Access Official Technical PDF

Share Summary:LinkedInEmail

A III-V TFET heterojunction design seeks to address a central challenge in tunneling field-effect transistors: producing sufficient on-state current while limiting off-state leakage. A 2020 device study examined a structure that combines heterojunctions at both the source and channel with III-V semiconductor materials. Two-dimensional simulations indicated that this arrangement can shorten the tunneling path and improve modeled current characteristics compared with a conventional III-V TFET.

These results are relevant to engineers evaluating alternative transistor architectures, but they should be read in context: they are simulation results for a specific device structure, not evidence of production-level performance. The study also shows why gate-length scaling must be considered alongside headline on-current and subthreshold-swing values.

How a TFET controls current

A tunneling field-effect transistor is a gated p-i-n device operated in reverse bias. Rather than relying on conventional carrier injection over a barrier, it uses the gate voltage to control band-to-band tunneling (BTBT) between the source and channel. When the device is off, the source valence band and channel conduction band are separated by a substantial energy barrier and tunneling distance, restricting current. Applying gate voltage shifts the channel energy bands, opening a tunneling window through which electrons can move from the source valence band into the channel conduction band.

Effective tunneling depends on several conditions working together: suitable energy-band alignment, a strong electric field at the source-channel interface to reduce the barrier, and sufficient overlap between electron and hole wave functions. Material selection and junction geometry therefore influence the probability of tunneling, and consequently the transfer characteristics of the device.

Why combine source and channel heterojunctions?

Silicon TFETs can face low on-current compared with MOSFETs, while III-V materials offer high carrier mobility as an alternative material platform. Separately, heterojunctions can be used to engineer the tunnel junction and reduce the barrier height or tunneling distance. The proposed structure combines these approaches: III-V materials provide the device platform, while heterojunctions at both the source and channel target the tunneling bottleneck.

In the simulated design, the channel includes InAs and InAsAlSb layers, while the source includes AlGaSb and GaSb layers. These form the channel and source heterojunctions, respectively. The device also uses an InAsAlSb drain. The modeled structure has a 30 nm channel length, a 10 nm body thickness, a 10 nm channel width and a 2 nm gate oxide; ZrO2 serves as the gate insulator, and aluminum forms the gate electrode.

According to the study, the two added heterojunctions reduce the tunnel barrier and distance. The material properties are also described as supporting higher tunneling probability. The proposed structure exhibits two transmission resonances, which the study identifies as another factor contributing to improved on-current relative to the conventional III-V TFET used for comparison.

What the simulated performance indicates

For a 30 nm gate length and 1 V drain bias, the reported approximate values for the heterojunction TFET are:

  • On-current: 10-3 A/µm
  • Off-current: 10-13 A/µm
  • On-current/off-current ratio: 1010
  • Subthreshold swing: 30 mV/decade

The study reports that on-current is about one decade higher than in the conventional III-V TFET comparison, while off-current remains around 10-13 A/µm. It attributes the improved current and subthreshold swing to increased tunneling probability. Transconductance at high gate voltage is also reported to be 450% higher than in the comparison device, consistent with the higher on-current in the simulations.

Subthreshold swing describes the gate-voltage change needed to alter drain current by a decade over the defined measurement range. The study calculates an average value between off-state and threshold current, with threshold voltage defined using a constant-current method. This definition matters when interpreting or comparing reported swing values, since measurement conventions affect the result.

Gate-length scaling brings a trade-off

The simulations also varied gate length while keeping the other modeled factors similar. On-current increased as gate length decreased from 90 nm to 15 nm. However, off-current was more strongly affected by scaling. The study links the rise in leakage to short-channel effects, including drain-induced barrier lowering (DIBL) and charge sharing.

As off-current rises, the on/off-current ratio falls and subthreshold swing degrades. This is an important design consideration: improving the on-state through a shorter tunneling path does not remove the need to control leakage as device dimensions shrink. Evaluating a TFET therefore requires reviewing the full transfer behavior across gate lengths, not just the best reported on-current or swing.

Implications for device evaluation

The proposed III-V TFET heterojunction structure illustrates how band engineering can be used to target tunneling efficiency. Its simulated results suggest that combining source and channel heterojunctions may improve on-current and switching characteristics relative to the conventional III-V TFET considered in the study. At the same time, the gate-length analysis underscores the sensitivity of off-current, on/off ratio and subthreshold swing to scaling.

For technical decision-makers, the practical takeaway is to treat these figures as a device-design result requiring further evaluation, not as a standalone indicator of manufacturability or system-level suitability. Comparisons should account for bias conditions, device dimensions, current definitions and short-channel behavior. In this design space, the central challenge is balancing tunneling efficiency with leakage control as the geometry is scaled.

Access Official Technical PDF

About The Author