2026 IEEE NSREC TECHNICAL PROGRAM
SESSION A SCHEDULE
PUERTO RICO CONVENTION CENTER, SAN JUAN, PR
TUESDAY, JULY 21, 2026
SESSION A
Ballroom A
SINGLE-EVENT EFFECTS: MECHANISMS AND MODELING
9:05 AM
SESSION INTRODUCTION
Chair: Jeffrey Warner (Northrop Grumman)
A-1
9:10 AM
Low-loss Vertical β-Ga2O3 High-Power Diodes with Engineered Contacts: SEB Assessment and Field-Plate Effects
S. Islam1, P. Das2, D. Ball1, J. Speck3, E. Farzana2, D. Fleetwood1, R. Schrimpf1
1. Vanderbilt University, USA 2. Iowa State University, USA 3. University of California, USA
Vertical β-Ga₂O₃ diodes with engineered Pt/PtOₓ/thin-Pt Schottky contacts exhibit reduced turn-on voltage. Single-event burnout is experimentally evaluated on structures with and without field plates. TCAD analysis provides insight into the responsible mechanisms.
A-2
9:25 AM
Effect of Multi-Soft-Program Scheme on Single-Event Upset in 3D Charge-Trap NAND Flash Memory
J. Park1, V. Shastry1, M. Breeding2, S. Yi1
1. Texas A&M University, USA 2. Sandia National Laboratories, USA
Effects of multi-soft-program on retention and single-event upset in 176-layer 3D charge- trap (CT) NAND flash are investigated, clarifying the influence of trap occupancy in the CT layer on retention characteristics and heavy-ion-induced threshold voltage loss.
A-3
9:40 AM
Angular Effects on Single-Event Vulnerability at the 3- and 5-nm Bulk FinFET Nodes
J. Kronenberg1, S. Tolson1, X. Zhao1, Y. Xiong1, N. Pieper1, D. Ball1, B. Bhuva1
1. Vanderbilt University, USA
Angular single-event effects at 3- and 5-nm bulk FinFET nodes are presented. The relationship between angular incidence and SE vulnerability is explored as a function of technology scaling, critical charge, and particle LET.
A-4
9:55 AM
Study of Angular and Material Dependence of Single Event Effects in Enhancement- mode GaN HEMTs
P. Maloney1, A. Billa1, B. Bolton1, S. Hankinson1, S. Shorina1, H. Gingold1, J. Gray2, L. Massengill2, D. McMorrow3, D. Fleetwood2, E. Zhang1
1. University of Central Florida, USA 2. Vanderbilt University, USA 3. US Naval Research Laboratory, USA
Heavy-ion testing of enhancement-mode GaN HEMTs shows SELC and SEB thresholds depend strongly on incident angle and overlying metallization. SEM and SRIM reveal Cu interconnects amplify charge deposition, explaining reduced thresholds at near-normal incidence.
POSTER SESSION A
PA-1
SoC Processor and Memory Proton Testing Methodology and Results for the AMD Versal ACAP
G. Smith1, M. Wirthlin1, M. Brigham1
1. Brigham Young University, USA
A methodology is presented for testing multiple processors and 18 memories simultaneously on the Versal ACAP during a proton irradiation test. Post-run fault analysis is performed to identify the root cause of all processor failures.
PA-2
Impact of Structural and Numerical Constraints on Single-Event Upsets in ResNet50 Neural Networks
G. Qiao1, S. Feng1, G. Yu1, M. Tao2, Y. Chi1, J. Chen1, D. Luo1, J. Yang2, B. Liang1
1. National University of Defense Technology, China 2. Hunan University, China
Using software fault injection and pulsed laser experiments, this work analyzes ResNet50 failure behaviors under radiation, showing that combined structural and numerical constraints significantly enhance model robustness on AI chip.
PA-3
Heavy-Ion Testing of the AMD Versal Network-on-Chip
P. Drum1, A. George1, M. Cannon2, A. Kumar2, N. Myers2, A. Tabaczynski2, D. Lee2, N. Matter2, P. Thelen2
1. University of Pittsburgh, USA 2. Sandia National Laboratories, USA
Heavy-ion tests were performed on the Versal Network-on-Chip to determine cross sections and common errors for two general-use designs. Common errors included correctable parity errors and packet misrouting errors.
PA-4
Proton Radiation Testing of AI Models on the AMD Versal Deep Learning Processing Unit (DPU)
J. Brown1, H. Allan1, J. Goeders1, M. Wirthlin1
1. Brigham Young University, USA
Fault injection and proton radiation testing of YOLO and ResNet models on the AMD Versal DPU reveal six failure modes, model-dependent error rates, and the predictive capability of fault injection
PA-5
10:10 AM
Evaluation of Single-Event Effects on Swin Transformer Tracking Model for Sub-20nm FinFET-based AI Chips
Z. Li1, S. Feng1, G. Yu1, M. Tao2, Y. Chi1, J. Chen1, D. Luo1, J. Yang2, Y. Liu1, F. Luo1, J. Mo1, B. Liang1
1. National University of Defense Technology, China 2. Hunan University, China
Pulsed laser and heavy-ion experiments are conducted on a Swin-Transformer-based edge tracking system. Different structural variants exhibit distinct SEFI/SEU occurrence characteristics, revealing structure-dependent reliability behaviors under radiation- induced single-event effects. – BREAK SESSION B SINGLE-EVENT EFFECTS: Devices and Integrated Circuits SESSION INTRODUCTION Chair: Ruben García Alía (CERN)