Showing papers in "IEEE Transactions on Nuclear Science in 1999"

Radiation tolerant VLSI circuits in standard deep submicron CMOS technologies for the LHC experiments: practical design aspects

Abstract: We discuss design issues related to the extensive use of Enclosed Layout Transistors (ELT's) and guard rings in deep submicron CMOS technologies in order to improve radiation tolerance of ASIC's designed for the LHC experiments (the Large Hadron Collider at present under construction at CERN). We present novel aspects related to the use of ELT's: noise measured before and after irradiation up to 100 Mrad (SiO/sub 2/), a model to calculate the W/L ratio and matching properties of these devices. Some conclusions concerning the density and the speed of IC's conceived with this design approach are finally drawn.

Depth identification accuracy of a three layer phoswich PET detector module

Abstract: We describe a PET detector module that provides three levels of depth-of-interaction (DOI) information. The detector is a 9/spl times/9 array of 2 mm/spl times/2 mm/spl times/12 mm deep phoswich crystal elements, each consisting of 4 mm long LSO (entrance layer), GSO (middle layer) and BGO (exit layer) crystals joined optically together end-to-end. The BGO exit layer is directly coupled to a miniature position-sensitive photomultiplier tube (PSPMT). Delayed charge integration, a method that exploits differences in the light decay times of these scintillators, is used to determine the layer-of-interaction. DOI accuracy, measured by scanning a slit source of 511 keV radiation along the length of the module was 86% for the LSO layer, 80% for the GSO layer and 84% for the BGO layer. Energy resolution at 511 keV was 19% for LSO, 21% for GSO and 40% for BGO. Apparent gain differed between layers in the ratios 2.7:1.9:1.0 (LSO:GSO:BGO). Crystal separation was good between crystals in the LSO layer, acceptable between crystals in the GSO layer and poor between crystals in the BGO layer due, primarily, to the pronounced spatial nonlinearity of the PSPMT. The delayed charge integration method, however, does appear suitable for obtaining multi-level depth information when DOI effects are particularly significant, e.g. in very small ring diameter PET scanners for small animal imaging.

Prospects for time-of-flight PET using LSO scintillator

Abstract: We present measurements of the timing properties of lutetium orthosilicate (LSO) scintillator crystals coupled to a photomultiplier tube (PMT) and excited by 511 keV photons. These crystals have dimensions suitable for use in PET cameras (3/spl times/3/spl times/30 mm/sup 3/). Coincidence timing resolution of 475 ps fwhm is measured between detectors utilizing two such crystals, significantly worse than the 300 ps fwhm predicted based on first principles for small crystals and measured in 3 mm cubes. This degradation is found to be caused by the scintillation light undergoing multiple reflections at quasi-random angles within the scintillator crystal, which has two effects. First, it slows down the effective information propagation speed within the crystal (to an effective n/spl circ/=3.9-5.3). Since the incident annihilation photon travels with n=1, information from interactions at different depths arrives at the PMT with different time delays. Second, the random nature of the reflection angles (and path lengths) introduce dispersion and so a 10%-90% rise time of 1 ns to the optical signal.

Nonionizing energy loss (NIEL) for heavy ions

Abstract: The concept of nonionizing energy loss (NIEL) has been found useful for characterizing displacement damage effects in materials and devices. Published tabulations, however, are limited with respect to target materials, particle types and energies. In this paper we show how the NIEL database can be significantly expanded to include heavy ions in the coulombic limit by using the Monte Carlo code SRIM. The methodology used to extract NIEL from SRIM is described. This greatly adds to the number of materials and incident particles for which the NIEL concept can be applied. To show that values so derived are consistent with previous calculations, we compare alpha particle NIEL for GaAs derived from SRIM with a direct analytical calculation. The SRIM code is limited in that only coulombic interactions are considered. General rules of thumb are also described which permit prediction of NIEL for any target material over a large energy range. Tabulated values of NIEL for alpha particles incident on Si, GaAs and InP are presented.

Radiation effects on advanced flash memories

Abstract: Radiation tests of advanced flash memories including, multi-level flash technology, are compared with results from previous generations. Total dose failure levels are comparable to or lower than those of older technologies, but are likely still caused by degradation of the internal charge pump. Small numbers of read errors were observed during single event tests of the multi-level devices that appear to be caused by shifts in the sense amplifier detection levels or cell threshold shifts rather than loss of electrons off the floating gate.

