Studies of co-implanted helium and hydrogen with an intermediate annealing step for thermal splitting of bonded silicon to oxide-coated wafers

Abstract

Layer transfer by thermal or mechanical splitting of implanted silicon wafers can often require processing at elevated temperatures before bonding and annealing. The possibility has been investigated here of making two implants, the first defining a damage plane for splitting, the second being made after high-temperature processing. Helium and hydrogen implanted wafers were annealed at various temperatures, Ta, this being followed by a second implantation of either gas. Thermal splitting of bonded wafer pairs was studied for both implants to the same projected range, Rp, and with hydrogen implanted to Rp=25% less than helium. In the former case, complete thermal splitting occurred for Ta⩽550°C; for Ta>550°C, wafers failed to split or delaminated at the bond interface. When hydrogen was implanted to a shallower depth complete thermal splitting occurred for Ta⩽500°C. The split occurred at the hydrogen Rp for Ta<400°C and at the helium Rp for Ta=or>400°C. Above Ta=500°C a decreasing percentage of the total silicon area was transferred. At 400°C measurements suggested that splitting occurred at both helium and hydrogen Rp planes and it is believed that helium platelet formation at this temperature is an important factor. For Ta<400°C, helium, a proportion of which may have diffused towards the peak of the hydrogen distribution, is believed to react and produce splitting there during bond-annealing. For Ta>400°C, sufficient hydrogen is thought to diffuse during bond-annealing towards the peak of the helium distribution and react there to produce the weakest splitting plane. For Ta>500°C, loss of helium, annealing out of strain and break-up of platelets into bubble-cavities reduces the probability of thermal splitting. When hydrogen was implanted first, complete thermal splitting occurred for Ta⩽300°C but was inconsistent at 400°C and above.