noble-gas type and pressure . . . . . . . . 199 9.11 Distributions of the times between pulses during Ar scintillation . . . 202 9.12 Distributions of the times between pulses during Kr scintillation . . . 203 9.13 Distributions of the times between pulses during Xe scintillation . . . 203 9.14 Triplet-state lifetimes of Ar, Kr, and Xe excimers . . . . . . . . . . . 204 9.15 Relative uncertainties in . . . . . . . . . . . . . . . . . . . . . . . . 204 xii A.1 Noble-gas scintillation signal from 241Am -particle emitter . . . . . . 212 B.1 Diagram of the FUV re ectometer . . . . . . . . . . . . . . . . . . . . 215 B.2 The FUV re ectometer . . . . . . . . . . . . . . . . . . . . . . . . . . 215 B.3 Head-on view of the FUV re ectometer . . . . . . . . . . . . . . . . . 216 B.4 The re ectometer prepared for installation at BL-4 . . . . . . . . . . 216 xiii List of Abbreviations 10B(n; )7Li Boron-neutron capture reaction BL-4 Beamline 4 boPET biaxially-oriented Polyethylene Terephthalate BRDF Bidirectional-Re ectance Distribution Function BT2 NCNR Neutron Imaging Facility C/T Counter/Timer CCD Charge-Coupled Device CNST Center for Nanoscale Science and Technology CVD Chemical Vapor Deposition DHS Department of Homeland Security DoD Department of Defense DOE Department of Energy EBPVD Electron-Beam Physical Vapor Deposition END Excimer-based Neutron Detection FUV Far Ultraviolet GUI Graphical User Interface LAND Lyman-Alpha-based Neutron Detection LET Linear Energy Transfer MCA Multichannel Analyzer MCNP Monte Carlo N-Particle transport code MRI Magnetic Resonance Imaging MRSEC Materials Research Science and Engineering Center MUTR Maryland University Training Reactor NBSR National Bureau of Standards Reactor NCNR NIST Center for Neutron Research NG-6 Neutron Guide 6 NGBD Noble-Gas-Based Detector NIF Neutron Imaging Facility NISP Nanoscale Imaging Spectroscopy Laboratory NIST National Institute of Standards and Technology NNSA National Nuclear Security Administration PCD Pulse-Charge Distribution PHD Pulse-Height Distribution PMT Photomultiplier Tube PSF Point Spread Function PSL Photostimulated Luminescence PVD Physical Vapor Deposition RF Radio Frequency RGA Residual Gas Analyzer RMS Root Mean Square RPM Radiation Portal Monitor RTV Room Temperature Vulcanizer xiv SEM Scanning Electron Microscope SF Spontaneous Fission SNM Special Nuclear Material SRIM Stopping and Range of Ions in Matter SURF Synchrotron Ultraviolet Radiation Facility TAC Time-to-Amplitude Converter TRIGA Training, Research, Isotopes, General Atomics TRIM Transport of Ions in Matter UMD University of Maryland XPS X-Ray Photoelectron Spectroscopy XRD X-Ray Di raction xv Chapter 1 Introduction The goal of this research is to provide a fundamental understanding of noble-gas excimer scintillation as a means of neutron detection. This introductory chapter reviews various applications of neutron detection and the demand for alternative detection methods, including a discussion of the helium-3 shortage and the disad- vantages of current detectors. Additionally, a brief introduction to excimer-based neutron detection accompanies an overview of the research project. Further de- tails on current neutron detection methods, and other background relevant to the research, are presented in the following chapter. 1.1 Applications of Neutron Detection Neutron detection is an essential tool in a wide range of scienti c and industrial applications. It provides both a means of analyzing nuclear reactions and inter- rogating material properties. Starting with Fermi?s Chicago Pile, in 1942, neutron detectors have been used in every ssion reactor core to monitor reactor power levels by measuring neutron ux. These detectors, individually or in arrays, must often cover a ux range of 10 orders of magnitude, and provide spatial ux pro les, as well 1 as signals for safety trip channels [1, 2, 3, 4]. Today, nuclear fusion research relies on neutron yield and energy distribution measurements to diagnose fusion plasma behavior. These measurements, obtained with neutron detectors, enable estimates of fusion plasma ion velocity, ion density, and reaction rates [5, 6, 7, 8, 9]. A variety of disciplines use neutron scattering techniques to investigate material structure, composition, and dynamics, including biological membrane behavior, magnetism, atomic motion, and material stresses. These methods require large detectors, with small spatial resolutions, to identify where neutrons impinge on a detector?s surface. By creating images or patterns, indicative of how neutrons are scattered, absorbed, or transmitted, a great deal may be inferred about the nature of a sample [10, 11, 12, 13, 14]. Neutrino research, dark matter detection, and nuclear physics also rely on neutron detection as a scienti c tool [15, 16, 17]. For example, neutrons may interact in large cosmic-ray detection experiments and generate trou- blesome backgrounds. Neutron detectors are frequently used to identify and veto these signals [18]. Neutron detection plays a role in a number of industrial applications. Oil- and gas- well prospectors use neutron detectors for pulsed-neutron well logging. By attaching neutron generators and multiple neutron detectors to wireline logging packages, it is possible to determine the porosity of the strata surrounding a borehole. In non-shale formations, neutron porosity is indicative of hydrogen content due to hydrogen?s e cacy as a neutron moderator. This enables such methods to identify the location of gas and oil reserves [19, 20, 21]. 2 Most recently, neutron detection has become a critical instrument of the Department of Homeland Security (DHS). Special nuclear material (SNM), which includes the ssile isotopes uranium-233, uranium-235 and plutonium-239, and other isotopes associated with such material (238U, 240Pu), undergo spontaneous ssion (SF) with non-zero probabilities, providing distinct and detectable neutron signatures (see Table 1.1). This material could potentially be used in a nuclear weapon or dirty bomb, so the DHS has an interest in tracking its exchange and movement. In the U.S., more than 1,400 radiation detecting portal monitors (RPMs), stationed at ports and border crossings, contain neutron detectors to screen cargo and vehicles for SNM. Abroad, 2,000 U.S.-deployed RPMs are used for the same purpose. The DHS, the Department of Energy (DOE) and the Department of Defense (DoD) plan to deploy an additional 3,500 RPMs within the next 6 years [22]. However, due to the helium-3 shortage, these e orts, along with the neutron detection requirements of all other scienti c and industrial endeavors, will rely on alternative neutron detection methods to meet their needs [23, 24, 25, 26, 27]. Nuclide Half-life (years) Branching ratio Neutron emission rate (g 1s 1) 235U 7:04 108 7:0 10 11 1.86 1:0 10 5 238U 4:47 109 5:4 10 7 2.07 0:0136 239Pu 2:41 104 4:4 10 12 2.16 0:0220 240Pu 6569 5:0 10 8 2.21 920 252Cf 2.638 3:09 10 2 3.73 2:3 1012 Table 1.1: Neutron emission rates of various actinides found in SNM [28]. = the average number of neutrons per spontaneous ssion. 3 Figure 1.1: Left: Cherenkov radiation illuminating the MUTR reactor core, where power is monitored with various neutron detectors. Center: Neutron radiograph of a camera taken with a spatially sensitive neutron detector [29]. Right: A radiation portal monitor scanning cargo for SNM [30]. 1.2 Demand for Alternative Methods For the past 30 years, helium-3 (3He) proportional counters have been considered the gold standard in thermal neutron detection [22]. However, today, 3He is in short supply and readily available alternatives do not o er the stability, sensitivity, and discrimination capability of 3He detectors [31]. While 3He is a stable isotope, its abundance is very low, at 1.34 ppm of all natu- rally occurring helium [32]. Enriched 3He is obtained predominantly through the radioactive decay of tritium, 3H! 3He + T 1 2 = 12:3years (1.1) and is made available as a byproduct of tritium stockpiles, maintained by the Na- tional Nuclear Security Administration (NNSA) for nuclear weapons refurbishment. Due to decommissioning e orts, the US halted tritium production in 1988. The sup- ply of 3He has, therefore, signi cantly diminished and continues to decline [22]. 4 In 2010, the demand for 3He was greater than 76,000 L, with annual domestic production at 8000 L, and total domestic supply at 48,000 L [33]. The detector- related 3He requirements of the DOE, the DHS, and the DoD over the next 5 years is estimated at 100 kliter [34]. A separate estimate by GE Reuter Stokes projects the total demand, including safeguards, non-governmental, and scienti c needs at 65 kliter per year, while total supply is less than 20 kliter per year [31]. Along with neutron detection, 3He has other uses including MRI lung imaging, cryogenics, and fusion fuel, further diminishing its availability for use in detectors. In the absence of 3He, currently available neutron detection alternatives include boron-lined proportional counters, boron-tri uoride proportional counters, ssion chambers, lithium-6 scintillators, and semiconductor detectors, among others. Many of these detector types are discussed in the following chapter. The disadvantages of these detectors, such as high cost, toxicity, poor discrimination, poor detection e ciency, and poor durability, have prevented them from replacing 3He [35]. Boron- tri uoride (BF3) proportional counters, for example, have been in use as long as 3He. Yet, BF3 is a toxic and corrosive gas, and, therefore, undesirable as a practical detection medium. In light of the 3He shortage, and the shortcomings of other neutron detectors, this thesis investigates the phenomenon of noble-gas excimer scintillation, which provides the basis for an alternative means of neutron detection. 5 1.3 Excimer-Based Neutron Detection Excimer-based neutron detection (END) relies on the same conversion mechanism as most traditional, thermal-neutron detectors. First, a neutron is absorbed by a nucleus with a large neutron absorption cross section, in the form of a solid thin lm or a gas. This results in an exothermic reaction and the emission of energetic charged-particle reaction products. Due to their mass and charge, these particles are subject to high rates of linear energy transfer (LET), and, therefore, readily deposit their energy within a compact detector volume. Examples of these reactions include, 3He + n! 3H + p+ 0:764 MeV ( 0 = 5330 b) (1.2) 6Li + n! 3H + + 4:78 MeV ( 0 = 940 b) (1.3) 10B + n! ( 7Li + + 2:792 MeV 7Li + + 2:310 MeV ( 0 = 3840 b) (1.4) Unlike proportional-tube detectors, END does not depend on the collection of free charge to form an electrical pulse { a result of gas ionization by charged-particles accelerating across an electric potential. Rather, as a scintillation technique, END implements an optical signal as the characteristic neutron signature. By surrounding or mixing the absorbing target medium with a noble gas, the charged-particle reac- tion products induce the formation of noble-gas excimers as they dissipate kinetic energy. 6 Noble-gas excimers are loosely-bound, diatomic molecules that exist only in an ex- cited electronic state. In the ground state, noble-gas atoms are repulsive. There- fore, excimers are short-lived and quickly decay by emitting far-ultraviolet (FUV) photons, with wavelengths between 120-180 nm. The emission spectra of argon, krypton, and xenon excimers appear in Figure 1.2. Figure 1.2: Emission spectra of Ar, Kr, and Xe excimers, produced by thyratron modulator excitation. Redrawn from [36]. Peak intensities have been normalized to 1. This mechanism of FUV photon emission, known as excimer scintillation, provides a unique means of detecting neutrons. A single neutron absorption may liberate MeV of energy due to the exothermic nature of reactions like those shown in Eq. 1.2 - 1.4. Since excimer formation requires 20 - 70 eV [37, 38], depending on experimental conditions, a complete conversion of energy from neutron capture could theoretically produce 105 excimer photons. In this way, exothermic neutron absorption under a noble gas background generates a burst of scintillation light, which constitutes a 7 neutron detection signature. Preliminary experiments have previously demonstrated that this transfer of energy is perhaps greater than 30% e cient [39]. E ectively, END is a combination of processes from two conventional neutron de- tection methods { (1) gaseous excitation and ionization followed by (2) scintillation and light capture. Nonetheless, a neutron detector based on excimer scintillation has many potential advantages over traditional detectors: highly e cient signal, fast response time, wide dynamic range, exible geometry, physically robust, low voltage, low pressure, and helium-3 free. 1.4 Research Objectives This research project investigates noble-gas excimer scintillation following the boron- neutron capture reaction in boron-10 thin lms. Speci cally, experiments were per- formed to determine (1) the excimer scintillation yield per neutron absorption for various noble-gas types (Ar, Kr, Xe), gas pressures (50 - 800 torr), and lm thick- nesses (0.3 - 1.2 m), and (2) the slow component of the excimer lifetime for various gas types and gas pressures. In addition, computer modeling of charged-particle transport was implemented to simulate energy transfer and photon emission under the same set of experimental conditions. Based on these experimental results and computer simulations, a number of conclusions are drawn regarding the fundamen- tal nature of excimer scintillation from the 10B(n, )7Li reaction and the potential to employ this phenomenon in a deployable neutron detector. 8 Preliminary research for this project took place at the National Institute of Stan- dards and Technology (NIST) Center for Neutron Research (NCNR) in Gaithers- burg, MD. However, due to the construction of a new guide hall at the NCNR, starting in April 2011, research was transferred to the Maryland University Train- ing Reactor (MUTR), in College Park. Many other facilities, including the Syn- chrotron Ultraviolet Radiation Facility (SURF), the Center for Nanoscale Science and Technology (CNST), and the Neutron Imaging Facility (NIF) at NIST, as well as the Maryland NanoCenter, and the Maryland Materials Research Science and Engineering Center (MRSEC), were also essential for the study. This thesis is broken into chapters, corresponding to the various elements of the project. Background on neutron detection, the boron-neutron capture reaction, charged-particle transport, and noble-gas excimers is presented in Chapter 2. The preliminary experiments at the NCNR are discussed in Chapter 3. The excimer scintillation experiments at the MUTR are discussed in Chapter 4. Fabrication and characterization of the boron-10 thin- lm targets are discussed in Chapter 5. The photomultiplier tube calibration is discussed in Chapter 6. The development and characterization of a thermal-neutron beam at the MUTR are discussed in Chapter 7. Computer modeling of charged-particle transport and excimer photon collection is discussed in Chapter 8. The results of the experiments and their analysis are discussed in Chapter 9. Appendices on -particle scintillation measurements, and the construction of an FUV re ectometer are also provided. 9 Chapter 2 Background The excimer scintillation experiments were based on knowledge from a number of topics which have been previously studied in depth. This chapter provides a brief background on some of these topics including a discussion of current neutron detec- tion techniques, the boron-neutron capture reaction, charged-particle transport and noble-gas excimers. 2.1 The Neutron The neutron is one of two constituents of all atomic nuclei, known as nucleons. Ernest Rutherford rst conceived the neutron in 1920, as a form of nuclear glue, to account for the stability of the atomic nucleus, despite the electromagnetic repulsion of positively-charged protons. James Chadwick then demonstrated the existence of the neutron in 1932, after observing the ejection of energetic protons from para n, when bombarding nearby beryllium with alpha particles [40]. Noting similar obser- vations in previous experiments performed by Bothe, Curie, and Joliot, Chadwick posited that the phenomenon was the result of massive, neutral particles, radiating from the (Be+ ) interaction [41]. Following this discovery, Fermi?s Chicago Pile, 11 and the Manhattan Project asserted the the neutron?s profound impact and abrupt rise to the scienti c forefront. Today, the understanding of the neutron has greatly evolved since Chadwick?s dis- covery. The neutron is no longer considered an elementary particle; rather, it is known to be a hadron composed of three quarks (juudi) and binding, strong force gauge bosons, called gluons [42]. Although the neutron has a neutral charge, it has a non-zero magnetic dipole moment as a result of this internal structure [43]. Addi- tionally, while bound neutrons in stable nuclei are stable, free neutrons decay with a mean lifetime of 881:5 1:5 s [44]. Some properties of the neutron are still under investigation, such as the possibility of a non-zero neutron electric dipole moment [45], and others, like the neutron lifetime, are being re ned to higher precision. 2.2 Neutron Interactions Fundamentally, radiation detection relies on (1) the interaction of radiation within a detection medium, (2) the formation of free charge following an interaction, and (3) the collection of charge to produce an electrical signal [1]. While ionizing ra- diation may be observed through the direct creation and collection of free charge, the neutron is uncharged and non-ionizing, and therefore more elusive. Neutron detectors, thus, rely on the conversion of neutrons into secondary radiations, often in the form of heavy charged particles. These conversions occur most frequently as one of two types of neutron interaction: elastic scattering (n; n), or exothermic ab- 12 Figure 2.1: Two types of neutron interactions used for neutron detection. Elastic scatter- ing (left) is most e ective in detection of energetic neutrons, while exothermic absorption (right) is better suited to convert thermal neutrons. sorption (n; ), (n; p), (n; ssion). A diagram of these two interaction types appears in Figure 2.1. Other interactions, like radiative capture (n; ) are used in passive neutron detection, such as foil activation, yet are not practical in active detectors, due to the di culties of discriminating neutron signals from background. Inelas- tic scattering (n; n0) may also contribute to interactions within detectors, but cross sections for this interaction type are typically at least an order of magnitude below elastic scattering cross sections [46]. A neutron-interaction cross section is a measure of the neutron?s probability of interaction with a speci c nucleus. Di erent nuclei have vastly di erent interaction cross-sections, according to their nuclear potential, as shown in Figure 2.2. Nuclei with large cross-sections are ideal for use as neutron detection media. Additionally, interaction cross sections depend on incident neutron energy. Cross sections for neutrons at thermal equilibrium with ambient temperatures (E = 0:025 eV) are often more than 3 orders of magnitude higher than those for fast neutrons (E = 1 MeV). 13 Neutron kinetic energy (E) is related to neutron velocity (v) by the relationship, E = 1 2 mv2 (2.1) where m is the mass of the neutron (939.565 MeV/c2). A large number of nuclei are classi ed as 1=v absorbers because their neutron-absorption cross sections are inversely proportional to incident neutron velocity at thermal and, sometimes, ep- ithermal energies [2]. Due to this relationship, ssion neutrons, with an average energy of 1 MeV, are more di cult to detect than thermal neutrons. In many appli- cations, the physical and practical challenges of fast-neutron detection have spurred more interest in the development of detectors based on thermal-neutron capture [23]. Neutron moderators, such as polyethylene, are often incorporated into neutron detector packages to slow high energy neutrons, and increase their probability of interaction, particularly when sources of interest include ssile material. Interaction cross sections may be used to determine the neutron interaction rate within a target or detector. For a monodirectional neutron beam, this relationship takes the form of the Beer-Lambert law of exponential attenuation, I = I0 e N x (2.2) where I is the intensity of the transmitted beam, I0 is the intensity of the incident beam, N is the atom density of the target, is the microscopic cross section of the 14 Figure 2.2: Energy-dependent neutron interaction cross sections of various nuclei over cold and thermal energy regions [47]. 15 target, and x is the thickness of the target. The interaction rate is then determined by the equation, F = I0(1 e N x) (2.3) Following an interaction, a neutron indirectly creates ionizing radiation through a transfer of energy { either from a fast neutron to a recoiling nucleus, or from nuclear binding energy to charged-particle reaction products. This secondary radiation may be detected in a proportional tube, scintillator, semiconductor, or other traditional detector con guration. In the case of exothermic absorption, written as X(n; b)Y, the reaction Q-value, which mostly appears as the total kinetic energy of the reaction products, may be calculated using the equation, Q = (mX +mn mY mb)c 2 (2.4) where mn and mX are the masses of the neutron and the neutron-absorbing nucleus, mb and mY are the masses of the reaction products, and c is the speed of light (931.5 MeV/u). In the case of elastic scattering, the energy imparted to the recoiling nucleus is given by, EA = E + E 0 2 p EE 0 cos( ) A (2.5) where E is the energy of the incident neutron, E 0 is the energy of the scattered neutron, is the scattering angle of the scattered neutron, and A is the atomic mass number of the scattering nucleus. 16 Ideally, the amount of energy deposited within a detector by each neutron inter- action should be large in comparison to other types of radiation, and, therefore, distinguishable from background events. Thus, nuclei with large neutron-reaction Q-values, or large maximum recoil energies, make the best neutron detection media. In recoil-type detectors, for example, the maximum energy transferred to a nucleus in an elastic scattering collision is inversely proportional to that nucleus? mass. From Eq. 2.5, this is given by the equation, E 0min = A 1 A+ 1 2 E (2.6) where E is the energy of the incident neutron, E 0 is the energy of the scattered neutron, and A is the atomic mass number of the scattering nucleus. Accordingly, low atomic mass materials make optimal detection media for fast neutron detec- tors. Both the probability of interaction, and the average energy deposition per inter- action, must be taken into account when designing an e ective neutron detector. Other characteristics, such as background sensitivity, cost, and durability, must also be considered. Some of the most common detector con gurations and their proper- ties are discussed in the following section. 17 2.3 Neutron Detection Neutron detectors come in several varieties. The most popular are based on gas- lled proportional tubes, scintillators, and semiconductors. Detectors are often com- pared on the basis of application speci c requirements, which may include: intrinsic detection e ciency, collection e ciency, non-neutron event discrimination, energy resolution, spatial resolution, signal response times, and physical integrity. Three de nitions will be useful in this discussion of detectors, and the subsequent interpretation of the results from the excimer scintillation experiment. Intrinsic detector e ciency (or quantum e ciency) is de ned as the probability of detect- ing a given quanta of radiation incident on a detector, and is represented by the equation, int = number of pulses recorded number of radiation quanta incident on detector (2.7) Collection e ciency (or absolute e ciency) considers both the intrinsic e ciency of a detector and its geometrical con guration with respect to a radiation source, and is given by the equation, abs = number of pulses recorded number of radiation quanta emitted by source (2.8) 18 Gamma ray discrimination (or rejection e ciency) is a detector?s ability to positively identify a neutron among various types of radiation interactions, and is represented by the equation, reject = number of false neutron counts number of gamma rays incident on detector (2.9) 2.3.1 Gas- lled neutron detectors Gas- lled neutron detectors include 3He and BF3 proportional counters, helium (4He) and methane (CH4) recoil detectors, 10B-lined proportional counters and s- sion chambers. In this detector con guration, a neutron interacts with a gas or a solid coating, contained within an aluminum or stainless steel cathode enclosure. The neutron transfers kinetic energy to a recoil nucleus (fast neutron), or initiates a reaction with energetic charged-particle reaction products (thermal neutron). This energy transfer initiates the ionization of a gas detection medium. Typically, about 30 eV of energy is required to form an ion pair [1]. An electric eld applied across the gas volume accelerates ionized electrons towards a central anode wire. The voltage di erence between the anode wire and the cath- ode enclosure determines the gas multiplication characteristics of the detector and, consequently, the magnitude of the electrical output signal. Figure 2.3 shows the various gas multiplication regions with respect to the applied voltage of a gas- lled detector. In the proportional region, primary electrons in gas- lled detectors gain 19 enough energy to create 103 to 105 secondary and tertiary electrons (Townsend avalanche). When this charge is collected on the anode wire at the center of the tube, an electrical pulse is generated. Typically, proportional tube neutron detec- tors are operated in the 1-3 kV range [48]. A diagram of this detector con guration appears in Figure 2.3. Figure 2.3: Left: Multiplication curves for gas ionization detectors showing the various regions of gas multiplication [49]. The abscissa values depend on gas type and detector characteristics. Upper right: Diagram of a gas lled proportional counter. Lower right: Photograph of a helium-3 proportional tube [4]. The mean free path of electrons in a proportional counter is 1-2 m. While secondary electrons are collected within 10 8 s, the drift velocity of the positive ions is three orders of magnitude lower. Response and recovery times of gaseous detectors are limited to a few microseconds due to this slow collection of positive charge [48]. Pressures of gas- lled detectors range between 1 and 20 atm. However, pressures in the higher end of this range can lead to charge recombination in some gases and, 20 therefore, impose limitations. Heavy gases, such as Ar or CO2, may be added to gas- lled detectors to reduce reaction-product ranges, enhance charge collection, and decrease charge collection time. Polyatomic gases may be added as quench gas to shorten the avalanche process, improve pulse height resolution, and prevent gamma-ray pile up. Non-neutron background events, such as Compton scattering of gamma rays, can plague these detectors with false positives. However, gamma rays have much lower LET than charged particles, especially in gaseous media, and give rise to lower amplitude output signals. To some extent, backgrounds may be dealt with through pulse-height discrimination. Typical portal-monitor 3He proportional counters have intrinsic e ciencies of 70- 80%, and gamma-ray rejection e ciencies of 10 5 [23]. Boron-tri uoride tubes are limited to pressures of 1 atm due to high-voltage biasing, and achieve intrinsic e ciencies of 20% that of 3He tubes [50]. Conversion-layer detectors, or lined detectors, have intrinsic e ciencies of 1-5% relative to 3He-tube e ciencies, but may be bundled to increase the collection e ciency of a single detector package [51]. In 10B and ssile-deposit lined detectors, the short ranges of charged-particle reaction products limit absorber layer thicknesses to 2 m. These wall-layer detectors use standard detector gases, such as P-10, which are chosen for ideal charge collection properties. Lithium is not widely used as a thermal neutron absorber in gas- lled detectors due to its chemical reactivity. Fast-neutron, gas- lled recoil detectors, using natural helium or methane, have intrinsic e ciencies of 1% for neutrons between 20 keV and 20 MeV, and often have faster response times (10-100 ns) than 21 thermal neutron tubes [48]. Modern approaches to proportional neutron counters include boron-lined straw tube detectors, and multi-wire proportional chambers with anode wire grids [52, 53]. 