How Do CNTFETs Work, and Why Are They So Promising?
Edis Osmanbasic posted on July 17, 2020 |
What to know about carbon nanotube field-effect transistors.
The structure of a carbon nanotube field-effect transistor (CNTFET). Source: Arvind R. Singh, Shandong University; Reference [1].
The structure of a carbon nanotube field-effect transistor (CNTFET). (Source: Arvind R. Singh, Shandong University; Reference [1].)

New technologies require faster processors, smaller integrated circuits, and less power consumption. Technology advancements such as 5G networks increase the pressure to improve smartphone battery life, spectral efficiency, and more. One potential solution is the use of carbon nanotube field-effect transistors (CNTFETs).

A CNTFET is a nano-scaled device that can provide low-power integrated circuits with high performance and high power density. Instead of the bulk silicon material used in traditional metal-oxide semiconductor field-effect transistors (MOSFETs), CNTFETs use carbon nanotubes (CNTs) in between the source and the drain of a MOSFET structure. This enables higher current carrier mobility, enabling CNTFETs to provide a superior drive current density.

The first simple CNTFET, reported in 1998, was manufactured by depositing single-wall CNTs from solution onto oxidized silicon wafers. The CNTs were synthesized by laser ablation and Si wafers were prepatterned with gold or platinum electrodes. Over time, the process has improved. Previously, CNTs were laid down on the weak contacts of source and drain electrodes. Now, the improved process patterns the electrodes on top of previously laid CNTs.

The contact between metal and nanotubes can be improved by using gold, titanium and carbon with a thermal annealing step. The thermal processing leads to the formation of titanium carbide (TiC) at the metal/nanotube interface, significantly reducing the contact resistance from several megaohms to approximately 30kΩ.

P-Type, N-Type and Ambipolar CNTFETs

Previously, all CNTFETs were p-type (conducting positive charge carriers) because contact doping technology by the adsorption of oxygen from the atmosphere was not well understood. Later, n-type CNTFETs (conducting electrons) were developed by promoting electron conduction when CNTFETs were annealed in a vacuum. Atmospheric oxygen near the metal and nanotube contacts affects the local bending of the conduction and valence bands in the nanotube via charge transfer. The Fermi level is also near the valence band, which makes injection of holes easier. Oxygen desorption at high temperature adapts the Fermi level near the conduction band, allowing the injection of electrons. By using thermal annealing, there is no threshold voltage shift when making n-type from p-type (which is not the case during a bulk doping process).

A back gated n-type nanotube transistor can be achieved by doping the CNT with potassium vapor (see below). The process can shift the Fermi level of the tube from the valence band edge to the conduction band edge by transferring the electrons from adsorbed potassium atoms to the nanotube, thus reverting the doping from p- to n-type. An intermediate state where both electrons and holes are allowed can also be achieved, resulting in ambipolar conduction and the creation of ambipolar CNTFETs.

Schematic diagram of the potassium doping setup.
Schematic diagram of the potassium doping setup.

The capability to make n-type CNTFETs is important because it enables the manufacturing of CNT-based complementary logic circuits.

Main Features of CNTFETs

Like MOSFETs, CNTFETs have three terminals: source, gate and drain. When the gate is on, the current transmits from the source to the drain through a semiconducting carbon nanotube channel. The segment between the drain/source and the gate is heavily doped to provide low resistance. CNTFETs have very promising I-V and transfer characteristics.

The main features of CNTFETs include:

  • Low pulverization probability because the carbon nanotubes are one dimensional. This feature allows a device to operate in the ballistic regime.
  • The CNT conducts on its surface where all the chemical bonds are saturated and stable. There is therefore no need for careful passivation of the interface between the nanotube channel and the gate dielectric.
  • The metal-nanotube contact Schottky barrier represents the active switching element.

