Effect of solution pH and adsorbent concentration on the sensing parameters of TGN‐based electrochemical sensor

Abstract

The response of trilayer graphene nanoribbon (TGN)‐based ion‐sensitive field‐effect transistor (ISFET) to different pH solutions and adsorption effect on the sensing parameters are analytically studied in this research. The authors propose a TGN‐based sensor to electrochemically detect pH. To this end, absorption effect on the sensing area in the form of carrier concentration, carrier velocity, and conductance variations are investigated. Also, the caused electrical response on TGN as a detection element is analytically proposed, in which significant current decrease of the sensor is observed after exposure to high pH values. In order to verify the accuracy of the model, it is compared with recent reports on pH sensors. The TGN‐based pH sensor exposes higher current compared to that of carbon nanotube (CNT) counterpart for analogous ambient conditions. While, the comparative results demonstrate that the conductance of proposed model is lower than that of monolayer graphene‐counterpart for equivalent pH values. The results confirm that the conductance of the sensor is decreased and V_g‐min is obviously right‐shifted by increasing value of pH. The authors demonstrate that although there is not the experimental evidence reported in the part of literature for TGN sensor, but the model can assist in comprehending experiments involving nanoscale pH sensors.

1 Introduction

Graphene as an extremely attractive material is composed of a two‐dimensional (2D) single atom thick layer of sp² ‐bonded carbon atoms arranged in a honeycomb lattice. Due to outstanding physiochemical and electronical properties of graphene nanoribbon (GN) and its derivatives, they have great potential applications in many scientific fields, such as energy technology, bioscience, and biotechnologies. GN reveals startling physical, mechanical, thermal, electrical, and chemical properties such as high surface area, strong mechanical strength, good thermal conductivity, excellent electrical conductivity, high charge carrier mobility, good optical transparency and ease of biological as well as chemical functionalisation of GN that leads to great opportunities for implementing into a broad area of sensor applications. GN preserves various appreciable properties of carbon nanotube (CNT), like a paucity of surface dangling bonds and a high mean‐free path for scattering. Whereas it minus many of CNT disadvantages, like insensitivity to atomic chirality and the possibility of top‐down lithographic patterning [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]. The structure of multilayer graphene (MLG) is obtained from the hybridisation of single‐layer states through interlayer hopping. Its main properties are taken already with considering hopping between the nearest‐neighbour carbons, which is stacked on top of each other in two neighbouring layers. In MLGs, stacking order offers a significant scarcely explored degree of freedom for adjusting its’ electronical properties. The MLGs reveal rich exquisite phenomena at low charge densities because of enhanced electronic interactions and competing symmetries. Trilayer graphene nanoribbon (TGN) as a MLG is characterised by two natural stable allotropes as ABA and ABC. TGN is a semimetal with a resistivity that reduces with rising electric field, a behaviour that is noticeably different from that of monolayer and bilayer graphene. In ABA or Bernal‐stacked TGN (B‐TGN), the atoms of the highest layer lie precisely above those of the bottom layer. While in ABC or rhombohedral‐stacked (r‐TGN), one sub‐lattice of the top layer lies on top of the centre of the hexagons in the bottom layer. This delicate difference in stacking order leads to an exhibitive distinction in band structure and transport features. At the Dirac point, B‐TGN becomes metallic, whereas r‐TGN remains insulating with an inherent interaction‐driven gap of 6 meV [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. In fact, ABA‐stacked TGN shows semi‐metallic behaviour with a tunable band overlap; however, ABC‐stacked TGN with a tunable band gap is known as a semiconductor which is selected for the proposed sensor. The use of electronic‐biology devices capable of monitoring and detecting biomolecular events can enable enormous advances in the detection and treatment of diseases for biomedical areas [31, 32, 33, 34, 35]. Ion‐sensitive field‐effect transistors (ISFETs) have been captivated recently the significant attention of researchers because of their properties such as small size, the possibility for mass production, short and steady response times [35]. By changing the gate material of ISFETs, depositing layers of appointive membrane or a bio‐recognition element on top of the gate, variance of selectivity can be achieved. The gate material is basically sensitive to pH changes because of its organic nature. All enzymes usually work in a narrow pH range, and they are very sensitive to pH changes. pH is an important factor in enzyme stability, and each enzyme includes an appropriate or optimal pH stability range [35]. Electrolyte‐gated graphene FETs for detecting pH and protein adsorption, and enzyme‐polyelectrolyte multilayer assemblies on reduced graphene oxide FETs for biosensing applications have been reported [36, 37]. Also, the construction of layer‐by‐layer (LbL) assembly of polyethylenimine and urease onto reduced graphene‐oxide (rGO)‐based FETs for the detection of urea has been presented [37]. This adaptable sensor platform simultaneously uses the pH dependency of liquid‐gated graphene‐based FETs and the local pH change generated by the catalysed hydrolysis of urea [37]. The use of an interdigitated microchannel resulted in FETs showing high pH sensitivity (20.3 μA/pH), low noise, and transconductance value up to 800 μS. The adjustment of rGO FETs with a weak polyelectrolyte improved the response of pH owing to its transducing features by electrostatic gating effects [37]. TGN‐based ISFET for pH sensing is investigated in this research. To this end, the detection mechanism of hydrogen ions adsorbing from solution to TGN surface is analytically studied. In other words, an analytical model for reaction between buffer solution of different pH and TGN is presented here which is based on the electron transfer between ion solutions and TGN. Fig. 1 shows the proposed structure of TGN‐based pH sensor consists of a drain and source, in which TGN is used as a channel material on a polyethylene terephthalate (PET) substrate (standard 4 in). The electrochemical detection of pH value is also performed by means of the ISFET characterisation technique. Integrated reference electrodes are fundamental elements in liquid channels for dynamical binding and unbinding experiments. TGN is the platform for integrated reference electrodes to probe the electrostatic potential in an aqueous electrolyte. The Ag/AgCl reference electrode with internal filling solution of 3 M KCl is absorbed into sample solution as a gate, to make the solution with stable electrical potential. The miniaturised Ag/AgCl reference gate electrode and the wires from drain and source are fixed to the outer frame to avoid the relative movement [38]. Ag/AgCl reference electrode is utilised as the top‐gate electrode to minimise environmental effects. In order to verify the accuracy of the proposed model, it is compared with recent reports on pH sensors, such as fabricated flexible single‐walled CNT (SWCNT) ISFET [38] and monolayer graphene‐based sensors [39]. The fabrication of graphene and SWCNT ISFETs by lithographic techniques with LbL bottom‐up construction has been studied [38, 39], which provides the product cost‐effective owing to the mass production and low‐cost polymer substrate. The poly is deposited via LbL self‐assembly between two electrodes patterned on a PET substrate and poly styrene sulphonate (PSS) is implemented for surface charge enhancement. The polymer substrate is biocompatible because of the material nature, which liquidates the requirement of dielectric layer [38]. There could be electrochemical side‐reactions by exposing wire surface to the solution, which may be detected as gate leakage current [38, 39]. As can be seen in Fig. 1 c, channel conductance of TGN‐based ISFET can be changed by binding of hydrogen ions in the solution to channel surface. Less hydrogen ions are adsorbed when the pH value of the solution increases. Thus, the sensor will be able to attract less ions, leading to the change of the conductance of TGN‐based ISFET.

Fig. 1.

The proposed electrochemical sensor

(a) The structure of TGN as a channel material ($y,z\le {\lambda }_D,x=L\gg {\lambda }_D;{\lambda }_D\simeq 10$ nm), (b) Schematic of the proposed TGN ISFET‐based sensor for pH detection, (c) Schematic of hydrogen ion adsorption processes by surface area of TGN