2.3.2 Scintillation-based neutron detectors Scintillation neutron detectors consist of a scintillation medium coupled to a photo- multiplier tube, CCD, photodiode, or other photon detector. In this con guration, a fast neutron elastically scatters in a scintillation medium, transferring some of its ki- netic energy to a recoil nucleus. Alternatively, a thermal neutron may be absorbed by a nucleus with a large absorption cross section, embedded within or coupled to the scintillation medium. An exothermic reaction then emits charged particles, which scatter and transfer energy to a recoil nucleus or a phosphor dopant. The transfer of energy, by either fast or thermal neutrons, excites a phosphor and gen- erates light. This light may then be collected by a photon detector. Scintillators are designed in the form of thin bers, crystalline or amorphous solids, liquids, or, in the case of excimer-based detection, gaseous media. Common neutron detection scintillators include liquid and solid organics (proton-recoil plastics, 10B, Gd, or 6Li loaded plastics), and solid inorganics (Li crystals, ZnS(Ag) compounds, lithiated glass, and glass bers). Fast-neutron scintillators have also been developed using 3He [54, 55, 56]. A diagram of a typical scintillation detector appears in Figure 2.4. 22 Figure 2.4: Left: Lithium-doped scintillating bers used in a large-area neutron detector [23]. Right: Diagram of a scintillation-based neutron detector using a solid scintillation crystal and a photomultiplier tube. Redrawn from [57]. Characteristics of ideal scintillators include (1) transparency to their own radiation, (2) high conversion e ciency between recoil energy and uorescent radiation, (3) short scintillation decay times, (4) linear conversion, and (5) glass-like indices of refraction [1]. Ranges of recoil nuclei and charged particles are signi cantly shorter in dense scintillators than in proportional tubes. As a result, fast response times, on the order of a few nanoseconds, make scintillators more attractive in speci c appli- cations [48]. One fundamental disadvantage of solid scintillation-based detectors are high sensitivities to gamma-ray backgrounds. In many cases, detection probabilities for neutrons and gamma rays are nearly comparable. However, some scintillators exhibit delayed uorescence (phosphorescence) from triplet-state excitations. These delayed emissions enable pulse-shape discrimination for distinguishing neutron ra- diation in mixed elds [58]. High-LET particles produce higher triplet densities and are, therefore, distinguishable from low-LET radiation, as shown by the distinct de- cay pulses in Figure 2.5 [59]. This same phenomenon may prove useful for gamma ray discrimination in future excimer-based neutron detectors. 23 Figure 2.5: Left: Potential curve diagram of uorescence and phosphorescence from singlet and triplet states [60]. Right: Plot of scintillation light intensity over time for particles with di erent LETs, demonstrating variable pulse shapes used to distinguish radiation types [61]. Scintillation-based neutron detectors have been developed predominantly for use in neutron radiography, rather than SNM detection [23]. Lithium-bearing solid plastics, for example, have been fabricated with intrinsic e ciencies near 100%, for thicknesses of 10 mm [62]. Inorganic scintillators are also widely used. Mod- ern ZnS:Ag/6LiF detectors have intrinsic thermal-neutron detection e ciencies of 40%. Similar detectors using ZnS:Ag/10B2O3 have been fabricated with intrinsic e ciencies of near 60% [63]. These scintillators are usually kept quite thin (1-2 mm) due to their relative opaqueness. Even more modern scintillators (ie. Cs2LiYCl6:Ce) have shown promise of neutron-gamma signal separation through pulse-height dis- crimination [23]. Some characteristics of common scintillators, including typical light yields, appear in Table 2.1. 24 Table 2.1: Characteristics of various thermal-neutron scintillators, including light yields per neutron (Yn) and gamma-ray (Y ) interaction, emission wavelengths ( h ), and scin- tillation decay times ( ) [64]. Host Dopant Density (g/cm3) Yn Y h (nm) (ns) 6Li-glass Ce 2.5 6000 4000 395 75 6LiI Eu 4.1 50,000 12,000 470 1400 6LiF/ZnS Ag 2.6 160,000 75,000 450 1000 LiBaF3 Ce,K 5.3 3500 5000 190-330 1/34/2100 LiBaF3 Ce,Rb 5.3 3600 4500 190-330 1/34/2400 6Li6Gd(11BO3)3 Ce 3.5 40,000 25,000 385 200/800 Cs26LiYCl6 Ce 3.3 70,000 22,000 380 1000 Cs26LiYBr6 Ce 4.1 88,000 23,000 389 89/2500 2.3.3 Semiconductor neutron detectors Semiconductor neutron detectors consist of a semiconductor layered with a thermal- neutron converter, or a fast-neutron recoil material. These detectors exploit the narrow band gap in speci c materials to generate electron-hole pairs. After a neutron is absorbed or scattered in a material coupled to the semiconductor, charged particles or recoil nuclei excite electrons from the valance band into the conduction band, leaving behind positively-charged holes. The motion of these electron-hole pairs in an applied electric eld constitutes the fundamental detection signature, with no need for further ampli cation. Neutron detectors of this type include planar, semiplanar (pyramidal), and pillared thin- lm geometries, as well as multisurface detectors. Common materials in these detectors include 6LiF and 10B, on Si or GaAs semiconducting substrates, and bulk detectors of CdTe, HgI, B4C and BN [62, 65]. A diagram of a semiconductor neutron detector appears in Figure 2.6. Semiconductor detectors are sought after for their ruggedness and fast response times. Typical electron-hole lifetimes in pure Si are on the order of 100 ns [68]. 25 Figure 2.6: Left: SEM image of 10B deposited on a Si pillar structure for a semiconductor neutron detector [66]. Right: Diagram of a semiconductor neutron dector [67]. In contrast to proportional tubes and some scintillation detectors, semiconductor neutron detectors do not require high voltage. Each neutron interaction has the potential to generate 1.5 106 electron-hole pairs [69]. However, much of the Q-value or recoil energy from a neutron interaction may be lost in a converter layer or contact layer, before the charged-particle reaction products reach the active semiconductor. Both the thickness of the conversion-layer lms, and the orientation of the detector with respect to the neutron source, a ect the average energy deposition within the active volume. Additionally, gamma-ray scattering in silicon is non-negligible, and has the potential to create signi cant background noise in mixed elds. Gamma-ray rejection e ciencies of semiconductor neutron detectors range between 10 2 and 10 6 [23]. Standard 2D layer designs in semiconductor neutron detectors typically have in- trinsic e ciencies of 2-5%. Compound devices using multiple thin- lm conversion 26 layers, or even multiple converter types, have achieved thermal e ciencies greater than 25% [67]. More recently, attempts to increase semiconductor detector e - ciency have turned to advanced semiconductor processing techniques to fabricate more exotic structures. Pillar-type detectors, for example, have been fabricated with intrinsic e ciencies greater than 50% [70]. Single-crystal diamond has been used in combination with 6Li to form a dual fast and thermal neutron detector with comparable e ciencies [71]. Despite their advantages, semiconductor neutron de- tectors are limited by size. The fabrication of semiconductor detectors with surface areas and collection e ciencies large enough for SNM-detection, still faces numerous practical challenges. 2.3.4 Other neutron detection methods A variety of neutron detectors do not fall into the above categories. However, some of the technologies are immature, undeveloped, or unsuited for most applications. Microchannel plates rely on electron showers in narrow capillaries of borosilicate glass [72]. Water-based Cherenkov detectors contain large volumes of water and suspended gadolinium particles (GdCl3), which, when activated, emit beta radi- ation to generate Cherenkov light [73]. Bubble chambers employ superheated or supertensioned uids, which nucleate upon interaction with energetic radiation [74]. Calorimeters with minuscule heat capacities (10 nJ/K) can detect neutron reactions through small changes in temperature [75]. An array of passive detection techniques also exist, including foils, dosimeters, and imaging plates. 27 Excimer-based neutron detection is an alternative method which has not yet received signi cant attention. The remainder of this chapter addresses topics that directly relate to the investigation of this technique. 2.4 Boron-Neutron Capture The excimer scintillation experiment at the University of Maryland used thin lms of boron-10 to convert neutrons into charged particles. Boron is a metalloid, found in nature in oxide form, with an abundance in the earth?s crust of 8.6 ppm by mass [76]. Boron-10 (10B) is one of two stable isotopes of boron, with an isotopic abundance of 19.9%, and a large thermal-neutron absorption cross section ( 0 = 3842 8 b) [32, 77]. A plot of the energy-dependent 10B absorption cross section appears in Figure 2.7. The cross section is structureless below 100 keV, an example of a nearly ideal 1=v absorber. Figure 2.7: Boron-10 neutron absorption cross section between 0.01 meV and 20 MeV [77]. 28 For 99.93% of thermal-neutron absorptions, 10B undergoes an exothermic neutron- capture reaction, resulting in the emission of an alpha particle and a 7Li ion, as shown in Figure 2.8. The products of the 10B(n; )7Li reaction are released with a total energy of either 2.792 or 2.310 MeV, depending on the nal state of the 7Li nucleus. In the rst case, the 7Li nucleus is emitted in the ground state. In the second case, it is emitted in an excited state, which decays with a 10 13 s half-life by emitting a 0.48 MeV gamma ray. The branching ratios for these two pathways are 6% and 94%, respectively. Following thermal-neutron absorption, the reaction products are emitted isotropically, in opposite directions, with equal momentum. This holds true for all neutrons of relatively low incident momentum. Thus, the total reaction Q-value is split among the products as 1.02 MeV (7Li) and 1.78 MeV ( ), for the 7Li ground-state pathway, or 840 keV (7Li) and 1.47 MeV ( ), for the 7Li excited-state pathway. Figure 2.8: Boron-neutron capture reaction. Due to its relative abundance, large absorption cross section, and large branching ratio to an exothermic reaction, 10B is an ideal nucleus for conversion in a thermal- neutron detector. In many instances, it is a suitable replacement for 3He. Following 29 the boron-neutron capture reaction, the kinetic energy of the reaction products may be used to generate detectable, secondary radiation. The transport of these charged particles is discussed in the following section. 2.5 Charged-Particle Transport Heavy charged particles, such as atomic nuclei, are a form of directly ionizing radia- tion, which interact continuously when passing through matter. The interactions to which they are subject include the long-range Coulomb force, elastic scattering via collisions with atomic electrons, inelastic scattering, resulting in the excitation of atomic electrons, and nuclear recoil. Typically, energetic charged particles interact thousands or tens of thousands of times before accumulating electrons to become a neutral atom. This leads to an abundance of ionized and excited atoms along a charged particle?s path, which may generate an electrical or optical signal within a radiation detector. Several de nitions are helpful to understand charged-particle transport, including stopping power, range, and straggling. Stopping power is de ned as the total energy lost per unit path length, or the rate of energy loss at a given point along a particle?s path [2]. Without accounting for energy loss due to nuclear interactions, linear stopping power may be represented by the relationship, S = dE dx col + dE dx rad (2.10) 30 where dE dx col represents energy loss due to collisions involving ionization and elec- tronic excitation, and dE dx rad represents energy loss by radiation. For heavy charged particles, radiative energy loss only becomes signi cant at energies in the GeV range. This component may, therefore, be ignored when dealing with energies in the realm of thermal-neutron reactions. The rst term in Eq. 2.10 is often referred to as a particle?s rate of linear energy trans- fer (LET). Heavy charged particles, such as those from the boron-neutron capture reaction, are considered high-LET radiation because they deposit their energy over very short distances. Classically, charged-particle energy loss for non-relativistic, high-energy ions can be described by the Bethe-Bloch stopping power relation, S = 1 N dE dx = 1 (4 0)2 4 z2e4 m0v2 NZ ln 2m0v2 J (2.11) where z is the charge number of the particle, e is the charge of the electron, m0 is the rest mass of an electron, v is the charged particle velocity, N is the atomic density of the material, Z is the atomic number of the material, 0 is the vacuum permittivity, and J is an experimentally determined quantity describing the average excitation potential of a representative atom in the material. For non-relativistic particles of mass m and energy E, particle velocity is given by the equation, v = r 2E m (2.12) 31 Range is de ned as the point at which charged particle ionization falls to zero. Ranges for speci c particle types, energies, and materials are often empirically tab- ulated, or calculated based on tabulated values, using scaling laws and the Bragg- Kleeman rule. If the range (R) of a particle is known, the stopping time of the particle may be calculated using the equation, T = R hvi = R K r m 2E (2.13) where hvi is the average particle velocity, and K is a factor used to account for non-uniform deceleration of the particle. Typically, because a particle?s energy loss is greater towards the end of its range, K is approximated as a fraction somewhat greater than 1=2 [1]. The maximum energy that can be transferred from a charged particle in a single collision with an electron, is about 1/500 of the particle energy per nucleon. Thus, the de ection of a charged particle by any one interaction is small, and its path will tend to be fairly straight until the end of its range. Speci c energy loss and charged particle ranges are often represented by characteristic Bragg curves. The Bragg curves of charged-particle reaction products from the boron-neutron capture reaction, traveling through boron, appear in Figure 2.9. The maxima toward the end of the -particle curves, represent the increase in the energy loss rate at the end of the particle?s path. Additionally, the small tails at the very end of these curves are due to range straggling, the variation in ranges of charged particles of a given 32 type and energy. This e ect is attributed to the stochastic nature of charged-particle interactions. Figure 2.9: Ranges of charged particles in boron [67]. Knowledge of charged-particle ranges in various media is useful for predicting the behavior of thermal-neutron detectors. For example, Wang et al. describe analytical calculations of e ciencies of multi-layer boron thin- lm semiconductor detectors using charged particle ranges [78]. It is also possible to simulate charged particle transport using Monte Carlo methods as statistical solutions for the Bethe-Bloch equation. In this research, the TRIM (Transport of Ions in Matter) code was used to simulate the transport of the charged particle products from the 10B(n; )7Li reaction, under conditions similar to those created in the scintillation experiments. The results of these simulations appear in Chapter 8. Excimer-based neutron detection relies on the transfer of energy between charged- particle reaction products and the surrounding noble-gas atoms to generate noble- gas excimers. A neutron detection signal is provided by photon emissions from these excimers, as discussed in the following section. 33 2.6 Noble-Gas Excimers Excimers were rst identi ed in 1927 by Lord Rayleigh, who documented broad continuum emissions from mercury vapor in the ultraviolet region of the electromag- netic spectrum [79]. Later, in the 1930s, Hop eld and Mulliken recorded emissions from the rst rare-gas excimers using high-pressure helium, and calculated poten- tial curves for excited excimer states. Excimer lasers, widely used today in medical surgeries and photolithography, were conceived in 1960, and, in 1970, stimulated emission of noble gas excimers was rst observed by Basov while irradiating liquid xenon with high-energy electrons [80, 81]. Ground state interactions of two closed-shell atoms are typically repulsive. The interaction of an excited closed-shell atom with the ground state of another closed- shell atom may, however, form a bound state. A noble-gas excimer is a dimer molecule of two noble-gas atoms that exists only in an excited electronic state (ie. Ar 2, Kr 2, Xe 2). When these excimers decay, they dissociate and emit far-ultraviolet radiation, characterized by broad continua, as shown by the spectra in Figure 2.10. The term excimer may also refer to excited dimers of other monatomic gases (ie. Hg 2), excited aromatic molecules, as well as mixed dimers, called exciplexes (KrXe*, ArF*, HgNH3*). In this thesis, excimer will refer exclusively to excited homonuclear dimers of the rare gases. This speci c type of excimer, when produced by the slowing down of charged particles, forms the basis of an optical neutron detection signal. Some of the relevant noble-gas properties appear in Table 2.2. 34 Table 2.2: Properties of noble gases, including atomic number (Z), atomic weight (A), density ( ) and mean free path (l) at STP, peak excimer emission wavelength ( peak), and peak excimer photon energy (Eh ) [76, 54]. Gas Z A (u) (kg/m3) l (nm) peak (nm) Eh (eV) Argon 18 39.95 1.761 72.29 128 9.57 Krypton 36 83.80 3.696 55.66 150 8.42 Xenon 54 131.3 5.897 40.55 175 7.02 2.6.1 Excimer scintillation As energetic charged particles travel through noble gases, they deposit their energy by excitation and ionization of the surrounding noble-gas atoms. These processes include, R + A ! e+ A+ +R0 (2.14) R + A ! A +R0 (2.15) e+ A+ ! A (2.16) where R is a charged particle, A is a noble-gas atom, e is an electron, A+ is an ionized noble-gas atom, R0 is a charged particle of reduced energy, and A is a noble-gas atom in an excited electronic state [54]. Excited noble-gas atoms may decay through the radiative emission of a far-ultraviolet photon, A ! A+ h a (2.17) However, at increased pressures (103{104 Pa), the probability of another atom re- absorbing this atomic photon (h a) greatly increases. These photons are said to become \trapped" within the gas volume [54]. Independently, the likelihood of 35 three-body collisions also increases at higher pressures. This leads to the excimer forming process, A + 2A! A 2 + A (2.18) where A 2 is a noble-gas excimer. This reaction occurs on a time scale of 10 11 to 10 12 s [54]. The excimers may then decay by emitting a photon, A 2 ! 2A+ h m (2.19) Variations in the amount of energy carried away by the third body, and the repul- sive nature of the molecular ground state, enable emissions to occur over a broad molecular continuum. Examples of these continua appear in Figure 2.10. The two principal maxima in these spectra are the result of excimer vibrational levels. The short wavelength maxima are attributed to transitions from high vibrational levels, and the longer wavelength maxima are attributed to the same transitions follow- ing vibrational relaxation of the molecular states [82]. Because molecular excimer emissions do not coincide with atomic absorption resonances of the same gas, no- ble gases are highly transparent to their own excimer radiation. This transparency makes excimer photons (h m) ideal constituents of a radiation detection signal. The atomic ground states of the noble gases beyond helium have the electron shell con guration s2p6 (1S0). The rst excited con guration, s2p5s, describes both sin- glet (1P1) and triplet (3P0;1;2) spin states. The combination of an excited noble-gas atom with a ground-state noble-gas atom may form the molecular states 1 +u or 3 +u . 36 Figure 2.10: Excimer emission spectra from bombarding Ar (left), Kr (center) and Xe (right) with 4 MeV protons at 50 and 400 torr [83]. Transitions from the high vibrational levels of the 1 +u and the 3 +u states to the repulsive ground state, give rise to photon emission in the primary continuum of the excimer emission spectra [84]. Following excitation, noble-gas excimer systems typ- ically evolve according to the ow patterns shown in Figure 2.11. However, the true complexity of these transitions is perhaps better illuminated by the argon molec- ular potential curves in the same gure. Relaxation occurs by transitions through a tightly nested set of electronic states. Following relaxation, the only available downward channel is by emission of FUV radiation. The numerous crossings of the excimer levels create a highly-e cient energy funnel toward this radiative pathway [81]. Spontaneous radiative decay times for simple excimers fall between 0.005{5 s [79]. Comparatively, dissociation times for very weakly bound molecular states may be as short as 10 13 s, on the timescale of molecular rotation [85]. Excimers are known to exhibit both fast and slow decay times, due to the di erent probabilities of transi- tion between the singlet and triplet excited states and the dissociative ground state. 37 Figure 2.11: Right: Energy ow pattern for the decay of rare gas excimers [81]. Left: Potential curves for the Ar 2 excimer [79]. Therefore, excimer scintillation exhibits a structured decay with two time constants, as shown by the oscilloscope traces in Figure 2.12. The classically-forbidden triplet- state transition may have a characteristic decay time two orders of magnitude longer than the singlet-state decay. Collisions play an important role in the decay process, making decay times strongly dependent on gas pressure [82]. At higher pressures, singlet-triplet transitions, aided by electronic and atomic collisions, occur with in- creased probability [79]. Therefore, both the formation and the decay of excimers are more rapid at high gas pressures. Numerous sources report fast and slow decay times for gas-phase noble-gas excimer scintillation. For example, Mutterer et al. describe a series of experiments in which 38 noble gases were bombarded by relativistic heavy ions [86]. Their results show scintillation rise times in argon and xenon gas of 10 ns, fast decay constants of 140 ns (Ar) and 50 ns (Xe) at 57 torr, and slow decay constants of 900 ns (Ar) and 250 ns (Xe) at 300 torr. Koehler et al. report rise times and decay times for argon and xenon at pressures between 0.1 and 100 atm, as plotted in Figure 2.12. Morikawa et al. have compiled an extensive list of gas-phase excimer time constant measurements, which appears in Figure 2.13. Figure 2.12: Left: Fast multiplier signals from scintillations of xenon at (a) 300 torr and (b) 150 torr, and in argon at (c) 57 torr [84]. Right: Time constants for (a) xenon, and (b) argon as a function of pressure, showing formation time constants ( ), as well as fast and slow decay time constants ( ) [87]. Knowledge of excimer scintillation yield is essential for understanding the e ciency of excimer radiation pathways, and for estimating the output amplitudes and energy 39 Figure 2.13: Literature values for excimer radiative decay times in a dilute gas [88]. Times are listed in nanoseconds. resolution of noble-gas-based detectors [38]. Yield characteristics of a scintillator are often represented by the value Wh , given by, Wh = E Y (2.20) where Wh is the average energy expended per scintillation photon, E is the energy deposited in the scintillator, Y is the number of photons generated in the scintilla- tor. While numerous sources report Wh values for liquid noble-gases [89, 90], and others report secondary scintillation yields in gas-scintillation proportional counters [91, 92], few sources provide Wh values for gas-phase, noble-gas scintillators. Chan- 40 drasekharan et al. report Wh values of 67.9, 61.2, and 55.9 eV, for Ar, Kr, and Xe, respectively, at unspeci ed gas pressures [37]. Saito et al. report values for Wh from experiments with an alpha source, at pressures between 1 and 10 atm, as shown in Table 2.3 and plotted in Figure 2.14. The lack of su cient data on noble-gas scintillation yields and the overall disagree- ment among the reported values, demonstrate the necessity for careful measurements of this parameter before developing a deployable detector based on gas-phase ex- cimer scintillation. A combination of features { (1) large light output, (2) fast decay, (3) transparency, (4) unique decay structure, (5) immunity to radiation damage { make noble gases particularly promising for use in radiation detectors [82]. A variety of existing noble-gas detectors are brie y discussed in the following section. Figure 2.14: Absolute excimer photon yields in argon, krypton, and xenon at pressures between 1 and 10 atm, due to 5.49 MeV alpha particles [38]. 41 Wh (eV/photon) Pressure (Pa) Argon Krypon Xenon 1:01 105 42:0 3:0 34:3 1:6 2:03 105 50:6 2:6 41:3 2:4 27:5 1:3 4:05 105 46:3 2:4 34:8 1:8 21:0 0:8 6:08 105 36:0 1:9 27:9 1:2 17:5 0:6 7:09 105 32:1 1:6 25:9 1:0 16:5 0:5 8:10 105 28:9 1:3 24:3 0:9 15:6 0:5 1:01 106 25:3 1:1 22:3 0:8 14:6 0:5 Table 2.3: Wh values of noble gases when subjected to 5.49 MeV -particle radiation [38]. 2.6.2 Noble-gas-based detectors Noble-gas-based detectors (NGBD) were rst developed in the 1940s, following ob- servations of high electron mobility in liquid and solid argon. Today, NGBD are currently used in a number of applications, including x-ray astronomy, dark-matter detection, neutrino detection, particle physics calorimetry and medical imaging [54]. These detectors include gas-, liquid-, and solid-phase detection media, in the form of scintillation detectors, ionization chambers, gas-scintillation proportional coun- ters, and two-phase electron emission detectors. NGBD rely on the direct inter- action of radiation with a noble gas detection media to precipitate photon emis- sion, electron emission, or both. A wealth of information on NGBD is available in [54, 84, 93, 94, 95]. Speci cally, information on 3He scintillators for fast neutron detection is provided in [54, 55, 56]. 42 Chapter 3 Review of Preliminary Experiments Prior to the excimer scintillation experiment at Maryland, which constitutes the ma- jority of this research project, a number of experiments were carried out at NIST, to demonstrate the viability of noble-gas scintillation as a means of neutron detec- tion. These experiments were conducted at the NIST Center for Neutron Research (NCNR), between Fall 2008 and Spring 2011. In April 2011, the NCNR was shut- down for expansion and reliability improvements, including the construction of a new guide hall, at which point all research was transferred to the MUTR. 3.1 The NIST Center for Neutron Research The NIST Center for Neutron Research (NCNR), located on the NIST campus in Gaithersburg, MD, houses the National Bureau of Standards Reactor (NBSR), along with more than 26 instruments for scienti c study using thermal and cold neutrons [96]. The NBSR consists of 30 highly-enriched fuel elements, which are cooled and moderated by heavy water surrounding the core. Operating at a thermal output of 20 MW, in 38-day cycles, the reactor has a peak in-core ux of 3:5 1014 cm 2s 1 [97]. Nine beamlines surrounding the core supply thermal-neutrons to various instruments 43 Figure 3.1: The NCNR guidehall [11]. within the reactor con nement building. Additionally, a 5-liter liquid-hydrogen cold source provides cold neutrons to 7 radial neutron guides and the instruments located in the NCNR guidehall. A photograph of the guidehall appears in Figure 3.1. A diagram of the reactor con nement and the guidehall appears in Figure 3.2. All data for the preliminary excimer scintillation experiments were collected on beamline 6 (NG-6), at experimental stations NG-6A and NG-6M. Starting in the reactor core, neutrons stream from the cold source thimble into the various neu- tron guides. An 18-cm gap between the upper and lower fuel regions, and multiple bismuth lters upstream of the neutron instruments, prevent gamma-ray radiation 44 from streaming into the guides. The guides themselves are coated with nickel-58 to achieve nearly total re ection of low-momentum, low-angle-of-incidence neutrons. This enables neutrons to be transported to the far end of the guides, 20 - 60 m from the source, with minimal losses [98]. Beamline NG-6A consists of two monochroma- tors: a pyrolytic graphite crystal and cooled slab of beryllium. These select 0.496-nm wavelength neutrons from a polychromatic neutron beam. Beamline NG-6M resides downstream of NG-6A, and contains an additional potassium-intercalated graphite monochromator, and additional lters for selecting 0.89 nm neutrons from the beam. At both locations, an aperture was used to collimate the monochromatic beam to a square pro le, 4 4 mm. During the experiments, the nominal neutron ux at the sample locations was (2:61 0:37) 105 cm 2s 1 [39]. Figure 3.2: Plan view of the NBSR (left) and the the NCNR guide hall (right), including the thermal and cold neutron instruments [11]. 