United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EMB Report 79-CKO-22
November 1979
Air
&EPA Coke Oven Emissions
Emission Test Report
Armco Steel
Houston, Texas
-------
COKE OVEN EMISSIONS
ARMCO, INC.
Houston, Texas
Prepared for the
U.S. Environmental Protection Agency
Emission Measurement Branch
Research Triangle Park, North Carolina 27711
Prepared and compiled by
Clayton Environmental Consultants, Inc
25711 Southfield Road
Southfield, Michigan 48075
TRW
Environmental Engineering Division
Progress Center
P.O. Box 13000
Research Triangle Park, NC 27711
-------
FOREWORD
Two firms prepared this report under contract
to the U.S. Environmental Protection Agency, there-
fore it is presented in two sections. Section I
was prepared by Clayton Environmental Consultants,
Inc., Southfield, Michigan and includes testing
results for benzene soluble organics, benzene,
and 02} CO, C02, as well as coke oven door emission
rates and visible emission data. Section II was
prepared' by TRW Energy Systems Group, Durham, North
Carolina and contains benzo (a)pyrene (B(a)P) sampling
data only, and immediately follows Appendix H of the
Clayton report.
-------
SECTION I - CLAYTON REPORT
BENZENE SOLUBLE ORGANICS STUDY
COKE OVEN DOOR LEAKS
Project No. 79-CKO-22
Contract No. 68-02-2817
Work Assignment No. 20
-------
TABLE OF CONTENTS
Page
List of Figures ±
List of Tables ii
1.0 Introduction 1
2.0 Summary and Discussion of Results 4
3.0 Process Description and Operation 12
4.0 Location of Sampling Points 13
5.0 Sampling and Analytical Procedures 18
APPENDICES
A. Project Participants
B. Field Data Sheets
B-l. Benzene Soluble Organic
Test Data Sheets
B-2. Sampling Summary Data
B-3. Visible Emission Data
Sheets
B-4. Summary of Visible Emissions
B-5. Fugitive Emission Observa-
tion Data Sheets
C. Benzene Soluble Organic Weights
by Fraction
D. GC Analysis
E. Detailed Summary of Sampling and
Analytical Procedures
E-l. Benzene Soluble Organic
E-2. Determination of Benzene
-------
E-3. Draft Method 109 (and Addendum
to Method 109)
E-4. Method 9
F. Example Calculation
G. Calibration Data
H. Field Audit Report
-------
LIST OF FIGURES
Figure Pa g e^
1.1 Elevation and plan view of pro- 3
cess/control system
A.I Inlet stack cros s-slect ion and -,
sampling location
4.2 Outlet stack cross-section and ^5
sampling location
4.3 Plan view of battery orientation 17
5.1 Benzene soluble organic sampling £i
train
5.2 Integrated bag sampling train 25
-------
LIST OF TABLES
Table Page
2.1 Benzene Soluble Organic Concentra- 5
tions and Emission Rates
2.2 Benzene Concentrations and Emission 7
Rates
2.3 Removal Efficiency of WESP 9
2.4 Summary of Fugitive Emission 10
Ob se rva tions
ii
-------
1.0 INTRODUCTION
The U.S. Environmental Protection Agency (EPA)
retained Clayton Environmental Consultants, Inc. to
determine the benzene soluble organic (BSO) fraction
of particulate, and benzene emissions from the inlet
and outlet of a Mikropul wet electrostatic precipitator
(WESP). In addition, stack visible emission and coke
oven door fugitive emission data were obtained. This
unit cleans the door leak and pushing emissions from
Battery Nos. 1 and 2 at the Armco, Inc. facility in
Hous ton, Texas.
The results of this study will be used in
research and development efforts for supporting
possible New Source Performance Standards for the iron
and steel industry, coke oven door emissions. This
study was commissioned as EMB Project No. 79-CKO-22,
Contract No. 68-02-2817, Work Assignment 20.
The testing program included the following:
(1) triplicate, simultaneous WESP inlet and
outlet samples for benzene soluble organics;
(2 ) simultaneous integrated bag samples for
benzene and Orsat analyses, at the WESP
inlet and outlet;
(3) visible emission observations (from the
WESP exhaust stack) recorded for the duration
of each BSO sample run; and,
-------
(4) visible emission observations of the coke
oven doors.
Auxiliary data included exhaust gas temperatures
and flowrates as determined from the traverses.
Figure 1.1 presents an elevation and plan view of
the process/control system as tested.
- 2 -
-------
I
u>
I
Plan view
Outlet
sampling
location
WESP
Outlet
sampling
location
WESP
ILki
Inlet
sampling
location
Elevation view
Inlet
sampling
location
Capture hood
Batteries 1 and 2
Batteries 1 and 2
Figure 1.1. Elevation and plan view of process/control system (not to scale)
-------
2.0 PRESENTATION OF RESULTS
BENZENE SOLUBLE ORGANICS
Table 2.1 presents the concentrations and
emission rates of benzene soluble organics
as determined from both the inlet and outlet of
the WESP. Concentrations are presented in grains
per dry standard cubic foot (gr/dscf) and milligrams
per dry standard cubic meter (mg/dscm). Emission
rates are expressed in pounds per hour (Ib/hr) and
kilograms per hour (kg/hr).
