Transient Annealing Research Articles

Stability study of laser irradiation of silicon diffused with arsenic

The stability of a metastable state produced by laser irradiation on arsenic-diffused Si is studied by measurement of the sheet resistance as a function of postlaser heat treatment between 300 and 900 °C. The good conductivity produced by laser irradiation decays at as low as 400 °C within 1 h. This preempts the advantage of good conductivity produced by transient annealing and sets a severe restriction on the application of transient processing.

Laser pulse annealing of ion-implanted GaAs

GaAs single-crystals wafers are implanted at room temperature with 400-keV Te+ ions to a dose of 1×1015 cm−2 to form an amorphous surface layer. The recrystallization of this layer is investigated by backscattering spectrometry and transmission electron microscopy after transient annealing by Q-switched ruby laser irradiation. An energy density threshold of about 1.0 J/cm2 exists above which the layer regrows epitaxially. Below the threshold the layer is polycrystalline; the grain size increases as the energy density approaches threshold. The results are analogous to those reported for the elemental semiconductors, Si and Ge. The threshold value observed is in good agreement with that predicted by the simple model successfully applied previously to Si and Ge.

Transient and Permanent Effects of Neutron Bombardment on a Commercially Available N-Buried-Channel CCD

A study of the effects of neutron bombardment on a commercially available n-buried-channel CCD (the Fairchild CCD 321A) has been performed. Both transient and permanent changes in dark current density and charge transfer inefficiency were examined. Considerable shortterm annealing of dark current was observed at room temperature for times up to ~104 sec following pulsed neutron irradiation, but negligible annealing of the transfer inefficiency was noted over the period from 1.5 sec to 1000 hours. This transient annealing difference suggests that transfer inefficiency increases are dominated by isolated, as opposed to clustered, defects under the present experimental conditions. Transfer inefficiency data obtained over the range 208-303° K yield trap energy levels with respect to the conduction band of 0.54 eV before irradiation and 0.33 eV after irradiation. The trap introduction rate was determined to be ~2.6 n-1 cm-1. In studies at cryogenic temperatures, a dependence of transfer inefficiency on clock frequency to the one-half power was noted. This dependence is attributed to a trap-limited diffusion process, with trapping occurring at the As donor level. Neutron irradiation at 77° K caused the transfer inefficiency to decrease, and a qualitative model for such behavior is given.

Explosive cladding of Cu/Cu systems: An electron microscopy study and a thermomechanical model

Transmission electron microscopy investigations were carried out on three explosively cladded Cu/Cu systems: equiaxial small grains, elongated large grains and single crystals. The investigation has shown that the cladding process leads to the formation of a ‘bond zone’ rather than to a ‘planar interface’. The ‘bond zone’ independent of the cladding system could be divided into a few ‘cladding affected regions’ symmetrically extending on both sides of the colliding surfaces. Each of the regions is characterized by a definite structural morphology. Selected area diffraction patterns show that in all three systems, regions characterized by similar structures exhibit the same zone axis, namely 〈111〉 or 〈011〉. The ‘metallurgical bond’ achieved in the cladding process is constituted of small equiaxed recrystallized grains which form the center of the ‘bond zone’. The array of structures in the ‘bond zone’ as well as the dislocation densities result from a combination of three basic processes taking place during explosion cladding: mechanical (deformation), thermal and mass transfer. A theoretical model has been suggested for the treatment of the effects of the thermomechanical process on the annihilation of dislocations and their density distribution after cladding and on the extent of mass transport during transient annealing. The theoretically predicted results were compared with the experimental results obtained by means of transmission electron microscopy. For the Cu/Cu systems the agreement was found to be reasonable. The model presented should enable thermomechanical processes in other similar explosion clad systems to be treated.

Tensile Properties and Annealing Characteristics of H. B. Robinson Spent Fuel Cladding

