United States Office of Air Quality EPA-450/3-83-017
Environmental Protection Planning and Standards September 1983
Agency Research Triangle Park NC 27711
Air
Review of New
Source Performance
Standards for Kraft
Pulp Mills
-------
EPA-450/3-83-017
Review of New Source Performance
Standards for Kraft Pulp Mills
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
September 1983
-------
This report has been reviewed by the Emission Standards and Engineering Division of the Office of Air Quality
Planning and Standards, EPA, and approved for publication, Mention of trade names or commercial products is not
intended to constitute endorsement or recommendation for use. Copies of this report are available through the Library
Services Office (MD-35), U.S. Environmental Protection Agency, Research Triangle Park, N.C. 27711, or from
National Technical Information Services, 5285 Port Royal Road, Springfield, Virginia 22161.
-------
TABLE OF CONTENTS
Page
LIST OF ILLUSTRATIONS viii
LIST OF TABLES ix
1. EXECUTIVE SUMMARY 1_1
1.1 BEST DEMONSTRATED CONTROL TECHNOLOGY 1-1
1.2 ECONOMIC CONSIDERATIONS AFFECTING THE NSPS 1-2
1.3 RECOMMENDATIONS ON REVISION OF CURRENT MSPS 1-3
1.4 .RECOMMENDATIONS ON EXTENSION TO THE NSPS 1-4
2. THE KRAFT PULPING INDUSTRY 2-1
2.1 INTRODUCTION 2-1
2.2 BACKGROUND INFORMATION 2-1
2.3 DESCRIPTION OF INDIVIDUAL PROCESS FACILITIES AND
THEIR EMISSIONS 2-4
2.3.1 Digester System 2-4
2.3.2 Brown Stock Washer System 2-7
2.3.3 Multiple-Effect Evaporator System 2-10
2.3.4 Recovery Furnace System 2-12
2.3.5 Smelt Dissolving Tank 2-17
2.3.6 Lime Kiln 2-17
2.3.7 Black Liquor Oxidation System 2-20
2.3.8 Condensate Stripping Sytem 2-21
2.4 INDUSTRY CHARACTERIZATION 2-21
2.5 SELECTION OF KRAFT PULP MILLS FOR NSPS CONTROL 2-21
2.6 REFERENCES 2-23
3. CURRENT STANDARDS FOR KRAFT PULP MILLS 3-1
3.1 AFFECTED FACILITIES 3-1
3.2 CONTROLLED POLLUTANTS AND EMISSION LEVELS 3-1
3.3 NSPS IMPACT ON EXISTING KRAFT PULP MILLS 3-2
3.4 STATE REGULATIONS 3_2
3.5 TESTING AND MONITORING REQUIREMENTS 3-3
3.5.1 Testing Requirements 3-3
3.5.2 Monitoring Requirements 3-4
4. STATUS OF CONTROL TECHNOLOGY 4-1
4.1 PARTICULATE CONTROL 4-1
4.1.1 Recovery Furnace 4_1
4.1.2 Smelt Dissolving Tanks 4-6
4.1.3 Lime Kiln 4_6
-------
TABLE OF CONTENTS (Continued)
Page
4.2 TRS CONTROL 4-7
4.2.1 Recovery Furnace System 4-7
4.2.2 Digester and Multiple-Effect Evaporator Sytems 4-8
4.2.3 Brown Stock Washer System 4-8
4.2.4 Black Liquor Oxidation System 4-9
4.2.5 Smelt Dissolving Tank 4-9
4.2.6 Lime Kiln 4-9
4.2.7 Condensate Stripping System 4-10
4.3 REFERENCES 4-11
5. COMPLIANCE TEST RESULTS 5-1
5.1 ANALYSIS OF NSPS COMPLIANCE TEST RESULTS 5-1
5.1.1 Recovery Furnace System 5-1
5.1.2 Smelt Dissolving Tanks 5-3
5.1.3 Lime Kiln 5-5
5.2 REFERENCES 5-8
6. COST ANALYSIS 6-1
6.1 UPDATE COST FOR THE AFFECTED FACILITIES 6-1
6.1.1 Particulate Control 6-3
6.1.2 TRS Control 6-3
6.2 REFERENCES 6-7
7. ENFORCEMENT ASPECTS 7-1
7.1 REFERENCES 7-3
8. ANALYSIS OF POSSIBLE REVISIONS TO THE STANDARDS 8-1
8.1 POSSIBLE REVISIONS TO NSPS 8-1
8.1.1 Recovery Furnace Systems 8-1
8.1.2 Smelt Dissolving Tanks 8-4
8.1.3 Lime Kiln 8-5
8.1.4 Digester and Multiple-Effect Evaporator Systems 8-7
8.1.5 Brown Stock Washer Systems 8-7
8.1.6 Black Liquor Oxidation Systems 8-8
8.1.7 Condensate Stripping Systems 8-8
8.1.8 Exemption for Brown Stock Washers and Black
Liquor Oxidation Systems 8-9
8.2 POSSIBLE REVISIONS TO MONITORING REQUIREMENTS 8-9
8.3 EXTENSION TO OTHER SOURCES 8-10
8.4 EXTENSION TO OTHER EMISSIONS 8-11
8.5 REFERENCES 8-13
IV
-------
TABLE OF CONTENTS (Concluded)
9. CONCLUSIONS AND RECOMMENDATIONS 9-1
9.1 REVISION OF THE CURRENT STANDARDS 9_1
9.1.1 Conclusions Based on Control Technology 9-1
9.1.2 Conclusions Based on Economic Considerations 9-1
9.1.3 Conclusions Based on Other Considerations 9-2
9.1.4 Recommendations on Revision of Current Standards 9-2
9.2 EXTENSION OF STANDARDS 9.3
9.2.1 Conclusions Based on Control Technology 9-3
9.2.2 Recommkendations on Extension of Standards 9-3
APPENDIX A. Kraft Pulp Mills Subject to the NSPS A-l
-------
LIST OF ILLUSTRATIONS
Figure Number Page
2-1 Kraft Pulping Process 2-3
2-2 Batch Digester Flow Sheet 2-8
2-3 Continuous Digester Flow Sheet 2-9
2-4 Continuous Diffusion Washer 2-11
2-5 Direct Contact (Conventional) Recovery
Furnace System with Black Liquor Oxidation 2-14
2-6 Indirect Contact Recovery Furnace System 2-15
2-7 Fluidized Bed Calcining System 2-19
• i •
vi
-------
LIST OF TABLES
Table Number
2-1 TRS Emissions from An Uncontrolled
907 Megagrams (1,000 Tons) Per Day
Kraft Pulp Mill 2-5
2-2 Typical Concentrations and Emission Rates
For Particulate Matter From Uncontrolled
Kraft Pulp Mill Sources 2-6
4-1 Design of Electrostatic Precipitators
Subject to the NSPS 4-2
*
4-2 Particulate Emission Data for the No. 4
Recovery Furnace at the Tacoma, Washington,
Mill, 1973 Through 1981 4-5
5-1 Compliance Test Results of Recovery
Furnaces Subject to NSPS 5-2
5-2 Compliance Test Results of Smelt Dissolving
Tanks Subject to NSPS 5-4
5-3 Compliance Test Results of Lime Kilns
Subject to NSPS 5-6
6-1 Capital Costs, Annualized Costs, and
Cost-Effectiveness 6-2
vn
-------
1. EXECUTIVE SUMMARY
The new source performance standards (NSPS) for kraft pulp mills were
promulgated by the Environmental Protection Agency (EPA) on February 23,
1978. These standards affect recovery furnaces, smelt dissolving tanks,
lime kilns, digester systems, multiple-effect evaporator systems, brown
stock washer systems, black liquor oxidation systems, and condensate
stripper systems. Affected facilities are those facilities which commenced
construction or modification after September 24, 1976.
The objective of this report is to review the NSPS for kraft pulp
mills, and to assess the need for revision on the basis of developments
that have occurred since the original standard was promulgated. The
following paragraphs summarize the results and conclusions of the analysis,
as well as recommendations with respect to EPA action in implementing
changes in the NSPS.
1.1 BEST DEMONSTRATED CONTROL TECHNOLOGY
The NSPS limits emissions of particulate matter and total reduced
sulfur (TRS) compounds. The particulate sources covered are the recovery
furnace, the smelt dissolving tank, and the lime kiln. These three facilities,
along with power boilers, account for virtually all of the particulate
matter emissions from a kraft pulp mill. Emissions of TRS are limited from
eight affected facilities: the digester system, the brown stock washer
system, the multiple-effect evaporator system, the black liquor oxidation
system, the recovery furnace, the smelt dissolving tank, the lime kiln, and
the condensate stripper system. These eight facilities account for virtually
all of the TRS emissions from a kraft pulp mill.
No changes have occurred in the selection of control technologies used
for either the particulate facilities or the TRS sources. The original
particulate standards were based on the use of an electrostatic precipitator
(ESP) for recovery furnaces, a venturi scrubber for lime kilns, and a wet
scrubber for smelt dissolving tanks. These are the control devices installed
on facilities subject to the NSPS. One ESP has been installed on a lime
kiln. ESP's on lime kilns were investigated during the NSPS development,
but were not used as the basis for the NSPS.
1-1
-------
The original TRS standards were based on incineration of the noncondensable
gas streams from the digesters, evaporators, washers, black liquor oxidation
systems, and condensate strippers; type of water used for smelt dissolving
tanks; process controls and either black liquor oxidation or noncontact
evaporators for recovery furnaces; and process controls, good mud washing,
and caustic scrubbing for lime kilns.
Most lime kilns subject to the NSPS have demonstrated compliance with
the TRS standard without caustic addition. Presently, the mills are not
required to continuously monitor the TRS emissions. Until continuous
monitoring data are available, the ability of the present control systems
to continuously achieve the TRS standard without caustic scrubbing cannot
be determined.
Fluidized bed calciners subject to the NSPS have demonstrated compliance
with both the particulate and TRS standards. Fluidized bed calciners are
covered by the NSPS, but emission data demonstrating their performance were
not included in the NSPS data base due to the limited use of fluidized bed
calciners at the time of the NSPS development.
Compliance test results for facilities subject to the NSPS show that
all are in compliance with the NSPS, except for two recovery furnaces and
one smelt dissolving tank which have not been able to meet the TRS standard.
1.2 ECONOMIC CONSIDERATIONS AFFECTING THE NSPS
The capital and annualized costs for each control system were estimated
during the NSPS development. During this review, those costs were updated
to February 1982 using the Chemical Engineering plant cost index. In
addition, the costs for the recovery furnaces were revised based upon
information supplied by the companies on actual recovery furnace system
operations. These updated and revised costs were compared with actual
industry costs on control systems used to achieve the NSPS. For all cases
for which actual industry data were available, the costs are similar.
The cost-effectiveness for each control system was estimated from the
updated and revised costs and the emission reduction resulting from applica-
tion of the control techniques. For each of the particulate control techniques,
the incremental cost effectiveness over the economic recovery level was less
1-2
-------
than $1,000 per ton. The cost effectiveness of the various TRS controls
ranges from zero dollars per ton for the smelt dissolving tank to $9,250
per ton for the black liquor oxidation system. The brown stock washer
system with a cost-effectiveness of $2,500 per ton is the only other
affected facility with a cost effectiveness above $1,000 per ton. However,
this cost was based on the use of a vacuum drum type washing system which
was in general use during the NSPS development. Many of the washing
systems subject to the NSPS are the diffusion types which are a closed
reactor and have a much lower gas volume. Based on capital costs obtained
from one mill, the cost effectiveness for control of diffusion type
washers is estimated to be $900 per ton of TRS controlled.
Since the NSPS was proposed, 14 new recovery furnaces, 16 new smelt
dissolving tanks, and 19 new lime kilns have started operation. The growth
rate in terms of pulp production averaged 3.5 percent per year between 1978
and 1981. The forecast is for a 2.3 percent decline in overall industry
production in 1982, but a 4.6 percent growth in 1983.
1.3 RECOMMENDATIONS ON REVISION OF CURRENT NSPS
There has been general compliance with the current NSPS with achievability
of the existing standards adequately demonstrated. Based upon technological
conclusions, it is recommended that the basis for the NSPS remain unchanged.
It is recommended that the TRS standard for smelt dissolving tanks be
raised to the level indicated by compliance data. It is also recommended
that the units of the TRS standard for smelt dissolving tanks be revised to
allow the use of EPA method 16A. It is recommended that the TRS standard
for black liquor oxidation systems be rescinded. A revision is also recommended
to rescind the requirement for monitoring the combustion temperature when a
device other than an incinerator is used to destroy TRS emissions.
