© 2021 IIETA. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
The combination of two techniques: lowbandgap semiconductor and linetunneling structure is an effective way to achieve the highest oncurrent in TFETs. In this paper, design of lowbandgap linetunneling TEFT and its analytical modeling of drain current equation is proposed. The previously suggested drain current equation for the lowbandgap linetunneling TEFT has been explained in a relatively complex form based on the minimum tunnel path that is an effective factor in determining bandtoband tunneling (BTBT). It has been simplified in this paper and reformulated based on gatetosource voltage. Important design factors such as source doping concentration, material and thickness of the gateinsulator were examined by simulation and numerical calculations based on the minimum tunnel path for two lowbandgap In_{0.88}Ga_{0.12}As and relatively highbandgap GaSb semiconductors. The comparison of the results obtained from simulations with the proposed analytical drain current model show a good agreement. Drain doping concentration, is an effective factor on the offstate current of lowbandgap TFET. This factor was examined in order to reduce the offcurrent.
analytical modelling, linetunneling, lowbandgap semiconductor, minimum tunnel path, tunnel fieldeffect transistor (TFET)
Nowadays, the downscaling of conventional MOSFET has led to increase in power consumption [14] that has been called as the power crisis [1, 3]. The mechanism of MOSFET transistors operation does not allow subthreshold swing less than 60 mV/dec [57]. Tunnel fieldeffect transistor (TFET) is defined as a semiconductor device in which the gate controls the sourcedrain current through modulation of BandtoBand Tunneling (BTBT) and BandtoBand Tunneling is a process in which electrons tunnel from the valence band through the semiconductor bandgap to the conduction band [8]. The TFET is suitable for low power consumption applications because of its low subthreshold swing and low offstate leakage current. [913]. However, Sibased TFETs have a low oncurrent due to their highbandgap [1416]. It has been proved that using lowbandgap semiconductors increases the oncurrent (I_{on}) of TFET [1720] although the lowbandgap in these semiconductors, reduces the undesirable tunnel length at the channeldrain side and increases significantly the offcurrent (I_{off}) [16]. Reducing the drain doping concentration has an important role in decreasing the tunneling of the drain side and controlling the offcurrent [16]. In the lowbandgap TFET the effect of drainsource voltage (V_{ds}) controlling is more due to its lowbandgap. When (V_{ds}) goes beyond the bandgap voltage (E_{g}/q), the undesirable tunnel length is significantly decreased. Using gatesource overlap structure or linetunneling structure is another effective way to increase the oncurrent in TFETs [1, 8, 21]. Combining the linetunneling structure and the lowbandgap semiconductor is an effective way to achieve the highest oncurrent in TEFT [22]. The tunnel path is defined as the physical path between two points corresponding to equal energy for the conduction and valence band respectively [8]. Since the minimum tunnel path (l_{min}) which is the minimum length of tunnel path in the depletion region, has a great effect in determining the amount of the linetunneling TFET current, important parameters of design such as: source doping concentration, material and the thickness of gateinsulator were investigated by simulation and numerical calculations based on the minimum tunnel path (l_{min}) for two semiconductors, lowbandgap In_{0.88}Ga_{0.12}As and relatively highbandgap GaSb. The previously proposed current equation for lowbandgap linetunneling TEFT [22] based on l_{min} is relatively complex. It has been simplified in this paper and it has also been reformulated based on the gatetosource voltage (V_{gs}). Then, a comparison has been made between the results obtained from the proposed model and the results obtained from the previous model [22] and the simulation results for several different source doping concentrations, gateinsulator with different materials and thicknesses, which showed a good agreement.
2.1 Structure and parameters
Figure 1 shows a schematic sketch of the line and pointtunneling TEFTs used in this study. Direct bandgap of 0.5 eV In_{0.88}Ga_{0.12}As with work function of 4.7 eV is considered as the lowbandgap semiconductor and direct bandgap of 0.72 eV GaSb with the work function of 4.12 eV is considered as the relatively highbandgap semiconductor. Table 1 shows the parameters of TFETs used in investigation.