Studies of the tensile properties of Zircaloy-4 spent fuel cladding and their change with both isothermal and transient annealing have been conducted. The cladding was obtained from spent fuel rods irradiated to a maximum fuel burnup of 30 MWd/kg in the Carolina Power and Light H.B. Robinson power reactor. The yield and ultimate strengths of the as-received material decreased linearly with temperature from room temperature to 42.7/sup 0/C (800/sup 0/F). Uniform elongation was unaffected by temperature over the same range, while total elongation increased sharply between 329 and 371/sup 0/C (625 and 700/sup 0/F). At 482/sup 0/C (900/sup 0/F), properties reflected annealing that occurred during the test. The tensile properties at 371/sup 0/C (700/sup 0/F) were found to be strain rate dependent. The strength properties increased in strain rate, while the total elongation decreased. Uniform elongation exhibited no effect of strain rate. Evidence of dislocation channeling was observed. When the spent fuel cladding was annealed, radiation anneal hardening was noted during early stages in the annealing process. Annealing of irradiation-induced strengthening occurred rapidly at temperatures above 538/sup 0/C (1000/sup 0/F) under isothermal conditions and below 704/sup 0/C (1300/sup 0/F) under transient annealing conditions for heating rates of 28/sup 0/C (50/supmore » 0/F)/s or less. Ductility increases lagged the strength changes during annealing. A ductility minimum, as measured by total elongation, is not reflected in reduction-of-area measurements. The annealing behavior of cold-worked Zircaloy cladding was found to be significantly different from the irradiated material. Annealing was accompanied by a change in the isotropy of deformation as determined from tube wall and diameter measurements. The as-irradiated cladding exhibited essentially isotropic reductions, as opposed to the anisotropic reductions measured for both annealed cladding and unirradiated Zircaloy tubing.« less

Properties of Fully Decoded 256 Bit PMNOS/PMOS/SOS/AL-GATE Memory in Radiation Environment

Tests were conducted to characterize a fully decoded 256 bit PMNOS memory in a nuclear environment, These tests were as follows: (1) transient photo-current tests using 20 ns and 100 ns electron pulses from the AFCRL Linac, (2) transient annealing tests using 4,5 ?s electron pulses from the same Linac and also 365 ?s to 2.5 ms gamma pulses from the Aberdeen Pulse Reactor with doses up to 6 ? 104 rads (Si) and neutron fluences up to 2.6 x 1013n/cm2 pulse, (3) survivability tests using 20 ns gamma pulse from the AFCRL Flash X-ray, and (4) total dose using Cobalt-60 source at ECOM, Evans Laboratory. The results show that predictions of failure levels based on measurement made on the elementary devices are reasonably accurate. In addition, the results show that total does hardness levels of 3 ? 105 rads (Si) can be achieved without the use of hardened insulators. Most of the various failure levels were determined by the PMOS radiation effects.

Unified Model of Damage Annealing in CMOS, from Freeze-In to Transient Annealing

Irradiation studies at 76°K are described which demonstrate that radiation-produced holes in SiO2 are immobile at this temperature. If an electric field of either polarity is present in the SiO2 during 76°K irradiation, to sweep out the mobile electrons, the holes will virtually all be trapped where created and produce a uniform positive charge density in the oxide. This predicts ΔVT ∝ dox2 , which is observed. The magnitude of the observed shift is consistent with 18 eV required per hole-electron pair generated. All CMOS transistors of the same oxide thickness exhibit the same initial ΔVT, for a given exposure level, regardless of oxide type or process. The rate of annealing differs, both as a function of oxidation process and oxide thickness. Thin oxides have two advantages: a smaller ΔVT (stable shift) and faster annealing. If a CMOS device is irradiated for sufficient time at 76°K to build in an appreciable field, further irradiation with zero gate-substrate bias will produce little additional change in VT, since the field in the oxide tends to keep all generated electrons in the oxide, where they recombine with trapped holes. Room temperature annealing following a pulsed gamma exposure occurs in two regimes. The first regime can be quite fast, and occurs prior to 10-4 seconds. The magnitude of this early-time recovery is both process dependent and thickness dependent.

Transient and steady-state radiation response of CMOS/SOS devices

In an investigation to characterize the radiation properties of a pilot line product capability for manufacture of MNOS memory arrays, a CMOS test vehicle proposed as the building block for the array electronics was tested in a nuclear environment. The test device was a CD4007 configuration fabricated with an unhardened SiO 2 gate insulator on a sapphire substrate. Test devices were made with aluminum and silicon gates, and test samples were selected to satisfy the requirement of operating satisfactorily with a source-drain voltage ≤21.5 volts. All tests were conducted with a drain-source voltage of 20 volts and gate biases in the range of 0 to 20 volts. The tests were as follows: (1) transient photo-current, (2) transient annealing, (3) total dose, (4) survivability, and (5) neutrons. Facilities at AFCRL were used for the first four and the Aberdeen Pulse Reactor for the last test. The tests were conducted in the order listed above on different batches of samples for each test.

