EPA REPORT NUMBER 72-PC-13
CD
o
AIR POLLUTION
EMISSION TEST
AMERICAN CAN COMPANY
Halsey, Oregon
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Emission Measurement Branch
Research Triangle Park. North Carolina
-------
SOURCE TEST REPORT
EPA No.: 72-PC-13
Particulate and Gaseous
Emissions From
A Kraft Pulp Mill
AMERICAN CAN COMPANY
Halsey, Oregon
EPA Contract No.:. 68-02-0232
Task No.: 10
Environmental Engineering, Inc.
2324 Southwest 34th Street
Gainesville,. Florida 32601
-------
TABLE OF CONTENTS
Page Number(s)
I. INTRODUCTION ..... . 1
II. SUMMARY AND DISCUSSION OF RESULTS ....... 2-8
TABLE 1 - TRS Daily Averages 4
TABLE 2 - Participate Emissions Recovery Furnace
Outlet 5
TABLE 3 - Participate Emissions Smelt Tank-..'.. . 6
TABLE 4 - 0,,, C09, and CO Daily Mean Concen-
trations 7
TABLE 5 - Nitrogen Oxide Concentrations
Recovery Furnace Outlet ....... 8
III. PROCESS DESCRIPTION AND OPERATION . . . ... . 9-20
Figure 1 - Kraft Pulping Process ........ 10
Figure 2 - Recovery Furnace System ....... 12
Figure 3 - Smelt Dissolving Tank and Scrubber . 15
TABLE 6 - Summary of the Recovery Furnace
Process Data . . 18
TABLE 7 - Summary of Process Data for the
Electrostatic Precipitator ...... 20
IV. LOCATION OF SAMPLING POINTS .......... 21-24
Figure 4 - Outlet of Electrostatic Precipitator
Recovery Furnace ........... 23
Figure 5 - Smelt Dissolving Tank ,,..,,.. 24
V. SAMPLING AND ANALYTICAL PROCEDURES . 25-39
Figure 6 - GC Gas Sampling System 27
Figure 7 - Barton Sampling System ........ 29
Figure 8 - Particulate and S02 Train ...... 32
Figure 9 - C02, Op, and CO Sampling System . . , 37
-------
I. INTRODUCTION .
In accordance with Section 111 of the Clean Air Act as amended
of 1970, the Environmental Protection Agency is charged with the
establishment of performance standards for new stationary sources
which may contribute significantly to air pollution. These standards
are based upon the best air pollution control technology that has been
demonstrated.
, This report presents the results of an extensive source testing
program conducted at American Can Company, Halsey, Oregon, July 13-21,
1972, to obtain data for a partial basis in consideration of new source
performance standards in the kraft pulping industry.
Stack emissions were measured from the chemical recovery boiler
for particulate, sulfur dioxide, reduced sulfur compounds, oxides of
nitrogen, carbon dioxide, carbon monoxide, and oxygen. Emissions from
the smelt dissolving tank vent were also measured for particulates, sulfur
dioxide, reduced sulfur, COp, O^, and oxygen. The recovery boiler is
the direct-fired type and exit gases are controlled with an electrostatic
precipitator. The smelt dissolving tank vent is controlled with a water
scrubber packed with plastic Pall rings.
Reduced sulfur compounds were measured by flame photometric gas
chromatography and coulometric titration. All other stack emissions
were measured with EPA reference methods.
-------
II. SUMMARY AND DISCUSSION OF RESULTS
Table 1 summarizes results of gaseous sulfur determinations
utilizing both flame photometric and coulometric detection systems.
All summary results are reported in terms of TRS as ^S. TRS is
defined as hydrogen sulfide plus methyl mercaptan
plus dimethyl disulfide; all compounds are reported as hydrogen
sulfide. It should also be noted that dimethyl disulfide (RSSR) con-
centrations, determined with the chromatographic system, are assumed
to yield twice those concentrations when considered as hydrogen sulfide.
Results from the particulate emission tests on thfe recovery
furnace are shown in Table 2. Emission rates were calculated on the
basis of the moisture content determined from a separate moisture test
instead of the condensed moisture in the impingers. The reason for
this is that the evaporation rate of the isopropanol in the impingers
was found to be excessive based upon previous tests.
Results from the particulate emission tests on the smelt tank
are shown in Table 3. Emission results from the smelt dissolving tank
were calculated by using the moisture content determined from the
assumption that the stack gases were saturated at the dry bulb temperature.
The third and fourth impingers, which contained 3% hydrogen
peroxide, were analyzed for sulfur dioxide by using barium perchlorate
titrations. The data are also included in Tables 2 and 3.
Complete particulate and_sulfur dioxide data are contained in
Appendix B. ;.-;:p.;:;,^U^2"^T^--- '_.,'-
-------
Daily mean concentrations for oxygen, carbon dioxide, and
carbon monoxide are presented in Table 4. The very low CCL and very
high Op concentrations experienced on July 13 and 14 were the result
of system leaks. On July 15, a new probe was installed and a thorough
system leak check was conducted. Subsequent concentrations are believed
to be representative of actual source concentrations. The results are
reported on a dry gas basis.
