Tunable homojunction field effect device-based artificial synapse circuit and implementation method thereof
A tunable homojunction field effect device-based artificial synapse circuit includes a first tunable homojunction field effect device M 1 , a second tunable homojunction field effect device M 2 , a third tunable homojunction field effect device M 3 , and a capacitor C; the tunable homojunction field effect device can exhibit the electrical properties of NN junction, PP junction, PN junction, and NP junction under the control of gate voltage; in the circuit, whether the device M 2 and the device M 3 are turned on rely on the combined action of presynaptic pulse and postsynaptic pulse; compared with the circuit structure of a traditional CMOS circuit scheme which exhibits neural synaptic functions of spike-time-dependent plasticity and continuously adjustable pulse-to-synaptic weight, the circuit in the present solution requires a greatly reduced number of devices and shows the feature of reconfigurable function, exhibiting a great advantage in constructing low-power, high-density integrated bionic chips for future neuromorphic applications.
1 . A tunable homojunction field effect device-based artificial synapse circuit, wherein the circuit comprises a first tunable homojunction field effect device M 1 , a second tunable homojunction field effect device M 2 , a third tunable homojunction field effect device M 3 , and a capacitor C;
the source S 2 of M 2 is connected to the gate electrode G 3 A of M 3 close to the source, which serves as a first input terminal and is used to input a presynaptic pulse, the source S 3 of M 3 is connected to the gate electrode G 2 A of M 2 close to the source, which serves as a second input terminal and is used to input a postsynaptic pulse; the drain D 2 of M 2 is connected to the gate electrode G 2 B close to the drain D 2 and the gate electrode G 1 A of M 1 close to the source S 1 , the drain D 3 of M 3 is connected to the gate electrode G 3 B close to the drain D 3 and the gate electrode G 1 B of M 1 close to the drain D 1 ; the capacitor C is connected between the two gate electrodes G 1 A and G 1 B of M 1 , the source S 1 and the drain D 1 of M 1 are respectively used to apply the source bias voltage V 1 and the drain bias voltage V 2 ;
the structures of M 1 , M 2 , and M 3 are the same, each comprising a substrate insulating material, a channel material layer, an insulating layer, and a metal electrode layer, the metal electrode layer comprises a drain electrode layer, a source electrode layer, a gate electrode layer A, and a gate electrode layer B, the gate electrode layer A and the gate electrode layer B are fabricated side by side on the substrate insulating material, and a gap is left between the gate electrode layer A and the gate electrode layer B to ensure electrical insulation therebetween, the insulating layer completely covers the gate electrode layer A and the gate electrode layer B, the drain electrode layer is placed on the left edge of the channel material layer above the gate electrode layer A, and the source electrode layer is placed on the right edge of the channel material layer above the gate electrode layer B, that is, the gate electrode layer A corresponds to the gate electrode G 1 B in M 1 , the gate electrode G 2 B in M 2 , and the gate electrode G 3 B in M 3 , and the gate electrode layer B corresponds to the gate electrode G 1 A in M 1 , the gate electrode G 2 A in M 2 , and the gate electrode G 3 A in M 3 .
2 . The tunable homojunction field effect device-based artificial synapse circuit according to claim 1 , wherein the channel material layer uses a low-dimensional semiconductor material with ambipolar field effect characteristics.
3 . The tunable homojunction field effect device-based artificial synapse circuit according to claim 2 , wherein the low-dimensional semiconductor material is silicon nanowires, carbon nanotubes, two-dimensional layered materials, or organic semiconductor thin film materials.
4 . The tunable homojunction field effect device-based artificial synapse circuit according to claim 2 , wherein the material with ambipolar field effect characteristics is an intrinsic semiconductor with a band gap ranging from 0.5 eV to 1.5 eV and a material thickness of less than 30 nm.
5 . The tunable homojunction field effect device-based artificial synapse circuit according to claim 1 , wherein the homojunction state comprises an NN junction, a PP junction, a PN junction, and an NP junction.
6 . A method for implementing a circuit function of the tunable homojunction field effect device-based artificial synapse circuit according to claim 1 , comprising:
applying bias voltages V 1 and V 2 to the source S 1 and the drain D 1 of M 1 , V 1 <V 2 , and measuring the current value I ds-M1-before passing through M 1 at this moment;
applying a presynaptic pulse and a postsynaptic pulse, recording the time difference Δt between a presynaptic pulse voltage and a postsynaptic pulse voltage;
measuring the current value I ds-M1-after passing through M 1 after applying the presynaptic pulse and the postsynaptic pulse, and calculating the synaptic weight change ΔW;
changing the time difference between the presynaptic pulse and the postsynaptic pulse input with the bias voltages V 1 and V 2 remaining unchanged to determine the change rule between ΔW and Δt, and simulating the biological anti-Hebbian learning rule.
7 . The method for implementing the circuit function of the tunable homojunction field effect device-based artificial synapse circuit according to claim 6 , wherein the change rule between ΔW and Δt comprises:
when Δt>0, generating a net negative-going effective pulse, reducing the electrical conductivity of M 1 , namely, reducing the strength of a synaptic connection, otherwise,
when Δt<0, generating a net positive-going effective pulse, enhancing the electrical conductivity of M 1 , namely, enhancing the strength of a synaptic connection, and when |Δt| increases, the absolute value of the net effective pulse will decrease, that is, the absolute value of the synaptic weight change |ΔW| decreases.
8 . The method for implementing the circuit function of the tunable homojunction field effect device-based artificial synapse circuit according to claim 6 , wherein the method further comprises:
exchanging the values of the bias voltages V 1 and V 2 and keeping unchanged and changing the presynaptic pulse and the postsynaptic pulse to determine the change rule between ΔW and Δt and then simulate the biological Hebbian learning rule.
9 . The method for implementing the circuit function of the tunable homojunction field effect device-based artificial synapse circuit according to claim 8 , wherein when Δt>0, a net positive-going effective pulse is generated, enhancing the electrical conductivity of M 1 , namely, enhancing the strength of a synaptic connection, otherwise,
when Δt<0, a net negative-going effective pulse is generated, reducing the electrical conductivity of M 1 , namely, reducing the strength of a synaptic connection.