Engineering topics

Hughart, David R.

Publications and source records attributed to Hughart, David R..

4 records

DOE OSTI2022

A Comparison of Radiation-Induced and High-Field Electrically Stress-Induced Interface Defects in Si/SiO 2 MOSFETs via Electrically Detected Magnetic Resonance

Here, we utilize electrically detected magnetic resonance (EDMR) measurements to compare high-field stressed, and gamma irradiated Si/SiO 2 metal–oxide–silicon (MOS) structures. We utilize spin-dependent recombination (SDR) EDMR detected using the Fitzgerald and Grove dc I-V approach to compare the effects of high-field electrical stressing and gamma irradiation on defect formation at and near the Si/SiO 2 interface. As anticipated, both greatly increase the concentration of P b centers (silicon dangling bonds at the interface) densities. The irradiation also generated a significant increase in the dc I-V EDMR response of E' centers (oxygen vacancies in the SiO 2 films), whereas the generation of an E' EDMR response in high-field stressing is much weaker than in the gamma irradiation case. These results likely suggest a difference in their physical distribution resulting from radiation damage and high electric field stressing.

42 ENGINEERING↗

DOE OSTI2022

Atomic-scale defects generated in the early/intermediate stages of dielectric breakdown in Si/SiO 2 transistors

In this study, electrically detected magnetic resonance and near-zero-field magnetoresistance measurements were used to study atomic-scale traps generated during high-field gate stressing in Si/SiO 2 MOSFETs. The defects observed are almost certainly important to time-dependent dielectric breakdown. The measurements were made with spin-dependent recombination current involving defects at and near the Si/SiO 2 boundary. The interface traps observed are P b0 and P b1 centers, which are silicon dangling bond defects. The ratio of P b0 /P b1 is dependent on the gate stressing polarity. Electrically detected magnetic resonance measurements also reveal generation of E' oxide defects near the Si/SiO 2 interface. Near-zero-field magnetoresistance measurements made throughout stressing reveal that the local hyperfine environment of the interface traps changes with stressing time; these changes are almost certainly due to the redistribution of hydrogen near the interface.

42 ENGINEERING↗

DOE OSTI2022

Probing the Atomic Scale Mechanisms of Time Dependent Dielectric Breakdown in Si/SiO 2 MOSFETs (June 2022)

Here we report on an atomic-scale study of trap generation in the initial/intermediate stages of time-dependent dielectric breakdown (TDDB) in high-field stressed (100) Si/SiO2 MOSFETs using two powerful analytical techniques: electrically detected magnetic resonance (EDMR) and near-zero-field magnetoresistance (NZFMR). We find the dominant EDMR-sensitive traps generated throughout the majority of the TDDB process to be silicon dangling bonds at the (100) Si/SiO2 interface (Pb0 and Pb1 centers) for both the spin-dependent recombination (SDR) and trap-assisted tunneling (SDTAT) processes. We find this generation to be linked to both changes in the calculated interface state densities as well as changes in the NZFMR spectra for recombination events at the interface, indicating a redistribution of mobile magnetic nuclei which we conclude could only be due to the redistribution of hydrogen at the interface. Additionally, we observe the generation of traps known as E’ centers in EDMR measurements at lower experimental temperatures via SDR measurements at the interface. Our work strongly suggests the involvement of a rate-limiting step in the tunneling process between the silicon dangling bonds generated at the interface and the ones generated throughout the oxide.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Ionizing radiation effects in SONOS-based neuromorphic inference accelerators

Here, we evaluate the sensitivity of neuromorphic inference accelerators based on Silicon-Oxide-Nitride-Oxide-Silicon (SONOS) charge trap memory arrays to total ionizing dose (TID) effects. Data retention statistics were collected for 16 Mbit of 40 nm SONOS digital memory exposed to ionizing radiation from a Co-60 source, showing good retention of the bits up to the maximum dose of 500 krad(Si). Using this data, we formulate a rate-equation-based model for the TID response of trapped charge carriers in the ONO stack, and predict the effect of TID on intermediate device states between ‘program’ and ‘erase’. This model is then used to simulate arrays of low-power, analog SONOS devices that store 8-bit neural network weights and support in situ matrix-vector multiplication. We evaluate the accuracy of the irradiated SONOS-based inference accelerator on two image recognition tasks – CIFAR-10 and the challenging ImageNet dataset – using state-of-the-art convolutional neural networks, such as ResNet-50. We find that across the datasets and neural networks evaluated, the accelerator tolerates a maximum TID between 10 krad(Si) and 100 krad(Si), with deeper networks being more susceptible to accuracy losses due to TID.

43 PARTICLE ACCELERATORS↗