The flowrate data presented in Table 2.1 show
a consistent pattern of approximately 14-percent lower
flowrates at the WESP outlet than at the inlet, for
each of the three runs. The flowrates determined by
TRW, which are presented in Section II, Table 2,
also tend to corroborate a higher flowrate at the inlet
than at the outlet.
Due to the flowrate difference, a review of all
pitot tube calibrations was conducted and these calibra-
tions were compared with calibrations made throughout
the life history of these pitot tubes. No irregularities
were detected. The pitot tubes and sampling lines had
all been leak checked in-field, the alignment of the
pitot tube with respect to the gas flow was verified,
- 4 -
-------
TABLE 2.1. BENZENE SOLUBLE ORGANIC CONCENTRATIONS AND EMISSION RATES
Sample
Location
Sample
Number
Date
1979
Stack Gas
Parameters
Flowrate
ds c f m
Temp
F
Concentration
gr/dscf
mg/dscm
Emission Rate
Ib/hr
kg/hr
Inlet
Outlet
1
2
3
Average
1
2
3
Average
10-4
10-5
10-5
10-4
10-5
10-5
176,000
174,000
172,000
174,000
153,000
149,000
148,000
150,000
95.0
105.0
109.0
103.0
79.6
77.3
66.9
74.6
0.010
0.017
0.020
0.016
0.007
0.011
0.011
0.010
22.8
37.8
45.0
35.2
16.0
25.4
24.9
22.1
15.0
24.6
28.9
22.8
9.18
14.1
13.8
12.4
6.83
11.2
13.1
10.4
4.16
6.41
6.27
5.61
-------
and adequate clearance was maintained between the pitot
tube openings and the sampling nozzle to eliminate flow
disturbances. These various procedures and checks, therefore,
substantiate the validity of the data.
The data suggest air exfiltration in the system
between the inlet and outlet sampling locations which
allows gases to escape the system. During the week
of testing, the WESP was continually being serviced,
so it is not beyond the realm of possibility that the
WESP structure or ductwork was a source of out-leakage.
If the outlet flowrates were as high as the
inlet, the emission rates of BSD and benzene would
be correspondingly higher and the WESP removal effi-
ciency would be lower.
Inlet
BSD concentrations at the inlet ranged from 0.010
to 0.020 gr/dscf (22.8 to 45.0 mg/dscm) and averaged
0.016 gr/dscf (35.2 mg/dscm). Emission rates ranged
from 15.0 to 28.9 Ib/hr (6.83 to 13.1 kg/hr), averaging
22.8 Ib/hr (10.4 kg/hr).
Outlet
Outlet BSO concentrations ranged from 0.007 to 0.011
gr/dscf (16.0 to 25.4 mg/dscm) and averaged 0.010 gr/dscf
(22.1 mg/dscm). Emission rates at the outlet ranged
from 9.18 to 14.1 Ib/hr (4.16 to 6.41 kg/hr), averaging
12.4 Ib/hr (5.61 kg/hr).
- 6 -
-------
BENZENE
Table 2.2 presents the results of the benzene
analyses. Concentrations are presented in parts per
million (ppm) with emission rates in Ib/hr and kg/hr.
Inlet
Benzene concentrations at the inlet ranged from
0.7 to 2.2 ppm and averaged 1.7 ppm. Emission rates
ranged from 1.6 to 4.6 Ib/hr (0.71 to 2.1 kg/hr) and
averaged 3.5 Ib/hr (1.6 kg/hr).
Outlet
Concentrations at the outlet ranged from 0.7 to
2 . 9a ppm and averaged 1.9 ppm. Emission rates ranged
from 1.4 to 5.3a Ib/hr (0.63 to 2.4a kg/hr) and
averaged 3.5 Ib/hr (1.6 kg/hr).
EFFICIENCY
Table 2.3 presents the removal efficiency for
the wet electrostatic precipitator relative to
benzene soluble organics and benzene emissions. Removal
efficiency for BSD ranged from 38.8 to 52.2-percent
and averaged 44.6-percent. Benzene removal was 12.5-
percent during Run 1 and 19.6-percent during Run 3,
aSince the volume obtained for this bag sample was much
smaller than those of the other samples, it is suspected
that the bag was leaking and therefore, this value may
be mis lead ing.
-------
TABLE 2.2 BENZENE CONCENTRATIONS AND EMISSION RATES
Sample
Location
Sample
Number
Concentration
ppm
Emission Rate
Ib/hr
kg/hr
1
Inlet 2
3
Average
1
Outlet 2
3
Average
0.
2.
2.
I.
0.
2.
2.
1.
7
1
2
7
7
9a
1
9
1
4
4
3
1
5
3
3
.6
.4
.6
.5
.4
.3a
.7
.5
0.
2.
2.
1.
0.
2.
1.
1.
71
0
1
6
63
4a
7
6
This result determined from small air sample,
possible leaky bag.
