The study was carried out from December 2008 to October 2009 in collaboration with a Project Expert Group consisting of selected experts with many years of experience in the non-ferrous metal sector in Europe (see annex III to this report).

1. In 2006, the EU non-ferrous metal industry consumed quantities of SF6 with a global warming potential of 3 million t CO2 equivalent: 1.8 million t CO2 eq. in the production of secondary aluminium and 1.2 million t CO2 eq. in the magnesium casting and recycling sector. While in the magnesium industry SF6 consumption is considered equal to SF6 emissions, only 1.5% of the applied SF6 quantities are considered emissions in the aluminium industry. The total global warming emissions of SF6 from the NF metal industry thus amounted to 1.22 million t CO2 eq. (thereof 0.04 million t CO2 eq. from the aluminium sector). The use prohibition under Art 8(1) of the F-Gas Regulation should, as of 1 January 2008, have already eliminated 0.7 million t CO2 eq. in magnesium die casting (foundries with SF6 consumption > 850 kg/a). The remaining sectors of the magnesium industry include die casting companies with SF6 consumption < 850 kg/a, sand casting foundries, and recycling plants. The total amount of SF6 consumed/emitted was 0.5 million t CO2 eq.

2. In the magnesium industry (die casting, sand casting, recycling), the surface of the hot metal melt must be protected against oxidation by cover substances. The gases SO2 and SF6 allow good protection up to temperatures of 800°C of the liquid metal. SF6 has a high global warming potential (GWP) of 22,800, and the chemical industry developed substitutes containing fluorine with GWP substantially lower than that of SF6. In the EU, HFC-134a became the most accepted new alternative to SF6, aside from the well-established cover gas SO2. Another new-developed cover gas, FK 5-1-12 (Novec-612^TM), started being used in industrial applications in the USA and in Japan in 2008.

3. In the aluminium sector, only one smelter in Europe is using SF6, though not as a cover gas but as a degassing agent to help eliminate impurities from the molten metal in the production of one special alloy. In 2008, the SF6 consumption of this plant amounted to 100 t/a. The operator claimed more than 98.5% decomposition of the gas in the hot melt, thus only the remaining 1.5% is released un-destroyed to the atmosphere. Waste gas measurement carried out in the course of this study confirmed the 1.5% emission factor.

4. A survey on the use of SF6 as cover gas in the EU magnesium industry and as degassing agent in the EU aluminium industry showed that 19 of the overall 53 magnesium die casting foundries used SF6 in quantities < 850 kg/a, causing emissions of 135 kt CO2 equivalent in 2008. In magnesium sand casting, SF6 is still the gas commonly used because of the extremely high melting temperatures and the open operation which require extra stable and non-toxic cover gas. In sand casting SF6 emissions were 228 kt CO2 eq in 2008, thus higher than emissions from die casting. In magnesium recycling, the mostly used cover gas for normal die casting alloys is SO2 with only one recycling plant still applying SF6 for normal die casting alloys in a quantity of 3,000 kg/a. Another recycler, who relies on HFC-134a for normal die casting alloys, uses also SF6 (3,000 kg) for the special alloys for which extremely high melt temperatures are needed. Thus, total SF6 emissions from recycling amounted to 137 kt CO2 eq. in 2008. In the production of one special aluminium alloy, the European plant, mentioned above, uses undiluted SF6 as degassing agent of the melt. Their SF6 emissions amounted to 3,000 kg (68 kt CO2 equivalent) in 2008.

5. There are several policy options, both regulatory and voluntary, to reduce cover gas emissions. In addition to the option of no action (business as usual), options of containment/recovery, partial or full prohibition of use, voluntary agreements, joint implementation mechanism come into question. Screening the options for technical feasibility reveals that containment and recovery, which were the basic options followed by the F-Gas Regulation for the main F-Gas using sectors (stationary refrigeration, air conditioning, fire protection etc), are not feasible solutions in this sector. This is because the foundry equipment of the EU non-ferrous metal industry already represents state-of-the-art technology, and further containment measures would not be effective. From a technical point of view, the only possible choice to further reduce emissions is substitution of SF6, i.e. by conversion of the plants to use cover gases SO2 or HFC-134a in die casting and recycling of die casting alloys. In sand casting and recycling of special magnesium alloys, the melting temperature is too high for the use of HFC-134a, whereas SO2 cannot be applied in such open application either because of its toxicity. The third alternative cover gas, the fluorinated ketone FK 5-1-12 (Novec-612^TM), showed promising results at high temperatures in laboratory trials but is not yet commercially available in Europe . As a consequence, at present, the replacement of SF6 is not feasible in sand casting and recycling of special magnesium alloys. This is also true for the use of SF6as a degassing agent in secondary aluminium production.