All gas concentrations are reported at five-minute intervals and
the maximum, minimum, and mean concentrations with the regression
coefficients for calibration curves are presented in Appendix C.
The results from the nitrogen oxide emission testing are summarized
in Table 5. Complete NO data are included in Appendix B.
J\ '
-------
TABLE 1
TRS DAILY AVERAGES .
COULOMETRIC AND FLAME PHOTOMETRIC DETECTION
AMERICAN CAN COMPANY
Halsey, Oregon
Date
7-13-72
7-14-72
7-15-72
7-17-72
7-18-72
7-19-72
7-20-72
7-21-72
System
Precipitator
Outlet
Precipitator
Outlet
Precipitator
Outlet
Precipitator
Outlet
Precipitator
Outlet
Precipitator
Outlet
Precipitator
Outlet
Smelt Tank
Outlet
Flame Ph
ppm(l)
1.60(2)
0.24(3)
0.51(4)
0.26(4)
0.33(4)
0.40(4)
0.34(4)
2.01(2)
otometric Detection
Ibs/hr
0.71
0.11
0.23
0.12
0.15
0.18
0.15
0.86
Ibs/ADTP*
0.051
0.0079
0.016
0.0086
0.011
0.013
0.011 .
0.062
Cot
ppm
0.34
0.74
0.67
0.72
0.52
0.32
0.35
0.35
lometric Detection
Ibs/hr
0.15
0.33
0.30
0.32
0.23
0.14
0.15
0.15
Ibs/ADTP*
0.011
0.024
0.022
0.023
0.016
0.010
0.011
0.011
(1) Parts per million by volume - dry gas basis
(2) H2S & RSR only
(3) RSR only
(4) RSR & RSSR only
*Based on 334.5 ATDP/day
-------
TABLE 2
PARTICULAR FMISSIONS
RECOVERY FURNACE OUTLET
Run
Date
Time Began
Time End
Barometric Pressure, In. Hg Absolute
Meter Orifice Pressure Drop, In. H20
Vol. Dry Gas @ Meter Conditions, ft
Average Gas Meter Temperature, °F.
Vol. Dry Gas @ S.T.P.*, ft
Stack Gas Moisture, % Volume
% C02
% 02
% CO
* N2
Average Stack Gas Temperature, °F.
Stack Pressure, In. Hg Absolute
Stack Gas Velocity @ Stack Cond., fpm
*
Stack Gas Flow Rate @ S.T.P. , scfm
Net Time of Test, min.
Percent Isokinetic
Particulate Concentrations, grains/scf
Front half and Filter
Total
Particulate Emissions, Ibs/hr
Front half and Filter
Total
Particulate Emissions, Ibs/ton
Front half and Filter
Total
S02 Emissions, Ibs/hr
3rd Impinger
4th Impinger _
**Dry, 70°F., 29.92^in. Hg
Moisture determined from separate test
Run #2
Run #3
7/13/72
1400
1755
30.25
0.93
122.664
83
121.321
**
20.4
12.0
8.0
0.0025
80
395
30.03
3459
85,167
176.5
114.3
0.120"
0.130
87.58
94.87
6.28
6.81
0.06
0.006
7/14/72
1245
1550
30.25
0.63
80.770
85
79.506
**
20.4
11.5
8.5
0.003
80
400
30.03
3428
83,820
144
93.3
0.111
0.123
79.78
88.44
5.72
6.34
Neg.
Neg.
7/15/72
1128
1500
30.10
0.81
91 . 567
88
89.268
**
20.4
11.3
8.5
0.0035
80
415
29.88 '
3555
85,579
144
0.132
0.139
96.55
102.10
6.93
7.32
0.005
0.03
-------
TABLE 3
' . ... PARTICULATE EMISSIONS
SMELT TANK
Run II Run #2 Run #3
Date 7/12/72 7/19/72 7/21/72
Time Began 1450 1435 1301
Time End 1705 1700 1600
Barometric Pressure, In. Hg Absolute 30.2 30.0 30.0
Meter Orifice Pressure Drop, In. H20 '1.57 1.51
Vol. Dry Gas @ Meter Conditions, ft 106.875 120.623
>
Average Gas Meter Temperature, °F. 93.5 101.0
Vol. Dry Gas @ S.T.P.*, ft 102.739 114.660
Stack Gas Moisture, % Volume 25.0*** 24.0*** 22.0***
% C02 0.06 0.06 0.06
% 02 , 20.0 20.0 20.0
% CO v 0.002 0.002 0.002
% N2 79.9 79.9 79.9
Average Stack Gas Temperature, °F. ,150 148 144
Stack Pressure, In. Hg Absolute ! 30.2 30.0 30.0
Stack Gas Velocity @ Stack Cond., fpm ' 2901# 3666# 3437#
4r ' jffc "fffc ' ^tSt
Stack Gas Flow Rate @ S.T.P. , scfm 7,442 7,442 7,442
Net Time of Test, min. : 132 132 132
Percent Isokinetic \ 120.7# 47.0* 68.0*
Particulate Concentrations, grains/scf . .