1-3
-------
1.4 RECOMMENDATIONS ON EXTENSION TO THE NSPS
It is recommended that standards for other processes and pollutants
not be considered at the present time. Recovery furnaces and lime kilns
can be sources of nitrogen oxides, sulfur dioxide, and carbon monoxide, but
control of these pollutants has not been demonstrated for the lime kiln
and cost-effective control of sulfur dioxide emissions has not been demonstrated
for recovery furnaces in the kraft pulp industry. It is recommended that
a study be performed to investigate the sulfur dioxide emission control
methods available for recovery furnaces and to determine whether regulation
is appropriate.
Two sources of emissions not covered by the present NSPS are power
boilers and treatment ponds. Power boilers will be regulated under standards
being proposed for industrial boilers. Methods of measuring TRS emissions
from water treatment ponds are not available.
1-4
-------
2. THE KRAFT PULPING INDUSTRY
2.1 INTRODUCTION
The United States Environmental Protection Agency (EPA) proposed new
source performance standards for kraft pulp mills under Section 111 of the
Clean Air Act on September 24, 1976 (41 FR 42012). These regulations were
promulgated on February 23, 1978 (43 FR 7568). The regulations applied to
recovery furnaces, smelt dissolving tanks, lime kilns, digester systems,
multiple-effect evaporator systems, black liquor oxidation systems, brown
stock washer systems, and condensate stripper systems, the construction or
modification of which commenced after September 24, 1976.
The Clean Air Act Amendments of 1977 require that the Administrator of
the EPA review and, if appropriate, revise established standards of perform-
ance for new stationary sources at least every 4 years. The purpose of
this report is to review and assess the need for revision of the existing
standards for kraft pulp mills based on developments that have occurred or
are expected to occur within the kraft pulping industry. The information
presented in this report was obtained from reference literature, discussions
with industry representatives, trade organizations, process and control
equipment vendors, EPA regional offices, and State and local agencies.
2.2 BACKGROUND INFORMATION
Manufacturing of paper and paper products is a complex process which
is carried out in two distinct phases: the pulping of the wood and the
manufacture of the paper. Pulping is the conversion of fibrous wood into
a "pulp" material suitable for use in paper, paperboard, and building
materials. Of the two phases involved in paper-making, the pulping process
is the largest source of air pollution. The kraft or sulfate pulping
process produces over 80 percent of the chemical pulp produced annually in
the United States. The remaining 20 percent of the chemical pulp is produced
by the sulfite and neutral sulfite semi-chemical (NSSC) processes. These
pulping processes were investigated as potential candidates for NSPS develop-
ment in a 1978 study entitled "Screening Study on Feasibility of Standards
2-1
-------
of Performance for Two Wood Pulping Processes." This study concluded that
no new growth was predicted at that time and, therefore, an NSPS was not
recommended to be developed.1
Pulp wood can be considered to have two basic components, cellulose
and lignin. The fibers of cellulose, which comprise the pulp, are bound
together in the wood by the lignin. To render cellulose usable for paper
manufacture,, the pulping process must first remove the lignin.
The kraft pulping process is shown in Figure 2-1. In the process,
wood chips are cooked (digested) at an elevated temperature and pressure in
"white liquor", which is a water solution of sodium sulfide (Na,,S) and
sodium hydroxide (NaOH). The white liquor chemically dissolves lignin from
the wood. The remaining cellulose (pulp) is filtered from the spent
cooking liquor and washed with water. Usually, the pulp then proceeds
through various intermittent stages of washing and possibly bleaching,
after which it is pressed and dried into the finished product (paper).
The balance of the kraft process is designed to recover the cooking
chemicals and heat. Spent cooking liquor and the pulp wash water are
combined to form a weak black liquor which is concentrated in a multiple-
effect evaporator system to about 55 percent solids. The black liquor is
then further concentrated to 65 percent solids in a direct-contact evaporator,
which evaporates water by bringing the liquor in contact with the flue
gases from the recovery furnace, or in an indirect-contact concentrator.
The strong black liquor is then fired in a recovery furnace. Combustion of
the organics dissolved in the black liquor provides heat for generating
process steam and converting sodium sulfate (Na^SOJ to Na2S. To make up
for chemicals lost in the operating cycle, salt cake (sodium sulfate) is
usually added to the concentrated black liquor before it is sprayed into
the furnace. Inorganic chemicals present in the black liquor collect as a
molten smelt at the bottom of the furnace.
The smelt, consisting of sodium carbonate (Na2CO~) and sodium sulfide,
is dissolved in water to form green liquor which is transferred to a
causticizing tank where quicklime (CaO) is added to convert the sodium
carbonate to sodium hydroxide. Formation of the sodium hydroxide completes
the regeneration of white liquor, which is returned to the digester system.
A calcium carbonate mud precipitates from the cauticizing tank and is
calcined in a lime kiln to regenerate quicklime.
2-2
-------
T
ca
r>
o,
CO
cc.
cc
NONCONDENSABLES
i
VENT GAS
WHITE LIQUOR
(NaOH + Na2S) ~~
— ^~
DIGESIER " " PULP k*
^ SYSTEM
VENT GAS
j VENT C
RECOVERY HEAVY BLAC
FURNACE BLACK LIQUO
SYSTEM LIQUOR OXIDAT
,^^..M. »._.... _...___. T A IM If
"* ••" 1 AIMK
(OPTIOI
-*• AIR
1 ~ T"
SMELT '
(Na2C03 + Na2S) A'R
1 VENT GAS
I i
SM
WATER — *- DISSO
TA
GREEN
(WHITE LIQUOR
/nrrvn r xn
• rnwnFNQATP
1 VINP bUIMUCDIOMI t
MK STRIPPER
SYSTEM
|
LIQUOR |
\ AIR
\ 1 1 L I/ I b L L 1 U ••
DIGESTER) CAUSTICIZING' LIME \
IAI\IK ~* ^
CALCIUM
. TAnnnwATF .
MUD
_ — pill p
PULP *"" IULI
WASHERS
—a. WATFR
— -*4:"B " VVMILIl
••
IA/FAK Rl APR 1 inUHP
iAS NONCONDENSABLES
i ,
K
««»^
MAI i - EVAPORATOR _
MAL) * SYSTEM
) VENT GAS
». TO TREATMENT POND
^*. CONDENSATE
STREAM i
1 VENT GAS
Figure 2-1. Kraft Puloina ProrP<;«:
-------
2.3 DESCRIPTION OF INDIVIDUAL PROCESS FACILITIES AND THEIR EMISSIONS
The atmospheric emissions from the kraft process include both gaseous
reduced sulfur compounds and particulate matter. The reduced sulfur
compounds are hydrogen sulfide (H2S), methyl mercaptan (CH3SH), dimethyl
sulfide (CH3SCH3), and dimethyl disulfide (CH3SSCH3). The particulate
matter emissions are primarily sodium sulfate from the recovery furnace,
sodium salts and calcium compounds from the lime kiln, and sodium compounds
from the smelt dissolving tank.
Hydrogen sulfide and the other organic sulfides, when taken as a
group, are called total reduced sulfur (TRS). They are extremely odorous
arid are detectable at a concentration of only a few parts per billion.
Thus, odor control is one of the principal air pollution problems in a
kraft pulp mill.
Summaries of values on typical gas flow rates, variations in uncontrolled
TRS concentrations, and uncontrolled emission rates per unit production for
the kraft process units subject to the NSPS are presented in Table 2-1.
These emission rates are for a 907 megagrams per day (1,000 tons per day)
kraft pulp mill. A summary of typical ranges in uncontrolled particulate
concentrations and emission rates from process units subject to the NSPS is
presented in Table 2-2.
2.3.1 Digester System
In a digester, wood chips are cooked with white liquor at elevated
temperatures of about 170° to 175°C and pressures ranging from 6.9 to 9.3 x
10 pascals (100 to 135 psig). There are two types of digester systems:
batch and continuous. Most kraft pulping is presently done in batch
digesters, although continuous digesters generally are being installed in
the industry for new pulping capacity.
In a batch digester, gases formed during digestion are vented to
"relieve" the digester and maintain proper cooking pressure. Relief takes
place more or less continuously during digestion. At mills pulping turpentine-
containing softwoods, the relief gases are first cooled to condense and
recover turpentine before venting. The condenser cooling water recovers
the heat and may be used in some other process. At the end of the cooking
cycle, the contents of the digester are transferred to an atmospheric tank
usually referred to as a blow tank. The entire contents of the blow tank
2-4
-------
Table 2-1. TRS EMISSIONS FROM AN UNCONTROLLED 907 MEGAGRAMS (1 000 Tons)
PER DAY KRAFT PULP MILL 2 '
Source
Recovery Furnace
Digester System
Multiple-Effect
Evaporator System
en
Lime Kiln
Brown Stock
Washer System
Black Liquor
Oxidation System
Smelt Dissolving Tan
Condensate Stripper
System
Typical
Exhaifst Gas
Flow Rate
m /a(acfm)
212 (450,000)
3 (6,200)
1 (2,200)
37 (79,200)
71 (150,000)
14 (30,000)
27 (58,100)
2 (4,000)
TRS Emission Range
ppm
18-1303
1525-30,000
92-44,000
3-613
-
3-335
5-811
-
g/kg ADP' (Ib/T ADP )
0.75-31 (1.5-62)
0.24-5.3 (0.47-10.5)
0.015-3.2 (0.03-6.3)
0.01-2.1 (0.02-4.2)
0.005-0.5 (0.01-0.9)
0.005-0.37 (0.01-0.73)
0.007-1.9 (0.013-3.70)
-
Average TRS Emission Rate
Ppm
550
9,500
6,700
170
30
35
60
5,000
g/kg ADP (Ib/T ADP)
7.5 (15.0)
0.75 (1.5)
0.5 (1.0)
0.4 (0.8)
0.15 (0.3)
0.05 (0.1)
0.1 (.0.2)
1.0 (2.0)
-------
Table 2-2. TYPICAL CONCENTRATIONS AND EMISSION RATES FOR
PARTICULATE MATTER FROM UNCONTROLLED KRAFT PULP MILL SOURCES3'1*
Recovery Furnace
Lime Kiln
Smelt Dissolving Tank
ro
i
en
Concentration
g/dscfm(gr/dscf)
3.10-22.79 (1.35-9.95)
5.85-33.96 (2.44-14.81)
0.89-13.60 (0.39-5.94)
Emission Rate
g/kg ADP (Ib/ton
22.1-283 (44.1-565)
6.4-49.9 (12.7-99.8)
0.1-11.9 (0.19-23.7)
-------
are normally sent to the pulp washers where the spent cooking liquor is
separated from the pulp. Steam and other gases that flash from the blow
tank are piped to a heat recovery unit. This blow of the digester is not
applicable to continuous digester systems. Figure 2-2 shows a batch
digester and its blow heat recovery system.
In a continuous digester, an uninterrupted flow of wood chips passes
through the unit where steam and cooking liquor are added. Before entering
the digester, the wood chips are fed to a pre-steaming vessel for wetting
and deaeration. Most existing units have a countercurrent wash zone in the
bottom of the digester. The wash liquor reduces the temperature of pulp-
liquor mixture resulting in a so-called "cold blow". The'spent liquor is
drawn off and transferred to a flash tank where the liquor is expanded, or
flashed. The flash steam from the flash tank is usually used to impregnate
the chips in the pre-steaming vessel. The pre-steaming vessel relief,
which contains the noncondensable gases from the flash tank and from the
pre-steaming vessel, then passes to a condenser. When pulping softwoods,
the condensate is separated in a decanter to recover turpentine. In some
cases, a continuous diffusion washing stage (described in the following
section) is integrated with the digester. A continuous digester arrangement
is presented in Figure 2-3.
The noncondensable gases from the relief system and the blow tank vent
contain TRS concentrations as high as 30,000 ppm.^ Both streams are sometimes
referred to as digester "noncondensables". TRS compounds formed in the
digester are mainly methyl mercaptan, dimethyl sulfide, and dimethyl
disulfide. Uncontrolled TRS emissions range between 0.24 and 5.25 g/kg ADP
(0.5 and 10.5 Ib/ton ADP) and average about 0.75 g/kg ADP (1.5 Ib/ton ADP)
at a concentration of 9,500 ppm.4 Operating variables that affect digester
TRS emissions and, therefore, account for the range of TRS emissions shown
in Table 2-1 include black liquor recycle rate, cook duration, cooking
liquor sulfidity (percentage of sodium sulfide to total alkali, Na?S and
NaOH, in white liquor), and residual alkali level.5
2.3.2 Brown Stock Washer System
Pulp from the digester system passes from the blow tank to the washing
plant. The two main types of washers used for removal of black liquor are
2-7
-------
(V)
cx>
12
13
14
15
SECONDARY
CONDENSER
COND.
