      Rationale For Chromium Compounds From Wool Fiberglass Manufacturing
We are regulating chromium compounds as a `brick issue' in the NESHAP.  Although the 1999 rule used PM as a surrogate for HAP metals, we believe certain forms of chromium compounds (chromic oxides) have volatile characteristics in the furnace process and are emitted more readily than the other HAP metals present in the PM of the furnace emissions.  Therefore we question whether PM is an appropriate surrogate for the HAP chromium compounds from this source category, but we have no reason to believe that PM is not an appropriate surrogate for other HAP metals from furnace processes. We are therefore proposing emission limits for chromium compounds in the RTR.
Furthermore, we make no distinction between furnaces at area sources and those at major sources.  We are therefore proposing to regulate furnace emissions at wool fiberglass area sources in the same way that we regulate furnaces at major sources in this source category.  We are listing the area source category "wool fiberglass manufacturing"  for regulation under this proposed rule.
The risk assessment concluded that the MIR  for wool fiberglass is 40-in-one million.  We recognize uncertainties inherent in the risk assessment methodology in the preamble to the proposed rule. When risks are between 1-in one-million and 100-in-one million, we generally consider other factors in our determination of whether the risks are acceptable and present an ample margin of safety.  Because glass furnaces must be rebuilt every 7-10 years, and because one facility emits 840 lb of hexavalent chromium per year due to furnace construction using high chrome refractories,  we examined what the risks would be if all wool fiberglass furnaces rebuilt their furnaces using high chrome refractories and as a result emitted at this level. We found under this hypothetical scenario that the risks would be 900 in one million from one facility and over 100 in one million from 7 other facilities.  
We believe industry has a strong economic incentive to rebuild glass furnaces using high chrome refractories because this practice extends furnace life almost two-fold.  Therefore, a furnace constructed of high chrome refractories would have to be rebuilt every 14-20 years instead of 7-10 years.  The cost of rebuilding a furnace lies mostly in the labor to rebuild, and not so much in the materials themselves.  This practice costs and additional 15% over the use of conventional refractories for furnace rebuilds.
Moreover, because most of the industry has phased out the use of formaldehyde-based binders in their manufacturing processes, sources that were once major are now area sources.  We have some information indicating that all but 2 facilities are area sources.  While we encourage and applaud the `green' choice across the industry to phase out sources of HAP, we are also concerned that in the near future, furnaces at both area and major sources may choose to rebuild their furnaces using high chrome refractories.  Although this does not change the risk assessment results in the RTR, the hypothetical scenario and the economic incentives presented by high chrome refractories inform our finding of a threat of adverse health effect for the area source finding.  
We do not fully understand the properties and all of the different types of chrome compounds that may be emitted from wool fiberglass furnaces at this time. We plan to issue 114 letters to the wool fiberglass industry to gather additional information on refractory composition and complete chrome and hexavalent chromium emission testing.  