Total variation regulated EM algorithm [SPECT reconstruction]

Abstract: An iterative Bayesian reconstruction algorithm based on the total variation (TV) norm constraint is proposed. The motivation for using TV regularization is that it is extremely effective for recovering edges of images. This paper extends the TV norm minimization constraint to the field of SPECT image reconstruction with a Poisson noise model. The regularization norm is included in the OSL-EM (one step late expectation maximization) algorithm. Unlike many other edge-preserving regularization techniques, the TV based method depends one parameter. Reconstructions of computer simulations and patient data show that the proposed algorithm has the capacity to smooth noise and maintain sharp edges without introducing over/under shoots and ripples around the edges.

Silicon detectors with 3-D electrode arrays: fabrication and initial test results

Abstract: The first three-dimensional detectors with n and p electrodes that penetrate through the silicon substrate have been fabricated. Some expected properties, including low depletion voltages, wide voltage plateaus before leakage current limits are reached, and rapid charge collection are reviewed. Fabrication steps and initial test results for leakage currents and infrared signal detection are covered. The authors conclude with a description of current work, including fabrication of active-edge detectors, ones with sensitive areas that should extend to their physical edge.

Determination of key parameters for SEU occurrence using 3-D full cell SRAM simulations

Abstract: A 3-D entire SRAM cell, based on a 0.35-/spl mu/m current CMOS technology, is simulated in this work with a DEVICE simulator. The transient current, resulting from a heavy ion strike in the most sensitive region of the cell, is studied as a function of the LET value, the cell layout and the ion penetration depth. A definition of the critical charge is proposed and two new methods are presented to compute this basic amount of charge only using SPICE simulations. Numerical applications are performed with two different generations of submicron CMOS technologies, including the determination of the sensitive thicknesses.

Probability model for worst case solar proton event fluences

Abstract: A predictive model of worst case solar proton event fluences is presented. It allows the expected worst case event fluence to be calculated for a given confidence level and for periods of time corresponding to space missions. The proton energy range is from >1 to >300 MeV, so that the model is useful for a variety of radiation effects applications. For each proton energy threshold, the maximum entropy principle is used to select the initial distribution of solar proton event fluences. This turns out to be a truncated power law, i.e., a power law for smaller event fluences that smoothly approaches zero at a maximum fluence. The strong agreement of the distribution with satellite data for the last three solar cycles indicates this description captures the essential features of a solar proton event fluence distribution. Extreme value theory is then applied to the initial distribution of events to obtain the model of worst case fluences.

SRAM based re-programmable FPGA for space applications

Abstract: An SRAM (static random access memory)-based reprogrammable FPGA (field programmable gate array) is investigated for space applications. A new commercial prototype, named the RS family, was used as an example for the investigation. The device is fabricated in a 0.25 /spl mu/m CMOS technology. Its architecture is reviewed to provide a better understanding of the impact of single event upset (SEU) on the device during operation. The SEU effect of different memories available on the device is evaluated. Heavy ion test data and SPICE simulations are used integrally to extract the threshold LET (linear energy transfer). Together with the saturation cross-section measurement from the layout, a rate prediction is done on each memory type. The SEU in the configuration SRAM is identified as the dominant failure mode and is discussed in detail. The single event transient error in combinational logic is also investigated and simulated by SPICE. SEU mitigation by hardening the memories and employing EDAC (error detection and correction) at the device level are presented. For the configuration SRAM (CSRAM) cell, the trade-off between resistor dc-coupling and redundancy hardening techniques are investigated with interesting results. Preliminary heavy ion test data show no sign of SEL (single event latch-up). With regard to ionizing radiation effects, the increase in static leakage current (static I/sub CC/) measured indicates a device tolerance of approximately 50 krad(Si).

A low altitude trapped proton model for solar minimum conditions based on SAMPEX/PET data

Abstract: We present a low altitude (below about 600 km) trapped proton model for solar minimum conditions (1994-1995), based on measurements made by the Proton/Electron Telescope onboard the SAMPEX satellite. Substantial differences are found with the low altitude part of the AP-8 MIN model: low energy fluxes appear to be overestimated by AP-8 MIN, while the new model predicts higher fluxes than AP-8 MIN for energies above about 30 MeV.

