Computational Modelling — An optoelectronic synapse based on Cu-BHT MOF for multi-wavelength optical logic gates and neuromorphic vision system

Measurement evidence

Computational Modelling

An optoelectronic synapse based on Cu-BHT MOF for multi-wavelength optical logic gates and neuromorphic vision system · Li Z., Zhang L., Chen Y. et al. · Applied Materials Today · 2025 · 102926

1 measurement group · 9 results

Reported values remain attached to the sample, method, conditions, extraction quality and source location that produced them.

CNN simulation using conductance data from Cu-BHT synaptic device

Ag/Cu-BHT/Ag planar optoelectronic synaptic device · Electrode

Two convolution layers, two pooling layers and fully connected layer; 60,000 MNIST training images, 10,000 test images; conductance data from Fig. S10d integrated in training.

Atmosphere: not applicable
Geometry: model uses Ag/Cu-BHT/Ag device conductance data
Context: pristine Cu-BHT film device data
Measurement source: p005-p006 / 5-6 · Results and discussion · Fig. 4; Fig. S10d; Fig. S11

Property	Reported value	Normalised value	Uncertainty	Origin and quality	Source
LTP nonlinearity beta_p, learning	4.8825	4.8825 dimensionless	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. S12
LTP nonlinearity beta_p, relearningMarked as a best value within this paper	4.4969	4.4969 dimensionless	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. S12
CNN recognition accuracy after 100 cycles, learningMarked as a best value within this paper	96.3%	0.963 fraction	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. 4d
Initial CNN recognition accuracy, learning	24.88%	0.2488 fraction	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. 4d
CNN recognition accuracy after 100 cycles, relearningMarked as a best value within this paper	96.6%	0.966 fraction	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. 4d; Fig. S11
Maximum conductance during learning pulsesMarked as a best value within this paper	20.37 uS	20.37 uS	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. S10d
Conductance retained after pulse withdrawal	12.07% of maximum value within 100 s	0.1207 fraction	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. S10d
Pulses needed to restore conductance during relearning	28 light pulses	28 pulses	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. S10d
Iterations per training cycle	1688 iterations	1688 iterations	—	Text Exact Reported	p005 / 5 · Results and discussion · Fig. 4d