Keywords

1 Introduction

In both chemical and biological processes, the measurement of pH is crucial [1, 2]. Although a glass electrode can measure the pH, it has some drawbacks, such as the fragility of glass and less flexibility, so it cannot be employed in the food industry. To measure pH, different types of field effect transistors are modified, such as ion sensitive field effect transistor (ISFET) and extended gate field effect transistor (EGFET), which is reported in the literature [3].

High sensitivity, selectivity, reliability, and affordability are the performance criteria for evaluating pH sensors. Recently, ZnO-based detectors have caught much attention because of the nontoxicity, thermal stability, electrochemical signaling, fast electron transfer, and biocompatibility nature of ZnO [4]. ZnO is a wide-band gap semiconductor and has a high ionic bonding. Zinc oxide as a sensing material is analyzed in many fields, such as chemical sensors, solar cells, electrochemical cells, biosensors, and light-emitting diodes [5, 6]. Apart from sensing hydrogen ion (H+), recently, ZnO-based nanorod FET has been fabricated for the detection of phosphate (\({\text{PO}}_{4}^{ - 3}\)), nitrate (\({\text{NO}}_{3}^{ - }\)), and potassium ions (K+) by K. S. Bhat et al. with high sensitivity and selectivity and low detection limits [7]. ZnO as a sensing layer is integrated with EGFET, and this device shows high voltage & current sensitivity (53.21 mV/pH and − 10.43 µA1/2), respectively [8]. To examine in a noninvasive manner, ZnO-based EGFET was fabricated by sol-gel process, and the device shows a voltage sensitivity of 63.15 mV/pH and linearity of 0.99 [9].

The fabrication of the semiconductor-based sensor is the succession of well- established Si Technology [10]. The silicon-based sensor is favorable because of its low cost, reproducibility, and controllable electronic response. However, this sensor cannot work at critical points such as high temperature and pressure. Wide-bandgap semiconductors such as ZnO can replace mature silicon technology because of their attractive properties (biocompatibility, piezoelectric, non- toxicity, stability, and label-free sensing) [11]. ZnO can absorb the protein or enzymes with a low isoelectric point (IEP) of 4.5 because of its high isoelectric point (9.5) [12].

To measure the pH among a variety of biosensors, ISFET is mainly used, which has multiple advantages such as rapid response, small size, low output impedance, highly reliable, and on-chip integration of biosensor arrays [13]. ISFET is employed to detect the enzyme, immune, DNA, and cell-based material used in the healthcare industry, environment monitoring, DNA detection, food industry, and biomedicine that can be easily integrated with microfluidic for sensing applications. In 1970, Bergveld first fabricated the ISFET; since then, continuous modifications have been going on to improve the performance and figure of merits (sensitivity & signal-to-noise ratio) [14].

To overcome the short channel effects, achieve high sensitivity, and attain high switching capability, a conventional sensing layer of ISFET is replaced by a new material ZnO. This structure is a field effect transistor compatible with the CMOS fabrication process. Portability, compactness, simplicity, cheap cost, and ease of manipulating the sensor based on FET enable in situ monitoring [7]. A FET can work as a biosensor by replacing a metal gate electrode with a sensitive layer, which reacts with the solution to be measured. The gate and reference voltage are supplied to provide biasing in the circuit.

In literature, ISFET is characterized through simulation, modeling, and fabrication [15, 16]. ISFET is an electrochemical sensor with chemical input in the form of ion concentration. In an ISFET at interfaces (electrolyte and electrolyte-oxide), an electrochemical reaction happens that helps detect ions. Gouy-Chapman explained this electrochemical reaction in terms of Stern double-layer capacitance [17].

In this paper, in conventional ISFET, gate oxide material SiO2 is accompanied by ZnO as a sensing layer in addition to junctionless geometry in the structure, significantly improving sensitivity. The design of ISFET is almost similar to MOSFET, where a reference gate replaces the top gate in the electrolyte solution with a sensing membrane. A stern layer surrounds the top gate layer with a capacitance of 20 µF/cm−2 and a thin membrane layer.

2 Model Simulation

The projected device’s schematic construction is depicted in Fig. 1, and the measurement of various layers in device design is represented in Table 1. The electrolyte solution is considered an intrinsic semiconductor because of similar properties of cations and anions with electrons and holes in the semiconductor. The electrolyte solution is simulated considering the dielectric constant of water (k = 78), electron & hole mobility of ZnO, and properties of the ZnO semiconductor (affinity, permittivity, energy bandgap, and density of states) [18]. The intrinsic semiconductor material consists of thermal-generated mobile carriers. The ion molar concentration of the electrolyte solution is defined by the density of states in the conduction and valence bands [19]. The electrolyte solution interacts with the oxide (sensing membrane) and generates surface charges.