Comparison of Hotelling observer models and human observers in defect detection from myocardial SPECT imaging

Abstract: The objective of this study was to compare defect detection performance using a channelized Hotelling observer with radially symmetric channel (RSC) and oriented channels (OC) to that found in a previously reported human observer ROC study (see K.J. LaCroix and B.M.W. Tsui, "An evaluation ofthe effect of nonuniform attenuation compensation on defect detection for Tc-99m myocardial SPECT images," J. Nucl. Med., vol. 38, p. 19P, 1997). The observer's task involved the detection of left-ventricular myocardial perfusion defects for Tc-99m sestamibi myocardial perfusion SPECT images. The images were reconstructed using the filtered-backprojection (FBP) algorithm without attenuation compensation or the maximum-likelihood expectation-maximization (ML-EM) algorithm with non-uniform attenuation compensation (AC). This was performed for cases with and without significant attenuation artifacts. The areas under the ROC curves for the Hotelling observers were calculated and compared to those for the human observers. The ML-EM reconstructed images (with AC) had high defect detectability across all anatomy types, while with FBP, a lower detectability was found for cases where the reconstructed images contained attenuation artifacts in the myocardium. Similar trends between the channelized Hotelling observer and those from the human observer study were found with both radially symmetric channel (RSC) and oriented-channel (OC) models. Further investigation of the channel models is needed to determine the number and alignment of orientations and the number and cutoff of frequency bands to improve agreement between human and Hotelling observer results.

3D image reconstruction for a Compton SPECT camera model

Abstract: Proposes a 3D image reconstruction algorithm for a 3D Compton camera being developed at the University of Michigan. The authors present a mathematical model of the transition matrix of the camera which exploits symmetries by using an adapted spatial sampling pattern in the object domain. For each projection angle, the sampling pattern is uniform over a set of equispaced nested hemispheres. By using this sampling pattern the system matrix is reduced to a product of a (approximately) block circulant matrix and a sparse interpolation matrix. This representation reduces the very high storage and computation requirement inherent to 3D reconstruction using transition matrix inversion methods. The authors geometrically optimize their hemispherical sampling and propose a 3D volumetric interpolation. Finally, the authors present a 3D image reconstruction method which uses the Gauss-Seidel algorithm to minimize a penalized least square objective.

Comparison of radiation damage in silicon induced by proton and neutron irradiation

Abstract: The subject of radiation damage to Si detectors induced by 24-GeV/c protons and nuclear reactor neutrons has been studied. Detectors fabricated on single-crystal silicon enriched with various impurities have been tested. Significant differences in electrically active defects have been found between the various types of material. The results of the study suggest for the first time that the widely used nonionizing energy loss (NIEL) factors are insufficient for normalization of the electrically active damage in case of oxygen- and carbon-enriched silicon detectors. It has been found that a deliberate introduction of impurities into the semiconductor can affect the radiation hardness of silicon detectors.

Simulation technologies for cosmic ray neutron-induced soft errors: Models and simulation systems

Abstract: The authors review two types of simulators for the analysis of cosmic ray neutron-induced soft errors (SEs). One of them is the neutron-induced soft error simulator (NISES). A recently proposed nuclear reaction theory forms the foundation for the nuclear reaction database used in NISES. The other simulator, the simplified simulator MBGR, is based on a modified version of the burst generation rate (BGR) model. Both simulators accurately simulate neutron-induced SE rates (SERs). MBGR actually provides an easier and quicker estimation of neutron-induced SER's than NISES. On the other hand, NISES covers more applications: it simulates neutron-induced charge collection, multiple-bit SE, and /spl alpha/-induced SE analysis.

