**3.1. Principle of operation**

then, the semiconductor industry has been on the lookout for novel devices which can effectively address the issues of scaling and depict performance which is superior to MOSFETs.

18 Design, Simulation and Construction of Field Effect Transistors

Over the past few decades, industries and researchers have proposed a number of devices as prominent alternatives to MOSFETs for low power applications. Most of these devices possess principles of operation which are different from MOSFETs. The International Technology Roadmap for Semiconductors in its document 'Beyond CMOS' published in 2015 reported the emerging devices based on structure or materials and charge/non-charge entity [4]. This include a number of devices like nanowire FET [5–7], carbon nanotube FET [8–10], graphene FET [11–13], TFET [14–16], spin FET [17–19] and negative gate capacitance FET [20, 21]. Of these devices, TFETs have gained concentrated focus for low power applications due to their fundamental fabrication methodologies being similar to MOSFETs, and their ability to achieve sub-60 mV/dec subthreshold swing and lower off currents than MOSFETs. TFETs operate by interband tunneling mechanism unlike thermionic emission in MOSFETs due to which the high energy tails of the Fermi distribution of carriers while moving from source to drain get curtailed, resulting in low subthreshold swings and off currents. Different architectures of TFETs have been proposed till date to improve their performance and increase the on currents. [22, 23], nanowire TFET [24, 25], heterojunction TFET [26, 27], III-V TFET [28], triple material gate TFET [29, 30], cylindrical TFET [31] and SOI TFET [32] are some of the widely

TFETs have found their uses in a wide range of low power applications like digital circuits and memory applications [33–35]. However, recently, the emergence of FET-based biosensors has projected TFETs as biosensors based on dielectric modulation in which the dielectric constant along with the charge of the biomolecules in the gate dielectric region affect the drain current [36]. The sensitivity of the biosensor in presence of biomolecules is defined with respect to a reference value. A number of geometries of TFETs has been proposed as dielectric-modulated biosensors, and the analyses of their sensitivities having dependence on device parameters

Section 2 of this chapter presents a brief report on the existing works on FET-based biosensors. In Section 3, the principle of dielectric modulation in TFETs and a reference architecture for TFETs as biosensors are discussed. Section 4 mentions the different physics-based models to be considered while simulating a TFET on a Technology Computer Aided Design (TCAD) tool. The different sensitivities are defined in Section 5. Section 6 mentions the various nonideal conditions that may possibly exist in case of FET-based biosensors. A circular gate TFET is analyzed as a dielectric-modulated biosensor through TCAD simulation in Section 7.

The compactness, compatibility in fabrication and label-free detection have made FET-based biosensors one of the promising area of interests. Generally, there are two methods to detect the presence of biomolecules: gating effect and dielectric modulation. The gating effect uses

Section 8 concludes the chapter and comments on future scope.

used structures.

have been reported [37–41].

**2. A brief survey**

A conventional homojunction TFET is a gated reverse-biased p-i-n structure [3]. As opposed to the thermionic emission in MOSFETs, the mechanism of transport in TFETs is band-to-band tunneling. In an n-TFET, when positive gate bias increases, the energy bands get suppressed as a result of which the width between the p + source valence band, and i-channel conduction bands reduces, thus facilitating the tunneling of electrons from the former to the latter as depicted in **Figure 1** [3]. This contributes to the drain current.

The tunnel barrier at the source-channel junction is modeled as a triangular barrier, and using WKB approximation, its tunneling probability is calculated as [47].

$$T(E) = \exp\left(-\frac{4\lambda\sqrt{2\,m^\*E\_c^{3/2}}}{3q\hbar(\Delta\Phi + E\_c)}\right) \tag{1}$$

where m\* is the effective mass, *EG* is the energy band gap at the source-channel tunnel junction, is the energy overlap of the bands at tunnel junction, *λ* is the screening tunneling length, q is the electronic charge, and ℏ is the reduced Planck's constant. The screening length *λ* is defined as [47].

**Figure 1.** Energy band diagram of a p-i-n TFET showing the on and off states.

$$
\lambda = \sqrt{\left(\varepsilon\_s / \varepsilon\_{ox}\right) t\_s t\_{ox}} \tag{2}
$$

the energy bands at the channel-drain junction get influenced so as to form a tunneling barrier. This results in flow of current by tunneling at this junction. This becomes the reason for anomalies in complementary TFET digital circuits. However, as far as label-free biosensing is concerned, this ambipolar current can be considered as a measure of sensitivity in TFETs.

