Search for: All records

NiO/Ga 2 O 3 heterojunction rectifiers were exposed to 1 Mrad fluences of Co-60 γ-rays either with or without reverse biases. While there is a small component of Compton electrons (600 keV), generated via the interaction of 1.17 and 1.33 MeV gamma photons with the semiconductor, which in turn can lead to displacement damage, most of the energy is lost to ionization. The effect of the exposure to radiation is a 1000× reduction in forward current and a 100× increase in reverse current in the rectifiers, which is independent of whether the devices were biased during this step. The on–off ratio is also reduced by almost five orders of magnitude. There is a slight reduction in carrier concentration in the Ga 2 O 3 drift region, with an effective carrier removal rate of <4 cm −1 . The changes in electrical characteristics are reversible by application of short forward current pulses during repeated measurement of the current–voltage characteristics at room temperature. There are no permanent total ionizing dose effects present in the rectifiers to 1 Mad fluences, which along with their resistance to displacement damage effects indicate that these devices may be well-suited to harsh terrestrial and space radiation applications if appropriate bias sequencesmore »are implemented to reverse the radiation-induced changes.« less

Strain localization in microelectronic devices commonly arises from device geometry, materials, and fabrication processing. In this study, we controllably relieve the local strain field of AlGaN/GaN HEMTs by milling micro-trenches underneath the channel and compare the device performance as a function of the relieved strain as well as radiation dosage. Micro-Raman results suggest that the trenches locally relax the strain in device layers, decreasing the 2DEG density and mobility. Intriguingly, such strain relaxation is shown to minimize the radiation damage, measured after 10 Mrads of⁶⁰Co-gamma exposure. For example, a 6-trench device showed only ∼8% and ∼6% decrease in saturation drain current and maximum transconductance, respectively, compared to corresponding values of ∼15% and ∼30% in a no-trench device. Negative and positive threshold voltage shifts are observed in 6-trench and no-trench devices, respectively, after gamma radiation. We hypothesize that the extent of gamma radiation damage depends on the strain level in the devices. Thus, even though milling a trench decreases 2DEG mobility, such decrease under gamma radiation is far less in a 6-trench device (∼1.5%) compared to a no-trench device (∼20%) with higher built-in strain.

Radiation damage in electronic devices is known to be influenced by physics, design, and materials system. Here, we report the effects of biasing state (such as ON and OFF) and pre-existing damage in GaN high electron mobility transistors exposed to γ radiation. Controlled and accelerated DC biasing was used to prestress the devices, which showed significant degradation in device characteristics compared to pristine devices under ON and OFF states after γ irradiation. The experiment is performed in situ for the ON-state to investigate transient effects during irradiation until the total dose reaches 10 Mrad. It shows that threshold voltage, maximum transconductance, and leakage current initially decrease with dosage but slowly converge to a steady value at higher doses. After 10 Mrad irradiation, the OFF-state device demonstrates larger R_ONand one order of magnitude increased leakage current compared to the ON-state irradiated device. The micro-Raman study also confirms that the ON-state operation shows more radiation hardness than OFF and prestressed devices. Prestressed devices generate the highest threshold voltage shift from −2.85 to −2.49 V and two orders of magnitude higher leakage current with decreased saturation current after irradiation. These findings indicate that high electric fields during stressing can generate defects by modifying strain distribution, andmore »higher defect density can not only create more charges during irradiation but also accelerate the diffusion process from the ionizing track to the nearest collector and consequently degrade device performances.

« less

Halide vapor phase epitaxial (HVPE) Ga 2 O 3 films were grown on c-plane sapphire and diamond substrates at temperatures up to 550 °C without the use of a barrier dielectric layer to protect the diamond surface. Corundum phase α-Ga 2 O 3 was grown on the sapphire substrates, whereas the growth on diamond resulted in regions of nanocrystalline β-Ga 2 O 3 (nc-β-Ga 2 O 3 ) when oxygen was present in the HVPE reactor only during film growth. X-ray diffraction confirmed the growth of α-Ga 2 O 3 on sapphire but failed to detect any β-Ga 2 O 3 reflections from the films grown on diamond. These films were further characterized via Raman spectroscopy, which revealed the β-Ga 2 O 3 phase of these films. Transmission electron microscopy demonstrated the nanocrystalline character of these films. From cathodoluminescence spectra, three emission bands, UVL′, UVL, and BL, were observed for both the α-Ga 2 O 3 /sapphire and nc-Ga 2 O 3 /diamond, and these bands were centered at approximately 3.7, 3.2, and 2.7 eV.

Localized residual stress and elastic strain concentrations in microelectronic devices often affect the electronic performance, resistance to thermomechanical damage, and, likely, radiation tolerance. A primary challenge for the characterization of these concentrations is that they exist over sub-[Formula: see text]m length-scales, precluding their characterization by more traditional residual stress measurement techniques. Here, we demonstrate the use of synchrotron x-ray-based differential aperture x-ray microscopy (DAXM) as a viable, non-destructive means to characterize these stress and strain concentrations in a depth-resolved manner. DAXM is used to map two-dimensional strain fields between the source and the drain in a gallium nitride (GaN) layer within high electron mobility transistors (HEMTs) with sub-[Formula: see text]m spatial resolution. Strain fields at various positions in both pristine and irradiated HEMT specimens are presented in addition to a preliminary stress analysis to estimate the distribution of various stress components within the GaN layer. [Formula: see text]-irradiation is found to significantly reduce the lattice plane spacing in the GaN along the sample normal direction, which is attributed to radiation damage in transistor components bonded to the GaN during irradiation.