3.2 Lyman-Alpha Experiment at the NCNR The initial noble-gas scintillation measurements set out to observe Lyman-alpha radiation following the 3He(n; tp) reaction, described by Eq. 1.2. With a Q-value 45 of 764 keV, this reaction was believed to produce Lyman-alpha photons (121.5 nm) through both charge transfer and excitation, in the processes, H+ + He! H(2p1) + He+ ! H(1s1) + He+ + h (3.1) H(1s1) + He! H(2p1) + He! H(1s1) + He + h (3.2) By joining a solar-blind photomultiplier tube to a scintillation cell lled with 3He (25 - 800 torr), and irradiating the cell in a neutron beam, Lyman-alpha radiation was observed at rates of up to 46 photons/neutron absorption [99]. Filters were later placed between the cell and the photomultiplier tube to con rm that the observed radiations were Lyman-alpha photons. The results of these experiments, at various pressures of 3He, are plotted in Figure 3.3. These measurements, their analysis, and modeling of the experimental processes are described in [99] and [100]. A description of an experimental apparatus similar to the apparatus used in these measurements, is provided in the following chapter. 3.3 Excimer-Based Experiments at the NCNR Following the Lyman-alpha experiment, attempts were made to increase the ob- served photon signal by introducing noble gases into the scintillation cell assembly. These experiments involved mixtures of 3He and heavier noble gases (Ar, Kr or Xe), as well as substitutions of 3He with alternative neutron absorbing materials, mixed with the same noble gases. The excimer emission signal from various neutron reac- 46 Figure 3.3: Lyman-alpha photons generated per neutron reacted with 3He, as a function of 3He pressure [99]. tions was measured to be a factor of a thousand more intense than light from the Lyman-alpha experiments. A series of spectroscopic measurements were performed to con rm that the observed signal was a result of excimer radiation [39]. Using various spectral lters (sapphire, CaF2, fused silica), the radiation was identi ed as broad-continuum, far-ultraviolet light, characteristic of noble-gas excimer emissions. By inserting these lters be- tween the cell window and the PMT, the short-wavelength cuto of the detector system was shifted from 115 nm to a longer wavelength: 122 nm for CaF2, 142 nm for sapphire, and 160 nm for fused silica. The Ar2* excimer emissions were detected only with the CaF2 lter in place; the other two lters were opaque to the emitted radiation. Emissions from Kr2* were detected through CaF2 and weakly through sapphire, while emissions from Xe2* were detected through all three lters. 47 The response of the photomultiplier tube to a range of monochromatic synchrotron light, while behind these lters, appears in Figure 3.4. The relative change in the observed scintillation signal from various noble gases, when placed behind these lters, appears in the same gure. Figure 3.4: Left: Photomultiplier tube response to a range of monochromatic ultraviolet synchrotron radiation behind various lters [39]. Right: Relative excimer emission signals observed passing through various lters under identical experimental conditions [39]. 3.3.1 Helium-3 and noble-gas mixtures The rst excimer-based neutron detection experiment consisted of a cell containing a mixture of 3He and a single heavier noble gas (Ar, Kr, or Xe). This cell was joined to a calibrated lter-detector package, sensitive to various bands of far ultraviolet (FUV) radiation, between wavelengths of 120 and 200 nm. Helium-3 was introduced into the cell to a pressure of 200 torr, and noble gases were introduced at additional pressures between 25 - 1000 torr. When exposed to a cold neutron beam at the NCNR, the assembly produced tens of thousands of FUV photons per reacted neu- 48 tron. This output suggests a conversion e ciency of >30%, between the neutron reaction Q-value and the excimer scintillation emissions. The results are plotted in Figure 3.5. It is likely that this mechanism is also present in mixtures of 3He and 4He or Ne, but the peak emissions of these excimer species are outside the spectral range of the detector (80 and 77 nm respectively). Figure 3.5: Left-hand scale: photons per neutron reacted with 3He at 200 torr, with 600 and 1200 torr of Xe, Kr, and Ar. Wide error bars correspond to random and systematic uncertainties in the derived photon signal. Right-hand scale: counts from which the left- hand scale is derived. Narrow error bars correspond to statistical counting uncertainties in signal and background [39]. In consideration of the recent 3He shortage, continued development of a neutron detector based on 3He was deemed impractical. It was hypothesized that similar results could be attained from any exothermic neutron absorption reaction with a Q-value comparable to the 3He(n; tp) reaction. 49 3.3.2 Preliminary thin- lm experiment The rst measurements using thin- lms were performed with an assembly similar to the 3He experiments, including a few minor alterations to accommodate the lm targets. Four di erent lms were used in the experiments: three containing boron, and one containing lithium. The rst boron sample was 200-nm thick, enriched 10B, found in storage from a previous application. The other two boron samples were natural boron, fabricated by RF sputtering at the NIST CNST. These samples were 250-nm and 500-nm thick. The fourth sample, containing lithium-6 carbonate (6Li2CO3), was fabricated at the University of Maryland Nanocenter, using thermal annealing to achieve thicknesses of 20 m. All lms were deposited on silicon. By placing the lms within the scintillation cell, and irradiating the cell in a cold- neutron beam, a large excimer scintillation signal was once again observed, on the order of many thousands of FUV photons per neutron absorption. Only xenon and argon gas were used, at pressures between 25 - 800 torr. However, the results of these experiments exhibited inconsistencies between the signals from the enriched and natural boron lm samples. This discrepancy may potentially be attributed to the stability of the sputtered natural-boron lms when exposed to air. Over the course of the experiments, visible changes were observed in the color and luster of the lms, perhaps indicating oxidation. These results led to careful fabrication and storage procedures in the subsequent experiment at the MUTR. 50 The ndings of the preliminary experiments at NIST provided a foundation for the excimer scintillation experiment at Maryland. The experiment at Maryland focused speci cally on observing the scintillation from thermal-neutron irradiation of enriched 10B thin lms, in the presence of Ar, Kr, and Xe. This experiment and its results are the subject of the following chapters. 51 Chapter 4 Excimer Scintillation Experiment at Maryland The excimer scintillation experiment at the Maryland University Training Reactor was designed to generate and observe far-ultraviolet noble-gas excimer emissions, as a result of the boron-neutron capture reaction. Measurements of these emissions were collected for di erent gas types, gas pressures, and boron target thicknesses. Careful calibrations and characterizations of the various components of the experi- ment complimented this scintillation data, to allow its interpretation as quantities which are relevant to the development of an excimer-based thermal-neutron detec- tor. The end results of the analysis include: 1. Scintillation yield per neutron absorption 2. Scintillation yield per unit energy deposited 3. Excimer lifetime Speci cally, this experiment examined excimer emission following thermal-neutron absorption by boron-10 thin lms, under pressures of argon, krypton, and xenon, between 50 - 800 torr. This chapter describes the experimental setup and procedures. The results of the experiments and the calibrations contributing to those results are discussed in the subsequent chapters. 53 4.1 The Maryland University Training Reactor The Maryland University Training Reactor (MUTR) is a TRIGA reactor located on the University of Maryland campus in College Park, MD. Initially fueled with highly-enriched, plate-type MTR fuel, the reactor achieved criticality in October of 1960, becoming the rst nuclear reactor in the state of Maryland [101]. In 1974, 93 TRIGA-type fuel elements (19% enrichment) were installed to replace the plate fuel, increasing the reactor power to 250 kW. Today, the reactor?s primary functions are operator training and scienti c research. Five experimental facilities allow access to the MUTR?s radiation elds, including a through tube, two beam ports, a pneumatic rabbit system, and a thermal column. The layout of the MUTR core, and the respective positions of these facilities, appear in Figure 4.1. Figure 4.1: Left: Plan view diagram of the MUTR core [101]. Right: The MUTR core from the reactor bridge [102]. For this experiment, the MUTR thermal column was utilized as a source of neutrons. A collimator insert was designed for this facility, to allow only a narrow directional beam of neutrons to stream past the boundary of the concrete containment. The thermal column earns its name from the large quantity of graphite separating the 54 reactor core and the inner face of its access cavity. Fission neutrons di using through this graphite volume are moderated to thermal energies with a high probability, making this facility ideal for the development of a thermal-neutron detector. A diagram of the thermal-column access cavity appears in Figure 4.2. The fabrication of the neutron-beam collimator, and the characterization of the thermal-column neutron beam are discussed in Chapter 7. Figure 4.2: Elevation view of the MUTR thermal column. The access plug is highlighted in red. The core centerline is labeled on the right. The edge of the concrete reactor containment appears on the left [101]. 4.2 Experimental Apparatus The equipment used in this experiment included various boron-10 thin- lm targets, a gas-scintillation cell, a photomultiplier tube, a gas-handling system, a ssion- chamber beam monitor, counting electronics for both the photomultiplier tube and 55 the beam monitor, a neutron-beam collimator, a neutron-beam aperture, lead and borated polyethylene shielding, and a lithium-glass beam block. 4.2.1 Scintillation cell The scintillation cell was designed to produce and detect excimer radiation when exposed to a neutron beam, by providing a high-purity noble-gas environment, surrounding a thin- lm neutron-absorbing target. The cell consisted of a 70-mm stainless-steel cube, with metal-seal ange ports on all sides, attached to a pho- tomultiplier tube (PMT). The PMT (Hamamatsu R6835) was a solar-blind head- on tube, tailored speci cally for FUV detection (115-200 nm), with a 23-mm CsI photocathode, and an integrated MgF2 window. The tube was contained in end- on housing (McPherson 658), and placed behind a second, discrete MgF2 window (MPF A4530) and a di erential volume, which was continuously evacuated for the duration of the experiments. Magnesium uoride is transparent to photons of wave- lengths >120 nm, and, therefore, provided a suitable transmissive barrier between the cell and the PMT. The discrete window on the scintillation cell was 2.5 mm thick. The calibration of the PMT/MgF2 detector package is discussed in Chapter 6. A diagram of the scintillation cell appears in Figure 4.3. Boron-10 thin- lm targets, 0.3 - 1.2 m thick, deposited on silicon substrates (6.45 cm2), were placed at the center of the cube beneath the PMT. The lms were angled at 45 with respect to the incident neutron beam to increase the pathlength 56 of the beam by a factor of p 2. The lms were held in place by a slotted aluminum cylinder (2.54 cm diameter). The cylinder, in turn, was held in place by a groove in the Con at ange at the bottom of the cube. The inner surface of this cylinder was coated with black copper-oxide (Ebonol C) to reduce re ections of FUV photons within the cell. The fabrication and characterization of the boron-10 lms are discussed in Chapter 5. Silica windows (0.5 mm thick), attached to opposing sides of the cell, provided a throughway for the thermal-neutron beam with little attenuation. To minimize backgrounds from stray FUV light, these windows were covered with aluminum foil during PMT operation. All anges on the scintillation cell were sealed with copper gaskets, with the exception of the seal between the PMT and the di erential volume, for which a Viton gasket was used. 57 Figure 4.3: Diagram of the scintillation cell. 58 Figure 4.4: The scintillation cell with one ange port open, exposing the thin- lm target and the slotted lm holder. 59 4.2.2 Gas-handling system A gas-handling system was designed to evacuate and ll the scintillation cell. This system consisted of various components connected by a stainless steel manifold: a turbomolecular pump, a diaphragm pump, gas lecture bottles, a digital pressure indicator, a vacuum gauge, a residual-gas analyzer, a gas puri er, a metering valve, a series of bellows-sealed valves, and VCR, Con at, and Swagelock interconnections. Diagrams of the gas-handling system appear in Figures 4.6 and 4.7. Prior to each data collection run, the turbomolecular pump (Pfei er TPH/TPU 180 HM) and the diaphragm pump (Vacuubrand MD4T) were used to evacuate the gas handling system and the scintillation cell to high vacuum (0.5 - 1:0 10 6 mbar). Vacuum pressure in the evacuated cell was measured with a cold-cathode vacuum gauge (Pfei er PKR251). Whenever the system was opened and resealed, such as to change thin- lm targets, it was baked at 60 C overnight to remove water and other adsorbed contaminants from the chamber walls. A residual-gas analyzer (SRS RGA-100) was used to perform a helium leak check of the system upon initial assembly, and to monitor contaminants in the system prior to lling with noble gas. A typical RGA scan appears in Figure 4.5. All noble gases were passed through a puri er (Microtorr MC1-902-F) designed to remove trace contaminants (H2O, O2, CO, CO2, H2, and hydrocarbons) to < 1 ppb. The gases used in these experiments were of research grade purity. Noble-gas purities and contaminants, determined by manufacturers? analyses, appear in Table 60 4.1. A metering valve (SS-MGVR4-MH) was used to control the ow of noble gases into the scintillation cell. Additional valves isolated the cell, the gas manifold, and the di erential volume. A digital pressure indicator (Omega DPI 705) with a silicon transducer was used to measure the pressures of the admitted gases to an accuracy of < 0:1% of a 30 psi full scale, or 1.55 torr. Figure 4.5: Typical scan of residual gases after gas-handling system evacuation and bake out. Gas Type Purity (%) Contaminants (< n ppm) Ar 99.9995 CO2 (0.1), CO (0.1), O2 (0.05), THC (0.1), H2O (0.05) Kr 99.999 Ar (1), CF4 (1), CO (1), CO2 (1), H2 (1), He (1) H2O (4), Xe (10), N2 (10), O2 (2), THC (0.1) Xe 99.999 Ar (1), CF4 (0.5), CO2 (1), H2 (2), H2O (0.004), Kr (5), N2 (1), O2 (0.10), THC (0.5) Table 4.1: Noble-gas purities and contaminant concentrations as listed by the manufac- turers. THC: total hydrocarbons. 61 Figure 4.6: Gas-handling system layout. The 5-way cross (highlighted in blue) appears in two orientations. Figure 4.7: Gas-handling system model. 62 4.2.3 Neutron beamline A new thermal-column insert was fabricated for use in this project. The insert was designed to allow only a narrow beam of neutrons (5.08-cm diameter) to stream beyond the reactor containment. The collimator consisted of a carbon-steel com- partmented frame and a central steel tube. Portions of the box were lled with a mixture of steel shot, para n wax, and boron-carbide (B4C) powder. Construction of this plug is described in Chapter 7. A diagram of the insert appears in Figure 4.8. During the scintillation measurements, neutrons were allowed to stream into the scintillation cell through the collimator opening. Before reaching the beam tube, neutrons propagating from the core rst di used through the thermal column, com- posed of water (0.953 cm), aluminum (0.635 cm), graphite (150 cm), and lead (5.08 cm). While both the graphite and the lead bricks in the thermal column atten- uate a signi cant portion of the gamma-ray radiation originating in the core, the gamma-ray dose rates measured in front of the scintillation cell were not insignif- icant (90 - 100 mR/hr). Gamma-ray dose rates in the vicinity of the PMT, just outside the beam, were measured at 5 - 10 mR/hr. Fitted lead bricks were placed around the photomultiplier tube to reduce direct gamma-ray interactions with the photocathode. Borated-polyethylene blocks were used as a neutron-beam stop be- hind the scintillation cell. Additional blocks were placed around the cell to reduce signal contributions from neutrons external to the beam. A remotely-operated beam 63 shutter, consisting of a lead brick and a thick boral plate, was mounted on tracks on the outer face of the collimator access plug. This shutter was used to block both neutrons and gamma-rays in the beam, when working with the experiment. Figure 4.8: Collimator insert for the MUTR thermal column. Black shading de nes segments lled with the shielding mixture. Beyond the outlet of the collimator plug, the neutron beam was further reduced to a 4-mm diameter by a Boral aperture, 8 cm in front of the scintillation cell. The neutron uence behind this aperture was monitored with a NIST-calibrated ssion-chamber reference detector. Data obtained with this detector are discussed in Chapter 7. The areal density of the 235U deposit (25S-5-4) in this detector was 458.073 g/cm2, with a relative uncertainty of 0.5%, as measured by Dr. David Gilliam. A beam block, consisting of a small slab of 6Li glass, was placed on a slide between the ssion-chamber beam monitor and the scintillation cell. This beam block presented little obstruction to gamma radiation within the beam, while ab- sorbing nearly all thermal neutrons. In this way, the beam block was used to isolate 64 the gamma-ray contribution to the PMT signal. Diagrams of these components and their arrangement appear in Figures 4.9 and 4.10. The gas-handling system, the scintillation cell, the ssion-chamber beam monitor, the beam aperture, the beam block, and the shielding were supported by an ex- truded aluminum frame (80/20). This frame was designed to enable centering of the scintillation cell within the neutron beam, and was xed to the reactor structure to prevent changes in the setup geometry. Photographs of the experimental setup in place at the MUTR appear in Figures 4.11 and 4.12. 65 Figure 4.9: Scintillation experiment con guration (close view). Figure 4.10: Scintillation experiment con guration (wide view). 66 Figure 4.11: The scintillation cell in the beamline. Neutrons propogate through the cell from left to right. Figure 4.12: The scintillation experiment beside the MUTR containment. 67 4.2.4 Electronics con guration Two sets of electronics were con gured to analyze pulses from the scintillation- cell photomultiplier tube. The rst set was used to count PMT pulses with a counter/timer (C/T), and to obtain pulse-charge distributions with a multichannel analyzer (MCA). These data were collected to calculate excimer scintillation photon yield. The second set was used to obtain distributions of the times between pulses from the PMT. These data were collected to determine the slow decay component of the excimer lifetime. A third set of electronics was used to gather pulse-height distributions from the ssion-chamber beam monitor. These data were collected to extrapolate the neutron-beam uence during the photon yield measurements. Schematics of these electronics appear in Figures 4.14, 4.16, and 4.17. With su cient energy, a photon impinging on a PMT photocathode may liberate an electron due to the photoelectric e ect. This photoelectron is directed by a fo- cusing electrode to a series of dynodes, held at successively increasing voltages. The dynodes amplify the signal through secondary emission, until charge is accumulated at the anode to form an electrical pulse. The total number of photon events may be determined by counting these pulses over a given period of time. During all of the excimer scintillation measurements, the scintillation-cell photomul- tiplier was operated at -2300 V, corresponding to a gain of 3 105 [103]. Pulses from the PMT were further ampli ed by a fast preamplifer (Ortec VT120A), with a non- inverting gain of 200, and a rise time of 1 ns. The ampli ed pulses were divided 68 to two outputs by a power splitter (Minicircuits ZFSC-2-4+). Oscilloscope traces of characteristic output pulses from these components appear in Figure 4.13. Figure 4.13: Oscilloscope traces of output pulses from the PMT (top left), the fast pream- pli er (top right), a single output from the splitter (bottom left) and both outputs from the splitter, with one output run through the TAC-setup delay line (bottom right). 50 scope termination. In the photon-counting con guration, one splitter output was connected to the C/T (Ortec 994), and the other was connected to the MCA (LeCroy qVt 3001). The MCA was used to obtain pulse-charge distributions (PCDs). Observation of these PCDs was necessary to ensure that multi-photoelectron pulses were not contributing signi cantly to the number of pulses counted by the C/T. The MCA was operated 69 in the charge mode (q-mode), with an integration width of 20 ns, and a resolution of 0.25 pC/channel, over 1024 channels. Data from the MCA were transferred to a laptop via a Perl script and a custom microcontroller board (Wiring v1.0). The digitization time of the MCA (12 s + 0.05 s/channel) was long with respect to the typical decay time of each scintillation event. Therefore, not every pulse from the PMT was collected by the MCA. The resulting PCDs were assumed to be only representations of the true distributions. The C/T was used to count the absolute number of PMT pulses. The speci cations of this module include a maximum count rate of 100 MHz, a pulse-pair resolution of <10 ns, and a xed discriminator threshold of -250 mV for negative polarity inputs. Figure 4.14: Photon-counting electronics. A time-to-amplitude converter (TAC) measures the time between \start" and \stop" pulses by generating an output pulse with an amplitude proportional to that time di erence. In the pulse-timing electronics, one splitter output was sent to the \stop" input of a TAC (Ortec 567), and the other was sent, through a delay line, to the \start" input of the same module. Two PMT pulses sent to this con guration cause the following sequence (1) the undelayed rst pulses arrives at \stop" and is ignored, (2) the delayed rst pulse arrives at \start", initiating the TAC, (3) the undelayed second pulse arrives at \stop" and TAC generates an output, (4) 70 the delayed second pulse arrives at \start" and restarts the TAC. The TAC output generated by a pair of PMT pulses is, therefore, proportional to the time di erence between the delayed rst pulse and the undelayed second pulse. An example of two TAC input pulses, originating from a single PMT pulse, appears in Figure 4.13. A schematic of this pulse timing behavior appears in Figure 4.15. The MCA, operated in the voltage mode (V-mode), was used to obtain pulse-height distributions (PHDs) of TAC output pulses. Based on the time-scale setting of the TAC, these PHDs provided distributions of the times between pulses. Figure 4.15: Visualization of TAC inputs and outputs. Three PMT pulses create two output pulses from the TAC. The rst TAC pulse (A) corresponds to the time di erence between the rst delayed PMT pulse (1d) and the second PMT pulse (2). The second TAC pulse (B) corresponds to the time di erence between the second delayed PMT pulse (2d) and the third PMT pulse (3). Figure 4.16: Pulse-timing electronics. 71 A ssion-chamber beam monitor was utilized as a reference neutron detector during the scintillation experiments. This detector houses a thin deposit of 235U (458.073 g/cm2) and a small, continuously-circulated volume of P-10 gas (90% argon, 10% methane). Neutrons that are absorbed in this deposit induce ssion with an 85% probability, and the resulting ssion products ionize the gas. The free charge may then be collected by an anode wire. Pulses from the detector were ampli ed by a charge-sensitive preampli er (Tennelec 2273) and a shaping ampli er (Ortec 671), before they were sent to a multichannel analyzer (Tracor Northern 7200). PHDs from this MCA were transferred to a computer using a Perl script. The detector was operated at +125 V. Figure 4.17: Fission-chamber beam-monitor electronics. 4.2.5 Electronics calibration Excimer scintillation events, as observed by the PMT, often consisted of multiple closely spaced pulses arriving within a window of 0.1-10 s. Oscilloscope traces of typical scintillation events appear in Figure 4.18. A series of simple calibrations were performed to verify the behavior and listed speci cations of the PMT electronics for events of this type. A function generator (Tektronix AFG 3101) was used to introduce pulses of varying amplitudes and frequencies into the preampli er, the 72 splitter, the C/T, and the MCA. The shape of a typical PMT pulse was emulated using the fastest settings of the tool: 5 ns rise time, 10 ns plateau, and 5 ns decay. In the amplitude regime, the low-amplitude thresholds of the C/T, and the MCA were observed to be -290 mV and -150 mV, respectively. Neither module exhibited an upper threshold for negative pulses up to -10 V. In the frequency regime, the C/T demonstrated 100% pulse collection at repetition rates up to the generator limit of 50 MHz. Finally, to verify the expected discrepancies between the C/T and the MCA, a pair of pulses were generated at a rate of 1 kHz. The spacing between each pulse in the pair was varied between 1 ms and 1 s. As expected, the MCA failed to count both pulses at pair spacings shorter than 20 s, due to its digitization time. Data from the C/T were, therefore, used to determine the absolute number of photon events. PCDs from the MCA were used only to verify the absence of multi-photoelectron pulses from the overall pulse population. Figure 4.18: Oscilloscope traces from typical scintillation events, showing multiple pulses within 0.5 s. 73 4.3 Measurements After designing, assembling, and testing the various components of the experimental apparatus, a sequence of measurements were made to determine (1) the excimer scintillation yield from thermal-neutron irradiation of boron-10 thin lms under Ar, Kr, Xe, and (2) the excimer lifetime under the same conditions. To measure the excimer scintillation yield, a boron lm sample was loaded into the scintillation cell, and the cell was evacuated and baked out overnight. With the reac- tor operating at 250 kW and the neutron beam propagating through the cell, PMT pulses were collected by the C/T and the MCA, for count periods of 200 seconds. During these measurements, data from the ssion-chamber beam monitor were si- multaneously collected over 1200-second count periods. Starting with an empty cell, a single noble gas was added to pressures of 50, 100, 150, 200, 300, 400, 600, and 800 torr. Two counts were recorded from the PMT at each pressure: one with lithium glass in front of the cell and one without. By collecting a pair of measurements for each set of experimental conditions, it was possible to isolate the component of the signal attributable to gamma-ray interactions. These measurements were taken consecutively to ensure the consistency of reactor conditions. Once data from the rst gas were gathered, the cell was reevacuated and the same process was repeated for the remaining gases. This entire procedure was repeated for four boron-10 targets of di erent thicknesses (300, 600, 900, 1200 nm). Background contributions from gamma rays were also measured by placing a \dummy" substrate 74 within the cell and allowing the mixed beam (neutrons and gammas) to propagate through it. This dummy substrate consisted of 2 nm of natural boron deposited on silicon. To measure the excimer lifetime, a 1200-nm boron lm was loaded into the scintil- lation cell and the cell was evacuated and baked out overnight. With the reactor, once again, operating at 250 kW, distributions of the times between PMT pulses were collected over 300-second count periods, for the same noble-gas types and the same range of pressures. Depending on the gas type and gas pressure, the TAC out- put range was adjusted in accordance with the observed distribution. The lithium glass beam block was also used to collect background distributions in this con gu- ration. The results and analysis of these measurements are discussed in Chapter 9. The calibrations, characterizations, and simulations contributing to this analysis are dis- cussed in Chapters 5 - 8. 