Fig. 1
A diagram presents the basic structure of Z n O IS-FET. A Z n O layer is placed on the silicon layer. An electrolyte solution between the source and the drain is placed on the Z n O layer. A reference electrode is provided on the electrolyte solution.

Schematic structure of proposed ZnO ISFET

Full size image
Table 1 Measurement of various layers of ZnO ISFET
Full size table

The structure consists of a ZnO semiconductor material as a sensing layer, enhancing the figure of merits. The simulation parameters and framework of the proposed device have been maintained coherently with our previous works [20]. A thin membrane layer is also considered between the electrolyte solution and the stern layer. The membrane layer is simulated as silicon with a high dielectric constant (k = 77.8), and the density of states in the conduction and valence band is (Nc = 4.16199 × 1025 and Nv = 3.74 × 1023), respectively. The stern layer is considered an oxide of thickness 1 nm and k = 21. Various models such as concentration-dependent mobility (conmob), field-dependent mobility (fldmob), Fermi-Dirac model (fermidirac), Shockley-Read-Hall recombination model (srh), and basic model are applied during the simulation of the proposed ISFET. The charge concentration is considered at the interfaces (electrode and electrode/oxide interface) to model the device accurately.

In an ISFET, the concentration of ions can be measured by electrical potential when the sensing layer interacts with an electrolyte solution. The sensitivity parameters can be derived from the site binding model [21]. In the proposed device, ZnO is the insulating layer; when this layer interacts with the pH solution following reaction occurs at the interface:

$${\text{ZnOH}} \leftrightarrow {\text{ZnO}}^{ - } + {\text{H}}^{ + }$$
(1)

The membrane layer incorporates the charged ions, which are situated between the electrolyte solution and the stern layer. The membrane layer assists in the alignment of the biomolecules to be detected in the sensing layer. Water in an ionized form is represented as

$${\text{H}}_{2} {\text{O}} \leftrightarrow {\text{H}}^{ + } + {\text{ OH}}^{ - }$$
(2)

The concentration of carrier is described as per Boltzmann Statistics considering holes equal to H ions and electrons equal to hydroxyl ion (p = [H+], n = [OH−], and Ec − Ef = Eg/2). The density of states in the conduction and valence band (Nc and Nv) is defined as

$$N_{c} = {\text{ne}}^{{\frac{Eg}{{2kT}}}}$$
(3)
$$N_{v} = {\text{pe}}^{{\frac{Eg}{{2kT}}}}$$
(4)

The link between density of states in the conduction and valence band and the pH change in the electrolyte solution is established by above equations.

3 Simulation Results and Discussion

The device is characterized at VGS = 0–2 V, VDS = 0.1 V, and the electrolyte solution contains pH = 3–11. Figure 2 shows the contour plot of potential along the device at all voltages equal to zero (VGS = VDS = 0 V). A horizontally cut line is drawn at the surface of ZnO to extract the potential.

Fig. 2
A heat map distributes the potential variation along the materials of the Z n O IS-FET. Horizontal cut lines are drawn at the surface of Z n O to extract the potential.

Contour plot of the variation of potential along the device

Full size image

The potential fluctuation in the channel region is illustrated in Fig. 3. The sensing performance of the ISFET depends upon the sensing membrane’s potential variation. The potential changes with pH according to the site binding model. The variation in potential further leads to variations in the drain-to-source current of ISFET. Site binding theory states that variations in the surface potential voltage at the interface of the aqueous electrolyte solution-sensing layer can result from the concentration of active binding sites on the oxide layer. The surface potential of the sensing layer and pH of the electrolyte solution are related through site binding model, and mathematically, it is represented as [22]:

$$2.303\left( {{\text{pH}}_{\text{pzc }} - {\text{pH}}} \right) = \frac{{q\varphi_{o} }}{kT} + \sin h^{ - } \frac{{q\varphi_{o} }}{kT}* \frac{1}{\beta }$$
(5)
Fig. 3
A multiline graph of potential versus horizontal distance presents the potential fluctuation in the channel region. The potential changes with changes in the p H values between 3 to 11.

Variation of potential along with pH solution

Full size image

where pHPZC is the pH value at the point of zero charge on the sensing layer, q is for the charge of a single electron, k stands for Boltzmann’s constant, T defines the temperature, and β is the sensitivity parameter that indicates the sensitivity of the oxide layer.