A series of nanoscale ISFET sensors developed held nanoparticles as a dielectric layer that collects the surface charge carriers, resulting in conductance changes in underlying semiconducting nanomaterial layer [38]. The penetration of ion species into the semiconducting layer is important, which would cause the change in electrical properties of the nanomaterial itself owing to proximal ionic composition [38]. The change in surface charge induced by the shift in ionic composition directly influences the conductance of TGN semiconducting layers. The variation of local pH and proximal ionic composition in the vicinity of TGN, and the direct electron transfer play the significant role of drain current change in the semiconducting layer. Specially, as the steady state reaches from the initial hydrogen ion depletion mode, because of negative gate voltage applied, the concentration of hydrogen ions is maintained constant that ions diffuse away to reach the electrochemical transducer layer or disappear towards the bulk solution owing to the buffering power of the sample solution [38]. On the other hand, the direct electron transfer might take place directly to TGN surface caused by drain‐to‐source voltage and/or gate voltage. The nanomaterial‐based ISFET can be immersed in the electrolyte and a silicone rubber can be placed on the ISFET to allow the surface of the nanomaterial channel to be filled with several buffer solutions and analytes for measurement and sensing [36, 37]. Electrolyte‐gated ISFETs have revealed excellent electrical properties as thin top‐gate insulators with high dielectric constants in ionic solutions. The electrical‐double layer in an electrolyte operates as a top‐gate insulator with a thickness determined by the Debye–Huckel equation, which is the thickness depends on the ionic strength with a concentration of several millimolars [36, 37]. It is notable that the high [KCL]/[HEPES] ratio ensured approximately a constant ionic strength independent of pH [37]. This paper is organised as follows: in Section 2, the proposed computational model for electrochemical pH sensor and sensing mechanism are presented in detail. Moreover, the obtained results and main findings of the model are investigated in Section 3, and concluding remarks are given in Section 4.

2 Proposed computational model

In this section, in order to understand the characteristics of the TGN‐based pH sensor and define its physical and electrical phenomena, the analytical model of the device is presented. According to the current–voltage characteristic of FET devices, the performance of the sensor in the saturation region can be evaluated through (1) [40, 41]. By assuming the substrate terminals and source are held in ground potential, the channel region has the resistor characteristics in small drain‐source voltage (V _DS).

$$I_D=G\times V_{\mathrm{DS}(\mathrm{sat})}=G\times (V_{\mathrm{gs}}-V_t)$$ (1)

where G is the conductance and V_t is the threshold voltage. According to the relation between the conductance and carrier velocity ($G=(\mathrm{nq}\mu /l)=\mathrm{nqv}/\mathrm{lE}$), (1) can be considered as the following equation:

$$I_D=\frac{\mathrm{nqv}}{\mathrm{lE}}\times (V_{\mathrm{gs}}-V_t)$$ (2)

where n is the carrier concentration, q is the el­ectron charge, l is the channel length, E is the band energy and v is the velocity of carriers. By substitution of the proposed model of carrier velocity for TGN [42] in (2), the I‐V characteristic of the proposed sensor is analytically modelled as (3);where $A=-6.2832\alpha$ , $B=14.3849{\alpha }^2\beta$ , $C=2.7444/\beta$ , $D=-9{\beta }^2$, $F=-0.1690{\alpha }^3/\beta$ , $E_{\mathrm{CO}}=E_c/k_{\mathrm{B}}T$, $N=F/(k_{\mathrm{B}}T){}^2$, $\alpha =v_f\mathrm{\Delta }/t_{\perp }\sqrt{2},$ $\beta =v_f^3/t_{\perp }\sqrt{2}\mathrm{\Delta }$ in which $v_f$, Δ, $t_{\perp }$ and E_c are the Fermi velocity, applied voltage, hopping energy and conduction band energy, respectively. Moreover, $k_{\mathrm{B}}$ is the Boltzmann constant, T is the temperature, m is the carrier effective mass, $x=(E_{\mathrm{TGN}}-E_c)/k_{\mathrm{B}}T$ and the normalised Fermi energy is defined as $\eta =(-E_c+E_f)/k_{\mathrm{B}}T$.

In the sensing mechanism of the proposed sensor, the gate‐source voltage is assumed as a function of pH value ($V_{\mathrm{gs}}\sim F_{\mathrm{pH}}^{\mathrm{\prime }}$).