The simulations were done using the Silvaco software package and using nonlocal tunneling.
Figure 1. Schematic structures of (a) linetunneling and (b) point tunneling, TFETs
Table 1. Parameters of TFETs
Gatesource overlap length 
80 nm 
Channel length 
50 nm 
Body thickness 
40 nm 
Source doping concentration 
8×10^{18} cm^{3} 
Drain doping concentration 
10^{18 }cm^{3} 
Channel doping concentration 
10^{16} cm^{3} 
Gate insulator material 
HfO_{2} 
Gate insulator thickness 
1 nm 
2.2 Analytical model
The analytical model used in this paper is linetunneling TFET analytical model presented in ref. [22] in which the minimum tunnel path is calculated as an indicator of linetunneling TFETs. In this model minimum tunnel path (l_{min}) and maximum tunnel path (l_{max}), are expressed in Eq. (1) and Eq. (2) as follows
${{l}_{\max }}=\sqrt{2\varepsilon {{E}_{g}}/{{q}^{2}}{{N}_{a}}}$ (1)
where, ε is the dielectric permittivity of the semiconductor, E_{g} is the bandgap energy of the semiconductor, N_{a} is the source doping concentration and q is the electron charge.
${{l}_{\min }}={{l}_{\max }}.(\sqrt{q{{\psi }_{s}}/{{E}_{g}}}\sqrt{q{{\psi }_{s}}/{{E}_{g}}1})$ (2)
where, ψ_{s} is the surface potential at the semiconductorinsulator interface. For small bodyeffect coefficient (γ), can be written as follows
${{\psi }_{s}}={{V}_{gs}}{{V}_{fb}}\gamma \sqrt{{{V}_{gs}}{{V}_{fb}}}$ (3)
where, V_{fb} is the flatband voltage. The drain current of the lowbandgap linetunneling TEFT is written based on l_{min} parameter as follows
$I=(Aq{{N}_{a}}E_{g}^{1/2}{{l}_{\max }}/12\varepsilon ).{{({{l}_{\max }}/{{l}_{\min }})}^{3}}.\exp (BqE_{g}^{1/2}{{L}_{\min }})$ (4)
In this paper, the coefficients $A=q^{2} \sqrt{m_{r}} / 18 \pi \hbar^{2}$ and $B=\pi \sqrt{m_{r}} / 2 \hbar q$ in Eq. (4), for linetunneling TFET based on In_{0.88}Ga_{0.12}As semiconductor are 2.64 × 10^{20} eV^{1/2}.V^{2}.cm^{1}.s^{1} and 15.44×10^{6} V.cm^{1}.eV^{3/2}, respectively. Here m_{r} and ћ are the reduced mass and the reduced Plank’s constant respectively.
2.3 Operation investigation
The lateral tunneling or pointtunneling occurs in parallel with the semiconductorgate insulator surface while vertical tunneling happens perpendicular with the semiconductorgateinsulator surface in the gatesource overlap. Vertical tunneling is also called linetunneling since the vertical tunneling area is similar to a line which becomes important in higher V_{gs} [23]. Figures 2(a) and 2(b) show the simulated I_{ds}V_{gs} characteristics of line and pointtunneling TFETs for In_{0.88}Ga_{0.12}As and GaSb. It can be seen that the linetunneling TEFT in both semiconductors has a higher oncurrent and smaller subthreshold swing than the pointtunneling TFET. To be exact, point subthreshold swings are 8.5 mV/dec and 15.84 mV/dec for In_{0.88}Ga_{0.12}As line and pointtunneling TFETs respectively, and 11.84 mV/dec and 32.75 mV/dec for GaSb line and pointtunneling TFETs respectively.
