Graphene sensor decorated with Pd nanoparticles was fabricated. Graphene films grown by chemical vapor deposition (CVD) were transferred to 90-nm-thick SiO2 on p+-Si substrates. Then, graphene was patterned into Hall bar structure using photo-lithography followed by oxygen plasma etching. The Cr/Au (10 nm / 100 nm) electrodes were deposited at terminals as well as back-side using electron-beam evaporation. Finally, 1-nm-thick Pd deposition was followed by 400 ℃ annealing in N2 for 2 hours to form Pd nanoparticles. The Hall effects were measured at constant drain voltage (VD) of 100 mV and magnetic field (B) of 0.43 T using a neodymium magnet under the sensor chip. Two sensors (devices A and B) fabricated through the same processes were characterized. The channel length/width of device A and B were 400/20 μm and 300/50 μm, respectively. In order to investigate carrier concentration and Hall mobility in hydrogen/nitrogen atmosphere, 1000 ppm hydrogen (nitrogen balance) and nitrogen gases were introduced at the flow rate of 500 mL/min onto the sensor devices.
We investigated sensing performance of Pd-decorated graphene against H2. From 0 to 3 min and from 6 to 9 min, sensors were under inert (N2) ambient. Whereas, from 3 to 6 min, 1000 ppm H2 was introduced. Resistance increase was observed under 1000 ppm H2. VG dependence of carrier concentration (Ns) was observed through Hall effect measurement. Because of the positive-shift of the Dirac point, Ns decreases as VG increases. The changes of Ns and Hall mobility (μH) during 3-min 1000 ppm H2 ambient were summarized in Table I. Note that VG of 0 and 40 V correspond to the larger and smaller Ns, owing to the shift of the Dirac point, respectively.
First, we will discuss Ns change under H2 ambient. For all devices and applied VG, Ns decreases under H2 ambient. In other words, ID-VG characteristics of graphene shift in the negative direction in the amount of −ΔVG. When |VG−Vdirac| is large, the ratio of Ns change ΔNs/Ns should be equal to −ΔVG/VG, where Vdirac is the VG corresponding to the Dirac point. ΔNs/Ns observed in device A can be understood using this relationship. Simple calculation shows |ΔNs/Ns| should be less than |ΔVG/VG|, when VG is close to Vdirac. Therefore, we consider that Vdirac of device B is smaller than that of device A. Next, Hall mobility change (ΔμH) will be discussed. As shown in Table I, μH also changes under H2 ambient. Assume that effective mass of hole is proportional to root of Ns [4] and that carrier scattering mechanisms are the same in H2 ambient as in N2 ambient. Then, it is shown that ΔμH/μH is −ΔNs/2Ns. As shown in Table I, the expected relationship were qualitatively observed. However, ΔμH/μH is always smaller than −ΔNs/2Ns. This is due to the increased Coulomb scattering induced by the hydrogen ions, which constitute the origin of negative shift of ID-VG characteristics under H2 ambient.
We have shown that not only carrier concentrations but also carrier mobility is greatly changed under H2 ambient. The physical mechanisms of mobility change is considered to be due to carrier-concentration-dependent effective mass and increased Coulomb scattering by hydrogen ions.
[1] J. Villatoro et al., Sensors and Actuators B, 110, 23-27, 2005.
[2] R.Kumar et al., Sensors and Actuators B, 209, 919-926, 2015.
[3] W.Wu et al., Sensors and Actuators B, 150, 296-300, 2010.