Electrical conduction and resistive switching in polymer and biodegradable nanocomposites

Abstract

Modern memory devices such as static random-access memory (SRAM), dynamic random-access (DRAM), and Flash memories demonstrated inevitable limitations, i.e., large cell size (50 − 120 F2) of SRAM, accompanied by current leakage; high operating voltages of 3 V and up to 6 V for DRAM and NOR Flash, respectively; DRAM capacity should sustain enough charges (there is a limit to how small the DRAM capacitor can be) and Flash need a novel array structure. Additionally, these current memory devices contribute significantly to the world’s earth pollution. These memories still use heavy metals such as Pb, which are harmful to humans. There is a demand for a next-generation random-access memory (RAMs) having fast read and write operations as the SRAM, high density and cost-benefit as the DRAM, and nonvolatility as the Flash. Furthermore, new memory device must be compatible with on-chip computing. Resistive switching memories (ReRAMs) are an emerging memory technology with prospects of combined benefit found in all current memories. Furthermore, ReRAMs can be fabricated using any material, including organic polymers and biological materials. This gives ReRAM environmentally friendly properties and compatibility with futuristic electronics, where special mechanical properties such as transparency and flexibility are important. In this study, we conducted intense research on electrical conduction and resistive switching in biodegradable polymers such as chitosan and polyvinylpyrrolidone, and in the process, we discovered, for the first time, resistive switching in raw cow milk. First resistive switching and conduction mechanisms in spin-coated devices consisting ofcadmium telluride/cadmium selenide (CdTe/CdSe) coreshell quantum dots embedded in a chitosan active layer sandwiched between (1) aluminium (Al) and silver (Ag) and (2) indium-doped tin oxide (ITO) and Ag electrodes were studied. Here, both devices exhibited bipolar memory behavior at low (+0.70 V ) voltage, enabling both devices to be operated at low powers. The devices displayed different switching mechanisms, i.e., conductive bridge mechanism in the Al-based device and space-chargelimited driven conduction filament attributed in the ITO device. Additionally, the Al-based device showed long retention (> 103 s) and a reasonable large (> 103) ON/OFF ratio. We also observed a sweeping cycle-induced reversal of the voltage polarity of the VSET and VRESET in the Al-based device, which is a new observation. Using the same composite but changing the film deposition method, i.e., now using the drop-casting method. All devices consisting of 0.96 wt%, 0.48 wt%, 0.32 wt% and 0.24 wt% CdTe/CdSe QDs to chitosan showed ‘O-type’ memory behavior with OFF-state current conduction mechanism attributed to the hopping mechanism. However, the ON-state current in each device followed a unique mechanism, such that Ohmic behavior was observed for the device with 0.96 wt%, while linear then hopping, space-charge limited, and lastly, hopping conduction mechanisms were attributed to devices with 0.48 wt%, 0.32 wt% and 0.24 wt%, respectively. Proving that memory behavior and conduction in these devices can be exploited by controlling the amount of CdTe/CdSe. Next, we investigated the effect of molybdenum(IV) sulfide (MoS2) on both conduction and memory behavior in polyvinylpyrrolidone (PVP) by fabricating various ReRAM devices using (1) plain MoS2 (device A), (2) plain PVP (device B), (3) PVP and MoS2 bilayer (device C), and (4) PVP +MoS2 nanocomposites with 10 wt% (device D), 20 wt% (device E), 30 wt% (device F) and 40 wt% (device G) MoS2 fabricated with Al and Ag as bottom and top electrodes, respectively. We did not observe switching in devices A and B. Device C showed a combination of bipolar and threshold switching at 0.40 V . Device G portrayed bipolar switching at 0.56 V . In Device C, space charge-limited conduction while Ohmic behavior followed by trapping of charge before switching was noticed in device G. Both devices C and G showed reasonably (≥ 102) ON/OFF ratio. In the nanocomposite devices, we observed that an increase in MoS2 content increased electrical conductivity in the Ohmic region, leading to threshold switching at 30 wt% (device F) and ultimately bipolar switching at 40 wt% (device G). These studies showed that both switching and conduction mechanisms are sensitive to the type and composition of the active layer in the devices studied. Next, we investigated resistive switching in chitosan/PVP composite as the active layers sandwiched between Al and Ag electrodes. ReRAMs with active layers consisting of 1 : 3, 1 : 1, and 3 : 1 chitosan to PVP ratios were studied. Asymmetric threshold switching with only the negative voltage bias was obtained for the device with a chitosan to PVP ratio of 1 : 3. The 1 : 1 chitosan to PVP ratio device showed optimal memory behavior with bipolar switching with low (0.28 V ) switching voltage in the first cycle, followed by asymmetric threshold switching during the second cycle and back to bipolar switching. We did not observe memory behavior in the 3 : 1 chitosan to PVP-based device. Electrochemical conduction metalization was attributed to the switching mechanism in the device with a 1 : 1 ratio of chitosan to PVP. Our results reveal the applicability of chitosan and PVP blend in memory device fabrication and that both the memory and switching can be exploited by varying the ratio of chitosan to PVP in the composite. Lastly, we fabricated the first resistive switching memory devices that use raw organic cow milk as active layers. Our devices comprised fat-free, medium cream, and full cream raw cow milk active layers sandwiched between ITO and Ag. All devices showed low switching voltages, with the medium fat milk-based device showing the lowest VSET = +0.45 V and VRESET = −0.25 V . Additionally, the medium fat-based device showed an ‘S-type’ memory mode attributed to the space-charge-limited conduction mechanism. Alternatively, fat-free and fill-cream-based devices both showed ‘O-type’ memory behavior attributed to hopping conduction. EDS analysis of all active layers revealed a relatively higher weight percentage of metallic ions in the medium fat milk film than in fat-free and full-cream milk films, which explains the different behaviors. These devices combine biodegradability and low power characteristics that are important for green computing.