Front half and Filter " " i 0.047 070630.078
Total . ' ', 0.050 0.066 0.083
Particulate Emissions, Ibs/hr ' \
Front half and Filter i 2.97 4.04 4.95
Total ! 3.18 4.22 5.35
Particulate Emissions, Ibs/tonADP ' - .
Front half and Filter 10.213 0.290 0.355
Total ~ | 0-228 0.303 0.384
S02 Emissions, Ibs/hr - \ Neg. Neg. Neg.
tory, 70° F., 29.92 in. Hg |
^Velocity measurements were affected by'tangential flow in the outlet stack
s flow was determined from inlet stack velocity measurements
Moisture determined by assuming stack gas was saturated
-------
TABLE 4
AMERICAN CAN COMPANY - HALSEY, OREGON
NO. 3 RECOVERY BOILER
02, C02, AND CO
' DAILY MEAN CONCENTRATIONS
Date
*7-12-72
*7-13-72
7-14-72
7-15-72
7-17-72
7-18-72
7-19-72
7-20-72
02 (%)
17.9
16.1
8.1
7.6
7.4
7.7
8.0
8.0
Mean Concentration
C02 (%)
3.4
4.7
12.3
12.4
13.3
12.7
12.0
12.4
(dry basis)
CO (ppm)
88
113
132
85
72
156
90
36
Results due to leak in probe and subsequent dilution with ambient air.
-------
TABLE 5
NITROGEN OXIDE CONCENTRATIONS
RECOVERY FURNACE OUTLET
Date
7/13/72
7/14/72
7/15/72
Time
1340
1630
1810
1230
1450
1600
1055
1335
1500
NOX, ppm
44
38
43
8
43
!6
44
48
63
8
-------
III. PROCESS DESCRIPTION AND OPERATION
The American Can operation at Halsey, Oregon is a complete pulp
and paper mill, producing about 300 tons of bleached kraft pulp per
day. From this, about 200 tons per day of tissue, towels, and napkins
are made; the remaining pulp, about 100 tons per day, is shipped to
other mi 11s.
Process Descri pti on
A. General
Kraft pulp is produced from wood as shown in Figure 1. The wood
used at Halsey is mostly fir, mixed'with less than 5 percent of hemlock,
silver spruce, and cedar. -All the .wood is residual material from
neighboring sawmills. Sawdust and chips are kept segregated and
processed in two separate, continuous digesters..
In the pulping process, wood is cooked in the digesters under
pressure at elevated temperature. The cooking chemicals (a water
solution of sodium hydroxide and sodium sulfide called "white liquor")
chemically dissolve the wood lignin. The freed wood cellulose, or
pulp, is filtered from the spent liquor arid washed. After being
bleached, most of the pulp is made into paper.
The balance of the process is designed to recover cooking
chemicals. Spent cooking liquor and the pulp wash water are combined
for treatment. The combined stream, called weak black liquor, is
concentrated in multiple-effect evaporators. The final two effects
-------
Vent gas
Konconciens shies
1
Wood
5-invite liquor -
~^ fHaOH -v Ka?S).-
DIGESTER
SYSTEM
(KaO
Pulp
&
tu
s
o
UJ
RECOVERY
FURKACE
SYSTEM
^"
Smelt
--- Vlater
SKELT
DISSOLVIMS
I Mi K ;
*
-w'nito liquor
(recycle to
digester)
Green( Liquor
CAUSTICIZIKG
Pulp
Vlater
Ğ,*-
Weak Black Liquor
Noncondensables
, ---'
)vicSansables
! CONCENTRATORS
HOLTIFEE
EFFECT
EVAPORATOR !
SYSTEM
cal ci U;TI
carbonate
-i VT 1
O:\l- I
iA1 '.'
l-K/i! I
10
-------
are'specially designed to handle the .thickened black liquor. These
effects are called "concentrators" and replace the conventional direct
contact evaporator used in most mills. Liquor leaving the concentrators,
containing about 37 percent water, is fed to the recovery furnace. The
organic constituents (principally dissolved lignin) burn, and the heat is
used to generate process steam. Inorganic chemicals in the black liquor
collect at the bottom of the furnace as a molten smelt. The smelt is a
mixture of sodium carbonate and sodium hydroxide. After being tapped from
the furnace it is dissolved in water and transferred to a causticizing tank.
Lime added to this tank converts sodium carbonate to sodium hydroxide, completing
the regeneration of white liquor for recycle to the digesters. A calcium
carbonate mud precipitates from the causticizing tank, and is recycled to
a kiln to regenerate lime.
Two of the above process facilities were tested by EPA; the recovery
furnace system, and the smalt dissolving tank. These facilities are
*
described below.
B. Recovery Furnace System
The recovery furnace system consists of a recovery furnace and an
electrostatic precipitator. The system is shown in Figure 2.
The recovery furnace was designed by Babcqck and Wilcox to burn
1.2 million pounds of black liquor solids per day, which is equivalent !