Interaction of nuclear environment with MNOS memory device

This paper reports on the results of an investigation of the interaction of a nuclear environment with a metal-gate, silicon nitride, silicon dioxide, silicon-on-sapphire memory device. The test device was a 63-bit (7 words by 9 bits) three-terminal device. The tests were as follows: (1) transient photocurrent, (2) transient annealing, (3) total dose, (4) survivability, and (5) neutron tests. Facilities at AFCRL were used in the first four tests and at the Aberdeen Pulse Reactor for the last test. Samples were irradiated under power. Both interrogated and passive words containing “1” and “0” states were investigated. Transient photocurrent and annealing data were obtained with the Linac operating in the electron mode. Electron energy was 10 MeV and pulse widths of 20 nanoseconds and 4. 5 microseconds were used. Total dose data was obtained with a cobalt-60 source as well as the reactor. A Flash X-ray generator operating in the electron mode (20-nanosecond pulse) was used to determine survivability rates.

Rapid annealing in irradiated CMOS transistors

Data characterizing the transient annealing of the gate threshold voltage of contemporary CMOS transistors following exposure to pulsed ionizing radiation are presented and discussed. Devices tested during the study include those fabricated on both bulk silicon and silicon-on-sapphire substrates. Silicon dioxide and aluminum oxide gate dielectrics are evaluated. Leakage current phenomena associated with charge formation in dielectric substrates or dielectric isolation layers are also considered.

Transient annealing of defects in irradiated silicon devices : B. L. Gregory and H. H. Sander, Proc. IEEE58, No. 9, September (1970), p. 1328

Transient annealing of defects in irradiated silicon devices : B. L. Gregory and H. H. Sander, Proc. IEEE58, No. 9, September (1970), p. 1328

Circuit Applications of Transient Annealing

A method is presented for predicting transistor annealing factors, in continuously ON devices, from a simple electrical measurement on a device. A nomograph has been derived from experimental data which relates annealing factors at any time after exposure to the electron density in the base-emitter space-charge region. The electron density in this region can be obtained by a single measurement of VEE, at the transistor operating current of interest. The effects of transient annealing are investigated for two typical circuit situations (power-inverter circuits and logic circuits) where devices may be OFF during irradiation and subsequently required to turn ON. The Darlington circuits utilized in typical power-inverters prolong the annealing, if OFF when exposed. The propagation delay of a NAND gate series pair is increased considerably at early times after a neutron exposure. This additional delay can result in reduced pulse width or lost pulses.

Transient annealing of defects in irradiated silicon devices

The annealing of radiation-produced defects in semiconductor devices is discussed briefly for60Co gamma-ray and 1-MeV electron damage, and in detail for fast-neutron damage. The effects on the reordering processes of varying the material parameters and the irradiation conditions are considered. Transient annealing of neutron damage near room temperature has been investigated for a wide variety of devices, and the data are presented in generalized form to increase their usefulness to device and circuit designers. Based on the experimental results, physical models are suggested for the reordering processes which occur during the annealing of neutron damage. Electron density is shown to be the most important factor governing the rate of transient annealing. Annealing factors are estimated for very early times (1 µs) following neutron exposure. Suggestions are made to minimize the effects of transient annealing on devices.

Injection Dependence of Transient Annealing in Neutron-Irradiated Silicon Devices

Studies have been performed to explore accurately the injection level and temperature dependence of transient annealing in neutron-irradiated P- and N-type silicon. In P-type material, the annealing factor in the 0 to 0.1 second time interval is very sensitive to the minority carrier injection level. For example, by varying the injection level from 10-5 to 10-1 the annealing factor at 0.001 second can be reduced from 10 to approximately 2. In contrast to the P-type results, the injection dependence observed in N-type silicon is very small and, furthermore, is in the opposite sense; i.e., an increase in the injection level causes an increase in the annealing factor. However, this study shows that this seemingly different behavior can be correlated on the basis of the hole-to-electron ratios of the different material types and resistivities. Annealing measurements performed in the temperature range from 180°to 300°K reaffirm the 0.3 eV activation energy previously found in P-type silicon and establish a value of 0.17 eV for N-type silicon.

Transient Annealing in Sekiconductor Devices Following Pulsed Neutron Irradiation

Transient annealing following pulsed neutron exposure has been investigated in silicon transistors and solar cells as a function of both irradiation temperature and injection level. In addition, experiments incorporating both X and gamma ray irradiation have demonstrated that transient annealing is a bulk, not a surface, effect. The observed annealing following neutron irradiation is considerably slower at low temperatures than at room temperature. At 213°K thg density of annealable defects remaining at 10-10 seconds after exposure is approximately three times that observed at this same time for a 300°K irradiation. The injection studies show that at any temperatures from 76°K to 300°K the presence of minority carrier injection considerably speeds the transient annealing processes. A model is presented which explains many of the features observed in transient annealing. Various kinetic processes are discussed in the derivation of this model, including diffusion-limited and generation-limited processes, and the results of the model are compared with experimental data.