11 HOT WATER
10 COOLING WATER
(9) RELIEF GAS
(8) TURPENTINE
2) FOUL CONDENSATE
f?) BLOW GAS
HEAT
EXCHANGER
5 HOT WATER
4 WARM WATER
(5) BLOW CONDENSTATE BLEED
2 FRESH WATER
fT) PULP - LIQUOR
Points of Possible Odor Release are Encircled by O
Figure 2-2. BATCH DIGESTER FLOW SHEET3
-------
14 STEAM
15 COND.
PRESTEAMING
VESSEL
13 WHITE LIQUOR
12 CHIPS
D VENT
11 HOT WATER
4 10 COOLING WATER
f) TURPENTINE
2) POULCONDENSATE
16) FLASH STEAM
10J WEAK LIQUOR
PULP +LIQUOR
WASH LIQUOR
POINTS OF POSSIBLE ODOR RELEASE ARE ENCIRCLED BY Q
Figure 2-3. CONTINUOUS DIGESTER FLOW SHEET3
2-9
-------
rotary-drum vacuum washers and diffusion washers. Vacuum washers are the
most common type of kraft pulp washers, but diffusion washing is generally
integrated with the continuous digesters used in newer mills.
In a vacuum washing system, the pulp first passes through the knotter
which removes chunks of wood not digested during cooking. The pulp then is
washed countercurrently with water in several sequential stages. On
leaving each stage, the pulp is dewatered on a vacuum filter, and the water
drains into filtrate tanks. The washers are normally hooded to collect the
vapors that steam off the open washers.
Continuous diffusion washing usually takes place in a closed reactor.
Figure 2-4 illustrates the unit. Inside the wall of the unit are pairs of
cylindrical screen plates. Each pair of screen plates comprises a double-
sided screen. Pulp enters the bottom of the unit and passes slowly upward
between the screens. The wash medium, water, is introduced to the pulp by
a set of nozzles. Each nozzle travels a circular path through the pulp,
midway between adjacent screens. The wash water flows radially from each
nozzle in both directions to the adjacent screens, thus displacing the
chemical-laden water from the pulp. The extractant leaves the pulp via the
perforations in the screen plates and flows downward to the drainage system.
As the washed pulp rises above the screen section, it is removed by scrapers,
TRS emissions from the washers arise primarily from the vaporization
of the volatile reduced sulfur compounds. TRS compounds emitted are
principally dimethyl sulfide and dimethyl disulfide.4 Uncontrolled TRS
emissions from a vacuum drum washing system average about 0.14 g/kg ADP
(0.27 Ib/T ADP) at a concentration of 30 ppm. TRS emissions are affected
by the wash water source, water temperature, degree of agitation and
turbulence in the filtrate tank, and blow tank pulp consistency.6 TRS
emissions will increase significantly if contaminated condensate from the
digester and evaporator systems is used for washing. Higher temperatures
and agitation result in increased stripping of TRS during the washing.5
TRS emission data for diffusion type washers are presently not available.
2.3.3 Multiple-Effect Evaporator System
Spent cooking liquor from the digester system is combined with the
brown stock washer discharge to form weak (dilute) black liquor. Multiple-
effect evaporators are utilized to concentrate the weak black liquor from
2-10
-------
SCREEN
PULP OUTLET
^-3^^ ... *-RM '^S^^-^-T T' f;
WASH WATER INLET
SCRAPER
HYDRAULIC CYLINDER.
FLEXIBLE CONNECTOR-
EXTRACTANT COLLECTION HEADER
ULP LAUNDER
NOZZLE
WASHER WALL
RETENTION TOWER
Figure 2-4. CONTINUOUS DIFFUSION WASHER7
2-11
-------
an initial 12 to 18 percent solids to a final level of 40 to 55 percent
solids. Concentration of the black liquor is necessary to facilitate
combustion of the dissolved organic material in the recovery furnace.
Usually, five or six evaporation units (effects) make up the system. Each
effect consists of a vapor head and a heating element. Hot vapors from the
vapor head of a previous effect pass to the heating element of the following
effect. The effects are operated at successively lower pressures, which
causes a decrease in the boiling point of the liquor. Vapors from the
final effect are condensed rapidly enough to maintain a high vacuum. Two
types of barometric condensers are used: direct contact condensers and
surface condensers. Each type of condenser is equipped with a steam ejector
to remove noncondensables.
The evaporation system will also contain concentrators for the
final evaporation of black liquor to combustion strength if the associated
recovery furnace is not equipped with a direct-contact evaporator.
The noncondensable gases from a multiple-effect evaporator system
consist of air drawn in through system leaks and reduced sulfur compounds
that were either in the dilute black liquor or formed during the evaporation
process. TRS emissions are as high as 44,000 ppm.^ Uncontrolled TRS
emissions average about 0.5 g/kg ADP (1.0 Ib/T ADP) at a concentration of
6,700 ppm.1*
The type of condenser used can influence the concentration of TRS
emissions. Certain types of condensers (e.g., direct-contact) allow the
noncondensable gases and the condensing water to mix, which adds water to
the condensates and results in more noncondensable gases being dissolved in
this water, depending on the water temperature. This reduces the TRS
emissions from the system, but increases the sulfide level in the condensate.
Sulfidity and pH of the weak black liquor also have an effect on the TRS
emissions. Higher sulfidity levels result in higher TRS emissions. TRS
levels increase with decreasing pH levels.
2.3.4 Recovery Furnace System
The purposes of burning concentrated black liquor in the kraft recovery
furnace are: (a) to recover sodium and sulfur, (b) to produce steam, and
(c) to dispose of unwanted dissolved wood components in the liquor. In
most instances, liquor of 60 to 65 percent solids content will burn in a
self-supporting combustion.
2-12
-------
The recovery furnace theoretically is divided into three sections:
the drying zone, the reducing zone, and the oxidizing zone. The black
liquor is introduced to the furnace through spray guns located in the
drying zone. The heat in the furnace is sufficient to evaporate the
remaining water from the liquor. The dried solids fall to the hearth to
form the char bed.
Combustion of the black liquor char begins on the hearth of the
furnace. Air for combustion is supplied by a forced-draft system to the
reducing and oxidizing zones of the furnace. Since a reducing atmosphere
is required to convert sodium sulfate and other sodium-based sulfur compounds
to sodium sulfide, only a portion of the air required for complete combustion
is supplied to the char bed through the lower or primary air ports. The
heat released by the combustion in the zone is sufficient to liquefy the
chemicals in the char and to sustain the endothermic reduction. The liquefied
chemical, or molten smelt is continuously drained from the furnace hearth.
Air is admitted through secondary and tertiary air ports above the
primary zone to complete the combustion of the volatile gases from the char
in the furnace.
There are two main types of recovery furnace systems. The first type
employs a direct-contact evaporator to provide the final stage of evapora-
tion for the black liquor. This is accomplished by bringing the black
liquor in direct contact with the furnace's exhaust gases. This furnace
type is called a conventional or direct-contact system. A conventional
system is shown in Figure 2-5. The second type of recovery furnace system
employs an indirect-contact evaporator as the final evaporation stage; this
type is called a noncontact, direct-fired, or "low odor" system. A noncontact
system is shown in Figure 2-6. The majority of the furnace systems in
operation are the conventional type, although newer installations tend to
be noncontact recovery furnace systems.
In addition, so-called cross-recovery is practiced at several mills.
In practice, the waste liquor from neutral sulfite semi-chemical (NSSC)
cooking is combined with the black liquor from the kraft mill prior to
burning. The inorganic content of the NSSC liquor will join the bulk of
inorganics in the smelt from the furnace, substituting for the sodium
sulfate normally added in the kraft recovery cycle to cover losses of
chemicals.
2-13
-------
ro
i
K
Air
Sme", t
RECOVERY FURNACE
Combustion
Exhaust
Gas
DIRECT CONTACT
EVAPORATOR
->
PARTICULATE
CONTROL
DEVICE
T
Vent
Gases
BLACK LIQUOR
OXIDATION
TANK
Black Liquor
Figure 2-5. Direct Contact (Conventional) Recovery Furnace System With Black Liquor Oxidation
-------
rx>
i
Air
Smelt
1
1
RECOVERY FURNACE
r
Solids
_Combust1j)n_gas
INDIRECT
CONTACT '
EVAPORATOR
57%
PARTICULATE
CONTROL
OFVICF ...
Black Liquor
Exhaust
Gas
Figure 2-6. Indirect Contact Recovery Furnace System
-------
The particulate levels from a recovery furnace prior to a direct-
contact evaporator or control device normally range from 18 to 27 g/dscm
(8 to 12 gr/dscf).14 A direct-contact evaporator acts as a participate
control device and reduces the particulate emission from a furnace system
by about 20 to 40 percent.3 The particulate emissions from uncontrolled
recovery furnace systems average about 8.7 g/dscm (3.8 gr/dscf, 180 Ib/T
AUP).4 Data on particle size of uncontrolled emissions indicate that
90 percent of the particles from a direct-contact recovery furnace and
78 percent from a non-contact recovery furnace are 10 ym or less in size.8
The particulate matter emitted consists of sodium sulfate and sodium carbonate
and may contain small amounts of sodium chloride. Sodium chloride will be
present if the pulpwood has been stored in saline water or if the make-up
chemicals contain chloride impurities.
TRS emissions are generated both in the furnace and in the direct-
contact evaporator. The furnace-generated TRS concentration is as high as
several hundred ppm and as low as 1 ppm, depending on the furnace design
and operation. Recovery furnace TRS emissions are affected by the relative
quantity and distribution of combustion air, rate of solids (concentrated
black liquor) feed, spray pattern and droplet size of the liquor fed,
turbulence in the oxidation zone, smelt bed disturbance, and the combination
of sulfidity and heat content value of the liquor fed.2-14 The impact of
these variables on TRS emissions is independent of the absence or presence
of a direct-contact evaporator.k
TRS emissions generated in the direct-contact evaporator depend largely
on the concentration of sodium sulfide in the black liquor. Acidic gases
such as carbon dixoide in the flue gas can change the black liquor equilibrium,
resulting in the release of increased quantities of hydrogen sulfide and
methyl mercaptan.
Uncontrolled TRS emissions from conventional recovery furnace systems
range from 0.75 to 31 g/kg ADP (1.5 to 62 Ib/T ADP) and average about 7.5
g/kg ADP (15 Ib/T ADP).4 This is an average of about 550 ppm. TRS emissions
from uncontrolled indirect-contact furnace systems would be somewhat less
because these furnaces do not have direct-contact evaporators (which can
contribute up to 15 g/kg ADP [30 Ib/T ADP] of the total uncontrolled TRS
emissions shown for direct-contact furnace systems3).
2-16
-------
2.3.5 Smelt Dissolving Tank
The smelt dissolver is a large tank located below the recovery furnace
hearth. Molten smelt (sodium carbonate and sodium sulfide) that accumulates
on the floor of the furnace is dissolved in water to form green liquor in
the tank. The tank is equipped with an agitator to assist dissolution, and
a steam or liquid shatterjet system to break up the smelt stream before it
enters the solution. Contact of the molten smelt with the water causes the
evolution of large volumes of steam, which must be vented.
The sodium carbonate (Na2C03) in the green liquor is converted to
sodium hydroxide (NaOH) in the causticizing tank. This is done by adding
quicklime (CaO) to the liquor. The quicklime forms calcium hydroxide,
Ca(OH)2, which reacts with rta2C03; calcium carbonate precipitates out and
is converted back into quicklime in a lime kiln.
Particulate matter (finely divided smelt) is entrained in the vapor
that leaves the tank. Uncontrolled emissions may be as high as 4 g/kg ADP
(8 Ib/T ADP).1* Twenty-four percent of the particles from an uncontrolled
smelt dissolving tank are 10 ym or less in size.8
Because of the presence of a small percentage of reduced sulfur
compounds in the smelt, some of these odorous materials escape the tank
with the flashed steam. Uncontrolled TRS emissions are as high as 2.0 g/kg
ADP (811 ppm) and as low as nondetectable.4 The average is about 0.1 g/kg
ADP (0.05 Ib/T AOP).
Several factors affect the TRS emissions. Among these are the water
used in the smelt tank, turbulence of the dissolving water, scrubbing
liquor used in the particulate control device, pH of the scrubbing liquor,
and sulfide content of the particulate collected in the control device.6
The use of contaminated condensate in the smelt tank or the scrubber can
result in the stripping of TRS compounds into the gas stream.
