Surface States Hold the Key to Diamond’s Elusive n-Type Electron Gas

Diamond has long been the dream material for electronics that must survive conditions that would destroy silicon. It conducts heat better than any other bulk material, it can withstand enormous electric fields, and its wide band gap of 5.5 electron volts promises transistors that keep working at temperatures and power densities far beyond the reach of conventional semiconductors. Yet for all its virtues, diamond has a stubborn flaw that has frustrated engineers for decades: it is extraordinarily difficult to dope n-type, meaning to introduce donors that supply free electrons. Without a reliable source of electrons, the most powerful transistor architectures remain out of reach for diamond, no matter how spectacular its intrinsic properties may be.

A promising way around this roadblock has emerged in recent years in the form of heterostructures that pair diamond with cubic boron nitride, a wide-band-gap compound that is structurally compatible with diamond and nearly as hard. When the two crystals are joined, a two-dimensional electron gas, a sheet of mobile electrons confined to the interface, can form with a density of roughly ten trillion electrons per square centimeter. That is the kind of carrier density that underpins the high-frequency transistors in modern radar and communications systems built from gallium nitride. But a critical question has shadowed the field: where exactly do these electrons come from, and what controls how many of them appear at the interface?

A new theoretical study published in the Journal of Materials Science by Yifan Xu, Yao Li and colleagues at Xi’an University of Technology and Xidian University tackles that question head-on. Using technology computer-aided design, or TCAD, simulation software, the team systematically varied the properties of donor-like surface states, electronic trap levels that sit at the outer surface of the c-BN layer, and tracked how each change rippled through the band structure of the heterostructure and altered the density of the two-dimensional electron gas. Their simulations were run both with and without polarization charges, the bound sheet charges that arise in polar crystals and that are known to play a decisive role in the analogous aluminum gallium nitride and gallium nitride systems.

The central finding is that the energy level of the donor surface states, essentially how deeply those traps sit below the conduction band, is the dominant lever controlling the electron gas. When the team held the surface state density fixed at ten trillion per square centimeter and increased the donor ionization energy from 0.4 to 2.0 electron volts, the simulated electron gas density collapsed from 4.08 times ten to the twelfth per square centimeter down to just 1.27 times ten to the eleventh in the absence of polarization charges. In other words, a fivefold deepening of the donor level stripped away more than ninety-six percent of the interfacial electrons. Shallow donors, which ionize readily, feed the electron gas generously; deep donors, which hold onto their electrons, starve it.

Polarization charges change the picture in an instructive way. When the researchers included them in the c-BN barrier layer, the same increase in donor ionization energy still reduced the electron gas density, but far less dramatically: from 4.65 times ten to the twelfth down to 3.27 times ten to the twelfth per square centimeter, a decline of about thirty percent rather than a near-total collapse. The physical interpretation is that polarization charge provides an electrostatic foundation for the interfacial quantum well, so that even when the surface donors become stingy, a substantial reservoir of electrons remains confined at the diamond side of the junction. This robustness mirrors the logic that made gallium nitride transistors practical, where polarization doping does much of the heavy lifting.

Perhaps the most surprising result concerns the density of the surface states themselves. Intuition might suggest that more donor sites should mean more donated electrons, but the simulations show that once the surface state energy level is fixed, increasing the surface state density has only a minor effect on the electron gas density in both the polarized and unpolarized cases. The system, it turns out, is limited not by how many donors are available but by how easily they release their electrons and by the electrostatic boundary conditions of the whole stack. For device designers, that is a liberating conclusion: achieving a strong electron gas does not require engineering an impossibly precise surface state density, but it does demand control over the energetic position of the surface states.

The thickness of the c-BN barrier layer tells a similar story about the interplay between geometry and polarization. Without polarization charges, thinning the c-BN cap from 50 nanometers to 5 nanometers boosted the electron gas density substantially, and growing it thicker cut the density by 63.8 percent across the studied range, because the surface donors sit farther from the interface and their influence weakens with distance. With polarization charges included, that thickness dependence nearly vanished: the electron gas density changed by only 14.9 percent over the same 5-to-50-nanometer span. A polarization-anchored electron gas is therefore far less sensitive to the exact thickness of the barrier, which relaxes one of the most demanding constraints on epitaxial growth and should make manufacturing far more forgiving.

Temperature behavior is where the heterostructure delivers its most striking headline numbers. When the researchers raised the ambient temperature in their simulations from 300 kelvin, roughly room temperature, to 600 kelvin, the electron gas density climbed in a nearly linear fashion in every configuration they tested. For a shallow donor level of 0.4 electron volts, the density surged by 149.5 percent with polarization charges and 134.1 percent without them. Even for a much deeper donor level of 1.6 electron volts, the density still rose by 128.7 percent with polarization and 125.1 percent without. The mechanism is straightforward thermal physics: at higher temperatures, electrons trapped at the surface states are thermally excited over the donor barrier more efficiently, replenishing the interfacial sheet with additional carriers.

That temperature resilience is not a mere curiosity; it is the property that could set diamond electronics apart from every competing technology. Silicon devices begin to falter well below 600 kelvin, and even gallium nitride and silicon carbide, the current champions of wide-band-gap electronics, face mounting challenges as junctions heat up. A transistor whose carrier supply actually strengthens as it gets hotter, rather than degrading, inverts the usual thermal penalty and suggests that c-BN and diamond heterostructures could serve in jet-engine controllers, deep-well drilling electronics, spacecraft instrumentation and next-generation power grids where cooling is difficult and heat is unavoidable. The authors describe this as robust carrier activation and strong high-temperature operational potential, and the simulation numbers back that language up.

There are, of course, caveats that separate a simulation study from a finished device. TCAD modeling depends on assumed band alignments, state distributions and material parameters, and the real c-BN and diamond interface harbors structural complexities, from dislocations to adsorbates, that idealized simulations simplify. Experimental validation, in which surface treatments are deliberately tuned and the resulting electron gas is measured, will be the decisive test. Still, the study provides the field with something it has lacked: a quantitative map of which knobs matter. It says that the energy of donor surface states, not their number, is the master variable; that polarization charges buffer the system against imperfections in both donor depth and barrier thickness; and that heat, the traditional enemy of electronics, may actually be an ally for diamond. If those predictions survive contact with the laboratory, the long-sought n-type diamond transistor may finally be within engineering reach, and the hardest of all semiconductors may at last get the electronics it deserves.

Subject of Research: Donor surface state control of the two-dimensional electron gas in cubic boron nitride/diamond heterostructures

Article Title: Effect of donor surface states on 2DEG in c-BN/diamond heterostructures

Article References: Xu, Y., Li, Y., Su, K., Zhu, J., Zhang, J., Zhang, J., & Hao, Y. (2026). Effect of donor surface states on 2DEG in c-BN/diamond heterostructures. Journal of Materials Science, 61(42), 32995-33004. https://doi.org/10.1007/s10853-026-13673-0

Image Credits: AI Generated

DOI: 10.1007/s10853-026-13673-0

Keywords: diamond, cubic boron nitride, two-dimensional electron gas, n-type doping, donor surface states, polarization charges, heterostructure, TCAD simulation, high-temperature electronics, field-effect transistors, wide band gap semiconductors, Journal of Materials Science