Why SiC? Probably the first question anyone asks. "It's the material of the future" is that glib reply that we've all heard and groaned about as the haunting thoughts of the early days of GaAs rear their ugly head. As much as 30 years ago people used the phrase in talking about GaAs and only recently has it begun to realise its small niche market. It's a history the SiC community is eager to avoid. The motivation for work in SiC is founded in markets where the Si workhorse has reached the limits imposed by nature and not by technology. No new device design will overcome the simple laws of physics, although there are many who find this hard to accept! It is this security that is a refreshing thought when faced with the technological mountains that SiC represents. Of further comfort is that, despite the differences with Si there is much about the chemistry of SiC that is similar to its more well known constituent element. In facing SiC, the technologist has already a wealth of processes that with refinement can be used in device production. It is perhaps this fact as much as any other that has again ignited the interest of the semiconductor community.

SiC is in fact no newcomer to the scene, some might argue it is the great grandfather of all semiconductors. The first report of SiC came as long ago as 1824, by a Swede (or we wouldn't mention it!) by the name of Jöns Jacob Berzelius, of course at the time the properties and potential of this material were not understood. Next came the development of the electric smelting furnace and the growth by Acheson of SiC around 1885; it was Acheson who fist recognised it as a silicide of carbon and gave it the chemical formula SiC. SiC does not occur naturally in nature so it cannot be mined like other minerals, hence the need for these elaborate furnace techniques. The only occurrence of SiC in nature is found in meteorites, a gift from the stars as one enthusiastic researcher once put it! In fact, mineralogists call natural SiC moissanite after the man who first identified SiC in a meteorite in 1905.

The first application of silicon carbide was as an abrasive and cutting material, but already in 1907 the first SiC LED was produced. The next major development in SiC happened in 1955, when Lely proposed a new concept for growing high quality crystals. From this point the interest in SiC as an electronic material slowly began to gather momentum, the first SiC conference was held in 1958 in Boston. During the 60's and 70's the SiC cause was mainly taken up by researchers in the former Soviet Union. 1978 saw a further major step in the development of SiC, the use of a seeded sublimation growth technique, this was basically a modification of Lely's original method and is still referred to as the "modified Lely technique". This breakthrough led to the possibility for true bulk crystal preparation and in 1987 CREE Research was founded as the first, and until recently the only, commercial supplier of SiC substrates. With the availability of high quality substrates the SiC field has exploded. SiC LEDs have already been in commercial production for a number of years and we can expect them to be joined by other electronic devices in the not too distant future. SiC is fast approaching the maturity which will allow it to take its long promised place alongside Si with its own niche market.

SiC exists not as a single crystal type but as a whole family of crystals known as polytypes. Polytypes differ not in the relative numbers of Si and C atoms but in the arrangement of these atoms in layers. The polytypes are named according to the periodicity of these layers, for example one of the commonest polytypes is called 6H, this means a hexagonal type lattice with an arrangement of 6 different Si+C layers before the pattern repeats itself. In total more than 200 different polytypes of SiC have been shown to exist, some with patterns that do not repeat for hundreds of layers. The exact physical properties of SiC depend on the crystal structure adopted, some of the most common structures used are 6H, 4H and 3C, the final of these structures is the one cubic form of SiC.

Above all it is the physical properties of SiC that ring to its merit and excite even the dullest to its potential. SiC belongs to a class of semiconductors commonly known as 'Wide bandgap'. What this means is not easily explained in layman terms, what it implies is that amongst other things these materials are less sensitive to increased temperatures. Within limits there is no reason why a SiC device should not operate at 500ºC or higher, a realm that Si does not even approach. The thermal conductivity exceeds even that of copper; any heat produced by a device is therefore quickly dissipated. The inertness of SiC to chemical reaction implies that devices have the potential to operate even in the most caustic of environments. It is extremely hard, SiC is probably most familiar to people as the grit coating on sandpaper, this hardness again implies that SiC devices can operate under conditions of extreme pressure. Of importance to our nuclear and space age is the fact that SiC is extremely radiation hard and can be used close to reactors or for space electronic hardware. Less transparent are the properties of particular importance to the device design engineer, high electric field strength and high saturation drift velocity. Again the implication of these parameters is that devices can be made smaller and more efficient. SiC is a material with which it is possible to stretch the limits of conventional technology to its extremes. Its market is at present a fringe one, but as the technology is developed and improved its impact will extend to the full reaches of what is currently the Si domain.