2.3.6 Lime Kiln
The lime kiln is an essential element of the closed-loop system that
converts green liquor (solution of sodium carbonate and sodium sulfide) to
white liquor. The kiln calcines the lime mud (calcium carbonate which
precipitates from the causticizer) to produce calcium oxide (quicklime,
CaO). The quicklime is wetted (slaked) by the water in the green liquor
solution to form calcium hydroxide, Ca(OH)2, for the causticizing reaction.
2-17
-------
The kraft pulp industry typically uses large rotary kilns that are
capable of producing 36 to 360 megagrams (40 to 400 tons) of quicklime per
day. Lime mud is fed in at the elevated end as a 55 to 60 percent solid-
water slurry. The mud is contacted by hot gases produced by the combustion
of natural gas or fuel oil and proceeding through the kiln in the opposite
direction. Large motors turn the entire kiln at low speeds (1-2 rpm),
causing the lime to proceed downward through the kiln toward the high-
temperature zone (980 to 1090°C; 1800 to 2000°F) to discharge at the lower
end. As the mud moves along, it dries in the upper section, which may be
equipped with chains or baffles to give the wet mud better contact with the
gases. As the lime mud moves down farther, it agglomerates into small
pellets and finally is calcined to calcium oxide in the high-temperature
zone near the burner. Some of the rotary lime kilns recently installed
have been equipped with tube coolers for heat recovery.
Fluidized bed kiln systems have found limited application in the kraft
pulping industry. Figure 2-7 shows a fluidized bed system. The system is
divided into two basic parts: the drying system in which the mud is dried
to form dry, powdery feed for the calciner, and the calcining system which
produces the lime. The lime mud is fed to a cage mill disintegrator along
with calciner stack gases. The dried powder is swept upward in the gas
stream to a cyclone collection system. The dry powder collected in the
cyclones is discharged to a storage bin which feeds the calciner. Calcination
takes place in a two compartment reactor. The upper fluid bed is used for
calcination and the lower fluid bed cools the calcined product. A positive
displacement type blower is used to supply air to the reactor for fluidization,
which also serves as combustion air for fuel burning.
Particulate emissions consist principally of sodium salts, calcium
carbonate, and calcium oxide. The sodium salt emission results primarily
from sodium compounds that are retained in the lime mud because of less
efficient or incomplete washing. The calcium particles result from entrainment.
Uncontrolled particulate emissions average about 40 g/kg ADP (80 Ib/T ADP)
at a concentration of 22.2 g/dscm (9.7 gr/dscf).1* Data on particle size
indicate that 25 percent of the particles from an uncontrolled lime kiln are
10 urn or less in size.8
TRS emissions can be generated in the lime kiln proper and may also
be released from the liquor in the downstream scrubber which is normally
2-18
-------
SCRUBBER
Waler
FLUOSOLIDS
CALCINER
CALCINER
FEED
\
CALCINER
FEED BIN
FUEL (Gas or Oil)
COOLING COMPARTMENT
ENTRALIZED
CONTROL
PANEL
PELLETIZEDLIME
-Watef
FLUIOIZING
if AIR BLOWER
FILTER
aiBsaiHiaMWBi^^
SLAKER
Figure 2-7. FLUIDIZED BED CALCINING SYSTEM8
-------
installed to control particulate emissions. TRS emissions originating in
the lime kiln are affected by the oxygen content of the exhaust, the kiln
length to diameter ratio, the lime mud sulfide content, cold-end exit gas
temperature, and simultaneous burning of sulfur bearing materials contained
in the lime mud (e.g., green liquor dregs, the impurities resulting from
clarifying the green liquor).10 If digester and evaporator condensates are
used as scrubbing water, reduced sulfur compounds can be stripped into the
exit gas stream. If the scrubbing liquor contains sodium sulfide, H^S may
be released in the scrubber as a result of the equilibrium shift caused by
the absorption of CCL in the liquor.
Uncontrolled TRS emissions average about 0.4 g/kg ADP (0.8 Ib/T ADP)
at a concentration of 170 ppm. TRS emissions range between 3 and 600 ppm
(0.02 to 4.2 Ib/T ADP) depending on combustion characteristics of the
individual kilns.4
2.3.7 Black Liquor Oxidation System
Black liquor oxidation is applied to facilitate the control of TRS
emissions from direct contact recovery furnace systems. Black liquor
oxidation is designed to decrease the TRS emissions from the direct contact
evaporator by producing a negligible sodium sulfide concentration in the
black liquor by oxidizing the sodium sulfide in either weak or strong
black liquor to sodium thiosulfate or possibly higher oxidation states.
In those mills which oxidize black liquor, air is most often used. However,
molecular oxygen has been used instead of air at some mills. Sparging
reactors, packed towers, and bubble tray columns have been used in singular
or multiple stages to provide intimate contact between the liquor and air.
TRS emissions from the oxidation system are created by the stripping
of the reduced sulfur compounds from the black liquor by passing air
through the liquor. Uncontrolled TRS emissions (principally dimethyl
sulfide and dimethyl disulfide) are in the range of 0.005 to 0.37 g/kg ADP
(about 3 to 335 ppm) and average 0.05 g/kg ADP (35 ppm).4 Oxidation
systems that use only molecular oxygen have the advantage of emitting
virtually no off-gases because the .total gas stream reacts in the sparge
system.
2-20
-------
2.3.8 Condensate Stripping System
When digester and multiple-effect evaporator off-gases are condensed,
some TRS gases are partially dissolved in the condensate. To prevent the
release of kraft odor (TRS) from either the recycled water or the water
treatment ponds, the TRS compounds can be stripped from the digester and
multiple-effect evaporator condensates prior to being recycled or discharged
to the ponds. The two principal ways of stripping are air stripping and
steam stripping. Stripping can be performed in multistage (multiple tray)
columns with a large countercurrent flow of air or steam.
Actual TRS emission data are unavailable, but TRS emissions from
condensate strippers are expected to be high because the condensate contains
high concentrations of dissolved TRS compounds. Uncontrolled TRS emissions
are estimated to range from 0.3 to 5.2 g/kg ADP (0.15 to 2.6 Ib/T ADP) and
average 0.6 g/kg ADP (1.1 Ib/T ADP, 2,800 ppm).11
2.4 INDUSTRY CHARACTERIZATION
There are approximately 121 existing kraft pulp mills operating in the
United States. The locations of these mills are distributed among 28
States, with the Southeast, Northwest, and Northeast being the areas of
greatest density.
The main product of the kraft pulping industry is wood cellulose or
pulp. About 35,894,000 megagrams (39,574,000 tons) of kraft pulp were
produced in 1981.12 The first quarter 1982 market price for bleached kraft
pulp was 474 to 518 dollars per air dried megagram (430 to 470 dollars per
air dried ton).13
Between 1956 and 1975, the growth rate of the industry was 5.5 percent
per year. The growth rate averaged 3.5 percent per year between 1978 and
1981. The forecast is for a 2.3 percent decline in overall industry
production for 1982 and a 4.6 percent growth in 1983.13
2.5 SELECTION OF KRAFT PULP MILL FOR NSPS CONTROL
Kraft pulp mills were originally selected for NSPS development because
they can be significant sources of total reduced sulfur (TRS) compounds and
particulate matter. At the time of NSPS development, the nationwide
emissions of TRS exceeded 181,000 megagrams (200,000 tons) in 1973; emissions
of particulate matter totaled 363,000 megagrams (400,000 tons) during the
same year. The growth rate was projected to be 2.5 percent per year
2-21
-------
between 1975 and 1978, with a return to a higher rate by 1980.llf As stated
above, the actual growth rate between 1978 and 1981 was 3.5 percent per
year.
2-22
-------
2.6 REFERENCES
1. Screening Study on Feasibility of Standards of Performance for Two
Wood Pulping Processes, U.S. Environmental Protection Agency, Publication
No. EPA-450/3-78-111, November 1978.
2. Kraft Pulping - Control of TRS Emissions from Existing Mills, U.S.
Environmental Protection Agency, Publication No. EPA-450/2-78-003b,
March 1979.
3. Environmental Pollution Control—Pulp and Paper Industry, Part I Air,
U.S. Environmental Protection Agency, EPA-625/7-76-001, October 1976.
4. Atmospheric Emissions from the Pulp and Paper Manufacturing Industry,
EPA-450/1-73-002, September 1973.(Also published by NCASI as Technical
Bulletin No. 69, February 1974).
5. Control of Atmospheric Emissions in the Wood Pulping Industry,
Volume I, Environmental Engineering, Inc., Gainesville, Florida, and
J.E. Sirrine Company, Greenville, South Carolina, Contract No. CPA 22-69-18,
March 1970.
6. Factors Affecting Emission of Odorous Reduced Sulfur Compounds from
Miscellaneous Kraft Process Sources, NCASI Technical Bulletin No. 60,
March 1972.
7. Kamyr Continuous Diffusion Washer, Kamyr Inc., Glens Falls, New York,
Bulletin No. 600.
8. Particulate Mass and Particle Size Measurements at Representative Kraft
Pulp Mills, Volume 1 (Draft), U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina, January 1982.
9. Dorr-Oliver Fluosolids System For Lirne Mud Reburning in Connection
with Recausticizing in Kraft Mills, Dorr-Oliver Incorporated, Stamford,
Connecticut, Bulletin No. 7550-F.
10. Suggested Procedures for the Conduct of Lime Kiln Studies to Define
Emissions of Reduced Sulfur Through Control of Kiln and Scrubbing Operating
Variables, NCASI Special Report No. 70-71, January 1971.
11. Contaminated Condensate Stripping—An Industry Survey, McCance, K.E.
and Burke, H.G., Pulp and Paper Canada, Vol 81, No. 11, November 1980.
12. Paper Trade Journal, April 30, 1982.
13. Paper Trade Journal, May 30, 1982.
14. Standards Support and Environmental Impact Statement, Volume 1:
Proposed Standards of Performance for Kraft Pulp Mills, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, Publication
No. EPA-450/2-76-014a, September 1976.
2-23
-------
-------
3. CURRENT STANDARDS FOR KRAFT PULP MILLS
3.1 AFFECTED FACILITIES
The existing new source performance standards (NSPS) for new, modified,
and reconstructed kraft pulp mills limit emissions from recovery furnaces,
smelt dissolving tanks, lime kilns, digester systems, multiple-effect
evaporator systems, black liquor oxidation systems, brown stock washer
systems, and condensate stripping systems. The standards governing recovery
furnaces apply to both straight kraft recovery furnaces and cross recovery
furnaces.
3.2 CONTROLLED POLLUTANTS AND EMISSION LEVELS
The NSPS limits emissions of total reduced sulfur (TRS) and particulate
matter. The standards are as follows:
Particulate Matter
Recovery furnace—Exhaust gases discharged into the atmosphere shall
not contain particulate matter in excess of 0.10 g/dscm (0.044 gr/dscf)
corrected to 8 percent oxygen, and shall not exhibit 35 percent opacity or
greater.
Smelt dissolving tank—Exhaust gases discharged into the atmosphere
shall not contain particulate matter in excess of 0.1 g/kg black liquor
solids (dry weight) [0.2 Ib/ton black liquor solids (dry weight)].
Lime kiln—Exhaust gases discharged into the atmosphere shall not
contain particulate matter in excess of:
a) 0.15 g/dscm (0.067 gr/dscf) corrected to 10 percent oxygen,
when gaseous fossil fuel is burned.
b) 0.30 g/dscm (0.13 gr/dscf) corrected to 10 percent oxygen,
when liquid fossil fuel is burned.
TRS
0 Straight kraft recovery furnace—Gases discharged into the atmosphere
shall not contain TRS in excess of 5 ppm by volume on a dry basis, corrected
to 8 percent oxygen.
3-1
-------
0 Cross recovery furnace—Gases discharged into the atmosphere shall
not contain TRS in excess of 25 ppm by volume on a dry basis, corrected to
8 percent oxygen.
0 Smelt dissolving tank—Gases discharged into the atmosphere shall not
contain TRS in excess of 0.0084 g/kg black liquor solids (dry weight)
[0.0168 Ib/ton black liquor solids '(dry weight)].
0 Lime kiln—Gases discharged into the atmosphere shall not contain
TRS in excess of 8 ppm by volume on a dry basis, corrected to 10 percent
oxygen.
0 Digester system, brown stock washer system, multiple-effect evaporator
system, black liquor oxidation system, or condensate stripper system--
Gases discharged shall not contain TRS in excess of 5 ppm by volume on a
dry basis, corrected to 10 percent oxygen.
3.3 NSPS IMPACT ON EXISTING KRAFT PULP MILLS
The Clean Air Act applies to three general categories of pollutants
emitted from stationary sources. These are criteria pollutants, hazardous
pollutants, and designated pollutants. At existing sources, the first two
categories can be controlled under Sections 108-110, or 112 of the Act.
