2D materials in 3D transistors; electrochemical memristive mechanism; matching substrates for power electronics.
Researchers at the University of California Santa Barbara investigated 3D gate-all-around (GAA) transistors made using 2D semiconductors. They considered three different approaches to channel stacking: nano-sheet FETs, nano-fork FETs, and nano-plate FETs.
The nano-plate FET architecture, which exploits lateral stacking of 2D layers, was found to maximize the gains from the unique properties of atomically thin 2D materials like tungsten disulfide (WS₂), enhancing integration density by tenfold with iso-performance metrics.
Three-dimensional gate-all-around CMOS field-effect transistors in the form of a nano-sheet, a nano-fork, and a nano-plate, all uniquely enabled by two-dimensional layered semiconductors. G=gate, S=source, D=drain, and Lch=channel length (Credit: UCSB)
“By leveraging the unique physical and quantum mechanical properties of 2D materials, we can overcome many of the limitations associated with conventional 3D transistors designed with silicon,” said Kaustav Banerjee, a professor of electrical and computer engineering at UCSB, in a press release. “Our simulations show that the nano-plate transistors achieve significant improvements in energy efficiency and performance, with channel lengths scaled to sub-5nm dimensions.”
Using simulations that modeled critical factors such as energy-band non-parabolicity, finite bandwidth, contact resistance, and carrier mobility, the researchers determined that the 2D semiconductor-based 3D-FETs outperform their silicon counterparts in metrics such as drive current and energy-delay-product. They hope the work offers a blueprint for the integration of 2D materials into 3D transistor designs. [1]
Researchers from Forschungszentrum Jülich, RWTH Aachen University, and Peking University developed memristive components that are more robust, function across a wider voltage range, and can operate in both analog and digital modes. The memristor exhibits properties that could enable neuromorphic systems to learn continually, similar to neural plasticity in the brain.
“We have discovered a fundamentally new electrochemical memristive mechanism that is chemically and electrically more stable,” said Ilia Valov from the Peter Grünberg Institute (PGI-7) at Forschungszentrum Jülich in a press release. “Its unique properties allow the use of different switching modes to control the modulation of the memristor in such a way that stored information is not lost.”
Two main mechanisms have been identified for the functioning of bipolar memristors: electrochemical metallization (ECM) and valence change mechanism (VCM). ECM memristors change electrical resistance based on the formation of metallic filaments between two electrodes, while VCM memristors change resistance based on the movement of oxygen ions.
“Our new memristor is based on a completely different principle: it utilizes a filament made of metal oxides rather than a purely metallic one like ECM,” Valov explained in the release. This filament is formed by the movement of oxygen and tantalum ions and is highly stable—it never fully dissolves. “You can think of it as a filament that always exists to some extent and is only chemically modified.”
Additionally, the different oxidation states created by this filament conductivity modification mechanism (FCM) allow the memristor to be operated in a binary and/or analog mode, giving it a potential use in neuromorphic chips. The team implemented the memristive component in a model of an artificial neural network in a simulation and achieved a high level of accuracy in pattern recognition. [2]
Researchers from the National Renewable Energy Laboratory (NREL), the Colorado School of Mines, and Oak Ridge National Laboratory are working to improve the performance of aluminum gallium nitride (AlxGa1–xN) by growing it on optimized, lattice-matched substrate materials.
Growing AlxGa1–xN on lattice-mismatched substrates leads to dislocation, line defects that distort a lattice due to the misalignment of atoms and diminish performance. To address this, the team used electrically conductive, lattice-matched tantalum carbide (TaC) as a suitable substrate for AlxGa1–xN epitaxy.
“If we can engineer lattice-matched substrates to reduce the effect of dislocations, we can widen the range of sufficiently high-quality materials and build better, more energy-efficient power electronics,” said Dennice Roberts, a materials science researcher at NREL, in a statement. “Lattice matching is critical for high-quality epitaxial growth. We hypothesized that substrates from transition metal carbide and nitride families could enable desired conditions for AlxGa1–xN growth, not only because of ideal lattice matching but also because of ideal thermal and electrical conductivity properties. TaC and AlxGa1–xN are closely lattice-matched, TaC is highly conductive, and they display matched growth in size in response to changes in temperature.”
The researchers used TaC thin films as virtual substrates. High-temperature annealing of the TaC film created a flatter surface that facilitated the growth of higher-quality AlxGa1–xN.
Other researchers at NREL investigated how to form interfaces between materials with different crystal structures by looking at how individual atomic layers are stacked. [3]
[1] Pal, A., Chavan, T., Jabbour, J. et al. Three-dimensional transistors with two-dimensional semiconductors for future CMOS scaling. Nat Electron 7, 1147–1157 (2024). https://doi.org/10.1038/s41928-024-01289-8
[2] Chen, S., Yang, Z., Hartmann, H. et al. Electrochemical ohmic memristors for continual learning. Nat Commun 16, 2348 (2025). https://doi.org/10.1038/s41467-025-57543-w
[3] Roberts, D.M, Hachtel, J.A, Haegel, N.M, et al. Designing TaC Virtual Substrates for Vertical AlxGa1-xN Power Electronics Devices, PRX Energy (2024). https://dx.doi.org/10.1103/PRXEnergy.3.033007
Leave a Reply