The Challenge Of Industrialization Of Diamond Semiconductor

Currently, diamond semiconductors are at a critical stage of transitioning from research and development to practical applications. Although they have achieved certain application results in fields such as thermal conductive substrates and radiation detectors, they still face many challenges.

 

Material growth is the primary challenge for the industrialization of diamond semiconductors. The current mainstream 12 inch silicon wafers have achieved large-scale applications, which can significantly reduce the unit cost of chips, while the size of diamond single crystal substrates is much smaller than 8 inches, directly limiting chip integration and production. Small sized substrates not only fail to meet the high-density layout requirements of large-scale integrated circuits, but also drive up equipment depreciation, raw material consumption, and other shared costs, weakening price competitiveness.

 

There are also bottlenecks in the preparation technology. Chemical vapor deposition (CVD) is the mainstream method, but the growth rate is only a few micrometers to tens of micrometers per hour, which is difficult to match the efficient production needs of the semiconductor industry. It also requires precise control of multiple parameters, and the equipment and operating costs are high. Although high temperature and high pressure method (HTHP) can produce diamond, it is prone to introducing impurities and defects, and cannot be directly used in semiconductors. However, the crystal quality and uniformity of diamond prepared by CVD method still need to be improved.

 

In terms of doping technology, both p-type and n-type are in a dilemma. P-type doping mainly relies on boron atoms, but the ionization energy of boron is as high as 0.37eV, making it difficult to completely ionize at room temperature and resulting in extremely low carrier concentration. If heavy doping is carried out to increase the concentration, it will lead to an increase in lattice stress and surface defects, intensify electron hole recombination, and increase the device's turn-on voltage and on resistance.

In theory, n-type doping can use phosphorus atoms, but their atomic radius is much larger than that of carbon atoms, which can cause severe lattice distortion during doping. This distortion significantly increases the probability of carrier scattering, leading to a sharp decrease in mobility. Currently, it is still difficult to obtain high concentration and high-quality n-type doped diamond, which limits the application of related devices.

 

However, some experts predict that 4-inch diamond substrates are expected to achieve mass production in the next 3-5 years, and their excellent conductivity characteristics are expected to solve the global problem of the lack of efficient p-type devices in wide bandgap semiconductors.

 

In device manufacturing, traditional semiconductor processes have poor compatibility with diamond. In the photolithography process, the surface characteristics of diamond are special, and ordinary photoresist is difficult to uniformly adhere, which can easily lead to pattern distortion and uneven lines; In the etching process, diamond has extremely strong chemical stability, and most traditional etchants have weak effects, making it difficult to accurately control the etching depth and shape.

 

The superhard properties of diamond also pose challenges for processing. Silicon and silicon carbide polishing pads need to achieve atomic level flatness (roughness RMS ≤ 0.1nm), while diamond has extremely high hardness and ordinary grinding tools wear out quickly. Even with diamond grinding wheels, there are still problems such as low efficiency and easy thermal damage, making it difficult to meet the "substrate level" surface quality requirements.