Figure 2. I_{ds}–V_{gs} characteristics for point and linetunneling TFETs using (a) In_{0.88}Ga_{0.12}As, and (b) GaSb. (c) onstate energybands diagram of In_{0.88}Ga_{0.12}As and GaSb linetunneling TFETs. (d) Calculated onstate minimum tunnel path as function of gate voltage for In_{0.88}Ga_{0.12}As and Gasb linetunneling TFETs
In the linetunneling TFET, as V_{gs} is increasing, charge inversion is formed beneath the gate in the source region and electrons tunnel from the source to the newly inverted source region. When V_{gs} is close to the threshold voltage, this tunneling increases and decreases the subthreshold swing of linetunneling TFET. Furthermore, in the linetunneling TFET, both linetunneling and pointtunneling are done and there are more electrons in the channel; therefore, oncurrent of the linetunneling TFET becomes higher than pointtunneling TFET. Figure 2(a) and 2(b) also show that the linetunneling TEFT based on In_{0.88}Ga_{0.12}As lowbandgap semiconductor has higher oncurrent and smaller subthreshold swing than the linetunneling TFET based on GaSb highbandgap semiconductor. Figure 2(c) shows the onstate energybands diagram of the In_{0.88}Ga_{0.12}As and GaSb TEFTs. As it is seen in these energybands diagram, the In_{0.88}Ga_{0.12}As TEFT has a smaller tunnel length than the GaSb highbandgap TEFT. The smaller tunnel length results in more BTBT generation and consequently increase in the oncurrent.
Figure 3. I_{ds}–V_{gs} characteristics for linetunneling TEFTs using (a) In_{0.88}Ga_{0.12}As and (b) GaSb, for various drainsource voltages. (c) offstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TFETs for various drainsource voltages. (d) onstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TFETs for various gatesource voltages
Figure 2(d) shows the calculation of l_{min} as function of V_{gs} for two semiconductors In_{0.88}Ga_{0.12}As and GaSb. As it is observed, l_{min} is strongly dependent on E_{g}. The In_{0.88}Ga_{0.12}As linetunneling TEFT, has smaller l_{min} than GaSb. This issue causes sooner initiation of linetunneling that results in the improvement of subthreshold swing and higher oncurrent of the In_{0.88}Ga_{0.12}As linetunneling TEFT compared with GaSb linetunneling TEFT. Figures 3(a) and 3(b) show the dependence of the I_{ds}–V_{gs} characteristic of the simulated linetunneling TEFTs for different values of V_{ds} for two semiconductors. As V_{ds} goes higher than bandgap voltage (E_{g}/q), undesirable tunneling at channeldrain side increases because the tunnel length at the channeldrain side becomes thinner and it also increases offcurrent and subthreshold swing while, the increase of oncurrent is negligible. Figure 4(c) shows offstate energybands diagram for In_{0.88}Ga_{0.12}As TFET for different values of V_{ds}. In this figure, the narrowing of channeldrain side tunnel length is well seen. Moreover, increase in V_{ds} also increases the diode leakage current and helps increase the offcurrent. In order to avoid increase in the offcurrent and also improve the subthreshold swing, V_{ds} for both semiconductors is assumed to be equal to the bandgap voltage (E_{g}/q) of that semiconductor. Figure 2(d) shows that the variations of l_{min} with V_{gs} are exponential for both semiconductors and this shows the extreme dependence of l_{min} to V_{gs}. The decrease in l_{min} with increase in V_{gs} leads to increase in the oncurrent and improvement of the subthreshold swing. However, l_{min} is saturated and oncurrent almost remains stable for high V_{gs}. Figure 3(d) shows the onstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TEFT for different values of V_{gs}. Increase in V_{gs} results in a reduction of the tunnel length and it causes an increase in BTBT generation and consequently an increase in the oncurrent and improvement in the subthreshold swing. However, lowering of tunnel length is saturated and oncurrent almost remains constant for high V_{gs}. The value of V_{gs} was considered to be 1.5 times that of the bandgap voltage (E_{g}/q) of that semiconductor in order to assure the high oncurrent and desirable subthreshold swing.