Dielectric-Modulated TFETs as Label-Free Biosensors http://dx.doi.org/10.5772/intechopen.76000 21

As evident from Section 3.1, the objective of utilizing dielectric modulation for biosensing

• The biomolecules must be immobilized in the gate dielectric region. To realize this, a nanogap is required in the gate dielectric. The gate dielectric, therefore, is composed of two regions; apparently, it works like a dual gate dielectric device, where one gate dielectric has a fixed dielectric constant, and the other carries the dielectric constant of the biomolecules.

• The height of the embedded nanogap must be large enough to allow the entry of biomolecules into the cavity or nanogap. A minimum height of 10–11 nm is usually considered for

• In order to accommodate more biomolecules, a double gate structure with the above de-

• The TFETs must have appropriate source doping. A few biomolecules have dielectric constants of 2 or 3 [39–42], which is closer to the dielectric constant of 1. So, the geometry must be so designed that it is able to respond to immobilization of biomolecules having low

• For TFETs which utilize the ambipolar behavior of the devices, the drain doping concentra-

The most convenient way to analyze a biosensor is on a computational platform where the physics-based models applied to the architecture assist in analyzing its performance. There are a number of industrial simulators and most of them come equipped with provisions for defining a geometry and feeding it through iterations of selective models available in their

The embedded nanogap in a DM TFET is designed by substituting that region in the gate dielectric with a dielectric material (oxide) whose dielectric constant can be altered as per requirement [38]. For charged biomolecules, the charges are considered at the oxide-semiconductor interface. By varying the dielectric constant and the charge, the immobilization of

The models for TFETs must be chosen with care. Since TFETs usually have high source doping concentration to achieve large band bending at equilibrium, therefore, Fermi-Dirac statistics

**4. Simulation strategy for a DM TFET as a label-free biosensor**

requires a modified geometry with the following basic requirements

simulation analyses in nanoscale TFETs.

biomolecules may be mimicked appropriately.

signs may be employed.

dielectric constants as well.

tion is more important.

libraries.

**3.2. Geometry**

The dielectric modulation in a MOS-based device, and particularly TFETs is the alteration in the dielectric constant in the gate dielectric region of the device. Keeping other parameters constant, the effect of dielectric modulation in TFET is best explained by the parameters, *λ* and ΔΦ. A low dielectric constant decreases the gate-to-channel coupling as a result of which the tunnel width is more even at high gate voltages, as compared to the presence of higher gate dielectric constants at the tunnel junction [3, 47, 48]. This results in a low tunneling probability, and hence, low tunneling current. Lesser the gate-to-channel coupling, lesser is the amount of ΔΦ at the tunnel junction.

A 2-D Poisson's equation based model of TFET results in a closed form equation of drain current as [49].

$$I\_{DS} = \left. \mathcal{W} \, T\_{S,\mathcal{off}} \frac{E\_{\mathcal{G}}}{\mathcal{W}\_{t,\text{in}}^2} \frac{A\_{\mathcal{K}}}{B\_{\mathcal{K}}} \exp\left(-\frac{B\_{\mathcal{K}} E\_{\mathcal{G}}}{q \, \mathcal{W}\_{t,\text{in}}}\right) \tag{3}$$

where, *W* is the width of the device, *TSi*,*eff* is the effective Silicon body thickness, *EG* is the band gap of the semiconductor, *Wt,min* is the minimum tunneling width, *AK* and *BK* are material dependent constants. Equation (3) suggests the dependence of drain current on the minimum tunnel width which, in turn, is dependent on the dielectric constant of the gate insulator region near the tunnel junction. Therefore, as dielectric constant of the nanogap increases, *WT*,*min* decreases, and drain current increases.

Another prospect of utilizing dielectric modulation for biosensing is by placing the embedded nanogap towards the channel-drain junction of the TFET, and exploit its ambipolarity [45]. For a conventional n-TFET, when the drain voltage is positive, and gate voltage is negative, the energy bands at the channel-drain junction get influenced so as to form a tunneling barrier. This results in flow of current by tunneling at this junction. This becomes the reason for anomalies in complementary TFET digital circuits. However, as far as label-free biosensing is concerned, this ambipolar current can be considered as a measure of sensitivity in TFETs.
