c/EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EMB Report 81-8YC-13
September 1981
Air
Benzene
Coke Oven By-Product Plants
Fugitive Process Emissions
Emission Test Report
Vapor-Liquid Analyses
-------
HEADSFACE
BENZENE CONCENTRATION
OVER LIQUID SAMPLES FROM
COKE BY-PRODUCT PLANTS
EPA Contract 68-02-3544
Work Assignment 3
ESED Number 74/4j
September 1981
Prepared For:
Mr. Daniel Bivins
U. S. Environmental Protection Agency
OAQPS, EMB, ESED, MD-13
Research Triangle Park, NC 27711
-------
TABLE OF CONTENTS
Page
1.0 INTRODUCTION .......................... 1
2.0 RESULTS
3.0 DISCUSSION OF RESULTS . .
4.0 ANALYTICAL PROCEDURES
APPENDIX A - FIELD DATA SHEETS
APPENDIX B - LIQUID SAMPLE ANALYSIS
APPENDIX C - LABORATORY HEADSPACE DATA
-------
-1-
1.0 INTRODUCTION
Scott Environmental Services conducted a benzene sampling program
at seven coke by-product plants during the summer of 1980 for the U.S.
Environmental Protection Agency. Fugitive process emissions from sources
were determined by measuring the benzene concentration in the stack gas and
the gas flow rate and temperature. Process liquid samples were also collected
from each source.
A test program was conducted to determine if any correlation existed
between the benzene emission, rate from a process and the benzene concentration
in the headspace over a liquid sample from that process. If a correlation
existed it would allow emission rate estimates based on a simple laboratory
procedure rather than on field sampling procedures.
Laboratory tests were performed on liquid samples from eight con-
fined sources. The samples were heated in enclosed vessels and headspace
samples were extracted and analyzed by gas chromatography. Two separate
sets of experimental conditions were used. Initially the samples were
heated to process temperature in vented flasks. Due to the great variability
in the results a new procedure was devised wherein the samples were all
heated to 212°F while maintaining a constant pressure as thermal expansion
of the headspace gas occurred. No gas was vented.
The following sections present the results and analytical procedures
of the vapor headspace. sampling program.
-------
-2-
2.0 RESULTS
Table 2-1 presents the results of the headspace sampling at 212°F
(unvented) and the calculated ratios of headspace/liquid concentrations,
headspace/stack gas concentrations, and stack gas/liquid concentrations.
The results of the headspace sampling at process temperatures (vented) are
.shown in Table 2-2, with the ratios of headspace at 212°F/headspace at
process temperature and headspace at process temperature/stack concentration.
The two tar decanters are not included in Table 2-2 because the headspace
benzene concentrations were not measured at process temperature.
All the process samples used in this program were dipped from the
liquid surface through a hatchway, with the single exception of the tar
dehydrator at Burns Harbor, where the sample was collected from a pump at
the tank outlet..
-------
TABLE 2-1 HEADSPACE BENZENE CONCENTRATIONS AT 212°F
Source/Plant
Wash Oil Decanter (F)
Tar Decanter (BH)
Tar Decanter (B)
Tar Dehydrator-W (F)
Tar Dehydrator-E (F)
Tar Dehydrator (BH)
Tar Storage (P)
Tar Storage (M)
Sample
Composition
Oil
Aqueous
Aqueous
Heavy Tar
Heavy Tar
Whole Tar
Aqueous
Aqueous
Process
Temp.
t- (°F)
205
120
178
207
219
162
145
168
(S)
Stack
Benzene
Cone.
(ppm-V)
817
2473
1380
1757
2080
9133
630
1825
(L)
Liquid
Benzene
Cone.
(ppm-Wt )
44
92
5
621
276
1987
33
N.A.
(H)
Headspace
Benzene
Coric.
