1 Introdcution

The recent COVID-19 pandemic has again brought in forefront the importance of electronic biosensors in detection of biomolecules & biological warfare agents. The electronic biosensors may find applications in real time monitoring of airborne biomolecules in transport systems like aircrafts & metro rail and centrally air conditioned buildings like schools, hospitals, etc. Therefore, the need of the hour is to design and develop fast and accurate biosensors.

Biosensors use chemical reactions to detect the biochemical compounds such as antibodies, biological molecules, nucleic acids, enzymes etc. So they are widely used in many applications such as monitoring of diseases, food analysis, crime detection, environmental field monitoring, and also for the study of biomolecules interaction [1, 2].

Now a days Gallium Nitride (GaN) is emerging as one of the most promising material due to very promising properties. GaN has wide band gap of 3.4 eV and can operate at high temperature (> 300 °C) as compare to silicon, which is being used to fabricate conventional devices. Apart from this they have high electric field, polarization nature, distinctive chemical stability in the biological atmosphere, low level of toxicity to the living cells, and the ability to integration with other electronic devices [3,4,5,6,7]. AlGaN/GaN HEMT shows promising biological/chemical sensing applications due to the presence of 2DEG (Two-dimensional electron gas) at the hetero interface. Any biomolecules can get easily attached to AlGaN barrier layer that varies the surface charges at the AlGaN/GaN interface and as a counter effect channel properties are varied. So interface property [8] is very crucial parameter in determining the device functionality. So HEMT based gas sensors, mercury and chlorine ion sensors, pH sensors, DNA sensors, glucose sensor, cancer biomarker detector, terahertz sensors, strain sensors etc. have been developed [9, 10].

Extensive work has been done by various researchers on FET based biosensors for detection of various biomolecules. Schwarz et al. have reported a novel HEMT‐based hybridization sensor using photochemical functionalization to detect DNA [11]. Khazanskaya et al. have reported field-effect transistor biosensor for ammonium ions and urea detection [12]. Chen et al. used Au-gated AlGaN/GaN HEMTs to detect c-erbB-2, which is a breast cancer marker [13]. Kannan reported GaN HEMT to detect pH and biomarkers for the detection of biomolecules using MOS transistor sensor [14]. Yantao et al. have invented an effective FET biosensor for Ebola Antigen detection [15]. Chung et al. demonstrated AlGaN/GaN HEMT hydrogen gas sensor at extreme environmental conditions [16]. Hu proposed AlGaN/GaN/AlN quantum-well MOS-HEMTs based on numerical simulations [17]. Recently, Seo et al. reported a field-effect transistor (FET)-based biosensing device by using graphene sheet for detecting SARS-CoV-2 in clinical samples [18].

In this paper, Simulation of GaN MOSHEMT based biosensor for detection of neutral biomolecule such as ChOx, protein, streptavidin and Uricase has been demonstrated. Although gate-all-around junctionless transistor (GAA-JLT) [19] and inversion mode MOSFETs [20] have been demonstrated in past for label-free electrochemical detection of neutral biomolecule. As compare to these devices GaN MOSHEMT devices have advantages of reduced gate leakage current density, high breakdown voltage and high electron mobility under both low and high transverse fields [21, 22]. The structure is analyzed by using dielectric modulation technique. Silvaco TCAD software is used for simulation of the structure.

Section 2 presents the device architecture and models used for simulation. Section 3 discusses the device current model used for simulation. Results including output and transfer characteristics, sensitivity and channel potential are discussed in Section 5.

2 Device architecture

Figure 1a highlights the schematic cross-section of AlGaN/GaN MOSHEMT. The structure has 1 µm thick GaN buffer layer, 1 nm AlN spacer, 20 nm oxide, 18 nm Al0.20Ga0.80 N barrier. The gate length (LG) is 5 µm, length between source and gate (LSG) is 1 µm and between gate and drain (LGD) is 2 µm. The structure is passivated by thin film of Si3N4 layer to reduce surface states effect. The complete length and width of the device is 9 µm and 100 µm, respectively. Contact taken at the gate terminal is Schottky and at source & drain is ohmic. Initially the region below the cavity was filled with SiO2 and later two cavities of 1.5 µm each are considered in region I and III underneath the gate electrode as shown in Fig. 1b. This region is confined on the top of AlGaN layer to induce and detect biomolecule. In this paper, neutral biomolecules like ChOx, Protein, Streptavidin and Uricase are introduced in the cavity. The dielectric constant of these biomolecules is given in Table 1.

