Laserdyne Labs has developed several processes to apply laser technology to manufacturing tasks typically conducted by using mechanical equipment. In the emerging SOI (Silicon on Insulator) field, a present technology (Smart-Cut®) uses a layer of Hydrogen implanted in the silicon lattice to generate a splitting region between host and handle wafer. The current process is using thermal annealing in an oven. Based on our previous experience in this field we are suggesting utilizing localized laser heat instead of the broad overall heating of the bonded wafers in an oven. First tests have shown that the process is feasible and is not restricted by geometric boundaries. Cleave planes as produced by the laser can be pieced together without any shift in the definition of the cleavage plane. Furthermore, first principle calculations suggest that the higher energy flux rate of the laser can be utilized to reduce the necessary Hydrogen dose, as well as provides more flexibility in terms of various dopant levels. This proposed process would result in lowering the costs of SOI wafers through our simple, more cost-effective process flow. The cleaved wafer surface will be smooth and can be used directly without any post-cleave mechanical polishing or edge treatment. Anticipated Benefits/Commercial Applications: Our work in SOI processing technology is aimed at helping drive down the cost of SOI production so significantly that it will become a mainstream wafer technology. For many years the semiconductor industry has used SOI wafers for specialized applications such as chips used in military hardware and in equipment for use in space. Our invention will be directed to an improvement of the Smart-Cut technology, which uses a high dose of Hydrogen implants to generate a splitting region on the bonded Silicon through formation of hydrogen bubbles during a thermal annealing cycle. The Smart-Cut process uses high dose implantation of light ions (usually Hydrogen or Helium). In a thermal annealing cycle gas bubbles are formed, which grow together and eventually create sufficient local stress to break off the entire implanted layer at a depth close to the implant depth. The necessary Hydrogen dose stands in direct relation with the cost and number of implanter equipment. The Hydrogen diffusion throughout the lattice is the limiting step in the traditional approach. As the higher flux rate of the laser can drastically improve the timing in the diffusion step, a wider range of exfoliation agents can be utilized, in reduced concentrations. Furthermore, a side benefit of the high flux density is that the gas filled void is typically smaller than in the conventional approach, reducing the surface roughness. It might be that a further improvement here can eliminate or reduce the polishing step after exfoliation. Smart Cut Process Despite the cost reductions realized through utilization of the Smart Cut Process, the process is still relatively expensive, with a high waste factor. Opportunities exist to generate value through further cost reductions and improvement in yields. One of the main cost drivers is still the machinery needed to perform the H+ implantation. If the H+ implant dose could be reduced below the current process minimum of 5 x 1016 ions/cm2, the throughput of the ion implantation devices could be increased proportionally, generating significant cost savings for SOI wafer manufacturers. Currently an ion implanter can process 20 wafers per hour at a dose of 5 x 1016. If Laserdyne Labs’ process is able to reduce the required H+ dose to 1 e16, the throughput would increase by a factor of five per machine, reducing the number of implanters required to meet demand to roughly 60 from 290 in the year 2008. This would represent a capital equipment savings of about $920 million to the industry. A second possible benefit of applying laser technology to wafer cleaving may be in the improvement of the as split surface roughness, now approximately 100 angstroms. This surface roughness is too great and must be reduced before the wafer can be used. The wafer requires chemical mechanical processing (CMP, which smoothes the wafer to approximately five angstroms of surface roughness. Although this is a simple process, it does reduce the overall yield. Eliminating the CMP step would reduce the cost of manufacturing an SOI wafer. A leading micro-chip manufacturer* is interested in funding Laserdyne Labs for research and development if the Phase I feasibility is approved. In preparation for Phase II, Laserdyne Labs will utilize funds from Fast Track matching corporate investments. The commitment for Fast Track investment from Central Florida Innovation Corporation (CFIC) and Capital Partners of Brevard are contingent upon Phase II award.
Keywords: SOI, Laser, Dose reduction, Dopant levels, Cleaving, Annealing