75 Chapter 5 Boron-10 Thin-Film Targets The excimer scintillation experiment at Maryland utilized thin lms of isotopically enriched boron-10 as neutron absorbing targets. These lms were fabricated at the NIST Center for Nanoscale Science and Technology (CNST), by physical vapor deposition in an electron-beam evaporator. The lms were later characterized using various methods to determine their density, thickness, chemical content, physical structure, and long-term stability. These characterizations took place at both UMD and NIST. 5.1 Fabrication Boron lms have previously been fabricated by many methods including sputter- ing [104, 105, 106], chemical vapor deposition (CVD) [107, 108, 109], physical va- por deposition (PVD) [110], thermionic vacuum arc [111, 112], electroplating [113], pulsed-laser deposition [114, 115], and even synchrotron-radiation-induced deposi- tion [116]. Past research has examined a variety of boron-containing compounds (BN, B4C, MgB2), and a variety of substrates (Si, Fe, Al, NaCl, GaAs, diamond). In this project, boron lms were fabricated strictly by physical vapor deposition, us- 77 ing isotopic 10B of 92% enrichment. The lms were deposited on plain silicon (111), as well as silicon treated with an adhesion layer of titanium or chemically-deposited natural boron. Several sets of lms were fabricated across a range of thicknesses (0.3 - 2.5 m) to examine the e ects of absorber thickness on scintillation yield from the 10B(n; )7Li reaction. Physical vapor deposition encompasses a number of techniques that rely on physical processes to deposit thin- lms under vacuum. Electron-beam evaporation (EBPVD) was used to fabricate the 10B targets for the excimer scintillation experiment. During EBPVD, an electron beam heats a small amount of source material to the point of vaporization. The electron beam, the source, and the source crucible, reside in a vacuum chamber. A portion of the vaporized source material adheres to sample substrates, which are also situated in the vacuum chamber. Both the power of the electron beam, and the distance between the source and the substrates, determine the rate of deposition. A diagram of a typical EBPVD chamber appears in Figure 5.1. Before fabricating the boron lms, it was necessary to prepare substrates upon which the lms were deposited. The primary substrates were silicon wafer pieces, 2:54 2:54 cm, with thicknesses of 525 m. These pieces were scored from 10 cm wafers, using a dicing saw (Microautomation BKSIDE-02) in the UMD FabLab. Each wafer was cleaned with piranha etch (H2SO4 and H2O2), rinsed in deionized water, and dipped in hydro uoric acid to remove the native oxide from its surface. The wafers were then broken into squares and heat sealed into aluminized-boPET 78 Figure 5.1: Left: Diagram of an electron-beam evaporator. Redrawn from [117]. Right: Denton EBPVD tool used for boron lm fabrication [118]. pouches within a nitrogen glove box, to prevent their surfaces from reoxidizing. Additional silicon substrates were provided by Dr. Lis Nanver of the Delft University of Technology. These substrates, fabricated at Delft, were chemically deposited with a 2-nm lm of natural boron, providing an epitaxial surface for the enriched 10B to adhere to during the EBPVD process. The 10B lms were fabricated in the NIST CNST NanoFab clean room. The EBPVD tool (Denton In nity 22) is pictured in Figure 5.1. Prior to loading the substrates, the back of each sample was marked with a diamond scribe as a means of identi ca- tion. The samples were then mounted to holders and suspended in a carousel rack at the top of the vacuum chamber. Granular 10B source material was also loaded 79 into 25 cc graphite crucibles at the bottom of the chamber. The 10B source material is pictured in Figure 5.2. After evacuating the chamber to a pressure < 5:0 10 6 torr, the source material was premelted using the electron beam. A low-power premelt prevented boron fragments from dispersing in the chamber during the higher power depositions. An internal shutter blocked the substrates from boron vapor throughout this phase. Following the premelt, the current of the electron beam was increased according to a preset program, and the shutter was opened to begin the deposition. Both the rotation speed of the carousel, and the rastering pattern of the electron beam on the source, were tuned to achieve uniform deposits. A quartz crystal oscillator within the vacuum chamber was used to regulate the deposition rate, and to measure the cumulative deposit thickness. As the mass of deposited material on a quartz oscillator increases, its vibration frequency decreases. In this way, the thickness of each deposit was calculated using a preprogrammed material density, and known characteristics of the crystal. The power of the electron beam was limited to 3.5 kW during the depositions, and the acceleration voltage was held constant at 10 kV. The various program parameters of the EBPVD tool, during these depositions, appear in Table 5.1. The various sets of 10B lms will be referred to by their nominal thicknesses, as programmed into the deposition tool. A thermocouple attached to the carousel was used to determine the temperature of the samples within the chamber. By the end of the longer depositions, this temperature was typically 175 C. For comparison, the melting point of boron is 80 2075 C at atmospheric pressure [76]. Following the depositions, the samples were allowed to cool below 60 C before they were removed. The lms were then packaged in a desiccator, transported to a nitrogen- lled glovebox, and sealed in boPET pouches until further use. An image of one set of boron lms, immediately after deposition, appears in Figure 5.2. Table 5.1: Tool settings during the boron- lm depositions and titanium adhesion-layer depositions. Rise, soak, and predeposit parameters describe the rate at which the electron beam current is increased prior to the depositions. Electron-beam powers are listed as percentages of the 10-kW tool maximum. Material Parameter Titanium Boron-10 density (g/cm3) 4.50 2.54 acoustic impedance (g/cm2s) 14.06 22.7 rise time to soak (min) 2 4 soak time (min) 1 1 soak power (%) 12 14 rise time to predeposit (min) 1 1 predeposit time (min) 1 0.5 predeposit power (%) 13.5 16 max deposit power (%) 25 35 deposition rate ( A/s) 1 1/2.5/5 tooling factor (%) 136 136 Several sets of lms were fabricated across three visits to the CNST NanoFab. The rst group, fabricated in July 2011, included lm thicknesses of 1.5, 2.0, and 2.5 m, deposited on pure silicon. These lms su ered peeling and delamination from the substrates due to intrinsic stress. An image of the delaminated lms appears in Figure 5.3. Both the lack of literature on boron- lm fabrication > 1 m, and the experience of the sta at the CNST, indicate that the practical upper limit of boron- lm fabrication by EBPVD lies between 1 - 2 m. Thinner depositions were next attempted. 81 Figure 5.2: Left: Boron-10 source material. Right: 600-nm 10B lms mounted to the deposition holder. The second group of lms, fabricated in August 2011, included lm thicknesses of 300, 600, and 900 nm. These lms were deposited on silicon treated with a 10-nm adhesion layer of titanium. The use of this adhesion layer came by recommendation from the CNST sta , as a means to reduce stress. The titanium was deposited in the same EBPVD tool. The parameters of the titanium depositions appear in Table 5.1. This set of boron lms did not show visible signs of physical stress or delamination. Since the stress in the lms was naturally lower, as a result of their reduced thicknesses, the stress-reducing contribution of the titanium adhesion layer remains unknown. A third group of lms was fabricated in March 2012, at thicknesses of 1.2 and 1.5 m. These lms were deposited on the epitaxial substrates provided by Dr. Nanver. None of these lms delaminated, though the 1.5 m lms developed visible signs of 82 stress within a day of fabrication, in the form of bubbling and fracturing of the lm surface. Ultimately, only the lms which demonstrated prolonged physical stability (the 300, 600, 900 nm lms from August, and the 1200 nm lms from January) were used as targets in the excimer scintillation experiment. Figure 5.3: Delamination of 2.5 m lms, immediately after deposition. 83 5.2 Characterization Several characterizations of the boron thin- lm targets were performed to determine their neutron absorbing properties (content, thickness, density), and the stability of those properties over the course of the scintillation measurements. The charac- terization techniques included neutron imaging, pro lometry, x-ray photoelectron spectroscopy (XPS), x-ray di raction (XRD), scanning electron microscopy (SEM), ellipsometry, and stress measurement. 5.2.1 Neutron imaging The thermal-neutron absorption properties of the 10B thin lms were measured at the neutron imaging facility (NIF/BT2) at the NCNR. This facility uses an intense, collimated beam of thermal neutrons to obtain radiograph images from a spatially sensitive neutron detector. Neutrons from the NBSR pass through a cooled, single- crystal bismuth lter, a series of apertures, and an evacuated ight tube, before impinging on a sample, and a position-sensitive detector. A digram of the neutron imaging facility appears in Figure 5.4. The NIF detector consists of a 6Li conversion layer, a ZnS scintillation layer, and an x-ray imager (Varian PaxScan 2520) made of amorphous silicon. The imager has an active area of 19:3 24:2 cm and a pixel size of 127 m2. The uence of the neutron beam during the boron lm imaging was 5:31 106 cm 2s 1 at the sample position. A diagram of the detector and the sample location appears in Figure 5.5. 84 Figure 5.4: Plan view of the NCNR neutron imaging facility [119]. Each boron lm used in the scintillation experiments was placed in a holder and mounted at the sample position in the imaging station. A photograph of a boron sample and the sample holder appears in Figure 5.6. Upon opening the beam shutter, the neutron beam illuminated each sample and the detector collected a series of 1800 images, with 1-second exposures. This measurement was repeated, without a sample or a sample holder in the beam, to obtain a at- eld image. All of the images were then corrected for the point spread function (PSF) of the detector system with assistance from Dr. Daniel Hussey. The PSF is a systematic additive background, attributed to di use light in the scintillation screen of the detector [120]. Following the PSF corrections, the images of each sample (1800 per sample) were averaged to form a single image. By dividing the averaged at- eld image from the averaged image of each boron lm, the fractional absorption of each sample was determined. The resulting images appear in Figures 5.7 - 5.11. From the Beer- 85 Figure 5.5: Diagram of the NIF detector station. The detector is shown in an exploded view. Lambert law (see Eq. 2.2), the value of each pixel (P ) in these images represents fractional absorption, in the form, P = 1 I I0 = 1 e x (5.1) where I0 is the intensity of the incident neutron beam, I is the intensity of the neutron beam after passing through the sample, is the macroscopic cross section of the sample, and x is the thickness of the sample. The value of II0 is determined by dividing each pixel value in the sample image by the corresponding pixel value in the empty at- eld image. The colormap of each image in Figures 5.7 - 5.11 depicts the neutron absorption fraction between 0 - 10%. Rearranging Eq. 5.1, each pixel in the neutron image acquires a value, x = ln(1 P ) (5.2) 86 By averaging the values of ( x) over the surface of each lm, the areal density of the lms (( A)10) may be determined, with the equation, ( A)10 = ( x) h iNIF M10 NA (5.3) where ( x) is the average value of ( x) determined from the neutron image, h iNIF is the e ective cross section of 10B in the NIF neutron beam, M10 is the molar mass of 10B (10.0129 g/mol [76]), and NA is Avogadro?s number (6:022 1023 mol 1). The value of h iNIF may be determined with a continuous weighted average, in the form, h iNIF = R 10( ) ( ) d R ( ) d (5.4) where 10( ) is the neutron-wavelength-dependent, microscopic neutron-absorption cross section of 10B, and ( ) is the neutron-wavelength distribution of the NIF beam, normalized for each wavelength bin d d . The distribution d d was previously measured at the NIF with a neutron chopper and time-of- ight spectrometry. The neutron-wavelength distribution of the NIF beam appears in Figure 5.12. In solving Eq. 5.4, the neutron-wavelength parameter space was used, rather than the neutron- energy parameter space, to avoid ampli cation of noise in the fast and epithermal regions of the d d data. The value obtained from Eq. 5.4 is h iNIF = 3936:1 b. By comparison, the 2200 m/s cross section of 10B is 3842 8 b [77]. The thermal neutron absorption cross section of silicon is small ( 0 = 2:16 b), as demostrated by the nearly invisible substrate in Figure 5.11. Therefore, the absorption of each sample 87 was attributed completely to absorption by 10B. The values of ( A)10, obtained from Eq. 5.3 and 5.4, appear in Table 5.2. The uncertainties in ( A)10 are dominated by the uncertainty in h iNIF. Dr. Daniel Hussey has estimated that the uncertainties in the NIF measurements of d d , lie between 5-10%. A conservative 10% uncertainty has therefore been associated with the values of ( A)10 in the nal analysis in Chapter 9. However, the relative vari- ations (root mean square (RMS) values) in the measurements of ( x) are 1 %. Additionally, these RMS values are nearly identical for all of the measured samples (see Table 5.2). Consequently, the RMS values represent a statistical limit imposed by counting time or baseline noise of the NIF detector, and are not correlated to density variations in the 10B samples. They do, however, provide an upper limit for this density variation. Nominal Thickness (nm) ( x) RMS ( A)10 ( g/cm2) 300 0.01740 2:39 10 4 73.50 600 0.03321 2:39 10 4 140.31 900 0.04951 2:39 10 4 209.13 1200 0.06155 2:39 10 4 260.02 substrate 4:80 10 4 2:39 10 4 2.028 Table 5.2: Areal densities of the 10B lms obtained though the analysis of neutron imaging data. Nominal lm thicknesses indicate the input to the EBPVD deposition tool at the time of lm fabrication. Measurements of an epitaxial substrate, deposited with 2-nm of natural boron, are also listed. 88 Figure 5.6: Boron-10 lm and holder used for neutron imaging Figure 5.7: Thermal-neutron image of a 300-nm 10B lm. The colorscale of the image represents fractional absorption between 0 - 10%. 89 Figure 5.8: Thermal-neutron image of a 600-nm 10B lm. The colorscale of the image represents fractional absorption between 0 - 10%. Figure 5.9: Thermal-neutron image of a 900-nm 10B lm. The colorscale of the image represents fractional absorption between 0 - 10%. 90 Figure 5.10: Thermal-neutron image of a 1200-nm 10B lm. The colorscale of the image represents fractional absorption between 0 - 10%. Figure 5.11: Thermal-neutron image of a 2-nm layer of natural boron on a silicon substrate. The colorscale of the image represents fractional absorption between 0 - 10%. 91 Figure 5.12: Neutron-wavelength distribution of the NIF beam. Secondary peaks are the result of Bragg di raction due to the Bi-crystal lter. 92 5.2.2 Pro lometry A contact pro lometer (Bruker Dektak XT) located in the CNST NanoFab, was used to measure the thicknesses of the boron lms. This tool moves a stylus laterally across a sample to determine the pro le of its surface based on the vertical de ection of the stylus. During the evaporation process, the sample clips holding the substrates in the EBPVD tool shielded one corner of each substrate from the boron vapor. This resulted in a bare spot of silicon on each sample, across which each boron lm thickness was measured. The clip silhouettes are visible in the neutron images in Figures 5.8 - 5.10. For each boron sample, several scans of the silicon-boron step were performed at di erent angles and positions. The parameters of these scans appear in Table 5.3. Typical data from the scans are plotted in Figure 5.13. The step heights and rough- nesses of each lm were calculated using the Vision 64 software, associated with the pro lometer. These outputs were then averaged across several scans of the same sample. The average step heights are within 10% of the nominal lm thicknesses, as programmed in the deposition tool. Two di erent parameters for roughness were determined, Ra and Rz. The value Ra is the average variation of all data points in a scan from the mean of those data points. The value Rz is the average of the di erences between the ve highest peaks and valleys in a scan. The results appear in Table 5.4. 93 Table 5.3: Pro lometer scan parameters. Parameter Value Scan length 400 m Lateral scan resolution 0.044 m Stylus force 0.9 mgf Stylus radius 0.7 m Vertical range 6.5 m Vertical resolution 1.0 A Table 5.4: Film thicknesses and roughnesses from pro lometry. h = average step height. h = relative di erence between h and nominal thickness. Nominal thickness (nm) h (nm) h (%) Ra (nm) Rz (nm) 300 278.4 7.2 6.44 42.1 600 586.2 2.3 5.01 38.7 900 889.0 1.2 3.64 39.6 1200 1151 4.0 3.86 23.6 Figure 5.13: Pro lometer scans of several 10B thin- lm targets. 94 5.2.3 X-ray photoelectron spectroscopy X-ray photoelectron spectroscopy (XPS) is a technique used to determine the chem- ical composition of surfaces. In XPS, a sample is placed under vacuum and ir- radiated with a collimated beam of soft x-rays. Atoms within 5 - 10 nm of the target surface are ionized by these x-rays, and the resulting electrons are focused towards an electron-energy spectrometer. The spectrometer eliminates all but a narrow passband of electrons, according to their kinetic energies. The mean energy of this passband is tunable to allow scanning across the electron-energy spectrum. After passing through the spectrometer, electrons within the passband impinge on an electron detector. Together, the analyzer, the detector, and the associated elec- tronics measure the distribution of ionized electrons as a function of their kinetic energies. The kinetic energies of the electrons (EKE) impinging on an XPS detector are given by the equation, EKE = Eh E EBE (5.5) where Eh is the energy of the incident x-rays, E is the work function of the spectrometer, and EBE is the electron binding energy of the sample. Since Eh and E are known, and the distribution of EKE is measured, the distribution of EBE may be derived from this equation. Electron binding energies are characteristic of speci c materials, and they are accurately tabulated for a large number of elements and compounds. A distribution of EBE may be used to identify the chemical composition 95 of a sample surface. A diagram of a typical XPS setup appears in Figure 5.14. Figure 5.14: Left: Schematic of an XPS tool. Redrawn from [121]. Right: Kratos AXIS 165 used for boron lm analysis [122]. Prior to the excimer scintillation experiment, XPS was used to investigate (1) po- tential oxidation of the boron lms, (2) sources of nitrogen and oxygen impurities in the lms, and (3) the depth pro les of these impurities. The measurements were performed at the MRSEC, with assistance from Dr. Karen Gaskell. The XPS tool (Kratos AXIS 165) uses a monochromatic (1.486 keV) aluminum x-ray source and a 165-mm concentric hemispherical analyzer. A photograph of this tool appears in Figure 5.14. The base vacuum pressure for these measurements was 3 10 9 torr. All measurements were collected at a 90 takeo angle (electron emission angle with respect to the sample surface). In the rst set of measurements, four boron lm samples were analyzed for surface oxidation. Three of these boron samples (300, 600, 900 nm thicknesses) were stored on a workbench, open to atmosphere, for 3.5 months after fabrication. The remain- ing boron sample (300 nm) was stored in a nitrogen- lled glove box over the same 96 period. The concentrations and peak centroids derived from the EBE distributions, appear in Table 5.5. For comparison, binding energy peak centroids and widths for various boron compounds appear in Figure 5.15. The EBE distributions for both 300 nm lms (stored in N2, and stored in atmosphere) appear in Figures 5.16 and 5.17. The boron peaks isolated from these distributions appear in Figure 5.18. As shown in Figure 5.15, the binding energy peak of boron-oxide (B2O3) is quite distinct from the elemental boron peak. The EBE distributions, obtained from the XPS measurements, show no signs of a minor B2O3 peak from any of the boron samples. The isotopic enrichment of the samples could contribute to the shift of the major boron peak (187.5 eV) from the tabulated value (189.5 eV). While oxygen is apparent in the spectra (531.7 eV), it does not arise from metal oxides (529.5 eV). Rather it most likely comes from hydroxides (531.5 eV). Additionally, a small amount of nitrogen appears in all of the samples (400.3 - 401.5 eV), most closely matching the EBE of ammonium salts (400.5 - 403 eV). The carbon peak arises from adventitious carbon, a contaminant found on nearly every freshly exposed surface [123]. Ultimately, no evidence of oxidation was observed in any of the samples, however, the presence of oxygen and nitrogen in the lms was further investigated with XPS. 97 A second set of measurements was performed to analyze the 10B source material, the melted source material from the EBPVD chamber, and the crucible that contained the melted boron. All three of these samples were shown to contain both oxygen and nitrogen, in amounts comparable to the boron lms. The concentrations and peak centroids from these spectra appear in Table 5.6. Figure 5.15: Typical EBE and XPS peak widths of boron compounds [124]. To determine the depth pro le of nitrogen and oxygen throughout the boron lms, one of the samples (600 nm) was sputtered with an argon ion gun (50.84 A/cm2) within the XPS chamber. Sputtering removes atoms from the surface of a target by momentum exchange between an energetic ion and an atom in the material. During this process, XPS measurements were collected every 300 s. The sputtering rate of boron was not found in the literature, therefore the ultimate sputter depth is un- known. The measured concentrations are plotted, as a function of sputtering time, in Figure 5.19. Clearly, the nitrogen, oxygen, and carbon, found in the earlier measure- ments, exist predominantly at the sample surface. After only 300 s of sputtering, 98 the concentrations of these contaminants were reduced to < 2%. At 6000 s, the sputtering was stopped and the sample was allowed to sit overnight in the vacuum chamber at a base pressure of 3 10 9 torr. When measurements were resumed, the same contaminants reappeared. Even in ultrahigh vacuum, trace amounts of the contaminants adhered to the sample surface, demonstrating an a nity of these compounds for boron. Table 5.5: Concentrations (%) and EBE peak centroids (eV) of surface contaminants in boron lms. Sample Boron Nitrogen Oxygen Carbon 300 (N2) 78.94/187.7 0.44/401.5 7.03/531.8 13.58/284.8 300 77.86/187.5 0.72/400.5 8.18/531.7 13.24/284.8 600 78.12/187.5 0.61/401.2 8.53/531.8 12.74/284.8 900 75.31/187.5 0.81/400.3 7.66/531.7 16.22/284.8 Table 5.6: Concentrations (%) and EBE peak centroids (eV) of surface contaminants in boron source materials. Sample Boron Nitrogen Oxygen Carbon Source granule 70.30/187.8 1.06/400.1 11.20/531.8 17.44/284.8 Boron melt 54.84/187.8 3.63/402.0 19.70/532.1 21.83/284.8 Crucible wall 65.76/187.5 0.42/401.6 12.56/531.9 21.26/284.8 99 Figure 5.16: EBE distribution for a 300-nm boron lm stored in atmosphere. KLL peaks originate from Auger electrons. Figure 5.17: EBE distribution for a 300-nm boron lm stored in N2 glovebox. KLL peaks originate from Auger electrons. 100 Figure 5.18: Boron (1s) electron peaks from XPS. Left: 300-nm lm stored at atmosphere. Right: 300-nm lm stored in an N2 glovebox. Figure 5.19: Contaminant concentrations in the 10B lms during sputtering. Sputtering and XPS measurements were stopped automatically at 6000 s and were resumed 8 hours later. 101 5.2.4 X-ray di raction The crystallinity of the boron lms was analyzed with an x-ray powder di ractome- ter in the X-ray Crystallography Lab at Maryland. X-ray di raction relies on the specular scattering of x-rays from the atoms in a crystal lattice. Bragg?s law is used to describe the conditions for constructive interference between photons re ected from two successive crystal planes, and is given by the equation, 2d sin = n (5.6) where d is the spacing between the di racting planes, is the angle of incidence of the x-rays, n is an integer, and is the x-ray wavelength. A diagram of Bragg scattering appears in Figure 5.20. By irradiating a sample with a collimated beam of x-rays, a scattered interference pattern may be formed on a detector opposite the sample. This pattern is dependent on . By scanning over , it is possible to determine the crystal structure of a sample based on Bragg peaks in the interference pattern. In powder di raction, x-rays are scattered from a sample of randomly oriented microcrystals. The resulting interference pattern takes the form of Debye- Scherrer cones, which create rings on a at plate detector. These rings may be used to determine the lattice spacing of a crystal [125]. The di ractometer used to analyze the boron lms was a Bruker Apex2, equipped with a CCD area detector. The only peaks visible in a scan of a 600 nm sample arise from the silicon substrate. The dominant peaks from tetragonal boron (2 = 20:3 ) 102 and alpha-rhombohedral boron (2 = 21:9 ) do not appear, indicating that the boron lms were amorphous. XRD spectra from numerous crystalline forms of boron are discussed in [126]. The XRD spectrum from the 600-nm sample, showing only characteristic silicon peaks, appears in Figure 5.21. Figure 5.20: Bragg scattering from a crystal lattice. Figure 5.21: X-ray di raction spectrum from a 600-nm boron lm. Prominent Bragg peaks result from x-ray scattering in the silicon substrate. 103 5.2.5 Scanning electron microscopy Images of the boron lms were taken using a scanning electron microscope (SEM), in the Maryland Nanoscale Imaging Spectroscopy and Properties Laboratory (NISP). These measurements were used to examine the boron lm surfaces, and to verify the thicknesses of the lms. In an SEM, electrons are thermionically emitted from a cathode lament and focused into a beam by a series of electron lenses. When this beam impinges on a sample, it produces secondary electrons, back-scattered electrons, x-rays, and light. Depending on the tool, one of these signals is collected with a detector. By rastering over the sample surface with the electron beam, a 2D image of the sample is obtained. A diagram of an SEM appears in Figure 5.22. The tool used to measure the boron lms (Hitachi SU-70) utilizes a ZrO/W Schottky emission electron source, and a silicon drift detector. Images from measurements of a 600 and 900 nm sample appear in Figures 5.23 and 5.24, respectively. The images include surface scans at magni cations between 1 mm and 1 m (50x - 10,000x), and pro le scans at a magni cation of 5 m. To obtain clean pro le scans, small samples (25 mm2) were cut from the larger lms with a diamond scribe. The cross- sectional images of the lms show a narrow layer of boron between adhesive tape and the silicon substrate. The thicknesses derived from these pro le images were 594 nm and 892 nm, a 1% di erence from the nominal thicknesses of the lms (600 nm and 900 nm, respectively). 104 Figure 5.22: Diagram of a scanning electron microscope [127]. 105 Figure 5.23: SEM images of a 600-nm boron lm. Pro le images appear on the bottom right. 106 Figure 5.24: SEM images of a 900-nm boron lm. Pro le images appear on the bottom right. 