A model of radiation induced leakage current (RILC) in ultra-thin gate oxides

Abstract: An analytical model of Radiation Induced Leakage Current (RILC) has been developed for ultra-thin gate oxides submitted to high dose ionizing radiation. The model is based on the solution of the Schrodinger equation for a simplified oxide band structure, where RILC occurs through electron trap-assisted tunneling. The values of the model parameters have been calibrated by comparing the transmission probabilities obtained in this model with those obtained through the WKB method in the actual oxide band structure. No free fitting parameter has been introduced, and all physical constant values have been selected within the values found in literature. Different trap distributions have been considered as candidates, but the comparison between simulated and experimental curves have indicated that a double gaussian distribution in space and in energy grants the best fit of the experimental results for different ionizing particles, oxide fields during irradiation, radiation doses, and oxide thickness. Excellent matching has been found for both positive and negative RILC by using a single trap distribution. The trap density linearly increases with the radiation dose and decreases with the oxide field during irradiation. The trap distribution is spatially symmetrical in the oxide, centered in the middle of the oxide thickness, and is not modified as the cumulative dose increases.

Development of a small animal PET imaging device with resolution approaching 1 mm

Abstract: This paper describes progress in the design and construction of a single-plane PET tomograph having a spatial resolution approaching 1 mm. The system consists of a 12-cm diameter ring with 360 LSO (Lu/sub 2/SiO/sub 5/) detectors viewed by 30 photomultiplier tubes. Thin (5-mm) crystals and a low energy threshold are used. Crystals are identified using both position arithmetic and energy criteria. To date, the system construction has been completed, system tuning has been carried out and imaging studies have begun.

BUSFET-a radiation-hardened SOI transistor

Abstract: The total-dose hardness of SOI technology is limited by radiation-induced charge trapping in gate, field, and SOI buried oxides. Charge trapping in the buried oxide can lead to back-channel leakage and makes hardening SOI transistors more challenging than hardening bulk-silicon transistors. Two avenues for hardening the back-channel are 1) to use specially prepared SOI buried oxides that reduce the net amount of trapped positive charge or 2) to design transistors that are less sensitive to the effects of trapped charge in the buried oxide. In this work, we propose a partially-depleted SOI transistor structure for mitigating the effects of trapped charge in the buried oxide on radiation hardness. We call this structure the BUSFET-Body Under Source FET. The BUSFET utilizes a shallow source and a deep drain. As a result, the silicon depletion region at the back channel caused by radiation-induced charge trapping in the buried oxide does not form a conducting path between source and drain. Thus, the BUSFET structure design can significantly reduce radiation-induced back-channel leakage without using specially prepared buried oxides. Total dose hardness is achieved without degrading the intrinsic SEU or dose rate hardness of SOI technology. The effectiveness of the BUSFET structure for reducing total-dose back-channel leakage depends on several variables, including the top silicon film thickness and doping concentration, and the depth of the source.

A scaleable, radiation hardened shallow trench isolation

Abstract: Shallow trench isolation (STI) is rapidly replacing LOCOS (LOCal Oxidation of Silicon) as the device isolation process of choice. However, little work has been done to characterize the radiation-hardness capability of devices built with STI. In this paper, some of the basics of STI devices are examined, such as the radiation response of unhardened STI devices. We discuss several issues affecting the total dose hardness of unhardened STI. These issues have critical implications for the hardness of CMOS built using STI in commercial foundries. Finally, data from hardened STI devices are presented. Total dose hardened STI devices are demonstrated on devices with gate widths down to 0.5 /spl mu/m.

Proton degradation of light-emitting diodes

Abstract: Proton degradation is investigated for several types of light-emitting diodes with wavelengths in the near infrared region. Several basic light-emitting diode (LED) technologies are compared, including homojunction and double-heterojunction devices. Homojunction LEDs fabricated with amphoteric dopants are far more sensitive to displacement damage than double-heterojunction LEDs, and are strongly affected by injection-enhanced annealing. Unit-to-unit variability remains an important issue for all LED technologies. For technologies, degradation of the forward voltage characteristics appears to be more significant than degradation of light output.

Dynamic SPECT imaging using a single camera rotation (dSPECT)

Abstract: Techniques to estimate dynamic parameters from the data acquired during a single rotation of a standard (single, dual and triple head) SPECT camera are being investigated. The performance of the constrained linear least squares (C-LS) method is being assessed. This method uses a simple assumption that the activity within each pixel at a given projection angle is less than or equal to the activity seen at the prior angle. In simulations, we have used an analytical heart model with single and bi-exponential activity decrease, with and without background activity and statistical noise in the data. Studies using a single, a 90/spl deg/ dual and a 120/spl deg/ triple head camera were modeled with a total acquisition time of 20 min, performing rotation over 1800. These simulations were followed by experimental scans of our dynamic phantom used with and without attenuation and subsequently a preliminary study involving a human subject was performed. Dynamic images from simulations and phantom experiments reconstruct well with an accuracy of reconstructed half-lives for most reconstructions within a 20% error range. Some streaking artifacts which are evident in images corresponding to early reconstruction times subside upon incorporating further constraints and regularizations.