-------
TABLE 2.3 REMOVAL EFFICIENCY OF WESP
Sample
Number
Removal Efficiency
Percent
Benzene Soluble
Organics
Benzene
38.8
12.5
42.7
52.2
19.6
Average
44.6
16.1
Not applicable}as there was more benzene
measured at the outlet than at the inlet
for this test run.
- 9 -
-------
averaging 16.1-percent. Sample 2 was not applicable for
this determination as more benzene was measured at the
outlet than at the inlet.
COKE OVEN DOOR EMISSION RATES
Table 2.4 presents a summary of the fugitive
emissions observations made during this study. The
leaking coke oven door emission rates (expressed as
a percent of the total doors) are presented for the
entire battery and both the coke side and push
side doors for each run for each observer. Also
included in this table are.the observed total number
of leaking doors which were combined over the number
of runs conducted at each site for each observer.
The emission rate was calculated as follows.
For each run, the total number of leaking oven doors
and leaking chuck doors were summed individually for
both the push side and coke side. These sums were
then divided by the total number of ovens on the
battery and then multiplied by 100 to determine the
percentage of leaking doors. To obtain the total
emission rate for the entire battery, first, the
number of leaking doors from all the runs for both
the push side and the coke side were summed. Next,
the total number of push side and coke side observation
runs was multiplied by the number of ovens in the
battery. Then the total number of leaking doors was
divided by this product.
- 10 -
-------
TABLE 2.4. SUMMARY OF FUGITIVE EMISSIONS OBSERVATIONS
Sample
Number
5SO Sample
No. 1
Observer3
1
2
3
Average
(SO Sample
No. 2
1
2
3
Average
ISO Sample
No. 3
1
2
3
Average
Totalb
No. of
Leaking
Doors
46
28
28
34
68
81
67
72
51
35
51
46
Total
Emi s s i on
Rate for
Entire
Battery
12
9.0
9.0
10
22
22
22
22
21
14
21
19
Coke Oven Door Emission Rate, Percent of Total Door
s
Run Number
1
PS CS
9.7 16
8.1
6.5
8.1 NA
18
16 29
18
17 NA
15
9.7
15 42
13 NA
2
PS CS
11
9.7 15
6.5
9.1 NA
25
19
24 32
23 NA
13 45
13
13
13 NA
3
PS CS
9.7
9.7
11 11
10 NA
16 31
19
18
18 NA
9.7
6.5 27
13
9.7 NA
4
PS CS
8.1 19
3.2
9.7
7.0 NA
23
16 31
16
18 NA
c
c
c
c
AVG
PS
13
7.7
8.4
9.7
20
18
19
19
13
9.7
14
12
CS
18
NA
NA
NA
NA
30
NA
NA
NA
NA
NA
NA
c*
The observers are as follows: (1) J. Breger; (2) A. Baecker; (3) D. Lazarevic.
These values are the combined number of leaking doors from all the runs per observer.
Due to darkness, Run 4 was not conducted.
NA - Not applicable.
-------
The coke-side door leak emissions were captured
by the shed system and ducted into the WESP. These
leaks, therefore, were the emissions quantitatively
measured for BSO and benzene.
From Table 2.4, it is evident that the coke side
door emission rate ranged from 11 to 19-percent during
BSO Sample No. 1, from 29 to 32-percent during Sample
No. 2, and from 27 to 45-percent during Sample No. 3.
This corresponds .to the pattern of emission rates of
BSO and benzene at the WESP inlet, which progressively
increased from Sample No. 1 to Sample No. 3.
Visible Emission Observations
Visible emissions from the WESP exhaust stack
were recorded during each BSO sample run. The readings
were summed and averaged over six-minute periods.
The summaries of visible emissions may be found in
Append ix B-4.
During Sample No. 1 the six-minute averages were
5-percent or less, with one excursion to 10-percent.
Averages over Sample No. 2 ranged from 6 to 24-percent.
Visible emissions were terminated before the end of
BSO Sample No. 3 due to darkness, however, the averages
were 20-percent or less, with two excursions to 34 and
40-percent. Generally, therefore, the visible emission
observations follow the progressively increasing trend
of BSO emission results from Sample No. 1 to Sample No. 3
- 12 -
-------
3.0 PROCESS DESCRIPTION AND OPERATION
To be supplied by EPA.
- 13 -
-------
4.0 LOCATION OF SAMPLING POINTS
INLET
The WESP inlet sampling location was a 75.5-
inch (191.8 cm) by 75.0-inch (190.5 cm) duct, located
approximately 17-feet (5.18 meters) downstream of
a 45-degree bend in the duct and 28-feet (8.53
meters) upstream of the WESP. This location provided
adequate upstream/downstream distances to disturbances.
The sampling platform was 37-feet (11.3 meters) above
ground level. The duct was accessed through four
3-inch (7.6 cm) ports along the vertical face.
Each traverse consisted of eleven sampling points.
Velocity pressures and temperatures were measured at
each of the 44 sampling points. Figure 4.1 depicts
the inlet sampling location along with the traverse
points and their respective distances from the inside
duct wall.
