Richard E. Dunn, Director Air Protection Branch 4244 International Parkway Suite 120 Atlanta, Georgia 30354 404-363-7000 MEMORANDUM January 6, 2020 To: Thru: From: Subject: James Boylan Byeong-Uk Kim Henian Zhang Modeling Analysis for Ethylene Oxide - UPDATE Sterilization Services of Georgia, Atlanta, Fulton County, GA On December 20, 2019, Sterilization Services of Georgia submitted updated model input parameters (location, height, diameter, and exit velocity for the dry bed reactor stack) to the Georgia Environmental Protection Division. The updated stack parameters were reviewed and deemed appropriate. This memorandum contains updates to the modeling memorandum dated December 19, 2019 (All changes made to the December 19 memorandum are italicized and bolded). GENERAL INFORMATION Air dispersion modeling of ethylene oxide was conducted by the Georgia Environmental Protection Division (GA EPD) to assess the impacts of ethylene oxide emissions from sources at Sterilization Services of Georgia (hereafter SSG) on ambient air surrounding the facility. Although this modeling is not required for issuance of an air quality permit, GA EPD followed the procedures described in GA EPD’s Guideline for Ambient Impact Assessment of Toxic Air Pollutant Emissions1 (hereafter “Georgia Air Toxics Guideline”). Computer models are used to predict the concentrations of toxic air pollutants (TAPs) being analyzed using facility information provided by the source and other information developed by GA EPD staff. The modeling results are compared to the 15-min, 24-hour, and annual Acceptable Ambient Concentrations (AACs). GA EPD’s 15-min and 24-hour AACs are derived from Occupational Safety and Health Administration (OSHA) permissible exposure limits. GA EPD’s annual AACs are derived from U.S. EPA’s risk values which are found in EPA’s Integrated Risk Information System (IRIS) database. Appendix B contains detailed calculations for the 15-min, 24-hour, and annual ethylene oxide AACs. GA EPD uses AACs as a screening tool to ensure that public health is protected. No further evaluation is needed if the modeled concentrations are below the corresponding AACs. If the modeled concentration is above the AAC, GA EPD requires the company to consider a reduction in pollutant emission rates, additional controls, and/or an increase in stack heights, followed by a site-specific risk assessment. After performing a site-specific risk assessment, if it is infeasible for the applicant to comply with the AAC, the Director at his/her discretion may approve control technology which reflects the maximum degree of reduction in emissions of hazardous air pollutants that the Director determines is achievable by the source, provided that such control technology is no less effective than the level of emission control which is achieved in practice by the best controlled similar source. 1 https://epd.georgia.gov/air-protection-branch-technical-guidance-0/toxic-impact-assessment-guideline This memo discusses modeling results and the input data used to perform the ethylene oxide dispersion modeling. Emissions for the current scenario and a proposed scenario that includes additional controls were modeled. The current scenario models the impact of SSG’s emissions prior to installation of the back vent controls. The proposed scenario models the impact of SSG’s emissions after the back vent controls are installed. The back vent controls are required to be installed and operational on or before December 31, 2019 by the amended permit that was issued on November 7, 2019. With the current scenario, the modeled maximum ground-level concentration (MGLC) for the 15-min averaging period was below its corresponding AAC. However, the MGLC for the 24-hour averaging period and the modeled annual averaged ground-level concentrations across the 5-year period (AAGLCs) at the three closest residential areas exceeded their corresponding AACs. With the proposed scenario, the MGLCs for the 15-min and 24-hour averaging periods were below their corresponding AACs. However, the modeled AAGLC at one of the three closest residential areas exceeded the annual AAC. The results are summarized in the following sections of this memorandum. INPUT DATA 1. Meteorological Data – Hourly meteorological data (2014 to 2018)2 were generated by GA EPD. Surface measurements were obtained from the Hartsfield-Jackson Atlanta International Airport, Atlanta, GA. Upper air observations were obtained from the Atlanta Regional Airport – Falcon Field, Peachtree City, GA. These measurements were processed using the AERSURFACE (v13016), AERMINUTE (v15272), and AERMET (v18081) with the adjusted surface friction velocity option (ADJ_U*). 