107 After a series of characterizations, the boron lm targets were determined to be amorphous, stable to oxidation, and representative of the thicknesses programmed into the deposition tool. While surface contaminants (oxygen and nitrogen con- taining compounds) were found on both the lms and the source materials, these contaminants were demonstrated to be predominantly absent from the bulk of the deposits. The areal densities derived from the neutron imaging data will be used in Chapters 7 and 9 to determine the rate of neutron capture during the excimer scintillation measurements. 108 Chapter 6 Photomultiplier Tube Calibration The intrinsic e ciency of the photon detector package, used in the excimer scintil- lation experiment, was measured at the Synchrotron Ultraviolet Radiation Facility (SURF) at NIST. This package consisted of the solar-blind PMT, the PMT hous- ing, the di erential volume, and the discrete MgF2 window. The calibration focused speci cally on the FUV region (130 - 210 nm) common to the excimer continua of the heavy noble gases. The intrinsic e ciency of the detector was required to derive absolute yield values from the scintillation measurements in Chapter 9. 6.1 The Synchrotron Ultraviolet Radiation Facility The Synchrotron Ultraviolet Radiation Facility (SURF III) is a compact electron storage ring located on the NIST campus in Gaithersburg, MD. SURF is a stable source of synchrotron light (see Fig. 6.1), ranging from the infrared to the soft x-ray regions, with a peak output in the extreme ultraviolet [128, 129]. Whereas conven- tional light sources (gas discharge, arc lamp, plasma) have limited spectral ranges plagued by line structure, synchrotron radiation is continuous, providing an excellent source for photodetector calibrations [130, 131]. Furthermore, the characteristics of 109 synchrotron light are easily controlled. Variation of synchrotron electron-beam en- ergy changes the resulting photon spectrum. Variation of synchrotron electron-beam current changes the photon-beam intensity. SURF electron energies range between 100 and 400 MeV, and electron beam currents range from 10 pA (1 electron/s) to 1000 A (1011 electrons/s). The wavelength-dependent radiant power of SURF at various operating energies appears in Figure 6.2. Figure 6.1: As ultrarelativistic electrons accelerate radially in a magnetic eld, they emit a collimated beam of photons, known as synchrotron radiation, perpendicular to the di- rection of acceleration [132]. Figure 6.2: SURF III radiant power at various operating energies [131]. Also shown are the output spectra from a 30 W deuterium lamp and a 3000 K blackbody. 110 Tangential beamlines built around synchrotrons and storage rings direct photon radiation to experiments. SURF hosts a total of thirteen beamlines. Beamline 4 (BL-4), used to calibrate the PMT, was constructed primarily for UV detector calibrations and UV optical material characterization. A schematic of BL-4 appears in Figure 6.4. After light leaves the SURF storage ring and enters BL-4, two grazing-incidence, fused-silica mirrors image the beam onto the entrance slit of a two-meter, normal- incidence monochromator. This monochromator consists of a normal-incidence mir- ror and a curved grating with 600 lines/mm. It eliminates all but a small passband of the incident radiation (0.7 nm, at a wavelength of 200 nm), enabling precision wavelength-dependent measurements. Spectral scans are performed by rotating and translating the monochromator along the bisector of the angle between the entrance and exit slits [133]. Beyond the exit slit, monochromatic light is reimaged in the mir- ror box by two Al-MgF2 mirrors, before propagating into the detector box containing an x-y linear motion stage and the detector under test. The PMT detector package was mounted to this detector box for calibration against an absolutely-calibrated photodiode. 111 Figure 6.3: The Synchrotron Ultraviolet Radiation Facility at NIST. BL-4 appears second from the right [134]. Figure 6.4: Diagram of BL-4 at SURF. The photon beam path is highlighted in ma- genta. Positions of the photodiode and the photomultiplier tube during the calibration are labeled. Redrawn from [130]. Not to scale. 112 6.2 Calibration Apparatus In addition to the beamline components discussed in Section 6.1 and the associated beamline electronics, the calibration setup included the solar-blind PMT (Hama- matsu R6835), a discrete MgF2 window, a di usion-type np-junction photodiode (IRD AXUV-100G), a manual x-y translation feedthrough, a number of vacuum adapters, and the electronics necessary for PMT pulse counting (fast preampli er, power splitter, multichannel analyzer, microcontroller, counter/timer, and high- voltage power supply). These electronics and their settings were identical to those used in the scintillation yield measurements, described in Chapter 4.2.4. The rele- vant speci cations of the photomultiplier tube appear in Table 6.1. Parameter Value Spectral response 115 - 200 nm Photocathode material Cs-I Photocathode diameter 23 mm Number of dynodes 11 Anode dark current 0.03 nA Anode pulse rise time 2.8 ns Electron transit time 22 ns Table 6.1: Speci cations of the Hamamatsu R6835 photomultiplier tube [103]. During the scintillation experiment at the MUTR, the PMT was attached to the scintillation cell behind an evacuated di erential volume and an MgF2 window. These components were con gured identically during the PMT calibration. The resulting intrinsic e ciency measurements account for photoabsorption by the dis- crete MgF2 window, photoabsorption by the MgF2 window coupled to the PMT housing, the quantum e ciency of the PMT, and the thresholds of the counting 113 electronics. These components were mounted on an x-y translation feedthrough to allow scanning of the photon beam over the surface of the detector. This xture was then mounted behind the BL-4 detector box. The PMT calibration xture appears in Figure 6.5. Figure 6.5: The PMT calibration xture. From right to left: PMT (black housing), dif- ferential volume, MgF2 window, x-y translation stage, vacuum adapters. The photodiode and the BL-4 detector box are not shown. The photodiode was placed upstream of the PMT within the BL-4 detector box, on a remotely-operated x-y linear-motion stage. This photodiode was previously ab- solutely calibrated against measurements from a cryogenic radiometer. The active surface of the photodiode (1 cm2) is composed of n-type silicon. Incident photons with an energy equal to or greater than the silicon bandgap energy (1.1 eV) may ex- cite electrons within this active area into the conduction band. Following excitations near the n-p junction, charge ow is induced by an internal electric eld. Photocur- 114 rent may then be measured with an ammeter in an external circuit to determine the photon ux incident on the detector. A schematic of an np-type photodiode appears in Figure 6.6. With the PMT and the photodiode in place, the detector box, the mirror box, and di erential volume were evacuated to high vacuum (10 4 mbar) to prevent photoabsorption by O2. Figure 6.6: Schematic of an np-type VUV photodiode [135]. 6.3 Measurements Due to the steady decay of the SURF electron-beam current after initial beam injection, high beam-current photodiode measurements were made before dumping electrons to collect PMT measurements at lower beam currents. Using the x-y stage within the detector box and one of the BL-4 ammeters, a spatial scan of the photodiode was performed to verify its uniform response, and to center its active area on the photon beam. The incident photon ux was measured over the wavelength range 130 - 210 nm, in 5-nm increments, at electron beam currents of 50, 40, 30, 20, and 10 mA. These measurements were used to determine a linear relationship 115 between the electron-beam current in the SURF storage ring and the photon ux in the BL-4 detector box at each wavelength. Both the spatial position of the photodiode and the wavelength selection of the monochromator were controlled and automated by LabView. Following these measurements, the SURF electron-beam current was reduced to 10, 5, and 1 A, successively. At these lower beam currents, the photon ux in the BL-4 detector box is within the operating range of the PMT. The photodiode was then removed from the photon beam, allowing the beam to strike the PMT detector package. Spectral scans over the same wavelength region were repeated at these reduced currents, and the PMT signal was measured using the associated electron- ics. During operation at 10 A, a spatial scan of the detector was performed at a single wavelength (170 nm) using the x-y translation feedthrough. Under the same conditions, counts were recorded at a number of PMT voltages between -1700 and -2500 V. The results of these measurements appear in the following section. 116 6.4 Results and Analysis The intrinsic e ciency of the PMT/MgF2 detector package ( pmt( )) is de ned by the equation, pmt( ) = nh ( ) Nh ( ) th (6.1) where nh ( ) is the number of photon pulses observed from the PMT at wavelength , Nh is the rate at which photons strike the PMT/MgF2 detector package at wavelength , and th is the time over which PMT pulses were recorded. The number of photon pulses observed from the PMT is given by the equation, nh = nh total nh bkgd (6.2) where nh total is the total number of pulses counted over time th , and nh bkgd is the number of those pulses which are attributed to backgrounds (dark current). Measurements of nh total were recorded by scanning over the FUV wavelength region on SURF BL-4, and counting pulses from the PMT at increments of 5 nm. A single measurement of nh bkgd was taken prior to the nh total measurements, while the BL-4 shutter was closed. A plot of nh ( ) at a SURF electron beam current of 10 A appears in Figure 6.7. 117 The uncertainty in nh arises strictly from counting statistics, and may be deter- mined with the equation, nh = q 2nh total + 2 nh bkgd (6.3) The standard deviation ( ) of a counting measurement with a value n and a Gaussian parent distribution is given by, = p n (6.4) Therefore, nh = p nh total + nh bkgd (6.5) Accordingly, the relative uncertainty in nh is given by, nh = p nh total + nh bkgd (nh total nh bkgd) (6.6) A plot of the relative statistical uncertainties in nh ( ) appears in Figure 6.8. 118 Figure 6.7: Net counts from the PMT during an FUV spectral scan at SURF. Electron beam current Ibc = 10 A. Count time t = 10 s. Figure 6.8: Relative statistical uncertainties in nh ( ) from a spectral scan of the PMT at SURF. Electron beam current Ibc = 10 A. Count time t = 10 s. 119 Values of Nh ( ) were extrapolated from photodiode measurements of the photon ux in the SURF BL-4 detector box. The photon ux at each wavelength ( ) in the region of interest, was determined at several SURF electron-beam currents (10-50 mA) within the operating range of the photodiode. These values were calculated with the equation, Nh ( ) = Ipd( ) pd( )h e (6.7) where Nh is given in units of s 1, Ipd( ) is the average current from the photodiode at wavelength , pd( ) is the responsivity of the photodiode at wavelength , h is the energy of a photon of wavelength (and frequency ), and e is the electron charge (1:60218 10 19 C or J/eV). The BL-4 software automatically records photodiode dark current at the beginning of each data run, and subtracts this value from all measurements of Ipd. The responsivity of the photodiode over the FUV region, appears in Figure 6.9. After calculating values of Nh at each wavelength and several SURF electron-beam currents, the values were t by linear regression to determine the relationship be- tween the SURF electron-beam current and the BL-4 photon ux at each wave- length. This relationship takes the form, Nh ( ; Ibc) = k( ) Ibc (6.8) 120 where k( ) is the ux coe cient (slope of the t) at each wavelength ; and Ibc is the SURF electron beam current. Values of k( ), calculated from the photodiode measurements, are plotted in Figure 6.10. The uncertainties in the ux coe cients were derived from the mean square error of each regression. A linear t (when forced through zero) is given by, y^i = kxi (6.9) where y^i are the tted values of the dependent variable, xi are the explanatory values of each measurement, and k is the t coe cient (slope). The residuals of the t are given by, yi y^i (6.10) where yi is the measured value at xi. The sum of the squared residuals is then, SSR = NX i=1 (yi y^i) 2 (6.11) Accordingly, the mean square error of the regression is given by, ^2 = SSR N 2 (6.12) 121 where N is the number of measurements. Finally, the standard deviation in the slope parameter of the t is given by, k = ^ 2 s 1 P (xi x)2 (6.13) where x is the mean of the explanatory values. The relative uncertainties in k( ) are plotted in Figure 6.11. The uncertainties are greater at shorter wavelengths due to the reduced throughput of the monochromator in this region. Figure 6.9: Responsivity of an absolutely-calibrated IRD AXUV-100G photodiode ( pd( )) in the FUV region [136]. 122 Figure 6.10: Photon ux coe cients (k( )) of SURF BL-4 over the FUV region. Figure 6.11: Relative uncertainties in k( ). 123 By extrapolating the regressions from Eq. 6.7 to electron-beam currents in the mi- croampere range, values of Nh ( ) were determined for each PMT measurement. The intrinsic e ciency of the PMT package was then obtained by substituting nh ( ) and Nh ( ) into Eq. 6.1. A plot of pmt( ) appears in Figure 6.12, at three electron-beam currents. For comparison, the spectral response curve supplied by the manufacturer appears in Figure 6.13. The relative uncertainty in pmt( ) was calculated by propagating the uncertainties in nh and k( ), in the form, pmt = s nh nh 2 + k k 2 (6.14) The relative uncertainties in pmt( ) are plotted in Figure 6.14. Figure 6.12: Intrinsic e ciency of the PMT/MgF2 detector package in the FUV region. Fits used in Eq. 6.15 are also shown. 124 Figure 6.13: Spectral response of the PMT supplied by the manufacturer. Figure 6.14: Relative uncertainties in pmt( ). 125 Because excimer emissions occur over broad continua, an e ective PMT e ciency was determined for each noble-gas type used in the scintillation experiments. Each e ective e ciency was calculated with a continuous weighted average, in the form, h pmti = R t( ) s( ) d R s( ) d (6.15) where t( ) is a cubic spline t of the discrete values of pmt( ), and s( ) is the wavelength distribution of the noble-gas excimer. The ts ( t( )) appear in Figure 6.12. Values of s( ) were obtained by digitizing excimer emission spectra from [36]. These digitized spectra appear in Figure 1.2. Values of h pmti for Ar, Kr, and Xe appear in Table 6.2. For comparison, values of the 10 A spline t ( t( )) at the discrete peak excimer emission wavelengths of each noble gas, also appear in Table 6.2. The uncertainties in h pmti were derived from the average value of the relative uncertainties in pmt( ), h pmti = pmt( ) = 3:49% (6.16) Gas peak pmt( peak) (%) h pmti (%) Ar 128 2:204 0:108 1:648 0:058 Kr 150 3:094 0:047 3:144 0:110 Xe 175 1:674 0:033 2:607 0:091 Table 6.2: List of (1) intrinsic e ciencies of the PMT/MgF2 package at peak excimer wavelengths, and (2) e ective intrinsic e ciencies as determined by Eq. 6.15. The results from the spatial scan of the PMT detector package appear in Figures 6.15 and 6.16. These measurements were performed at an electron-beam current of 10 A. The SURF electron-beam current-lifetime product is nominally 1.5 A 126 hr. Therefore, in the microampere range SURF provides an incredibly stable light source, with a lifetime of 17.7 years at 10 A. The calibration xture and the port it was mounted to were not directly centered on the photon beam. Therefore, only half of the active detector area was measured due to the limits of travel of the x-y translation stage (2:5 2:5 cm). The detector response is shown to be relatively uniform over an 18-mm diameter, with a tailing response out to the manufacturer- speci ed 23-mm diameter. This reduced response towards the edges of the detector package could be the result of irregularities in either MgF2 window at their periphery, or an indication of a reduced PMT detection e ciency for photons that strike the edges of the photocathode. The spatial response of the detector package will be addressed by the collection e ciency modeling in Chapter 8. The results from the voltage scan appear in Figure 6.17. This scan was also per- formed at an electron-beam current of 10 A. Ideally, the PMT should be operated in a at region of the voltage curve. Small changes in detector voltage in this plateau region, result in only small changes in the detector response. During both the spec- tral calibration of the PMT and the excimer scintillation experiment, the PMT was operated at -2300 V, just below the maximum rating of -2500 V. 127 Figure 6.15: Spatial response of PMT/MgF2 detector package. Beam current Ibc = 10 A. = 170 nm. Count time t = 2 s. Figure 6.16: Spatial response of PMT/MgF2 detector package. Ibc = 10 A. = 170 nm. Count time t = 2 s. 128 Using SURF as a stable source of FUV synchrotron light, the e ciency of the PMT/MgF2 detector package was determined for the continua of heavy noble-gas excimer emissions (1.6 - 3.1%). Additionally, a suitable detector voltage was identi- ed (-2300 V) for scintillation photon counting. Lastly, the spatial uniformity of the detector was characterized for interpretation in the collection e ciency modeling of the scintillation cell. Figure 6.17: PMT response as a function of operating voltage. Ibc = 10 A. = 170 nm. Count time t = 10 s. 129 Chapter 7 Neutron Beam Characterization A thermal-neutron beamline did not exist at the MUTR prior to the excimer scintil- lation experiment. To obtain a thermal-neutron beam, it was necessary to construct and install a collimator insert in the MUTR thermal column. Subsequently, this beam was characterized in several ways to determine the neutron interaction rate in the excimer scintillation cell. 7.1 Neutron Beam Collimator The MUTR core hosts ve experimental facilities, shown in Figure 4.1. Of these facilities, the thermal column is the best suited for thermal-neutron experiments due to the large volume of graphite (1.50 m) that separates its access cavity from the core. Neutrons di using through this graphite are e ciently thermalized by scattering interactions with carbon nuclei. The cavity beyond the graphite volume allows access to these thermal-neutrons for experimental irradiations. Before the scintillation experiment, a collimator insert was fabricated for the thermal- column access cavity. This insert allows only a narrow beam of neutrons (5.08 cm diameter) to stream past the reactor containment. The use of a collimated beam 131 during the excimer scintillation measurements enabled operation of the experiment at close proximity, and precise control over the number of neutrons interacting within the apparatus. The collimator insert was designed using the original MUTR blueprints, and mea- surements from an existing insert, constructed by Ian Gi ord. A carbon-steel frame for the insert was fabricated at the UMD Aerospace Machine Shop. The frame consists of 1=2" steel plate, joined with nickel-free welding rods. A beam tube at the center of the frame provides a void through which neutrons pass. The outer face of the frame was tapped with a matrix of threaded holes for a xing a beam-shutter assembly. The top plates on both sections of the frame were attached with hex bolts so that the frame could be lled with shielding material. The design and dimensions of this frame appear in Figure 7.1. Photographs of the frame appear in Figures 7.2 and 7.3. Figure 7.1: Collimator insert design and dimensions. Left: Outer face. Center: Side view. Right: Inner face. The innermost and outermost sections of the collimator frame were lled with a mixture of para n wax, steel shot, and boron carbide (B4C) powder. This ll mix- ture attenuates gamma-ray and neutron radiation outside of the beam tube. Prior 132 to lling, the seams of the frame were sealed with room-temperature vulcanizing (RTV) silicone to prevent leakage of the molten wax. The frame was then wrapped with heating tape and berglass insulation. During the lling, a heat gun was used to constantly heat the mixture, allowing the slurry of wax and B4C to ll the spaces in the steel shot. Approximately 140 kg of steel shot was used to ll the outer segment, and 70 kg was used to ll the inner segment. The volume fraction of the ll material is estimated to be 50% shot, 15% B4C, and 35% wax. A sample of the ll mixture appears in Figure 7.4. The lled collimator frame appears in Figure 7.5. A beam shutter, composed of Boral and lead, was a xed to the outer face of the collimator with the matrix of tapped holes. This shutter was used to block radiation leaving the beam tube. The shutter was held in place by an aluminum cradle, mounted on extruded aluminum tracks. The cradle was connected to a toggle lever by pulleys and wire rope, allowing manipulation of the shutter position from the experimental oor. Additionally, a microswitch was installed beneath the shutter seat to illuminate a safety light during beam operation. Photographs of the shutter assembly appear in Figure 7.6. Following the construction and installation of this collimator, the radiation dose rates in the vicinity of the thermal column were measured to verify the safety of nearby experimenters. The neutron beam was then characterized for use in the scintillation experiments. 133 Figure 7.2: Thermal-column collimator insert. Figure 7.3: Collimator insert frame, showing the section partitions and the beam tube. 134 Figure 7.4: Sample of the collimator ll mixture, composed of wax, steel shot, and boron carbide powder. Figure 7.5: Collimator insert lled with shielding mixture. 135 Figure 7.6: Collimator shutter assembly in the closed (left) and open (right) positions. 136 7.2 Fluence Measurements During the excimer scintillation experiment, a NIST-calibrated ssion chamber was used to continuously monitor the uence of the thermal-column neutron beam. For all scintillation measurements, the MUTR was operated in automatic mode, at a power level of 250 kW. The ssion chamber was placed in front of the scintillation cell and behind a beam aperture (see Fig. 4.9). The aperture was used to reduce the pro le of the neutron beam to a size (12.6 mm2) signi cantly smaller than both the ssion chamber deposit (127 mm2) and the boron targets (625 mm2). Pulses from the ssion chamber were collected with an MCA, as described in Chapter 4.2.4. A photograph of the detector appears in Figure 7.7. Figure 7.7: The ssion-chamber beam monitor. The rate at which a thin target absorbs a beam of neutrons may be derived from Beer-Lambert?s Law of neutron absorption, F = I0(1 e ax) (7.1) 137 where I0 is the intensity of the incident neutron beam, a is the macroscopic cross section of the sample, and x is the thickness of the sample. For ax 1, Eq. 7.1 may be approximated by, F = I0 (1 (1 ax)) = I0 a x (7.2) Futhermore, a x = N a x = A NA M a (7.3) where N is the atom density of the target, a is the microscopic absorption cross section of the target, A is the areal mass density of the target, NA is Avogadro?s number (6:0221 1023 mol 1), and M is the molar mass of the target material. Thus, Eq. 7.2 becomes, F = A NA M a I0 (7.4) By rearranging Eq. 7.4, the rate at which neutrons propagate through the scintil- lation cell (Nbeam) may be determined based on the ssion chamber measurements (nn), with the equation, Nbeam = nn tn M235 h 235i ( A)235NA (7.5) where Nbeam is the beam uence in units of Hz, tn is the length of each neutron count period, M235 is the molar mass of 235U (235.0439 g/mol), h 235i is the e ective microscopic absorption cross section of 235U in the MUTR thermal-column beam, 138 ( A)235 is the areal density of the ssion chamber deposit (458.073 g/cm2), and is a self-absorption correction factor for the ssion chamber deposit (1.03756). The factor accounts for ssions in the 235U deposit in which the ssion fragments are absorbed before ionizing the P-10 gas. The value of this factor was provided by Dr. Je rey Nico. The mass of the ssion chamber deposit and its uncertainty were provided by Dr. David Gilliam. Likewise, Eq. 7.4 may be used to determine the rate at which neutrons in the MUTR thermal column beam are absorbed by a boron thin- lm target, Nn = Nbeam h 10i ( A)10NA M10 p 2 (7.6) where h 10i is the e ective microscopic absorption cross section of 10B in the MUTR thermal-column beam, ( A)10 is the areal density of the boron thin- lm targets, M10 is the molar mass of 10B (10.0129 g/mol), and is the fraction of the neutrons in the beam which are not absorbed between the ssion chamber and the boron target. The areal densities of the boron targets were derived from the neutron imaging data, discussed in Chapter 5.2.1. The path length of the neutron beam through the boron targets was increased by tilting the targets at a 45 angle. The factor p 2 is included to account for this increased path length. 139 The absorption rate of the boron targets may be determined directly from the ssion chamber measurements, by combining Eq. 7.5 and 7.6, Nn = nn tn h 10i ( A)10M235 h 235i ( A)235M10 p 2: (7.7) The net counts from each ssion chamber measurement (nn) is given by the equa- tion, nn = nn total nn bkgd (7.8) where nn total is the total number of pulses counted over time tn, and nn bkgd is the number of those pulses attributed to non-neutron events. Values of nn were deter- mined by identifying a threshold just above the noise peak in each ssion chamber pulse height distribution, and subtracting the counts below that threshold from each measurement of nn total. Typical pulse height distributions (PHDs) from the ssion chamber appear in Figure 7.8. The noise peaks in these PHDs are highlighted in black. Plots of nn over each data collection day appear in Figures 7.9 - 7.12. Due to the need for su cient counting statistics, the time base of the neutron-count period was set longer than the time base of the photon-count period, tn = 1200 s and th = 200 s. Neutron-count periods often overlapped multiple photon-count periods, and sometimes ended or started in the midst of photon-count periods. Over the course of the reactor day, measurements of nn typically decreased by 5-10%, as shown in Figures 7.9 - 7.12. This decrease in the count rate (and the neutron beam uence) was the result of (1) spectrum hardening due to the gradual increase in 140 the temperature of the pool water and the reactor containment, and (2) ux pro le changes in the core due to movements of the reactor control rods. Overcoming ssion poisons (negative reactivity insertion) throughout the course of the reactor day requires the addition of positive reactivity, in the form of pool water heating and control rod withdrawal. Because these additions are gradual, the changes in nn throughout the day were also assumed to be gradual. To determine a value of nn corresponding to each photon count, the measurements of nn were t by linear regression of the form, n^n(t) = knt+ n^0 (7.9) where n^n(t) are the tted values of the neutron count measurements over time t, kn is the slope of the t, t is the starting time of of each photon count during the data collection day, and n^0 is the tted intercept of the dependent variable, relative to some pivot date-and-time, t = 0. The same pivot date was used throughout the analysis of nn. The linear ts (n^n(t)) appear in Figures 7.9 - 7.12. The uncertainties in n^n(t) were rst determined by propagating the uncertainties in kn and n^0. The uncertainties in kn were determined using the residuals of the ts, as described in Eq. 6.9 - 6.13. The uncertainties in n^0 were also determined using the residuals of the ts, with the equation: n^0 = ^ s 1 N + x2 P (xi x)2 (7.10) 141 The overall relative uncertainties in the values of n^n(t) are given by the equa- tion, n^n(t) = s kn kn 2 + n^0 n^0 2 (7.11) The uncertainties calculated with Eq. 7.11 are listed in Table 7.1 for each data collection day. This method of deriving uncertainties in the ts did not produce consistent results. Speci cally, the outlying value of n^n(t) on March 23 is the product of a very at linear t (low absolute values of kn and n^0). This at t is the result of operating with a cold (low negative reactivity) core, following a period of reactor down time during Spring Recess. With low initial ssion poison concentrations, it is possible to continuously run secondary cooling over the course of the reactor day, leading to less heating of the reactor pool and fewer compensating manipulations of the control rods. Under these conditions, the uence in the thermal-column neutron-beam was observed to be nearly constant in comparison to other data collection days. A simpler method for determining the uncertainties in n^n(t) was chosen to avoid this biasing against at- ux days. The standard deviation in each set of counts was used to determine n^n(t), with the equation, n^n(t) = v u u t 1 N 1 NX i=1 ((nn)i nn)2 (7.12) 142 where nn is the mean value of the counts from that day, and N is the number of measurements of nn. Relative uncertainties in n^n(t) determined by this method also appear in Table 7.1. These values were used for the nal analysis. Date kn n^0 kn kn n^0 n^0 n^n(t) 1 n^n(t) 2 T ( C/hr) 3.09.12 -1739 17787 8:36 10 5 0.0239 0.0239 0.0438 5.53 3.14.12 -1228 19568 1:36 10 4 0.0258 0.0258 0.0317 2.96 3.16.12 -950.1 17774 2:37 10 4 0.0391 0.0391 0.0223 6.51 3.23.12 -11.95 3396.6 0.0279 0.3069 0.3082 0.0161 0.51 Table 7.1: Values and uncertainties from ts of the ssion chamber counts. The uncer- tainty values were determined by (1) linear regression analysis techniques described in Eq. 6.9 - 6.13 and 7.10, and (2) standard deviations of each set of counts. The rate of increase of the reactor pool temperature ( T ) is also shown for each reactor day. Following the tting of the measured values nn, Eq. 7.7 is reduced to, Nn = n^n(t) tn h 10i ( A)10M235 h 235i ( A)235M10 p 2: (7.13) The remaining unknowns in this equation are the factors h 10i, h 235i, and . E ec- tive microscopic cross sections (h i)must be used in Eq. 7.13, because the MUTR neutron beam is not monoenergetic. An e ective cross section may be calculated with a continuous weighted average of the form, h i = R (E) (E) dE R (E) dE (7.14) where (E) is the energy-dependent microscopic cross section of the material, and (E) is the energy-dependent ux of the neutron beam. Unfortunately, the energy distribution of the thermal-column neutron beam has not been directly measured (this distribution may be estimated with modeling, as discussed in Section 7.4). 