"Where are we now?" is usually the last question people ask, it is after all more glamorous to deal with the "might be" rather than the "is at the moment". It is really only in the late eighties that the SiC train really began again to gather substantial momentum, in that time its progress has been startling. The technological criteria that have plagued the development of most wide bandgap materials, and for most continue to do so, can be summarised as follows:

• The availability of a suitable substrate or bulk material.
• The growth of single crystal films over large enough areas and of high quality.
• Well controlled doping, both n and p-type.
• An appropriate and available dielectric, as needed for example in MIS devices.
• A developed technology for contacts and etching.

For SiC, all of these problems have to a large extent been solved although in some cases the technology is not yet at commercial production standard. Probably the best gauge of the maturity of material development is to look at the range and performance of the devices that have been made using it.

The power handling capabilities and high temperature operation make SiC an ideal material for the production of rectifiers. Both Schottky and p-n diode rectifiers have been produced. A record 4.5 kV p-n device was manufactured and demonstrated here at the Industrial Microelectronics Centre. Schottky devices have been demonstrated up to 1 kV. Operating temperatures in excess of 350ºC have also been shown. Other two terminal devices that have been made are SiC photo-diodes for the UV and SiC light emitting diodes. The latter has already been successfully marketed for a number of years. The photo-diode is of particular interest in that it can be used in such harsh environments, a property of particular interest for example for the in-situ monitoring of engine combustion or for monitoring from satellites. The SiC photodiode has been shown to achieve some four orders of magnitude higher sensitivity than its Si counterpart.

Switching devices constitute perhaps the largest current area of device activity. Of these by far the largest effort has concentrated on uni-polar devices i.e. devices that rely on the transport of one carrier type only for their operation. These include Field Effect Transistors in their many guises, JFET, MOSFET and MESFET.

MOSFETs dominate Si switching applications, it is in the high power, high temperature end of the market that SiC is expected to triumph. The operation of a Si power MOSFET is limited by the low doping and large thickness needed to support the high electric field, as a result the on-state resistance of such a device is large, limiting the current handling capabilities. The higher field strength of SiC implies that the on-state resistance can be lowered by some two orders of magnitude. A perhaps significant value of 13 is the derived figure for the improvement in power handling capabilities of a SiC device over its Si counterpart. Clearly this will have a major impact on the size, efficiency and application of power electronics.

Devices have been manufactured in all of the principal polytypes 3C, 6H and 4H. Initial work focused on 3C material due to the superior transport properties, although due to the poor quality of the material recent attention has switched to the 6H and 4H polytypes. A notable result even from this early work was the operation of a MOSFET up to 650ºC. The largest remaining concern in the development of the MOSFET is the relatively poor quality of the Silicon Dioxide dielectric on SiC as compared to Si. In particular the interface roughness leads to a significant degradation in the channel mobility leading to a low transconductance. Significant progress is, however, being made in this direction in particular using deposited dielectrics rather than the thermally grown oxide.

MESFETs have received a great deal of attention due to their application in high frequency, high power systems. They do not require the high quality dielectric needed in MOS devices and can be manufactured with much smaller dimensions. Again the physical and electrical properties of SiC allow much higher power densities to be achieved, an extremely important requirement for signal generation in such applications as radar or mobile telecommunications.

As SiC technology matures we can expect the design and complexity of SiC devices to increase, already there is work on SiC integrated circuits. Without question the increasing reliability of SiC devices will allow it to begin to reach its real potential in areas where no other semiconductor can cope with the environmental conditions. The harsh conditions of Space is one area in particular where SiC devices will demonstrate their merit. Ironically the material that first came from the stars is exactly the same material which may help us to reach out to the stars. What ever the final use for SiC, it is clear that SiC is a "material for the future".

The author would like to acknowledge with thanks the numerous discussions with collegues at IMC and the University of Linköping which have contributed to the writing of this article.

Chris Harris

The pictures are copyright Dr. Virgil B. Shields and used with permission. They originate from his What is Silicon Carbide? pages.