$$I_D=\frac{n_{\mathrm{TGN}}q}{l{E}_{\mathrm{TGN}}}\frac{{\int }_0^{\eta }{\displaystyle \frac{\sqrt{{\displaystyle \frac{2}{m}}}{(k_{B}T)}^{3/2}{x}^{1/2}\mathrm{d}x}{\left\{A-{\displaystyle \frac{B}{{(k_{B}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}-C{(k_{B}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}\right\}(1+\exp (x-\eta ))}}}{{\int }_0^{\eta }{\displaystyle \frac{(k_{B}T)\mathrm{d}x}{\left\{A-{\displaystyle \frac{B}{{(k_{B}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}-C{(k_{B}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}\right\}(1+\exp (x-\eta ))}}}.(V_{\mathrm{gs}}-V_t)$$ (3)

$$V_{\mathrm{gs}(\mathrm{withpH})}=\frac{{\alpha }^{\mathrm{\prime }}}{F^{\mathrm{\prime }}}{V}_{\mathrm{gs}(\mathrm{withoutpH})}$$ (4)

where (${\alpha }^{\mathrm{\prime }}$) is introduced as pH sensing factor and different values of pH is presented in the form of ($F^{\mathrm{\prime }}$) parameter. Thus, the molecules adsorbed on TGN surface is modelled by iteration method [8] as

$${\alpha }^{\mathrm{\prime }}=\mathrm{a}{{\mathrm{F}}^{\mathrm{\prime }}}^2+\mathrm{b}{\mathrm{F}}^{\mathrm{\prime }}+c$$ (5)

From extracted data, a, b, c parameters can be obtained [38, 39]. The interaction between pH molecules and surface of TGN has a significant influence on the I‐V characteristic of the sensor through various mechanisms like electronic doping and electrostatic gating. Eventually, in order to analytical prediction of the sensor performance, the I‐V characteristic of the TGN‐based pH sensor is obtained as (6).

$$I_D=\frac{n_{\mathrm{TGN}}q}{\mathrm{l}{\mathrm{E}}_{\mathrm{TGN}}}\frac{{\int }_0^{\eta }{\displaystyle \frac{\sqrt{{\displaystyle \frac{2}{m}}}{(k_{\mathrm{B}}T)}^{3/2}{x}^{1/2}\mathrm{d}x}{\left\{A-{\displaystyle \frac{B}{(k_{\mathrm{B}}T)2/3{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}-C{(k_{\mathrm{B}}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}\right\}(1+\exp (x-\eta ))}}}{{\int }_0^{\eta }{\displaystyle \frac{(k_{\mathrm{B}}T)\mathrm{d}x}{\left\{A-{\displaystyle \frac{B}{{(k_{\mathrm{B}}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}-C{(k_{\mathrm{B}}T)}^{2/3}{\left\{D(x+E_{\mathrm{CO}})+\sqrt{N+{(x+E_{\mathrm{CO}})}^2}\right\}}^{2/3}}\right\}(1+\exp (x-\eta ))}}}.\left(\frac{\mathrm{a}{{\mathrm{F}}^{\mathrm{\prime }}}^2+\mathrm{b}{\mathrm{F}}^{\mathrm{\prime }}+c}{F^{\mathrm{\prime }}}{V}_{\mathrm{gs}(\mathrm{withoutpH})}-V_t\right)$$ (6)

3 Results and discussion

By utilising the proposed model in the preceding section, obtained results are indicated in this part. Fig. 2 shows all theoretical I_D –V _DS characteristics of TGN‐based electrochemical sensor when the concentration of hydrogen ions changes from pH = 5 to pH = 9. V _DS is scanned from 0 to −1.0 V with a step of 100 mV. Apparently, the drain current decreases by increasing the pH value. In fact, the proposed model points out pH value dependence of I_D –V _DS characteristic. It is notable that the pH is characterised based on the molecular protonation/deprotonation of carboxylated TGN to illustrate reduction in drain current with increment in pH. The charged ions dissolved in pH buffers can also deteriorate the positive threshold voltage and low I_on /_off . The charged ions can be also localised positively with high negative gate voltage, suppressing the ionic conducting effect. At low gate voltage, V _DS plays a main role of driving charged species for ionic conductor. Saturation effects are discovered in drain current because of the gate voltage that suppresses the ionic current and localises charged ions [38]. The pH response of TGN ISFET and fabricated flexible SWCNT ISFET sensors [38] are shown in Fig. 3 for different values of pH (5, 7, 9). As depicted in Fig. 3, the presented model points out strong concentration dependence of I_D –V _DS characteristic indicating that the effect of pH value increase will influence the I_D .

Fig. 2.

