MEMORANDUM
 
DATE:	July 1, 2023

FROM:	Donna Lee Jones, Katie Boaggio, Kevin McGinn, and Ned Shappley, U.S. Environmental Protection Agency; and Gabrielle Raymond, RTI International

TO:		Coke Oven Batteries NESHAP (40 CFR Part 63, subpart L) Project File

SUBJECT:	 Fugitive Monitoring at Coke Oven Facilities
--------------------------------------------------------------------------------
      
      
1.0	INTRODUCTION 

This memorandum describes the fugitive testing and results as part of the technology review of the National Emission Standards for Hazardous Air Pollutants (NESHAP) for Coke Oven Batteries (40 CFR part 63, subpart L). 
The NESHAP for Coke Oven Batteries (COB) was promulgated on October 27, 1993 (58 FR 57898) at 40 CFR part 63, subpart L, and set maximum achievable control technology (MACT) standards for leaks of coke oven emissions (COE) from oven doors, lids, and offtakes as determined by measurements of visible emissions (VE) made according to EPA Method 303; and for COE and PM emissions from charging. The Coke Oven Batteries NESHAP limits the number of seconds of VE during charging coke into ovens at coke facilities; and the percent leaking doors, lids, and offtakes. Coke facilities were required by the 1990 Clean Air Act (CAA) amendments (CAA § 112(i)(8)) to choose to be regulated under one of two regulatory tracks: either a MACT track or lowest achievable emissions rate (LAER) track, by 1993 (58 FR 57898). The 1993 COB NESHAP established increasing stringent standards for doors, lids, offtakes, and charging in subsequent years. The residual risk and technology review (RTR) for COB, subpart L MACT track, was promulgated on April 15, 2005 (72 FR 1992). The RTR for COB subpart L LAER track was to be done by 2020 but was not done at that time. Only one of the remaining nine by-product coke oven facilities is a MACT track facility (Cleveland Cliffs in Middletown, OH). The other eight by-product facilities are on the LAER track. Four of the five existing HNR coke oven facilities are on the MACT track with only one, SC-Vansant-VA, on the LAER track. The docket ID for Coke Oven Batteries NESHAP is EPA-HQ-OAR-2003-0051 (subpart L).

According to section 112(d)(6) of the CAA, the EPA is required to perform a technology review of MACT standards every eight years. This memorandum describes the fugitive testing at coke oven facilities as part of the technology review of Coke Ovens Batteries NESHAP (40 CFR part 63, subpart L). Fenceline fugitive standards are being proposed as a result of this review. Included here are summaries of the coke facilities' fugitive data. Detailed information about the fugitive data is included in the appendices along with Excel files used to make the emissions and associated calculations. 



                                   CONTENTS
1.0	INTRODUCTION	1
2.0	BACKGROUND	2
3.0	CAA SECTION 114 REQUESTS	3
4.0	CALCULATIONS AND DATA	6
4.1	Calculated Fenceline Data	6
4.2	Range in Average Concentrations, High to Low	7
5.0	RESULTS OF COKE FUGITIVE MEASUREMENTS 	7
6.0	FENCELINE TESTING COSTS	8
APPENDIX A	A-1
TO-13A & 15A CHEMICAL LISTS	1
APPENDIX B	B-1
MONITORING PLANS, TEST PLANS, AND QAPPS	1
APPENDIX C	C-1
FACILITY FUGITIVE DATA	1
APPENDIX D	D-1
FACILITY FUGITIVE DATA CALCULATIONS	1
APPENDIX E	E-1
VISUAL SUMMARY OF LOCATIONS OF MEASURED FUGITIVE CONCENTRATIONS	1

