J.G. Kelly

Neutron Damage Equivalence for Silicon, Silicon Dioxide, and Gallium Arsenide

Displacement- and ionization-energy transfers to Si, SiO2, and GaAs as functions of incident neutron energy were calculated using new cross section data and fine group structure in the NJOY code system. Neutron spectra determinations for several reactor neutron environments were made using new activation cross sections and a new technique with the SAND II code. Measurements of carrier removal rates in GaAs and of Si transistor gain degradation were made in representative neutron environments. Experimental results are compared to damage ratios predicted with the new spectra and NJOY displacement functions. For fission-like spectra, calculated Si damage ratios are in good agreement with those determined with ASTM E722-85 and with measured transistor damage ratios. Significant differences are found between Si NJOY and ASTM E722-85 for 14-MeV-to-reactor neutron damage ratios where NJOY gives better agreement with experimental data reported in the literature. In GaAs 14-MeV-to-reactor experimental damage ratios are smaller than predicted by calculated displacement ratios. This suggests that a more complex model of damage for majority carrier removal in GaAs is required. The use of incorrect damage functions is shown to adversely affect simulation fidelity in some representative neutron environments.

Neutron damage equivalence in GaAs

A 1-MeV neutron damage equivalence methodology and damage function have been developed for GaAs based on a recoil-energy dependent damage efficiency and the displacement kerma. This method, developed using lifetime degradationIn GaAs LEDs in a variety of neutron spectra, is also shown to be applicable to carrier removal. A validated methodology, such as this, is required to ensure and evaluate simulation fidelity in the neutron testing of GaAs semiconductors. >

Dose Enhancement Effects in MOSFET IC's Exposed in Typical 60Co Facilities

The responses of CMOS dosimeters sensitized to ionizing radiation by ion implantation have been used to demonstrate dose enhancement of 55 percent when exposed in typical 60Co facilities. Pairs of these ICs, one type with an alumina lid over the silicon chip and the other with a gold-kovar lid were used to evaluate this effect. Additional tests with a 1.3 mm thick lead filter show that the enhancement is predominately induced by low energy components in the radiation fields.

Benchmarking the Sandia Pulsed Reactor III cavity neutron spectrum for electronic parts calibration and testing

The SPR (Sandia Pulsed Reactor) III bare cavity spectrum and integral parameters have been determined with 24 measured spectrum sensor responses and an independent, detailed, transport calculation. This environment qualifies as a benchmark field for electronic parts testing. The measured SPR III cavity spectrum is characterized by a very smooth curve that predicts the measured activities with a standard deviation of 3.5% for the 24 sensors. Within the limitations of the computational model, the spectrum is also very close to that provided by MCNP (Monte Carlo Neutron Photon) calculations. >

Energy dependence of neutron damage in silicon bipolar transistors

2N2222A transistors were exposed in neutron fields ranging in energy from thermal to 14 MeV. Spectrum characterization data are reported for the three fast reactors used in hardness testing. The neutron energy dependence of transistor damage response was measured and compared with predictions based on ASTM standards and recent compilations of the silicon kerma factors calculated using NJOY. For all the environments considered, the hardness parameters calculated using the NJOY displacement kerma factors are significantly smaller than those calculated using the ASTM standard values, using the same value of 95 MeV-mbarn for the kerma factor at 1 MeV. The NJOY calculations actually show that this value is too high, so that if the standard is revised the 1-MeV value should also change. The spectrum evaluations for the fast reactors improve the agreement between the observed and predict hardness ratios, provided consistent methods are used for the unfolding. The thermal reactors and the moderated californium source show consistent results only when thermal and epithermal neutrons are removed by boron filters. The ASTM standards should be revised to include this precaution as well. >

User`s manual for SNL-SAND-II code

Sandia National Laboratories, in the process of characterizing the neutron environments at its reactor facilities, has developed an enhanced version of W. McElroy`s original SAND-II code. The enhanced input, output, and plotting interfaces make the code much easier to use. The basic physics and operation of the code remain unchanged. Important code enhancements include the interfaces to the latest ENDF/B-VI and IRDF-90 dosimetry-quality cross sections and the ability to use silicon displacement-sensitive devices as dosimetry sensors.

Use of silicon bipolar transistors as sensors for neutron energy spectra determinations

Recent reevaluation of the neutron displacement damage function for silicon qualifies it as a sensor for spectra determinations. This development is especially useful in the critical energy region from 0.2 to 2.0 MeV where, in the absence of fission foils, there is a shortage of response functions needed to define spectra satisfactorily. How silicon bipolar devices can be used to improve neutron spectra determinations and therefore to better predict the displacement damage induced in devices is described. >

Simulation fidelity issues in reactor irradiation of electronics-reactor environments

Fast burst reactors, such as SPR III, and pool-type reactors, such as the University of Illinois TRIGA and the Sandia Annular Core Research Reactor, are used as sources of neutrons and/or gamma rays in transient radiation effects in electronics research and testing. The authors identify the important simulation fidelity issues, compare radiation environments produced in representative reactors, and describe custom environments that have been achieved to better meet experimenter requirements. >

A rigorous treatment of self-shielding and covers in neutron spectra determinations

Most neutron spectrum determination methodologies ignore self-shielding effects in dosimetry foils and treat covers with an exponential attenuation model. This work provides a quantitative analysis of the approximations in this approach. It also provides a methodology for improving the fidelity of the treatment of the dosimetry sensor response to a level consistent with the users spectrum characterization approach. A library of correction functions for the energy-dependent sensor response has been compiled that addresses dosimetry foils/configurations in use at the Sandia National Laboratories Radiation Metrology Laboratory.

Effect of new cross section evaluations on neutron spectrum determination

Several neutron cross section libraries, such as ENDF/B-VI and IRDF-90, have been made available to the dosimetry community. Recommendations are made for the source selection of reaction cross sections that vary significantly among the libraries. In general, integral parameters from spectra obtained from unfold/adjustment codes using the cross sections will not significantly change. A 61-reaction compendium of dosimetry cross sections drawn from existing evaluations has been compiled for use at the Sandia National Laboratories Radiation Metrology Laboratory. This dosimetry library (SNLRML) is recommended for use in spectrum determination with unfold/adjustment methods. >