At 212°F
(ppm-V)
1404
2388
336
4261
1390
12385
2550
1402
H/L
Ratio
32
26
64
6.9
5.5
6.3
70.0
—
H/S
Ratio
1.7
1.0
0.3
2.4
0.7
1.4
3.7
0.8
S/L
Ratio
18.6
26.9
230
2.9
7.4
4.7
19.1
—
(F) = Fairless
(BH) = Burns Harbor
(B) = Bethlehem
(P) = Pueblo
(M) = Honessen
-------
TABLE 2-2 HEADSPACE BENZENE CONCENTRATIONS AT PROCESS TEMPERATURES
Source/Plant
Wash Oil Decanter (F)
Tar Dehydrator-W (F)
Tar Dehydrator-E (F)
Tar Dehydrator (BH)
Tar Storage.(P)
Tar Storage (M)
Process
Temp.
t (°F)
205
207
219
162
160
169
Ht
Headspace*
Benzene
Cone.
at t (ppm-V)
Avg.
179
1273
521
2346
75
252
Max.
189
1273
521
5296
193
622
H
Headspace**
Benzene
Cone.
at 212°F
(ppm-V)
1404
4261
1557
12385
2309
1402
H/Ht
Ratio
7.8
3.3
3.0
5.3
30.8
5.6
Ht/Stack
Ratio
0.22
0.72
0.25
0.26
0.12
0.14
*Vented headspace procedure
**Confined headspace procedure
-------
-5-
3.0 DISCUSSION OF RESULTS
The lack of any close correlation between the headspace benzene
concentration and the source concentration measured in the field tests of
multi-phase processes is due to the many other variables which affect the.
benzene emitted by these sources. These variables include:
1. Type of Process - continuous, steady state; continuous,
variable-, batch.
2. Nature of Material - tar, oil, aqueous, combination of
tar and aqueous.
3. Dynamic residence time of continuous processes.
4. Degree of agitation of process material.
5. Concentration of water and light organics in tars and oils.
6. Thickness of aqueous layer in aqueous organic systems.
7. Process headspace volume vs. process surface area.
8. Temporal variations in temperature, liquid level,
liquid composition, throughput, etc.
9. Ambient temperature and pressure.
The liquid samples chosen for this test program were those, collected
from confined sources, i.e. covered and vented to the atmosphere, because
the conditions affecting open sources would introduce many more variables
into the. data analysis.
Each of the above factors has an impact on the concentration and
rate of benzene emissions. A detailed discussion of the role of each variable
is beyond the scope of this task, but a brief- discussion of a tar decanter
will demonstrate the complex nature of these systems. Tar decanters are
used to separate mixtures of tar and ammonia liquor from primary coolers and
battery collector mains into liquor, tar and sludge. The tar decanter tested
at Bethlehem Steel, Bethlehem, Pennsylvania yielded benzene concentrations
of 720 to 1980 ppm in the vent during three test runs. The flushing liquor
sampled from the surface (180°F) which'contained 5 ppm (wt) benzene gave a
headspace benzene concentration of 386 ppm at 212°F. The incoming material
from the primary cooler contained approximately 20% tar and 80% flushing
liquor. The flushing liquor contained 16 ppm (wt) benzene and the tar 1800 ppm
(wt). Headspace gas from this material contained 41,000 ppm benzene which
-------
-6-
is 100 times the concentration in the corresponding surface liquor (no tar)
analysis. This demonstrates the transfer of benzene from the tar to the
flushing liquor and then to the headspace in the laboratory analysis. In
the actual process, the tar decanter contains a tar layer beneath the
surface flushing liquor yet the measured benzene concentration was 20 to 50
times less than in the laboratory test of the tar/liquor sample. In fact,
the measured process concentrations average only 2 to 5 times the laboratory
result from the liquor alone. The most likely factor limiting the actual
process emission is the thickness (depth) of the liquor layer. The thickness
c-f the liquor layer was not measured but can be assumed to be at least 1 to
2 feet, In the laboratory test, the liquor layer in the flask was less than
one inch thick, thus facilitating transfer of the benzene in the tar to the
headspace. If the tar decanter were operated as a static (batch) process,
it would be expected that the benzene concentration would approach the
41,000 ppm value obtained in the lab test. However, during normal operations
as a continuous flow system with.-a short dynamic residence time, the benzene
concentration remained less than 2,000 ppm.
The above data demonstrate that the benzene emissions from a com-
plex multi-phase system cannot be estimated from simple laboratory tests.