3.1 Source doping concentration
Figures 4(a) and 4(b) show the dependence of the linetunneling TEFT current on the source doping concentration for two In_{0.88}Ga_{0.12}As and GaSb semiconductors. As the source concentration increases, the offcurrent decreases due to increased recombination of the carriers participating in the offcurrent; In addition, the oncurrent increases and the subthreshold swing decreases. However, it can be seen that for heavy doping concentrations, the subthreshold swing increases again. The increase in the source doping concentration decreases the pointtunneling onset voltage while it increases the linetunneling onset voltage and also the oncurrent of linetunneling [23]. For high source doping concentrations when V_{gs}becomes near the threshold voltage, vertical tunneling is more effective in determining the subthreshold swing conditions. However, for heavy source doping concentrations, the linetunneling onset voltage increases and lateral tunneling becomes dominant in determining the subthreshold swing conditions. Using a source with heavy doping concentration provides the considered on and offcurrent; but increases the subthreshold swing. Figure 4(c) shows the onstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TFET. It can be seen that increasing the source doping concentration causes decrease of the tunnel length that results in an increase in the oncurrent. However, in heavy doping concentrations the decrease in the tunnel length is saturated and the tunneling generation and consequently the oncurrent stays approximately constant. Figure 4(d) shows the calculated l_{min} for different values of the source doping concentration for two semiconductors. With increase in the source doping concentration, l_{min} decreases and becomes saturated in heavy doping concentrations and the oncurrent stays approximately constant. Thus, the source doping concentration has been considered to 8×10^{18} cm^{3} in order to improve the conditions of subthreshold swing.
Figure 4. I_{ds}–V_{gs} characteristics for linetunneling TEFTs using (a) In_{0.88}Ga_{0.12}As and (b) GaSb, with various source doping concentrations. (c) onstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TFETs for various source doping concentrations. (d) Calculated onstate minimum tunnel path against source doping concentration for In_{0.88}Ga_{0.12}As and GaSb linetunneling TFETs
3.2 Gateinsulator material
Figures 5(a) and 5(b) show the I_{ds}V_{gs} characteristics for three different gateinsulators of Si_{3}N_{4}, Al_{2}O_{3} and HfO_{2}, for two In_{0.88}Ga_{0.12}As and GaSb semiconductors. As it can be seen, the gateinsulator material does not affect the offcurrent but with increasing the dielectric constant, the oncurrent and the subthreshold swing improve. Figure 5(c) shows the onstate energybands diagram of the linetunneling TEFT for two different gateinsulators of HfO_{2} and Si_{3}N_{4}. Using the gateinsulator with the higher dielectric constant causes a more powerful capacitor and more bending of energybands; thus tunneling begins at a lower voltage and the oncurrent increases and the subthreshold swing decreases. Calculating the minimum tunnel path in Figure 5(d) shows that using the gateinsulator with higher dielectric constant decreases the l_{min} in lower V_{gs} and causes the tunneling to begin in a lower voltage. The smaller l_{min} at higher dielectric constants justifies the higher oncurrent and the lower subthreshold swing of TEFTs.
Figure 5. I_{ds}–V_{g}_{s} characteristics for linetunneling TFETs using (a) In_{0.88}Ga_{0.12}As, and (b) GaSb, for different gateinsulators. (c) onstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TFETs for two different gateinsulators. (d) Calculated onstate minimum tunnel path as function of gate voltage for In_{0.88}Ga_{0.12}As and GaSb linetunneling TFETs for different gateinsulators
3.3 Gateinsulator thickness
Changing the thickness of the gateinsulator does not have much effect on the offcurrent. However, as it is seen in Figure 6(a) and 6(b) that using a thinner gateinsulator has improved the oncurrent and the subthreshold swing. Figure 6(c) shows the onstate energybands diagram of the In_{0.88}Ga_{0.12}As linetunneling TEFT for two different gateinsulator thicknesses. For a thin gateinsulator thickness, the bending of the energybands has increased because of the higher gate capacitor. Also, l_{min} increases linearly with the increase in gateinsulator thickness in the calculations done as shown in Figure 6(d). However, the reduction of the gateinsulator thickness is a weak function in reducing the minimum tunnel path; while a thin gateinsulator initiates tunneling at a small V_{gs} and improves the oncurrent and the subthreshold swing.
