Coal-Derived Carbon Nanodots Offer a Primer for Atomically Thin Transistors

Silicon has carried the digital age for more than half a century, but the material itself is beginning to set the limits of what transistors can do. As engineers shrink devices toward atomic dimensions, silicon’s own physics starts to work against them, degrading performance and efficiency in ways that no amount of clever design can fully overcome. A promising escape route lies in two-dimensional materials, crystalline sheets just one atom thick that retain excellent electronic behavior at scales where silicon falters. Yet the very feature that makes these sheets so attractive also makes them notoriously difficult to build into working devices. Now, researchers at the University of Illinois Urbana-Champaign’s Grainger College of Engineering, working with the U.S. Department of Energy’s National Energy Technology Laboratory, report a surprisingly elegant fix, and it comes from one of the world’s oldest energy resources: coal.

The team, led by Qing Cao, associate professor of materials science and engineering, found that carbon nanodots processed from bituminous coal can serve as an atomic-scale interfacial binder for two-dimensional semiconductors such as molybdenum disulfide and graphene. These nanodots are graphene-like particles measuring roughly one to five nanometers across and only a single atomic layer thick. When spread across the surface of a 2D semiconductor, they act as nucleation centers from which an ultrathin insulating dielectric layer can be grown, without chemically hybridizing with the electronic channel beneath. The result is a clean interface between the semiconductor and its insulator, along with dielectric films that are thermally stable and thin enough for the most demanding transistor designs. The work was published in the journal Nature Communications.

“Devices made from 2D materials are very attractive because they can continue scaling to dimensions where silicon begins to encounter significant performance challenges,” Cao said. “But without a 3D chemical structure, it is very difficult to directly grow the insulating dielectric layers on top required for transistor operation. It’s a very old bottleneck, and we’ve found an answer from an unexpected place: carbon nanodots made from processed coal that act as a primer coat for growing thin, high-quality dielectrics on 2D materials with a clean interface, a crucial step to manufacturing transistors.”

To understand why this bottleneck has persisted for so long, it helps to consider what happens at the surface of a semiconductor. Silicon atoms at a surface carry what materials scientists call dangling bonds: free electrons that are not tied up in chemical bonds the way atoms in the bulk of the crystal are. In conventional manufacturing, those dangling bonds are actually useful, because they provide chemical attachment points that allow insulating films to be deposited through well-established processes such as chemical vapor deposition and atomic layer deposition. But when silicon is scaled to ultrathin dimensions, those same dangling bonds and associated interface defects begin to scatter and slow down the electrons that carry current, eroding the material’s performance precisely where scaling demands it most.

Two-dimensional materials such as molybdenum disulfide sidestep that problem entirely. Because every atom in the sheet is fully bonded within the plane, there are no dangling bonds at the surface, and current-carrying electrons can move with far less scattering even in atomically thin channels. That absence is a blessing for device performance but a curse for fabrication. “The lack of dangling bonds is very attractive from a scaling perspective, but, at the same time, it is also a major integration obstacle to overcome,” Cao said. “With silicon, the dangling bonds allow insulating films to be grown using the well-established processes of chemical vapor deposition and atomic layer deposition. But with 2D materials, which have no free chemical bonding sites, figuring out how to grow layers on them is a longstanding problem.” Without a reliable way to lay down a high-quality insulator, 2D semiconductors have remained largely confined to laboratory demonstrations rather than production lines.

The carbon nanodots resolve this paradox through a clever division of chemical labor. Their flat, graphene-like faces interact with 2D materials through the van der Waals force, a gentle physical attraction that requires no chemical bonding and therefore leaves the semiconductor’s electronic structure untouched. Meanwhile, the functional groups decorating the edges of each nanodot can form genuine chemical bonds with oxide insulators deposited on top. In effect, each nanodot is a two-faced molecular adapter: one side grips the 2D semiconductor without disturbing it, and the other side anchors the dielectric. The nanodots thereby bind the two layers together while preserving the pristine interface that gives 2D devices their performance advantage.

