Radiation Response Of Devices Research Articles

Charge Trapping in Al2O3/$\beta$ -Ga2O3-Based MOS Capacitors

Trapping characteristics of MOS structures with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\beta $ </tex-math></inline-formula> -Ga <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> substrates and Al <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> gate dielectrics are evaluated via constant-voltage stress and X-ray irradiation. Traps that affect bias-induced charging are located primarily in the Al <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> dielectric layer, and are distributed broadly in time and/or energy. Stress-induced flatband voltage shifts are reduced by N <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> annealing. Trap-assisted tunneling is shown to be responsible for the observed gate leakage. Hole trapping in the Al <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> dielectric layer dominates device radiation response. The relatively modest radiation-induced charge trapping observed in these devices is promising for the potential future use of Al <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> / <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\beta $ </tex-math></inline-formula> -Ga <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> O <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> devices in a space environment.

Editors' Choice—On the Radiation Tolerance of AlGaN/GaN HEMTs

The radiation tolerance of AlGaN/GaN high electron mobility transistors (HEMTs) fabricated on high quality, low threading dislocation density (TDD) ammonothermal GaN and hydride vapor phase epitaxy GaN substrates was studied and compared to the radiation response of devices on SiC substrates where the TDD is 104 times higher. Hall and transport measurements were performed as a function of 2 MeV proton fluence. The threading dislocation density had no effect on the radiation response. Comparing the results with published data reveals that almost all irradiated GaN-based HEMTs respond to radiation damage similarly regardless of differences in initial film quality, device structure, aluminum mole fraction, etc. AlGaAs/GaAs HEMTs are also shown to behave similarly but are around ten times more sensitive to radiation damage than GaN-based HEMTs. Known values of the displacement energy thresholds in GaN and GaAs are used to calculate that 36% fewer defects are created in GaN than in GaAs, which is too small to cause a 1000% difference in radiation sensitivity between GaN- and GaAs-based HEMTs. An alternative explanation is proposed in which the piezoelectric field at the AlGaN/GaN interface causes scattered carriers to be reinjected into the 2DEG channel, thereby mitigating some of the harmful radiation effects.

Characterization of Radiation Damage in Carbon Nanotube Field-Effect Transistors

<?Pub Dtl=""?> Carbon nanotube field-effect transistors with <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">${\rm Al}_{2}{\rm O}_{3}$</tex></formula> gate dielectrics have been characterized before and after 1 MeV electron irradiation. Pre-irradiation interface trap densities determined by charge pumping measurements were on the order of <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$10^{11}~\hbox{cm}^{\hbox{-}2}\hbox{eV}^{\hbox{-}1}$</tex> </formula> . Following irradiation to fluences of <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">$10^{16}~\hbox{e/cm}^2$</tex></formula> and <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex Notation="TeX">$10^{17}~\hbox{e/cm}^2$</tex></formula> , the measured interface trap densities decreased significantly for all devices. This result is largely attributed to the removal of molecular adsorbates that act as trap sites on the nanotubes (de-doping) while the devices are irradiated under vacuum. Negative shifts in the transistor <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$I\hbox{-}V$</tex></formula> curves with irradiation indicate radiation-induced hole trapping in the gate dielectric, with some recovery after room temperature annealing. No significant changes were observed in the Raman spectra following irradiation, indicating the intrinsic structure of the carbon nanotubes remained unchanged. These results suggest that the radiation response of these devices is dominated by charge trapping in the oxide layers and at the carbon nanotube surfaces. Furthermore, the charge pumping technique can be used as a tool for probing surface adsorbates in carbon nanotube devices, and elucidating their role in the device radiation response.

Radiation Effects in Carbon Nanoelectronics

We experimentally investigate the effects of Co-60 irradiation on the electrical properties of single-walled carbon nanotube and graphene field-effect transistors. We observe significant differences in the radiation response of devices depending on their irradiation environment, and confirm that, under controlled conditions, standard dielectric hardening approaches are applicable to carbon nanoelectronics devices.

