Boosting Neuromorphic Synapses through Alloying-Induced Trap Engineering in Two-Dimensional Mo1-xWxS2

Area of Science:

• Materials Science
• Condensed Matter Physics
• Nanotechnology

Background:

• Two-dimensional transition metal dichalcogenides (TMDs) offer tunable electronic properties for neuromorphic applications.
• Defect-mediated charge dynamics in TMDs are crucial for memory functions in electronic devices.

Purpose of the Study:

• To engineer Mo1-xWxS2 alloys via chemical vapor deposition.
• To tailor lattice strain, defect density, and optoelectronic properties for enhanced neuromorphic functionality.
• To investigate the relationship between alloy disorder, defects, and synaptic device performance.

Main Methods:

• Alloying engineering of Mo1-xWxS2 using chemical vapor deposition.
• Structural and spectroscopic analyses (e.g., XRD, Raman spectroscopy) to characterize material properties.
• Fabrication and characterization of synaptic field-effect transistors (FETs) to evaluate device performance.

Main Results:

• Homogeneous distribution of Mo and W atoms in Mo1-xWxS2 alloys, inducing strain fields that enhance charge trapping.
• Composition-dependent bandgap tuning and exciton lifetime modulation, with near-equiatomic MoWS2 showing strong non-radiative recombination.
• Synaptic FETs demonstrated large hysteresis (up to 22 V), high interface trap densities, and robust low-frequency noise.
• Near-equiatomic MoWS2 devices emulated biological synaptic plasticity, including paired-pulse facilitation, learning-forgetting cycles, and potentiation/depression.
• Achieved 87.95% recognition accuracy in CNN simulations, highlighting practical neuromorphic potential.

Conclusions:

• Alloy engineering of Mo1-xWxS2 provides a versatile platform for defect-strain coupling in 2D materials.
• The developed MoWS2-based synaptic devices exhibit promising characteristics for next-generation neuromorphic computing.
• These findings pave the way for advanced 2D material-based artificial intelligence hardware.