Modulation of Conductivity and Contact Resistance of RuO2 Nanosheets via Metal Nano-Particles Surface Decoration

We studied the variation in electrical conductivity of exfoliated RuO2 nanosheets and the modulation in the contact resistance of individual nanosheet devices using charge transfer doping effects based on surface metal nanoparticle decorations. The electrical conductivity in the monolayer and bilayer RuO2 nanosheets gradually increased due to the surface decoration of Cu, and subsequently Ag, nanoparticles. We obtained contact resistances between the nanosheet and electrodes using the four-point and two-point probe techniques. Moreover, the contact resistances decreased during the surface decoration processes. We established that the surface decoration of metal nanoparticles is a suitable method for external contact engineering and the modulation of the internal properties of nanomaterials.

We studied the electrical transport properties of individual RuO 2 nanosheet devices that underwent surface treatment with metal nanoparticles to enhance their electrical conductivity. We measured the variations in electrical conductivity as a result of the surface charge transfer doping of Cu and Ag nanoparticles in monolayer and bilayer RuO 2 nanosheets as a function of temperature. Moreover, we performed a quantitative analysis of the contact resistances between the nanosheet and electrodes, which had been overlooked in past studies, using two different measurement configurations, with fourterminal devices. We demonstrated a methodology to modulate the contact properties of devices and materials using RuO 2 nanosheets as a conductive material.

Materilas
We mixed potassium carbonate (K 2 CO 3 from Wako Pure Chemical Corporation, Osaka, Japan) and ruthenium dioxide (RuO 2 from Wako Pure Chemical Corporation) powders in a 5:8 ratio and pelletized the mixture. The pelletized mixture was heated to 850 • C for one day in a N 2 atmosphere, followed by stirring in water at room temperature for one day. The potassium ruthenates (K x RuO 2 ) were filtered and stirred into 1 M hydrochloric acid (HCl from Wako Pure Chemical Corporation) aqueous solution for 3 days. H + ions substituted K + during this process, resulting in hydrogen ruthenates (H x RuO 2 ). We mixed H x RuO 2 (4 g) with tetrabutylammonium hydroxide (TBAOH, Sigma-Aldrich, St Louis, USA) in water (1 L) and stirred the aqueous solution at room temperature for 14 days. The RuO 2 nanosheet was exfoliated and stabilized using tetrabutylammonium ions (TBA + ). Figure 1a illustrates a low-magnitude transmission electron microscopy (TEM, FEI Titan 80-300) image of a 2D nanosheet fabricated using the proposed method. We performed scanning transmission electron microscopy (STEM, FEI Titan 80-300) to study the atomic arrangement of the RuO 2 nanosheet (Figure 1b). The interplanar d-spacing in the electron diffraction (ED) pattern and inter-atomic distance in STEM were equivalent to the d-spacing between the monolayer RuO 2 nanosheets obtained from our calculations [28]. The synthesis and exfoliation of RuO 2 nanosheets were described in detail elsewhere [28,29,31].

Metal Nanoparticle Surface Decorations
We employed surface charge transfer, using the surface decoration of metal nanoparticles, to modulate the electrical transport properties of RuO 2 nanosheets [29,30]. Copper (Cu) nanoparticles were decorated on the nanosheets using 0.05 M Cu acetate (Cu(OOCCH 3 ) 2 > 99.999%, Alfa Aesar, Haverhill, MA, USA) at room temperature for one day and washed with deionized water to remove the residual particles. The nanoparticle decorated RuO 2 nanosheet was immersed in 0.05 M sodium borohydride (NaBH 2 > 99.99, Aldrich, St Louis, MO, USA) for 2 min to reduce the copper oxide (CuO x ) formed as a by-product of the surface decoration process using Cu acetate. Figure 1c,d illustrate the surface of a RuO 2 nanosheet before and after the Cu nanoparticle decoration process. We carried out the subsequent Ag nanoparticle decoration following the same processes, using 0.05 M Ag acetate (CH 3 COOAg > 99%, Aldrich, St Louis, MO, USA) for additional carrier density modulations of the RuO 2 nanosheets.