OUTLET
The WESP outlet sampling location was a 95.5-
inch (242.6 cm) I.D. stack located approximately 56-
feet (17.1 meters) downstream of the nearest disturbance
(fans) and 15-feet (4.57 meters) upstream from the top
of the stack. This provided adequate upstream/downstream
distances to disturbances. The sampling platform was
80-feet (24.4 meters) above ground level. The stack
- 14 -
-------
75.0"
11
4- 4 -f
4 -t- 4
1
4- •»- 4-
44 4^-444 t •»• 4 t
4-44444-4 4444
4 t 4 4 4 4
11
44-1-
Plan view
3
75.5"
Ul
i
Point
1
2
3
4
5
6
7
8
9
10 ..
11
Distance
(Inches )
3
10
17
24
30
37
44
51
58
65
72
.4
.3
.2
.0
.9
.7
.6
.5
.3
.2
.1
I
From
capture
hood
Inlet sampling
location
Elevation view
Inlet sampling
location
28' i 17'
From
capture
hood
Figure 4.1. Inlet stack cross-section and sampling location (not to scale)
-------
was accessed through two 3-inch (7.6 cm) ports located
at a 90-degree separation about the stack circum-
ference .
Each traverse consisted of ten sampling points.
Velocity pressures and temperatures were measured
at each of 20 sampling points. Figure 4.2 is a
diagram of the outlet sampling location with each
of the traverse points and their respective distances
from the inside stack wall.
OBSERVER LOCATION FOR FUGITIVE EMISSIONS OBSERVATIONS
Figure 4.3 presents a plan view of coke oven Battery
Nos. 1 and 2. Battery orientation is presented,
along with the designation of coke side and push side,
and oven door numbers.
For safety reasons, observations were made from
outside the pusher machine and quench car tracks,
placing observers 15 to 35-feet away from the battery.
On the coke side, observers sometimes stood in the
quench car tracks to obtain a better view of the oven
doors. This resulted in an extremely dangerous
situation since the movement of the quench car had
to be watched constantly. All observations were made
from ground level with the guidance of an Armco, Inc.
representative.
- 16 -
-------
Plan view
Outlet
sampling
location
WESP
95.5-inch I.D.
Point
1
2
3
4
5
6
7
8
9
10
Distance
(Inches )
2
7
13
21
32
62
73
81
87
93
.5
.8
.9
.6
.7
.8
.9
.6
.7
.0
15'
Elevation view
80
95.5-inch I.D,
Outlet
sampling
location
WESP
Figure 4.2. Outlet stack cross-section and sampling location (not to scale)
-------
N
CS
1
1
52
Battery 1
52
53
Battery
53
68
2
68
PS
Figure 4.3. Plan view of battery orientation,
- 18 -
-------
5.0 SAMPLING AND ANALYTICAL PROCEDURES
BENZENE SOLUBLE ORGANICS
Sampling was conducted in accordance with EPA
Reference Methods 1 - 4, as outlined in the Standards
of Performance for New Stationary Sources (Federal
Register, 40CFR60, December 23, 1971, as amended
through August 18, 1977) and the EPA draft method
Benzene Soluble Organics July 3, 1978.
Triplicate samples were extracted isokinetically
and simultaneously from the .inlet and outlet of the
Mikropul wet electrostatic precipitator. At the
inlet, 44 points were sampled for three minutes
each, while at the outlet, 20 points were sampled
for seven minutes per point.
Prior to sampling, each duct was divided into
equal areas and exhaust gas velocities and temperatures
were measured at their centers. Velocity pressures
were obtained, using a calibrated S-Type Pitot tube
and an inclined 0 to 10-inch water gauge manometer.
Temperatures were measured with an iron-constantan
(Type J) thermocouple attached to a calibrated pyrometer,
Preliminary moisture determinations were made at both
locations each using a Method 4 sampling train. An
exhaust gas grab sample was obtained from the inlet
and analyzed by the Orsat method for gas composition.
- 19 -
-------
Exhaust gas flowrates and the nozzle sizes required
to maintain isokinetic sampling rates were then
calculated from these preliminary determinations.
Each BSO sampling train consisted of a sharp,
tapered, stainless steel nozzle; a heated TeflonO^
probe and flexline at the inlet, a heated glass
probe at the outlet; an empty modified Greenburg-
,Smith impinger; an unheated 110-mm Type A glass-fiber
filter in a glass filter holder with a thermocouple
positioned at the outlet; one modified and one
standard Greenburg-Smith impinger each containing
150-ml of distilled water; two modified Greenburg-
Smith impingers, the first empty, the second con-
taining 200-300 grams of silica gel with a thermocouple
positioned to monitor the temperature at the impinger
outlet; an umbilical cord; a leak-free vane axial-
vacuum pump with a vacuum gauge; a calibrated dry gas
meter equipped with bimetallic inlet and outlet
thermometers; and a 0 to 10-inch water gauge manometer
connected to a calibrated orifice-type flowraeter. The
impingers were immersed in an ice bath to maintain the
impinger temperature at +70F.