The third category consists of pollutants that are (or may be) harmful to
public health or welfare but are not or cannot be controlled under Sections
108-110 or 112. Section lll(d) requires control of existing sources of
such pollutants whenever standards of performance (for those pollutants)
are established under Section 111 for new sources of the same type. The
NSPS developed and promulgated for kraft pulp mills include emission limits
for TRS and particulates. TRS is a designated pollutant and, therefore,
under Section lll(d), States are required to develop TRS emission standards
for the affected facilities. Since the EPA determined that TRS emissions
will be considered a we!fare-related pollutant, States have greater flexibility
in establishing plans for the control of TRS.
3.4 STATE REGULATIONS
A survey was conducted of current State air quality regulations controlling
kraft pulp mills. State standards governing particulate matter from existing
kraft mills are generally less stringent than the Federal NSPS. State
opacity standards are, however, similar to the NSPS. State regulations for
3-2
-------
TRS emissions have been developed in accordance with Section lll(d) of the
Clean Air Act. The State standards are generally similar to the lll(d)
guidelines, which are identical to the NSPS except that no control is
required for BLO systems and brown stock washer systems, and the TRS level
for lime kilns is 20 ppm instead of 8 ppm.
3.5 TESTING AND MONITORING REQUIREMENTS
3.5.1 Testing Requirements^
Performance tests to verify compliance with the particulate and TRS
standards for the affected facilities must be conducted within 60 days
after achieving full capacity operation, but not later than 180 days after
the initial startup of the facility (40 CFR 60.8). The EPA reference
methods to be used in connection with the affected facilities include:
1. Method 5 for the concentration of particulate matter and the
associated moisture content.
2. Method 1 for sample and velocity traverses.
3. Method 3 for gas analysis.
4. Method 9 for visible emissions.
5. Method 16 for the concentration of TRS.
For Method 5, the sampling time for each run is at least 60 minutes
with a minimum sampling rate of 0.85 dscm/hr (0.53 dscf/min). Water is
used as the cleanup solvent instead of acetone in the sample recovery
procedure outlined in Method 5.
When determining compliance for the smelt dissolving tank, Method 2,
for velocity and volumetric flow rate, and the black liquor solids feed '
rate are also used. The following equation is used to determine the TRS
emission rate:
E = (CH2SFH2S + CMeSHFMeSH + CDMSFDMS + CDMDSFDMDS) (QSd) / 8LS
wnere: E = mass of TRS emitted per unit of black liquor solids (g/Kg)
CH2S = average concentration of hydrogen sulfide (H2S), ppm
CMeSH = average concentration of methyl mercaptan (MeSH), ppm
CDMS = average concentration of dimethyl sulfide (DMS), ppm
CDMDS = average concentration of dimethyl disulfide (DMDS), ppm
FH s = 0.001417 g/(m3 ppm)
FMeSH = °-00200 g/(m3 ppm)
3-3
-------
= 0.002583 g/(m3 ppm)
FDMDS = °-003917 9/(m3 Ppm)
Qsd = standard dry volumetric stack gas flow rate (dscm/hr)
3LS = black liquor solids feed rate, kg/hr
Method 17 (in-stack filtration) may be used as an alternate method for
Method 5 for determining compliance with the NSPS, provided' that a constant
value of 0.009 g/dscm (0.004 gr/dscf) is added to the results of Method 17
and the stack temperature is no greater than 205°C (400°F). Water is also
used as the cleanup solvent in the sample recovery procedure instead of
acetone, which is outlined in Method 17.
Method 16A (Impinger Technique) has been proposed as an alternative
method for Method 16 for determining compliance with the NSPS. This method
utilizes the impinger collection method and barium-thorin titration producer
outlined in EPA Method 6. Method 16A measures all reduced sulfur in contrast
to the four reduced sulfur compounds specified in Method 16. Therefore,
Method 16A may give higher results than Method 16, but the EPA expects any
difference to be insignificant.
All concentrations of particulate matter and TRS resulting from these
measurements are corrected to 10 volume percent oxygen for lime kilns and
to 8 volume percent oxygen for recovery furnaces.
3.5.2 Monitoring Requirements
A continuous monitoring system is required to monitor and record the
opacity of gases discharged into the atmosphere from the recovery furnace.
For lime kilns and smelt dissolving tanks using a scrubber, a monitoring
device is required for the continuous measurement of the pressure loss of
the gas stream through the scrubber and of the scrubbing liquid supply
pressure to the scrubber.
Continuous monitoring systems are required under the NSPS to monitor
and record the concentration of TRS emissions on a dry basis and the percent
of oxygen by volume on a dry basis in the gases discharged from recovery
furnaces and lime kilns. However, the TRS monitoring system requirement
will not be effected until performance specifications for the monitoring
system are promulgated. The performance specifications were proposed
on July 20, 1981 (46 CFR 37287).
3-4
-------
A monitoring device which measures the combustion temperature at the
point of incineration is required for the noncondensable gas streams from
the digester system, multiple-effect evaporator system, brown stock washer
system, black liquor oxidation system, and condensate stripper system. A
temperature monitor is not required if the incineration device is a lime
kiln or recovery furnace subject to the NSPS.
3-5
-------
4. STATUS OF CONTROL TECHNOLOGY
The methods of emission control being employed on process facilities
subject to the NSPS are presented in this chapter. A comparison is also
presented between these control systems and the control systems which were
used as a basis for the NSPS. In general, the control equipment being used
by the industry is that upon which the data base for the NSPS was developed.
To develop this information, the EPA contacted kraft pulp mills, State
agencies, and EPA Regional offices.
4.1 PARTICULATE CONTROL
4.1.1 Recovery Furnace
All recovery furnaces subject to the NSPS employ electrostatic
precipitators (ESP) as their particulate control devices. The ESP has been
extensively used in the kraft pulp industry. Initially, ESP's were primarily
installed for recovery of the soda ash suspended in the recovery furnace
flue gas for economic purposes.
Both direct contact and non-contact type furnaces have been installed
under the NSPS. The physical arrangements and design details of ESP's
installed to control emissions from recovery furnaces subject to the NSPS
are presented in Table 4-1. ESP's installed to control particulates from
non-contact furnaces tend to be larger than ESP's on direct contact furnaces
because of the higher inlet loadings and finer and lower density dust
associated with non-contact furnaces. Although the purpose of the direct
contact evaporator is to concentrate black liquor, it may also remove up to
50 percent of the particulate matter from the gas stream. The collector
plate area to air flow ratios or SCA's (square feet of plate area per 1,000
cubic feet per minute of gas volume) reported for ESP's on non-contact furnaces
range from 438 to 587 and average 500. The only reported SCA for a direct
contact system is 453. These design SCA values are higher than the design
SCA values of the ESP's tested by the EPA during the NSPS development
program. The design SCA values for the EPA-tested ESP's ranged from 346 to
441 and averaged 374 for direct contact systems and 412 for non-contact systems.
4-1
-------
ro
Table 4-1. Design of Electrostatic Precipitators
Subject to the NSPS1-11
Physical Arrangement
ESP
A
B
C
D
F
G
H
I
P
Q
S
Furnace
Design
Air
Type Bottom Wire
Controlled Type Type
NC
DC
NC
DC
NC
NC
NC
NC
NC
NC
NC
Wet
Wet
Wet
Wet
Wet
Dry
Dry
Dry
Dry
Wet
Wet
Weighted
Weighted
Weighted
Rigid
—
--
Rigid
Weighted
Weighted
Weighted
Rigid
Shell No. of
Type Chambers
Heated
Heated
Heated
Heated
--
--
Insulated
Heated
Heated
Heated
Heated
2
3
2
2
2
2
2
2
2
2
2
No. of
TR Sets
12
9
14
6
10
6
8
8
12
6
8
SCA
478
453
587
--
—
461
438
510
542
506
472
Velocity
(ft/s)
2
3
3
3
2
3
2
2
3
.5
.2
.3
--
--
.2
.84
.3
.7
.0
.8
Parameters
Effi- Power Current
ciency Input Input
1%) kVA/lOOOcfm mA/lOOOcfm
99
99
99
99
-
99
99
99
99
99
99
.8 1.9 22.2
.8 2.5
.8 3.5 34.3
.7 2.2
-
.7
.7 2.8 33.4
.8 2.4 31.2
.8
.7
.7
NC = Non-contact type
DC = Direct contact type
-------
In 1979, the Technical Association of the Pulp and Paper Industry
(TAPPI) Air Committee conducted surveys on electrostatic precipitators used
to control emissions from direct contact furnaces and non-contact furnaces.8^9-10
The surveys reviewed new facilities which started up between 1973 and 1977
for direct contact furnaces and 1974 and 1978 for non-contact furnaces, and
presented the design and performance results of a second generation of sale
cake electrostatic precipitators.
The 1979 surveys indicate:12-ltt
The industry has made significant progress in the control of particulate
matter from recovery furnaces.
The heated steel shell design ESP has displaced the tile shell
design.
An ESP for application to non-contact furnaces require about 40 percent
more SCA than ESP's on similar direct contact furnaces.
Maintenance costs for ESP's applied to direct contact furnaces are
about $38,000/year compared to costs of about $76,000/year for ESP's
applicable to non-contact furnaces.
Problems, as reported in the 1974 survey, of salt cake buildup on
rappers, inlet vane pluggage, damper problems, unit undersizing, and "snowing"
(the emission of large white particles resembling snowflakes) have been
solved.
The newer ESP installations on non-contact furnaces have higher reliability
and higher particulate removal efficiencies.
Wet-bottom ESP's have a significant reduction in yearly downtime
over dry-bottom ESP's.
During this review, EPA personnel visited five kraft mills, contacted
four EPA regional offices, several State agencies, and sent information
requests to eight mills. The result of these contacts generally confirmed
the 1979 study results, except that the ESP's on non-contact furnaces subject
to the NSPS have about 15 percent, instead of 40 percent, more SCA than
ESP's on direct contact furnaces subject to the NSPS.
During the NSPS development program, the industry voiced concerns
about the potential for gradual ESP deterioration and potential maintenance
problems.
4-3
-------
Concerning the problem of gradual deterioration of precipitator
performance, the St. Regis' Tacoma, Washington, mill, which was tested as
part of the NSPS development program, was contacted to obtain information
on maintenance costs and ESP performance.15 The recovery furnace and ESP
started operation in 1973. The result of the EPA test conducted in 1974
was 0.003 gr/dscf (0.007 g/dscm).1^ jhe result of the latest quarterly
test (September 1982) reported to the State agency was 0.034 gr/dscf
(0.081 g/dscm), corrected to 8 percent oxygen.1'' Table 4-2 presents the
results of monthly and quarterly testing done by the company for the ESP
since startup in 1973 through September 1982. The data reported for the
period from startup to December 1980 were collected using the Washington
State sampling train, which includes both the front and back half fractions.
The NSPS is, however, based only on the front half fraction which, based
on the EPA test at St. Regis, is less than half of the total sampling
train result. Starting in March 1981, only the front half fraction is
reported. The data shown in Table 4-2 indicate that, with maintenance,
the ESP is still capable of achieving the NSPS level after 9 years of
operation. The main maintenance problems for the St. Regis' dry-bottom
ESP have been with the rappers and the screw conveyors. Total maintenance
for the unit has averaged 913 manhours over a five year period (1977-1981).
This mill has also reported no corrosion problems with this 9-year-old
ESP. Information obtained from mills with furnaces subject to the NSPS
indicate no major operating or maintenance problems. Maintenance is
usually performed on an "as needed" basis since the mills have the ability
to isolate either chamber of the ESP without shutting down the furnace
operation. Physical inspection of the ESP's are generally conducted once
or twice per year when the mill is shutdown for general maintenance.
These units subject to the NSPS have generally been in operation less
than 3 years.
The only other control device presently being installed is an alkaline
scrubbing system.18 These units, however, have not been installed on NSPS
recovery furnaces, but on older furnaces downstream from precipitators.
The installation of these units has been confined to upgrading existing
systems. Since the system is also designed to control TRS emissions, it
will be more fully discussed in Section 4.2.1.