Scintillation properties of RbGd/sub 2/Br/sub 7/:Ce. Advantages and limitations

Abstract: The scintillation properties of RbGd/sub 2/Br/sub 7/ crystals, doped with Ce/sup 3+/ concentrations of 0.02, 0.11, 0.88, 2.05, 4.1, and 9.8%, are studied under X-ray and /spl gamma/-quanta excitations. For the RbGd/sub 2/Br/sub 7/ sample doped with 9.8% Ce, the authors measured a light yield of 56000/spl plusmn/6000 photons per MeV of absorbed /spl gamma/-ray energy with a main decay time of 43/spl plusmn/1 ns, using a Hamamatsu R1791 photomultiplier (PMT), a /sup 137/Cs radioactive source, and a shaping time of 10 /spl mu/s. A time resolution of 790/spl plusmn/10 ps was measured for the RbGd/sub 2/Br/sub 7/:9.8% Ce compound, using BaF/sub 2/ as second scintillator, two XP2020Q PMTs, a /sup 22/Na source, and an energy threshold set at E/spl ges/511 keV. With the R1791 PMT, an energy resolution of 4.1% (FWHM over peak position) for the 662-keV full absorption peak has been observed for two crystals of 7/spl times/4/spl times/2 mm/sup 3/ and 15/spl times/5/spl times/1 mm/sup 3/ with 4.1 and 9.8% Ce content, respectively. Moreover, the nonproportional responses of three RbGd/sub 2/Br/sub 7/:Ce compounds with different concentrations (0.11, 2.05, and 9.8%) were studied revealing an almost-constant light output response from 17.4 keV to 1 MeV. These properties are compared to three other well-known scintillators: NaI:Tl, CsI:Tl, and Lu/sub 2/SiO/sub 5/:Ce.

Identification of degradation mechanisms in a bipolar linear voltage comparator through correlation of transistor and circuit response

Abstract: The input bias current (I/sub IB/) of the National LM111 voltage comparator exhibits a non-monotonic response to total dose irradiations at various dose rates. At low total doses, below 100 krad(SiO/sub 2/), increased I/sub IB/ is due primarily to gain degradation in the circuit's input transistors. At high total doses, above 100 krad(SiO/sub 2/), I/sub IB/ shows a downward trend that indicates the influence of compensating circuit mechanisms. Through correlation of transistor and circuit response, the transistors responsible for these compensating mechanisms are identified. Non-input transistors in the circuit's input stage lower the emitter-base operating point voltage of the input device. Lower emitter-base voltages reduce the base current supplied by the input transistors, causing a moderate "recovery" in the circuit response.

Neutron induced single-word multiple-bit upset in SRAM

Abstract: The single-word multiple-bit upset (SMU) frequency for nine commercial static random access memories (SRAM) have been evaluated at eight different neutron energies; 0-11 MeV, 14 MeV, 22 MeV, 35 MeV, 45 MeV, 75 MeV, 96 MeV, 160 MeV. The SRAM types used at these experiments have sizes from 256 Kbit up to 1 Mbit, with date-codes ranging from 9209 up to 9809. The result showed a slightly rising dependence on the neutron energy. Also experiments at two neutron energies, 45 MeV and 96 MeV, were performed where the supply voltage influence on the SMU-rate was studied. Five device types were used at 96 MeV and the supply voltage was changed between 5 V, 3.3 V and 2.5 V. At 45 MeV three devices at 5 V and 3.3 V were irradiated. The experiments showed a relation between the amount of total upset and SMU that indicates no clear supply voltage dependence.