The relatively fev; confined, multi-phase systems included in the field test
program and the limited information available on process variables preclude
any further search for a means of estimating emissions from such sources.
Further work would have tc be based on more extensive field data
and on complex phase transfer and boundary layer theory and experiment.
-------
-7-
4.0 .ANALYTICAL PROCEDURES
Two separate sets of experimental conditions were used to measure
headspace benzene from samples collected from confined sources. Initially
headspace benzene concentrations were measured with the liquid sample at
process temperature in a vented flask. Experimental procedures were subse-
quently revised such that headspace benzene was measured with the liquid
sample at 212°F without any venting. 212°F was chosen because it provided
a convenient way to maintain a stable sample bath temperature.
For the first series of determinations at process temperature,
25 cc of the liquid sample was placed in a graduated glass vessel which
was fitted with a ground glass joint. The vessel top had a single port
which was covered by a small rubber septum. A 20 gauge syringe needle was
inserted in the septum for the purpose of relieving pressure and simulating
the process vent. The samples were heated by immersing the vessel in a
beaker of water which was maintained at the process temperature. Samples
were allowed to equilibrate in the water bath for approximately 15 minutes
prior to extraction of headspace gas samples for injection into the chromatograph.
From the triplicate liquid samples collected from each process, an
aliquot of a single sample was aniyzed for the tarry materials and a composite
of all three samples was analyzed for the less viscous liquids. Replicated
determinations were not done during this phase of the analysis. Chromatographic
data for consecutive injections of the same sample show a significant degree
of variability.
This lack of replicability was attributed to problems in maintaining
a constant bath temperature and also to the open vent in the top of the vessel,
which was releasing benzene from the vessel and consequently changing the
concentration of benzene in the headspace. The experimental procedure was
revised such that, all samples were heated to 212°F in a boiling water bath,
and a 20 cc or 50 cc syringe was used in place of the open syringe needle, to
relieve pressure without releasing benzene as the sample was heated. This
created a more stable equilibrium in the sample vessel prior to extraction of
a headspace sample for analysis. Chromatographic data for this second pro-
cedure had a higher degree of precision than did the previous analysis.
-------
-8-
For the chromatographic analysis of most samples it was necessary
to dilute samples before injection into the chromatograph. Dilutions were
typically by a factor of 10 or 20.
All gas samples were analyzed using a Shimadzu GC Mini 1 gas
chromatograph equipped with dual flame ionization detectors,.dual elec-
trometers, heated sample loop and a backflush system.
The samples were drawn into a glass syringe and introduced to the
sample loop inlet. The samples once in the sample loop were allowed to come
to atmospheric pressure by waiting 15 seconds prior to injection. The
following chromatographic conditions were maintained:
Column Temperature (isothermal) - 100°C
Injector and Detector Temperature - 200°C
5 ml Sample Loop, Temperature - 50°C
Carrier Gas Flow Rate - 32 cc/min.
Hydrogen Flow Rate - 40 cc/min.
Air Flow Rate 240 cc/min.
Analysis Time 5 min.
Detector - . Flame Ionization
The columns used for laboratory analysis were:
A - Scrubber Column
• 10% FFAP on Supelcoport 80/100
1/8" x 1 m Stainless Steel
B - Analytical Column
20% SP-2100, 0.1% Carbowax 1500
100/120 Supelcoport
1/8" x 10' Stainless Steel
-------
APPENDIX A
FIELD DATA SHEETS
LIQUID SAMPLES
-------
Page B-10
PROJECT 1906 BENZENE/BaP PRESURVEY
Plant
SAMPLE DATA
Process
Tgr J
Date ~7
Sample No.
Time Sampled
Sample Type: (Liquid/ Air
Sample Temperature
©
T" O
Ambiient 'Temperature
Description of Sampling Location:
0
Sample No.
Sample Type: .Liquid Air
^l*~ __ —**^
Sample Temperature Q pi C-
Ambient Temperature
Description of Sampling Location:
Time Sampled
/
/^U/LxUL^e.
1)
/vA-i c^wxjtf-M, &~f~
Sample No.
