Keith Christman - Sun Aug 7 10:43:35 2005

The Vinyl Institute strongly disagrees with this approach to estimating a mercury emissions factor for PVC production. As we pointed out in our comments on the draft report, mercury cell chlorine is very rarely used to produce PVC in the US.

There are a number of problems with the approach used here. First, this approach allocates all the mercury emissions to the chlorine co-product from the chlor-alkali process and none of it to the sodium hydroxide (caustic) or hydrogen. Certainly in an LCA, some of the mercury emissions should be allocated to the sodium hydroxide and hydrogen co-products on either a mass basis or economic valuation basis. A strong argument can be made to allocating half or more mercury emissions to the caustic because mercury cell technology was developed to produce higher purity “rayon grade” sodium hydroxide. Chlorine from mercury cell technology is essentially the same as chlorine from other technologies but caustic from mercury cell is a preferred product required by some industries. The Task Group should allocate less than half of the mercury emissions to chlorine.

Second and more importantly, the approach used here falsely assumes that Ethylene Dichloride (EDC) is produced from unknown or generic chlorine technology. This assumption is contradicted by the fact that EDC is produced in the US from captive co-located chlorine sources. As discussed in VI’s comments on the draft report, we know which plants use mercury cell produced chlorine to produce EDC in the United States. Thus, rather than assuming that PVC is produced using generic chlorine with an average amount coming from mercury cell processes, as is done in the data request, the Task Group should use TRI mercury emissions data from the one mercury cell chlor-alkali plant that actually is used for manufacturing EDC. VI provided a report from CMAI indicating that only one mercury cell chlorine plant is co-located with an EDC plant. (Chlorine and EDC facilities have been Co-located to virtually eliminate the very high costs of transporting large volumes of chlorine to an EDC facility) This is a PPG facility in Lake Charles, Louisiana. Furthermore, PPG has announced plans to “eliminate use of mercury at Lake Charles, LA” in 2007 with the installation of membrane technology to replace this capacity. http://biz.yahoo.com/bw/050804/45754.html?.v=1.

If the Task Group feels that it must develop a mercury emissions factor for vinyl, we recommend that the Task Group use the Mercury emissions from TRI for this PPG Lake Charles plant to estimate a mercury emissions factor for vinyl production. This one plant released 1228 pounds of mercury and mercury compounds to the open environment in 2003. Allocating half of this mercury to chlorine (and half to caustic) results in 614 pounds of mercury allocated to chlorine at this facility. According to CMAI, 30 percent of chlorine produced at the site is used for EDC production and chlorine on site is blended with chlorine produced from other processes. Therefore, only 30% of the mercury allocated to chlorine at the site should be allocated to chlorine used in vinyl. Utilizing the TRI data for this facility and market estimates from CMAI for the facility, a more accurate emissions factor for mercury associated with vinyl production is 0.01 grams of mercury per tonne of PVC produced in the United States. (Please note that past TRI reports utilized mercury emissions estimates derived from a generic EPA formula for estimating fugitive mercury cell emissions. Preliminary data from tests of new equipment providing real time monitoring of mercury cell fugitive emissions indicates that actual emissions are substantially lower than estimates derived from the generic EPA formula. We have not adjusted the emissions factor for this new information. Thus, the estimated emissions factor is likely to be higher than actual emissions)

VI will post the spreadsheet showing these emissions calculations and the CMAI report showing that chlorine from only one mercury cell plant is likely to be used in EDC and PVC production.

If the Task Group feels it must develop a mercury emissions factor for vinyl production, the Task Group should also develop mercury emissions factors for alternative materials which will result in mercury emissions to the environment. Alternative materials will result in mercury emissions through their use of chlorine and caustic derivatives (generic emissions factors developed by the Task Group would be more appropriate) in their processes and from their energy use including coal combustion and their use of electricity derived from coal combustion.

The Task Group may also be interested in the Chlorine Institute Report to EPA on Mercury which shows total TRI emissions declining to 6.8 tons in 2004.

Tom Lent - Sun Aug 7 19:34:03 2005

The use of TRI data to calculate mercury emissions does not account for the problem that the Task Group clearly identified in the original report of large discrepancies between TRI reporting and actual mercury purchases. See our response to the Dioxin EF inquiry for more on the limitations of a TRI based analysis.