Optimization of Junction Resistance in Solution-Processed 2D Nanosheet Networks

Abstract

Networks of solution-processed two-dimensional (2D) nanosheets are a promising materials platform for printable and flexible electronics, combining the unique physicochemical properties of atomically thin crystals with scalable processing into large-area films. Such networks can retain electrical performance under repeated strain, making them attractive for wearable and implantable technologies. However, their electrical conductivity is typically reduced by several orders of magnitude relative to that of individual nanosheets due to the presence of inter-nanosheet junctions. Charge transport across these junctions proceeds by charge hopping, giving rise to a so-called junction resistance that forms the principal bottleneck in network conductivity. Lowering this resistance is therefore a central challenge for realizing functional nanosheet network devices. This thesis explores strategies to reduce junction resistance through three complementary approaches: (i) preparation of high aspect ratio nanosheets by electrochemical exfoliation (EE), (ii) assembly of aligned and homogeneous networks by liquid interface deposition (LID), and (iii) modelling of network charge transport in the AC regime to extract junction resistance and determine its dependence on nanosheet dimensions. Cathodic EE using alkylammonium intercalants was shown to yield large-area, thin TMD nanosheets (MoS2, WS2, and WSe2) while preserving their desirable semiconducting 2H phase. Compared to shear exfoliation, EE produced nanosheets with an order of magnitude higher aspect ratio (~100 vs. ~15), directly addressing the requirement for conformal sheets to form conductive junctions. Networks of these nanosheets, assembled by LID, were incorporated into TFTs electrolytically gated with 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EMIM-TFSI) as the ionic liquid dielectric. The resulting devices exhibited average mobilities of 10.7 ± 0.9 cm^2 V^-1 s^-1 (MoS2), 9.1 ± 2.3 cm^2 V^-1 s^-1 (WS2), and 2.0 ± 0.2 cm^2 V^-1 s^-1 (WSe2) with on/off ratios of (2.6 ± 0.4) x 10^3, (3.4 ± 0.6) x 10^3, and (4.2 ± 1.8) x 10^4, respectively. These figures of merit are competitive with alternative semiconducting material systems, such as carbon nanotubes and metal oxides. Importantly, device performance was retained on a flexible polyethylene terephthalate (PET) substrate, with a WSe2 TFT on PET maintaining high mobility (1.9 cm^2 V^-1 s^-1) and on/off ratio (~ 3 x 10^3) over 1000 bending cycles. To refine the LID technique, a procedure was established to optimise single nanosheet layer deposition by varying nanosheet ink concentration and interfacial areal mass load. Using electrochemically exfoliated (EE) graphene as a reference system, several figures of merit for film quality were defined. Optimal deposition was identified at low ink concentration (0.02 g L^-1) and high areal mass load (21 mg m^-2). These parameters were then applied to assemble multilayer films by sequential deposition of monolayer-enriched EE graphene onto glass. Conductivity-thickness data for these films was fitted using a novel percolative scaling model, yielding a high bulk conductivity ( σ_B = 1.4x10^5 S m^-1) and low bulk transition thickness (t_x = 12 nm). As transparent conductors, these networks compared favourably with graphene nanosheet films in the literature, with a representative sample combining 85% optical transmission and sheet resistance of 4.1 kΩ □^-1. Finally, analogous EE graphene films were fabricated on Kapton to realise a flexible transparent conductor. Although the device maintained conductivity (~3.5 x 10^3 S m^-1) over hundreds of bending cycles, the sheet resistance (15.7 kΩ □^-1) was higher and the transmission (~50%) much lower than the films on glass, likely due to the reduced post-processing temperature required by the polymer substrate. Finally, a model linking junction resistance to nanosheet dimensions was hypothesised and tested. To enable this, a liquid cascade centrifugation protocol was developed that could selectively isolate fractions of EE MoS2 with distinct thickness distributions, including populations of substantially thick nanosheets. These size-selected dispersions were then deposited into films by both LID and spray coating, and they were electrically characterized by impedance spectroscopy to extract junction resistance. While a clear correlation between junction resistance and nanosheet thickness was not established, the results revealed that networks assembled by LID consistently exhibited substantially lower junction resistance than those prepared by spray coating, reinforcing its suitability for producing low junction resistance networks and hence high-performance nanosheet network devices.