                                       
2.0	BACKGROUND
There are 14 coke plants in the U.S., owned by six companies, located in 8 states. Nine of the 14 facilities are by-product (ByP) facilities and five are HNR facilities, where four of the five HNR facilities recover heat from the process. Table 1 lists the 14 coke oven facilities, their owners and locations, acronyms used in this memorandum for each facility, coke and facility type, and typical annual production.
In the coke production process, coal is heated in ovens in the absence of air to temperatures approaching 2,000°F (1,100[o]C), which drives off most of the volatile organic constituents of the coal, as gases and vapor, to form coke, which consists almost entirely of carbon. Coal is charged into the top or end of the oven, depending on the process, and the finished coke is pushed out using doors on both ends of the ovens. The hot coke product is cooled in quench towers that use water spray to cool hot coke in rail cars that drive into the bottoms of quench towers. The coke product is used either in blast furnaces in the integrated iron and steel industry (II&S) or at iron and steel foundries that produce iron molds for shaping molten metal. Coke oven temperatures are slightly higher to produce blast furnace coke than foundry coke. Coking lasts for 15 to 18 hours to produce blast furnace coke in the ByP process, 24 to 48 hours in the heat and nonrecovery recovery (HNR) process, and 25 to 30 hours to produce foundry coke in ByP process. 

The coke ovens are arranged in a series of adjacent ovens called batteries, where one battery may have up to 100 ovens and multiple batteries may be used by one facility. For the 14 coke facilities in the Coke Oven 
                         Table 1. Coke Oven Facilities
                                 Facility Name
                                Parent Company
                                     City 
                                     State
                                 Coke Process
                              Currently Operating
ABC
ABC Coke
Tarrant
                                      AL
                                      ByP
                                      Yes
BLU 
Bluestone 
Birmingham
                                      AL
                                      ByP
                                      No
CC-BurnsHarbor
Cleveland-Cliffs
Burns Harbor
                                      IN
                                      ByP
                                      Yes
CC-Follansbee
Cleveland-Cliffs 
Follansbee
                                      WV
                                      ByP
                                      No
CC-Middletown
Cleveland-Cliffs 
Middletown
                                      OH
                                      ByP
                                      No
CC-Monessen
Cleveland-Cliffs 
Monessen
                                      PA
                                      ByP
                                      Yes
CC-Warren
Cleveland-Cliffs 
Warren
                                      OH
                                      ByP
                                      Yes
DTE/EES
DTE Vantage
Detroit 
                                      MI
                                      ByP
                                      Yes
SC-Indiana Harbor
SunCoke Energy
East Chicago
                                      IN
                                      HNR
                                      Yes
SC-Haverhill
SunCoke Energy
Franklin Furnace
                                      OH
                                      HNR
                                      Yes
SC-Granite City
SunCoke Energy
Granite City
                                      IL
                                      HNR
                                      Yes
SC-Middletown
SunCoke Energy
Middletown
                                      OH
                                      HNR
                                      Yes
SC-Vansant
SunCoke Energy
Vansant
                                      VA
                                      HNR
                                      Yes
USS-Clairton
United States Steel
Clairton
                                      PA
                                      ByP
                                      Yes


Risk Modeling Database, the number of batteries per facility ranged from 1 to 10, and the total number of ovens at the facility ranged from 56 to 708, with median values of 3 batteries per facility and 126 ovens total at the facility. Ovens at ByP facilities are tall and narrow (approximately 20 ft/6 m by 1.3 ft/0.4 m; whereas, at HNR facilities, the ovens are short and wide (approximately 7.8 ft/2.4 m by 12 ft/3.6 m). ByP ovens are operated at positive pressure, whereas HNR ovens operate under negative pressure. 
The organic gases and vapors that evolve from the ovens are removed through an exhaust system: for by-product (ByP) plants, gases are sent to a chemical recovery plant that recovers chemicals and other by-products for sale and also cleans the coke oven gas (and is regulated under 40 CFR part 61, subpart L NESHAP); for nonrecovery plants (i.e., no chemical recovery) with heat recovery, the oven gases are sent to a heat recovery steam generator that produces power for sale and also to units that perform gas cleaning; for nonrecovery plants without heat recovery, oven exhaust gases are released to the air through waste heat stacks.