Effects of device aging on microelectronics radiation response and reliability

We review recent work that shows that, for several kinds of devices built by several different manufacturers, MOS and bipolar device radiation response can change significantly with pre-irradiation elevated temperature stress and/or with aging time after device fabrication or packaging. The primary effects that have been observed are related to changes in interface-trap charge buildup, but some effects are also noted for charge trapped in the oxide. On the basis of a wide array of experimental data and first-principles calculations, these changes are attributed to the motion and reactions of hydrogenous species, e.g., protons and water. Similar reactions with hydrogen can also affect the reliability of MOS devices and integrated circuits, as illustrated here for the case of negative bias-temperature instability for high- κ alternative gate dielectrics. Heating the device accelerates some kinds of aging effects, but reverses the effects of others. These observations are consistent with a dual role for hydrogen in improving or degrading device performance, depending on its particular incorporation in the device and subsequent motion and reactions. These results emphasize that measurements used to provide estimates of MOS radiation response and long-term reliability can vary significantly with time. In some cases, typical margins built into radiation response and reliability tests may be inadequate to account for aging-related changes in device performance.

Elimination of Enhanced Low-Dose-Rate Sensitivity in Linear Bipolar Devices Using Silicon-Carbide Passivation

The type of final chip passivation layer used to fabricate linear bipolar circuits can have a major impact on the total dose hardness of some circuits. It is demonstrated that National Semiconductor Corporation linear bipolar devices fabricated with only an amorphous silicon carbide passivation layer do not exhibit enhanced low-dose-rate sensitivity (ELDRS), while devices from the same production lot fabricated with other types of passivation layers are ELDRS sensitive. SiC passivation possesses mechanical, electrical and chemical properties that make it compatible with linear device fabrication processes. These properties of SiC passivation layers, combined with the excellent radiation response of devices passivated with SiC, make SiC passivation layers a very attractive choice for devices packaged in either ceramic or plastic-encapsulated packages for use in space environments

The impact of mechanical stress on the total-dose response of linear bipolar transistors with various passivation layers

In this work, we investigate how externally applied mechanical stress impacts the total dose hardness and enhanced low-dose-rate sensitivity of linear bipolar ICs fabricated with different passivation layers. To determine the effects of externally applied stress and/or radiation exposure on the current gain, various compressive or tensile stresses were applied to lateral PNP bipolar transistors at the wafer level, using a cantilever method. The devices were then exposed to 10-keV X-rays or Cs-137 /spl gamma/-rays. The effect of externally applied stress on pre and post irradiation gain degradation is found to be relatively minor compared to the effects of changes in passivation layers on the radiation response of devices.

Predicting NMOS device radiation response at different dose rates in γ-ray environment

Linear response theory was used to predict threshold-voltage shifts for CC4007-NMOS device at dose rates of 0.1, 2.3, 44 and 91rad(Si)/s. These predictions were compared with the threshold-voltage shifts obtained after 60Co irradiations actually performed at these dose rates, and the agreement is excellent. Also, we use the linear response theory to predict the response of CC4007-NMOS device during radiation and 25℃ annealling for high and low-dose-rates. According to the predictions, under the same conditions of radiation and annealling, the threshold-voltage shifts caused by high dose-rate irradiation after room temperature annealling were equal to that by low-dose-rate irradiation within the limit of error, but the total times for both were the same. Finally, the failure doses for CC4007-NMOS device at different dose rates of irradiation were predicted also using the linear response theory.

Modeling BJT radiation response with non-uniform energy distributions of interface traps

Radiation-induced oxide defects that degrade electrical characteristics of BJTs can be measured with the use of gated diodes. The buildup of defects and their effect on device radiation response are modeled with computer simulation.

Effect of Key Processing Steps on Post Radiation Midgap Voltage

A variety of processing steps were investigated to determine the steps that affect a device's radiation response. Although emphasis is traditionally placed on gate oxide and thermal processing, the physical structure during processing and energetic processes can play a role as well. We find that processing subsequent to the gate oxide layer can have a significant effect on the radiation hardness of a device. Optimizing a process for radiation hardness requires investigation into the entire process and not just optimization of the gate oxide or a set of individual processing steps.

Radiation evaluation of commercial ferroelectric nonvolatile memories

Ferroelectric (FE) on complementary metal-oxide-semiconductor (CMOS) 4-bit nonvolatile memories, 8-bit octal latches (with and without FE), and process control test chips were used to establish a baseline characterization of the radiation response of CMOS/FE integrated devices and to determine whether the additional FE processing caused significant degradation to the baseline CMOS process. Functional failure of all 4-bit memories and octal latches occurred at total doses of between 2 and 4 krad(Si), most likely due to field-oxide effects in the underlying CMOS. No significant difference was observed between the radiation response of devices with and without the FE film in this commercial process. >