I_D ‐V_DS characteristic of TGN‐based electrochemical sensor for different pH values (5–9)

Fig. 3.

Representative I_D ‐V_DS curves of TGN and SWCNT ISFET‐based sensors [38]

(a) pH = 5, (b) pH = 7, (c) pH = 9

The sensitivity of the pH sensors under values of 5, 7, 9 are investigated in Fig. 3. Before the concentration of hydrogen ion is changed in the solution, a natural solution with a buffer (pH 7) is added in the electro‐active membrane to measure the dependence of drain current versus drain–source voltage. First of all, the TGN ISFET and flexible SWCNT ISFET sensors are characterised at pH 5 buffer to exhibit an essential functionality of FET. It is apparently observed in Fig. 3 a that higher drain current flows at a lower pH at the fixed V_gs  = −1.5 V through the V _DS tested. The field‐effects can be considerably seen at pH 5 buffer. Also, different pH values corresponding to the different controlling factors are considered in Figs. 3 b and c. Apparently, as the sensors are exposed to different pH values ($F_{\mathrm{pH}}^{\mathrm{\prime }}=7,9$), the different drain currents are resulted. It can be evidently observed that the lower currents are obtained as the sensors are exposed to more pH value. It is noteworthy that in SWCNT resistors where the gate electrode is not used, current–voltage characteristic is in parabolic relationship. However, saturation effects in ISFETs are found in current because of the V_g , which localises charged ions and represses the ionic current [38]. The drain current values of TGN/CNT sensors for different pH values are summarised in Table 1, which are tested at fixed V _DS  = −1.0 V and V_g  = −1.5 V. As specified in Table 1, the I_D for both sensors decreases with increasing the pH value.

Table 1.

Drain current values of TGN/CNT sensors for different pH values at V_DS  = −1.0 V and V_g  = −1.5 V

pH value	5	6	7	8	9
ID (TGN biosensor), µA	586.836	499.536	427.309	367.473	346.499
ID (CNT biosensor), µA	337.89	302.8	273.45	254.0	245.34

Fig. 4 shows the transfer curves of the electrolyte‐gated TGN ISFET indicating the conductance is a function of gate voltage (V_g ) for different pH buffer solutions of 5 to 10. A bipolar FET characteristic can be observed within an operation voltage of <0.8 V. This characteristic demonstrates that the type of charge carriers can be tuned from holes (p‐type region) to electrons (n ‐type region) by the liquid gate that controls the electrochemical potential (Fermi energy) of TGN carriers. The conductivity is influenced by the increased number of channel carriers. TGN conductance is minimal at the transition between the electron and hole regime. This point is named as the charge‐neutrality point (CNP). The conductance curve is approximately symmetric around V _CNP or V _g‐min, where the density of carriers (electron and hole) is equal. It saturates at large carrier concentration at a value of 300_320 μS, corresponding to a built in series resistance of ≃3.1 kΩ. The minimum conductance value G_min ≃210 μS is changing slightly with time. V _g‐min is susceptible to the immobilisation of probe pH and hybridisation of complementary target pHs. By immobilisation of probe pHs on the TGN surface, TGN conductance decreases and V _g‐min is obviously right‐shifted. In total, V _g‐min shift can be considered as a good indicator for pH detection. pH‐dependent conductance without the gate reference electrode reveals that the local pH in the proximity of TGN plays a role of the substitutive V_g which is improved by external electric voltage. The transfer characteristics of TGN‐based electrochemical pH sensor in solutions with constant ionic strength and pH ranging from 5 to 10 can be investigated from Fig. 4. While the slope remained almost unchanged for both sides of the V _CNP, indicating that the mobility of charge carriers is nearly independent of the pH. A dominant shift of V _CNP to more positive values of gate voltage is observed with increasing pH [37]. It is notable that the increased conductance in ISFET‐based pH sensor can be related to the increased negative charge around the nanomaterial channel, since the hole is the carrier in this condition. Also, increased conductance can be interpreted as the attachment of hydroxide ion that acts as electron scavenger. These phenomena can occur on the nanomaterial channel and the results demonstrate that ISFETs can be used as pH sensors [36]. Graphene‐based devices for measuring pH, edge effects on the pH response of GNFETs and suspended GN‐ISFETs formed by shrink lithography for pH/cancer biomarker sensing have been investigated recently [43, 44, 45]. In G‐based pH sensor, standard photolithography has been utilised to specify source and drain electrodes which are produced by a bilayer of Ti (5 nm) and Au (60 nm) [39]. There are two extra fabrication steps for the operation of the sensor in an electrolyte environment. First, a liquid channel is located above the graphene which is determined in a photoresist layer by UV lithography. Second, the device is mounted on a chip carrier and wire bonded and then an epoxy layer is deposited above the contact pads including the bonding wires. Furthermore, the liquid channel and the outer epoxy layer are applied. As substrate, pieces of p‐doped Si covered with a 300‐nm‐thick layer of thermal SiO₂ are used [39].

