**5. Sensitivity parameters**

The performance of a TFET or any MOS-based device is reflected through its electrical parameters. The most commonly used parameters for defining sensitivity are the drain current and threshold voltage. Subthreshold swing (SS) may also be considered for defining a sensitivity parameter; however, the measurement of SS is dependent on the orders of the logarithmic scale over which the drain current is measured. This value or the measurement may not be consistent as the drain current changes with the variation in the dielectric constant of the immobilized biomolecules. Sensitivities are usually defined with respect to reference values. The reference value in case of a dielectric-modulated biosensor is considered when the nanogap is devoid of biomolecules, and hence, is assumed to be filled with air with a dielectric constant of 1 without any charges at the oxide-semiconductor interface. The drain current based sensitivity is mathematically expressed as [44].

$$\text{Sensitivity}\_{\prime} \text{S}\_{1} = \left. \frac{I\_{D,k}}{I\_{D,k+1}} \right|\_{V\_{\text{cl}}} \tag{4}$$

**6. Non-idealities in dielectric-modulated biosensors**

**6.1. Steric hindrance**

**6.2. Probe placement**

**6.3. Fabrication issues**

continuous as in steric hindrance.

the tunnel junction is least affected by it.

**7. A circular gate TFET as a DM biosensor**

While simulating a geometry of a DM TFET as on a TCAD tool to assess its biosensing capacity, it is convenient to assume that the nanocavity is completely filled with the biomolecules. However, in practical cases, the issues of steric hindrance and probe placement do not allow the embedded nanogap to be completely filled [39]. As a result, partially filled nanogaps are formed.

Dielectric-Modulated TFETs as Label-Free Biosensors http://dx.doi.org/10.5772/intechopen.76000 23

In case of steric hindrance, the biomolecules which get immobilized first prevent the further entry of biomolecules. In fact, there is a hindrance to the biomolecules which are likely to get immobilized in the nanogaps, resulting in partial hybridization. In order to account for this on a TCAD tool for simulation, different patterns of immobilization inside the nanogap are assumed, like increasing, decreasing, concave and convex profiles of placement [39]. As explained in Section 2, this can be designed by defining different heights of gate dielectric

In order to immobilize the biomolecules, probes or receptors are used in the nanogap. The placement of probes in the nanogap for immobilization of biomolecules may not be continuous throughout, and this may result in partially filled nanogaps [39]. For simulation, this may be considered in a similar manner as the steric hindrance except that the profiles shall not be

The nanogap which is formed in a DM biosensor is carved out by forming a native oxide first, and then etching out the native oxide [38]. This method, however, is challenging, and practically, there is possibility that damages result from the process. These anomalies include creation of trap centers at the interface or incomplete etch of the native oxide along with traps in the residual gate dielectric [38]. The phenomenon of tunneling in a TFET is highly dependent on its source-channel tunnel junction, and the alignment of the gate with the junction. During fabrication, the gate edge may be displaced from the junction. This may result in an overlap or an underlap depending on whether the gate shifts towards the source or the channel respectively. In case of a gate-source overlap, the characteristics of the TFET generally improve as the shifted gate can now influence the energy bands in the source-channel junction with better control. However, in case of an underlap, the gate edge moves away from the junction, and

This section presents a geometry of TFET, a Circular Gate TFET (CG) as a dielectric-modulated biosensor, and discusses some of the results. The CG TFET has a non-uniform gate in the

material mimicking the biomolecules according to the profile of biomolecules.

where *I D*,*k* and *I D*,*k*=1 are the values of drain currents when the nanogap is filled with biomolecules and the nanogap is unfilled. The values must be measured at the same gate voltage so as to get a justified value of sensitivity.

The threshold voltage has a dependence on the dielectric constant of the gate dielectric and charge at the semiconductor-oxide interface. The shift in threshold voltage with the immobilization of biomolecules with different dielectric constants may be taken up as a sensitivity parameter. It is mathematically expressed as

$$\text{Sensitivity}\_{\prime} \mathcal{S}\_{\prime\_{\prime}} = \mathcal{V}\_{\prime,k=1} - \mathcal{V}\_{\prime,k} \tag{5}$$

where *VT*,*<sup>k</sup>* and *VT*,*k*=1 are the threshold voltages for the cases when the nanogap is filled with biomolecules and when the nanogap is completely unfilled.

There are many threshold voltage extraction methods for MOSFETs, and TFETs. Of them, the Linear Extrapolation (LE) Method is the most widely used [51, 52]. According to this extraction principle, the intercept on the gate voltage axis made by the tangent to the drain current curve corresponding to the maximum value of *gm* <sup>=</sup> *<sup>d</sup> <sup>I</sup> <sup>D</sup>*/*<sup>d</sup> VGS* is defined as the threshold voltage. Although this extraction method may result in change in threshold voltage with change in range in gate voltage, however, we have considered a fixed range of gate voltage in all the cases of comparison. So, we have used this method of threshold voltage extraction from simulation transfer characteristics.