Figure 6. I_{ds}–V_{gs} characteristics for linetunneling TFETs using (a) In_{0.88}Ga_{0.12}As, and (b) GaSb, for various Gateinsulator thicknesses. (c) onstate energybands diagram of In_{0.88}Ga_{0.12}As linetunneling TFETs for various Gateinsulator thicknesses. (d) Calculated onstate minimum tunnel path against Gateinsulator thickness for In_{0.88}Ga_{0.12}As and GaSb linetunneling TFETs
Since lowbandgap linetunneling TFET current of Eq. (4) is complicated, in this section we simplify the current equation for In_{0.88}Ga_{0.12}As lowbandgap semiconductor by appropriate approximations, without losing accuracy. By inserting Eq. (2) in the preexponential term (l_{max}/l_{min})^{3} of Eq. (4) we have the following:
${{({{l}_{\max }}/{{l}_{\min }})}^{3}}={{[1/(\sqrt{q{{\psi }_{s}}/{{E}_{g}}}\sqrt{q{{\psi }_{s}}/{{E}_{g}}1})]}^{3}}$ (5)
If we multiply $1 /\left(\sqrt{q \psi_{s} / E_{g}}\sqrt{q \psi_{s} / E_{g}1}\right)$ in Eq. (5) by $\left(\sqrt{q \psi_{s} / E_{g}}+\sqrt{q \psi_{s} / E_{g}1}\right) /\left(\sqrt{q \psi_{s} / E_{g}}+\sqrt{q \psi_{s} / E_{g}1}\right)$, after approximation and simplification we have the following:
${{({{l}_{\max }}/{{l}_{\min }})}^{3}}\cong {{(2q{{\psi }_{s}}/{{E}_{g}}+1/9)}^{2}}4.16$ (6)
On the other hand, by inserting Eq. (2) in the exponential term in Eq. (4) and multiplying exponential argument by: $\left(\sqrt{q \psi_{s} / E_{g}}+\sqrt{q \psi_{s} / E_{g}1}\right) /\left(\sqrt{q \psi_{s} / E_{g}}+\sqrt{q \psi_{s} / E_{g}1}\right)$, we have,
$\exp (BqE_{g}^{1/2}{{l}_{\min }})=\exp [BqE_{g}^{1/2}{{l}_{\max }}\sqrt{{{E}_{g}}/q}/(\sqrt{{{\psi }_{s}}}+\sqrt{{{\psi }_{s}}{{E}_{g}}/q})]$ (7)
Since the fraction term in Eq. (7) is rather complicated, it can be approximated as:
$\exp (BqE_{g}^{1/2}{{l}_{\min }})\cong \exp (B{{q}^{1/2}}{{E}_{g}}{{l}_{\max }}/2\sqrt{{{\psi }_{s}}{{E}_{g}}/2q})$ (8)
If $k_{1}=A q N_{a} E_{g}^{1 / 2} l_{\max } / 12 \varepsilon$ and $k_{2}=B q^{1 / 2} E_{g} l_{\max } / 2$ by inserting Eq. (6) and Eq. (8) in Eq. (4), we have,
$\begin{aligned} I \cong & k_{1}\left[\left(2 q \psi_{s} / E_{g}+1 / 9\right)^{2}4.16\right] \\ & . \exp \left(k_{2} / \sqrt{\psi_{s}E_{g} / 2 q}\right) \end{aligned}$ (9)
By inserting Eq. (3) in Eq. (9), we have,
$\begin{aligned} I \cong & k_{1}\left[\left(2 q\left(V_{g s}V_{f b}\gamma \sqrt{V_{g s}V_{f b}}\right) / E_{g}+1 / 9\right)^{2}4.16\right] \\ & \left.. \exp \left[k_{2} /\left(V_{g s}V_{f b}\gamma \sqrt{V_{g s}V_{f b}}\right)E_{g} / 2 q\right)^{1 / 2}\right] \end{aligned}$ (10)
Figure 7. The I_{ds}–V_{gs} characteristics for In_{0.88}Ga_{0.12}As linetunneling TEFTs of equation (4), and (10) and results of simulation for (a) the main parameters of the paper, (b) three different values of source concentration, (c) three different gateinsulators, and (d) three different thicknesses of gateinsulators
