5.4.3.2. Perfluorocarbons

PFCs, fully fluorinated hydrocarbons, have extremely long atmospheric lifetimes (2600 to 50,000 years) and particularly high radiative forcing (Table 5-7). The production of aluminum is thought to be the largest source of PFCs (CF₄ , and C₂F₆) emissions. These emissions are generated, primarily, by the anode effect, which occurs during the reduction of alumina (aluminum oxide) in the primary smelting process as alumina concentrations become too low in the smelter. Under these conditions, the electrolysis cell voltage increases sharply to a level sufficient for bath electrolysis to replace alumina electrolysis. This causes substantial energy loss and the release of fluorine, which reacts with carbon to form CF₄ and C₂F₆.

In 1990, the total annual global primary aluminum production was 19.4 Mt. Secondary aluminum production from recycling accounted for 21.5% of the total consumption in 1990. The production statistics from the World Bureau of Metal Statistics (1997) show that the total aluminum production was 27.5 Mt, and recycling has increased to 25.6%, or by about 3.5 percentage points, in 10 years.

The scenarios developed by Fenhann (2000) adopt a methodology of projecting future aluminum demand based on:

• Aluminum consumption elasticity with respect to GDP.
• Use of alternative assumptions concerning recycling rates.
• Varying emission factors to reflect future technological change.

These assumptions are altered to be in consistent with the four SRES scenario storylines described in Chapter 4.

For instance, in Fenhann (2000) the aluminum consumption elasticity varies between 0.8 and 0.96, and the range of increases in aluminum recycling rates varies between 1.5 and 3.5 percentage points per decade. The PFC emission factor varies according to the aluminum production technology used. The default emission factor from the Revised IPCC Guidelines (IPCC, 1997) is 1.4 kgCF₄ /t aluminum. However, Harnisch (1999) gives evidence that the average specific emissions of CF₄ per ton of aluminum has decreased from about 1.0 kg to 0.5 kg between 1985 and 1995. Accordingly, an emissions factor of 0.8 kgCF₄/t was used for 1990 and this was assumed to decrease to 0.5 kg CF₄/t in the future. This is also in agreement with the value of 0.51 kgCF₄/t recommended by the IPCC Expert Meeting on Good Practices in Inventory Preparation for Industrial Processes and the New Gases (January 1999, Washington, DC). The same sources also agree on an emission factor for C₂F₆ that is 10 times lower than that for CF₄ . This assumption was also used in the calculations presented here (Table 5-8).

Aluminum production is being upgraded from highly inefficient smelters and practices to reduce the frequency and duration of the anode effect. Since aluminum smelters are large consumers of energy, the costs of these modifications are offset by savings in energy costs and are therefore assumed to occur in all scenarios. The ultimate reduction of the anode effect frequency and duration was assumed to reach the same level in all the SRES scenarios. However, scenarios vary with respect to the rate of introducing the underlying modifications. It is technically possible to reduce the anode emissions by a factor of 10 (EU, 1997). This technically feasible reduction can be achieved by changing from the Söderberg cells currently in use to more modern pre-bake cells. It is assumed that this will happen in the A1 and B1 family scenarios, in which specific emissions of 0.15 kgCF₄/t are achieved by 2040 in the OECD90 region and by 2090 in the other regions. In the A2 and B2 family scenarios the same specific emissions are achieved later in the century in the OECD90 region and not until after 2100 in the other regions.

PFCs are consumed in small amounts in such sectors as electronics (tracers), cosmetics, and medical applications. However, the only emissions included in Fenhann (2000) beyond aluminum production were PFCs (as CF₄) from semiconductor production. In all SRES scenarios the emission estimates used are those given by Harnisch et al. (1999) of 0.3 kt CF₄ per year in 1990, 1.1 ktCF₄ in 2000, 1.0 ktCF₄ in 2010, and constant thereafter. The use of these estimates reflects the voluntary agreement, in April 1999, of the World Semiconductor Council, which represent manufacturers from Europe, Japan, Korea , and the US, among others. According to this agreement, manufacturers have adopted the emission reduction target for PFCs of 10% absolute reduction from 1995 emission levels by 2010. This target encompasses over 90% of the total semiconductor production (WMO/UNEP, 1999). The total PFC emissions in the four SRES scenario families cover a range from 24 to 97 kt PFC in 2100 (Table 5-8; Fenhann, 2000).