2. Source Data – Emission release parameters and emission rates for the current and proposed scenarios were provided by SSG and reviewed by the GA EPD Stationary Source Permitting Program (see Tables A1 and A2 of Appendix A for details). In the proposed scenario, the new dry bed outlet stack is expected to be located at (723866 E, 3734447 N) with a diameter of 0.508 m, height of 20.32 meters, exit temperature of 325 K, and exit velocity of 13.97 m/s, per email communication with SSG on December 20, 2019. 3. Receptor Locations – Discrete receptors with 50-meter intervals were placed along the property line. For the proposed scenario, receptors extend outwards from the property line at 100-meter intervals on a Cartesian grid to approximately 3 km and at 250-meter intervals to approximately 6 km. This domain (approximately 12 km by 12 km) is sufficient to capture the maximum impact from the proposed scenario. For the current scenario, additional receptors were added to the proposed scenario domain at 1500-meter intervals from approximately 6 km to approximately 16 km. This domain (approximately 33 km by 33 km) is sufficient to capture all concentrations above 0.00033 g/m3. Additional receptors were placed at the three closest residential areas. All receptor locations are represented in the Universal Transverse Mercator (UTM) projections, Zone 16, North American Datum 1983. 4. Terrain Elevation – Topography was found to be generally flat in the site vicinity. Terrain data from the USGS 1-sec National Elevation Dataset (NED) were extracted to obtain the elevations of all sources, buildings, and receptors by the AERMAP terrain processor (v18081). 5. Building Downwash – The potential effect for building downwash was evaluated via the “Good Engineering Practice (GEP)” stack height analysis and was based on the building parameters submitted by SSG (Table A3 in Appendix A) using the BPIPPRM program (v04274). The BPIPPRM model 2 https://epd.georgia.gov/air-protection-branch-technical-guidance-0/air-quality-modeling/georgia-aermet-meteorological-data 2 was used to derive building dimensions for the downwash assessment and the assessment of cavityregion concentrations. AIR TOXICS ASSESSMENT The impacts of facility-wide ethylene oxide emissions were evaluated according to the Georgia Air Toxics Guideline. The 15-min, 24-hour, and annual AACs were reviewed based on OSHA Permissible Exposure Limit (PEL), OSHA Total Weight Average (TWA) PEL, and U.S. EPA IRIS Risk Based Air Concentration (RBAC) according to the Georgia Air Toxics Guideline. For this assessment, GA EPD used the annual AAC derived according to the Georgia Air Toxics Guideline (see Appendix B for details). The EPA’s 2014 National Air Toxic Assessment (NATA) used a higher annual AAC value (see Appendix C for details). The modeled 1-hour, 24-hour, and annual ground-level concentrations were calculated using the AERMOD dispersion model (v19191). Table 1 summarizes the MGLCs and the AAC levels for the two scenarios. The 15-min MGLC is based on the 1-hour MGLC multiplied by a factor of 1.32. The 15-min MGLCs were below the corresponding AAC for both scenarios. The 24-hour MGLC exceeded the 24-hour AAC for the current scenario but did not exceed the 24-hour AAC for the proposed scenario. The annual MGLCs exceeded the corresponding AAC with both scenarios. Figure 1 shows the spatial distribution of the AAGLCs for the current scenario. Figure 2 shows a close-up of Figure 1 with the closest three residential areas labeled (R1, R2, and R3). R1, R2, and R3 represent the closest residential home within a group of homes or subdivisions. Figure 3 shows the spatial distribution