Types of CNTFETs

CNTFETs can be classified according to different criteria. When classified by current injection methods, there are two CNTFET types: Schottky barrier CNTFETs (SB-CNTFETs) that use metallic electrodes to form Schottky contacts, and CNTFETs with doped CNT electrodes that form Ohmic contacts (similar to the MOSFET design). The contact type determines the current transport mechanism and CNTFET output characteristics. In SB-CNTFETs, the current means tunneling of electrons and holes from the potential barriers at the source and drain junctions. The barrier width is controlled by the gate voltage, which thus controls the current.

The Ohmic contact CNTFET type uses the n-doped CNT as the contact. The doped source and drain regions behave just like MOSFETs. The potential barrier is formed at the middle of the channel, and the current is controlled via modulation of the barrier height (controlling the gate voltage).


CNTFETs can be fabricated as a single-wall CNT (SWCNT) channel between two electrodes, a multi-wall CNT (MWCNT), or a coaxial CNTFET. MWCNT CNTFETs have a complex structure, which limits their potential. The shells can interact with each other. In addition, only the outer shell effectively contributes to electrical transport. In coaxial geometry, the gate contact wraps all around the channel (CNT), thus providing better electrostatics and very good control of carrier transport. Metal-CNT contact type plays a crucial role in the transistor output characteristics.

There are four typical CNTFET designs: back gate CNTFETs, top gate CNTFETs, wrap-around gate CNTFETs, and suspended CNTFETs.

Back gate CNTFETs are the earliest design that uses prepatterning parallel metal strips across a silicon dioxide substrate and SWCNT arranged on top. CNTs together with metal strips (one metal strip source contact and one drain contact) create a rudimentary field-effect transistor. The silicon oxide substrate presents the gate and includes a metal contact on the back. The metal electrodes are made of metals compatible with silicon, such as titanium (Ti) or cobalt (Co). Since the side-bonding configuration has the weak van der Waals coupling of the devices to the noble metal electrodes, this CNTFET type has high contact resistance (≥1 MΩ).  

Side view of a CNT arranged on a silicon oxide substrate prepatterned with source and drain contacts.
Side view of a CNT arranged on a silicon oxide substrate prepatterned with source and drain contacts.

Top gate CNTFET design requires a more advanced fabrication process compared to the back gate design. SWCNTs are arranged onto a silicon oxide substrate. Each CNT is located and isolated by using an atomic force or scanning electron microscope. Then, high-resolution electron-beam lithography is used to pattern source and drain contacts. The lower contact resistance is achieved via a high temperature anneal step in which adhesion between the contacts and the CNT is improved. After this step, a thin top gate dielectric is deposited on top of the nanotube using evaporation or atomic layer deposition. The final step is placing the top gate contact on the gate dielectric.

 The top gate CNFET with a P++ Si wafer substrate.
The top gate CNFET with a P++ Si wafer substrate.

The main difference between the top and back gate designs is the fabrication process. In the case of the top gate design, the CNTFET arrays on the same wafer because the gate contacts are electrically isolated from each other. A higher electrical field with a lower gate voltage can be achieved in the top gate design due to the thin gate dielectric. Because of those features, top gate CNTFETs are preferred over the back gate design, despite their complex fabrication process.

Wrap-around gate CNTFETs (or gate-all-around CNTFETs) have an improved design over the top gate device. In this design, the entire nanotube volume is gated, while with the top gate design only the CNT closer to the metal gate contact is gated. This innovation improves the CNTFET electrical performance and reduces the leakage current.

Wrap-around gate CNTFET. Source: Wikimedia from author Popproject3.
Wrap-around gate CNTFET. (Source: Wikimedia user Popproject3.)

Suspended CNTFET design avoids placing the CNT over a trench, reducing contact with the substrate oxide and thus improving device performance. Fabrication methods to suspend the CNT over trenches use catalyst particles that are transferred onto a substrate.