143 However, the 1=v nature of the deposit (235U) and the target (10B) cross sections, and the thermal nature of the neutron beam, enable the elimination of (E) from this calculation. First, Eq. 7.14 may be rewritten in terms of velocity rather than energy, where (v) = n(v)v, h i = R (v)n(v) v dv R n(v) v dv (7.15) In the 1=v absorption region of (v), / v 1 and, therefore, the product v is a constant. Equation 7.15 may then be rewritten as, h i = v R n(v) dv R n(v) v dv (7.16) This reduction holds true as long as contributions from neutrons outside the 1=v energy region are negligible [41]. The absence of epithermal and fast neutrons from the thermal-column neutron beam was veri ed by measuring the neutron count rate with a lithium-glass beam block in front of the ssion chamber. While the beam block (7.325 mm thick, 6.3% 6Li) eliminates 99.998% of thermal neutrons, it only eliminates 81.84% of neutrons at 1 eV, and 16.37% of neutrons at 100 keV. Measure- ments with this beam block in place revealed count rates of 0.018 Hz. Comparatively, count rates without the beam block were typically around 2.66 Hz. Additionally, the absence of non-thermal neutrons is demonstrated by the energy-distribution modeling, discussed in Section 7.4. 144 According to Eq. 7.16, the ratio of the two e ective cross sections in Eq. 7.7 may be reduced to, h 10i h 235i = 10 v R n(v) dv R n(v) v dv R n(v) v dv 235 v R n(v) dv = 10 235 = (7.17) where is simply the ratio of the two cross sections at energies (or velocities) in the 1=v region. While is not perfectly constant below 1 eV, it does not vary signi cantly in the thermal region. Even a 50% uncertainty in the neutron energy at which to select , from a centerpoint of E = 0:025 eV, results in a variation of by only 2.61%. The value of at E = 0:025 eV (2200 m/s) was selected for this analysis ( = 6:567). A relative uncertainty of 2.61% was associated with this value. The cross sections of 10B and 235U are plotted in Figure 7.13. The value of over thermal energies is plotted in Figure 7.14. Following this simpli cation, Eq. 7.7 is reduced to, Nn = n^n(t) tn ( A)10M235 ( A)235M10 p 2: (7.18) The value represents the fraction of neutrons which were not absorbed between the ssion chamber reference monitor and the boron thin- lm targets. This accounts for absorption by: the ssion chamber deposit and housing, the air gap between the ssion chamber and the scintillation cell, the silica entrance window of the cell, the foil covering the silica entrance window, the gas in the cell, and the aluminum cylinder coated with Ebonol-C. Thicknesses, macroscopic cross sections (2200 m/s), 145 and transmission fractions (T ) for these components, are listed in Table 7.2. Trans- mission fractions were determined using Beer-Lambert?s Law. The resulting value of is 0.995. This value assumes noble gas pressures of 1 atm. The value does not account for neutrons that were scattered out of the beam. For each scintillation measurement, values ofNn were calculated using measurements from the ssion-chamber beam monitor and Eq. 7.18. Values of Nn are later used to determine scintillation yields per neutron absorption, in Chapter 9. The average neutron absorption rate of each 10B target, when irradiated in the thermal-column beam, appears in Table 7.3. Component Material Thickness (mm) a (cm 1) T deposit 235U 2:41 10 4 33.5 0.9992 housing Al 1 0.0139 0.9986 air mixture 75 3:83 10 5 0.9997 foil Al 0.05 0.0139 0.9999 window SiO2 0.5 1:41 10 3 0.9999 noble gas Ar, Kr, Xe 35 (1:82; 67:4; 66:3) 10 5 0.9999, 0.9976, 0.9977 cylinder Al 1 0.0139 0.9986 coating Cu 0.012 0.322 0.9996 Table 7.2: Thicknesses, macroscopic cross sections, and transmission fractions of the ma- terials between the ssion-chamber reference detector and the boron thin- lm target. Gas values were calculated at atmospheric pressure. Nominal Thickness (nm) ( A)10 ( g/cm2) Nn (Hz) 300 73.50 91.7 600 140.31 170.5 900 209.13 256.2 1200 260.02 332.9 Table 7.3: Average neutron absorption rates of each 10B target, as determined by Eq. 7.18 and measurements from the ssion-chamber beam monitor. tn = 1200 s. 146 Figure 7.8: Pulse-height distributions from the ssion chamber. Noise peaks are repre- sented in black. Bottom-right distribution shows a measurement taken with Li-glass in front of the detector. Count time t = 1200 s. 147 Figure 7.9: Neutron counts on 3.09.12. Count time tn = 1200 s. Figure 7.10: Neutron counts on 3.14.12. Count time tn = 1200 s. 148 Figure 7.11: Neutron counts on 3.16.12. Count time tn = 1200 s. Figure 7.12: Neutron counts on 3.23.12. Count time tn = 1200 s. 149 Figure 7.13: Cross sections of 10B and 235U over the thermal energy region. Figure 7.14: Value of over the thermal energy region. 150 7.3 Beam Pro le Measurement A beam pro le measurement was performed to verify the limited divergence of the neutron beam between the ssion-chamber reference monitor and the 10B targets. Dysprosium foils were irradiated at two locations in the beam (1) directly behind the beam aperture, and (2) 12 cm behind the aperture, in the approximate position of the boron targets. An imaging plate (photostimulable phosphor) was then ex- posed by the activated foils, and processed to produce images of the neutron beam pro le. Dysprosium-164 has an isotopic abundance of 28.18%, and a large thermal-neutron absorption cross section ( 0 = 2981 b) [77, 32]. When irradiated with neutrons, this isotope undergoes the reaction, n+164 Dy!165 Dy (7.19) 165Dy!165 Dy + + 1:25 MeV (7.20) from which 165Dy decays with a half-life of 2.334 h. The natural dysprosium foils used in this measurement were 10 10 cm, with areal densities of 0.0213 g/cm2. The foils were irradiated separately for 45 minutes each, at a reactor power of 200 kW. With the exception of the 4-mm beam aperture, the scintillation cell and the other components of the scintillation experiment were removed from the beamline during the irradiations. 151 Following both irradiations, the activated foils were transferred to the imaging plate. When -particles (or other energetic radiation) pass through an imaging plate, they excite electrons in the phosphor. Some of the excited electrons become trapped in the phosphor lattice until stimulated by a light source, at which point their deexcitation releases additional photons (photostimulated luminescence). These -particles are said to \expose" the imaging plate. Activities of 0.1 min 1mm 2 can be measured by exposures of less than 1 hour [137]. The Dy foils exposed the imaging plate for 120 minutes. The foils were then removed from the imaging plate and the plate was transported to NIST, where its photostimulated luminescence (PSL) was measured by an image reader (FujiFilm BAS 1800-II). This image reader produced a digital scan of the stored images, with 50 m resolution. Two sections of the measured image correspond to the neutron beam pro le at each Dy foil location. These sections appear in Figures 7.15 and 7.16. The color value of each pixel in these images is directly proportional to the amount of PSL observed by the image reader at that location. The rst image shows the neutron beam pro le directly behind the aperture. The Dy foil in this location was irradiated rst, and was allowed to decay during the second foil irradiation. Therefore, the resulting image was corrected for this decay time, with the factor 1 e t = 1:2495 (7.21) 152 where is the decay constant of 165Dy (0.00495 min 1), and t is the decay time (45 min). This correction assumes that the PSL from the imaging plate is linearly proportional to the deposited energy. The second image shows the neutron beam pro le 12 cm behind the aperture. This was approximately the distance between the aperture and the boron targets dur- ing the scintillation experiment. This image was not corrected since the foil was transferred to the imaging plate directly after irradiation. The beam spot in Figure 7.15 is slightly larger than the 4-mm aperture. This is due to the \blooming" e ect of the imaging plate exposure [137]. The isotropic nature of -decay causes some electrons to be emitted at glancing angles to the imaging plate, producing a blurring e ect at the boundaries of the image. The beam spot in Figure 7.16 has a diameter of approximately 7 mm. Together, these measurement demonstrate that the diameter of the neutron beam spread by approximately 3 mm over the 12 cm distance between the aperture and the boron lm. The 7-mm beam spot in Figure 7.16 is still signi cantly smaller than the boron target size (25 25 mm). 153 Figure 7.15: Neutron beam pro le immediately behind the aperture. 154 Figure 7.16: Neutron beam pro le 12 cm behind the aperture. 155 7.4 Energy-Distribution Modeling To derive the uence of the MUTR thermal-column neutron beam, based on mea- surements from the ssion-chamber beam monitor, it is necessary to estimate the energy distribution of the beam, and the e ective cross section of 235U in the beam (h 235i). These parameters were not required to calculate the number of neutrons absorbed by the boron thin lms, as demonstrated in Section 7.2. Nonetheless, they provide useful benchmarks for future experiments that will use the thermal-column beam for neutron irradiations. Monte Carlo N-Particle (MCNP) transport code is a general-purpose, continuous- energy, coupled neutron/photon/electron transport code developed at Los Alamos National Laboratory [138]. The code implements random number generation to ob- tain solutions for complex radiation transport calculations. These calculations typ- ically lack deterministic solutions and, therefore, require statistical simulation. De- signing an MCNP simulation involves de ning geometry, materials, particle sources, and output types. Energy-dependent cross sections for a wide range of materials are provided in the MCNP data libraries. An MCNP model of the MUTR, developed by Ali Mohamed, was modi ed to simulate the thermalization of neutrons passing through the thermal column. The details of the original model are described in [139]. The geometry of the modi ed MUTR MCNP model appears in Figure 7.17. This geometry includes the reactor core, the reactor pool, the concrete containment 156 structure, and the various components of each experimental facility, speci cally the thermal-column graphite, and the collimator insert. In modifying the model from [139], the thermal-column graphite was arti cially partitioned to demonstrate the thermalization of the neutron spectrum between the core and the access cavity. Each labeled cell in Figure 7.17 represents a ux tally volume within the model. During the simulation, neutrons that enter a particular cell, make a contribution to that cell?s tally equal to, C = WT V (7.22) where T is the particle track length through that cell, V is the volume of the cell, and W is the weight of the particle. Analytically, if (r; E; ) is the energy and angular distribution of the neutron uence as a function of position within the model, ux tallies measure a statistical solution to the equation, FT = 1 V Z V dV Z E dE Z 4 d (r; E; ) (7.23) Particle weights (W ) are determined according to the importance value given to the cell through which the particle is traveling. Variance reduction techniques often exploit cell importances to amplify particle populations in a particular segment of the model geometry. In the MUTR simulation, for example, cell importances were assigned according to the proximity of each cell to the thermal-column access cavity. More information on cell importances and geometry splitting techniques is available in [140]. 157 Figure 7.17: Geometry of the MUTR MCNP model, showing a cross sectional view of the thermal-column facility. The geometry has been arti cially partitioned to sample the neutron spectrum at various depths in the graphite moderator. The numbered segments correspond to the cells that were used as tally volumes for the simulations. By creating a histogram of tally contributions according to particle energy, energy distributions may be obtained from each tally volume. Neutron energy distributions ( (E)) from the various segments of the thermal-column graphite, appear in Figure 7.18. These distributions have been normalized to the peak value of the distribution closest to the core. The thermalization of the neutron population is demonstrated by the thermal-to-fast ux ratio at various distances from the core. This parameter is given by the equation, R = th fast (7.24) 158 where th is the number of neutrons below the cadmium cuto (0.6 eV), and fast is the number of neutrons above the cadmium cuto . Values of R in the thermal- column graphite appear in Figure 7.19 as a function of distance from the core. Using Eq. 7.14, the microscopic cross section of 235U, and the energy-dependent ux of the thermal column access cavity, the e ective cross section of 235U in the thermal- column neutron beam (h 235i) was calculated to be 483.8 b. For comparison, the 2200 m/s cross section of 235U is 585.1 b. By substituting h 235i into Eq. 7.5, the average uence of the thermal column neutron beam was determined for each data collection day, with the equation, NTC = Nbeam A (7.25) where NTC is the average uence in units of cm 2s 1, Nbeam is the average uence behind the beam aperture in units of Hz, and A is the area of the beam aperture opening (0.126 cm2). Values of NTC appear in Table 7.4. For comparison, the MUTR Safety Analysis Report lists the maximum thermal ux in the fuel as 5:0 1012 cm 2s 1 [101]. Date nn (Hz) NTC (cm 2s 1) 3.09.12 2.427 35,297 3.14.12 2.515 36,577 3.16.12 2.569 37,357 3.23.12 2.606 37,894 Table 7.4: Average count rates from the ssion chamber (nn) and average total neu- tron uence in the MUTR thermal-column beam (NTC), as determined from nn and the modeled energy distributions. 159 Figure 7.18: Neutron energy distributions at various depths in the MUTR thermal column. Cell numbers correspond to the geometry displayed in Figure 7.17. Figure 7.19: Neutron ux ratios across the MUTR thermal column. 160 With a newly constructed collimator, a neutron beam was obtained from the MUTR thermal-column. Several methods were implemented to characterize this beam. Measurements from the ssion-chamber beam monitor were analyzed to determine the rate of absorption in the boron thin lm targets (90 - 330 Hz). Foil activation was used to demonstrate the limited divergence of the beam between the aperture and the target position. MCNP simulations were used to demonstrate the thermal nature of the energy distribution, and to derive estimates of the total neutron ux (3.5 - 3:8 104 cm 2s 1). 161 Chapter 8 Modeling In conjunction with the excimer scintillation measurements, computer modeling was employed to simulate charged-particle transport from the 10B(n; )7Li reaction, and FUV photon emission within the scintillation cell. The results of these models were used to determine the average energies of charged particles escaping the boron lms, the ranges and stopping times of the charged particles in the heavy noble gases, and the photon collection e ciency of the experimental apparatus. The incorporation of these results into the analysis of the scintillation measurements is discussed in Chapter 9. 8.1 Charged-Particle Transport The TRIM (Transport of Ions in Matter) code is a collection of programs that cal- culate the stopping and range of ions in matter through statistical simulations of ion-atom collisions [141]. These simulations provide quantum mechanical approxi- mations of Coulomb collisions, long range interactions, and the charge state of each ion as it passes through a target. User inputs to TRIM include ion charge, mass, energy (10 eV - 2 GeV), emission position, and angle, as well as a 3D description of 163 the target geometry and composition. TRIM outputs include the nal distribution of ions within the target, and descriptions of the ions? energy loss and trajecto- ries while slowing down. The code has been actively developed since 1983 and is available at [142]. For this research, TRIM was used to simulate charged-particle emission from boron thin lms, and the slowing down of those charged particles in noble gases. Custom TRIM inputs were generated in the TRIM.dat le format, using a MATLAB script to choose random numbers for the various entries. The TRIM.dat inputs specify the emission positions, directions, and starting energies of and 7Li pairs, according to the characteristics of the 10B(n; )7Li reaction. Emission directions were chosen isotropically. Starting energies were chosen according to the branching ratios of the 10B(n; )7Li reaction (6% to E = 1777 keV and E7Li = 1015 keV, and 94% to E = 1470 keV and E7Li = 840 keV). Emission positions were chosen according to the neutron absorption depth pro le and the neutron beam width at the locations of the boron lms. The absorption depth pro le was sampled using an inverse transform of the form, u = 1 e x u 2 [0; 1] (8.1) where u is a random number from a uniform distribution, is the macroscopic absorption cross section of 10B, and x is the absorption depth. The input generator allowed variation of the emission-position distribution based on the thicknesses of the boron lm targets. The 45 tilt of the lm was not considered during these simulations. Target geometries and densities were speci ed in the TRIM graphical 164 user interface (GUI). The targets described the thicknesses and density of the boron lms, and the types and pressures of the noble gases. All simulations were run in the \quick KP damage" mode for 20,000 particle histories. Typical plots of particle trajectories appear in Figure 8.1. Figure 8.1: Particle trajectory plots from TRIM simulations of the 10B(n; )7Li reaction under 400 torr of Xe. Depth plot depicts the boron lm at the left-hand border. Transverse plot depicts a head-on view of the boron lm. The rst set of TRIM simulations estimated the average kinetic energy of charged- particles escaping the boron lms. These simulations were run for a range of lm thicknesses (0.1 - 2.0 m). The results appear in Figure 8.2. These values represent the average energy available to generate noble gas excimers, following each neutron capture. They demonstrate that, on average, 60-85% of the 10B(n; )7Li Q-value (2.792 MeV) is deposited in the boron lm before the charged particles escape into the gas. From these results, the average escape energies (Eesc) were determined 165 for each boron sample used in the scintillation experiment. The thicknesses of the boron samples were calculated from the NIF data, with the equation, x = A (8.2) where A is the areal density of the boron lm, as determined in Chapter 5.2.1, and is the default density of boron in the TRIM library (2.350 g/cm3). The corresponding values of Eesc appear in Table 8.1. Figure 8.2: Charged-particle escape energy as a function of lm thickness. Accounts for both and 7Li from the 10B(n; )7Li reaction. Nominal Thickness (nm) A ( g/cm2) x (nm) Eesc (keV) 300 73.5 312.7 932.2 600 140.3 597.0 797.2 900 209.1 889.8 685.5 1200 260.0 1106.4 615.9 Table 8.1: Boron lm thicknesses, as calculated with Eq. 8.2, and average charged-particle escape energies, determined by TRIM simulation. 166 The second set of TRIM simulations modeled (1) the radial range of the charged particles, (2) the stopping time of the charged particles, and (3) the ionization den- sity in the gas targets, as a function of noble-gas type and pressure. Modeling these parameters is necessary to understand the rate and duration of excimer formation, and the geometry of the photon emission volume. The gas types and gas pressures de ned in these simulations were identical to the conditions established in the scin- tillation measurements. The simulations considered a single target thickness of 300 nm. The radial range and stopping times determined by TRIM appear in Figure 8.3, as a function of gas pressure. Radial range is given by the equation, Rr = PN i=1 p y2i + z 2 i N (8.3) where y and z are the lateral projections of each particle on a plane parallel with the boron lm, and N is the number of particle histories. These range values were used to estimate the size of the photon emission volume in Section 8.2. The stopping times were calculated by substituting the range values into the equation, T = Rr K r mMp 2Eesc (8.4) where Rr is the charged particle range, K is an adjustment coe cient (0.6) described in Chapter 2.5, m is the number of nucleons per charged particle (averaged to 5.5), Mp is the mass of the proton (938.272 MeV/c2), and E is the average escape energy 167 from a 300-nm lm (0.9382 MeV). The stopping times in Figure 8.3 are only order-of- magnitude estimates, since the and 7Li particles were not modeled independently. Ionization density plots appear in Figures 8.4 - 8.6 for Ar, Kr, and Xe, at a number of gas pressures. The plots show the two-dimensional shape (x-y projection) and the ionization density (z projection) of the volume in which the charged particles deposit their energy. They are contour plots, with a contour drawn at every step of 3 10 6 eV/ ion A. Since the mean free path of Ar, Kr, and Xe atoms at STP are < 100 nm (see Table 2.2), the shape and density of the excimer photon source may be assumed to resemble these ionization volumes, at least for pressures close to 1 atm. The results of the TRIM simulations provide estimates of the amount of energy available to generate noble gas excimers (600 - 930 keV), the time over which that energy is deposited in the noble gas (1-15 ns), and the size, shape, and ionization density of the volume in which the energy is deposited. 168 Figure 8.3: Charged-particle range and stopping time in Ar, Kr, and Xe, as a function of gas pressure. Includes contributions from and 7Li particles. 169 Figure 8.4: Ionization density in argon at various pressures, from neutron irradiation of a 300-nm 10B lm. Each right-hand, upright plane represents the surface of the lm. 170 Figure 8.5: Ionization density in krypton at various pressures, from neutron irradiation of a 300-nm 10B lm. Each right-hand, upright plane represents the surface of the lm. 171 Figure 8.6: Ionization density in xenon at various pressures, from neutron irradiation of a 300-nm 10B lm. Each right-hand, upright plane represents the surface of the lm. 172 8.2 Photon Emission and Re ection Chapter 6 describes measurements of the intrinsic e ciency of the PMT/MgF2 de- tector package. This parameter is needed to determine the fraction of photons impinging on the detector which create a countable pulse. Before calculating the absolute scintillation yield, it is also necessary to determine the fraction of all scin- tillation photons which impinge on the detector. This fraction will be referred to as the collection e ciency of the apparatus, and is given by the equation, cube = number of radiation quanta incident on detector number of radiation quanta emitted by source (8.5) This de nition has been slightly modi ed from Eq. 2.8 to account for all photons incident on the detector, not just those that create a recorded pulse. Additionally, the e ective collection e ciency determined by statistical modeling will be referred to as h cubei. A point source approximation is the simplest way to determine cube. This approxi- mation assumes that photons are emitted isotropically from the exact center of the boron target, and is calculated with the equation, cube = 4 = 1 2 1 d p d2 + a2 (8.6) where is the solid angle subtended by the detector, d is the distance from the source to the detector photocathode (8.58 cm), and a is radius of the detector 173 photocathode (1.15 mm). The result from this calculation is cube = 0:4432%. This simpli cation does not take into account the form of the photon emission source, the re ectivities of the surfaces within the apparatus, or the spatial uniformity of the PMT/MgF2 detector package. To better approximate the true collection e ciency of the excimer scintillation cell, a Monte Carlo model of the cell was developed using MATLAB. These simulations incorporated both the geometry of the apparatus, and the re ectivity of its various surfaces over the FUV region. A drawing of the model geometry appears in Figure 8.7. Figure 8.7: Photon-emission model geometry. The model rst generates the emission position of a photon (x), from a uniform and random distribution describing a solid hemisphere. The radius of the hemisphere is an input parameter of the model. A useful resource on sphere point picking is available at [143]. The hemisphere is tilted at a 45 angle with respect to the 174 detector plane due to the tilt of the boron target. An isotropic vector direction (u) for the photon is then generated from a uniform random distribution. A line drawn from the emission position in the vector direction is represented by the parametric equations, x = x0 + ut (8.7) y = y0 + vt (8.8) z = z0 + wt (8.9) where (x0; y0; z0) are the components of x, (u; v; w) are the components of u, and t is distance along the line. The intersection of that line with a cylinder is calculated with the equation, t = (u x wz0) + p (u x wz0)2 (1 w2)(x20 + y 2 0 a2) 1 w2 (8.10) where a is the radius of a cylinder parallel to, and centered on, the z axis. The radius of the cylinder (1.15 cm) describes both the radius of the photocathode, and the radius of the copper-black FUV absorber. According to the location of this intersection, the photon hits one of three surfaces within the model. If the inter- section is above the plane of the detector surface, the photon strikes the detector. If the intersection is below the plane of the boron lm, the photon strikes the lm. If the intersection is between those two planes, the photon strikes the copper-black cylinder. 