Fig. 4.

Conductance‐voltage characteristic of TGN‐based electrochemical sensor for different pH values (5–10)

According to results of ref [39], very small V_g shifts in the transfer characteristic of GFET as the pH of the buffer is changed. The low gate voltage‐shift of GFET can be further increased if the surface of graphene is covered with a thin Al‐oxide layer. If a hydrophobic fluorobenzene layer is utilised instead, the opposite situation will happen. As can be seen in Fig. 5, by varying proton concentration in solution through the controlling factors, conductance‐voltage characteristic of the TGN ISFET can be controlled. The proposed TGN device exposes lower conductance compared to that of monolayer graphene‐counterpart for equivalent ambient conditions [39]. According to Fig. 5, the conductance decreases by increasing the pH value. It is demonstrated that the transfer characteristic of TGN (V _g‐min) also shifts to the right when the pH of the buffer is changed. Adoption of electrostatic gating effect by adsorbed charge species explains the shift of conductance‐voltage curve of the sensor in Figs. 4 and 5. If electrostatic gating mechanism can explain our results, positive shift of V_g would be expected. Table 2 shows the effect of pH value of the buffer solution on the conductance of TGN and G‐based sensors, which are tested at V _g‐min. It can be stated that, by increasing the pH value, the conductance of both sensors decreases.

Fig. 5.

PH response of the TGN and flexible fabricated G‐based sensors [39] for different concentrations

(a) pH = 5, (b) pH = 7 (c) , pH = 9, (d) pH = 10

Table 2.

Conductance of TGN/G sensors for different pH values at V_g‐min

pH value	5	6	7	8	9	10
G (TGN‐biosensor), µs	227.81	224.81	223.81	220.81	216.81	209.81
G (G ‐biosensor), µs	267.2	259.72	258.26	255.74	247.42	234.1