of the AAGLCs for the proposed scenario with the closest three residential areas labeled. Table 2 contains the AAGLCs for the current and proposed scenarios at the three closest residential areas (R1, R2, and R3). For the current scenario, R1, R2, and R3 are all above the annual AAC. For the proposed scenario, only R1 is above the annual AAC. Table 1. Modeled MGLCs for the Current and Proposed Scenarios and their Respective AACs. MGLC (g/m3) MGLC (g/m3) Averaging Period AAC (g/m3) Current Scenario Proposed Scenario Annual 0.59 0.0094 0.00033 24-hour 3.54 0.069 1.43 15-min 12.25 0.24 900 Table 2. Risk Analysis for Residential Areas with Modeled AAGLCs for the Current and Proposed Scenarios. Receptor Ratio of Ratio of UTM Zone:16 AAGLC AAGLC Modeled Modeled AAGLC* AAGLC* Residen (g/m3) to (g/m3) to Averaging AAC 3 3 tial AAC AAC (g/m ) (g/m ) Period (g/m3) Easting Northing 3 Areas Current Proposed (g/m ) (g/m3) (meter) (meter) Scenario Scenario Current Proposed Scenario Scenario R1 724,111.95 3,734,192.72 0.06645 0.0014 R2 723,267.63 3,732,490.42 0.00094 0.00002 R3 722,627.75 3,734,805.43 0.00732 0.00016 Annual 0.00033 *AAGLC is the annual averaged ground-level concentration across the 5-year period. 3 201.4 4.2 2.8 0.06 22.2 0.5 Figure 1. Contours of modeled annual averaged ground-level concentrations across the 5-year period (in g/m3) for the current scenario overlaid on a Google Earth map. 4 Figure 2. A close-up of Figure 1 with the closest residential areas labeled (R1, R2, and R3). Figure 3. Contours of modeled annual averaged ground-level concentrations across the 5-year period (in g/m3) for the proposed scenario overlaid on a Google Earth map with the closest residential areas labeled (R1, R2, and R3). 5 CONCLUSIONS The dispersion modeling analyses for ethylene oxide show exceedances of the annual AACs with the current and proposed scenarios. Site-specific risk assessments show that the modeled annual average ground-level concentrations across the 5-year period are above the annual AAC at the three closest residential areas (2.8 – 201.4 times) with the current scenario and above the annual AAC at one residential area (4.2 times) with the proposed scenario. For the current scenario, the modeled 15-min maximum ground-level concentration did not exceed its respective AAC, but the modeled 24-hour maximum ground-level concentration did exceed its respective AAC. For the proposed scenario, the modeled 15min and 24-hour maximum ground-level concentrations did not exceed their respective AACs. 6 Appendix A Emissions and Model Input Parameters Table A1. Emission Rates Current Scenario (Total Emission 1339.5 lbs/yr) Chamber 1 Use average lb/hr EO to Vacuum Pumps 95% Scrubber Inlet average lb/hr 5.752 14.379 EO to Aeration Chamber 2 Use average lb/hr 4% Scubber Efficiency 99.99% 3.632 EO to Back Vents 1% Chamber 3 Use average lb/hr Oxidizer Inlet average lb/hr 0.605 5.752 Oxidizer Efficiency 99.98% Emission Point Stack #1 EP1 Stack #2 EP2 Stack #3 EP3 Stack #4 EP4 Stack #5 EP5 lb/hr 0.0575 0.0363 0.0575 0.00144 0.000121 Proposed Scenario (Total Emission 26.9 lbs/yr) Chamber 1 Use average lb/hr EO to Vacuum Pumps 95% Scrubber Inlet average lb/hr 5.752 14.379 EO to Aeration Chamber 2 Use average lb/hr 4% Scubber Efficiency 99.99% 3.632 EO to Back Vents 1% Chamber 3 Use average lb/hr Oxidizer Inlet average lb/hr 0.605 5.752 Oxidizer Efficiency 99.98% Dry Bed System Inlet average lb/hr 0.151 Dry Bed Efficiency Emission Point Stack #1 EP1 Stack #4 EP4 Stack #5 EP5 lb/hr 0.00151 0.00144 0.000121 99.0% 8 Table A2. Stack Parameters Oxidizer Emission Points A1:K29Outlet Stack Datum Coordinates (SE Corner of Facility) Source ID Emission Type EP1 Point EP2 Point EP3 Point EP4 Point EP5 Point Latitude: o 33 43' 33" X (m) (from datum) Y (m) (from datum) Stack Base Elevation (m) Stack Height (m) Temperature (K) Exit Velocity (m/s) Diameter (m) -23.9 54.7 234.7 22.86 322.039 11.843 0.39 -22.6 57.1 234.7 22.86 322.039 11.843 0.39 -23.5 55.5 234.7 22.86 322.039 11.843 0.39 Scrubber 1 -23 56.3 234.7 22.86 294.261 43.465 0.076 Oxidizer Outlet Stack -14.6 78.7 234.7 13.716 388.706 9.521 0.762 X - Length (m) 10.1 8.4 6.1 123.4 Y - Length (m) 