Fig. 1
figure 1

a Schematic of AlGaN/GaN MOSHEMT with nanocavity for biomolecules. b Magnified view of cavity region introduced below gate region

Full size image
Table 1 Dielectric constant of biomolecules
Full size table

Considered neutral biomolecules do not react with the SiO2 interface. They only introduce the dielectric in the region and the structure is like gate-stack. Therefore, the presence of neutral biomolecules can be modeled by introducing material with a dielectric constant corresponding to neutral biomolecules. This approach has been calibrated with experimental results [23].

All simulations are carried out in Silvaco TCAD software. The models used are polarization model, Shockley–Read–Hall (SRH) recombination mode, field dependent drift velocity (FLDMOB) model, concentration dependent mobility (CONMOB), albrct.n and 8.85 × 1012 cm−2 of interface charges are considered at the AlGaN/GaN interface.

3 Device current model

As shown in Fig. 1b region I and III is empty cavity to be filled with biomolecule and region II is oxide layer. The capacitance for different regions is given by Eqs. (1) and (2)

$$C_{{{\text{SiO}}_{{2}} }}= \frac{{\varepsilon_{{{\text{SiO}}_{{2}} }} }}{{t_{{{\text{ox}}}} }}$$
(1)
$$C_{{{\text{Bio}}}} =\frac{{2\varepsilon_{{{\text{Bio}}}} }}{{t_{{{\text{Bio}}}} }}$$
(2)

where, εSiO2, εBio, tox and tBio are permittivity and thickness of SiO2 and biomolecule, respectively. The thickness of oxide and biomolecule layer is kept constant for simplicity.

The total capacitance of the region I, II and III is given by Eq. (3)

$$C_{{{\text{Total}}}} = C_{{{\text{SiO}}_{{2}} }} + C_{{{\text{Bio}}}}$$
(3)

Total capacitance of the MOSHEMT device is given by Eq. (4)

$$\frac{1}{{C_{{{\text{MOSHEMT}}}} }} = \frac{1}{{C_{{{\text{HEMT}}}} }} + \frac{1}{{C_{{{\text{total}}}} }}$$
(4)

where, CHEMT is the capacitance of the Device [24]. The drain current for the device is given by Eq. (5) Ref.[25],

$$I_{{D = - q\mu \frac{{W_{{\text{G}}} }}{{L_{{\text{G}}} }}\left( {\begin{array}{*{20}c} {\frac{2}{s}\gamma_{0} \left( {n_{{{\text{drain}}}}^{\frac{5}{s}} - n_{{{\text{source}}}}^{\frac{5}{s}} } \right) + v_{{{\text{th}}}} \left( {n_{{{\text{drain}}}} - n_{{{\text{source}}}} } \right)} \\ { + \frac{q}{{2C_{{{\text{MOSHEMT}}}} }}\left( {n_{{{\text{drain}}}}^{2} - n_{{{\text{source}}}}^{2} } \right)} \\ \end{array} } \right)}}$$
(5)

γo = 2.12 × 10–12 V m4/3 is constant. nsource and ndrain are charge concentration at source and drain, respectively. WG is width and LG is gate length. µ is the low field mobility.

4 Results and discussions

Figure 2 shows the Id-Vd characteristics of the GaN based MOSHEMT having gate length of 5 μm without biomolecule. The core current models used in the simulation are able to reproduce the device characteristics very well. The self-heating effects are not considered in the simulation.

Fig. 2
figure 2

Output characteristics of AlGaN/GaN MOSHEMT without biomolecules at different gate voltages

Full size image

The biomolecules has been introduced in cavities of Wcavity = 20 µm and Lcavity = 1.5 µm on both sides of oxide layer. Figure 3a shows the output characteristics of the device. When biomolecule is introduced in to the cavity, there is enhancement in the carrier concentration in the channel region which causes increase in the drain current. Here Maximum increase of 18 mA in drain current is observed for Uricase. gd vs Vd plot is shown in Fig. 3b. Since Id-Vd characteristics of the device are changing by presence of biomolecule in the cavity in similar way variation in gd is also observed.