5.4.3.3. Sulfur hexafluoride

SF₆ is an extremely stable atmospheric trace gas. All studies concur that this gas is entirely anthropogenic. Its unique physico-chemical properties make SF₆ ideally suited for many specialized industrial applications. Its 100-year GWP of 23,900 is the highest of any atmospheric trace gas. In 1994, atmospheric concentrations of SF₆ were reported to rise by 6.9% per year, which is equivalent to annual emissions of 5,800t SF₆ (Maiss et al., 1996).

According to several sources (Kroeze, 1995; Maiss et al., 1996; Victor and MacDonald, 1998), about 80% of SF₆ emissions originate from its use as an insulator in high-voltage electrical equipment. The remaining 20% of the present global SF₆ emissions (1200 tons per year) are emitted from magnesium foundries, in which SF₆ is used to prevent oxidation of molten magnesium. The global annual production of magnesium is about 350,000 tons (US Geological Survey, 1998), and developing countries account for about 15% of the total. SF₆ is also used to de-gas aluminum, but since SF₆ reacts with aluminum, little or no atmospheric emissions result from this process.

Major manufacturers of SF₆ agreed voluntarily to co-operate on the compilation of worldwide SF₆ sales data by end-use markets. Six companies from the US (three), Japan, Italy, and Germany participated in the data survey. The companies do not expect the total sales for magnesium foundries to increase before 2000 (Science & Policy Services Inc., 1997). Based on this information, the 1996 statistical production values were used for the year 2000 in the formulation of the scenarios reported in Fenhann (2000). Future production was projected assuming the same consumption elasticity to GDP as for aluminum (see discussion above). In 1996, about 41% of the world magnesium was produced in the US; of this, only 16% was processed in foundries for casting that resulted in emissions of SF₆ (Victor and MacDonald, 1998). Since the distribution of world foundry capacity appears to be roughly similar to that of world magnesium production, Fenhann (2000) assumes that, presently, 16% of the produced magnesium is processed in foundries across all regions. Relating this amount of the processed magnesium to the aforementioned emission of 1200t SF₆ per year yields an emission factor of 21kg SF₆ per ton of magnesium processed in foundries. The demand for magnesium in automotive applications as a strong lightweight replacement for steel is growing quickly. Hence, it is expected that the fraction of total magnesium production processed in foundries by 2050 will grow to between two to three times the present level.

As mentioned above, no less than 80% of SF₆ emissions (or 4600 tons of SF₆ per year at present) originate from the use of SF₆ as a gaseous insulator in high-voltage electrical equipment. The unique ability of SF₆ to quench electric arcs has enabled the development of safe, reliable gas-insulated high-voltage breakers, substations, transformers, and transmission lines. The demand for such electrical equipment is assumed to grow proportionally to electricity demand (Victor and MacDonald, 1998; Fenhann, 2000) with an emission factor of 132.6t SF₆/EJ electricity. Fenhann (2000) used preliminary electricity generation projections from the four SRES marker scenarios and assumed additional various other potentials for emission reductions that result from more careful handling, recovery, recycling, and substitution of SF₆ . Reduction rates vary in the different SRES scenario storylines; the detailed assumptions are reported in Fenhann (2000). The SF₆ emissions for the four scenarios given in Fenhann (2000) range from 7 to 25kt SF₆ in 2100. The main driver is electricity consumption, since the bulk of emissions originate from electric power transmission (for transformers).

SF₆ is also emitted from other minor sources, but for the purposes of this report it is assumed that uncertainty ranges factored into the alternative scenario formulations cover the emissions from these sources.

IPCC Homepage