Time Sampled
Sample Type: Liquid Air
Sample Temperature
Ambient Temperature
Description of Sampling Location:
-------
Plant
Page B-13 '
PROJECT 1906 BENZEN'E/BaP PRESURVEY
SAMPLE DATA
Process
Date
Sample N<
. TD
Sample Type: (Liquid) Air
Sample Temperature
Time Sampled
•&h<
,->
Description of Sampling Location:
Sample No.
Time Sampled
Sample Type:
Sample Temperature
Air
Ambient Temperature
Description of Sampling Location:
£ od If
Sample No.
Sample Type: Liquid Air
Sample Temperature
Ambient Temperature
Time Sampled
Description.of Sampling Location:
-------
PROJECT 1957 07 0181
SAMPLE DATA
Plant f?(/£/V'£ /'fa rbo/Z Process "Trj £ d pf r{ /i ,-f-ff T Date
~T~/I ?> 7%, x- ^ ^-( 7 ^ ^ ' ") Time Sampled ^V i .-< J
Sample Type:^ Liquid! Air
Sample Temperature ) (^ 0 /
Ambient Temperature
/•
Co
Description of Sampling Location:
f -f-> -/^
AL-Cg^U/r (*;gn I G^'1' i
Sample No. 1 gU
-------
Page B-6
PROJECT 1906 BENZENE/BaP PRESURVEY
SAMPLE DATA
Plant
Process
"•>/"~ Date
Sample No.
Time Sampled
Sample Type: Liquid Air
Sample Temperature
Ambient Temperature
Description of Sampling Location:
. 0 .,/ ',.'
• _;. . v-'J / ' & (_, -
."'..•. / ) '
Sample No.
Time Sampled -c--',.
0
Sample Type: Liquid/ Air
Sample Temperature
Ambient Temperature
Description of Sampling Location:
Sample No.
Time Sampled
Sample Type: (Liquid\ Air
Sample Temperature
\
'•''Ambient Temperature
Description of Sampling Location:
,
'^|. /.'^.i
-------
PROJECT 1906 3E.'!ZENE/3uP PRESLJRVEY
SAMPLE DATA
Page B-14
Plant
f(? g I "fa i\J.i>,4^ Process ~TOj\. JUljg A\Q .'Grf . Date ?///
' Sample No.
'4-
Time Sampled
6 '
Sample Type: \TIquid Air
m
Sample Temperature
Ambient Temperature
Description of Sampling Location:
(
Sample
f o I
i\v,'j- . ^ I C/A"5) V' ___ ^fime Sampled (p .' "SL
Sample Type: ^Liquid Air
Sample Temperature (f\
Ambient Temperature
Description of Sampling Location
ion: JL^fltJ
! 60--5
Sample No.
KUu-
\r
.•
Time Sampled
(0
Sample Type: iLiquid Air
Sample Temperature
^ I <5?
^ j 1
Ambient Temperature
Description of .Sampling Location:
-------
Plant
PROJECT 1906 BENZENE/BaP PRESURVEY
SAMPLE DATA
Process
Page B-9
Date
Sample No.
Q
e C^t(V ^ | 1,3 Time Sampled M
Sample Type: VLjLquid ) Air
Sample Temperature $ Oj \
Ambient 'Temperature Q ] |
Description of Sampling Location:
P P
Sample No.
Time Sampled
Sample Type: Liquid Air
Sample Temperature
Ambient Temperature
Description of Sampling Location:
Sample No.
Sample Type: Liquid Air
Sample Temperature
Ambient Temperature
Time Sampled
Description of Sampling Location:
-------
Plant
Page B-8
PROJECT 1906 BENZEME/BaP PRESURVEY
SAMPLE DATA
Process
Sample No. I £r j V\
Time Sampled
Sample Type: (Liquid/ Air
Sample Temperature f f ( „
Ambient Temperature
Description of Sampling Location:
t
Sample No. /*--« U •
Time Sampled
Sample Type: \Liquid / Air
Sample Temperature
Ambient Temperature
Description of Sampling Location:
rff
.Sample No.
Sample Type: [Liquid ) Air
Sample Temperature
Ambient Temperature
Description of Sampling Location:
Tine Sampled
\3 /r?
