Investigation of the Salt Concentration Dependence of Water-Gated Field Effect Transistors (WG-FET) Using 16-nm-Thick Single Crystalline Si Film ^†

Abstract

This paper presents the effect of NaCl concentration on the operation of a water-gated field effect transistor (WG-FET) that uses 16-nm-thick single crystalline silicon (Si) film. In WG-FET, electrical double layer (EDL) formed at the water/silicon interface behaves as gate dielectric and this fluidic interface makes WG-FET a suitable device for sensing applications. Characteristics of EDL and the threshold voltage of WG-FET depend on the molarity of solution. Increasing the molarity of NaCl solution from 0.5 to 65 mM changes the threshold voltage from 360 to 465 mV. Accordingly, drain current of the WG-FET device changes with NaCl concentration.

1. Introduction

Water-gated field effect transistors (WG-FET) that use 16-nm-thick mono-Si films as channel layers promise a new platform where chemical/biological sensors and their read-out circuits can be integrated together. By using electrical double layer capacitance (EDLC) as gate insulator, high channel control can be achieved at low gate voltages with cheap and easy device fabrication. Ultra-thin and high mobility characteristics of the channel layer result in high surface sensitivity in sensor applications and on-site amplification in their read-out circuits.

Applied gate potential in this device induces a voltage across the EDL capacitance, which causes accumulation of holes on the surface of p-type thin Si film as shown in Figure 1. This accumulation type device contains no p-n junctions, which reduces the fabrication cost and light sensitivity. Otherwise, electron-hole pairs can be created in these junctions in the presence of light and introduce noise to the system. I_DS current of the WG-FET is determined by EDL capacitance and the threshold voltage. Similar to Si nanowire FETs [1] and ion-sensitive FETs (ISFET) [2], both threshold voltage and EDL capacitance, hence I_DS, change with pH and the ion concentration of the solution as given by equations in Section 3. In this work, the effect of NaCl concentration on the operation of these WG-FETs is investigated for the first time.

2. Materials and Methods

WG-FET device is fabricated using a silicon-on-insulator (SOI) wafer. Fabrication steps are illustrated in Figure 2. First, 16 nm thick Si device layer is patterned to form the active region of the transistor on top of the 145-nm-thick SiO₂. Then, 200-nm-thick aluminum (Al) is thermally evaporated and patterned to build source and drain electrodes. After thermal annealing at 500 °C, ohmic contacts are obtained between Al and Si films. SiO₂ in the field regions are insulated with photoresist (PR) for repeatability enhancement [3]. Source and drain electrodes are also insulated to prevent galvanic corrosion of Al in NaCl solution.

A solution droplet is placed on top of the active region and a probe is immersed as the gate electrode. A device under test is shown in Figure 3. Contact pads are connected to the wires using silver epoxy paste.

3. Results and Discussion

Characteristic I_DS-V_DS curves of the WG-FET with de-ionized (DI) water are presented in Figure 4, showing the typical transistor characteristics of the WG-FET used in this work. Applied voltages are limited to 0.6 V to eliminate any electrochemical reaction in NaCl solution, which keeps EDL operations in non-Faradaic region. Transfer characteristic curves of the transistor (I_DS-V_GS) are measured at V_DS = V_DD for different molarities, as shown in Figure 5a. Increasing NaCl concentration reduces the on-current of the WG-FET.

Threshold voltages for each NaCl molarity are extracted from the I_DS-V_GS curves and plotted in Figure 5b. These results can be explained as below with the complex dependence of EDL capacitance and threshold voltage on the ion concentration of the solution.

The WG-FET contains two EDL capacitances, one at the surface of probe gate electrode and the other on active channel. These are connected in series, hence when a potential is applied to gate electrode, only some portion of it drops across the EDL of the active channel, which we refer as the effective gate voltage V_{GS_eff}. The current passing through the WG-FET in the saturation region can be characterized by the equation

$${\mathrm{I}}_{\mathrm{DS}}=\frac{1}{2}\frac{\mathrm{W}}{\mathrm{L}}{\mathsf{\mu }}_{\mathrm{p}}{\mathrm{C}}_{\mathrm{edl}}{\left({\mathrm{V}}_{\mathrm{GS}\_\mathrm{eff}}-{\mathrm{V}}_{\mathrm{T}}\right)}^2,$$