The pH‐responsive I_D for TGN and SWCNT ISFET‐based sensors are depicted in Fig. 6 a. The drain current value of the sensors decreased as the increase of pH value in the range of 5–9. Moreover, TGN‐based sensor exposes higher current compared to that of CNT‐counterpart for analogous pH values. It is noteworthy that the current of the electrochemical sensors is dependent on pH, which demonstrates that molecular protonation/deprotonation plays an important role of the electrical property change in nanomaterials. A comparison of the pH‐responsive conductance for TGN and G‐based sensors at V _g‐min is also indicated in Fig. 6 b. It can be observed that the conductance decreases with the increase of the solution pH and reaches the minimum value at pH 10. Furthermore, TGN sensor exposes lower conductance compared to G‐based sensor for the equivalent conditions. The performance of the nanomaterial‐based ISFET, including repeatability, reproducibility, stability, recovery and sensitivity, can be investigated. In this sense, the use of solution‐gated nanomaterial‐based ISFET looks like to be a more beneficial approach as this method consists the following attractive characteristics: real‐time response, operation in aqueous solutions at very low voltages and higher sensitivity than conventional electrochemical techniques due to their inherent amplification feature [37]. The nanomaterial‐based ISFET pH sensor sensitivity reveals an appreciable value of relative standard deviation (RSD), demonstrating good device‐to‐device reproducibility. It should also be pointed out the nanomaterial‐based ISFET pH sensor exhibited a good long‐term stability and a fast response [37]. The relatively large pH sensitivities for nanomaterial can be realised by considering nanomaterial of different quality. A high transconductance (ΔI _ds /ΔV_g ) is a basic feature for the development of ISFETs with improved sensitivity for sensing applications [37]. As the sensitivity depends on the transconductance, the reasons for a higher sensitivity are related to: (1) the use of ISFET characterisation technique, where the nanomaterials‐based ISFETs indicated significant electrical stability in pH sensing and gate voltage influences on the oxidation of hydrogen peroxide, (2) the use of interdigitated microchannels with large channel width (W) and short channel length (L), and (3) use of flexible polymer substrate, where the bioactivity of immobilised enzyme can be maintained [37]. Fig. 7 a illustrates the dependence of V_CNP versus pH. For all data obtained in subsequent evaluations, a consecutive curve is fitted to the data points and employed to extract the reference voltage V _CNP for which G (V _ref) is minimal. Apparently, V _CNP shifts with increasing pH to greater voltages. In other words, V _CNP shifts quite noticeably when the pH is changed and the transfer curve of the sensor is expected to shift to the right. The extracted pH sensitivity is 7 mV/pH, which is smaller than the Nernst limit of 60 mV/pH at room temperature. The weak pH sensitivity of the proposed sensor is not a surprise for TGN whose surface has saturated carbon bonds. For that reason, specific binding of ions is not expected in the ideal case of TGN clean surface. However, there are some defects such as hydroxyl and carbonyl groups along the edges and on the surface, which are induced through the transfer and device fabrication. These defects can react with the protons in the electrolyte, yielding a spurious pH. Therefore, samples treated in this method exhibit increased pH sensitivity [39]. The low gate‐shift of a FET can be further decreased by adding aromatic benzene‐like molecules onto the nanomaterial surface of the sensor. The fluorobenzene molecules are able to suppress the chemical activity of residual free bonds on the as‐grown nanomaterial surface [39]. In this sense, the transfer curves do not shift at all when changing the pH. An appreciable response of nanomaterial‐based ISFET to different pH solutions can be also deduced as the surface is covered with a thin inorganic oxide layer [39]. The surface is functionalised with hydroxyl (OH) groups to render it hydrophilic. In fact, adding OH groups to the nanomaterial surface increases the pH sensitivity. Whereas, a higher sensitivity is deduced for the sensor after having applied a thin Al₂ O₃ coating by atomic layer deposition (ALD). A sensitivity of ~40–50 mV/pH is expected for an ideal Al₂ O₃ layer with a large density of hydroxyl groups [39]. However, the mechanism of sensing remains unclear, as investigated by the