I
to a pulp production rate of 300 tons per day. Hot black liquor is j
"f
sprayed into the furnace through nozzles located near the furnace j
bottom. Occasionally, when extra heat is needed or when the black
liquor supply is temporarily interrupted, natural gas or oil is burned.
11 ,
-------
ro
Combustion
Air
Recovery Furnace
L
Combustion
Gases
54*
Ji
Mix
Tank
Dust
Electrostatic
P re ci pit a tor
Black
Liquor
-> Smelt
Stack
Figure 2. Recovery Furnace System at the American Can Company Mill in Halsey, Oregon,
-------
Preheated combustion air is blown in at three levels. Beginning
with .the lowest levels, the streams are called primary, secondary,
and tertiary air. The vent gases from the pulp washers, .which contain
odorous gases, are mixed with the combustion air and burned in the furnace.
To utilize the black liquor heat of combustion, the recovery furnace
is constructed with water walls and contains many boiler tubes. Steam
is produced at the rate of about 200,000 pounds par hour, providing part
of the process requirements. The remaining steam needs are met with a
A
conventional gas-fired power boiler. .
The boiler tubes in the recovery furnace accumulate participate from
the combustion gases. These deposits are continually removed by blowing
steam over the tubes. About 10 percent of the steam produced in the
recovery furnace is used in the scot blowers.
The last stages for concentrating black liquor prior to burning are
special evaporators called concentrators. The steam heated concentrators
replace the direct contact craporators used in most other mills. Two
concentrators are used in series., with a third as a standby replacement.
The units are rotated frequently and cleaned while on standby.
Combustion gases leaving the recovery furnace are cleaned in an
electrostatic precipitator and then exhaust from a 300 foot stack.
(Gases from the plant's lime kiln are also discharged through this stack.)
The precipitator was constructed by Research Cottrell. It has a dry bottom
with two parallel chambers and three electrical control sections in each
chamber. The design efficiency is 99,5 percent.
-- ' : . 13 .' . "... -'.:-;; . ',
-------
The material collected in the precipitator is principally a
mixture of sodium sulfate and sodium carbonate. These compounds are
valuable process chemicals. After being removed from the hoppers,"
the precipitated dust is dumped into a mix tank and dissolved in black
liquor for recycle to the furnace. Vent gases from the mix tank are burned
in the furnace to control odors.
Additional sodium sulfate^ to make up for stack losses, is dissolved
into the black liquor just before firing. The normal makeup rate is about
*
700 pounds of sodium sulfate per hour.
*
C. Smelt Dissolving Tank
*
Molten smelt formed in the bottom of the recovery furnace is tapped
off into a tank filled with water. The smelt dissolves in the water to
form "green liquor". The green liquor is then pumped from the tank for
further processing.
Contact with the hot smelt thoroughly agitates the receiving water.
and causes the formation of large amounts of steam. The steam is vented
to the atmosphere through a scrubber to remove participates. The smelt
dissolving tank and scrubber are shown in Figure 3.
The scrubber is packed with plastic cyclinders known as Pall rings.
An induced draft fan draws the gases up through the packing where
they are scrubbed counter/currently with water. The scrubbing water is
a blend of several discharge streams in the mill, but the major portion
is weak wash (water that was used to wash the mud being charged to the
linie kiln). Two water streams leave the scrubber. One stream is recycled
'by blending it with the incoming water;.the second stream goes to the
smelt dissolving tank. This-second stream serves as feed water for the
production of green liquor.
14
-------
A
Weak
Wash
Pall Ring
Packing
Scrubber
Fan
Scrubber
Discharge
o
- Sampling
0
Ports
^>x. ' sS?'
_, \ ''Inlet
Sampling
Ports
Smelt
Dissolving
Tank
Figure 3.. Smelt Dissolving Tank and Scrubber at the American
Call Company Mill in Halsey, Oregon.
15
-------
Process Operat ion
A. General
The purpose of the test program was to measure emission levels
during normal plant operation. Process conditions were carefully
observed, and testing was done only when the test facility appeared to
be operating normally. During the tests, important operating conditions
were monitored and recorded on process data sheets. The process data
are summarized below.
"B. Recovery Furnace .
During the tests, major furnace operating conditions were monitored
from the control room. Half-hourly readings were recorded on process
data sheets. The records and a key explaining the entries are included in
the appendix.
Steam production, black liquor feed rate, and saltcake addition rate
were each monitored on two instruments. One instrument, a chart recorder
gave instantaneous values; the second instrument indicated running totals.
On July 15, comparisons were made between the two instruments. Average
flow rates were computed by dividing the difference in integrator readings
by the time between readings. The average chart flow rate for the same
time interval was determined by inspection and compared. The integrator
readings were found to be 1 percent lower than the chart readings for
steam production and black liquor flow rate. For salt cake audition rate,
however, the integrator readings were 38 percent less than the chart
readings. The Company indicated that the chart readings (item 20 on the
process data sheets) are more reliable.
16
-------
Samples of black liquor (as fired) ware taken during the tests.