In Figure 7 that shows a comparison between the drain current from Eq. (4) and the drain current from simplified Eq. (10) and the results of drain current obtained from the simulation for various different source doping concentrations and the gateinsulators with different materials and thicknesses indicates a good agreement. This indicates that the approximations made are correct.
5.1 Drain doping concentration
Figure 8. I_{ds}–V_{gs} characteristics for linetunneling TEFTs using (a) In_{0.88}Ga_{0.12}As and (b) GaSb, for various drain doping concentrations. (c) offstate energybands diagram of In_{0.88}Ga_{0.12}As and GaSb linetunneling TFETs. (d) offstate energybands diagram of In_{0.88}Ga_{0.12}As TFETs for various drain concentrations
The drain doping concentration does not have any effect on the TEFT’s oncurrent. Controlling the drain doping concentration plays an important role in controlling the offcurrent in TEFTs [16]. Figures 8(a) and 8(b) show the I_{ds}–V_{gs} characteristics of the linetunneling TEFTs for different values of drain doping concentration for two semiconductors. It can be seen that with increase in the drain doping concentration, the offcurrent of linetunneling TEFT and its subthreshold swing have increased while the oncurrent is fixed. The TEFT energybands diagram in the offstate have been drawn for two In_{0.88}Ga_{0.12}As and GaSb semiconductors in Figure 8(c). It can be seen that in the In_{0.88}Ga_{0.12}As line tunneling TEFT, the tunnel length of the channeldrain side is smaller; that results in the undesirable tunneling of the channeldrain side being more and this issue justifies its higher offcurrent. Figure 8(d) shows the TEFT energybands diagram of In_{0.88}Ga_{0.12}As in the offstate for different amounts of drain concentration. For high doping concentrations, the tunnel length of the channeldrain side decreases that results in an increase in the undesirable tunneling of electrons from the drain side and increases the offcurrent considerably. Increase in the offcurrent also increases the subthreshold swing. Using the drain with lower doping concentration decreases the offcurrent and improves the subthreshold swing. Although we must consider that in very low doping concentrations, a good ohmic connection is not established well on the drain side [3].
It was shown that using the lowbandgap linetunneling TFET increases the oncurrent and improves the subthreshold swing. Important design factors such as source doping concentration, the material and thickness of the gateinsulator were considered by simulation and numeral calculations based on the minimum tunnel path. The factor that affects the offcurrent of the linetunneling TFET is the drain doping concentration and its effect on the offcurrent was specified. The drain current equation of lowbandgap linetunneling TFET was reformulated in a simpler form based on the gatesource voltage and it was shown that simplified equation is an agreement with the proposed drain current equation and simulation results.