4-4
-------
Table 4-2. PARTICIPATE EMISSIONS DATA FOR THE NO. 4 RECOVERY FURNACE
AT THE TACOMA, WASHINGTON, MILL, 1973 THROUGH 198115-17
Date
Aug. 1973
Sept. 1973
Oct. 1973
Nov. 1973
Dec. 1973
Jan. 1974
Feb. 1974
Mar. 1974
Apr. 1974
May 1974
June 1974
July 1974
Aug. 1974
Sept. 1974
Oct. 1974
Nov. 1974
Dec. 1974
Jan. 1975
Test
Result
(gr/dscf)
0.0192
0.0011
0.0240
0.0410
0.0101
0.0150
0.0100
0.0130
0.0140
0.0120
0.0120
0.0061
0.0133
0.0069
0.0171
0.0096
0.0143
0.0218
Feb. 1975 0.0309
Mar. 1975 0.0499
Apr. 1975 0.0213
May 1975 0.0411
June 1975 0.0349
ON STRIKE
Oct. 1975 No Test
Nov. 1975 No Test
1
Dec. 1975 0.0269
Date
Jan. 1976
Feb. 1976
Mar. 1976
Apr. 1976
May 1976
June 1976
July 1976
Aug. 1976
Sept, 1976
Oct. 1976
Nov. 1976
Dec. 1976
Jan. 1977
Feb. 1977
Mar. 1977
Apr. 1977
May 1977
June 1977
July 1977
Aug. 1977
Sept. 1977
Oct. 1977
Nov. 1977
Dec. 1977
Jan. 1978
Feb. 1978
Test
Result
(gr/dscf)
0.0445
0.0167
0.0112
0.0365
0.0114
0.0069
0.0288
0.0139
0.0061
CURTAILMENT
OF MILL
0.0217
0.0106
0.0697
0.0117
0.0182
0.0171
0.0112
0.0301
0.0104
0.0257
0.0381
0.0254
0.0216
0.0235
0.0282
0.0358
Date
WENT TO TEST
PER QUARTER
* GR/SCF CON
June 1978
Sept. 1978
Dec. 1978
Mar. 1979
i June 1979
Sept. 1979
Dec. 1979
Mar. 1980
June 1980
Sept. 1980
Dec. 1980
Mar. 1981
June 1981
Sept. 1981
Dec. 1981
Mar. 1982
June 1982
Sept. 1982
Test
Result
(gr/dscf)
ING ONCE
VERTED TO 7%
0.0341
0.0186
0.0315
0.0413
0,0333
0.0675
0.0356
0.0345
0.0274
0.0476
0.0804
0.0306
0.0321
0.0836
0.0298
0.026
0.0523
0.0342
4-5
-------
4.1.2 Smelt Dissolving Tanks
The exhaust gases from smelt dissolving tanks subject to the NSPS are
vented through wetted fan type scrubbers or low pressure drop venturi
scrubbers (4-10 inches of water) for particulate removal. Weak wash is
used as both the dissolving medium and scrubbing medium at those facilities
for which information is available. These control systems are comparable
to those tested by the EPA during the NSPS development.
4.1.3 Lime Kiln
All lime kilns subject to the NSPS, except one, are controlled with
venturi scrubbers, with pressure drops ranging from 4,230 to 8,460 pascals
(17 to 34 inches of water). The lower pressure drops are associated wtth
venturi scrubbers controlling emissions from fluidized bed calciners.
Pressure drops of venturi scrubbers installed on rotary lime kilns range
from 6,220-8,460 pascals (25-34 inches of water).
One mill has installed an electrostatic precipitator to control
emissions from the rotary lime kiln subject to the NSPS. This ESP has a
design SCA of 462 and a design outlet loading of 0.05 g/dscm (0.02 grains/dscf).
The ESP was installed to achieve additional reduction over the NSPS because
of the location of the mill.
The use of ESP's on lime kilns was investigated during the original
NSPS development, and ESP's were considered capable of achieving lower
particulate emissions than venturi scrubbers with pressure drops of about
7,470 pascals (30 inches of water). However, it was not considered feasible
to base the NSPS on ESP's, since the industry commented that any noncondensable
gases not combusted in the kiln might cause explosions in the ESP. Furthermore,
a venturi scrubber was considered capable of controlling both particulate
and TRS emissions.
Research has been done on the use of a fabric filter (baghouse) as an
alternative control technique for kraft lime kilns.^ A six-month pilot
operation, conducted at a pulp mill in Canada, indicated that fabric filters
are able to remove particulate at an efficiency of 99.9 percent. The
conclusions of this research were that fabric filters can be used, the
energy requirements are less than scrubbers, and the economics can be
superior to a wet scrubber. However, full-scale systems using baghouses
have not been installed so long-term performance data (emissions, maintenance,
etc.) are not available.
4-6
-------
4.2 TRS CONTROL
4.2.1 Recovery Furnace System
TRS emissions from a recovery furnace system can originate in the
recovery furnace itself, or in the direct contact evaporator, if this type
of evaporator is used. Most recent recovery furnace systems which have
been installed are the non-contact design. In these furnaces, the furnace
flue gases never directly contact the black liquor and TRS cannot be formed
in the evaporator. The TRS emissions from the recovery furnace itself are
controlled by maintaining proper process conditions.
TRS emissions that normally result from a direct-contact evaporator
are controlled by either black liquor oxidation or conversion to a non-contact
system. Black liquor oxidation inhibits the reactions between the combustion
gases and black liquor that normally generate hydrogen sulfide. This is
accomplished by oxidizing the Na2S to ^$203 in the black liquor before it
enters the direct contact evaporator. Air is the normal oxidizing agent, but
molecular oxygen is used at one mill. In the non-contact system, direct
contact between furnace gases and black liquor is eliminated, and hydrogen
sulfide is prevented from forming.
The direct-contact and non-contact furnaces subject to the NSPS are
using the same control technology as those tested by the EPA for development
of the MSPS.
One recovery furnace subject to the NSPS is a cross recovery furnace.
It was determined during the NSPS development that cross recovery furnaces
cannot be operated to achieve the same TRS levels as straight kraft furnaces
because of the higher sulfur content and lower heat content of the cross
recovery liquor and the restriction on excess oxygen to oxidize the relative
large quantities of sulfur compounds given off. A restriction of oxygen is
necessary to prevent development of a sticky dust which can plug up the ESP
and render furnace operation impossible.20
An add-on control technique for controlling TRS emissions from recovery
furnaces is an alkaline adsorption system with carbon activated oxidation
of the scrubbing solution.1** This control technique has not been installed
on recovery furnaces subject to the NSPS, but has been installed on older
existing furnaces. Reductions in particulate and SO? emissions are also
reported. This technique has been applied to control TRS emissions on
4-7
-------
existing furnaces or cross recovery furnaces which do not have the combustion
control capability for low TRS emissions. This technique is being promoted
as an alternative to BLO, replacing an existing furnace, modifying an
existing furnace to non-contact, or reducing the load on furnaces that are
not capable of achieving the necessary TRS regulations.
4.2.2 Digester and Multiple-Effect Evaporator Systems
Control of TRS emissions from the digester and multiple-effect evaporators
are considered together, since their emissions are normally combined for
treatment. The noncondensable gases are incinerated in either a lime kiln,
power boiler, or separate incinerator. Combustion of noncondensable gases
in a lime kiln or separate incinerator provides nearly complete destruction
of TRS compounds. Incineration of the gases was the basis for the NSPS.
The lime kiln has been the usual source for incinerating these gases.
However, at least two of the NSPS lime kilns are equipped with tube coolers
for heat recovery, and the mills are reluctant to pass the noncondensable
gases through these coolers because of the possibility of an explosion
resulting in damage to the coolers. The portion of the combustion air
going directly to the kiln is not sufficient to accommodate the noncondensable
gas stream from an equivalent size pulp mill. Therefore, these mills are
incinerating the gases in either a power boiler or separate incinerator.
4.2.3 Brown Stock Washing System
Incineration of the vent gases is the only control technique employed
on the brown stock washing systems subject to the NSPS. Incineration is
the basis for the NSPS. Diffusion washers are generally being installed.
Diffusion washing usually takes place in a closed reactor, and ideally
there is no air involved. Therefore, the vent gases released are very
small, when compared with the vacuum drum washers.21 In fact, two diffusion
brown stock washing systems subject to the NSPS were not required by the
permitting agencies to be controlled by incineration because the washers
are totally enclosed and are only open to the atmosphere to prevent over
pressure or vacuum conditions during filling and emptying procedures.22
Since this is a continuous process, the filling and emptying occurs on an
infrequent basis. The vent gases from other diffusion washers subject to
the NSPS are being incinerated in either a lime kiln, recovery furnace or
separate incinerator.
4-8
-------
Two vacuum drum washing systems are subject to the NSPS. The vent
gases from one system are used as combustion air in a power boiler. The
other vacuum drum washer is not controlled by incineration, but was
granted an exemption from the standard because the mill did not have
available an incineration device which could safely handle the gases
economically. This vacuum drum washer was installed as a new fourth
stage washer on an existing washing system.
4.2.4 Black Liquor Oxidation Systems
There are only two black liquor oxidation (BLO) systems subject to the
NSPS. One BLO system uses molecular oxygen to oxidize the black liquor.
There are no vent gases from this closed system and, therefore, no TRS
emissions. The vent gases from the other BLO system go through a condenser
and preheater, and are used as combustion air in a power boiler. The
combustion of these gases reportedly does not result in any significant
increase in fuel requirements or boiler operating problems.1 Incineration
of the vent gases and the use of molecular oxygen were the control techniques
investigated during the NSPS development.
4.2.5 Smelt Dissolving Tank
There are no special TRS control devices for smelt dissolving tanks.
TRS emissions are governed by process conditions; that is, the presence of
reduced sulfur compounds either in the smelt or the water. The principal
control option available is the choice of water in the smelt dissolving
tank or the particulate control device. This control technique was the
basis for the NSPS. All 11 of the smelt tanks subject to the NSPS, for
which information is available, use weak wash as the dissolving and
scrubbing medium.
4.2.6 Lime Kiln
The NSPS for TRS emissions from lime kilns is being achieved by
utilizing process controls and good lime mud washing. At one lime kiln
installation caustic is being added to the scrubber water. The use of
caustic scrubbing in combination with process control and good mud washing
was the basis for the NSPS. Several mills have designed the particulate
scrubber with the capability to add caustic, if necessary. These mills
have achieved compliance, based on a source test, without caustic addition,
4-9
-------
but feel that once the continuous monitoring requirements are promulgated,
caustic addition might be necessary to continuously achieve the NSPS
during normal variations in the lime kiln process.
4.2.7 Condensate Stripping System
Only three condensate stripping systems are known to be subject to the
NSPS. In both cases, the vent gases are incinerated. One of these incin-
erates the gases in a separate incinerator, while the other two incinerate
the gases in a lime kiln. Incineration is the control technology upon
which the NSPS was based.
4-10
-------
4.3 REFERENCES
May 20r11982P°rt " Champ10n Internat1onal Corporation, Courtland, Alabama,
2. Letter from C. F. Johnson, Continental Forest Industries to
Don R. Goodwin, U.S. EPA, dated May 21, 1982.
3. Trip Report - Union Camp Corporation, Montgomery, Alabama, May 19, 1982.
n\ PP^ fTi0™1? $. Dillard, Jr., Nekoosa Paper Inc., to Don R. Goodwin,
U.S. EPA, dated June 1, 1982.
5. Letter from Grey Forte, Tennessee Department of Public Health to
James Eddinger, U.S. EPA, dated March 16, 1982.
t0 JmeS Eddin9er, U.S. EPA,
M'S PPAtt6H ITA ?' L!nd?e£' International Paper Company, to Jim Eddinger,
U.S. EPA, dated October 4, 1982.
?'nt TKiP ?fp°!'no; Buckeye Cellulose Corporation, Oglethorpe, Georgia,
oeptemDer ib, I9o^..
!L. T?'eP5?ne.c?"^rsation between Ike Core of Continental Forest Industries,
Hopewell, Virginia, to James Eddinger, U.S. EPA, on August 11, 1982.
ion
I?' F1na1 Survey Results for Direct Contact Evaporator Recovery Boil
f^r.^rc'lgg'v^f-No; 3' He"derSOn- J' b' S'rri"e C°^"y-
13- F"jnal Survey Results for Noncontact Recovery Boiler Electrostatic
Precipitators; Part 1, Precipitator Design Features, J. S. Henderson
J. E. Sirnne Company, TAPPI, December 1980, Vol. 63, No. 12.
J4- .Fl!na1 Survey Results for Noncontact Recovery Boiler Electrostatic
Precipitators; Part 2, Precipitator Performance Results. J. S. HpndPrsnn
J. E. Sirrine Company, TAPPI, January 1981, Vol. 64, No. 1.
15. Trip Report - St. Regis Paper Company, Tacoma, Washington, May 25, 1982
EPA ReortNeS-!Prt " "' ^ **' ^^"^ T3C°ma' """Ingt™.
4-n
-------
17. Information received from C. F. Kleeberg, EPA Region X, Olympia,
Washington, April 29, 1983.