Each .day's sample was composed of several portions taken at approximate
intervals of 90 minutes. To inhibit vaporization of water, the hot
black liquor was drawn through a cooling coil. This procedure reduced
the temperature from about 259 to about 206°F. The samples were analyzed
for heat content, percent solids, and pH. Results of the analyses are
given in the Summary of Test Results section of this report.
A green liquor sample was taken from the smelt dissolving tank
^
during each day of testing. The samples were analyzed to determine
reduction ratio, a measure of the conversion of sodium sulfate to
sodium sulfide within the recovery furnace. Analytical results are
given in the Summary of Test Results section of this report.
Sulfidity Bevels during the tests were reported by ilia Company
to be about 22 percent. Normal levels for this mill are reportedly 21
to 22 percent. These low sulfidities account in part for the relatively
low levels of SC^ measured from the recovery furnace.
The observed ranges of major operating conditions during the tests
are sumniaried in Table 6. Black liquor feed rate was very steady
between 128 and 132 gallons per minute. The black liquor solids content
ranged narrowly between 62.0 and 63.9 percent. Black liquor heat content,
determined from six composite samples, varied from 3697 to 3946 BTU per
pound of liquor (wet basis). Total steam production (including what was
used for blowing soot) was normally abcut 210,000 pounds per hour but on
one occasion reached 245,000. Reduction .ratio determined from seven
samples ranged from 79.8 to 87.6 indicating a somewhat low degree of
smelt reduction in the furnace. "
17
-------
Table 6. SUMMARY OF THE RECOVERY ;:URNACE PROCESS DATA
Operating Condition
Black Liquor Feed Rate
Black Liquor Solids Content'0'
Black liquor Heat Content'^
Steam Production -c'
Reduction Ratio'd'
Sulfidity^
Units
GPM
wt. %
BTU/lb liquor
1000 Ib/hr
%
%
Ranae During Test
(July 12-21, 1972)
128 - 132
62.0 - 63.9
3697 - 3946 '
198 - 245
79.8 - 87.0
22
Item Number
On Process
Data Sheets
10
14
-
1
_
-
(a) Black liquor as fired.
(b) Black liquor as fired. Results of six composite samples. Wet basis,
(c) Total production including soot blowing steam.
(d) 100 (Na2S)/(Na2S + Na2 S04). Results of seven samples.
(e) 100 (Ma2S)/(Na2S + NaOH). Company reports.
18
-------
As far as known from the process records and discussions with the
operators, the furnace was operated normally during the tests.
C. Electrostatic Precipitator
The primary voltage and secondary current for each control section
of the precipitator are displayed in the recovery furnace control room.
Readings were taken on the half-hours during testing, and recorded with
the furnace operating conditions on the process data sheets (Appendix).
the readings are summarized in Table 7 for July 13, 14, and 15, the
""days on which participate emissions were sampled.
As far as known from the records and from conversations with the
operators, the precipitator was operating normally during the tests.
D. Smelt Disserving Tank
Particulate emissions from the smalt dissolving tank were measured
on July 12, 19 and 21, 1972. TRS and SQ0 emissions were measured July 21.
During the tests records were kept on the recovery furnace, whose operation
is closely related to the dissolving tank. These records are included
with the other furnace operating data in the appendix.
Green liquor samples taken from the dissolving tank were analyzed
for reduction ratio. The determinations for July 19 and 21 were 84.5
and 87.0 percent respectively. (No sample was taken on July 12,}
The records show that furnace operation during tests on the dissolving
tank were approximately the same as on the other days. As far as known
from thsse records, the smelt dissolving tank was operated normally during
the tests. .... . -. ..-..-,: .-,-....,.,..
19
-------
Table 7. SUMMARY OF PROCESS DATA FOR THE ELECTROSTATIC PRECIPITATOR
Operating Condition
Primary Voltage
Unit 1
: Secondary Current
Primary Voltage
Unit 2
Secondary Current
Primary Voltage
Unit 3
Secondary Current
| During Tests
Units (July 13, 14, 15, 1972]
Volts
Amps
Vol ts
Amps
Volts
Amps
260 - 350
0.4 - 0.9
300 - 380
0.7 - 1.5
320 - 360
1.3-1.8
20
-------
IV. LOCATION OF SAMPLING POINTS
Recovery Furnace
The outlet stack from the electrostatic precipitator on the
recovery furnace was sampled at the rectangular duct entering into
the vertical stack. Figure 4 shows relative dimensions.
The traverse points sampled in each of the three ports are
as follows:
Distance From
Sample Point No. Inside Wall, Inches
1 . 4 7/16.
2 8 13/16
3 13 3/16
4 17 9/16
5 21 15/16
6 26 5/16
7 3§ 11/16
8 35 1/16
' 9 39 7/16
10 43 13/16
11 48 3/16
12 52 9/16
Smelt Dissolving Tank
The outlet stack from the scrubber on the smelt dissolving tank
was sampled at the vertical portion as shown in Figure 5. Because of
tangential flow in the outlet stack, velocity measurements were also
made in the stack 24.3 feet above the dissolving tank.
21
-------
The traverse points used in both the inlet and the outlet
stack are as follows:
Distance From Inside Hall, In.