J i
-------
APPENDIX B
LIQUID SAMPLE ANALYSIS
-------
Source/Plant
Tar Decanter (B)
Tar Decanter (BH)
Tar Dehydrator (BH)
Tar Storage (P)
Liquid Sample Analysis
Sample Location Date
Liquor on Surface 7/8/80
7/8/80
Liquor on Surface 9/23/80
Tar from Outlet 9/23/80
Pump
Liquor on Surface 10/7/80
Wash Oil Decanter (F) Oil on Surface 9/10/80 1640
Tar Dehydrator-W (F) Tar-on Surface 9/11/80 1820
Tar Dehydrator-E (F) Tar on Surface 9/11/80 1825
Time
1525
1600
1525
1420
1700
1640
1820
1825
Sample
Temp.(F°)
176
180
120
162
145
205
207
219
Benzene*
Cone . (ppm-wt)
5.6
4.2
181.6
58.1
35.5
1956
1834
2171
16.8
75.4
6.2
42.2
40.3
50.3
629
599
634
282
286
261
*Note: The three concentrations shown for most sources were obtained from three
separate liquid samples collected^consecutively.
-------
APPENDIX C
LABORATORY HEADSPACE DATA
-------
ClIROMATOCRAPHIC ANALYSIS LOG
Project No, /?'£_\J.- y WV Y'A .-.)
-Sv, ^0P ^
J>
Y,, •:./„ ,: -,-'
• ••, • - ••-.
^.0.**" c^/^
c-' ^
~— ' C? .- •••:. \ 1
A
'• '.' "'< ' ' - - ,•
--. --,,..,
•
Teak
Height/Area
&.*-
2
/ /
vO.I
• - - / ' "'
?v^ '
PP.
•
Concentration
Factor
, *
.
/£. 3
/ ? ' 1 ~\
1 ^
.^.S? '
4
Concentration
-
/.o70 ^/o
•
r^"
1 i 1 ^> ,{ •* (O
*
Cc:r. meats
i
,07^ _f _.;
'6t'°"' ^': '
! L!t>' p^ - .
1
*
*
Ml 51 .,n>,~,
* 'i
\ ^ (^ | */ ^
•
.
-------
CHROMATOGRAPHIC ANALYSIS LOG
o.c
Project No, //I';/
Date
"Analyst
Time
,
•
Sample Identification
' •''•••'' •> (. ' :•' •.-''.'. 0 „ ,J
r. '. ( -. <• -. •/""'
'•-ooO N...^xx iO\
< . a
•/ / o .
( '•> *
^.-...^
,'°
f ' '"* ~"
•^ .^o )f'^ ' -i
' ^^^ •
-YW-^0. ^. rt- //P
'^^x'c'';"r^//f!>
^.v.a.V
Peak
Height/Area
/5-"
f(J
^
*-^* *~»
Concentration
Factor
.
Q '.?/
o /^ — "^
o ,'' (I"'"
l'5lo(-r-
s*
^^Tf_ .
P o'&c. |T. .
i^:; .
-------
CHROMATOCRAPHIC ANALYSIS LOG
Project No,
Date
"Analyst
Time
• v
•,
V
Sample Identification
f }
— r^,^- -f— ' ./•- ^ -s
^ ' c" '- ' ' \ ' ' ° '
' ^0 . ,.Jo 0
A
^i,Q ~o.cpr_
/ > // ^ ^
•
•T'0i,jY.aAT ^>>fl-:.r,,
Sv"^- 7^^ """
^ V *
.^XcU- //jo o .F>/
Peak
Height/Area
,'C? •( '
Concentration
Factor
•
•
/o.^:
c" •;;"'>-:> «..
,
Concentration
-
ni/i A.O
-
•
1 '^ =•" y"°
*
Co:r.rr.C:nts
•
orfp.v>
.-
/•*-. :: •:'
... , IT_
;^;;;:/V
"" r- . •>- •-
I i
.
- 1
1
-------
CHROMATOGRAPHIC ANALYSIS LOG
Project Ko,_
Date
"Analyst
Time
Sample Identification
Height/Area
Concentration
Factor
Concentration
CornniGnt
fO
1 !
------- |