18. Letter and Enclosure from Ronald A. Jolicoeur, Teller Environmental
Systems, to Don R. Goodwin, U.S. EPA, dated September 13, 1982.
19. Fabric Filters for Kraft Mill Lime Kilns, D. R. Lewis, Centurion
Engineering Ltd., Pulp & Paper Canada, vol. SO, No. 12, December 1979.
20. A Report on the Study of TRS Emissions from a NSSC-Kraft Recovery Boiler,
Container Corporation of America, March 9, 1977.
21. Environmental Pollution Control-Pulp and Paper Industry, Part I Air,
U.S. Environmental Protection Agency, EPA-bZ5/7-/6-001, October 1976.
22. Letter from Alan M. Lindsey, International Paper Company, to
James Eddinger, U.S. EPA, dated December 17, 1982.
4-12
-------
5. COMPLIANCE TEST RESULTS
EPA regional offices, State agencies, and kraft pulp mills were
contacted to obtain compliance test information for new, modified, or
reconstructed facilities. Test data for recovery furnaces, lime kilns, and
smelt dissolving tanks were specifically requested.
The results of the survey show that there are 14 new recovery furnaces,
20 new lime kilns, and 16 new smelt dissolving tanks in operation which are
subject to the NSPS. In addition, two existing recovery furnaces are subject
to the NSPS since both were modified from a direct contact type to an
indirect contact type furnace after the NSPS became effective. One of the
new recovery furnaces is a cross recovery furnace, and two of the new lime
kilns are fluidized bed calciners. Several new digesters, multiple-effect
evaporators, brown stock washers, condensate strippers, and black liquor
oxidation systems have been constructed since the NSPS became effective,
but since the control technique employed is incineration in another process
source, there is no compliance testing required and, therefore, no data are
available. Compliance testing for these sources is not required if the
gases are combusted in a lime kiln or recovery furnace subject to the NSPS
or combusted in another device with a minimum combustion temperature of
1,200°F and a residence time of at least 0.5 second.
5.1 ANALYSIS OF NSPS COMPLIANCE TEST RESULTS
5.1.1 Recovery Furnace System
The results of compliance tests obtained from new and modified recovery
furnaces are summarized in Table 5-1. Compliance test results for particu-
late emissions obtained on 13 recovery furnaces indicate compliance with
the NSPS, with emissions ranging from 0.007 to 0.080 grams per dry standard
cubic meter (g/dscm) [0.003 to 0.035 grains per dry standard cubic foot (gr/
dscf)]. Compliance test results for TRS emissions obtained on 11 recovery
furnaces range from 0.1 to 30 parts per million (ppm). The data indicate
that all but two are in compliance with the NSPS. Two additional installations
which are in the startup phase have not yet complied with the TRS standard.
No Method 9 opacity data were supplied with the compliance test reports.
5-1
-------
Table 5-1.
Compliance Test Results of
Recovery Furnaces Subject to NSPS 1-
Ml 1 1
A
B
C
D
r
1_
F
G
H
I
P3
Q
T
U
V
Typel
NC
DC
NC
DC
NO?
NC
NC
NC
NC
NC
NC
NC
NC2
NC
Parti cul
Control
ESP (SCA=478)
ESP (SCA=453)
ESP (SCA=587)
ESP (NA)
ESP
ESP (NA)
ESP (SCA=461)
ESP (SCA=438)
ESP (SCA=510)
ESP (SCA=542)
ESP (SCA=506)
ESP
ESP (NA)
ESP
1
| NSPS
ates
Level
gr/dscf
0.013
0.006
0.007
0.006
0,003
0.035
0.006
0.027
0.012
0.013
0.033
0.018
0.011
= 0.044
TRS
Control
Process
BLO
Process
BLO
--
Process
Process
Process
Process
Process
Process
Process
—
Process
NSPS =
Level
(ppm)
4.3
0.1
4.4
2.8
--
2.2
2.3
1.8
4.0
3.8
10
--
30
5 ppm
1 NC = Non-Contact, indirect-contact type
DC = Direct-contact type
- Furnace was modified from direct-contact to non-contact
3 Cross recovery furnace
5-2
-------
The design surface area-to-volume (SCA) ratios for the ESP's installed
to control emissions from the new recovery furnace systems ranged from 438
to 587 square feet of collecting plate area/1,000 actual cubic feet per
minute of exhaust gas and averaged 497. The design SCA's of the ESP's
tested by the EPA during the NSPS development ranged from 346 to 441 and
averaged 393.
'One mill reportedly had problems in initially complying with the TRS
standard.15 At this mill, it was found that the original design of the wet
bottom ESP using unoxidized black liquor resulted in an increase in TRS
concentrations as the furnace gas passed through the ESP. The TRS increase
was traced to the air flow in the ESP which promoted contact between the
flue gas and the unoxidized black liquor. Modifications were made to the
duct work and internal baffling to reduce this contact by redirecting the
gas flow. This recovery furnace is now in compliance. The two recovery
furnaces not in compliance plus two additional new installations with
wet-bottom ESP's have also reported problems in achieving the NSPS due to
an increase in TRS emissions through the ESP.16.*7 Modifications similar to
those discussed above have been made to these ESP's, and although the TRS
emissions have been reduced, they have not met the NSPS. Additional
modifications to achieve compliance are being investigated by the companies
and vendors.
Since the EPA has not promulgated specifications for continuous
monitoring systems, long-term TRS data showing whether these facilities are
continuously achieving the NSPS are not available.
5.1.2 Smelt Dissolving Tanks
The results of compliance tests obtained on 12 new smelt dissolving
tanks are summarized in Table 5-2. Either a venturi or wetted fan type
scrubber is employed for particulate control, with weak wash as the scrubbing
and dissolving medium in all cases. The particulate emissions ranged from
0.025 to 0.095 gram/kilogram of black liquor solids (g/kg BLS) [0.05 to
0.19 pound/ton of black liquor solids (T BLS)] and the TRS emissions ranged
from 0.001 to 0.016 g/kg BLS (0.002 to 0.032 pound/T BLS). The compliance
test data indicate that all the smelt dissolving tanks are in compliance
with the NSPS except for one mill which has not been able to comply with
the TRS standard.
5-3
-------
Table 5-2.
Compliance Test Results of
Smelt Dissolving
Tanks Subject to NSPS 1-12,26
Mill
A
D
c
D
C
F
G
H
I
P
Q
T
Particulates
Control
Wetted Fan
Wetted Fan
Scrubber
Venturi
Venturi
Wetted Fan
Venturi
Venturi
Venturi
Wetted Fan
Venturi
Level
(#/TBLS)
0.18
0.09
0.05
0.19
0.1
0.16
0.06
0.133
0.115
0.137
0.1
0.13
NSPS =0.2
TRS
Level
Water Used (#/TBLS)
Weak Wash 0.014
Weak Wash 0.004
Weak Wash 0.002
Weak Wash 0.007
Weak Wash 0.004
Weak Wash
Weak Wash 0.032
Weak Wash 0.008
Weak Wash
Weak Wash 0.013
Weak Wash 0.013
0.0168
5-4
-------
The smelt dissolving tank which is not in compliance with the TRS
standard is utilizing the same control technology as the other installations.
This mill has attempted to reduce the TRS emissions using various scrubbing
liquids, but has not found a solution to continuously achieve the NSPS.13
A continuous monitor was installed to help analyze the problem. Results of
the monitoring indicate that the TRS standard can be achieved, but not on a
continuous basis.19 The monitoring data ranged from 2.8 to 14.7 ppm
(0.0047 to 0.0247 #/T BLS) and averaged 5.5 ppm (0.0092 #/T BLS), on a
12-hour average basis. Of the 27 12-hour averages reported, 92.6 percent
were below the NSPS level.
Another mill also reportedly had problems in complying with the NSPS.
At this mill, the stack height resulted in too much draft, causing a portion
of the exhaust to bypass the scrubber. A damper was installed to bleed in
cool air to reduce the exhaust temperature and the draft. This corrected
the problem, and the smelt dissolving tank is now in compliance.4
5.1.3 Lime Kiln
The results of compliance tests obtained on lime kilns subject to the
NSPS are summarized in Table 5-3. The particulate emissions ranged from
0.027 to 0.14 g/dscm (0.012 to 0.063 gr/dscf) for gas-fired kilns, and
0.079 to 0.169 g/dscm (0.034 to 0.074 gr/dscf) for oil-fired kilns. The TRS
emissions ranged up to 7.3 ppm. The test data indicate that all the lime
kilns are in compliance with the NSPS.
All but one of the lime kilns are controlled by a venturi scrubber,
with design pressure drops of 4,233 to 8,217 pascals (17 to 33 inches of
water). The pressure drops of the scrubbers tested by the EPA during the
NSPS development ranged from 4,233 to 7,968 pascals (17 to 32 inches of
water).
Two of the lime kilns subject to the NSPS are fluidized bed calciners.
The NSPS covers all types of lime kilns (i.e., rotary kilns, fluidized bed
calciners). However, during the NSPS development only rotary kilns were
tested by the EPA because fluidized bed calciners were in limited used in
the kraft pulping industry. Fluidized bed calciners were in operation at
only four mills and their production rates were under 150 tons per day of
quicklime. The two fluidized bed calciners subject to the NSPS have
production rates of about 236 megagrams (260 tons) per day. As shown in
Table 5-3, both fluidized bed calciners are in compliance with the NSPS.
5-5
-------
Table 5-3.
Compliance Test Results of
Mill Type*
Gas-Fired Kilns
A Rotary
8 Rotary
C Rotary
J Fluidi zed Bed
M Rotary
N
0
Q Rotary
R Rotary
T Rotary
Oil -Fired Kilns
D Rotary
E
I Fluidized Bed
K Rotary
L Rotary
Lime Kilns Subj
Particul
Control
Venturi (AP-32")
Yenturi (AP-33")
Venturi (AP-28")
Yenturi
Venturi
Venturi
Venturi
Yenturi (AP-28")
Yenturi
Venturi (AP-28")
Venturi
Venturi (AP-20")
ESP (SCA-461)
Venturi (AP-30")
NSPS =
ect to NSPS 1-
ates
Level
gr/dscf
0.055
0.046
0.039
0.050
0.012
0.063
0.061
0.043
0.037
0.013
0.031
0.034
0.074
0.035
0.081
= 0.067 (Gas)
0.13 (Oil)
4,b,«,ll,2U-Zb
TRS
Level
Control (ppm)
Process 4.5
Process 0.5
Process 5.0
Process 2.0
7.3
1.75
4.5
Process 5.3
5.9
Process 2.7
Process 6.8
Caustic 3.8
Process 6.7
4.6
8.0
5-6
-------
An ESP is used to control participate emissions from one lime kiln.
This lime kiln has not been compliance tested for TRS emissions. Monitoring
results indicate that the TRS standard can be achieved, but presently not
on a continuous basis. The kiln operates with a low cold-end temperature
for energy conservation; and caustic scrubbing is not an option, because
the ESP is used for particulate control. The company is trying to achieve
the NSPS by good lime mud washing. If the NSPS is not achieved with mud
washing, the company is considering using lime mud oxidation to achieve the
NSPS.25
Only one kiln installation is using caustic addition for TRS control.
Several are designed to add caustic, if necessary. Process controls and
good mud washing are the techniques used at the other lime kiln installations
for TRS control.
The EPA has not promulgated specifications for continuous monitoring
systems, so long-term TRS data showing the performance of these TRS control
systems (both caustic and noncaustic systems) are not available.
5-7
-------
5.2 REFERENCES
1. Memo from Jam's Forehand, EPA Region VI, to Jim Eddinger, EPA, dated
June 29, 1982.
2. Letter from C. F. Johnson, Continental Forest Industries, to Don R.
Goodwin, U.S. EPA, dated May 21, 1982.
3. Letter and enclosures from Philip H. Speir, Hammermill Paper Company,
to Jim Eddinger, U.S. EPA, dated August 19, 1982.
4. Trip Report - Union Camp Corporation, Montgomery, Alabama, May 19, 1982.
5. Letter from Grey Forte, Tennessee Department of Public Health, to
James Eddinger, U.S. EPA, dated March 16, 1982.
6. Trip Report - Champion International Corporation, Courtland, Alabama,
May 20, 1982.
7. Memo from Frank W. Lilley, EPA Region I, to Stanley T. Cuffe, U.S. EPA,
dated March 19, 1982.
3. Letter from Les Oakes, Georgia Department of Natural Resources, to
James A. Eddinger, U.S. EPA, dated April 19, 1982.
9. Telephone Conversation between Gary Gross, EPA Region III, and
James Eddinger, U.S. EPA, on June 22, 1982.