The variation of energy in the conduction band and valence band at different pH values ranging from 3 to 11 is shown in Fig. 4. An Ag/AgCl reference electrode is considered, which has stability against variations in the electrolyte solution. The reference electrode is dipped in the pH solution to measure the current at different pH values. The reference electrode supplied a bias to the sensing layer, and the drain current was measured. The drain-to-source voltage value is kept at a low value (0.1 V) to minimize the leakage current. At zero gate voltage, leakage current is almost negligible (10−13 A/µm), and at high gate voltage VGS = 2 V, the current is in the range of (10−5 A/µm). The con- centration of ions varies with pH, so the current is because ions are exchanged between an aqueous electrolyte solution and the sensing surface layer. The device also exhibited a high ION/IOFF current ratio, i.e., 10−8, as shown in Fig. 5. By dividing the highest on-state current by the most negligible off-state current, the ION/IOFF ratio is calculated.

Fig. 4
A multiline graph of energy versus horizontal distance presents the energy fluctuation in the conduction band and the valence band. The energy changes with changes in the p H values between 3 to 11.

Energy band diagram of the proposed device in the conduction and valence band

Full size image
Fig. 5
A multiline graph of drain current versus gate voltage presents the rising trends at V d s equal to 0.1 V, with changes in the p H values between 3 to 11.

Transfer characteristics of a ZnO-based ISFET

Full size image

At low pH values, the concentration of hydrogen H+ ions is higher in the solution, accumulating in the oxide interface. More positive voltage is provided to the gate electrode when more positive hydrogen ions are present, boosting the drain current.

The transconductance curve of a ZnO ISFET is shown in Fig. 6 at gate voltage VGS = 0–2 V in 0.01 steps, with the variation in pH values ranging from 3 to 11. The transconductance (gm) is calculated by differentiating Fig. 5 at the constant drain-to-source voltage VDS = 0.1 V. gm values increase at a particular gate voltage, and then, it starts decreasing. The transconductance is highest (0.342 mS/µm) at VGS = 1.9 V and pH = 11. A high value of gm denotes the greater drain current change for a given surface charge change.

Fig. 6
A multiline graph of transconductance versus gate voltage presents the 9 bell-shaped trends for the p H values between 3 to 11. The V d s is equal to 0.1 V.

Transconductance characteristics along with the variation in pH values

Full size image

4 The Figure of Merits (FOMs)

The proposed device is characterized by various figure of merits (sensitivity, signal-to-noise ratio, and transconductance-to-current ratio).

4.1 Sensitivity

The Nernst equation defines the sensitivity of an ideal biosensor, i.e., 59 mv/pH at room temperature. The Nernst equation is an equation, as shown below, that describes the link between electric potential of the cell membrane and the ionic concentrations that occurred during the electrochemical reaction.

$$E = E_{O} + 2.303\frac{RT}{{nF}}\,{\text{pH}}$$
(6)

where E and Eo are the measured and standard potential, respectively, R defines the universal gas constant, T denotes the absolute temperature, n stands for the number of moles of electrons transferred in the electrochemical reaction, and F is the Faraday constant.

The sensitivity of an ideal biosensor is calculated through the Nernst equation which is given as [23]:

$$S_{{{\text{Ideal}}}} = 2.303 \frac{kT}{q} \approx 59 {\text{mv/pH}}$$
(7)

where kT/q defines by the thermal voltage.

In FETs, sensitivity is calculated from FET’s characteristics, such as transconductance and the on-off ratio of drain current. The average sensitivity is estimated by the change in threshold voltage and drain current with respect to pH. Threshold voltage varies with pH solutions, as does the concentration of carriers in the solution, which further influences the change in drain current. The threshold voltage sensitivity is defined as

$$SV_{{{\text{th}}}} = ~\frac{{V_{{{\text{th}}}} \left( {{\text{pH}}2} \right) - ~V_{{{\text{th}}}} \left( {{\text{pH}}1} \right)~~}}{{{\text{pH}}2 - ~{\text{pH}}1}} = ~\frac{{\Delta V_{{{\text{th}}}} }}{{\Delta {\text{pH}}}}$$
(8)

The device’s sensitivity is tested with the ZnO layer at various pH solutions (pH = 3–11). The gate voltage ranges from 0 V to 2 V, while the drain voltage is maintained at 0.1 V during the simulation. The electrolyte and oxide interface’s properties affect a sensor’s sensing ability.