large variation of reported pH sensitivities, ranging from a value of 5 mV/pH to 99 mV/pH. In the previous literature, there was a disagreement in the obtained sensitivity of graphene‐based pH sensors. The relatively large values of pH sensitivity reported in the literature can be attributed to the different quality of graphene [39]. A pristine graphene would not respond to pH, whereas a defective one will be suitable for this application. In fact, a defect‐free graphene is the ideal platform for reference electrodes to probe the electrostatic potential in an aqueous electrolyte and is not suited for pH sensors. In other words, the clean graphene does not sense the chemical potential of protons or proton concentration in the solution. It has been demonstrated that the transfer characteristic of as‐prepared single layer graphene shifts weakly when the pH value is changed, and pH sensitivity of 6 ± 1 mV/pH has been measured [39]. Moreover, a pH sensitivity of ~20 mV/pH for suspended graphene FET‐based sensor with reduced noise has been presented [39]. On the other hand, a super‐Nernstian response of 99 mV/pH for single‐, double‐, and triple‐layer graphene‐based sensors has been reported [39]. This superior value can be attributed to the large gate leakage current of the graphene FET‐based pH sensors. Also, a comprehensive study on the effect of the electrolyte composition on liquid‐gated CNT and graphene FETs has been investigated [39]. A pH sensitivity of ~9 and ~12 mV/pH for CNT and graphene in 1M KCl solution has been reported. However, if the concentration of KCl is reduced, a large sensitivity near to 50 mV/pH has been observed [39]. Fig. 7 b shows the time response of the TGN ISFET conductance to successive addition of pH buffers without the gate voltage applied which is investigated based on sensor response time testing theory and modelling [46, 47]. In the atmosphere where the nanomaterial conductance is susceptible to micro‐environment like ambient air flow and wetting status, the conductance is in transition [38]. Once the pH buffer solution is applied, the TGN conductance decreases exponentially. This pH‐dependent conductance without the gate reference electrode demonstrates that the local pH in the proximity of TGNs plays a significant role of the substitutive V_g that is ameliorated with applied external voltage. When measuring equipment, an electric sensor saturates at the time it is measuring a signal of interest, some information can be lost and it cannot be recovered afterwards. However, if the sensor does not saturate, it would be possible to recover even very small signals from large noisy ones by means of post processing techniques [48]. Therefore, it is critical that the sensor has an adequate dynamic range to prevent the saturation under any potential operating condition [48]. Sensor hysteresis, as one of the static characteristics of physical and chemical sensors, reveals the differences in sensor outputs achieved from the same input through the two opposite processes, in which one the input increases and the other one decreases [49]. It is noteworthy that the hysteresis and transfer characteristics occurs based on the movement of hydrogen ions according to the variation of hold time (T_H ) and delay time (T_D ) [50]. As the bias is applied to liquid gate voltage, slow movement of hydrogen ions in the region of liquid gate above the nanomaterial surface is considered as a delay component. Since T_H increases, the time of moving hydrogen ions forwards the channel is increased resulting in increase of drain current. This is because T_H is able to control the effect of slow hydrogen ion movement and the initial condition of the sensor [50]. The more hysteresis gap or difference of drain current between forward sweep and reverse sweep decreases in the hysteresis characteristic as the more T_D increases. Specifically, the time of moving hydrogen ions forwards the channel is increased in the case of forward sweep. Moreover, in the case of reverse sweep, the time of moving hydrogen ions towards the opposite direction of channel is increased [50]. These phenomena are caused by the slow movement of hydrogen ion; therefore, the appropriate time is required to get the better measure result. It can be concluded that the intrinsic properties of rGOs can be exploited for biosensing purposes. RGOs are enormously advantageous for pH sensing due to their defective structures and the remaining functional groups [37]. Any sign of pH response should be taken as a sign of imperfection in the graphene layer.

