Is C-Axis Czochralski the Wave of the Future for Sapphire?

The Czochralski process has been the dominant method of growing sapphire crystals for about 50 years ever since it displaced the Verneuil process.  It produces crystals of high quality at a reasonably low price.  In the past decade there has been an explosive growth in the consumption of sapphire substrates primarily driven by the demands of the GaN blue LED market.  This application requires crystals oriented such that the substrate surface is the c-plane, a basal plane of the hexagonal structure.  Sapphire crystals form a strong facet on the c-plane, and growth in that direction generally results in crystals with high defect densities, particularly dislocations and low angle grain boundaries.  To overcome this drawback, the usual methodology is to grow the crystal in the a-direction and then core drill rods perpendicularly which are then sliced into c-plane substrates.  For all crystal growth techniques commonly employed for sapphire (Czochralski, Kyropoulos, HEM), this approach suffers from poor material utilization.  Although this has generally been viewed as an acceptable trade-off in the manufacturing process as long as 2” substrates were the dominant market, as substrate diameters have increased towards 6” and 8”, this compromise is no longer seen as a viable alternative because of the low material utilization and the high energy consumption of the growth process.  Recently, interest in the Czochralski process has been rekindled because of its greater material utilization when crystals are grown in the c-axis direction.  This has been abetted by developments in GaN epitaxy which overcome the issues related to substrate quality.  Growth of large diameter c-axis sapphire crystals by the Czochralski process is not without its own issues.  Scaling to large diameter crucibles requires an abandonment of the traditional approach of using induction heated iridium crucibles and operating in an unsealed growth chamber.  Now the manufacture of large refractory metal crucibles combined with the thermodynamic interactions of the metal crucible, insulation (generally graphite) and oxide (sapphire) become important considerations in the successful implementation of the process.

I have already seen x-ray topographs of 6” CZ wafers cut from c-axis crystals.  Although the quality does not compare to a-axis material, it is highly likely that this will have little impact on the GaN epilayer and the subsequent LED yield and performance.  I expect to see increased activity in this area in the future.