For the calculation of the threshold voltage, a constant current method (10−7 * W/L) is used at VDS = 0.1 V. Figure 7 represents the voltage sensitivity in the acid medium (pH = 3, 4, 5, and 6) and alkaline medium (pH = 8, 9, 10, and 11). From the voltage sensitivity graph, it can be understood that the device shows 60 mV/pH threshold voltage sensitivity, near the Nernst limit (59 mV/pH).

$${\text{Current Sensitivity }}\left( {\Delta I_{{{\text{DS}}}} } \right) = \frac{{I_{{{\text{var}}}} - I_{{{\text{cons}}}} \left( {{\text{pH}} = 7} \right)}}{{I_{{{\text{cons}}}} \left( {{\text{pH}} = 7} \right)}} = \frac{{\Delta I_{D} }}{{I_{D} }}$$
(9)
Fig. 7
2 scattered plot of voltage sensitivity versus p H presents the linear failing trends for sensitivity and linear fit. A. in the acidic solution. B. In the alkaline solution.

Threshold voltage sensitivity

Full size image

The current sensitivity of the device in mathematical form is represented in above equation and plotted in acid and alkaline mediums shown in Fig. 8.

Fig. 8
2 scattered plots of current sensitivity versus p H. A. Presents linearly failing trends for sensitivity and linear fit in the acidic solution. B. Presents linearly rising trends for sensitivity and linear fit in the alkaline solution.

Current sensitivity

Full size image

4.2 Signal-to-Noise Ratio (SNR)

SNR measures the detection ability of a device, and it is defined as the ratio of usable signal to noisy signal. SNR can also be represented by power spectral density (PSD), and in mathematical form, it is described as [24]:

$${\text{SNR}} = \frac{{\delta I_{{{\text{us}}}} }}{{\delta I_{{{\text{ns}}}} }} = \frac{{\delta I_{{{\text{us}}}} }}{{\sqrt {\smallint S_{I} dF} }}$$
(10)
$$S_{I\left( f \right)} = \frac{{S^{2} \alpha }}{fN}$$
(11)

where SI is the power spectral density of the drain current, f is the frequency, N is the carrier concentration, and α is the Hooge constant.

Figure 9 represents the SNR curve of the proposed device at VDS = 0.1 V and VGS = 2 V, and it shows a sharp peak at VGS = 1.25 V. SNR follows the transconductance, and higher SNR shows a lower detection limit of the sensor [25].

Fig. 9
A line graph of signal-to-noise ratio versus gate voltage presents a bell-shaped curve. The value of v d s is 0.1 volts.

Signal-to-noise ratio of ZnO-based ISFET

Full size image

4.3 Transconductance-To-Current Ratio (gm/IDS)

gm/IDS is also applied as a sensing parameter for FET devices to observe the change in ion concentration [26]. gm/IDS is an analog performance indicator. The transconductance-to-current ratio characteristics of the ZnO ISFET obtained by sweeping gate voltage is shown in Fig. 10. These characteristics follow the pattern of transconductance. It means gm increases at a fixed voltage, attains a peak value, and then starts decreasing with the increasing gate voltage.

Fig. 10
A multiline graph of G m by I d versus gate voltage presents the 9 trends for the p H values between 3 to 11. The g m increases at a fixed voltage, attains a peak value, and then falls with the increase in gate voltage.

Transconductance-to- current ratio curve at various pH values

Full size image

The proposed device is demonstrated through simulation with excellency for pH application and, in future, can be fabricated through a simpler fabrication process. The ZnO-based ISFET can be employed in various fields, including life science, point-of-care testing (POCT) devices, medicine, and product safety. The ever-growing application of ZnO-based devices leads to the accumulation of zinc oxide (ZnO) nanomaterials that cause severe consequences to environment and human health. High-care of insulation and disposal must be taken care of while incorporating such devices in health care and environmental monitoring applications.

5 Conclusion

In this paper, the impact of ZnO as a sensing layer of a junctionless ISFET is studied over the various parameters (potential, drain current characteristics, transconductance, and sensitivity), and the results proved that ZnO is an effective sensitive material for detecting H+ ions in an ISFET. The drain current characteristics show the pH-dependent behavior of a ZnO-based ISFET and confirm that this device can be used as a pH sensor. Polycrystalline ZnO-based TFT pH sensors show Nernst limit sensitivity besides the advantages of ease of fabrication and CMOS process compatibility. The simulated device shows excellent potential as a pH sensor.