The drawback of this design is its limited options for gate dielectric (air or vacuum). Moreover, only short CNTs can be used as nanotubes because the longer ones will stretch in the middle and could potentially touch the metal contact (creating a short-circuit). While this type of design is not suitable for commercial use, it is convenient for researching the intrinsic properties of a clean CNT.

CNTFETs are still a new technology—one with a lot of potential for improvement. Currently, the most popular designs are back gate and top gate CNTFETs. Some semiconductor companies (such as Infineon Technology) have introduced the next-generation design of vertical CNTFETs (VCNTFETs).

A vertical CNTFET. Source: S. J. Wind et al; Reference [4].
A vertical CNTFET. (Source: S. J. Wind et al; Reference [4].)

I-V characteristics

The current-voltage (I-V) characteristic curves represent a transistor’s operating characteristics—the relationship between the current flowing through the device and the applied voltage across its terminals. The figure that follows illustrates the drain I-V characteristics of CNTFETs. The saturation current at gate-source voltage VGS = 0.5V is approximately 6μA [2]. Saturation drain current from drain I-V characteristics depends on the temperature. Drain saturation currents slightly decrease when the CNTFET is cooled down. The curve is also determined by the CNTFET conductance, width, length, mobility of carriers, and gate capacitance.

Drain current-voltage characteristics of planar CNTFET. Source: Ram Babu; Reference [2].
Drain current-voltage characteristics of planar CNTFET. (Source: Ram Babu; Reference [2].)

When the gate and the source voltages of SB-CNTFETs increase, the Fermi level of the CNT becomes closer to the conduction band. The band lowering effect develops barriers at CNT-metal junction. The electrons with high potential will cross the barrier and flow into the tube. The current through the nanotube is limited by the thermionic current component.

When the gate voltage VGS=0V, the current increases linearly with the drain voltage VDS (the thermionic current is linearly dependent on the drain voltage). Applying positive gate voltage induces a heavy charge on the channel, significantly increasing the tunneling through the barrier compared to the thermionic current component. The current increases almost quadratically, is highly sensitive to the drain voltage, and is controlled by manipulating the barrier height at the contacts.


CNTFETs are up-and-coming devices that provide dense, high performance, and low power circuits. CNTFET is a rapidly developing technology due to its outstanding electrical characteristics. The large Ion: Ioff, high current drive, and carbon nanotube’s other properties increase the possible applications of CNTFETs in the semiconductor industry. They are the most promising alternative for conventional transistors. It is expected that with the same power consumption, they will be three times faster than silicon-based transistors.

In comparison to traditional silicon technology creating structures with minimum diameters reaching 90nm, SWCNTs have diameters between 0.4 and 5nm. Semiconducting SWCNTs have extremely high charge-carrier mobility—higher than silicon by a factor of 200. CNTs can withstand extremely high current densities of up to 1010A/cm2 (compared to the current density of copper, approximately 107A/cm2) [3].

Semiconducting carbon nanotubes (CNTs) are an ideal substitute for silicon  due to their exceptional carrier mobility, significant mean free path, and improved electrostatics at nanoscales. As the one-dimensional transport properties increase the gate control, simultaneously fulfilling gate leakage constraints, they also allow for a more comprehensive gate insulator choice. Thus, CNTs can overcome the short channel effects, and the valence bands and symmetry of the conduction give these devices the upper hand for additional applications. When applied in CNTEFs, CNTs can assist in providing high-speed ballistic CNTFETs.

In theory, CNTFETs have the potential to reach the terahertz regime when compared to standard semiconductor technologies. Nevertheless, this field is still at an early stage, and for the time being, researchers should remain focused on lowering the process variation.

[1] Design and Analysis of CNTFET-Based SRAM. Arvind R. Singh. Shandong University.

[2] Carbon nanotubes field-effect transistors: A review. International Journal of Electronics and Communications, Busi, Ram Babu. (2010).


[4] Vertical scaling of carbon Nanotube Field-effect transistors using top gate electrodes. Applied Physics Letters. May 2002. S. J. Wind et al.

Recommended For You