MEMORANDUM


To:		OSWRO Docket (EPA-HQ-OAR-2012-0360)
From:		Paula Hirtz, Refining and Chemicals Group 
      Terri Hollingsworth, Air Toxics Assessment Group
      Office of Air Quality Planning and Standards, 
      U.S. Environmental Protection Agency 
CC:		Penny Lassiter, Leader, Refining and Chemicals Group 
      Kelly Rimer, Leader, Air Toxics Assessment Group	
Date:		January 9, 2015
Subject:	Additional Facilities Identified During the OSWRO Comment Period
--------------------------------------------------------------------------------


The EPA received information during the comment period on the proposed amendments to the Off-Site Waste and Recovery Operations (OSWRO) National Emission Standards for Hazardous Air Pollutants (NESHAP) identifying four sources that are subject to the OSWRO NESHAP but were not included in the risk assessment supporting the proposal. These facilities are E. I. DuPont de Nemours and Company in Orange, Texas; E. I. DuPont de Nemours and Company in Axis, Alabama; Eastman Chemical Company in Longview, Texas; and Eastman Chemical Company in Kingsport, Tennessee. Three of the facilities manufacture chemicals and chemical intermediates, and the fourth facility located in Axis, Alabama manufactures pesticides. We also reviewed the Title V permits for these facilities, and we confirmed they are subject to OSWRO NESHAP and will therefore be affected by the new requirements of the OSWRO NESHAP amendments.

In evaluating and developing standards under section 112(f)(2), the EPA applies a two-step process to address residual risk. In the first step, we determine whether risks are acceptable, and in the second step of the process, we consider whether the emission standards provide an ample margin of safety. To determine whether to conduct additional risk modeling for the newly identified OSWRO facilities, we first considered whether risk acceptability and ample margin of safety for the OSWRO source category, including the four additional facilities, could be determined using existing information. The existing information available to the EPA includes previously conducted OSWRO source category modeling results, facility-wide modeling results conducted for these facilities in our evaluations of risks due to emissions from other source categories, Title V permits and information obtained through discussions with facility representatives. We determined that OSWRO source-category specific modeling would be conducted for these facilities if facility-wide risks were near levels considered unacceptable and OSWRO operations constituted a majority of those emissions. However, we also determined that if either the facility-wide risks were not near levels considered unacceptable or if OSWRO emissions did not constitute a majority of emissions causing risks near levels considered unacceptable, additional modeling would not be necessary. 

The EPA contacted representatives at these companies who indicated that these facilities are primarily chemical and pesticide manufacturing plants. In addition to the manufacturing operations, the three chemical manufacturing plants also process limited volumes of company-generated off-spec or contaminated product as off-site waste. Company representatives stated that these facilities, as manufacturers, do not commercially accept or process off-site waste from other companies. The pesticide manufacturing plant processes a limited amount of off-site waste with a VOHAP concentration below the OSWRO 500ppmw threshold from a single adjacent company. Email correspondence with these company representatives is included in the docket for this rulemaking. Consistent with our findings at proposal, we expect emissions from the OSWRO source category to make only a small contribution to facility-wide risks at manufacturing facilities and to contribute more significantly to facility-wide risks at facilities that commercially accept off-site waste, such as hazardous waste treatment facilities and cement manufacturing facilities. The results of modeling conducted prior to the OSWRO proposal indicated that seven of the 38 modeled facilities had an OSWRO source category contribution of 50 percent or more to facility-wide risk. Five of these facilities were hazardous waste treatment facilities and two were cement manufacturing facilities. Considering these modeling results, we expect the OSWRO contribution to facility-wide risk to be less than 50 percent at the additional facilities, which do not commercially accept off-site waste.

These four facilities also contain sources that are subject to other NESHAP.  As such, the EPA previously modeled the facility-wide risks for these four facilities as part of the risk reviews for the other NESHAP. The EPA modeled these facility-wide risks in the "Residual Risk Assessment for 7 Source Categories: Printing and Publishing, Marine Tank Vessel Loading Operations, Pharmaceuticals Production, Epichlorohydrin Elastomers, Polybutadiene Rubber, Styrene Butadiene, Nitrile Butadiene" promulgated in January 2011 and the "Residual Risk Assessment for 7 Source Categories: Pesticide Active Ingredient Production, Polyether Polyols Production, Polymers and Resins IV -- Styrene Acrylonitrile, Polymers and Resins IV  -  Polystyrene, Polymers and Resins IV  -  Polyethylene Terephthalate, Polymers and Resins IV -- Methyl Methacrylate-Butadiene Styrene, Polymers and Resins IV -- Acrylic Butadiene Styrene" promulgated in January 2014.  A review of the risk assessment results for these facilities indicate that the maximum facility-wide cancer risks due to emissions of hazardous air pollutants (HAP) range from 6-in-1 million to 40-in-1 million. These risks are relatively low when compared to the upper end of the range of acceptability of 100-in-1 million. The maximum facility-wide non-cancer risks due to HAP emissions range from 0.08 to 1. In addition, the results show that the facility-wide cancer and non-cancer risks are attributed to HAP emissions from non-OSWRO processes. As discussed above, OSWRO processes are minor operations at these facilities, and it is reasonable to conclude that risks due to OSWRO processes at these facilities would be well below these facility-wide risk values and would likely be similar to or less than the previously modeled maximum risks for the OSWRO source category. See Table 1 below for the facility-wide risk results for each facility.