Getting the nanodots into position required their own ingenuity. “The carbon nanodots function as nucleation centers from which the dielectric layer grows, so they need to be packed as tightly as possible,” Cao said. The team achieved this with a technique called Langmuir-Schaefer assembly. The nanodots were first dispersed in an organic solvent, and drops of that liquid were placed onto a water surface. Driven by surface tension, the liquid spread into a thin layer; as the solvent evaporated, the nanodots were left floating on the water. The researchers then mechanically pushed them together, forming a tightly packed monolayer that could be transferred onto the 2D semiconductor surface. That monolayer then seeded the growth of a smooth, ultrathin, high-dielectric-constant oxide using atomic layer deposition, a workhorse technique already ubiquitous throughout the semiconductor industry.

The resulting films did not merely work; they exceeded benchmarks. The dielectric layers surpass the industry targets laid out by the IEEE in the International Roadmap for Devices and Systems, achieving an equivalent oxide thickness of 0.6 nanometers and a leakage current density below 0.1 milliamperes per square centimeter. Those figures matter because a thinner equivalent oxide allows a transistor’s gate to control the channel more effectively at a given voltage, while low leakage keeps standby power consumption in check. The team went further and demonstrated complete devices: field-effect transistors and integrated logic gates fabricated with the new process operate at around 0.5 volts. That is substantially lower than the roughly 0.6- to 0.8-volt range typical of today’s advanced silicon technologies, and lower operating voltage translates directly into less energy required for every computation a chip performs.

Perhaps the most striking aspect of the work is its industrial pragmatism. Coal is normally regarded as a dirty material, while microelectronics fabrication demands extraordinary cleanliness, so a significant portion of the study was devoted to characterizing the nanodots and verifying their chemical purity. “Normally, coal is thought of as dirty, and microelectronics need to be extremely clean,” Cao said. “We spent a great deal of effort making sure that the materials we derived have an extremely high level of purity. We made sure no impurities incompatible with standard silicon CMOS manufacturing environments were present in the monolayer.” Because the dielectric growth step relies on atomic layer deposition, a standard manufacturing technique, the process is designed to slot into existing fabrication flows rather than require an entirely new infrastructure. Cao’s group has previously collaborated with NETL on applications of coal-derived materials in microelectronics, and the unique chemical structures and physical properties of carbon found in coal are now forming the basis for a new class of advanced materials, an unexpected bridge between an old energy resource and the most advanced electronic technologies.

The project also drew support from the Center for Advanced Semiconductor Chips with Accelerated Performance at Illinois Grainger Engineering, which fosters connections with semiconductor industry partners including IBM, Intel and the Taiwan Semiconductor Manufacturing Company, a signal that the work is being positioned with real manufacturing adoption in mind. Contributors to the study include Sunny Wong, Fufei An, Yu Wu, Kaijun Yin and Jian-Min Zuo of the University of Illinois Urbana-Champaign, along with Viet Hung Pham, Yanxiao Li, Robert Thompson, Junseok Lee, Yuan Gao, Congjun Wang and Christopher Matranga of the National Energy Technology Laboratory. The study, “Integration of high-κ oxide on 2D semiconductors with carbon-dot monolayer assembly as van der Waals interfacial layer,” appeared in Nature Communications on 7 October 2026. “Researchers have known that two-dimensional semiconductors can outperform ultra-thin silicon, but integrating high-quality dielectric layers has remained a major challenge,” Cao said. “By using carbon nanodots as an atomic-scale interface, we have demonstrated a practical pathway that could help bring these materials into future electronic technologies.” If the pathway holds at scale, the humblest of fossil fuels may end up helping to write the next chapter of computing.

Subject of Research: Coal-derived carbon nanodots as an interfacial layer for growing high-quality dielectrics on two-dimensional semiconductors

Article Title: Coal-derived carbon nanodots enable next-generation 2D electronics

Article References: Coal-derived carbon nanodots enable next-generation 2D electronics. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: carbon nanodots, two-dimensional materials, molybdenum disulfide, dielectric layers, atomic layer deposition, transistors, coal-derived materials, molybdenum disulfide electronics, van der Waals interface, semiconductor manufacturing, low-voltage logic, Nature Communications