(1)

where W is the channel width, L is the channel length, µ_p is the mobility of holes in thin-Si, C_edl is the EDL capacitance on the active channel and V_T is the threshold voltage. C_edl is a function of ion concentration. V_T depends on both ion concentration and pH as shown in Equations (2)–(7). These complex dependences explain the change of the WG-FET current with different solutions. EDL capacitance can be given as [4]

$${\mathrm{C}}_{\mathrm{edl}}=\sqrt{\frac{2z^2{q}^2\epsilon {\epsilon }_0{n}^0}{\mathrm{kT}}}\cos \mathrm{h}\left(\frac{{\mathrm{zq}\mathsf{\Psi }}_0}{2\mathrm{kT}}\right),$$

(2)

where z is the signed charge of ion, q is the unit charge, ε is the relative permittivity of the solution, ε₀ is the permittivity of air, n⁰ is the number concentration of ion, k is the Boltzmann’s constant, T is absolute temperature and Ψ₀ is the electrostatic potential at the surface with respect to bulk solution. This equation simplifies to Equation (3) for dilute aqueous solutions of NaCl at 25 °C where |z| = 1 [4].

$${\mathrm{C}}_{\mathrm{edl}}=228\sqrt{{\mathrm{I}}_0}\cosh \left(19.5{\mathsf{\Psi }}_0\right),$$

(3)

where I₀ is the ion concentration. Threshold voltage of WG-FET can be expressed as [2]

$${\mathrm{V}}_{\mathrm{T}}={\mathsf{\Psi }}_{\mathrm{bulk}}-{\mathsf{\Psi }}_0+{\mathsf{\chi }}^{\mathrm{sol}}-\frac{{\varphi }_{\mathrm{Si}}}{\mathrm{q}}+\frac{{\mathrm{Q}}_{\mathrm{acc}}}{{\mathrm{C}}_{\mathrm{edl}}}$$

(4)

where Ψ_bulk is the potential of the bulk solution, χ^sol is the surface dipole potential, ϕ_Si is the work function of Si and Q_acc is the accumulation charge density. The change of the surface potential Ψ₀ with respect to the bulk pH is given in Equation (5) [5].

$${\mathsf{\Delta }\mathsf{\Psi }}_0=-{\mathsf{\Delta }\mathrm{pH}}_{\mathrm{B}}2.3\frac{\mathrm{kT}}{\mathrm{q}}\mathsf{\alpha },$$

(5)

where α is a dimensionless sensitivity parameter varying between 0 and 1, which is [5]

$$\mathsf{\alpha }={\left(\frac{2.3{\mathrm{kTC}}_{\mathrm{edl}}}{{\mathrm{q}}^2{\mathsf{\beta }}_{\mathrm{int}}}+1\right)}^{-1}$$

(6)

where β_int is the intrinsic buffer capacity given as [5]

$${\mathsf{\beta }}_{\mathrm{int}}=2.3a_{H{s}^{+}}{N}_s\frac{K_b{a}_{H{s}^{+}}^2+4K_a{K}_b{a}_{H{s}^{+}}+K_a{K}_b^2}{{\left(K_a{K}_b+K_b{a}_{H{s}^{+}}+a_{H{s}^{+}}^2\right)}^2}$$

(7)

where a_Hs⁺ is the activity of H⁺ at the active silicon layer surface related to the bulk activity, N_S is the number of sites per unit area, K_a and K_b are dimensionless dissociation constants.

Our results can be explained with the complex dependence of EDL capacitance and threshold voltage on the ion concentration of the solution as discussed above [4]. The changes in the threshold voltages of our devices with NaCl molarity agree with the study of Bergveld on the effect of NaCl concentration on ISFETs, which are depletion/inversion mode bulk Si devices with large reference electrodes, and have SiO₂ insulation [2].

4. Conclusions

It is concluded that concentration of NaCl solution used in WG-FET effects its characteristics by changing threshold voltage and EDL capacitance on the surface of thin Si film. WG-FETs can be both chemical/biological sensors and transistors, leading to an integrated platform where sensors and their read-out circuits are together.