To further consider whether to conduct additional risk modeling for these facilities, we compared the previously modeled facility-wide emissions, which were based on the emissions and emissions release characteristics contained in the 2005 National Emissions Inventory (NEI), to current facility-wide emissions, which are based on the 2011 NEI, for the HAP driving both cancer and non-cancer risks. The results of this comparison show that all four facilities experienced a net decrease in risk-driving HAP emissions (see Table 2 below).  Company representatives indicated that these decreases are attributable to new product raw material changes and reduction in HAP in a HAP-containing waste stream. We therefore expect the current facility-wide risks for these facilities to be lower than the previously modeled risk values presented in Table 1. This emissions comparison further supports our expectation that risks from the OSRWO source category at these facilities are well below the level that we generally deem acceptable. 

As the facility-wide risks at these four additional facilities are relatively low, the OSWRO operations are not the predominant activity at the facilities and are not expected to contribute significantly to the facility-wide risk, and the current emissions for the HAP driving cancer and non-cancer risk at each of these facilities is lower than when previously modeled, we determined that additional modeling to include these facilities is not necessary. Based on this information, we have also concluded that the risks from these four facilities do not change our decision regarding risk acceptability for the OSWRO source category. Further, it was determined at proposal that the amended NESHAP, including the proposed revisions to the tank and equipment leak standards, provides an ample margin of safety to protect public health. The modeling results for the OSWRO source category conducted prior to proposal of the NESHAP amendments showed that maximum cancer risks were 9-in-1 million and the maximum non-cancer target organ-specific hazard index (TOSHI) was 0.6, based on actual emissions. The projected OSWRO source category risks for these facilities are similar to or less than these risk values, and the information regarding the risks from these four facilities does not change our conclusions, presented for comment in the July 2, 2014 federal register notice, regarding the ample margin of safety for the OSWRO source category.    

Table 1. Facility - wide Risk for Facilities Identified during the Comment Period
                                       
                                   Facility
                                       
                                   Location
                                       
                                  Cancer MIR
                                1 in a million
                                       
                               Non-Cancer TOSHI
                                       
             Source Categories Driving Cancer and Non-Cancer Risk
                                       
                                    Source
DuPont
Orange, TX
                                     40[a]
                                      0.9
Boilers
Previous RTR modeling 
Marine Tank Vessel Loading
DuPont 
Axis, AL
                                       6
                                     0.08
Pesticide Active Ingredient
Previous RTR modeling
Pesticide Active Ingredient
Eastman Chemical Company
Longview, TX
                                      30
                                      0.7
Hazardous Organic NESHAP, Polyether Polyols, Hazardous Waste Incineration
Previous RTR modeling Polyether Polyols[2]
Eastman Chemical Company
Kingsport, TN
                                      10
                                       1
Hazardous Waste Incineration
Previous RTR modeling Polyethylene Terephthalate in Polymer & Resins IV[2]
 a As shown in Table 2, the facility-wide nickel emissions modeled in the Marine Tank Vessel Loading RTR are more than 2 orders of magnitude higher than current nickel emissions from the facility, and both the cancer and non-cancer risks from nickel would be significantly lower if additional modeling were performed.
   Table 2. Comparison of Previously Modeled Emissions vs Current Emissions
                                       
                              Facility, Location
                                       
HAP Driving Cancer Risk
                                   Emissions
                                (tons per year)
                                       
HAP Driving    Non-Cancer Risk
                                   Emissions
                                (tons per year)

                                       
                                       
                                    Modeled
                                       
                                    Current
                                   % change
                                       
                                       
                                    Modeled
                                       
                                    Current
                                   % change
DuPont,
Orange, TX
nickel
1.95
0.0034
-100
nickel
1.95
0.0034
-100
DuPont, Axis
AL
hydrazine
0.009
0
-100
chlorine
0.13
0
-100





hydrogen chloride
65.8
20.1
-69
Eastman Chemical Company, Longview, TX
naphthalene
6.01
11.0
+45
chlorine
5.81
2.47
-57

ethylene oxide 
9.93
8.25
-17
maleic anhydride
4.02
0
-100

benzene
27.4
8.89
-68
propionaldehyde
31.1
15.2
-51

1,3-butadiene
4.20
5.25
+20




Eastman Chemical Company, Kingsport, TN
arsenic 
0.359
0.331
-8
chlorine
19.5
11.2
-43

chromium (VI)
0.0973
0.0156
-84
hydrogen chloride
1200
355
-70

acetaldehyde
32.5
38.5
+16
antimony
2.81
0.0154
-99