Focused Ga + ion milling of lightly Si-doped, n-type Ga 2 O 3 was performed with 2–30 kV ions at normal incidence and beam currents that were a function of beam voltage, 65 nA for 30 kV, 26 nA for 10 kV, 13 nA for 5 kV, and 7.1 nA for 2 kV, to keep the milling depth constant at 100 nm. Approximate milling rates were 15, 6, 2.75, and 1.5  μm 3 /s for 30, 10, 5, and 2 kV, respectively. The electrical effects of the ion damage were characterized by Schottky barrier height and diode ideality factor on vertical rectifier structures comprising 10  μm epitaxial n-Ga 2 O 3 on n + Ga 2 O 3 substrates, while the structural damage was imaged by transmission electron microscopy. The reverse bias leakage was largely unaffected even by milling at 30 kV beam energy, while the forward current-voltage characteristics showed significant deterioration at 5 kV, with an increase in the ideality factor from 1.25 to 2.25. The I–V characteristics no longer showed rectification for the 30 kV condition. Subsequent annealing up to 400 °C produced substantial recovery of the I–V characteristics for all beam energies and was sufficient to restore the initial ideality factor completely for beam energies up to 5 kV. Even the 30 kV-exposedmore »rectifiers showed a recovery of the ideality factor to 1.8. The surface morphology of the ion-milled Ga 2 O 3 was smooth even at 30 kV ion energy, with no evidence for preferential sputtering of the oxygen. The surface region was not amorphized by extended ion milling (35 min) at 5 kV with the samples held at 25 °C, as determined by electron diffraction patterns, and significant recovery of the lattice order was observed after annealing at 400 °C.« less

We present a review of the published experimental and simulation radiation damage results in Ga 2 O 3 . All of the polytypes of Ga 2 O 3 are expected to show similar radiation resistance as GaN and SiC, considering their average bond strengths. However, this is not enough to explain the orders of magnitude difference of the relative resistance to radiation damage of these materials compared to GaAs and dynamic annealing of defects is much more effective in Ga 2 O 3 . It is important to examine the effect of all types of radiation, given that Ga 2 O 3 devices will potentially be deployed both in space and terrestrial applications. Octahedral gallium monovacancies are the main defects produced under most radiation conditions because of the larger cross-section for interaction compared to oxygen vacancies. Proton irradiation introduces two main paramagnetic defects in Ga 2 O 3 , which are stable at room temperature. Charge carrier removal can be explained by Fermi-level pinning far from the conduction band minimum due to gallium interstitials (Ga i ), vacancies (V Ga ), and antisites (Ga O ). One of the most important parameters to establish is the carrier removal rate formore »each type of radiation, since this directly impacts the current in devices such as transistors or rectifiers. When compared to the displacement damage predicted by the Stopping and Range of Ions in Matter(SRIM) code, the carrier removal rates are generally much lower and take into account the electrical nature of the defects created. With few experimental or simulation studies on single event effects (SEE) in Ga 2 O 3 , it is apparent that while other wide bandgap semiconductors like SiC and GaN are robust against displacement damage and total ionizing dose, they display significant vulnerability to single event effects at high Linear Energy Transfer (LET) and at much lower biases than expected. We have analyzed the transient response of β -Ga 2 O 3 rectifiers to heavy-ion strikes via TCAD simulations. Using field metal rings improves the breakdown voltage and biasing those rings can help control the breakdown voltage. Such biased rings help in the removal of the charge deposited by the ion strike.« less

As electronic systems become larger and more complex, detection of the most vulnerable regions (MVR) to radiation exposure becomes more difficult and time consuming. We present a heuristic approach where the mechanical and thermal aspects of devices are exploited to quickly identify MVRs. Our approach involves the topological mapping of two device conditions. The first condition identifies regions with the highest mechanical strain or density of defects and interfaces via thermal wave probing and phase analysis. The second condition identifies regions with high electrical field. It is hypothesized that the region with the highest thermal wave penetration resistance and electrical field will exhibit the highest sensitivity to incoming radiation for single events and potentially, total ionizing dose. Our approach implements a simplistic design that improves analysis time by ∼2–3 orders of magnitude over current radiation sensitivity mapping methods. The design is demonstrated on the well-studied operational amplifier LM124, which shows agreement with the literature in identifying sensitive transistors–QR1, Q9, and Q18–with relatively high phase percentile values (>70%) and ΔT percentiles (>50%), satisfying conditions for elevated radiation susceptibility. This is followed by experimental results on a static random access memory (HM-6504) and a Xilinx Artix-7 35 T system on a chip.