Papers, new IP, and people moves from the AMPIC Lab.
2026
P
Paper
CMOS-Compatible Thermo-Optic Switch with Undercut Resistive Heater
N. Afsary, S. Mishra and V. Saxena, CLEO 2026. A thermo-optic switch whose resistive heater sits over an undercut cavity, so heat stays near the waveguide and little power is needed to tune the phase. It is built in a CMOS-compatible flow, and measurements come from our foundry-fabricated test PIC.
A new University of Delaware licensing opportunity. Doped silicon replaces metal heaters in silicon nitride photonic circuits, removing metallic optical loss while a thermally isolating undercut keeps power low and thermal crosstalk small.
A new University of Delaware licensing opportunity. Heavily doped silicon ribs beside the waveguide act as side-contact Joule heaters for volatile phase-change materials, giving nanosecond-scale switching in a CMOS-compatible process.
Low-Power Silicon Programmable Photonic Meshes With Undercut Heaters
N. Afsary, S. Mishra, and V. Saxena. Programmable photonic meshes whose thermo-optic heaters use a substrate undercut, so large meshes can be tuned at low power.
Sub-mW low-latency silicon thermo-optic phase shifter using a foundry substrate undercut heater
N. Afsary, S. Mishra, and V. Saxena. A silicon thermo-optic phase shifter that tunes with under a milliwatt and switches quickly, using a foundry-available substrate undercut.
Accuracy-Efficiency Trade-offs of Temporal Enrichment in Spiking Neural Networks
S. Mandal, Y. Ren, and V. Saxena. A study of how enriching the temporal information in spiking neural networks trades accuracy against computational efficiency.
Programmable CMOS Izhikevich Neurons for Bio-realistic Neuromorphic Computing
J. Chen, S. Mandal, Y. Ren, and V. Saxena. Programmable CMOS neuron circuits based on the Izhikevich model that reproduce a range of biological firing behaviors.
A CCO-based Leaky Integrate and Fire Neuron in 130nm CMOS/ReRAM Technology for SNNs
A. Dorzhigulov and V. Saxena. A leaky integrate-and-fire neuron built around a current-controlled oscillator in 130 nm CMOS, for spiking networks with ReRAM synapses.
A Hybrid CMOS Photonic RF Channelizer: System Design and Analysis
A. Moussa and V. Saxena, IEEE MWSCAS 2025, East Lansing, MI. System-level design and analysis of an RF channelizer that splits a wideband signal into channels using CMOS electronics together with photonic components.
CMOS-RRAM Neuromorphic Accelerators Using Multi-bit Neurons
V. Saxena and A. Moussa, IEEE MWSCAS 2025, East Lansing, MI. Mixed-signal neuromorphic accelerators that pair CMOS neurons with RRAM synapses, using multi-bit neurons to improve accuracy and efficiency.
A 7x4 Silicon Photonic Reconfigurable Optical Analog Processor with Algorithmic Calibration
V. Saxena and M. J. Shawon, OFC 2024. A programmable mesh of silicon photonic components that works as an analog optical processor, with an algorithm that calibrates it automatically.
Optical Interconnects Using Hybrid Integration of CMOS and Silicon-Photonic ICs
V. Saxena, A. Kumar, S. Mishra, S. Palermo and K. Lakshmikumar, IEEE TCAS II: Express Briefs. A tutorial brief on how CMOS electronics and silicon photonics are integrated for high-speed optical links.
A sub-1pJ/bit laser power independent 32Gb/s silicon photonic EAM driver in 65nm CMOS
S. Mishra and V. Saxena. A 65 nm CMOS driver for a silicon photonic electro-absorption modulator, running at 32 Gb/s on under 1 pJ/bit regardless of laser power.
A Silicon Photonic Reconfigurable Optical Analog Processor (SiROAP) with a 4x4 Optical Mesh
M. J. Shawon and V. Saxena, ISSCC 2023. A reconfigurable 4x4 mesh of silicon photonic elements, controlled by CMOS electronics, for analog optical signal processing.
Spiking CMOS-NVM Mixed-Signal Neuromorphic ConvNet with Circuit- and Training-optimized Temporal Subsampling
A. Dorzhigulov and V. Saxena, Frontiers in Neuroscience, vol. 17, 2023. A spiking convolutional network built from CMOS and non-volatile memory, with temporal subsampling tuned in both circuit and training.
Jubayer graduated in summer 2023 with a dissertation on large-scale reconfigurable silicon photonic ICs for RF photonics and optical signal processing.