6. The conversion of magnesium die casting foundries subject to the 2006 F-Gas Regulation has proved that SF6 replacement by one of the two available alternative cover gases is technically feasible in this application. The lessons learned are that conversion to SO2 requires extensive technical restructuring of the gas delivery system because of the toxicity and corrosiveness of the gas. When changing to HFC-134a, normally the existing gas delivery system can be re-used, and only a few adjustments must be made (at gas mixing station and furnaces). In exceptional cases, however, it can be necessary to stabilise the surface of the melt additionally by means of special devices (converters) for ingot feeding. In these cases, conversion to HFC-134a is more complex and costly than conversion to SO2.

Therefore, the study is discussing and refining only policy options that are based on replacing SF6 as a cover gas in die casting and in recycling of normal die casting alloys.

7. The following seven policy options are qualified for substitution of SF6 by conversion to SO2 or HFC-134a and are selected for more in-depth analysis of their environmental, economical and social impacts.

Option 1: No policy action for magnesium and aluminium industry

Option 2: Full SF6 prohibition in magnesium die casting

Option 3: Revision of the 850 kg/a threshold in Mg die casting (reduction to 100 kg/a)

Option 4: Joint implementation mechanism for magnesium die casting

Option 5: Full prohibition of SF6 in recycling of magnesium die casting alloys

Option 6: Voluntary agreement to replace SF6 in recycling of die casting alloys

Option 7: Joint implementation mechanism for recycling of die casting alloys.

In die casting, the technical choices HFC-134a or SO2 exist within each option. Operators will have to decide for conversion to either one of these two cover gases. In recycling, only the technical solution SO2 is considered as the operator rejects conversion to HFC-134a as an alternative to SF6.

8. Assessment of the environmental impacts of the policy options reveals several aspects:

• Without political action, SF6 emissions from the NF metal industry will increase at a growth rate of 1% per year from 570 to 640 kt CO2 equivalent by 2020.

• All reduction options have the potential to cut SF6 emissions at a range of 212 to 229 kt CO2 eq. by 2020. This amount translates into about one third of the no-action SF6 emissions from the NF metal industry in 2020.

• The reduction potential achieved by substituting SF6 with SO2 is higher compared to HFC-134a. The use of HFC-134a (GWP=1,430) would also create global warming emissions in the range of 6% of SF₆ emissions.

• All options for SF6 replacement cause emissions of acidic waste gas (SO2, HF). As these emissions range below the legal threshold for waste gas concentration limits, considering the high environmental benefit of the replacement SF6for climate protection, they are considered acceptable.

9. Based on technical experiences gained by die casting foundries subject to the 2006 F-Gas Regulation within the conversion, the study estimates the annual additional costs (vs. SF6) arising from conversion to HFC-134a or SO2 in the 19 die casting foundries which consume < 850 kg SF6 per year, and from conversion to SO2 in the only recycling plant of die casting alloys. Conversion to SO2 generally requires new equipment so that the investment costs are high compared to conversion to HFC-134a. In contrast, the gas costs are lower when using SO2. Assessment of the economic and social impacts shows:

• The additional annual costs to be paid by the plants range from 0.0 to 0.5% of their annual turnover from magnesium products. This financial burden is considered acceptable for the industry.
• As a consequence, the expenses for new equipment, gas, and the license fee in case of HFC-134a do not cause job risks in the plants, but.
• New employment positions at equipment manufacturers, gas distributors, and the license holder would not be created either.

10. All options analysed are consistent with the EU policy on fluorinated greenhouse gases, and do not imply notable economic and social trade-offs. The specific abatement costs for the analysed reduction options vary between a minimum of -0.39 €/t CO2 eq. and a maximum of only 0.91 €/t CO2. The absolute abatement costs are very low in each case and the options do not significantly differ in efficiency (cost effectiveness).

The options are equivalent in coherence and cost effectiveness. Reliability of one particular option to reduce emissions hence becomes the key criterion for ranking. In this perspective, the non-regulatory policy options are ruled out. Full ban of the use of SF6 in both die casting and recycling of die casting alloys are the most effective policy options. The study recommends these two reduction options to policy makers.

11. With regards to technical choices for implementing those options, HFC-134a solutions are shown to be less effective in emission reduction than the application of SO2. The GWP of HFC-134a is relatively high (1,430), the potential of the HFC-134a solution to reduce emissions hence decreases. The relatively high costs for conversion result from the fee that users have to pay to the license holder of HFC-134a.

Sensitivity analysis reveals that the disadvantage of HFC-134a in emission reduction considerably decreases if the destruction of the gas over the melt is taken into account. When taking into account decomposition during the process, the GWP of HFC-134a after use would effectively drop to around 400. Furthermore, the disadvantage in annual costs significantly is balanced if the license fee is lowered. The conversion to HFC-134a would be more cost effective than conversion to SO2 if the license fee was reduced to half of the amount currently charged per tonne of magnesium produced. In contrast, it cannot be excluded that the propensity of operators of large magnesium foundries to choose HFC-134a as technical alternative to SF6 is lowered as a consequence of increasing prices of HFC-134a.