Nano‐Memristors with 4 mV Switching Voltage Based on Surface‐Modified Copper Nanoparticles

Abstract

Memristors, i.e., short for resistor with memory, are electronic devices that can switch their electrical resistance between different levels depending on the external voltages applied. [1] The resistive switching (RS) phenomenon was first observed in the late 1960s in a two-terminal cell based on a phase-change material sandwiched by two electrodes, [2] and nowadays it has been observed in many other materials (e.g., metal-oxide, magnetic, and ferroelectric thin films). [3] The performance of memristors is characterized by the voltage, energy, and time that they need to switch, as well as by the resistance of each state, the stability of the state resistance over time (i.e., retention time), and the maximum number of cycles that it can switch (i.e., endurance). [4,5] Depending on the materials employed to fabricate the memristor its main figures-of-merit may vary, allowing its use in multiple applications, such as non-volatile memory, advanced computation, radiofrequency switches, and entropy source for encryption systems. [3,6] The use of memristors in the field of bioelectronics is highly desired because they could enormously simplify the structure of integrated circuits required for spiking signal detection [7] and event-driven operation [8] -i.e., two key functions in multiple products, such as e-skin and monitoring systems-plus they may enable more sophisticated functionalities, such as in-memory computing, [9] and neuromorphic computation by reverse engineering the brain. [10] As an example, reference [11] employed bio-compatible organic memristors to monitor the signal activity of neocortical pyramidal neurons from rats, and reference [12] reported a three-neuron brain-silicon network to emulate transmission and plasticity properties of real synapses. However, in such studies, a patch-clamp integrated circuit was required to amplify the bioelectronic signal from the living cells (i.e., 50-100 mV) to the operational voltages of the memristors (i.e., >2 V), which increased the complexity of the overall system. Therefore, fabricating memristors capable to operate at low bio-compatible voltages is highly attractive to simplify the circuitry and reduce the energy consumption of bioelectronic systems, which may enable their use in battery-less and selfpowered implants.Many studies have attempted the fabrication of memristors with ultralow switching voltages (see Table S1, SupportingThe development of memristors operating at low switching voltages <50 mV can be very useful to avoid signal amplification in many types of circuits, such as those used in bioelectronic applications to interact with neurons and nerves. Here, it is reported that 400 nm-thick films made of dalkyl-dithiophosphoric (DDP) modified copper nanoparticles (CuNPs) exhibit volatile threshold-type resistive switching (RS) at ultralow switching voltage of ≈4 mV. The RS is observed in small nanocells with a lateral size of <50 nm -2 , during hundreds of cycles, and with an ultralow variability. Atomistic calculations reveal that the switching mech...

Recent progress in bio-voltage memristors working with ultralow voltage of biological amplitude

“…2b). A lower V set in the range of several mV was observed in Cu-based BMRs using Pt/DDP-CuNPs/Au 54 and Au/Cu/ ZnO-ZnS/ZnS/Pt/Ti 47 structures. The former showed volatile properties with the 4 mV V set , while the latter showed the feasibility of achieving nonvolatility with a larger V set .…”

“…51 Other thin semiconducting materials such as InSe 52 and GeSe 53 were also used to serve as the dielectrics in Ag-electrode and Cu-electrode BMRs, respectively. Some BMRs do not directly use Ag or Cu electrodes but preintroduce these elements into the dielectric layer, 54 which in effect can be considered pertaining to the same category. The elements were introduced either in a reduced metal form in composites such as Ta 2 O 5 -Cu, 55 AgNWs-TiO 2 -PVA, 56 and DDP-CuNPs, 54 or an oxidized ion form in compounds such as Ag 2 S, [57][58][59] AgI, 60 and Ag x AsS 2 .…”

“…Some BMRs do not directly use Ag or Cu electrodes but preintroduce these elements into the dielectric layer, 54 which in effect can be considered pertaining to the same category. The elements were introduced either in a reduced metal form in composites such as Ta 2 O 5 -Cu, 55 AgNWs-TiO 2 -PVA, 56 and DDP-CuNPs, 54 or an oxidized ion form in compounds such as Ag 2 S, [57][58][59] AgI, 60 and Ag x AsS 2 . 61 Pre-introduction of the elements into the dielectric layer can still facilitate voltage reduction in some BMRs already having Ag or Cu electrodes.…”

Probing switching mechanism of memristor for neuromorphic computing

“…Figure 6(a) reported a two-terminal memristor used for mimicking biological synapses [140]. Compared with traditional electrical test techniques, i.e., probe station + semiconductor parameter analyzer, the main advantage of CAFM is that by using the tiny conductive AFM probe (radius<20 nm), we are able to obtain typical MIM memristive cell with nano sizes of ∼50 nm 2 , and achieve in situ detection of the synaptic and neuronal behaviors by CAFM [141,142]. It is noted that in the above MIM cell, the conductive tip is served as a top electrode, and the contact area (between the tip and insulating layer) equals the effective size of a single device.…”

Multiple Switching Modes of NiOx Memristors for Memory-Driven Multifunctional Device Applications

“…This trend is further accelerating these days with explosively increases in the accumulated data sizes . As a result, a memory-driven parallel computing framework that enables efficient data processing in terms of energy and speed has emerged as a next-generation computing technology. , As an essential building block for this parallel computing system, memristors have gained attention owing to their excellent analog/digital switching performance and competitive nanoscale applicability. − The dynamic reconfiguration of internal states in the memristor, modulated by the history of applied external stimuli (electrically or optically), can mimic both synaptic and neuronal functionalities in the cortex neural network of the human brain, enabling parallel computation in an energy-efficient manner . In addition, it can switch to a binary state, further extending its device applicability.…”