While conducting each sample run, the temperatures
of the filter holder and the last impinger were monitored
and maintained below 104F (40C) and 70F (20C), respec-
tively. The probes were connected to the rest of the
sampling train with ball and socket joints, stainless
- 20 -
-------
steel at the inlet and glass at the outlet. Teflon®
tape was used on all connection fittings up to the
filter holder inlet to eliminate the possibility of
contaminating the sample with stopcock grease. Stop-
cock grease was used on all remaining glassware compo-
nents. A schematic of the sampling train is depicted
in Figure 5.1.
Each sampling train was checked for leakage
before and after each sample run, in accordance
with the requirement that the initial leak rate
shall not exceed 0.02 cubic feet per minute (cfm)
at 15-inches of mercury vacuum. The final leak rate
was checked at or above the greatest vacuum which
occurred during the run. At the inlet, the probe
assembly was moved to each sampling point, where the
velocity pressure and temperature of the exhaust gas
was measured and recorded. At the outlet, the
sampling train glassware was connected directly to the
probe and the assembly moved to each point.
At each individual sampling point, an isokinetic
sampling rate was calculated and the sampling flowrate
was adjusted accordingly, using an orifice-type meter
which indicated instantaneous flowrates. Isokinetic
sampling rates were maintained within 10-percent
of true isokinecity for any velocity pressure
measured. An insulating asbestos mitten and duct tape
- 21 -
-------
—J Heated probe
S-type Pitot
tube
110-mm
Type A glass-
fiber filter
/ 3
' T-Thermocouple
Thermocouple
Inclined
manometer
Dry
trap
150-ml
distilled
water
Dry 200-300g
trap silica
gel
Ori.fice
Thermometers
~^
\ \.
Bypass
valve
Vacuum
line
Main
valve
Vacuum
gauge
Inclined
manometer
Vacuum
pump
Figure 5.1. Benzene soluble organics sampling'train.
-------
were positioned around the probe assembly in each
sampling port to maintain a relatively positive
seal.
The testing program was designed to measure
non-pushing emissions only. Therefore, sampling
ceased during the pushing cycle. Pushes were
monitored by a Clayton Environmental Consultants, Inc.
observer and direct communications were maintained
between the observer and the sampling teams. For
the purpose of this study, push duration was considered
to be from the time the coke was sighted emerging
from the oven until the shed had been relatively
cleared of pushing emissions (after the quench car
had exited the shed area).
Following the final leak check, the sampling
trains were moved to a relatively dust-free area for
•sample transfer. Any condensate collected before the
filter was measured and collected in a glass sample
bottle. The probe, probe extension (inlet only),
initial condensate trap, and front-half of the filter
holder were rinsed and brushed, initially with acetone
and secondly, with benzene. The rinsings were collected
in separate glass sample bottles with Teflon® lined
caps. The volumes of the impingers were measured and
- 23 -
-------
increases recorded as condensate. The silica gel
was weighed and the gain recorded as condensate.
The impinger solutions were not saved beyond
volume determinations. The filter was transferred
to its original Petri dish and sealed. All bottles
were sealed with tape and liquid levels marked.
Thus, at the end of each sample run, the following
fractions were available for BSO analysis:
(1) condensate, when collected, before the filter;
(2) acetone rinsings of the probe, probe
extension (inlet only), initial condensate
trap, and front-half of the filter holder;
(3) benzene rinsings of the probe, probe
extension (inlet only), initial condensate
trap, and front-half of the filter holder;
and,
(4) 110-mm glass-fiber filter.
In the laboratory, each bottle was checked for
leakage and volumes measured. Fraction 1 was then
extracted in a separatory funnel three times with 50-ml
of benzene. The extract was then filtered through a
Whatman® 40 filter into a tared 250-ml beaker. The
filtrate was then dried at room temperature to a residue,
Fraction 2 was dried at room temperature in a tared
250-ml beaker. The residue was then extracted with
50-ml of benzene and set in an ultrasonic bath for one
- 24 -
-------
hour. The extract was then filtered through a Whatmarr
40-filter into a tared 250-ml beaker. The filtrate
was then dried at room temperature to a constant
weight. Fraction 3 was dried at room temperature
in tared beakers and the residue weighed until
constant. Fraction 4 was extracted with benzene in
a Soxhlet extractor for six hours. The extract
/R\
was then filtered through a Whatman6' 40 filter into
a tared beaker. The filtrate was then dried at
room temperature to residue. All weighings were
performed on analytical balances with sensitivities
of 0.1 milligram. A summary of weights by fraction
appears in Appendix C.
INTEGRATED BAG SAMPLING (BENZENE AND ORSAT)
An integrated bag sample was withdrawn from the
WESP inlet and outlet ducts simultaneously with each
BSO sampling run utilizing the train described by
EPA Method 110 and depicted in Figure 5.2. Sampling
was conducted during steady operation of the battery,
not during push times. An evacuated 96-liter Saran®
bag, especially treated to reduce permeability, was
placed inside an insulated steel drum. The drum was
then gradually evacuated, thereby filling the
/BS
Sararf^ bag. A rotameter was placed in-line to control
the actual sample flowrates, as. shown in Figure 5.2.