Device Fabrication
We diluted the chemically exfoliated RuO 2 solution in deionized (DI) water in a 1:20 ratio. Before dispersing the diluted RuO 2 nanosheets, we used the O 2 plasma method (COVANCE, Femto Science Inc.) for hydrophilic surface treatment on thermally oxidized silicon (SiO 2 /Si) substrates. Moreover, the alignment marks were patterned using the photolithography process (MDA-400S, Midas System) during our experiment. We determined the thickness of the nanosheets using atomic force microscopy (AFM, XE-150, Park Systems) on the SiO 2 /Si substrates to test the monolayer and bilayer RuO 2 nanosheets  [30]. The nanosheets dispersed on the substrate were patterned for transport measurements using electron-beam lithography (VEGA 3, Tescan and NPGS, JC Nabity Lithography Systems) and the lift-off process (Figure 2c,d). The open patterns were exposed to inductively coupled Ag plasma for 3 min, followed by deposition of Ti (10 nm)/Au (150 nm) using an ultra-high vacuum etching and sputtering system (custom-made) to improve the electrical contact between the nanosheet and electrodes [32]. Figure 2e illustrates a scanning electron microscopy (SEM, JEOl-7800F, JEOL Ltd.) image of a four-terminal device with individual monolayer RuO 2 nanosheets. We measured the resistances of the individual monolayer RuO 2 nanosheets using four-point and twopoint probe measurement techniques at constant current (2182 Nanovoltmeter and 236 Source Measure Unit, Keithley) to obtain their electrical transport properties. In the fourpoint probe measurement, the voltage drop due to contact resistance was excluded by physically separating the voltage measurement circuit (inner two electrodes) from the applied current circuit (outer two electrodes). On the other hand, the two-probe resistance value includes the contact and measurement circuit resistance because the voltage drop was measured across the overall circuit [32,33]. Figure 2f illustrates the I-V curves and resistances of a monolayer RuO 2 nanosheet obtained using two different measurement configurations. We calculated the electrical conductivity of the nanosheets using the following formula: σ = L/(R·w·t), where σ, L, R, w, and t are the electrical conductivity, length, resistance, width, and thickness, respectively. Moreover, we used a closed-cycle cryostat (X-1AL, ARS) under high-vacuum conditions to measure all transport properties, including the temperature dependence [33].

Metal Nanoparticle Surface Decorations
We employed surface charge transfer, using the surface decoration of metal nanoparticles, to modulate the electrical transport properties of RuO2 nanosheets [29,30]. Copper (Cu) nanoparticles were decorated on the nanosheets using 0.05 M Cu acetate (Cu(OOCCH3)2 > 99.999%, Alfa Aesar, Massachusetts, USA) at room temperature for one day and washed with deionized water to remove the residual particles. The nanoparticle decorated RuO2 nanosheet was immersed in 0.05 M sodium borohydride (NaBH2 > 99.99, Aldrich, St Louis, USA) for 2 min to reduce the copper oxide (CuOx) formed as a by-prod- resistances of a monolayer RuO2 nanosheet obtained using two different measuremen configurations. We calculated the electrical conductivity of the nanosheets using the fol lowing formula: σ = L/(R·w·t), where σ, L, R, w, and t are the electrical conductivity, length resistance, width, and thickness, respectively. Moreover, we used a closed-cycle cryosta (X-1AL, ARS) under high-vacuum conditions to measure all transport properties, includ ing the temperature dependence [33].

