top of page



Goal: Model and simulate the neuromorphic behavior of a RRAM-based circuit realized for neuromorphic applications.

• A LTspice based memristor model is developed in collaboration with UAB (Universitat Autonoma de Barcelona).
• To test if the memristor model can be used in a neuromorphic circuit, certain basic measurement was done such as LTP and STDP.
• Using PySpice (combination of python and NGspice) is being used to realize the neuromorphic circuit.


Neuromorphic application of RRAM

RRAM is an emerging nonvolatile memory that depicts synaptic characteristics hence it can be used to replace the weights in a neural network.

• Ruthenium-based RRAM is being characterized for the application.
• A circuit is developed that can do basic pattern recognition.
• To test the extent of how well the device will function on a real-time application some simulations are also being done on a platform called NeuroSim.


RRAM characterization

• RRAM and CBRAM are one of the most promising non-volatile memories.
• The devices were fabricated in the cleanroom of CGU.
• The devices were characterized for their non-volatile nature using the B1500a semiconductor analyzer.
• Later for testing the neuromorphic characteristics (LTP, LTD, STDP, SRDP) of the devices, I programed B1500a.
• Various time-dependent and rate-dependent plasticity were discovered in the devices.
• Applications include in-memory computing, neuromorphic computing, TRNG, etc.


Flexible Li-ion cell

Objective: To develop a pouch cell that can be used in wearable electronics.

• We synthesized Graphene using Tour’s method.
• We indigenously developed Glove box and spot-welding machine needed for the project.
• During this process I have learned using multiple equipments such as tubular furnace, muffel furnace, centrifuge, ultrasonic cleaner etc.

Lithium Ion Cell 3.png

Multi-junction Solar cell

• In generic solar cells, a single semiconductor is used for extracting energy from sunlight.
• Stacking multiple semiconductors increase the overall efficiency of the solar cell.
• We used AMPS-1D software to simulate multijunction solar cell.

bottom of page