Sample Point No. Inlet Outlet
1 1 3/8
2 2 1/2 1 3/32
3 51/417/8
^ 4 71/2 23/4
5 10 35/8
6 13 1/4 41/2
7 16 1/2 5 1/2
8 20 3/4 6 11/16
9 26 7/8 7 15/16
10 43 3/8 9 3/8
11 49 3/8 11 1/8
12 53 1/2 13.3/4.
13 56 7/8 20 3/4
14 59 3/4 23-11/32
15 62 1/2 25 1/8
16 64 3/4 26 1/2
17 67 1/2 27 13/16
18 69 28 15/16
19 29 15/16
20 30 7/8
21 31 3/4
22 32 19/32
23 33 3/8
24 34 1/8
22
-------
Particu1 ate
Sampling
Ports
Sample Port
4" Pipe Nipple
OUTLET -OF ELECTROSTATIC PRECIPITATOR
s ' RECOVERY FURNACE
Figure 4
23
-------
Vİ
l/\x-V--\x"^'-\-J
Gas Sampling Port
Participate Sampling Ports (Scrubber Outlet)
Inlet From Scrubber
Flow
Scrubber Inlet _1
Sampling Port ~*
-70 'ID
Scrubber By-Pass
With Damper
Tank
~*r~
SMELT DISSOLVING TANK VENT
v" Figure 5
24
-------
V. SAMPLING AND ANALYTICAL PROCEDURES
Chromatographic Sampling System
Figure 6 illustrates the system which was employed in conveying
the gases from the source to the sensing equipment. The stainless steel
probe and Teflon sampling line were maintained at temperatures exceeding
the dew point of the source gases. The sampling line consisted of an
insulated, electrically heated 1/4-inch Teflon tube. The sample gases
were transmitted to the heated dilution box where they were split into
two separate streams. One stream was conveyed to the vacuum source and
wasted to minimize lag time in the sampling line. The remainder of the
flow was diluted with nitrogen by an amount sufficient to lower the dew
point of the gases below ambient temperature. A portion of this diluted
sample was injected into the chromatograph through the Gas/Liquid
Chromatograph (GLC) sampling valve. The remainder of the diluted gas
was wasted through the vacuum source.
Chromatographic Analysis .
Gaseous sulfur concentrations were determined with a Tracer
Model 250 Gas/Liquid Chromatograph. This unit is equipped with a flame
photometric detector which is specifically for sulfur compounds. Two
analytical columns were utilized in the separation and analysis of the
gaseous sulfur compounds. One was a 36-foot by 1/8-inch OD Teflon
column packed with polyphenyl ether liquid phase on a solid support of
grannular Teflon with stripper column. The second column, constructed
of identical materials, was 8 feet long. Both columns were operated at
50°C.
25
-------
The 36-foot column was utilized for analyzing hydrogen sulfide,
sulfur dioxide, and methyl mercaptan while the 8-foot column facilitated
the analysis of dimethyl sulfide and dimethyl disulfide.
The chromatograph was calibrated for hydrogen sulfide, sulfur
dioxide, methyl mercaptan, dimethyl sulfide, and dimethyl disulfide,
using the spinning syringe technique.
26
-------
Stack
tNJ
--J
Glass
Wool
Flow
Dilution
Mi trogen
Heated Sampl|e
1/4Teflon
Ji I :,
i I xf
GC GAS SAMPLING SYSTEM
Figure 6
. (5?) Carri er Gas
1 ^^ (N2)
Gas
Chromatograph
GC
Sampling Valve
Vacuum Pump
RM: Rotameter
: Metering Valve
-------
Coulometric Detector (Barton Titrator)
Figure 7 illustrates the system which was employed in conveying
the gases from the source to the Barton Titrator. The stainless steel
probe and Teflon sampling line were maintained at temperatures exceeding
the dew point of the stack gases. The sampling line was the same as
the sampling line used with the GLC. The sample gases were transmitted
to the Barton Titrator by a vacuum source.
Barton Titrator
Total reduced sulfur (TRS) concentrations were analyzed using
a Barton Titrator, Model 400. Furnace gases were scrubbed through a
3% solution of potassium acid phthalate (KHP) which removes sulfur dioxide
and a large fraction of water vapor from the sample gases. The sample
gas was then introduced to a coulometric titration cell which utilizes
hydrobromic acid (HBr) as an electrolyte. The electrolytic cell
generates bromine from the HBr electrolyte which reacts with the
sulfur compounds entering the titration cell. The quantity of current
required to generate the excess bromine, to consume the sulfur compound,
is proportional to the gaseous sulfur concentrations introduced. The
current required to operate the titration cell is sensed and trans-
mitted to a recorder where a continuous readout is accomplished. The
recorded output is converted to TRS concentrations, as hLS from cali-
bration data generated with the "spinning syringe" technique.
28
-------
ro
Stack
Flow
Heated
Sample Line
1/4" Teflon
so2
Scrubbers
Data
Recorder
Barton
Titrator
BARTON SAMPLING SYSTEM
Figure 7
Flow
Meter
v
Micro
Metering
Valve
Vacuum Pump
-------
Sampling Procedure for Particulate Emissions
Prior to performing the actual participate emission tests,
certain preliminary stack parameters had to be determined for the
r
stack gases. This preliminary data included the average stack gas
temperature, velocity head, moisture content, stack dimensions, and
number of sampling points.