- 25 -
-------
I
to
n Stainless steel sampling line
•
Rotameter
Stainless
steel
probe
96-liter
J| Saran
Teflon tubing
Needle
valve
Insulated steel drum
Figure 5.2. Integrated bag sampling train.
-------
Upon filling, the bag was removed and transferred
to a laboratory for immediate gas chromatographic
(GC) analysis for benzene content and Orsat
analysis for gaseous composition.
Benzene concentrations were determined in
accordance with EPA Method 110, "Determination of
Benzene from Stationary Sources", delineated in
Appendix E-2. Gas chromatographic field analyses
were performed utilizing an Analytical Instrument
Development (AID) Model 511, portable gas chromato-
graph with a flame ionization detector and a 6' x
1/8" stainless steel column packed with 1.75-percent
Bentone and 5-percent SP1200 on 100/120 mesh Supelcoport.
The following operation conditions were maintained
for all analyses: 85C oven, 105C detector, 99C gas
sampling loop with 1-ml capacity, and 15 ml/min zero
nitrogen carrier gas. Prior to sample analysis, the
EPA required that the analyst accurately identify the
concentration of two audit cylinder standards (one low
concentration standard in the range of 5 to 20 ppm
benzene, and one high concentration cylinder in the
range of 100 to 300 ppm benzene). Each measured concen-
tration agreed to within +10% of the actual concentration
as required. The Field Audit Report can be found in
Appendix H. Samples were then analyzed and peak areas were
measured using a compensating planimeter. The sample
chromatograms had to apparent peaks, which were completely
resolved.
- 27 -
-------
Following the GC analyses, each integrated bag
sample was analyzed by the Orsat method for carbon
dioxide, oxygen, and carbon monoxide concentrations,
as specified in EPA Method 3. These results were used
to calculate the molecular weight of the process gas.
FUGITIVE EMISSIONS
Visible emission observations were performed in
accordance with EPA Draft Method 109 (and Addendum
to Method 109), Determination of Visible Emissions
from Coke Oven Batteries, Part C. These observa-
tions were conducted simultaneously with each BSO sample,
Several modifications to the method were made due to
difficulties encountered during the testing program.
Draft Method 109 requires one observer, however,
three observers were used for this study. All three
observers were to traverse together either the coke
or push side, then move to the opposite side of the
battery to inspect the remaining doors and complete
the run. Due to insufficient lighting, only one
observer traversed the coke side per run. All three
observers traversed the push side of the battery.
Two of the observers started their traverse simulta-
neously from opposite ends of the battery. The
third observer started the traverse from either end,
not less than one minute nor more than two minutes
- 28 -
-------
after the first two observers began traversing. A run
consists of traversing both the coke and push side
of a battery. Four runs were conducted during BSD Sample
Nos. 1 and 2 each. Due to darkness, only three runs
were conducted during BSD Sample No. 3.
The coke side was covered by a shed which captured
door leak and pushing emissions. This shed allowed
very little entry of natural light and several of the
electric lights, located within the shed, were inoperable,
Therefore, the darkness made reading of the coke doors
extremely difficult. Since there was no feasible way
of obtaining proper lighting, the EPA Technical Manager
decided that observers would use a high powered lantern
light to aide in viewing the top of the doors. Those
doors located at the outermost ends of the battery were
easiest to view since more light entered these areas.
The light intensity from the lantern was such that
the beam had to be moved around the jamb area of each
oven door (from top to bottom) to view the entire door.
Since there was only one lantern and observers.
were not allowed to traverse the battery in a group,
only one observer viewed the coke side per run.
Therefore, each observer read the coke side every third
run.
_ 29 -
-------
Jamb, buckstay, and lintel leaks were documented
by the observers, in addition to oven door and chuck
door leaks from the push side. Distinguishing between
these various types of leaks for the coke side was
impossible due to the lighting problem. The observers,
when entering the shed from bright sunlight, had to
wait several minutes before starting a traverse to
allow for eye adjustment.
Several other problems were encountered during
this study which made observations of the coke oven
battery doors extremely difficult. Obstructions,
such as push cars, door cars, quench cars, and
other equipment located on the battery, resulted
in frequent delays. Some interruptions were caused
•\
by plant personnel taking breaks. The workers would
leave the equipment in front of the oven doors,
making observations in those areas impossible.
Using a lantern light created several problems.
Fine dust particles, always present in the battery
area, were accentuated by the light beam. It was
difficult at times to determine if the oven door was
actually leaking through this intensified haze.
The wind also created some problems. Dust, which
had settled in the battery area, along with smoke from
- 30 -
-------
oven doors, would be carried sometimes across the entire i
battery, obscuring the vision of the remaining oven ;
doors. !
Observers had to view approximately 15 oven doors
at an angle ranging from 0 to 45-degrees. The bin,
where the quenched coke is dumped, was located in front
of these doors. No one was allowed in front of this
bin due to lack of clearance from the quench car.
