Decreasing the Release of PFCs

As of February 1998, 22 semiconductor manufacturers had signed a voluntary agreement with the U.S. Environmental Protection Agency to reduce the total amount of global warming compounds emitted into the atmosphere. The goal of the Memorandum of Understanding is to emit less than 1995 levels by the year 2000 -- a substantial undertaking as emissions tend to increase by 25% annually. A recent meeting in Kyoto, Japan, led to an agreement by the United States to reduce overall greenhouse gas emissions by the year 2012 to at least 7% below 1990 levels of 0.2 million metric tons of carbon equivalent. Table 1 shows descriptions of PFCs used in semiconductor processing along with their global warming potentials (effects relative to an equivalent mass of CO₂ ).

Economic data indicate that abatement may prove best for small PFCusers(<7500 lb/year (3500 kg/yr) PFCs purchased), capture addresses the needs of medium-sized users (7500-15,000 lb/yr (3500-7000 kg/yr)) and recycling is the best solution for large semiconductor fabs purchasing >15,000 pounds of PFCs per year.

Abatement, capture and recycling

Abatement techniques include catalytic, combustion, microwave and plasma technologies. Concerns associated with point-of-use abatement techniques include costs to purchase and maintain the equipment, costs to treat the additional fluoride waste generated by the systems, significant water requirement and needs to additionally treat byproducts such as CO₂, CH₄ , HF and NO_x. For these reasons, POU abatement may only prove economically feasible for small-scale users.

To allow high capture efficiencies and maximum concentration of greenhouse gases, PFC exhaust must be pretreated to remove particles and water-soluble components. Capture techniques include carbon adsorption, cryogenic absorption, membrane technologies and pressure swing adsorption. Under optimized conditions, membrane technology has enabled 95% recovery and 95% nitrogen rejection ratio for CF₄ and C₂F₆. Unfortunately, other chemicals typically present in the exhaust stream, such as CHF₃ and NF₃, are not recovered at the same efficiencies because of the nature of the molecules. In addition, the concentration of the product mix, particularly if it is high in nitrogen content, will increase disposal costs.

The most promising solution for reducing PFC emissions in large fabs is gas recycling. At SEMI's conference on PFC emissions in July 1997, Mike Mocella of DuPont reported the feasibility of recycling C₂F₆ at a cost commensurate with the composition of the captured PFC mixture provided.

SEMATECH is currently evaluating a state-of-the-art cryogenic capture and recycling facility. Its performance relies on the effectiveness of the pretreatment system to remove contaminants using, in this case, co-current and countercurrent water scrubbers with caustic injection to remove COF₂ (carbonyl fluoride), SiF₄ (silicon tetrafluoride), TEOS and corrosive materials. Molecular sieve absorber columns are used to remove CO₂ and H₂O. The PFCs are then continuously absorbed in a wash fluid while the diluent gases (primarily nitrogen) are vented. The wash liquid is transferred to a stripping column that separates the cryogenic wash fluid from the PFCs. The PFC mixture is stored in a tank until it reaches sufficient quantity to separate the gases using a distillation column. Purification techniques are designed to yield gases of starting material quality, which will be tested, bottled in cylinders and reused.

Following the evaluation of the cryogenic systems' ability to pretreat the PFC waste will be SEMATECH's evaluation of capture and recovery efficiencies. The advanced system is expected to have >99% capture efficiency and >99% concentration efficiency, which will qualify it for proposed tax incentives in the United States for recycling systems.

The elimination of PFC use altogether depends on the use of alternative chemistries for etching oxide and nitride films from wafers and process chambers. Hoover explained that "some materials, like ClF₃, provide clear-cut benefits. In fact, Praxair is actively responding to a great deal of interest for that product." Other alternatives include hydrofluorocarbons, iodofluorocarbons, hydrofluoroethers, trifluoroacetic anhydrides, chlorine trifluoride, dilute NF₃ and perfluoropropane (C₃F₈). "It often takes many years to gain broad acceptance of processes employing new alternative materials, and the variables that must be considered are complex," Hoover said.