175 Based on which surface the photon strikes, it is either collected, re ected, or ab- sorbed. A photon that strikes the detector, adds a score to the detector tally. In this model, the detector is divided into two scoring regions, determined by the spatial scan measurements from Chapter 6.4. These regions appear in Figure 8.8. The inner region (9-mm radius) demonstrated a uniform response with only 3% variation from the mean. The outer region (11.5-mm radius) demonstrated a response that was, on average, 57% of the mean value in the uniform region. In the model, photons striking the uniform region are given a score of 1, and photons striking the outer region are given a score of 0.57. Figure 8.8: Scoring regions overlayed on the spatial response of the PMT, from Chapter 6.4. A photon that strikes the boron lm is specularly re ected with a 35% probability. The re ectivity of boron (%B = 0:35) was obtained from [144]. A plot of boron re ectivity values over the FUV region appears in Figure 8.9. When the photon is 176 specularly re ected, a new line is drawn from the point of re ection, in the direc- tion, u^s = 2(u^n u^i)u^n u^i (8.11) where u^n is the direction normal to the lm surface, and u^i is the direction of the incident photon. The intersection of the new line with the cylinder is then determined with Eq. 8.10. If that intersection occurs above the detector plane, the photon strikes the detector. If not, the photon is absorbed. If the photon is not specularly re ected, it is also absorbed. Figure 8.9: Boron re ectivity over the FUV region [144]. Various data series correspond to boron lms of di erent thicknesses. A photon that strikes the copper-black cylinder, is di usely re ected with a 1% probability. The di use re ectivity of copper black (%Cu = 0:01) is described in [145]. Two plots of bidirectional-re ectance distribution functions (BRDF) for cop- per black appear in Figure 8.10. When the photon is di usely re ected, a new isotropic direction is generated from a uniform random distribution, and a line is 177 drawn from the point of re ection in that direction. The intersection of that line with the cylinder is then determined with Eq. 8.10. If that intersection occurs above the detector plane, the photon strikes the detector. If not, the photon is absorbed. If the photon is not di usely re ected, it is also absorbed. Figure 8.10: Copper-black BRDF data measured at 121.6 nm [145]. Left: 15 angle of incidence. Right: 45 angle of incidence. By simulating the histories of millions of photons, it is possible to determine the frac- tion of all scintillation photons impinging on the detector. The collection e ciency may be calculated with the equation, h cubei = ninner + Couter nouter N (8.12) where ninner is the number of photons which strike the inner detector region, nouter is the number of photons which strike the outer detector region, Couter is the fractional score given to photons which strike the outer region (0.57), andN is the total number of photons generated by the model. For the geometry and re ectivities speci ed above, the output of the model for 10 million particle histories is h cubei = (0:5115 0:0051)%. The statistical uncertainty in this value was derived by determining 178 h cubei after every run of 1 million particle histories, and calculating the standard deviation in the outputs from 10 runs. The photon-emission model was veri ed by replicating the conditions of the point source approximation, and comparing the statistical solution to the analytical so- lution of Eq. 8.6. To achieve this, Couter was set to 1, the re ectivities %B and %Cu were set to 0, and the radius of the hemispherical emission source was set to 1 mm. The resulting output of the model was h cubei = (0:4429 0:0056)%, which di ers by only 0.07% relative to the analytical solution (0.0003% absolute di erence). An early iteration of the model was also compared to a model constructed with the FRED optical analysis code. FRED is a ray tracing code developed by Photon Engineering LLC [146]. The output of the FRED model di ered by 7% from the output of the initial MATLAB code. The FRED simulations were performed by Dr. June Tveekrem at NASA?s Goddard Space Flight Center. There are several simplifying assumptions associated with the photon collection model (1) the excimer photon emission volume is assumed to be hemispherical, (2) the distribution of the emissions within this volume are assumed to be uniform, (3) the re ectivities of boron and copper black are assumed to be precisely known, (4) contributions of secondary and tertiary re ections are assumed to be negligible, and (5) refraction by the MgF2 windows is assumed not to alter the photon path. As a result of these assumptions, it was necessary to ascribe an uncertainty to the value h cubei. The total uncertainty in h cubei was determined by identifying uncertainties in the inputs of the model. The model was then run for an array of 179 input values, %B 2 [0:30; 0:40], %Cu 2 [0:005; 0:01], h 2 [8:33; 8:83] cm. These ranges were chosen according to the range of re ectivity values in [144] and [145], and a physical measurement of the photocathode height. The results of these simulations are plotted in Figures 8.11 and 8.12. The highest and lowest values of h cubei were used to determine a relative uncertainty in h cubei of 9.69% (relative). Several of the simulation results and their inputs appear in Table 8.2. Overall, the collection e ciency model incorporates a more realistic approach than the point source approximation, by accounting for re ective surfaces, a non-discrete emission source, and a non-uniform detector surface. While the model does not incorporate every complexity of photon transport within the apparatus, its uncer- tainty may be estimated by assigning uncertainties to the model inputs. The nal value of h cubei obtained with the model is 15% larger than the analytical point source solution. The value h cubei = 0:5115% will be used in the analysis on the excimer scintillation measurements in the following chapter. r (cm) %B (%) %Cu (%) h (cm) Couter n runs h cubei (%) h cubei (%) 1.0 35 1 8.58 0.57 106 10 0.5115 1.00 1.0 40 1 8.33 0.57 106 10 0.5611 1.30 1.0 30 0.5 8.83 0.57 106 10 0.4651 0.69 1.0 0 0 8.58 0.57 106 10 0.3852 2.09 0.1 0 0 8.58 0.00 106 10 0.2746 1.69 0.1 0 0 8.58 1.00 106 10 0.4429 1.26 Point Source 0.4432 Table 8.2: Outputs from the collection e ciency model. r = radius of hemispherical emission volume. %B = specular re ectivity of boron. %Cu = di use re ectivity of copper black. h = height of photocathode from lm centerline. Couter = score value of outer detector hits. n = number of photons per run. h icube = collection e ciency. = relative statistical uncertainty in h cubei from standard deviation of outputs from all runs. Point source value assumes uniform photocathode radius a = 11:5 cm and h = 8:58 cm. 180 Figure 8.11: Collection e ciency as a function of %B and h. Figure 8.12: Collection e ciency as a function of %B and %Cu. 181 Chapter 9 Results and Discussion In this chapter, the results of the excimer scintillation measurements are presented along with the analysis of the data. The characterizations and simulations discussed in Chapters 5 - 8 provide several elements of this analysis, including the areal den- sities of the boron targets (( A)10), the intrinsic e ciency of the PMT detector package (h pmti), the neutron absorption rate of the boron targets (Nn), and the collection e ciency of the experimental apparatus (h cubei). Ultimately, the excimer scintillation yield and lifetime are derived from the scintillation data. These quan- tities are directly relevant to the development of a thermal-neutron detector based on excimer scintillation. 9.1 Excimer Scintillation Yield Scintillation yield is an essential parameter for optimizing the geometry and col- lection e ciency of a scintillation detector. It is also necessary to assess the dis- crimination capabilities of a detection medium. At the MUTR, noble-gas excimer scintillation yield from the 10B(n; )7Li reaction was measured as a function of gas type, gas pressure, and 10B target thickness. 183 The excimer scintillation yield per neutron absorption is given by the equation, Y = Nh Nn (9.1) where Nh is the rate at which excimer photons were generated in the scintillation cell, and Nn is the rate at which neutrons were absorbed in the boron targets. Here, Nh = nh h pmtih cubei th (9.2) and, from Chapter 7, Nn = n^n(t) tn ( A)10M235 ( A)235M10 p 2 (9.3) The rate at which photons were generated in the scintillation cell (Nh ), was de- termined based on measurements from the photomultiplier tube. The number of excimer photons observed as pulses from the PMT (nh ), over time th , was cal- culated with two consecutive, but separate measurements from the PMT and the associated counting electronics. This value is given by the equation, nh = nh total nh bkgd (9.4) where nh total is the number of pulses counted over time th , and nh bkgd is the number of those pulses which are attributed to gamma-ray backgrounds and dark current. The rst value comes from a measurement taken while a mixed neutron and 184 gamma-ray beam propagated through the scintillation cell. The second value comes from a measurement taken with a lithium-glass beam block in front of the apparatus, which eliminated thermal neutrons from the beam. For comparison, nh bkgd was also measured by placing an epitaxial substrate (2 nm natural boron) in the cell, and irradiating it with a mixed beam. The count time for all scintillation measurements was th = 200 s. Plots of nh , and the components nh total and nh bkgd, appear in Figures 9.1 - 9.3, for each gas type and lm thickness. The results of the substrate irradiations appear in these gures as well. The statistical uncertainties in nh were calculated with Eq. 6.5. The relative statistical uncertainties appear in Figure 9.4. The three sources which contributed to nh bkgd were (1) dark current, (2) gamma- ray interactions with the PMT and solid components of the scintillation cell, and (3) gamma-ray interactions with the noble gases in the scintillation cell. The rst two sources produced a baseline count rate that was independent of gas pressure in the cell. The third source was dependent on noble gas type and pressure. This background signal behavior is apparent in both the data from the substrate irradi- ations, and the data from irradiations with the lithium beam block in place. The contributions to nh total are described in Figure 9.5. A multichannel analyzer (MCA) was used to ensure that multi-photoelectron pulses did not signi cantly contribute to the pulses which were counted by the counter/timer. None of the experimental conditions produced peaks in the pulse-charge distribu- tions (PCDs) outside of the single-photoelectron peak. Typical PMT PCDs appear 185 in Figure 9.6. The gaussian peak in each of these distributions represents the single- photoelectron peak. The remaining factors in Eq. 9.2 are the intrinsic e ciency of the PMT detector package (h pmti), and the collection e ciency of the apparatus (h cubei). Together, they de ne the fraction of all excimer photons generated within the apparatus which were counted by the PMT electronics. The derivations of these factors are described in Chapters 6 and 8, respectively. The derivation of Nn by Eq. 9.3 is described in Chapter 7. Based on Eq. 9.2 and 9.3, Y was determined as a function of gas type, gas pressure, and target thickness. Plots of Y appear in Figures 9.7 - 9.9. The results demonstrate yields of 5200 - 6000 photons/neutron from argon, 7500 - 11,000 photons/neutron from krypton, and 9600 - 14,200 photons/neutron from xenon, at pressures of 800 torr. The yields decrease as the target thicknesses increase because they have been normalized per neutron absorption. As shown by the TRIM simulations, charged particles from the 10B(n; )7Li reaction lose less energy in thinner lms. Thus, while thicker lms absorb a larger fraction of the neutron beam, the resulting scintillation yields per neutron capture are lower due to this energy loss. Additionally, the scintillation yields plateau around 600 torr of argon, 400 torr of krypton, and 300 torr of xenon. At these pressures, the charged particles deposit nearly all of their energy within the line of sight of the PMT. This behavior agrees with the TRIM simulations of ionization density (see Fig. 8.4 - 8.6). 186 The total uncertainty in Y may be calculated by propagating the uncertainties of each of its factors, with the equation, Y = v u u t NX i=1 ( i)2 (9.5) where Y is the total relative uncertainty in Y , and i is the relative uncertainty in each factor of Y . A description of each factor of Y and its associated uncertainty appears in Table 9.1. The total relative uncertainty in Y , as determined by Eq. 9.5, falls between 14.7 - 15.5% depending on the experimental conditions. This uncertainty is re ected by the error bars in Figures 9.7 - 9.9. 187 F acto r Descriptio n V alue(s ) (% ) Sourc e N h n h excime r photo n cou nt s 100 0 - 80,00 0 (O ) 0. 5 - 5 (O ) sci ntillatio n measureme nt s h p m ti PM T in trinsi c e cienc y 1. 6 - 3. 1 (% ) (O ) 3.4 9 PM T cal ibratio n an d excime r sp ectr a h cu b e i apparatu s collectio n e cienc y 0.511 5 (% ) 9.6 9 photo n emissio n an d re ectio n si m ulation s t h photo n cou nt tim e 20 0 (s ) { N n n^ n (t ) tte d neutro n cou nt s 273 0 - 315 0 1. 6 - 4. 4 (O ) ssion- cha m b er measureme nt s 1 0 B /23 5 U therma l cros s sectio n rati o 6.56 7 2.6 1 ana ly si s of cros s section s fro m [77 ] ( A ) 1 0 area l densi ty of 1 0 B target s 73. 5 - 260. 0 ( g/c m 2 ) (O ) 10 neutro n im agi ng ( A ) 23 5 area l densi ty of 23 5 U de p osi t 458.07 3 ( g/c m 2 ) 0. 5 measureme nt by D avi d Gillia m M 1 0 mola r mas s of 1 0 B 10.012 9 (g/mol ) { [76 ] M 23 5 mola r mas s of 23 5 U 235.043 9 (g/mol ) { [76 ] transmissio n fractio n 0.99 5 { transmissio n calculatio n self-absorptio n facto r 1.0375 6 { calculatio n by Je re y Nic o t n neutro n cou nt tim e 120 0 (s ) { T abl e 9.1 : V alue s an d relati ve uncertai ntie s of th e variou s input s to Eq . 9.1 . (O ) stand s fo r \o n th e orde r of" , du e to m ultipl e value s of th e sam e parameter . 188 Figure 9.1: Counts from scintillation of argon during irradiation of a 600-nm (top left), 900-nm (top right), and 1200-nm (bottom right) 10B targets, and an epitaxial substrate (bottom left). Count time th = 200 s. 189 Figure 9.2: Counts from scintillation of krypton during irradiation of a 300-nm (top left), 600-nm (top right), 900-nm (middle left), and 1200-nm (middle right) 10B targets, and an epitaxial substrate (bottom). Count time th = 200 s. 190 Figure 9.3: Counts from scintillation of xenon during irradiation of a 300-nm (top left), 600-nm (top right), 900-nm (middle left), and 1200-nm (middle right) 10B targets, and an epitaxial substrate (bottom). Count time th = 200 s. 191 Figure 9.4: Relative uncertainties in nh for argon (top left), krypton (top right), and xenon (bottom) data. 192 Figure 9.5: Contributions to nh total from (1) excimer scintillation signal, (2) pressure- dependent gamma-ray backgrounds, and (3) dark current and pressure-independent gamma-ray backgrounds. 193 Figure 9.6: Pulse-charge distributions from excimer scintillation of xenon, showing only a single-photoelectron peak. 0 (top left), 100 (top right), 200 (middle left), 300 (middle right), 600 (bottom left), and 800 (bottom right) torr during irradiation of a 1200-nm 10B lm. Distributions are shown from measurements with (nh total) and without (nh bkgd) the lithium glass beam block. 194 Figure 9.7: Scintillation yield from 10B(n; )7Li in argon. Data series correspond to 10B target thicknesses. 195 Figure 9.8: Scintillation yield from 10B(n; )7Li in krytpon. Data series correspond to 10B target thicknesses. 196 Figure 9.9: Scintillation yield from 10B(n; )7Li in xenon. Data series correspond to 10B target thicknesses. 197 Determining the amount of energy required to produce an excimer photon (Wh ) is necessary to evaluate the energy conversion e ciency of the excimer signal. Ad- ditionally, Wh may be used to derive the energy resolution of an excimer-based detector. Based on Y , Wh values were calculated using the average escape energies from Table 8.1 and the equation, Wh = Eesc Y (9.6) where Eesc is the average kinetic energy of the charged particles upon escaping the boron target. The results appear in Figure 9.10. Average values of Wh appear in Table 9.2. Equation 9.6 assumes that Eesc is fully deposited in the gas. This assumption is valid only at pressures in the plateau region, as described in the previous section. At lower pressures, charged particles may lose energy through collisions with the walls of the copper black cylinder, and, therefore, the scintillation signal is diminished. This \wall e ect" is well known in proportional-tube neutron detectors [1]. By dividing the energies of the excimer photons (9.57, 8.42, and 7.42 eV, for Ar, Kr, and Xe) by the Wh value, the conversion e ciency of the excimer emission signal may be obtained. For the Wh values at 800 torr, these e ciencies are 7.7, 10.4, and 11.8% for Ar, Kr, and Xe, respectively. 198 Figure 9.10: Wh values as a function of noble-gas type and pressure. Data from all thin- lm targets are included. Wh (eV/photon) Pressure (torr) Argon Krypon Xenon 200 227.2 120.0 77.8 400 144.1 86.9 66.7 600 122.9 80.9 64.9 800 124.8 81.3 63.1 Table 9.2: Average values of Wh from the 10B(n; )7Li reaction. Determined from scin- tillation measurements and TRIM simulations. 199 9.2 Excimer Lifetime Measurements of the excimer scintillation decay time following the 10B(n; )7Li re- action are necessary to determine the pulse-shape discrimination capability and the optimum signal integration width of an excimer-based neutron detector. These mea- surements were collected with a time-to-amplitude converter (TAC), which provided distributions of the times-between-pulses during excimer scintillation for a range of gas pressures (100 - 800 torr) and gas types (Ar, Kr and Xe). The distributions were analyzed to derive the triplet-state (slow) decay time of each gas. In this particular con guration, the delay on the \start" line (38 ns) was long with respect to typical fast excimer decay times (4 ns). Therefore, the fast decay was assumed to be unresolved in these distributions. This assumption is justi ed by the absence of multi-photoelectron pulses in the yield measurement PCDs. Coin- cidentally, typical fast excimer decay times are very close to the PMT pulse width (3-4 ns). If multiple pulses originating from singlet-state excimer decays of a single scintillation event, were frequently observed, they would inevitably form a multi- photoelectron peak in the PCDs. Since such a peak was not observed, it was assumed that the TAC rarely observed multiple pulses from singlet-decays of the same event. Furthermore, because neutron capture events in the experiment were signi cantly separated in time (< 340 Hz) with respect to excimer decay times (fast and slow), it was assumed that the times-between-pulses from separate scintillation events far exceeded the TAC range and, therefore, did not create TAC pulses. 200 Each decay time calculation required two TAC measurements, one with a mixed beam propagating through the cell (total distribution), and one with the lithium beam block in front of the cell (background distribution). The counts in each channel of the background distributions were subtracted from the counts in each channel of the corresponding total distributions. The range settings of the TAC were mapped to the channel numbers of the MCA to translate the abscissa values into times- between-pulses. The delay time of the delay line (38 ns) was then added to these values. Typical distributions from the TAC appear in Figures 9.11 - 9.13. Note that some of the distributions in these gures were collected at di erent full range settings on the TAC (500 ns - 20 s), leading to variation in the data point spacing and peak amplitude of each distribution. The changes in the order of the distribution amplitudes in Figures 9.11 - 9.13, with respect to change with respect to gas pressure, may be explained by these range settings. Following time calibration, the TAC data were t to an exponential distribution of the form, y^(t) = Ae( t) (9.7) where y^(t) is the value of the t as a function of the time-between-pulses t, A is an arbitrary multiplier based on the amplitude of the distribution, and is the decay constant of the distribution. Examples of these ts appear in Figures 9.11 - 9.13. The time constants obtained from this tting were then related to the excimer lifetime, with the equation, = 1 (9.8) 201 Values of the excimer triplet-state lifetime ( ), as a function of gas pressure, ap- pear in Figure 9.14. The uncertainties in these values were determined using the 95% con dence bounds in the t parameter , as determined by the MATLAB t function. The relative uncertainties in are plotted in Figure 9.15. The decrease in at higher pressures is the result of an increased frequency of collision-induced triplet-singlet transitions [79]. The values of calculated from these scintillation data distinctly agree with many of the tabulated values in [84] (see Fig. 2.12), particularly around 400 torr. Figure 9.11: Distributions of the times between pulses, and exponential ts of those distributions. Data collected during irradiation of a 1200 nm 10B target in argon. Count time t = 300 s. 202 Figure 9.12: Distributions of the times between pulses, and exponential ts of those distributions. Data collected during irradiation of a 1200 nm 10B target in krypton. Count time t = 300 s. Figure 9.13: Distributions of the times between pulses, and exponential ts of those distributions. Data collected during irradiation of a 1200 nm 10B target in xenon. Count time t = 300 s. 203 Figure 9.14: Triplet-state lifetimes of Ar, Kr and Xe excimers, as a function of gas pressure, resulting from irradiation of a 1200 nm 10B target. Figure 9.15: Relative uncertainties in . 204 9.3 Conclusions From the scintillation measurements and their interpretation based on absolute cal- ibrations, it is possible to draw a number of conclusions regarding the behavior of noble-gas excimer scintillation. Most prominently, exothermic neutron capture gen- erates a large excimer scintillation signal due to the large Q-value of the 10B(n; )7Li reaction, the ability of charged-particle reaction products to transfer that energy by electronically exciting noble-gas atoms, and the small amount of energy required to form an excimer molecule. The yields of the heavier gases are greater due to the lower excitation energies of these gas species [54]. The yields increase with pressure as a result of (1) an increased frequency of excimer-forming collisions, and (2) a higher density of ionization within the line of sight of the photon detector package. The excimer yields plateau at pressures at which nearly all of the charged-particle energy escaping the targets is deposited in view of the detector. This is con rmed by the TRIM simulations of ionization density. Additionally, the yields decrease with increasing target thickness due to the larger average energy loss of charged particles escaping thicker lms. The starting pressures of the signal plateau regions are depen- dent on the geometry of this speci c experiment { the radius of the photocathode, and the distance of the chamber walls from the location of charged-particle emission. However, the yields in the plateau region are dependent only on the properties of the neutron absorbing targets (though not on the remaining detector geometry). 205 Despite this speci city, the yields quantitatively describe relationships between pa- rameters of a basic con guration, and these relationships provide a foundation for optimizing a practical detector. Excimer triplet-state lifetimes decrease with increasing gas pressure as a result of an increased frequency of decay-inducing collisions between excimer molecules and sur- rounding noble-gas atoms (collisional broadening) [79]. Furthermore, while charged particles from neutron capture produced large populations of triplet-state excimers, gamma-ray radiation did not produce such phosphorescence, and only exhibited fast decays characteristic of singlet-state excimers. This is attributed to the high-LET of charged particles and the low-LET of gamma-rays. This scintillation behavior is well known in other scintillator types [59]. The cross sections of electronic excitation are vastly di erent for heavy ions and electrons produced by gamma-ray interactions [147]. The distinct scintillation decay behavior from the two radiation types will enable gamma-ray rejection based on pulse-shape discrimination [58]. The observed excimer scintillation signals from the 10B(n; )7Li reaction are compa- rable to the yields of many liquid and solid neutron scintillators [64]. These yields demonstrate that excimer scintillation is a practical means of neutron detection. The observed slow triplet-state decay of 10B(n; )7Li-induced excimers agrees well with literature values [88]. This behavior can be used in a practical detector to discriminate neutron interactions from gamma-ray interactions. Furthermore, the low densities of noble-gases translate to inherently lower gamma sensitivites than solid scintillators. Other advantages of a practical excimer-based neutron detector 206 include: immunity to radiation damage, simpli ed requirements (no need for high voltage, high-pressure gases), exible geometry (form- tting scintillation media, in- terchangeable neutron absorbers), and availability (helium-3 free). Several characterizations and absolute calibrations have enabled the interpretation of the scintillation measurements in terms of absolute excimer photon yield and excimer scintillation lifetime. The results provide a fundamental understanding of the properties of noble-gas scintillation from the 10B(n; )7Li reaction. These results will contribute to the development and optimization of a deployable neutron detector based on noble-gas excimer scintillation. Furthermore, the characterization and simulation methods described by this research provide a framework upon which to develop and characterize a practical detector. Additionally, this research has produced a characterized thermal-neutron-beam facility at the MUTR, which may be used for future detector development. 207 9.4 Future Work This research lays the foundation for a number of projects related to the development of an excimer-based neutron detector. These projects involve the optimization of de- tector e ciency and implementation of background discrimination techniques. The photon collection e ciency of a practical excimer-based neutron detector can be vastly improved. Wavelength shifters, FUV re ectors, modi ed geometry, and increased photodetector e ciency are all means of achieving this improvement. The absolute yield measurements from this thesis provide a method to derive the total collection e ciency of a complex apparatus, and to assess changes in that e ciency as a result of changes to the detector design. Additionally, implementing a spatially- sensitive photon detector (ie. CCD) may improve the understanding of the photon source geometry. A large photon collection e ciency will enhance the overall e - ciency of the detector, and may enable the di erentiation of neutron captures from background events. The neutron collection e ciency of an excimer-based neutron detector can be im- proved by investigating new absorber types and geometries. For example, by layering multiple lms in an array, thermal-neutron absorption may be increased to nearly 100%. Optimization of this con guration will require a balance between neutron absorption (number of lms and lm thickness) and light collection ( lm escape en- ergy and lm spacing). The measurements in this thesis describe scintillation yields from various lms which have been carefully characterized. These measurements 208 may be used to implement an array of boron lms to increase neutron absorption. Additionally, boron lms may be replaced with more exotic targets, such as a hon- eycomb or a porous foam of neutron absorbing material. These types of absorbers may provide even further leverage for optimizing neutron absorption and light col- lection. The techniques used for ionization density modeling in Chapter 8 prescribe a means to estimate the size and distribution of the photon emission source from such a neutron absorber. Discriminating gamma-ray events and dark counts from neutron events is essential for a practical neutron detector. While these signals were measured and subtracted from the scintillation data in this research, they were not identi ed on an individ- ual basis. As a means of discrimination, a future excimer-based neutron detector may implement an MCA integration time based on the scintillation lifetime mea- surements in this thesis, and pulse-shape discrimination techniques. Similarly, the lifetime measurements will be useful for developing digital pulse processing algo- rithms to discriminate backgrounds in a more sophisticated arrangement. 209 Appendix A Scintillation from 241Am -particle Emission Prior to the excimer scintillation experiment, an -particle check source was used to demonstrate the scintillation of noble gases due to bombardment by heavy charged- particles. These measurements qualitatively veri ed the operation of the scintillation cell by generating and detecting FUV photons. They required no use of the thermal- column neutron beam or the MUTR. The scintillation from 241Am -particles was measured in a con guration nearly identical to the apparatus described in Chapter 4. An 241Am -particle source was placed at the bottom of the scintillation cell, with its active surface facing the PMT/MgF2 detector package. The cell was then sealed, evacuated, and lled with noble gases (Ar, Kr, and Xe) at pressures between 25-800 torr. The FUV cylinder and the boron target were removed from the cell during these measurements. Americium-241 decays by the reaction, 241Am! 237Np + + 5:49 MeV (A.1) with a half-life of 432.6 years [32]. This reaction has the potential to generate noble- gas excimers through the same type of energy transfer as the 10B(n; )7Li reaction. 