Reconfigurable Ferroelectric Field-Effect Transistors with Single-Walled Carbon Nanotube Channels

 

The rapid development of the internet and artificial intelligence has accelerated data generation, imposing greater demands on data processing. To process data efficiently, it is necessary to reduce device dimensions and operating voltage, thereby lowering energy consumption.

However, with current field-effect transistor (FET) technology, such as silicon complementary metal-oxide-semiconductor (CMOS) transistors, these improvements have met with physical limitations: when transistors become too small, electrical control issues arise. More specifically, size reduction is hindered by effects associated with short channel length; while lowering the voltage below 1 V is constrained by the Boltzmann limit, which depends on the constant ratio between the interaction potential and the mean free path of charge carriers.

Considering these limitations, reconfigurable transistors—capable of changing their operation after fabrication—offer a promising alternative. A special case is reconfigurable FeFETs, which can operate as either p-type or n-type transistors as needed.

A research team from Korea, the U.S., and China developed this type of devices using highly aligned single-walled carbon nanotubes (SWCNTs) as semiconductor channels and an innovative ferroelectric material (aluminum–scandium nitride). These devices exhibit ambipolar carrier characteristics with high and well-balanced ON-state currents (~270 μA μm−1 at a drain voltage of 3 V) and ON/OFF ratios greater than 105, along with wide memory windows and excellent retention capability. Furthermore, they feature ternary memory capability (able to store -1, 0, or +1 instead of just 0 and 1). This means that more compact and efficient circuits can be built compared with those based on conventional silicon.

Further information in: Nature Communications

Homoepitaxial Growth of MoS₂ in a Rhombohedral Stacking

 

Two-dimensional (2D) materials possess unique properties that make them suitable for implementation in various technological devices. Manipulating and modifying the stacking order between layers is a way to tune their properties.

Two-dimensional transition metal dichalcogenides (TMDs) are important candidates for transistor downscaling. These materials are typically stacked in an arrangement known as 2H, consisting of two monolayers rotated by 180°. Inducing an alternative artificial stacking (for example, the rhombohedral stacking known as 3R, in which three monolayers are shifted without rotation) imparts exotic properties such as ferroelectricity, superconductivity, among others. However, inducing this type of stacking is not straightforward, due to the high thermodynamic stability of the 2H stacking.

Scientists from Chinese universities have succeeded in obtaining large-scale (on the order of centimeters) molybdenum disulfide (MoS₂) with 3R stacking. Using chemical vapor deposition (CVD), they deposited a MoS₂ monolayer onto a sapphire substrate. Once the first monolayer was deposited, the nucleation process of the subsequent monolayer became crucial to achieving the desired stacking, since once formed, it cannot be rotated.

This study proposes, through density functional theory (DFT) calculations, that defects known as Mo antisites promote 3R stacking over 2H. Experimentally, the presence of such defects was demonstrated in samples with 3R stacking, while no such defects were observed in samples with 2H stacking, corroborating the proposed growth mechanism. The obtained samples were shown to exhibit ferroelectricity via piezoresponse force microscopy (PFM).

Since 3R-MoS₂ is a semiconductor with ferroelectric characteristics, it is proposed for the fabrication of ferroelectric semiconductor field-effect transistors (FeS-FETs). Moreover, the advantage of obtaining high-quality, large-area samples ensures device reproducibility.

For more information go to:

nature materials