Source: TH
Subject: Science and Technology
Context: Scientists at the University of Cambridge have developed a new hafnium-based memristor that mimics human brain synapses to process and store data simultaneously.
About Memristor:
What it is?
- A memristor (a portmanteau of memory and resistor) is a two-terminal electronic component that regulates the flow of electrical current in a circuit while also remembering the amount of charge that has previously flowed through it.
- It is considered the fourth fundamental circuit element, alongside the resistor, capacitor, and inductor.
How it Works?
- Variable Resistance: Unlike a standard resistor with fixed resistance, a memristor’s resistance changes depending on the voltage applied.
- The Memory Effect: When the power is turned off, the memristor retains its last resistance state. When power is restored, it remembers that state, acting as non-volatile memory.
- Cambridge Innovation (P-N Junction): Traditional memristors use a filament that breaks/forms unpredictably. The Cambridge team used a p-n junction interface. By pushing ions with low-voltage pulses, they smoothly raise or lower the energy barrier for electrons, making the device’s behavior more predictable and energy-efficient than filament-based versions.
Key Features:
- Brain-Inspired (Neuromorphic): Like a biological synapse, it fuses memory and computation in the same location, mimicking the brain’s architecture.
- Ultra-Low Power: Requires a million times less current to switch states compared to conventional oxide memristors, leading to a 70% reduction in energy use.
- Linearity & Plasticity: Exhibits linearity (proportional input/output) and spike-timing-dependent plasticity, meaning the connection strengthens or weakens based on the timing of signals, just like real neurons.
- Scalability: Made of Hafnium Oxide, a material already widely used in the semiconductor industry (CMOS transistors), making it easier to integrate into existing chip manufacturing lines.
- Durability: Can endure tens of thousands of switching cycles.
Applications:
- Artificial Intelligence: Running massive AI models and Neural Networks with a fraction of the current energy cost.
- Edge Computing: Enabling powerful AI processing on small devices (like smartphones or sensors) without needing constant cloud/server connectivity.
- Neuromorphic Computing: Building brain-on-a-chip systems that can perform complex pattern recognition and learning locally.
- Non-Volatile Memory: Creating faster, denser, and more energy-efficient alternatives to current Flash or DRAM memory.