211 The activity of the source used in these measurements was 9:5 104 min 1 2 sr 1. The yield electronics were used to count scintillation pulses from these emissions. The resulting PMT signal appears in Figure A.1. Both the frequency of -particle emission (1600 Hz) and the energy of each emission (5.49 MeV) were larger during these measurements than during the thermal-neutron irradiations (90 - 330 Hz, and 0.62 - 0.93 MeV). Correspondingly, the signals from all three gases in Figure A.1 are signi cantly larger than the signals in Figures 9.1 - 9.3. The measurements have not been compared in terms of absolute yield, due to the di erent collection e ciencies of the two con gurations. Without the copper-black FUV-absorbing cylinder, simulating re ections within the cell becomes considerably more complex. Figure A.1: Noble-gas scintillation signal from 241Am -particle emitter. Y-axis represents counts from the PMT. Count time th = 200 s. 212 Appendix B Construction of an FUV Re ectometer When modeling the collection e ciency of the scintillation cell (see Chapter 8.2), re ectivity values for the boron and copper black surfaces were taken from [144] and [145], respectively. To verify these re ectivity values, an FUV re ectometer was constructed at SURF. This re ectometer was designed for use at BL-4, described in Chapter 6.1. A diagram of the apparatus appears in Figure B.1. Photographs of the re ectometer appear in Figures B.2 - B.4. The re ectometer consists of a stainless steel cube (150 mm) with metal-seal ange ports on all sides. Inside the cube, a sample holder and a detector arm rotate independently about a common axis, perpendicular to the incident photon beam. The sample holder is designed to t a 2:5 2:5 0:05 cm sample. The sample is held in place on two sides by exible mounting clips. Two photodiode detectors, of the same type used in the PMT calibration (IRD AXUV-100G), are secured to opposite ends of the detector arm. The detectors are mounted on breadboards perpendicular to the length of the arm, facing in the direction of the incident photon beam. The detector closest to the SURF storage ring measures the intensity of the incident photon beam. The second detector measures the intensity of the re ected photon beam. A collimator is placed around the re ected detector to prevent the interference from 213 di use photon re ections within the chamber. An electrical feedthrough on one side of the chamber connects these detectors to the BL-4 ammeters. Two rotating feedthroughs on opposite sides of the chamber allow the sample holder and the detector arm to be rotated independently. The feedthroughs enable mea- surement of the sample re ectivity as a function of incident and re ected angle. Additionally, the BL-4 monochromator allows these re ectivities to be measured at a range of wavelengths in the FUV region. A lubricant-free ceramic bearing was inset at the junction of the sample holder and the detector-arm shaft to facilitate smooth rotation of these components. Additionally, the anges connecting the feedthroughs to the chamber were sealed with Viton gaskets, to ensure ush alignment of the concentric shafts. The re ectometer is connected to the BL-4 detector box with several vacuum adapters and an x-y translation feedthrough. A 4-mm slit in the sample holder provides a means of aligning the rotation axis with the photon beam, using the re ected photodiode and the x-y translation feedthrough. A vacuum-safe window on the top ange of the chamber will assist in this alignment. The re ectometer is currently awaiting beam time at SURF. It will be used to measure the FUV re ectivities of the boron-10 thin lms, copper black coating (Ebonol-C), and any future components of an excimer-based neutron detector. 214 Figure B.1: Diagram of the FUV re ectometer. Faded detector arm represents the incident measurement/alignment position. Not to scale. Figure B.2: The FUV re ectometer. 215 Figure B.3: Head-on view of the FUV re ectometer. The sample holder is empty, exposing the alignment slit. Figure B.4: The re ectometer prepared for installation behind the BL-4 detector box. The x-y translation feedthrough required for alignment is shown. 216 Bibliography [1] G. F. Knoll, Radiation Detection and Measurement. John Wiley and Sons, 1979. [2] J. R. Lamarsh and A. J. Baratta, Introduction to Nuclear Engineering. Pren- tice Hall, 3rd ed., 2001. [3] T. Jaques, H. Ballinger, and F. Wade, \Neutron detectors for reactor instru- mentation," P IEEE, vol. 100, pp. 110{116, May 1953. [4] General Electric, \Nuclear reactor instrumentation." http://www.ge-mcs.com/en/nuclear-reactor-instrumentation/ neutron-monitoring-instrumentation.html. [5] R. Hartke, D. Symes, F. Buersgens, L. Ruggles, J. Porter, and T. Ditmire, \Fusion neutron detector calibration using a table-top laser generated plasma neutron source," Nucl. Instrum. Meth. A, vol. 540, pp. 464{469, 2005. [6] S. K. Lee, B.-H. Kang, G.-D. Kim, and Y.-K. Kim, \Fast neutron detector for fusion reactor KSTAR using stilbene scintillator," arXiv, December 2011. [7] T. Elevant, B. Wolle, and A. Weller, \Proposed neutron diagnostics for Wen- delstein 7-X stellarator," Rev. Sci. Instrum., vol. 70, no. 1, pp. 1185{1189. [8] D. Klir, J. Kravarik, P. Kubes, K. Rezac, E. Litseva, K. Tomaszewski, L. Karpinski, M. Paduch, and M. Scholz, \Fusion neutron detector for time- of- ight measurements in z-pinch and plasma focus experiments," Rev. Sci. Instrum., vol. 82, no. 3, pp. 033505{7, 2011. [9] Y. Jian-Lun, , W. Gen-Xing, Z. Qi-Hua, W. Shu-Huai, L. Zhong-Li, Y. Hong- Qiong, T. Zheng-Yuan, W. Shi-Bin, and W. Fan, \Neutron detector for fusion burn history measurements," Acta Phys. Sin.-Ov. Ed., vol. 8, pp. 768{775, October 1999. [10] NIST Center for Neutron Research, \How are neutrons useful?." http://www. ncnr.nist.gov/why neutrons.html, June 2010. [11] B. Hammouda, Probing Nanoscale Structures { The SANS Toolbox. NIST, 2010. [12] J. Bu et, J. Clergeau, R. Cooper, J. Darpentigny, A. D. Laulany, C. Fermon, S. Fetal, F. Fraga, B. Guerard, R. Kampmann, A. Kastenmueller, G. McIn- tyre, G. Manzin, F. Meilleur, N. Rhodes, L. Rosta, E. Schooneveld, G. Smith, 217 H. Takahashi, P. V. Esch, T. V. Vuure, and K. Zeitelhack, \Advances in detec- tors for single crystal neutron di raction," Nucl. Instrum. Meth. A, vol. 554, pp. 392{405, 2005. [13] M. Johnson, S. Manolopoulos, N. Rhodes, E. Schooneveld, R. Turchetta, and M. Daymond, \Silicon APS detectors for neutron scattering," Nucl. Instrum. Meth. A, vol. 501, pp. 72{79, Mar. 2003. [14] H. Tanaka, T. Nakamura, H. Yamagishi, K. Soyama, and K. Aizawa, \A two- dimensional gas detector with individual readouts for neutron detection with a high spatial resolution and fast temporal response," Rev. Sci. Instrum., vol. 76, no. 9, p. 093302, 2005. [15] S. Burgos, J. Forbes, C. Ghag, M. Gold, V. Kudryavtsev, T. Lawson, D. Loomba, P. Majewski, D. Muna, A. S. Murphy, G. Nicklin, S. Paling, A. Petkov, S. Plank, M. Robinson, N. Sanghi, N. Smith, D. Snowden-l t, N. Spooner, T. Sumner, J. Turk, and E. Tziaferi, \Studies of neutron detec- tion and backgrounds with the DRIFT-IIa dark matter detector," Astropart. Phys., vol. 28, pp. 409{421, December 2007. [16] D. Chandler, \New detector will aid dark matter search," MIT TechTalk, vol. 53, December 2008. [17] S. Monk, M. Joyce, Z. Jarrah, D. King, and M. Oppenheim, \A portable energy-sensitive cosmic neutron detection instrument," Rev. Sci. Instrum., vol. 79, no. 2, p. 023301, 2008. [18] P. Mosteiro, \Cosmogenic neutron background in dark matter detectors," pre- sentation, Princeton University, 2012. [19] M. Oraby, K. Verghese, and R. Gardner, \Investigation of an improved- sensitivity neutron-porosity oil-well logging tool," Nucl. Instrum. Meth. A, vol. 299, no. 1-3, pp. 674{681, 1990. [20] J. Reijonen, \Nuclear tools for oil eld logging-while-drilling applications," AIP Conf. Proc., vol. 1336, no. 1, pp. 433{436, 2011. [21] W. W. Givens and W. R. Mills, \Directional epithermal neutron detector," patent, United States Patent O ce, April 1986. [22] T. M. Persons and G. Aloise, \Neutron detectors: Alternatives to using helium-3," tech. report GAO-11-753, United States Government Accountabil- ity O ce, September 2011. 218 [23] R. Runkle, A. Bernstein, and P. Vanier, \Securing special nuclear material: Recent advances in neutron detection and their role in nonproliferation," J. Appl. Phys., vol. 108, no. 11, p. 111101, 2010. [24] R. V. Ginhoven, R. Kouzes, and D. Stevens, \Alternative neutron detector technologies for homeland security," tech. report PNNL-18471, Paci c North- west National Laboratory, 2009. [25] R. C. Byrd, J. M. Moss, W. C. Priedhorsky, C. Pura, G. Richter, K. Sager, and W. Scarlett, \Nuclear detection to prevent or defeat clandestine nuclear attack," IEEE Sens. J., vol. 5, p. 593, August 2005. [26] R. T. Kouzes, E. R. Siciliano, J. H. Ely, P. E. Keller, and R. J. McConn, \Passive neutron detection for interdiction of nuclear material at borders," Nucl. Instrum. Meth. A, vol. 584, pp. 383{400, January 2008. [27] J. Hall, S. Asztalos, P. Biltoft, J. Church, M.-A. Descalle, T. Luu, D. Manatt, G. Mauger, E. Norman, D. Petersen, J. Pruet, S. Prussin, and D. Slaughter, \The nuclear car wash: Neutron interrogation of cargo containers to detect hidden SNM," Nucl. Instrum. Meth. B, vol. 261, no. 1-2, pp. 337{340, 2007. [28] J. K. Shultis and R. E. Faw, Radiation Shielding. American Nuclear Society, 2000. [29] Institute of Atomic and Subatomic Physics, \Neutron radiography of a camera." http://www.ati.ac.at/ neutropt/experiments/Radiography/ radiography.html, July 2012. [30] Oak Ridge National Lab, \Preparing for the threats," Oak Ridge National Laboratory Review, vol. 39, no. 1, 2006. [31] R. Kouzes, \The helium-3 supply problem," tech. report PNNL-18388, Paci c Northwest National Laboratory, April 2009. [32] National Nuclear Data Center, \Chart of nuclides database." http://www. nndc.bnl.gov/chart/, April 2012. [33] D. Dixon, \Helium-3 shortage could mean nuke detection disaster," Wired, April 2010. [34] G. Aloise and T. M. Persons, \Managing critical isotopes: Weaknesses in DOE?s management of helium-3 delayed the federal response to a critical sup- ply shortage," tech. report GAO-11-472, Government Accountability O ce, May 2011. 219 [35] A. Simpson, S. Jones, M. Clapham, and S. McElhaney, \A review of neutron detection technology alternatives to helium-3 for safeguards applications," INMM Proc., July 2011. [36] R. Hu man, J. Larrabee, and Y. Tanaka, \Rare gas continuum light sources for photoelectric scanning in the vacuum ultraviolet," Appl. Optics, vol. 4, pp. 1581{1588, December 1965. [37] R. Chandrasekharan, M. Messina, and A. Rubbia, \Detection of noble gas scintillation light with large area avalanche photodiodes (LAAPDs)," Nucl. Instrum. Meth. A, vol. 546, no. 3, pp. 426{437, 2005. [38] K. Saito, H. Tawara, T. Sanami, E. Shibamur, and S. Sasaki, \Absolute num- ber of scintillation photons emitted by alpha particles in rare gases," IEEE T. Nucl. Sci., vol. 49, no. 4, pp. 1674{1680, 2002. [39] P. P. Hughes, M. A. Coplan, A. K. Thompson, R. E. Vest, and C. W. Clark, \Far-ultraviolet signatures of the 3He(n; tp) reaction in noble gas mixtures," Appl. Phys. Lett., vol. 97, no. 23, pp. 234105{3, 2010. [40] J. Chadwick, \Possible existence of a neutron," Nature, vol. 129, no. 3252, p. 312, 1932. [41] K. S. Krane, Introductory Nuclear Physics. USA: John Wiley and Sons, 2 ed., 1988. [42] R. A. Dunlap, The Physics of Nuclei and Particles. Belmont, CA: Thomson, Brooks/Cole, 2004. [43] CODATA, \The NIST reference on constants, units and uncertainty." http: //physics.nist.gov/cuu/Constants/, June 2011. [44] J. Beringer and et al. (Particle Data Group), \The review of particle physics," Phys. Rev., vol. D86, p. 010001, 2012. [45] R. Golub and P. Hu man, \Search for a neutron electric dipole moment," J. Res. Natl. Inst. Stan., vol. 110, May-June 2005. [46] S. Mughabghab, Atlas of Neutron Resonances: Resonance Parameters and Thermal Neutron Cross Sections. Oxford, UK: Elsevier, 5 ed., 2006. [47] University of Cambridge: Materials for Nuclear Power Generation, \Plot of neutron cross sections." http://www.doitpoms.ac.uk/tlplib/nuclear materials/cross section.php, May 2012. [48] T. Crane and M. Baker, Passive Nondestructive Assay of Nuclear Material, ch. 13. NRC, 1991. 220 [49] Stony Brook Dept. of Physics and Astronomy, \Gas multiplication curves." http://felix.physics.sunysb.edu/allen/252/PHY251 Geiger.html, Au- gust 1998. [50] R. Kouzes, J. Ely, A. Lintereur, E. Siciliano, and M. Woodring, \BF3 neu- tron detector tests," tech. report PNNL-19050, Paci c Northwest National Laboratory, 2009. [51] A. Lintereur, R. Kouzes, J. Ely, L. Erikson, and E. Siciliano, \Boron-lined neutron detector measurements," Tech. Rep. PNNL-18938, Paci c Northwest National Laboratory, 2009. [52] R. Kouzes, \Boron-lined straw-tube neutron detector test," tech. report PNNL-19600, Paci c Northwest National Laboratory, August 2010. [53] C. Morris, T. Bowles, J. Gonzales, R. Hill, G. Hogan, M. Makela, R. Morten- son, J. Ramsey, A. Saunders, S. Seestrom, W. Sondheim, W. Teasdale, H. Back, R. Pattie, A. Holley, A. Young, L. Broussard, K. Hickerson, J. Liu, M. Mendenhall, B. Plaster, R. Mammei, M. Pitt, R. Vogelaar, R. Rios, and J. Martin, \Multi-wire proportional chamber for ultra-cold neutron detection," Nucl. Instrum. Meth. A, vol. 599, pp. 248{250, February 2009. [54] E. Aprile, A. Bolotnikov, A. Bolozdynya, and T. Doke, Noble Gas Detectors. Berlin: Wiley VCH, 2006. [55] R. L. Aamodt, L. J. Brown, and G. M. Smith, \High pressure 3He gas scintil- lation neutron detector," Rev. Sci. Instrum., vol. 37, pp. 1338{1340, October 1966. [56] M. S. Derzon, D. R. Slaughter, and S. G. Prussin, \A high-pressure 3He gas scintillation neutron spectrometer," IEEE T. Nucl. Sci., vol. 33, February 1986. [57] Bruker AXS, Inc., \Diagram of a scintillation neutron detector." http:// bruker-axs.de/ leadmin/xr ntro/sec1 6.html, May 2012. [58] P. Reeder, A. Peurrung, R. Hansen, D. Stromswold, W. Hensley, and C. Hub- bard, \Detection of fast neutrons in a plastic scintillator using digital pulse processing to reject gammas," Nucl. Instrum. Meth. A, vol. 422, no. 1-3, pp. 84{88, 1999. [59] S. Youse , L. Lucchese, and M. Aspinall, \Digital discrimination of neutrons and gamma-rays in liquid scintillators using wavelets," Nucl. Instrum. Meth. A, vol. 598, no. 2, pp. 551{555, 2009. 221 [60] Wikipedia, \Diagram of uorescence and phosphorescence." http://en. wikipedia.org/wiki/Phosphorescence, June 2012. [61] J. Tain, \Neutron detection," presentation, Instituto de Fisica Corpuscular, February 2005. [62] J. Uher, 3D Neutron Detectors. PhD thesis, Academy of Sciences of the Czech Republic, 2007. [63] T. Kojima, M. Katagiri, N. Tsutsui, K. Imai, M. Matsubayashi, and K. Saka- sai, \Neutron scintillators with high detection e ciency," Nucl. Instrum. Meth. A, vol. 529, pp. 325{328, 2004. [64] C. van Eijk, A. Bessiere, and P. Dorenbos, \Inorganic thermal-neutron scin- tillators," Nucl. Instrum. Meth. A, vol. 529, pp. 260{267, 2004. [65] Z. W. Bell, D. Carpenter, S. Cristy, V. Lamberti, A. Burger, B. F. Wood eld, T. Niedermayr, I. D. Hau, S. E. Labov, S. Friedrich, W. G. West, K. R. Pohl, and L. van der Berg, \Neutron detection with cryogenics and semiconductors," Phys. Status Solidi B, vol. 2, no. 5, pp. 1592{1605, 2005. [66] A. Conway, L. Voss, C. Reinhardt, R. Gra , T. Wang, R. Nikolic, N. Deo, and C. Cheung, \Silicon based thermal neutron detectors," presentation, Lawrence Livermore National Laboratory, 2011. [67] D. McGregor, M. Hammig, Y. Yang, H. Gersch, and R. Klann, \Design consid- erations for thin lm coated semiconductor thermal neutron detectors," Nucl. Instrum. Meth. A, vol. 500, pp. 272{308, 2003. [68] A. Conway, T. Wang, N. Deo, C. Cheung, and R. Nikolic, \Numerical sim- ulations of pillar structured solid state thermal neutron detector e ciency and gamma discrimination," tech. report LLNL-JRNL-405025, Lawrence Liv- ermore National Laboratory, June 2008. [69] T. Mason, \Neutron detectors for materials research," presentation, Oak Ridge National Laboratory, 2011. [70] C. Guardiola, J. Rodriguez, C. Fleta, D. Quirion, and M. Lozano, \Portable silicon neutron detector system," Proc. Span. Conf. Elec. Dev., 2011. [71] S. Almaviva, M. Marinelli, E. Milani, G. Prestopino, A. Tucciarone, C. Verona, G. Verona-Rinati, M. Angelone, D. Lattanzi, M. Pillon, R. Montereali, and M. Vincenti, \Thermal and fast neutron detection in chemical vapor deposition single-crystal diamond detectors," J. Appl. Phys., vol. 103, no. 5, p. 054501, 2008. 222 [72] G. W. Fraser, J. F. Pearson, and O. S. Al-Horayess, \Thermal neutron imaging using microchannel plates," Proc. SPIE, vol. 1737, pp. 298{307, 1993. [73] S. Dazeley, A. Bernstein, N. Bowden, and R. Svoboda, \Observation of neu- trons with a gadolinium doped water Cherenkov detector," Nucl. Instrum. Meth. A, vol. 607, pp. 616{619, 2009. [74] D. Jordan, J. Ely, A. P. L. Bond, J. Collar, M. Flake, M. Knopf, W. Pitts, M. Shaver, A. Sonnenschein, J. Smart, and L. Todd, \Neutron detection via bubble chambers," Appl. Radiat. Isotopes, vol. 63, no. 5-6, pp. 645{653, 2005. [75] T. Niedermayr, I. Hau, T. Miyazaki, S. Terracol, A. Burger, V. Lamberti, Z. Bell, J. Vujic, S. Labov, and S. Friedrich, \Microcalorimeter design for fast-neutron spectroscopy," Nucl. Instrum. Meth. A, vol. 520, no. 1-3, pp. 70{ 72, 2004. [76] Wolfram Alpha. http://www.wolframalpha.com/, June 2012. [77] G. Hale and P. Young, \ENDF/B-VII." http://www.nndc.bnl.gov/exfor/ endf00.jsp, April 2006. [78] Z. Wang and C. Morris, \Multi-layer boron thin- lm detectors for neutrons," Nucl. Instrum. Meth. A, vol. 652, no. 1, pp. 323{325, 2011. [79] M. Hutchinson, \Excimers and excimer lasers," Appl. Phys., vol. 21, pp. 95{ 114, 1980. [80] J. Birks, \Excimers," Rep. Prog. Phys., vol. 38, no. 8, 1975. [81] C. K. Rhodes, ed., Excimer Lasers. Berlin: Springer-Verlag, 2nd ed., 1984. [82] M. Mutterer, P. Grimm, H. Heckwolf, J. Pannicke, W. Spreng, and J. Theobald, \Response of gas scintillators to heavy charged particles," Lect. Notes Phys., vol. 178, pp. 63{79, 1983. [83] T. Stewart, G. Hurst, T. Bortner, J. Parks, F. Martin, and H. Weidner, \Pro- ton excitation of continuous emission in the noble gases," J. Opt. Soc. Am., vol. 60, pp. 1290{1297, October 1970. [84] A. Policarpo, \Light production and gaseous detectors," Phys. Scripta, vol. 23, pp. 539{549, 1981. [85] M. Dantus, R. Bowman, and A. Zewail, \Femtosecond laser observations of molecular vibration and rotation," Nature, vol. 343, pp. 737{739, February 1990. 223 [86] M. Mutterer, J. Theobald, and K.-P. Schelhaas, \A low-pressure noble-gas scintillation counter for heavy-ion detection," Nucl. Instrum. Meth., vol. 144, pp. 159{166, 1977. [87] H. Koehler, L. Federber, D. Redhead, and P. Ebert, \Vacuum-ultraviolet emis- sion from high-pressure xenon and argon excited by high-current relativistic electron beams," Phys. Rev. A, vol. 9, no. 2, 1974. [88] E. Morikawa, R. Reininger, P. Gurtler, and V. Saile, \Argon, krypton and xenon excimer luminescence: From the dilute gas to the condense phase," J. Chem. Phys., vol. 91, no. 3, 1989. [89] T. Doke, A. Hitachi, J. Kikuchi, K. Masuda, H. Okada, and E. Shibamura, \Absolute scintillation yields in liquid argon and xenon for various particles," Jpn. J. Appl. Phys., vol. 41, pp. 1538{1545, 2002. [90] R. Bernabei, P. Belli, A. Incicchitti, and D. Prosperi, \Liquid noble gases for dark matter searches: A synoptic survey," arXiv, May 2009. [91] C. Monteiro, L. Fernandes, J. Lopes, L. Coelho, J. Veloso, J. dos Santos, K. Giboni, and E. Aprile, \Secondary scintillation yield in pure xenon," J. Instrum., vol. 2, May 2007. [92] M. Miyajima, K. Masude, Y. Hoshi, T. Doke, T. Takahashi, T. Hamada, S. Kubota, A. Nakamoto, and E. Shibamura, \A self-triggered liequid xenon drift chamber by the use of proportional ionization or proportional scintilla- tion," Nucl. Instrum. Meth., vol. 160, pp. 239{246, 1979. [93] E. Aprile, \Liquid xenon detectors for particle physics and astrophysics," Rev. Mod. Phys., vol. 82, no. 3, pp. 2053{2097, 2010. [94] T. Dias, F. Santos, P. Rachinhas, F. Borges, J. dos Santos C.A.N. Conde, and A. Stau er, \Xenon-neon gas proportional scintillation counters: Experimen- tal and simulation results," J. Appl. Phys., vol. 85, no. 9, p. 6303, 1999. [95] M. Plum, E. Bravin, J. Bosser, and R. Maccaferri, \Nitrogen and xenon gas scintillation cross-section, spectrum and lifetime measurements from 50 MeV to 25 GeV at the CERN PS and Booster," Nucl. Instrum. Meth. A, vol. 492, pp. 64{90, 2002. [96] NIST Center for Neutron Research, \NCNR Instrumentation." http://www. ncnr.nist.gov/instruments/index.html, April 2012. [97] R. E. Williams, \The NIST research reactor and cold neutron source," pre- sentation, NIST Center for Neutron Research, June 2007. 224 [98] J. Nico, M. Arif, M. Dewey, T. Gentile, D. Gilliam, P. Hu man, D. Jacobson, and A. Thompson, \The fundamental neutron physics facilities at NIST," J. Res. Natl. Inst. Stan., vol. 110, no. 3, pp. 137{144, 2005. [99] A. K. Thompson, M. A. Coplan, J. W. Cooper, P. P. Hughes, R. E. Vest, and C. W. Clark, \Observation of the 3He(n; tp) reaction by detection of far- ultraviolet radiation," J. Res. Natl. Inst. Stan., vol. 113, March-April 2008. [100] J. W. Cooper and P. P. Hughes, \The detection of Lyman-alpha radiation formed by the slowing down of protons and tritons produced by the 3He(n; tp) reaction { a model study," J. Res. Natl. Inst. Stan., vol. 114, no. 3, pp. 185{ 199, 2009. [101] Maryland University Training Reactor, Safety Analysis Report, February 2000. [102] T. Koeth, \Photograph of the MUTR core." [103] Hamamatsu Photonics, Japan, Photomultiplier Tube R6835, 2000. [104] W. Otano-Rivera, L. J. Pilione, J. A. Zapien, and R. Messier, \Cubic boron ni- tride thin lm deposition by unbalanced magnetron sputtering and DC pulsed substrate biasing," J. Vac. Sci. Technol. A, vol. 16, no. 3, pp. 1331{1335, 1998. [105] Y. Tzeng and H. Zhu, \Electron-assisted deposition of cubic boron nitride by RF magnetron sputtering," Diam. Relat. Mater., vol. 8, pp. 1402{1405, 1999. [106] J. Deng, B. Wang, L. Tan, H. Yan, and G. Chen, \The growth of cubic boron nitride lms by RF reactive sputter," Thin Solid Films, vol. 368, no. 312-314, 2000. [107] C. Rohr, J.-H. Boo, and W. Ho, \The growth of hexagonal boron nitride thin lms on silicon using single source precursor," Thin Solid Films, vol. 322, pp. 9{13, 1998. [108] T. Ichiki and T. Yoshida, \Preparation of cubic boron nitride lms by low- pressure ICP-CVD," Appl. Phys. Lett., vol. 64, no. 851, 1994. [109] H. Sachdev, F. Muller, and S. Hufner, \Formation of boron-based lms and boron nitride layers by CVD of a boron ester," ChemInform, vol. 42, no. 28, 2011. [110] S.-H. Park, H.-S. Kim, and J.-H. Ha, \Fabrication of a GEM-based gas detec- tor for thermal neutron measurement," J. Korean Phys. Soc., vol. 49, no. 5, pp. 1939{1943, 2006. 225 [111] T. Akan, N. Ekem, S. Pat, U. Issever, M. Balbag, M. Cenik, R. Vladoiu, and G. Musa, \Boron thin lm deposition by using thermionic vacuum arc technology," Mater. Lett., vol. 61, pp. 23{26, 2007. [112] G. Musa, R. Vladoiu, V. Ciupina, C. Lungu, I. Mustata, S. Pat, T. Akan, and N. Ekem, \Characteristics of boron thin lms obtained by TVA technology," J. Optoelectron. Adv. M., vol. 8, no. 2, pp. 617{620, 2006. [113] G. Yukin, \The mechanism of electroplating with boron," Met. Sci. Heat Treat., vol. 13, no. 8, pp. 662{664, 1971. [114] Z.-F. Song, S.-Z. Yi, Z.-Y. Chen, L. Song, and J. Shen, \Study on boron- lm thermal neutron converter prepared by pulsed laser deposition," Appl. Radiat. Isotopes, vol. 69, no. 443-447, 2011. [115] S. ichi Aoqui, H. Miyata, T. Ohshima, T. Ikegami, and K. Ebihara, \Prepa- ration of boron carbide thin lm by pulsed KrF excimer laser deposition pro- cess," Thin Solid Films, vol. 407, pp. 126{131, 2002. [116] F. Perkins, M. Onellion, S. Lee, D. Li, J. Mazarowski, and P. Dowben, \Synchrotron-radiation-induced deposition of boron and boron carbide lms from boranes and carboranes," Appl. Phys. A, vol. 54, pp. 442{450, 1992. [117] Brigham Young University, \Diagram of electron beam evaporation." http: //www.cleanroom.byu.edu/metal.phtml. [118] National Institute of Standards and Technology, \NIST CNST NanoFab Equipment Metal Deposition." http://www.cnst.nist.gov/nanofab/nanofab equipment/metal dep.html. [119] D. Jacobson, D. Hussey, E. Baltic, T. Udovic, J. Rush, and R. Bowman, \Neutron imaging studies of metal-hydride storage beds," in Asian Hydrogen Energy Conference, vol. 35, pp. 12837{12845, 2009. [120] D. Hussey, E. Baltic, K. Coakley, and D. Jacobson, \Improving quantitative neutron radiography through image restoration." unpublished. [121] Saclay Institue of Matter and Radiation, \Diagram of x-ray photoelec- tron spectroscopy." http://iramis.cea.fr/en/Phocea/Vie des labos/Ast/ast sstechnique.php, July 2012. [122] UMD Department of Chemistry and Biochemistry, \Photograph of XPS tool." http://www.chem.umd.edu/researh/sharedinstrumentation/ xrayphotoelectronspectroscopy. 226 [123] H. Piao and N. McIntyre, \Adventitious carbon growth on aluminium and gold-aluminium alloy surfaces," Surf. Interface Anal., vol. 33, pp. 591{594, 2002. [124] J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben, Handbook of X-Ray Photoelectron Spectroscopy. Minnesota, USA: Physical Electronics, 1995. [125] B. Cullity, Elements of X-Ray Di raction. Reading, Massachusettes: Addison- Wesley Publishing Company, 1956. [126] J. V. Bhardwaj, X-ray di raction study of the structure of boron. PhD thesis, Rice University, Houston, TX, December 1979. [127] The University of Iowa, \Diagram of an SEM." http://www.uiowa.edu/ cmrf/ methodology/sem/index.html, July 2012. [128] U. Arp, C. Clark, A. Farrell, E. Fein, M. Furst, and E. Hagley, \Synchrotron ultraviolet radiation facility: SURF III," Rev. Sci. Instrum., vol. 73, no. 3, pp. 1674{1676. [129] U. Arp, R. Friedman, M. L. Furst, S. Makar, and P. S. Shaw, \SURF III - an improved storage ring for radiometry," Metrologia, vol. 37, no. 5, pp. 357{360, 2000. [130] P.-S. Shaw, K. R. Lykke, R. Gupta, T. R. O?Brian, U. Arp, H. H. White, T. B. Lucatorto, J. L. Dehmer, and A. C. Parr, \Ultraviolet radiometry with synchrotron radiation and ultraviolet radiometry," Appl. Optics, vol. 38, no. 1, pp. 18{27, 1999. [131] U. Arp, C. Clark, L. Deng, N. Faradzhev, A. Farrell, M. Furst, S. Grantham, E. Hagley, S. Hill, T. Lucatorto, P. Shaw, C. Tarrio, and R. Vest, \SURF III : A exible synchrotron radiation source for radiometry and research," Nucl. Instrum. Meth. A, vol. 649, pp. 12{14, 2010. [132] J. A. Samson and D. L. Ederer, Vacuum Ultraviolet Spectroscopy. Academic Press, 2000. [133] D. Ederer, B. Cole, and J. West, \A high throughput 2 m normal incidence monochromator for SURF-II," Nucl. Instrum. Meth., vol. 172, no. 1-2, pp. 185{ 190, 1980. [134] U. Arp, \Photograph of the SURF." [135] L. Can eld, Vacuum Ultraviolet Spectroscopy, ch. Photodiode Detectors. Aca- demic Press, 2000. 227 [136] R. Vest, \Photodiode Responsivity Data." [137] Z. Chowdhuri, Absolute Neutron Measurements in Neutron Decay. PhD thesis, Indiana University, June 2000. [138] X-5 Monte Carlo Team, MCNP { A General Monte Carlo N-Particle Transport Code, Version 5. Los Alamos National Laboratory, February 2008. [139] A. B. Mohamed, Flux Maps Obtained from Core Geometry Approximations: Monte Carlo simulations and benchmark measurements for a 250kW TRIGA reactor. PhD thesis, University of Maryland, 2009. [140] J. Shultis and R. Faw, An MCNP Primer. Kansas State University, 2006. [141] J. Ziegler, J. Biersack, and M. Ziegler, SRIM - The Stopping and Range of Ions in Matter. James Ziegler, 2008. [142] J. Ziegler, \SRIM Software." http://www.srim.org. [143] E. W. Weisstein, \Sphere point picking." http://mathworld.wolfram.com/ SpherePointPicking.html, July 2012. [144] M. Vidal-Dasilva, M. Fernandez-Perea, J. A. Mendez, J. A. Aznarez, and J. I. Larruquert, \Electron-beam deposited boron coatings for the extreme ultraviolet," Appl. Optics, vol. 47, pp. 2926{2930, June 2008. [145] M. P. Newell and R. A. M. Keski-Kuha, \Bidirectional re ectance distribution function of di use extreme ultraviolet scatters and extreme ultraviolet ba e materials," Appl. Optics, vol. 36, pp. 5471{5475, August 1997. [146] Photon Engineering, \FRED Software." http://www.photonengr.com/software/, July 2012. [147] B. Dolgoshein and B. Rodionov, \The noble gas scintillation mechanism," Elem. Part. and Cosm. Rays, no. 2, pp. 170{198, 1972. 228