A diffusion furnace room in a wafer factory in England, before the orange-glowing mouth of a horizontal quartz tube a clear quartz glass boat thirty centimetres long on a quartz loading paddle, two lower rails with closely spaced ground grooves, above them two round rods with fine cross notches, fused cross struts and end frames, a small round eye at one end, in the grooves fifty mirror-grey silicon wafers ten centimetres across standing upright, a hand in a white glove at the paddle end, behind it a furnace cabinet of plain smooth cream panels with four stacked tube mouths, soft yellow light from above.
What it was used for
The quartz boat holds wafers upright and spaced while they are oxidised or doped in a tube furnace. The material is fused quartz: by the data-sheet values for GE Type 214, nominally 99.995 weight percent silicon dioxide with typically under 50 ppm impurities - the oxide of the same silicon the wafer is made of, in glassy form. A European patent on diffusion tubes describes the use in quartz tubes closed at both ends into which gas is admitted, and says fused quartz can withstand process temperatures in excess of 1200 degrees Celsius.
What was new
What is characteristic is that the part works above its own limits. The strain point of fused quartz is 1120 degrees by the GE data and the annealing point 1215 degrees - at diffusion temperatures, by GE, the material slowly sags, which a low OH content only reduces. Beta-cristobalite, on the same data, generally does not form below 1000 degrees; nucleation occurs only at the free surface and is usually started by alkali or metal contamination, and thermal cycling through the inversion at about 250 degrees can cause spalling or mechanical failure. The barb: the recommended remedy is to cause the devitrification deliberately - 1200 degrees, rotating the tube 90 degrees every 2 hours for the first 30 hours, until a protective cristobalite layer has formed.
What the sources leave open: The values refer to quartz tubes and to fused quartz as a material; GE expressly names Type 214 rod as material for wafer carriers, while the patent speaks only of tubes. Nothing solid was found on service life, unit numbers or prices of boats. The patent deals with sodium diffusion through the tube wall and does not name silicon carbide as an alternative.