Neutron Depth Profiling Measurements and GEANT4 Simulation Comparison for Intel-SEA2 Borophosphosilicate Glass (BPSG) Sample
Open Access
- Author:
- Ucar, Dundar
- Graduate Program:
- Nuclear Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- None
- Committee Members:
- Kenan Unlu, Thesis Advisor/Co-Advisor
Kenan Unlu, Thesis Advisor/Co-Advisor - Keywords:
- NIST
PennState
Borophosphosilicate Glass (BPSG)
Geant4
Neutron Depth Profiling method
NDP
simulation
spectrum
RSEC - Abstract:
- Neutron depth profiling (NDP) is a widely used near-surface neutron analysis technique to determine the concentration versus depth profile of several technologically important light elements in almost all solid materials. The most commonly analyzed light elements are boron, lithium, and nitrogen; but several other elements can also be analyzed. NDP method is based on the energy measurements of the charged particles escaping from the surface of the sample material. Energy measurements are mostly performed by using semiconductor detectors. By using the stopping power of the sample material, depth profile of the analyzed element can be obtained by making a linear transformation of the measured energy spectrum. A few micrometer of the material can be analyzed nondestructively, and on the order of 10 nm depth resolution can be obtained depending on the material type with NDP method. In this study, Intel-SEA2 borophosphosilicate glass (BPSG) sample was experimentally analyzed at Pennsylvania State University Neutron Depth Profiling (PSU-NDP) Facility and National Institute of Standard Technology Cold Neutron Depth Profiling (NIST-Cold NDP) Facility to calculate the 10B depth profile inside of the material. NDP measurements were obtained by using only silicon PIN Photodiode detector at PSU-NDP facility, and by using both silicon PIN Photodiode and surface barrier detector at NIST-Cold NDP facility. PSU-NDP Facility is a part of Radiation Science and Engineering Center (RSEC) of Penn State University. In this facility, NDP experiments suffered from some specific problems. These were mainly signal noise due most likely to ground loop formation on the experimental setup and gamma content of the neutron beam. As a consequence of these problems, the measured energy spectrum of reaction products obtained from BPSG sample at PSU-NDP facility were not satisfactory enough to analyze the 10B depth profile. Therefore, NDP experiments were repeated at NIST-Cold NDP Facility. In this facility, the gamma content of the neutron beam is much less than the one at PSU-NDP facility, and also there is no signal noise problem on the experimental setup. Therefore, better results were obtained from the NDP measurements of the BPSG sample. The thickness of the BPSG layer inside of the sample was calculated as 858 nm with surface barrier detector and 866 nm with silicon PIN Photodiode detector. Energy and concentration calibrations were made by using NIST standard reference samples. To verify the experimental results, NDP measurements performed at PSU-NDP facility with Intel SEA2 BPSG sample were simulated by using Geant4 code. Simulation was performed at LION-XO PC cluster of Penn State University (PSU) at 15 different computer nodes. In the simulation model, it was assumed that the thickness of the BPSG layer in the silicon wafer was the measured value, which is 858 nm, by using Tennelec surface barrier detector at NIST-Cold NDP Facility. Geant4 code was very successful to predict the 1472 and 1776 keV alpha peaks in the measured energy spectrum. The net area difference between the measured and predicted alpha peaks was less than 1%. It was also successful for 1013 keV lithium peak, but not for 840 keV lithium peak. Net area difference for that peak between the measured and predicted spectra was calculated as 36%. The problem might be the cross section data set used by the code to simulate the lithium ion transport in the silicon material. Since depth profiling calculation was made by using the 1472 keV alpha peak and it was well predicted by the Geant4 code, predicted depth profile perfectly fits to the measured one. The thickness of the BPSG layer inside of the sample was calculated as 856 nm, which is very close to the measured result of 858 nm.