A 
Parameter of Kane’s model, eV^{1/2}.V^{2}.cm^{1}.s^{1} 
B 
Parameter of Kane’s model, V.cm^{1}.eV^{3/2} 
E_{g} ћ I_{off} 
Bandgap energy of the semiconductor, eV Reduced Plank’s constant, eV.s Off current of TFET, A/µm or A 
I_{on} l_{max} l_{min} m_{r} N_{a} N_{p} q V_{ds} V_{fb} V_{gs} 
On current of TFET, A/µm or A Maximum tunnel path, m Minimum tunnel path, m Reduced mass, Kg Source doping concentration, cm^{3} Drain doping concentration, cm^{3} Electron charge, C Draintosource voltage, V Flatband voltage, V Gatetosource voltage, V 
Greek symbols 

$\gamma$ 
Bodyeffect coefficient, $\sqrt{V}$ 
$\varepsilon$ 
Dielectric permittivity of materials 
ψ_{s} 
surface potential at the semiconductorinsulator interface, V 
[1] Hu, C. (2008). Green transistor as a solution to the IC power crisis. Proceeding of 9th IEEE International Conference on SolidState and IntegratedCircuit Technology, Beijing, pp. 1620. https://doi.org/10.1109/ICSICT.2008.4735116
[2] Ionescu, A.M., Riel, H. (2011). Tunnel fieldeffect transistors as energyefficient electronic switches. Nature, 479(7373): 329337. https://doi.org/10.1038/nature10679
[3] Boucart, K. (2010). Simulation of doublegate silicon tunnel FETs with a highk gate dielectric (No. THESIS). EPFL. https://doi.org/10.5075/epflthesis4729
[4] Zhao, H., Chen, Y., Wang, Y., Zhou, F., Xue, F., Lee, J. (2011). InGaAs tunneling fieldeffecttransistors with atomiclayerdeposited gate oxides. IEEE Transactions on Electron Devices, 58(9): 29902995. https://doi.org/10.1109/TED.2011.2159385
[5] Choi, W.Y., Park, B.G., Lee, J.D., Liu, T.J.K. (2007). Tunneling fieldeffect transistors (TFETs) with subthreshold swing (SS) less than 60 mV/dec. IEEE Electron Device Letters, 28(8): 743745. https://doi.org/10.1109/LED.2007.901273
[6] Koester, S., Lauer, I., Majumdar, A., Cai, J., Sleight, J., Bedell, S., Koswatta, S. (2010). Are Si/SiGe tunneling fieldEffect transistors a good idea? ECS Transactions, 33(6): 357361. https://doi.org/10.1149/1.3487566
[7] Guo, P., Yang, Y., Cheng, Y., Han, G., Pan, J., Ivana, Yeo, Y.C. (2013). Tunneling fieldeffect transistor with Ge/In0. 53Ga0. 47As heterostructure as tunneling junction. Journal of Applied Physics, 113(9): 094502. https://doi.org/10.1063/1.4794010
[8] Vandenberghe, W.G., Verhulst, A.S., Groeseneken, G., Soree, B., Magnus, W. (2008). Analytical model for a tunnel fieldeffect transistor. Melecon 2008. Proceedings of the IEEE 14th Mediterranean Electrotechnical conference. Ajaccio, France, pp. 923928. https://doi.org/10.1109/MELCON.2008.4618555
[9] Wang, P.F., Hilsenbeck, K., Nirschl, T., Oswald, M., Stepper, C., Weis, M., Hansch, W. (2004). Complementary tunneling transistor for low power application. SolidState Electronics, 48(12): 22812286. https://doi.org/10.1016/j.sse.2004.04.006
[10] Bhuwalka, K.K., Schulze, J., Eisele, I. (2004). Performance enhancement of vertical tunnel fieldeffect transistor with SiGe in the δp+ layer. Japanese Journal of Applied Physics, 43(7R): 4073. https://doi.org/10.1143/JJAP.43.4073
[11] Seabaugh, A.C., Zhang, Q. (2010). Lowvoltage tunnel transistors for beyond CMOS logic. Proceedings of the IEEE, 98(12): 20952110. https://doi.org/10.1109/JPROC.2010.2070470