The stack gas temperature^as determined by using bimetallic
thermometers and a pyrometer. .'.'..
The approximate stack gas moisture content selected for
setting the nomograph was based upon previous tests made'on the same
boiler. The final moisture content used in calculating the stack
emissions from the recovery furnace was based upon the amount of conden-
sate collected in the impingars and the silica gel from a separate
moisture test. The moisture content of the scrubber outlet gases from
the smelt dissolving tank used for calculating emission rates was de-
termined by assuming that the gases were saturated at the dry bulb
temperature.
The sampling points selected and the respective stack gas
velocities were determined by using Methods No. 1 and 2 of the Federal
Register (Vol. 36, No. 247, December 23, 1971). Velocity head measure-
ments were made by using a calibrated S-type pi tot tube with an inclined
manometer.
Because of the excessive gas turbulence at the sampling
location for the scrubber outlet from the smelt dissolving tank, velocity
measurements were made at the scrubber inlet so that a gas flow rate
30
-------
could be determined at the outlet. It was assumed that there were no
leaks in the system and that the gas flow rate at the outlet was equal
to the inlet. Therefore, all emission rates were based upon this one
gas flow rate.
The pi tot tube was rotated at each traverse point in the
smelt tank scrubber outlet stack so that the approximate direction of the
gas flow could be determined. After-this preliminary determination,
each point was sampled with the nozzle aligned to the direction of the
k
upstream gas flow. After the first test run, the EPA project officer
requested that we sample about 50% below isokinetic conditions so that
the particulate emissions would be biased high rather than low.
The sampling train configuration used during the tests con-
sisted of the following: a stainless steel nozzle; a heated glass-
lined probe; a heated glass-fiber filter; two Greenburg-Smith impingers
with tips, each containing 100 ml of 80% isopropanol; two Greenburg-
Smith impingers without tips, each containing 100 ml of 3% hydrogen
peroxide; one Greenburg-Smith impinger without a tip, containing about
200 grams of silica gel; a flexible sample line; an air-tight vacuum
pump; a dry-test meter; and finally a calibrated orifice with an in-
clined manometer (see Figure 8). Velocity head measurements were
conducted simultaneously with the sampling at each point so that each
point could be sampled isokinetically.
The impinger portion of 'the sampling train was iced down to
collect the condensables, and to determine the actual stack gas moisture.
31
-------
4 I TJ
rr
20
1. Stainless steel nozzle
2. Glass-lined probe (heated)
3. Heated box (250°F)
'4. Glass-fiber filter and holder
5. Ice bath
6. Impinger with Tip, 100 ml of 80% Isopropanol
7. Impinger with Tip, 100 ml of 80% Isopropanol
8. Impinger without Tip,.100 ml of 3% H,00
/J L. C.
9. Impinger without Tip, 100 ml ,ft
. of-3% H909
c. C. -*:
10. Impinger with 200 grams
of Silica Gel
11. Thermometer
12. Flexible sample line
13. Vacuum gauge
14. Coarse valve
15. Fine valve
16. Vacuum pump
17. Drg-test meter
18. Calibrated orifice
19. Inclined manometer
20. S-type pi tot .tube
FIGURE 8
PARTICIPATE AND S02 TRAIN
-------
Sample Recovery and Analyses of Particulates
Sample recovery from the particulate train was accomplished
by separating train components into the following containers:
Container No. 1 - The previously weighed glass-fiber filter
was placed into this container, then sealed and labeled.
Container No. 2 - All port%ns of the train from the nozzle
through the front half of the filter holder were rinsed with
acetone and the contents placed into a glass container, then
sealed and labeled.
Container No. 3 - The volume of liquid from the first and
second impingers was measured and the contents placed into
a glass container. Also, all sample-exposed surfaces between
the filter and third impinger were rinsed with 80% isopro-
panol and placed into this container, then sealed and labeled.
Container No. 4 - The volume of liquid from the third and
fourth impingers was measured and the contents placed into
separate glass containers. All glassware between the second
and fifth impingers was then rinsed with deionized, dis-
tilled water and then added to each respective container.
The liquid samples were then sealed and labeled. Only one
sample container was used for both impingers used in the
smelt dissolving tank sampling. -
Container No. 5 - The previously weighed silica gel was re-
moved from the fifth impinger and placed into the original
polyethylene jar and sealed.
33
-------
The samples recovered were analyzed as follows:
Container No. 1 - The filter and all loose material in the
sealed petri dish were transferred to a tare glass weighing
dish, desiccated, and dried to a constant weight.
Container No. 2 - The acetone washings were transferred to a
tared beaker and evap&mted to dryness at ambient temperature
and pressure. It was desiccated and dried to a constant weight.
Container No. 3 - The contents were transferred to a tared
beaker, and then evaporated at 212°F. The residue was desic-
cated and dried to a constant weight.
