What a bandgap is
In a semiconductor, electrons occupy energy levels grouped in bands. The valence band holds electrons that are bound in the crystal’s bonds. The conduction band holds electrons free to carry current. Between them lies a forbidden range of energy called the bandgap. An electron must gain at least that much energy, from heat, light or an electric field, to jump into the conduction band.
Bandgap is usually quoted in electronvolts (eV). Silicon, the workhorse of electronics, has a bandgap of about 1.1 eV. Silicon carbide and gallium nitride are around 3 eV or a little more, and are called wide bandgap semiconductors. Diamond sits higher still, at about 5.5 eV, making it an “ultra-wide” bandgap material.
Why a large gap is useful
A wider gap changes how a device behaves in several ways.
- Higher breakdown field. A strong electric field can rip electrons across the gap and cause a sudden flood of current. A wider gap generally means a higher field is needed to do that. A device can therefore be thinner for the same voltage, which lowers its resistance and wasted power.
- Better high-temperature operation. Heat randomly lifts electrons across the gap. In silicon, this starts to ruin device behavior at moderately high temperature. A wider gap pushes that limit much higher.
- Lower leakage. Fewer thermally freed carriers means less unwanted current when the device is supposed to be off.
- Radiation tolerance. Strongly bonded, wide-gap crystals tend to cope better with radiation, the same property that makes diamond interesting as a detector material.
These traits point to power electronics: switches and converters for electric vehicles, power grids, industrial drives and high-frequency radio equipment. Silicon carbide and gallium nitride already serve these markets. Diamond is the longer-term candidate.
What diamond adds
Beyond the large gap, diamond has other features that look attractive on paper.
Outstanding thermal conductivity. Power devices produce heat, and diamond removes it better than almost any other solid. A diamond device could in principle run at higher power density with simpler cooling. This ties into thermal management, a major limit on power electronics generally.
High carrier mobility. In high-quality material, electrons and holes can move quickly, which helps with switching speed and efficiency.
Chemical and mechanical robustness. It stands up to high temperatures and many harsh environments, which supports its potential as a material for high-temperature materials in electronics.
Combining these, researchers often describe diamond as a possible “ultimate” semiconductor for extreme conditions. The word to notice is “possible.”
The obstacles
Diamond electronics are still mostly at the laboratory and early prototype stage. Several hurdles explain why.
Doping is difficult. A useful semiconductor needs both p-type (hole-conducting) and n-type (electron-conducting) regions. In diamond, boron doping produces p-type behavior and is well established. But the boron acceptor level is fairly deep, so at room temperature only part of the dopant contributes charge carriers, though this improves at higher temperature or heavier doping. N-type is harder: phosphorus and nitrogen donors sit deep in the gap, and nitrogen is especially poor for conduction. Good n-type diamond has been demonstrated mainly under specific growth conditions and with phosphorus, and it remains a research challenge.
Wafer size and cost. Silicon is made into wafers many centimeters across at low cost. Single-crystal diamond plates are generally small, often a few millimeters to a couple of centimeters across, and expensive. Some groups are working on larger wafers and on growing diamond on other substrates, but defect densities are higher in such heteroepitaxial layers. Quality depends heavily on how seed crystals and substrates are prepared.
Defects and impurities. Dislocations and unwanted impurities degrade devices. Growth by chemical vapor deposition has improved purity a great deal, but electronic-grade control remains demanding.
Contacts and processing. Making low-resistance electrical contacts and patterning devices on such a hard, inert material is difficult. Etching, implantation and polishing all need specialized methods.
Surface effects. Hydrogen-terminated diamond surfaces can form a thin conducting layer when exposed to air, which has been used for transistors but behaves differently from conventional bulk doping, and stability can be a concern.
What has been demonstrated
Researchers have built diamond Schottky diodes, field-effect transistors, and other test devices that show high breakdown voltages and operation at high temperature. Diodes are further along than switches, in part because they need only one doped type. Devices using the conductive hydrogen-terminated surface layer have been demonstrated as well. These are research results rather than products you can buy at scale, and performance figures depend on specific samples.
Where it fits among its competitors
Silicon carbide and gallium nitride are commercial today. Silicon carbide is used heavily in high-voltage switches, and gallium nitride in fast, lower-voltage converters and radio-frequency amplifiers. Gallium nitride devices can even be paired with diamond as a heat-spreading substrate, which is a nearer-term way to use diamond in electronics without making diamond transistors.
For the foreseeable future, the most realistic roles for diamond in electronics are as a thermal substrate, a sensor or detector material, and a research platform for ultra-high-power concepts. Full diamond power devices may eventually matter in niches where voltage, temperature and radiation exceed what silicon carbide can handle, but when or whether that happens is uncertain, and it depends largely on whether cost, wafer size and n-type doping can be solved.