Worst case total dose radiation response of 0.35 /spl mu/m SOI CMOSFETs

Abstract: Through experimental results and analysis by TSUPREM4/MEDICI simulations, the worst case back gate total dose bias condition is established for body tied SOI NMOSFETs. Utilizing the worst-case bias condition, a recently proposed model that describes the back n-channel threshold voltage shift as a function of total dose, TSUPREM4/MEDICI simulations, and circuit level SPICE simulations, a methodology to model post-rad standby current is developed and presented. This methodology requires the extraction of fundamental starting material/material preparation constants, and then can be utilized to examine post-rad stand-by current at the device and circuit level as function of total dose. Good agreement between experimental results and simulations is demonstrated.

Single event effects in static and dynamic registers in a 0.25 /spl mu/m CMOS technology

Abstract: We have studied single event effects in static and dynamic registers designed in a quarter micron CMOS process. In our design, we systematically used guardrings and enclosed (edgeless) transistor geometry to improve the total dose tolerance. This design technique improved both the SEL and SEU sensitivity of the circuits. Using SPICE simulations, the measured smooth transition of the cross-section curve between LET threshold and saturation has been traced to the presence of four different upset modes, each corresponding to a different critical charge and sensitive area. A new architecture to protect the content of storage cells has been developed, and a threshold LET around 89 MeV cm/sup 2/ mg/sup -1/ has been measured for this cell at a power supply voltage of 2 V.

Continuous charge restoration in semiconductor detectors by means of the gate-to-drain current of the integrated front-end JFET

Abstract: A continuous reset method for discharging the leakage current and the signal charge in semiconductor radiation detectors with on-chip electronics has been studied and experimentally verified. The charge collected at the output electrode of the detector is discharged by means of the gate-to-drain current of the front-end JFET integrated in the detector itself. The suitable value of gate current is reached by means of a "weak" avalanche breakdown mechanism which occurs in a high-field region of the transistor channel. When the JFET is operated in a source follower configuration, this mechanism is self-adapting to new values of leakage or signal currents.

MOSFET dosimetry of an X-ray microbeam

Abstract: A metal-oxide-semiconductor field effect transistor (MOSFET) has been used as a radiation monitor to map the profile of a variety of X-ray radiation microbeams. The present work studies the role of the topology of the MOSFET gate oxide when mapping radiation microbeams. The "edge-on" MOSFET has been introduced for microbeam mapping and the spatial resolution of the "edge-on" MOSFET is investigated. Comparison is made with other mapping techniques including, Gafchromic/sup TM/ film, radiographic film, ionization chamber. The results clearly demonstrate the superiority of the "edge-on" MOSFET when applied to mapping of narrow radiation beams. The spatial resolution of the "edge-on" MOSFET is estimated to be 1 /spl mu/m and appears to be limited by the width of the gate oxide thickness.

Ab initio calculations of H/sup +/ energetics in SiO/sub 2/: Implications for transport

Abstract: The process of H/sup +/ diffusion in SiO/sub 2/ is studied using first-principles density-functional theory. This is done by examining stable bonding sites, migration pathways, and barrier heights of H/sup +/ in /spl alpha/-quartz. The H/sup +/ is found to bind to the bridging oxygen site in a variety of ways, but always displaying an attraction to neighboring oxygen atoms. This attraction provides a mechanism for understanding the diffusion pathways and barrier heights for this system. The relative stability of H/sup +/ in SiO/sub 2/ is also discussed in the context of transport.

Reconstruction of multi-energy X-ray computed tomography images of laboratory mice

Abstract: A new X-ray computed tomography (CT) system is being developed at Oak Ridge National Laboratory to image laboratory mice for the purpose of rapid phenotype screening and identification. One implementation of this CT system allows simultaneous capture of several sets of sinogram data, each having a unique X-ray energy distribution. The goals of this paper are to (1) identify issues associated with the reconstruction of this energy-dependent data and (2) suggest preliminary approaches to address these issues. Due to varying numbers of photon counts within each set, both traditional (filtered backprojection, or FBP) and statistical (maximum likelihood, or ML) tomographic image reconstruction techniques have been applied to the energy-dependent sinogram data. Results of reconstructed images using both algorithms on sinogram data (high- and low-count) are presented. Also, tissue contrast within the energy-dependent images is compared to known X-ray attenuation coefficients of soft tissue (e.g. muscle, bone, and fat).