Sensing Responses Based on Transfer Characteristics of InAs Nanowire Field-Effect Transistors

Nanowire-based field-effect transistors (FETs) have demonstrated considerable promise for a new generation of chemical and biological sensors. Indium arsenide (InAs), by virtue of its high electron mobility and intrinsic surface accumulation layer of electrons, holds properties beneficial for creating high performance sensors that can be used in applications such as point-of-care testing for patients diagnosed with chronic diseases. Here, we propose devices based on a parallel configuration of InAs nanowires and investigate sensor responses from measurements of conductance over time and FET characteristics. The devices were tested in controlled concentrations of vapour containing acetic acid, 2-butanone and methanol. After adsorption of analyte molecules, trends in the transient current and transfer curves are correlated with the nature of the surface interaction. Specifically, we observed proportionality between acetic acid concentration and relative conductance change, off current and surface charge density extracted from subthreshold behaviour. We suggest the origin of the sensing response to acetic acid as a two-part, reversible acid-base and redox reaction between acetic acid, InAs and its native oxide that forms slow, donor-like states at the nanowire surface. We further describe a simple model that is able to distinguish the occurrence of physical versus chemical adsorption by comparing the values of the extracted surface charge density. These studies demonstrate that InAs nanowires can produce a multitude of sensor responses for the purpose of developing next generation, multi-dimensional sensor applications.


Transient Current in Acetic Acid at V GS = -1V
Figure S1 shows a sensing experiment performed with acetic acid and an InAs multi-NWFET. The top trace shows the relative vapour concentration set-point of the rotameters during the exposure. In this experiment, the back-gate was set to a constant DC bias of -1V, which was the threshold voltage for this device.
All the features of the exposure as discussed in Section 3.3 of the main text are presentthat is, an initial drop in current followed by a gradual rise. However, in this case, the baseline continues to rise significantly past the initial values of source-drain current. We believe this to be due to a constant variation in the nanowire capacitance due to adsorption. This change in capacitance alters the electrostatic coupling of the back-gate to the NW channel, which leads to a change in operating point of the FET. In this case, it becomes difficult to compare the changes in conductance at different vapour concentrations, since the change in electron density is affected by both the change in gate coupling and the influence of adsorption.  Figure S2 shows the transfer curves presented in the Sections 3.1 and 3.2 of the main text plotted with a log scale for IDS. From these plots, the value of subthreshold swing (SS) can be extracted as the inverse of the slope of the log-linear region of the transfer curve. Additionally, the effect of analyte adsorption on values of Ioff are readily apparent on this scale.
The analysis in Section 3.2 extracts values of SS for the transfer curves in acetic acid. While it is not clear that these curves show a change to subthreshold behaviour (and a value for VTh), we observe a slight inflection which is coincident with the shoulder of the curves in N2.
The shoulder indicates the change to subthreshold regime, and, in acetic acid, we take this as an indication that the transfer behaviour of the underlying NW is being masked by the effect of adsorption. Hence, the same values of VTh as in N2 are employed in the formula calculations.

Continuous Transient Current Measurement from Figure 3
Figure S3 shows the continuous current-time trace that was shown in cropped form in Figure 3 of the main text. As in Figure S1, the top trace shows the relative vapour concentration used to set the rotameters. In the second period (ca. 100 -120 min), the rotameter deviated low from its set point for an extended length of time before correction. However, the current response thereafter agrees with the trend from other periods exposed to the same vapour concentration.
The three curves shown in Figure 3b are taken from the 4 th to 6 th period (ca. 170 -230 min), as the timing of the exposures are the most similar. That is, in these three instances, the time spent in N2 prior to exposure, the time exposed to acetic acid vapour, and the time required for the device to return to a baseline after removal of vapour from the chamber.
In the third period of exposure (ca. 140 -150 min), we observe a greater drop in current once acetic acid is added to the chamber. This is likely related to the greater amount of time spent in N2 as compared to other periods of exposure (i.e. 20 min vs. 10 min). In the main text of the paper, this is attributed to time allowed for desorption of hydrogen from the acid-base reaction, and thus the re-activation of some surface sites for the proposed reaction.