Fig. 6.

pH‐response of the electrochemical sensors

(a) pH‐responsive drain current (I_D ) for TGN and CNT‐based ISFET sensors tested at fixed V_DS  = −1.0 V and Vg  = −1.5 V, (b) pH‐responsive conductance of TGN and G‐based ISFET sensors tested at V _g‐min

Fig. 7.

Effect of solution pH on the charge‐neutrality point and conductance‐time response

(a) V_CN P‐pH characteristic of the TGN‐based sensor, (b) Conductance–time response curve of the sensor for successive injecting pH

This pH dependency is affected by two factors: the interaction of functional groups at the surface of rGOs, for instance –OH and –COOH groups, with H⁺ ions of the electrolyte causing rise to a change in the charge density of surface, and the change of the Gouy–Chapman diffuse electric double layer resulting electrostatic gating effects [37]. The most of the research efforts in nanomaterial‐based ISFET biosensors focused on the immobilisation of biomolecules by covalent binding to a linker molecule or functional groups of nanomaterial [37]. The findings of this research confirmed that the presented sensor shows a similar trend with experimental results [38, 39]. Nanomaterial‐based ISFET can be useful for chemical and biological sensors, since the ion‐responsive electrical signals are obtained. It is concluded that the present work would be suitable for being employed in different applications of bionanotechnology.

4 Conclusion

In this research, a computational model of a TGN‐based sensor is presented which can be utilised in pH applications. The electrochemical detection of pH value, and adsorption effect on the conductance and current‐voltage (I‐V) characteristics of the TGN sensor are analytically studied using the ISFET characterisation technique. Then, the proposed model is compared with CNT and monolayer graphene pH sensors. The TGN model with high specificity and sensitivity exposes higher current compared to that of CNT counterpart. While the comparative results reveal that the conductance of the model is lower than that of monolayer graphene‐counterpart. Moreover, the findings of this work prove that the conductance and current of the TGN‐based sensor are decreased and V _g‐min is right‐shifted by increasing value of pH. This research emphasised that obtained results are now suitable for being employed in different applications of bionanotechnology such as bio‐nanosensors.