Container No. 4 - The liquid contents were shaken, and then
a 25 ml aliquot of each container was pipetted into separate
250 ml Erlenmeyer flasks. One hundred ml of isopropanol,
plus two to four drops of thorin indicator was added to each
V sample. The samples were titrated with barium perchlorate
to a pink end point. Another duplicate sample and blank was
titrated in the same manner as the first sample. Samples
were analyzed at the plant site.
Container No. 5 - The spent silica gel was weighed at the site
and recorded.
The filter from Container No. 1, and the beakers from Containers
No. 2 and 3 for each run were sent to the EPA project officer after the
initial analysis for additional analyses.
34
-------
Sampling System for Oxygen, Carbon Dioxide, and Carbon Monoxide
Figure 9 is a schematic diagram of the sampling system.
Source gases were drawn continuously through a glass-lined probe and
polyethylene tubing to a moisture trap consisting of silica gel im-
pingers immersed in an ice bath. Valves on the pressure side of the
vacuum pump controlled the flow of sample gas to the detectors. A
bleed valve was provided to maintain adequate purging of the sample
line. Gases to the oxygen and carbon monoxide detectors were passed
through an ascarite bed to remove carbon dioxide which potentially
interferes with the NDIR determination of carbon monoxide. Sample to
the carbon dioxide detector was diluted with nitrogen to accommodate
the range requirements of the detector.
35
-------
A Beckman model F-3 paramagnetic oxygen analyzer capable of
measuring 0 - 25% oxygen was used for Op detection. Beckman NDIR
models 315-B and 315-A, respectively, were used in determining carbon
dioxide and carbon monoxide concentrations. The instrument configura-
tions were 0-5% C02 and 0 - lOOOppm CO. All instruments were supplied
with a low, medium and high range which were calibrated separately. The
detectors were switched on and allowed to run continuously 24 hours per
day for the entire sampling interval.
^ ' . -'.'
Calibration
All calibration gases were supplied and analyzed by Matheson
Gas Products, Inc., Morrow, Georgia, and La Porte, Texas. The calibra-
tion procedure was conducted prior to sampling each morning and was re-
peated at the end of each day. Nitrogen was introduced into each
instrument and the zero control was adjusted to obtain a steady "zero
trace" on the recorder. Appropriate standards were then passed into
the instruments at less than 100 cc/minute. The gain controls for each
range were adjusted to provide maximum deflection and accuracy.
Daily Operation
Each morning after calibrating the instruments, charging the
traps, and checking the probe, the system was assembled as shown in
Figure 9. The dilution to the carbon dioxide detector was regulated
to provide an accurate deflection range at a total flow rate less than
TOO cc/minute. Flows for sample gas and dilution nitrogen were measured
with a bubble tube. The recorder traces were observed and the ranges
36
-------
From Stack
Ice Bath
j_
, Ascarite
Impinger
A
CO
Cal
Gas
A
co2
Cal
Gas
Silica
Gel
Midget Impinger
w/Silica Gel
To CO System
BleedV
T
Vacuum Pumo
To CO, 0 System
CO
NDIR
°2
Paramagnetic
Analyzer
Flowmeter
Dilution System
-CKh
co2
NDIR
C02, 02, and CO SAMPLING SYSTEM
Figure 9
37
Flow
Cal
Gas
Vent
Vent
Dilution
Nitrogen
-------
were adjusted throughout the sampling interval as required. Occasionally
sampling was interupted to obtain odor samples, charge traps, or to check
the system.
Data Reduction
A computer program, developed by Environmental Engineering,
Inc., was used to reduce recorder deflections to specific gas concentra-
tions. Average calibration deflections and respective calibration - gas
concentrations were entered for each range, and a second order regression
analysis was computed. The general form of the regression equation is:
y = A + Bx + Cx2
where y = measured gas concentration, % or ppm
A, B, C = regression coefficients as calculated
x = recorder deflection, millimeters.
Strip chart data was reduced to computer input sheets which
are included in the Appendix C. The carbon dioxide con-
centration was calculated using the above regression curve and the mean
dilution factor. Oxygen and carbon monoxide concentrations were deter-
mined in like manner and corrected for carbon dioxide absorption. The
correction factor for oxygen and carbon dioxide was determined from the
simultaneous carbon dioxide concentration using the relationship:
Actual 02, CO =
Rioo-%co2)
100 J
measured 02, CO
38
-------
Sampling Procedure for Nitrogen Oxides
Nitrogen oxide concentrations of the recovery furnace outlet
gases were determined by using the EPA Method 7, which is described
in the Federal Register (Volume 36, No. 247, December 23, 1971).
Essentially, the method consisted of collecting a grab sample
of the gas in an evacuated 2-liter flask containing a dilute sulfuric
acid-hydrogen peroxide absorbing solution. The sample remained in the
flask at least 16 hours, and was then placed in a glass storage bottle.
Sodium hydroxide (IN) was then added to the sample until alkaline. The
samples were taken back to the laboratory in Gainesville, Florida, and
measured colorimetrically using the phenoldisulfonic acid procedure.
39
------- |