Determining which doors were leaking and the type
of leak was very difficult, if not impossible at
times. This was especially true when oven doors
were leaking heavily, filling the entire area with
smoke.
At the request of the Technical Manager, Battery
Nos. 1 and 2 were observed as one, since only one push
car/quench car unit serviced both batteries.
Opacity Readings
Addendum to Draft Method 109 requires observers
to determine the opacity of the emissions at the lintel
area. Since exhaust hoods were located in this area
on the push side, opacity readings ware not recorded.
The aforementioned problems encountered on the coke
side prevented any reading of opacities, especially
since the lintel area was the darkest area of the battery,
-• 31 -
-------
VISIBLE EMISSIONS
Visible emissions from the WESP exhaust stack
were recorded for the duration of each BSO.sample
run. The observations were performed in accordance
with EPA Method 9 by a qualified visible emissions
observer. A summary of the visible emission data
is presented in Appendix B-4.
- 32 -
-------
SECTION. II - TRW REPORT
COKE OVEN EMISSION TESTING
Project No. 79-CKO-22
Contract No. 68-02-2812
Work Assignment No. 51
-------
TABLE OF CONTENTS
Page
List of Figures i
List of Tables i
1.0 Summary . 1
2.0 Sampling Locations and Location of 4
Traverse Points
3.0 Sampling Procedure 5
4.0 Laboratory Procedures 8
APPENDICES
A. Field Data Sheets
B. Analytical Results
-------
LIST OF TABLES
Number
1
2
3
4
B(a)P Test Results
B(a)P Calculations
B(a)P Test Results
Typical Elemental Analysis
for a Filter
Page
2
A-l
A-2
A-4
Number
1
2
LIST OF FIGURES
B(a)P Train
Battelle Trap
6
7
-------
1.0 SUMMARY
The results of the benzo (a)pyrene (B(a)P) test
conducted at the Houston, Texas location of Arm'co
Steel Corporation are presented here. The Environ-
mental Engineering Division of TRW, Inc. tested
simultaneously at this location with Clayton Environ-
mental Consultants, Inc.
The B(a)P trains were run simultaneously with the
benzene soluble organic (BSO) trains. This included
stopping and starting runs to avoid sampling during
a push. The inlet train sampled 44 points at three
minutes a point. During the last test, only thirty-
three points were sampled because the nozzle was
pulled off while removing the probe from the third
pprt. This did not affect the results of the test.
The outlet train sampled twenty points at seven
minutes a point. Because of interferences at the
outlet, it was not possible to sample points nine and
ten (see diagram on page A-27). For this reason, the
nozzle remained at point eight for twenty-one minutes
with readings taken every seven minutes. The results
from the inlet and outlet B(a)P test are listed in Table
1.
On Test Numbers 2 and 3, the outlet B(a)P values
are larger than the inlet values. A possible explanation
of this involves the wet precipitator. The potential
-------
TABLE 1. BaP TEST RESULTS
1-0
Test #
Date
Time
Meter Vol.(DSCF)
Stack Flow(DSCFM)
% Moisture
% Isokinetic
BdP(Lb/DSCF)
BaP (mq/DSCM)
BaP (Ib/hr)
BaP (kg/hr)
Stack Temp (OF)
BaP-I-1
10-2-79
1420-1756
45.987
163,692
0.8
97.7
5.731 x TO'9
0.0918
0.056
0.026
109.4
BaP-0-1
10-2-79
1400-1816
143.954
169,158
2.1
95.6
3.583 x 10"9
0.0574
0.036
0.016
88.1
BaP-I-2
10-5-79
0946-1322
51.591
179,849
0.6
101.3
6.455x 10"9
0.1034
0..070
0.032
116.7
BaP-0-2
10-5-79
0950-1322
137.477
163,358
1.4
94.6
1.191 x 10"8
0.1908
0.117
0.053
86.6
'
BaP-I-3
10-5-79
1655-2013
38.822 .
179,439
2.1
100.4
5.213 x 10"9
0.0835
0.056
0.025
127.5
BaP-0-3
10-5-79
1655-2100
133.210
157,826
3.2
94.9
1.834 x 10"8
0.2938 1
0.174 !
0.079
90.7 1
i
-------
situation exists that the clarified process water
used in the precipitator contains 0.0082 p.g/ml
B(a)P and that it is entrained by the precipitator
and carried out in the outlet stream.
-------
2.0 SAMPLING LOCATIONS AND LOCATION OF TRAVERSE POINTS
The sampling locations and traverse points are
exactly the same as were used for the BSO tests.
(Refer to Section 4.0 of the Clayton report).
-------
3.0 SAMPLING PROCEDURE
The sampling procedure used at the coke oven outlet
consisted of an EPA Method 5 train, modified in the
following manner (using EPA's draft B(a)P Method).
A Battelle trap loaded with XAD-2 res.in was inserted
between the heated filter, which was cut from General
Metal Works No. 25 Hi-Volume filters (see appendix
for typical Elemental Analysis), and the first impinger.