16.4 11.9 16.4 49.4 Angle of Rotation (deg) -28.5 -28.5 -28.5 -28.5 Description Chamber 1 Back Vent Stack Chamber 2 Back Vent Stack Chamber 3 Back Vent Stack Building Description (All are rectangluar sections) Appendage 1 Appendage 2 Appendage 3 Main Building Longitude: o 84 35' 02" Base Elevation (m) 234.7 234.7 234.7 234.7 Refereance Point (SW Corner) Y (m) X (m) (from datum) (from datum) 67.2 -20.2 62.3 -11.3 55.2 1.8 58.9 -108.5 Height (m) 6.25 6.25 6.25 7.32 Table A3. Locations of Buildings, Stacks, and Fence Line The following are spatial definitions of buildings, stacks, and the fence line. All locations are represented in the UTM projections, Zone 16. Main Building Easting SW 723757.5 NW 723781.8 NE 723889.3 SE 723865.0 Northing 3734441.8 3734484.8 3734424.0 3734381.0 Chamber 1 Room Appendage SW 723867.2 3734436.5 NW 723874.6 3734449.6 NE 723878.9 3734447.1 SE 723871.5 3734434.1 Scrubber Room Appendage SW 723851.8 3734445.2 NW 723857.5 3734455.2 NE 723864.7 3734451.1 SE 723859.1 3734441.1 Chamber 2 and Chamber 3 Room Appendage SW 723844.2 3734449.6 NW 723852.0 3734463.4 NE 723859.6 3734459.1 SE 723851.8 3734445.2 Chamber 1 Back Vent Stack 723840.0 3734434.7 Chamber 3 Back Vent Stack 723840.4 3734435.4 Scrubber Stack 723840.8 3734436.1 Chamber 1 Back Vent Stack 723841.2 3734436.8 Oxidizer Stack 723849.5 3734460.5 Lot (Fence) Line SW 723741.8 NW 723788.4 NE 723908.3 SE 723864.0 3734432.8 3734515.0 3734442.3 3734364.0 Appendix B GA EPD Calculation of the 15-min, 24-hour, and Annual AACs for Ethylene Oxide 11 GA EPD Calculation of the 15-min, 24-hour, and Annual AACs for Ethylene Oxide According to the GA EPD’s Guideline for Ambient Impact Assessment of Toxic Air Pollutant Emissions, the 15-min, 24-hour, and annual AACs for ethylene oxide are calculated as following: 15-min AAC The OSHA 15-min permissible exposure limit (PEL) for ethylene oxide is 5 ppm. To convert the PEL from ppm to mg/m3, the following conversion formula from the guidance is used: (5 ppm × 44.05 g/mol) / (24.45 L/mol) = 9 mg/m3 where, 44.05 is the molecular weight for ethylene oxide and 24.45 is the molar volume at 25oC and 760 mmHg. After applying a safety factor of 10 for acute sensory irritants, the 15-min AAC is calculated as: 15-min AAC = (9 mg/m3 × 1,000 μg/mg) / 10 (safety factor) 15-min AAC = 900 μg/m3 24-hour AAC The OSHA 8-hour Time Weighted Average (TWA) PEL for ethylene oxide is 1 ppm. To convert the TWA PEL from ppm to mg/m3, the following conversion formula from the guidance is used: (1 ppm × 44.05 g/mol) / (24.45 L/mol) = 1.8 mg/m3 where, 44.05 is the molecular weight for ethylene oxide and 24.45 is the molar volume at 25oC and 760 mmHg. After converting the 8-hour average weekly exposure to a 24-hour average weekly exposure and applying a safety factor of 300 for known human carcinogens, the 24-hour AAC is calculated as: 24-hour AAC = 1.8 mg/m3 × 1,000 g/mg × (8 hours/day × 5 days/week) 300 (safety factor) × (24 hours/day × 7 days/week) 24-hour AAC = 1.43 μg/m3 Annual AAC In the EPA Integrated Risk Information System (IRIS)3, the Inhalation Unit Risk (IUR) for ethylene oxide is 3×10-3 per µg/m3. Since ethylene oxide is carcinogenic to humans, it belongs to Group A4 with a cancer risk of 1/1,000,000. Therefore, the annual AAC is calculated as: Annual AAC = Cancer Risk / IUR = (1/1,000,000)/(0.003/μg/m3) Annual AAC = 0.00033 μg/m3 3 https://cfpub.epa.gov/ncea/iris/iris_documents/documents/subst/1025_summary.pdf https://www.epa.gov/fera/risk-assessment-carcinogenic-effects 4 12 Appendix C EPA Calculation of the Annual AAC for Ethylene Oxide 13 EPA Calculation of the Annual AAC for Ethylene Oxide According to EPA’s IRIS, inhalation unit risk (IUR) for ethylene oxide (EtO) is 3×10-3 per µg/m3 (as discussed in Appendix B). However, because of the elevated risk due to the mutagenic mode of action through early-life exposures, EPA multiplied the IUR by 1.6: Modified IUR for EtO = 3×10-3 per µg/m3 × 1.6 = 0.005/μg/m3 EPA’s NATA used (100/1,000,000) individual risk for the purpose of determining “acceptable risk” (AR) in their national assessment. AR Exposure Concentration = Cancer Risk / IUR = (100/1,000,000)/(0.005/μg/m3) = 0.02 μg/m3 14