3.0	CAA SECTION 114 REQUESTS

The EPA sent a CAA section 114 request in June 2022 for fugitive testing from five facilities: ABC, CC-BurnsHarbor, DTE/EES, USS-Clairton and SC-Haverhill. The five selected coke facilities for fugitive monitoring were required to perform sampling and analysis according to the methods specified in Table 2, at 3 months and 6 months from issue of the CAA section 114 request, to determine the facility fugitive emissions at fenceline and interior facility grounds. 
      Sampling by EPA Methods 325A and B (40 CFR Part 63, Appendix A) was required for benzene, toluene, ethylbenzene, xylenes, and 1,3 butadiene along the facility fenceline. The number and location of sampling sites was to be determined according to the specifications in section 8.2 of EPA Method 325A. The facilities performed thirteen 14-day integrated sampling episodes at each monitoring location, for a combined total of 182 days of sampling with EPA Method 325A. The collected samples were analyzed according to EPA Method 325B.

      In addition, selected facilities sampled at the fenceline with Compendium Methods TO-13A and TO-15A for VOC and PAHs, respectively, according to the specifications listed in Table 2. When siting the TO 
      Table 2. Fugitive Emission Testing and Analysis: Fenceline/Interior
                              Recommended Method
                           Number Tests and Location
                               Duration of Tests
                                    Timing
                                  To Analyze
                               Units of Measure
EPA Methods 325A/B
Number and location as determined by Section 8.2 of EPA Method 325A for ByP and HNR
Thirteen 14-day long sampling episodes (24 hours per day), 182 days total testing each location

BTEX
1,3 Butadiene
ug/dscm ppbv
EPA TO-15A 
Four locations (minimum) ByP and HNR
Upwind: 
  One (minimum) upwind ByP and HNR
Two downwind: 
  Downwind of batteries: 
     One for ByP, two for HNR 
  Downwind of CBRP: 
     One for ByP 
One or two within interior grounds: 
  One near CBRP for ByP 
  One in centroid space near batteries for both ByP and HNR
Seven 24-hour long samples, each location


Each 24-hour long sampling episode occurred on an approximately bi-weekly basis, within the 2-week period of the EPA Method 325B sampling
VOC analyte list in Table 1-1 of EPA Method TO-15A[.] (See Appendix A)
ug/dscm ppbv
EPA TO-13A[4]
Three locations (minimum):
Upwind: 
  One minimum both ByP and HNR
Two downwind: 
  Downwind of batteries: 
      One for ByP, two for HNR 
  Downwind of CBRP: 
      One for ByP One or two within interior grounds: 
      One near CBRP for ByP 
      One in centroid space near batteries for both ByP and HNR
Seven 24-hour long samples, each location


Each 24-hour long sampling episode occurred on an approximately bi-weekly basis, within the 2-week period of the EPA Method 325A/B sampling 
Analyte list in Table 1 of EPA Method TO-13A (See Appendix A)

Note: Samples were analyzed using GC/MS
ug/dscm ppbv


monitors, facilities considered the seasonal prevailing winds so that at least one monitor was placed at an upwind sampling location. If there was potential for upwind off-site contribution, facilities could choose to select additional upwind sampling locations. [Note: TO-13A and TO-15A sampling locations were not required to be collocated with an EPA Method 325A/B sampling location.] Facilities were required to perform seven 24-hour long samples at each TO monitor location for a total of at least 21 samples (3x7) for TO-13A and at least 28 samples (4x7) of TO-15A. 
      
      For TO-15A, facilities were required to sample at a minimum of four monitoring locations at the fenceline including at least one upwind location, as discussed above. At least one monitor was to be downwind of the batteries for both ByP and HNR. For ByP facilities, at least one of the facility fenceline sampling locations was required to be located immediately downwind (based on seasonal prevailing winds) of the coke byproduct chemical recovery plant (CBRP).
      