A thermostatically controlled water bath controlled
the temperature of the Battelle trap at 127F. The
Battelle trap was shielded from visible and ultra-
violet light by wrapping with aluminum foil. The
Battelle trap was capped after sampling and remained
capped until the analysis was performed. Methylene
chloride was used for the recovery of the sample from
the 316 stainless steel probe, glass filter holder
with 316 stainless steel filter support and connecting
glassware up to the Battelle trap.
The inlet train was identical to the outlet train
except that a heated Teflon tube was used between the
probe and the filter holder. The tube rinse was included
as part of the inlet sample. Figure 1 is a schematic
of the typical B(a)P train while Figure 2 shows a
Battelle trap as was used in the B(a)P train. All field
data sheets and analytical forms are included in the
appendix.
-------
Battelle trap
Filter
Water Bath
Controlled
127 °F
Pump —X-J
l
Umbilical Cord
-Figure 1. BaP. Sample Train,.
-------
Glass Water Jacket
Glass Fritted Disc
8mm- Glass Cooling Coil
Glass Wool
Figure 2. Battelle Trap.
-------
4.0 LABORATORY PROCEDURES
The volume of the rinse sample was recorded and
the sample was stored at 4C in an amber glass
bottle until the analysis was performed. If the
rinse sample was deeply colored or contained a large
amount of suspended material, it was diluted ten to one
with cyclohexane before it was analyzed.
The filter was extracted with 100-ml of cyclohexane
while the XAD-2 resin from the Battelle trap was
extracted with 250-ml of cyclohexane. The extraction
procedure placed the filter or XAD-2 into a single
thickness pre-extracted cellulose extraction thimble.
The thimble was then placed in a soxhlet extraction
apparatus and extracted for eight hours at five to
six cycles per hour. All this was done either behind
a yellow light-safe screen or under a yellow safe light.
At the end of the extraction, the extract volume was
recorded, and the extract stored in an amber bottle
at 4C until the analysis was performed. The thimble
was checked with a black light to confirm complete
extraction.
ANALYTICAL PROCEDURE
The samples are analyzed for B(a)P using the
fluorescence spectrophotometric procedure. This method
is preferred over the thin layer chromatographic (TLC)
method for low level B(a)P analysis, as the TLC method
-------
has only 0.01 the sensitivity of direct liquid measure-
ment. The benzc(a)pyrene method using the fluorescence
spectrophotometry was tailored to these samples. The
method originally chosen was intended to be thin layer
chromatography separation and measurement by scanning
in-situ with a scanning attachment for the fluorescence
spectrophotometer. This method lacked the sensitivity
required for the analyses.
The equipment used for this analysis was the
Aminco Model SPF-125 Spectrophotofluorometer with 7-mm
lightpath cell. This instrument accurately measures
concentrations of B(a)P as low as 0.0001 ppm. The
wavelength settings were 378-nm excitation and 403-
nm emission with respective slitwidth openings of 1-mm
and 5-mm. This instrument becomes extremely substance
specific at very narrow slit widths, as was used in
this analysis.
The spectrophotometer is equipped with a high
intensity xenon lamp which provides the excitation
energy. For B(a)P analysis, the best results are
obtained by setting the excitation wavelength and emission
wavelength to produce the maximum peak height. With a
narrow slit width, the specificity of the instrument is
greatly increased. The excitation wavelength is 378 nm.
The minimum entrance slit width used was 1-mm,, The
excitation energy is re-emitted as fluorescence of a
longer wavelength. For B(a)P, this wavelength is 403-nm.
-------
The exit slit width can be narrower than the entrance
slit width, as in this case, 0.5-mra. The fluorescence
is expressed as a relative intensity. The relative
intensity values are converted to B(a)P concentrations
by analyzing a set of known standards. These standards
are prepared by serial dilution of a 1000 \j.g/ml B(a)P
stock solution. This is prepared by dissolving 10-rag
of three times recrystallized B(a)P in 10-ml spectral
grade cyclohexane. This is stable for several months
if stored away from light at OC.
To determine the concentration of B(a)P in unknown
samples, it is necessary to plot a curve of the relative
intensities from the standards. The p.g/ml in the sample
is then determined by the sample's relative intensity
compared to the graph of the standards. The WESP
clarified process water was analyzed using the same
method as is used for the rinse of a B(a)P sampling train.
The results can be affected by temperature, humidity,
and light. Precautions are taken during sampling,
preparation, and analysis to keep the exposure to light
at a minimum. The optimum relative humidity is between
35-percent and 50-percent. The instrument is equipped
with a constant temperature cell compartment to avoid
instability and the possible loss of sensitivity which
could be caused by a change in sample temperature. All
glassware with which the sample comes in contact is cleaned
10
-------
by using a soapy water wash, 50-percent nitric acid
rinse, and a distilled, deionized water rinse,
respectively. When using fluorescence spectre-photo-
metry, only high quality quartz cuvettes are used.
No corks, rubber stoppers or lubricating agents are
used and care is taken so that impurities do not
contaminate the sample.
11
------- |