Increase in Electrical Conductivity
Figure 3a demonstrates the variations in the electrical conductivity of the pristine RuO 2 nanosheets due to the surface decoration treatment using Cu-nanoparticles followed by Ag-nanoparticles. We tested five-different monolayer and two bilayer nanosheets using the four-point probe measurement technique to determine the relationship between the surface doping process and the nanosheet thickness [28]. Although we observed differences in the conductivity of the nanosheets, the distributions of the conductivity values were consistent with the distributions observed in past studies using single-crystalline RuO 2 nanosheets [29]. Furthermore, the increase in the magnitude of conductivity was approximately constant with the nanosheets with the same number of layers, except for one nanosheet. The deviations in enhanced conductivity were equal to 0.08 and 0.03 (10 3 Ωm −1 ) for the 4-monolayer and 2-bilayer nanosheets, respectively (Figure 3b). The impact of Ag doping on the electrical conductivity of the nanosheets was approximately 10-times stronger than the impact of Cu doping in the both monolayer and bilayer nanosheets, but was weaker than the impact of Ag decoration without Cu decoration observed in past studies [29]. This was because the majority of the surface area of the nanosheets was covered with pre-decorated Cu nanoparticles, leaving insufficient surface area for decoration with Ag nanoparticles, which have a higher doping efficiency than Cu nanoparticles. In both metal nanoparticle doping processes, the conductivities of the monolayers were 10-times higher compared to the conductivities of the bilayers. This result agreed with the results from past studies [29].  Figure 3c demonstrates the change in electrical conductivity before and after the Cu decoration process as a function of the temperature obtained from the monolayer nanosheet, with the highest room-temperature conductivity when using the four probes. The increase in conductivity as a result of Cu doping was observed over the entire temperature sweep range, and the reproducibility of the temperature dependence was confirmed by the ramp-down and ramp-up tests. These facts demonstrate the high stability of the metal nanoparticle surface decoration doping. The low thermally excited noise level appearing at room temperature was eliminated at low temperatures in the most conductive nanosheet. Figure 3d demonstrates the change in temperature-dependent conductivity in the other nanosheet tested, with additional Ag doping. To directly observe the changes originating from the different doping processes, we calculated the normalized conductivity at room temperature, using the following equation: σnormalized = (σT − σT = 300 K)/σT = 300 K, where T is the temperature. The RuO2 nanosheets without metallic nanoparticle surface decoration exhibited a semiconducting behavior; that is, the conductivity decreased corresponding to a decrease in temperature, because of the low carrier density due to the small band overlap, despite the semi-metallic band structure of RuO2 [29,30]. The charge carriers transferred by the Cu nanoparticles increased the electrical conductiv-  Figure 3c demonstrates the change in electrical conductivity before and after the Cu decoration process as a function of the temperature obtained from the monolayer nanosheet, with the highest room-temperature conductivity when using the four probes. The increase in conductivity as a result of Cu doping was observed over the entire temperature sweep range, and the reproducibility of the temperature dependence was confirmed by the ramp-down and ramp-up tests. These facts demonstrate the high stability of the metal nanoparticle surface decoration doping. The low thermally excited noise level appearing at room temperature was eliminated at low temperatures in the most conductive nanosheet. Figure 3d demonstrates the change in temperature-dependent conductivity in the other nanosheet tested, with additional Ag doping. To directly observe the changes originating from the different doping processes, we calculated the normalized conductivity at room temperature, using the following equation: σ normalized = (σ T − σ T = 300 K )/σ T = 300 K , where T is the temperature. The RuO 2 nanosheets without metallic nanoparticle surface decoration exhibited a semiconducting behavior; that is, the conductivity decreased corresponding to a decrease in temperature, because of the low carrier density due to the small band overlap, despite the semi-metallic band structure of RuO 2 [29,30]. The charge carriers transferred by the Cu nanoparticles increased the electrical conductivity of RuO 2 nanosheets and the temperature dependence of conductivity. However, the Cu surface treatment could not overcome the temperature-dependent decrease in intrinsic carrier density. A charge exceeding the intrinsic carrier density was transferred during the Ag surface treatment at temperatures above 250 K. As a result, at room temperature, a metallic behavior with an increase in electrical conductivity with decreasing temperature was observed [32]. As the temperature decreased below 250 K, however, it changed to a semiconducting temperature behavior, due to the decrease in intrinsic carrier density. We established that the charge density transferred by the Cu-and subsequence Ag-nanoparticle decoration process was comparable to the internal carrier density. In this nanosheet, the noise level increased with decreasing temperature, indicating unstable electric contact between the nanosheet and electrodes. Figure 4a demonstrates the raw data measured by the four-and two-point probe techniques at a 10 nA current for the pristine monolayer RuO 2 nanosheet with the highest conductivity. The high reproducibility and low noise level, despite the infinitesimal measured, current indicated the stability of the electrical contact between the nanosheet and metal electrodes formed by the plasma etching process. We determined the contact resistance values by calculating the difference between the resistances of the two measurement configurations. The contact resistance in the pristine nanosheet increased from about 4.1 to 14.0 kΩ, corresponding to a decrease in temperature from 300 to 100 K, and accounted for approximately 40% of the total device resistance. The surface decoration with metal nanoparticles marginally reduced the semiconducting properties of the pristine RuO 2 nanosheet. As shown in Figure 4b and inset, the Cu doping effect decreased the temperature dependence of resistance and decreased the contact resistance, to approximately 3.6 kΩ at 300 K and 9.1 kΩ at 100 K. The proportion of contact resistance in the total device resistance was approximately equal to the proportion of contact resistance before Cu doping. We observed that the proportion remained unchanged (a deviation of 0.6%) at low temperatures. This was because of the increased sensitivity of the carrier density dependence owing to the decreased mobility change at low temperature [29]. The ratios of the resistance of doped nanosheets to the resistance of pristine nanosheets remained equal in both measurement configurations at 100 K. Moreover, the ratios of the contact resistance to the total resistance remained approximately the same (deviation less than 1% at 100 K and 3% at room temperature) after the doping process. These facts suggest that the contact resistance of RuO 2 nanosheets is determined by the carrier density over the entire temperature sweep range.