      For TO-13A, at least three fenceline locations were required to be tested including at least one upwind location, as discussed above. At least one of the three fenceline sampling locations were placed downwind from the batteries for both ByP and HNR. For ByP facilities, one monitor was placed downwind from the tar processing at the CBRP. 
      
      In addition to the fenceline monitoring, sampling at the interior of the facility was required with TO-13A and 15A at one or two locations, depending on the type of facility, as follows: (1) one monitor at the centroid of the CBRP for ByP facilities, and (2) one monitor at the center of the area next to the coke batteries for both ByP and HNR facilities. Facilities were required to perform seven 24-hour long samples at each of the two designated locations, for a total of 7 samples each for TO-13A and TO-15A for HNR and 14 samples (2x7) each for TO-13A and TO-15A for ByP at the interior facility locations (chemical plant for ByP and coke battery(ies) for both HNR and ByP). The interior sampling was designed to potentially identify the apportionment of compounds measured at the fenceline. After sending out the CAA request for fugitive testing, it was determined that interior measurements at the HNR facility was not warranted, because the apportionment would be solely due to the coke oven batteries since there is no CBRP.
      
      Each 24-hour long sampling period with the TO-13A and TO-15A methods was required to occur on an approximately bi-weekly basis and within the 2-week period of the EPA Method 325A/B samples. The summa canister inlets of the TO methods were located 1.5 to 3.0 meters above ground using a pole or other secure structure. The TO method canister sampling flowrate was to be maintained nominally at a constant flowrate during the sample period.
      
      Facilities were required to deploy a meteorological station (or use an existing available station) consistent with the requirements in Section 8.3 of EPA Method 325A of Appendix A of 40 CFR part 63. The meteorological data, including wind speed, wind direction, temperature, and barometric pressure on an hourly basis for each sampling episode were also submitted with the data. 
      
      The facilities were required to submit a "Fenceline/Interior Monitoring Plan" to EPA that included the planned sampling locations for approval prior to conducting any of the required measurements. The "Fenceline/Interior Monitoring Plan" depicted the following: (1) The facility's significant emission sources of benzene, if known; (2) a description of each significant emission source containing enough information so that the EPA would be able to properly determine if the emission sources potential impact to the fenceline emissions are such that the facility's proposed monitoring locations were appropriate; (3) each EPA Methods 325A/B sampling location(s); (4) each TO-13A and TO-15A sampling location(s); and (5) the location of the meteorological station. A "Test Plan" was developed by all facilities that included a simple facility schematic showing the coke batteries and major components of the CBRP, as applicable, that included, at minimum, the tar production and refining units, light oil production and refining units, cooling units, sulfur recovery/ desulfurization combustion unit, ammonium sulfate production units, napthalene processing units, waste units, and any other units included in the CBRP and a "Quality Assurance Procedure Plan (QAPP)" that ensures the quality of the data being produced. The Fenceline/Interior Monitoring Plans, Test Plans, and QAPPs for the five facilities providing fugitive data are in Appendix B. 
      
      The four selected ByP facilities were required to report the fenceline/interior monitoring data to the Agency at 3-months and 6-months from the date the EPA approved each facility's monitoring plan . The selected HNR facility was required to report the fenceline monitoring data to the Agency at nine weeks from the date the EPA approved the facility's monitoring plan. All meta data from all five facilities also were submitted that were needed by the EPA to calculate fugitive emissions. 
      