Fig. 3
figure 3

a Output characteristics b gd vs Vd of AlGaN/GaN MOSHEMT with different biomolecules introduced in the cavity

Full size image

Figure 4a–e shows the electron concentration contour. Figure 4a shows the no change of concentration when no biomolecule is introduced. As biomolecule is introduced, the more charges gets induced in the channel in the region beneath the cavity. The quantity of charge induced depends on the dielectric constant of the biomolecule as can be seen from Fig. 4b–e. Due to lowest value of dielectric constant in Uricase more charges are induced in the channel that is inferred from the rise of drain current in Fig. 3a.

Fig. 4
figure 4

Electron concentration contour (a) without biomolecule (b) ChOx (c) Protein (d) Streptavidin (e) Uricase

Full size image

Transfer characteristics and transconductance are shown in Fig. 5a, b. The positive shift in the threshold voltage is observed when biomolecule is introduced in the cavities. Because the presence of biomolecule induces charges in the channel thus results in the better formation of 2DEG in the channel at lower gate voltage. The Threshold voltage shifted from − 12 V for without biomolecule to − 5.5 V at VDS = 1 V for Uricase. However, with increase of dielectric constant for streptavidin, protein and ChOx negative shift in the voltage is observed as compared to Uricase the due to reduced induced charges. Similarly the change in the transconductance is also observed as shown in Fig. 4b. The maximum transconductance is 2 mA/mm for without biomolecule, 1.5 mA/mm for ChOx, 1.3 mA/mm for Protein, 1.2 mA/mm for streptavidin and 1 mA/mm for Uricase.

Fig. 5
figure 5

a Output characteristics b gd vs Vd of AlGaN/GaN MOSHEMT with different biomolecules introduced in the cavity

Full size image

Figure 6 shows the sensitivity parameter for different biomolecules. The maximum increase of 0.141 in sensitivity is observed when cavity length is varied from 1 to 2 µm on both each side of the oxide layer. This is due to reason, as cavity length increases, more area of the biomolecule is available to interact with underneath AlGaN functional layer. The maximum increase is observed for Uricase due to its lower dielectric constant. So for high sensitivity applications, the length of the cavity should be more.

Fig. 6
figure 6

sensitivity parameter for different biomolecules

Full size image

Figure 7 shows the potential along the channel at 4 µm below the AlGaN/GaN interface when biomolecule is introduced in the cavity region I&III underneath the gate. Permittivity of the biomolecules is different due to which the potential developed in the channel region changes. Uricase having lowest permittivity is showing potential variation of 10% as compared to 3% for ChOx. The comparison in parameter like on current and threshold voltage for MOSHEMT and GAA-JLT [26] for detection of neutral biomolecule is shown in Table 2. A large change in drain current and threshold voltage of MOSHEMT is observed for all biomolecules.

Fig. 7
figure 7

Channel potential for biomolecules along the channel length

Full size image
Table 2 Comparison between MOSHEMT and GAA-JLT [26]
Full size table

5 Conclusion

In this paper, the used model accurately predicts the effect of biomolecule species on the performance parameters of GaN based MOSHEMT. There is significant increase in the drain current and change gd is observed with the addition of biomolecule in the cavity. The electron concentration contour indicates notable increase of electron concentration with addition of biomolecule. The positive shift in the gate threshold voltage, maximum shift of ~ 54% is observed for Uricase due to lowest dielectric constant. Similarly the notable change in transconductance is also obtained. The sensitivity of the structure increases approximately four times with the increase in the cavity length due to more area of contact of the biomolecule with the underneath functional layer. The rise of 10% in channel potential is viewed when biomolecule (Uricase) is introduced in the cavity. MOSHEMT showed better sensitivity towards drain current and threshold voltage as compared to GAA-JLT. Therefore, it can be concluded that this GaN Based MOSHEMT gives significant change in performance parameters and can be used for biomolecule sensing applications effectively.