P.W. Marshall

Proton effects in charge-coupled devices

Basic mechanisms and ground-test data for radiation effects in solid-state imagers are reviewed, with a special emphasis on proton-induced effects on silicon charge-coupled devices (CCDs). For the proton fluxes encountered in the space environment, both transient ionization and displacement damage effects arise from single-particle interactions. In the former case, individual proton tracks will be seen; in the latter, dark-current spikes (or hot pixels) and trapping states that cause degradation in charge-transfer efficiency will be observed. Proton-induced displacement damage effects on dark current and charge transfer are considered in detail, and the practical implications for shielding, device hardening, and ground testing are discussed.

Particle-induced bit errors in high performance fiber optic data links for satellite data management

Experimental test methods and analysis tools are demonstrated to assess particle-induced bit errors on fiber optic link receivers for satellites. Susceptibility to direct ionization from low LET particles is quantified by analyzing proton and helium ion data as a function of particle LET. Existing single event analysis approaches are shown to apply, with appropriate modifications, to the regime of temporally (rather than spatially) distributed bits, even though the sensitivity to single events exceeds conventional memory technologies by orders of magnitude. The cross-section LET dependence follows a Weibull distribution at data rates from 200 to 1000 Mbps and at various incident optical power levels. The LET threshold for errors is shown, through both experiment and modeling, to be 0 in all cases! The error cross-section exhibits a strong inverse dependence on received optical power in the LET range where most orbital single events would occur, thus indicating that errors can be minimized by operating links with higher incident optical power. Also, an analytic model is described which incorporates the appropriate physical characteristics of the link as well as the optical and receiver electrical characteristics. Results indicate appropriate steps to assure suitable link performance even in severe particle orbits. >

High energy electron induced displacement damage in silicon

New measurements of displacement damage factors for electron-irradiated (4 to 53 MeV) bipolar silicon transistors have extended the correlation between nonionizing energy loss and damage factors reported previously another three orders of magnitude downward, to cover a total of six decades. To first order, the correlation remains linear for both n- and p-type silicon, but deviations are observed and explained in terms of differences in the fraction of initial vacancy interstitial pairs that recombines. These differences correlate linearly with the low-energy component of the PKA spectrum. Deep level transient spectroscopy measurements show oxygen- and dopant-related defect levels as well as divacancies. Defect concentrations scaled linearly with gain degradation, and no differences were seen between electron and proton plus neutron irradiated material. The results are consistent with a damage mechanism involving migration of vacancies to form well-separated stable defects. >

Heavy ion and proton-induced single event multiple upset

Individual ionizing heavy ion events are shown to cause two or more adjacent memory cells to change logic states in a high density CMOS SRAM. A majority of the upsets produced by normally incident heavy ions are due to single-particle events that causes a single cell to upset. However, for grazing angles a majority of the upsets produced by heavy-ion irradiation are due to single-particle events that cause two or more cells to change logic states. Experimental evidence of a single proton-induced spallation reaction that causes two adjacent memory cells to change logic states is presented. Results from a dual volume Monte-Carlo simulation code for proton-induced single-event multiple upsets are within a factor of three of experimental data for protons at normal incidence and 70 degrees.

Displacement damage extremes in silicon depletion regions

Measurements of proton-induced dark current increases in a Si CID (charge injection device) imager have been made following displacement damage by 12- and 63-MeV protons. Populations of 61504 pixels optimize statistics and make possible the first detailed study of rare events. To this end, extreme value statistics allow a quantitative treatment and lead to characterization of a rare device-dependent mechanism. Data comparing the response of two similar CID structures suggest that electric-field-enhanced emission is responsible for the largest dark current increases in the CID structure with the higher electric fields. Comparisons between observations and estimates based on new calculations of the recoil spectrum parameters demonstrate that the largest dark current increases can be predicted in the absence of high fields. In this case the inelastic recoil component of the recoil spectrum plays a dominant role in determining the large dark current increases. Implications for other materials are discussed. >

A comparison of Monte Carlo and analytic treatments of displacement damage in Si microvolumes

In this paper, we compare Monte Carlo and analytic calculations of displacement damage resulting from inelastic proton reactions in Si. These comparisons include the nonionizing energy loss rate, the mean recoil damage energy spectra, and their associated variance. In the limit of bulk material, both approaches are in good agreement. Sensitive volumes shrink and incident proton energies increase, the ranges of the spallation recoil fragments approach the smallest dimension of the microvolume, and the pixel-to-pixel damage variance increases rapidly. In this regime, a Monte Carlo approach is used to describe the damage energy distribution. Indeed, we show that such simulations predict the 63 MeV proton-induced dark current histograms more accurately than present analytic methods. The Monte Carlo code is also used to explore ground test fidelity issues for devices with small sensitive volumes. >

Single Event Upset cross sections at various data rates

We present data which show that Single Event Upset (SEU) cross section varies linearly with frequency for most devices tested. We show that the SEU cross section can increase dramatically away from a linear relationship when the test setup is not optimized, or when testing near the maximum operating frequency. We also observe non-linear behavior in some complex circuit topologies. Knowledge of the relationship between SEU cross section and frequency is important for estimates of on-orbit SEU rates.

Proton-induced displacement damage distributions and extremes in silicon microvolumes charge injection device

An analytic approach for determining the pixel-to-pixel distribution of particle-induced damage and damage extremes in microvolumes representative of focal plane array pixel geometries is presented. Comparisons between predicted and measured dark current distributions in a silicon charge injection device (CID) show excellent agreement for 12- and 63-MeV proton-induced damage. The calculated and measured damage extremes are compared using extreme value statistical analysis. The calculations reveal how high-energy recoils from proton-induced nuclear reactions strongly influence the pixel-to-pixel variation in damage as well as the damage extremes. A comparison between Si and GaAs pixels with equal volumes and equal 12-MeV proton fluences indicates that both the average damage and its variance are significantly greater in GaAs. >

Investigation of the oxygen-vacancy (A-center) defect complex profile in neutron irradiated high resistivity silicon junction particle detectors

The distribution of the A-center (oxygen vacancy) in neutron-damaged silicon detectors was studied using deep level transient spectroscopy. A-centers were found to be nearly uniformly distributed in the silicon wafer depth for medium-resistivity (0.1-0.2-k Omega -cm) silicon detectors and high-resistivity (>4-k Omega -cm) high-temperature (1200 degrees C) oxidized detectors. A positive filling pulse was needed to detect the A-centers in high-resistivity silicon detectors, and this effect was found to be dependent on the oxidation temperature. The A-center was not observed in a sample from a high-temperature oxidation with TCA having a very high carbon content. >

The generation lifetime damage factor and its variance in silicon

The generation damage factor and its variance for silicon are determined for proton energies of 12, 22, and 63 MeV, and for fission neutrons. Measurement of the variance is made for the first time using 61504 pixels of a charge-injection device. The variance is an intrinsic characteristic of damage due in part to differences in numbers and magnitudes of atomic recoils produced in each pixel. Calculations of the variance due to the recoil spectrum show that the magnitude of the relative variance is determined by Coulombic recoils. In fact, the experimental relative variance is almost two orders of magnitude less for fission-neutron-induced dark current increases than for the proton case. The relative variance was found to decrease as the proton energy was increased. The calculation shows that this trend is caused by the inelastic contribution to the relative variance of the recoil spectrum. >