4.0	CALCULATIONS AND DATA 

4.1	Calculated Fenceline Data

Before calculating the "delta C" (Δc), i.e., the difference between the highest and lowest measurements at the facility during the test period of the data received, a check was made that the below the method detection limit (BDL) values were properly handled by the reporting facilities for both benzene and naphthalene. The reported data for each measuring period and method at each monitor were evaluated as BDL using the following method, which was used in the fenceline monitoring requirements in the NESHAP for Petroleum Refineries (MACT) in 40 CFR part 63, subpart CC (sections 63.658(f)(1)(i)(A)&(B)): 
 If the lowest detected value of benzene/naphthalene is BDL, the owner or operator should use zero as the lowest sample result when calculating Δc. 
 If all sample results are BDL, the owner or operator should use the method detection limit as the highest sample result and zero as the lowest sample result when calculating Δc. 
This same data procedures were used for the four by-product coke facilities, which provided detection limits: ABC; CC-Burns Harbor, DTE/EES, and USS Clairton. For the heat recovery facility (SC-Haverhill), any value which was BDL was set to zero because these data already had been obtained before the CAA section 114 request. The test data, with the BDL changes for benzene and naphthalene data only, are located in Appendix C. 
After accounting for the BDL values, the Δc for each facility across all fenceline (only) locations for each measuring period and method was calculated for both benzene and napthalene. This resulted in one Δc per fenceline measurement period (3 or 6 months), method, and chemical. These Δc values are discussed below in Section 5.0. Appendix D shows the calculations for each facility and the results of the calculations, with individual files for each facility and a summary file that includes all facilities.
The four by-product facilities also provided interior measurements for benzene and naphthalene. The maximum value across all interior locations at each facility for each measuring period and method was calculated for both benzene and napthalene. Haverhill was not required to conduct interior measurements because of the absence of a byproduct chemical recovery facility; therefore, any fenceline concentration apportionment at Haverhill only would include the coke batteries. Appendix D also shows these calculations for each facility.
Table 3. Range in Benzene and Napthalene Data for Coke Ovens Facility's Fenceline Test Results 
    (Average Concentrations, No Δc Correction)  -  6 Months Data 2022-2023
                                 Coke Facility
            Lowest Benzene Fenceline Average Concentration (ug/m3)
           Highest Benzene Fenceline Average Concentration (ug/m3)
          Lowest Napthalene Fenceline Average Concentration (ug/m3)
          Highest Napthalene Fenceline Average Concentration (ug/m3)
ABC
                                     0.67
                                      13
                                     0.29
                                      1.0
CC-Burns Harbor
                                     0.48
                                      2.5
                                     0.15
                                     0.77
DTE/EES
                                      0.7
                                      3.1
                                     0.16
                                     0.57
SC-Haverhill
                                     0.44
                                     0.62
                                      0.0
                                     0.44
USS Clairton
                                      2.4
                                      32
                                     0.47
                                      9.8
  Note: Data from SC-Haverhill reflects only 3 months of data.


4.2	Range in Average Concentrations, High to Low
      The highest and lowest (average) fenceline and interior concentrations at each facility also were examined. See Table 3 for this comparison. For facilities located on a river (i.e., USS Clairton, DTE/EES, and CC-Burns Harbor), the highest concentrations were found along the river. The second highest value of 13 ug/m[3] at ABC was located approximately 300 meters from the nearest neighborhood south of the facility. The benzene 
and naphthalene concentrations (as averages) in the fenceline and interior data were plotted on maps to provide a visual summary of these data and are shown in Appendix E.

5.0	RESULTS OF COKE FUGITIVE MEASUREMENTS

A summary of the overall average Δc fugitive concentrations for the both the three and six-months of data after the 2022 CAA section 114 request was issued is shown in Table 4. Average benzene Δc concentrations ranged from 0.1 to 40 ug/m[3] and average naphthalene Δc concentrations ranged from 0.0 to 12 ug/m[3] at the fenceline. Inside the facility grounds, the measured concentrations (not Δc) were much higher than the fenceline values, with benzene concentrations inside facility grounds as high as 910 ug/m[3] and, for naphthalene, as high as 100 ug/m[3]. 

