Bidirectional Polarization-Controlled Nonvolatile Ferroelectric SnS/α-In₂Se₃ van der Waals Heterojunction for...

Area of Science:

• Materials Science
• Condensed Matter Physics
• Electrical Engineering
• Nanotechnology

Background:

• The traditional von Neumann architecture faces significant bottlenecks due to the physical separation of processing and memory units, limiting computational efficiency.
• Logic-in-memory (LIM) devices offer a promising solution by integrating data processing and storage, reducing data transfer delays and energy consumption.
• Ferroelectric field-effect transistors (Fe-FETs) are attractive candidates for LIM due to their nonvolatile memory characteristics and potential for reconfigurable logic.

Purpose of the Study:

• To develop a novel van der Waals (vdW) heterojunction-based ferroelectric field-effect transistor (Fe-FET) for integrated logic-in-memory applications.
• To demonstrate the feasibility of using distinct ferroelectric polarization states for nonvolatile memory and reconfigurable logic operations.
• To address the limitations of conventional computing architectures by enabling efficient data processing within memory.

Main Methods:

• Fabrication of a SnS/α-In2Se3 vdW heterojunction-based Fe-FET.
• Characterization of the Fe-FET's nonvolatile memory functionality, including endurance and retention properties.
• Demonstration of logic-in-memory capabilities by utilizing multiple resistance states controlled by ferroelectric polarization manipulation via voltage pulses.

Main Results:

• The fabricated Fe-FET exhibits distinct in-plane and out-of-plane ferroelectric polarization, enabling four stable resistance states.
• These resistance states demonstrate excellent cycling endurance and extended retention, confirming reliable nonvolatile memory operation.
• Fundamental logic operations (NOR, NAND, AND, OR) were successfully implemented by voltage-controlled manipulation of the ferroelectric polarization states.

Conclusions:

• The SnS/α-In2Se3 vdW Fe-FET successfully integrates nonvolatile memory and reconfigurable logic functions, achieving genuine logic-in-memory operation.
• This approach effectively eliminates data transfer bottlenecks and reduces energy consumption, paving the way for next-generation low-power electronic systems.
• The developed device architecture represents a significant advancement toward realizing highly efficient and dense computing systems.