Contact Resistance
We demonstrated these phenomena of the contact resistance in the Ag treatment with a higher doping effect. As shown in Figure 4c, the semiconducting properties of the RuO 2 nanosheets decreased during each doping process. This phenomenon was represented using the normalized resistance calculated from the following equation: R normalized = (R T − R T = 300 K )/R T = 300 K . The contact resistances were successfully extracted from the noisy raw data of the four-probe resistances by fitting the measured data ( Figure 4d). As shown in the most conductive nanosheet, the Cu nanoparticles decreased the contact resistance in this nanosheet with low conductivity, and an additional reduction was demonstrated through subsequent Ag nanoparticle decoration. In contrast, the contact resistance values of this nanosheet were found to be five-fold larger than that of the most conductive nanosheet. This is because of the incomplete surface treatment before the metallization of the electrodes, leading to an increase in the total resistance, as well as the amplified noise, by degrading the benefits of the four-probe technique.

Conclusions
We analyzed the changes in electrical conductivity and contact resistance of RuO2 nanosheets using charge transfer doping effects based on surface metal nanoparticle decoration. The electrical conductivity of RuO2 nanosheets increased gradually because of the surface decoration with Cu-and Ag nanoparticles. The additional carriers transferred from the metal nanoparticles to the RuO2 nanosheets decreased the semiconducting temperature behavior of the pristine RuO2 nanosheets, due to insufficient carrier density. We obtained the contact resistance between the nanosheet and electrodes from the resistances measured using the four-and two-point probe techniques. Moreover, the contact resistance decreased gradually during the surface decoration processes. From a quantitative analysis based on the two different measurement configurations, we observed that the resistance and carrier density of the nanosheets had a significant impact on the contact resistance. The surface decoration of metal nanoparticles for external contact engineering, as well as the modulation of the internal properties of nanomaterials, can be successfully used for future research in the field of 2D materials.

Conclusions
We analyzed the changes in electrical conductivity and contact resistance of RuO 2 nanosheets using charge transfer doping effects based on surface metal nanoparticle decoration. The electrical conductivity of RuO 2 nanosheets increased gradually because of the surface decoration with Cu-and Ag nanoparticles. The additional carriers transferred from the metal nanoparticles to the RuO 2 nanosheets decreased the semiconducting temperature behavior of the pristine RuO 2 nanosheets, due to insufficient carrier density. We obtained the contact resistance between the nanosheet and electrodes from the resistances measured using the four-and two-point probe techniques. Moreover, the contact resistance decreased gradually during the surface decoration processes. From a quantitative analysis based on the two different measurement configurations, we observed that the resistance and carrier density of the nanosheets had a significant impact on the contact resistance. The surface decoration of metal nanoparticles for external contact engineering, as well as the modulation of the internal properties of nanomaterials, can be successfully used for future research in the field of 2D materials.