The Silicon Carbide Race Begins

That is not a problem specific to silicon carbide.

The main reason why a long time was needed for developing methods to make silicon carbide crystals with few defects and methods for doping them, is that the methods that worked with silicon and with many other semiconductors would require extremely high temperatures to work with silicon carbide.

So either alternative methods using more manageable temperatures or techniques to handle higher temperatures had to be found.

The need for higher temperatures also delayed for some years the use of silicon compared to germanium, but the problems were much easier solved, because the temperatures needed for silicon processing, e.g. in epitaxy or diffusion furnaces are still relatively low, less than 1400 Celsius degrees.

The second problem of silicon carbide is shared with all the high-bandgap semiconductors, e.g. also with gallium nitride.

For silicon and other low-bandgap semiconductors there is a very wide choice of dopants that can convert the intrinsic semiconductor into either P or N semiconductor of any desired resistivity in a wide range.

For silicon carbide and the other high-bandgap semiconductors, most dopants are very inefficient, e.g. only a very small fraction of the dopant might be active, especially at low temperatures.

So it can be very difficult to make a doped semiconductor with low resistivities and it can also be very difficult to find a metal that will make good contact with the device, without exhibiting rectifying characteristics (like a diode).

Solving all of these material problems required a large number of costly experiments with various chemical elements so it is normal that commercial devices have appeared only after many years of research.

Moreover, an additional reason for the very long time-to-market for high-bandgap semiconductor devices is that nowadays the research is done with much higher secrecy than in the early years of the electronics industry, when the Silicon Valley was created.

In the beginning, essential patents, like the patent of the transistor, were licensed without restrictions to anyone who desired for small royalties. Now most patents do not have the purpose to gain money by licensing but to prevent the appearance of any competitors, so they either are not licensed at all or too high royalties are requested.

In the early years, until around 1970, all the results of semiconductor research were published, including very detailed recipes with what was tried in the lab. After reading a research article, you could go in your lab and try the same recipe and reproduce the published research.

So any innovation made by some company spread immediately to all the companies and the technology of germanium, then that of silicon advanced very fast.

Now, the research articles only seldom are detailed enough. They usually just report some successes without mentioning the more important parts, i.e. what was tried and failed and what is still wrong with their current approach and remains to be solved in order to be able to make usable devices.

So most companies had to duplicate all the research done by the others to be able to reach to the stage of being able to make commercial devices.

Few large electronics companies have spent money on the research for high-bandgap semiconductors, because it was a long-term project.

The research was done mostly by smaller companies dedicated to this, but most of those which have been successful have been bought by now by the big players.