[12] Ghosh, B., Akram, M.W. (2013). Junctionless tunnel field effect transistor. IEEE Electron Device Letters, 34(5): 584586. https://doi.org/10.1109/LED.2013.2253752
[13] Ionescu, A.M., De Michielis, L., Dagtekin, N., Salvatore, G., Cao, J., Rusu, A., Bartsch, S. (2011). Ultra low power: Emerging devices and their benefits for integrated circuits. IEDM 2011. Proceeding of the IEEE International Electron Devices Meeting, Washington DC, USA, pp. 16.1.116.1.4. https://doi.org/10.1109/IEDM.2011.6131563
[14] Nayfeh, O.M., ChlÉirighChleirigh, C.N., Hoyt, J.L., Antoniadis, D.A. (2008). Measurement of enhanced gatecontrolled bandtoband tunneling in highly strained silicongermanium diodes. IEEE Electron Device Letters, 29(5): 468470. https://doi.org/10.1109/LED.2008.920280
[15] Kao, K.H., Verhulst, A.S., Vandenberghe, W.G., Soree, B., Groeseneken, G., De Meyer, K. (2012). Direct and indirect bandtoband tunneling in germaniumbased TFETs. IEEE Transactions on Electron Devices, 59(2): 292301. https://doi.org/10.1109/TED.2011.2175228
[16] Toh, E.H., Wang, G.H., Samudra, G., Yeo, Y.C. (2008). Device physics and design of germanium tunneling fieldeffect transistor with source and drain engineering for low power and high performance applications. Journal of Applied Physics, 103(10): 104504. https://doi.org/10.1063/1.2924413
[17] Mayer, F., Le Royer, C., Damlencourt, J.F., Romanjek, K., Andrieu, F., Tabone, C., Deleonibus, S. (2008). Impact of SOI, Si1xGexOI and GeOI substrates on CMOS compatible tunnel FET performance. IEDM 2008. Proceeding of IEEE International Conference in Electron Devices Meeting. San Francisco, CA, USA, pp. 15. https://doi.org/10.1109/IEDM.2008.4796641
[18] Nayfeh, O.M., Hoyt, J.L., Antoniadis, D.A. (2009). StrainedSi1xGex/Si bandtoband tunneling transistors: Impact of tunneljunction germanium composition and doping concentration on switching behavior. IEEE Transactions on Electron Devices, 56(10): 22642269. https://doi.org/10.1109/TED.2009.2028055
[19] Shih, C.H., Chien, N.D. (2011). Sub10nm tunnel fieldeffect transistor with graded Si/Ge heterojunction. IEEE Electron Device Letters, 32(11): 14981500. https://doi.org/10.1109/LED.2011.2164512
[20] Toh, E.H., Wang, G.H., Chan, L., Sylvester, D., Heng, C.H., Samudra, G.S., Yeo, Y.C. (2008). Device design and scalability of a doublegate tunneling fieldeffect transistor with silicon–germanium source. Japanese Journal of Applied Physics, 47(4S): 2593. https://doi.org/10.1143/JJAP.47.2593
[21] Kim, S.H. (2012). Germaniumsource tunnel field effect transistors for ultralow power digital logic. Ph.D. dissertation. University of California, Berkeley, USA.
[22] Shih, C.H., Chien, N.D. (2014). Design and modeling of linetunneling fieldeffect transistors using lowbandgap semiconductors. IEEE Transactions on Electron Devices, 61(6): 19071913. https://doi.org/10.1109/TED.2014.2316217
[23] Vandenberghe, W.G., Verhulst, A.S., Groeseneken, G., Soree, B., Magnus, W. (2008). Analytical model for point and line tunneling in a tunnel fieldeffect transistor. 2008 International Conference on Simulation of Semiconductor Processes and Devices, Hakone, Japan, pp. 137140. https://doi.org/10.1109/SISPAD.2008.4648256