6.0	FENCELINE TESTING COSTS

For sampling and analysis of benzene only, a cost of $100 per tube was estimated, where the cost of the tube, including analysis, is approximately $75 and the costs associated with setting-up and recovering the tube, and reviewing and reporting the data are $25. Typically, the sorbent housing is covered by the analytical laboratory and their costs are relatively minimal.
The EPA Method 325A requires a minimum of 12 sampling locations, but depending on facility size and complexity, more sample locations may be required. To be conservative in the estimate of costs, 24 sample locations were used. 
Total Annual Cost Per Monitor Site (Facility) 
Tubes ($100 x 24 locations)		=	$2,400 per 14-day sampling period
26 x 14-day sampling periods		 = 	$62,400 annual cost ($2023) per facility


Table 4. Summary of 2022-2023 Results for Coke Ovens Fugitive Fenceline Average (Δc) and Interior Maximum Concentration Results for Benzene and Napthalene 
                              Method and Location
                                    DTE/EES
                                 USS Clairton
                                      ABC
                               SunCoke-Haverhill
                               CC Burns Harbor:

                                 Naphth-alene
                                    Benzene
                                 Naphth-alene
                                    Benzene
                                 Naphth-alene
                                    Benzene
                                 Naphth-alene
                                    Benzene
                                 Naphth-alene
                                    Benzene
                                 Fenceline[a]
325A&B[b]
                                       
                                       2
                                       
                                      40
                                       
                                      15
                                       
                                      0.1
                                       
                                       3
325A&B[c]
                                       
                                       4
                                       
                                      28
                                       
                                      17
                                       
                                      NA
                                       
                                       3
6-mo avg[d]
                                       
                                       3
                                       
                                      33
                                       
                                      16
                                       
                                      NA
                                       
                                       3
                                       
TO-15A[b] 
                                      0.4
                                       1
                                       9
                                      35
                                       1
                                       8
                                     0.00
                                      0.2
                                      1.1
                                       1
TO-13A[b] 
                                      0.6
                                       
                                      12
                                       
                                       1
                                       
                                     0.01
                                       
                                      0.4
                                       
                                   Interior
TO-15A[b]
                                      52
                                      49
                                      204
                                      620
                                       9
                                      100
                                      NA
                                      NA
                                      480
                                      910
TO-13A[b,][e]
                                      48
                                       
                                      100
                                       
                                      10
                                       
                                      NA
                                      NA
                                      97
                                       
[a] Difference the between measured value and background concentration, i.e., lowest value measured. This difference is called the "delta C" (Δc). See Appendices C and D for the data used and calculations made.
[b] Reflects 3 months of data collected in 2022.
[c] Reflects 3 months of data collected in 2023.
[d] Reflects 6 months of data collected from 2022 to 2023.
[e] Highest value across all 2022 sampling.


To account for any first year planning costs, total facility costs for testing were rounded up to $70,000 per facility, inclusive of analytical and labor costs. An additional cost of $36,745 for recordkeeping and reporting for each facility was estimated bringing the total to $106,74 per facility for fenceline testing. See the memorandum Coke Ovens Risk and Technology Review: Compliance Costs, for details of the cost analysis.



                                  APPENDIX A
                        TO-13A & 15A CHEMICAL LISTS
                                       
                                    [pdfs ]
                                       
                                       
                                       
                                  APPENDIX B
                    MONITORING PLANS, TEST PLANS, AND QAPPS

                      [individual facility plans, QAPPs]
                                       
                                       
                                       
                                  APPENDIX C
                            FACILITY FUGITIVE DATA 

                             [excel data file(s)] 

                                       
                                  APPENDIX D
                     FACILITY FUGITIVE DATA CALCULATIONS 

                 [python data file(s) and excel data file(s)] 
                                       
                                  APPENDIX E 
       VISUAL SUMMARY OF LOCATIONS OF MEASURED FUGITIVE CONCENTRATIONS 
                   (Missing units in top value in section c)

                                       
                             Top left (a): DTE/EES
Top middle (b): ABC
Top right (c): USS Clairton
Bottom left (d): CC Burns Harbor
Bottom right (e): SunCoke-Haverhill
Top left (a): DTE/EES
Top middle (b): ABC
Top right (c): USS Clairton
Bottom left (d): CC Burns Harbor
Bottom right (e): SunCoke-Haverhill
e)
e)
d)
d)
b)
b)
c)
c)
a)
a)