10. Letter from Eric J. Schmidt, Container Corporation of America, to
Don Goodwin, U.S. EPA, dated July 7, 1982.
11. Trip Report - Washington State Department of Ecology, Olympia, Washington,
May 24, 1982.
12. Memo and enclosure from Jam's Forehand, EPA Region VI, to Jim Eddinger,
U.S. EPA, dated December 22, 1982.
13. Letter and enclosures from A. M. Lindsey, International Paper Company,
to James Eddinger, U.S. EPA, dated December 20, 1982.
14. Letter and enclosure from John E. Pinkerton, NCASI, to James A. Eddinger,
U.S. EPA, dated April 8, 1983.
15. Telephone Conversation between Phillip Speir, Hammermill Paper Company,
and James Eddinger, U.S. EPA, on August 10, 1982.
16. Telephone Conversation between Keith Bentley, Georgia-Pacific Corporation
and James Eddinger, U.S. EPA, on September 28, 1982.
17. Meeting Report - Industry Representatives and EPA Personnel, Durham,
North Carolina, April 14, 1983.
5-8
-------
18. Letter and enclosures from David £. Peakes, Boise Cascade Corporation
to James Eddinger, U.S. EPA, dated December 9, 1982.
19. Letter from David E. Peakes, Boise Cascade Corporation, to Janes Eddinqer
U.S. EPA, dated November 8, 1982.
20. Total Reduced Sulfur Emission Compliance Test Performed at Willamette
Industries, inc., Hawsvnie, Kentucky, Harmon Lngineenng and testing
Auburn, Alabama. September 1981.
21. Compliance Testing of Particulate Emission from the No. 2 Lime Kiln
at the Willamette Industries, Inc. Bleached Pulp Plant, HawsvineTTentiJck.y,
Kenvirons, Inc., Frankfort, Kentucky, September 1981.
22. Letter from J. H. Millican, The Buckeye Cellulose Corporation to
Don Goodwin, U.S. EPA, dated June 7, 1982.
23. Letter from Ken Berryman, County of Shasta Department of Health Services
to James A. Eddinger, U.S. EPA, dated June 11, 1982.
24. Memo from Jim Herlihy, EPA Region X, to Jim Eddinger, U.S.EPA dated
March 24, 1982.
25. Trip Report - Longview Fibre Company, Longview, Washington, May 26, 1982.
26. H. A. Lewis, Jr., Arkansas Department of Pollution Control and
Ecology, to J. Eddinger, EPA:ISB, June 1, 1983, Compliance Test Reports.
5-9
-------
6.0 COST ANALYSIS
The purpose of this chapter is to present updated capital and annualized
costs for the control systems used to achieve the NSPS. The costs are
presented for each of the affected facilities in a 907 megagrams (1,000
tons) per day kraft pulp mill. The cost-effectiveness of the control
systems is also presented. All costs are in terms of February 1982 dollars.
The costs presented for achieving the particulate standards are the incremental
control costs over the economic recovery level. Actual industry costs
are presented as a comparison, where available.
The capital cost of a control system includes all the cost items
necessary to design, purchase, and install the particular device or system.
The capital cost includes the purchased costs of the major control device
and auxiliaries such as pumps, fans, and instrumentation; the equipment
installation cost including foundations, piping, electrical wiring, and
erection; and the cost of engineering, construction overhead, and contingencies,
The annual ized cost of a control system is a measure of what it costs
the company to own and operate the system. The annual ized costs include
direct operating costs such as labor, utilities, and maintenance; and
capital related charges such as depreciation, interest, administrative
overhead, property taxes, and insurance.
6.1 UPDATE COST FOR THE AFFECTED FACILITIES
The NSPS for kraft pulp mills cover particulate and TRS emissions.
The updated costs, including capital costs, annualized costs, and credits
for recovered particulate, and the cost effectiveness of achieving the NSPS
for each system are shown in Table 6-1. These costs were updated to
February 1982 using the Chemical Engineering plant cost index. The
original costs were presented in the NSPS support document and were
calculated based on the fourth quarter of 1975.1 The costs presented
for controlling particulate and TRS emissions from the recovery furnace
are revised costs,2 based upon information supplied by the companies on
6-1
-------
Table 6-1. Capital Costs, Annualized Costs, and Cost Effectiveness Of
Achieving the NSPS for 907 Megagrams Per Day Kraft Pulp Mill
en
tV)
Source
Participate sources
Recovery Furnace (DC)
Recovery Furnace (NC)
Smel t Tank
Lime Kiln
TRS Sources
Dig. & Evap.
Washers - Vacuum Drum
- Diffusion
Recovery Furnace (DC)
Recovery Furnace (NC)
B.L. Oxid. System
Lime Kiln
Cond. Stripper
Smelt Tank
Control
Device
ESP
ESP
Scrubber
Venturi
Scrubber
Inc.
Inc.
Inc.
Process
& BLO
Process
Inc.
Process, Mud
Washing &
Caustic
Inc.
Water
Capital
Costs
1,512,000
1,322,600
181,400
41,500
292,000
584 ,000
124,000
955,000
1,551,000
506 ,000
0
35,000
0
Annual
Cost
271,800
240,600
59,260
37,140
67 ,000
116,000
44,000
777,100
494,000
148,000
110,000
12,000
0
Emission
Reductions
Credits (T/Y)
(132,160) 1,180
( 93,070) 831
( 23,860) 213
( 8,460) 180
410
49
49
2,439
2,439
16
124
329
29
Cost
Effectiveness
($/T)
118
178
166
159
163
2,370
898
319
203
9,250
887
37
0
-------
actual recovery furnace system operations. Table 6-1 also presents the
emission reduction resulting from application of control techniques.
The particulate emission reductions are calculated based upon the difference
between the economic recovery levels and the NSPS levels. The TRS emission
reductions are calculated based upon the difference between the uncontrolled
levels reported in the NSPS support document and the NSPS levels.
6.1.1 Particulate Control
For each of the particulate control devices in Table 6-1, the costs
presented are the incremental control costs over the economic recovery
levels. Since the particulate is a valuable material, it is economical
to recover and recycle it to the process. The credits for recovered
particulate are calculated assuming that all the particulate from the
recovery furnace and smelt dissolving tank is sodium sulfate valued at
$112 per ton, and that the recovered particulate from the lime kiln is
valued at $47 per ton.3"7 In calculating the incremental control costs, the
control systems selected for economic recovery were a 95 percent efficient
ESP for the direct contact recovery furnace, a 97 percent efficient ESP
for the non-contact recovery furnace, a demister system (80 percent
efficient) for the smelt dissolving tank, and a venturi scrubber with a
3,740 pascals (15-inch) pressure drop for the lime kiln.8 The cost-
effectiveness values were calculated based on the incremental annual cost,
including credits, and the incremental emission reduction achieved by the
NSPS.
Actual capital costs were obtained from several companies for the
control devices associated with facilities subject to the NSPS. The capital
costs obtained were assumed to be from the year the unit started operation
unless actually stated by the company. The updated industry-reported data
ranged from $3 million to $6.4 million for ESP's controlling non-contact
furnaces.3»4»6>9 One mill reported a cost of $2.5 million for an ESP
controlling a direct-contact furnace.7 For comparison, the capital costs
estimated for achieving the NSPS are $5.1 million for a direct-contact
furnace and $5.9 million for a non-contact furnace.2
6-3
-------
Updated industry capital costs for scrubbers installed on four smelt
dissolving tanks ranged from 549,000 to $226,000. For comparison, the
capital cost estimated for a scrubber to achieve the MSPS is $229,000.2
Industry data were available for six lime kiln installations, with the
updated capital costs ranging from $75,000 to $660,000 dollars.3,4,6,7,10,11
The average of the six updated industry costs is $270,000. For comparison,
the capital cost estimated for a 7,470 pascals (30 inches of water) pressure
drop scrubber is $275,000.2
6.1.2 TR$ Control
For all TRS sources, no credits are estimated, because the TRS compounds
are not actually recovered in a control device. The cost-effectiveness of
the various TRS controls ranges from zero dollars for the smelt dissolving
tank to $9,250 for the black liquor oxidation system. The vacuum drum
type brown stock washing system is the only other affected facility with a
cost-effectiveness above $1,000 per ton.
Except for the recovery furnaces, lime kilns, and smelt dissolving
tanks, the TRS emissions from the affected facilities are controlled by
incineration in either a lime kiln, power boiler, or separate incinerator.
The cost of the incineration system consists of the necessary piping and
blowers to collect the gas streams, and delivery piping and controls to
inject the gases into the incineration point. The costs shown in Table 6-1
for brown stock washers are for a vacuum drum system and include hooding
costs. The costs shown in Table 6-1 for the black liquor oxidation system
include condensers.
The cost estimate shown in Table 6-1 for the direct contact furnace is
based on the cost of black liquor oxidation. The cost shown in Table 6-1
for the non-contact furnace is based on the incremental cost between the
non-contact furnace with a concentrator, and a direct contact furnace which
has a direct contact evaporator. Also included is a charge for the heat
loss of the non-contact furnace compared to the direct contact furnace.
The heat loss was calculated assuming that the flue gas is 40°F hotter than
the direct contact furnace flue gas. Included in the annual cost for the
6-4
-------
direct contact furnace is a charge for the heat loss resulting from the
oxidation of the black liquor. The heat loss was calculated assuming that
the heat content of the black liquor is reduced by 4 percent during black
liquor oxidation.
The control technique for reducing TRS emissions from the smelt dissolving
tank is to use water which is essentially free of dissolved TRS compounds
in both the smelt dissolving tank and the associated scrubber. No control '
costs are presented in Table 6-1, because this feature can be designed into
a new mill at essentially no cost.
The costs for the lime kiln are based on caustic addition to the
particulate scrubber and the cost of the energy needed to raise the cold-
end temperature of the kiln 55.5°C (100°F).
Presently, only two mills have BIO systems subject to the NSPS. One
mill uses oxygen as the oxidizing medium and has no vent gases. This mill
reported an operating cost of $508,000 per year for the oxygen. The other
mill incinerates the offgases in a power boiler. This 8LO system is the
only BLO system controlled by incineration in the industry. The annualized
costs of $148,000 for BLO systems shown in Table 6-1 is based on incineration
of the vent gases. The principal reason for the high cost-effectiveness
shown in Table 6-1 is the small amount (16 tons per year) of TRS controlled.
Of the eight TRS sources, the BLO system emits the least amount of uncontrolled
TRS from a typical pulp mill.
The high cost-effectiveness shown in Table 6-1 for the vacuum drum
type washer system is also the result of a low level of uncontrolled TRS
emissions. However, many mills are installing diffusion washers instead of
vacuum drum washers. The newer diffusion washers are a closed reactor, and
the vented gas volume is much less than that from a vacuum drum system.
The vent gases from the diffusion washer are usually incinerated with the
vent gases from the associated continuous digesters. The cost of controlling
a diffusion washing system would therefore be much less than the cost for
controlling a vacuum drum washing system because smaller ducts and no
hooding would be required. The updated capital cost reported by one mill
for controlling the vent gases from a diffusion washer with a capacity of
6-5
-------
743 megagrams (816 tons) per day is $101,000.5 This is about one-fifth the
capital cost estimated in Table 6-1 for controlling a vacuum drum washer
system. The annualized costs for a diffusion washing system are estimated
at about 344,000 which, assuming the TRS emission reduction for a vacuum
drum system, result in a cost-effectiveness of about $900 per ton of TRS
controlled.2
Actual industry data for controlling TRS emissions from digesters,
evaporators, BLO systems, and condensate stripping systems were not provided
and therefore cannot be compared to Table 6-1.
6-6
-------
6.2 REFERENCES
1. Standards Support and Environmental Impact Statement, Volume 1:
Proposed Standards of Performance for Kraft Pulp Mills, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, Publication No.
EPA-450/2-76-014a, September 1976.
2. Memo from James Eddinger, U.S. EPA, to Ken Durkee, U.S. EPA, dated
November 8, 1982.
3. Trip Report - Union Camp Corporation, Montgomery, Alabama, May 19, 1982.
4. Letter from Philp H. Speir, Hammermill Papers Company, to James Eddinger,
U.S. Environmental Protection Agency, dated July 22, 1982.
5. Trip Report - Longview Fibre Company, Longview, Washington, May 26, 1982.
6. Trip Report - Buckeye Cellulose Corporation, Oglethorpe, Georgia,
September 16, 1982.
7. Letter from David S. Dillard, Jr., Nekoosa Papers Inc., to Don R. Goodwin,
U.S. Environmental Protection Agency, dated June 1, 1982.
8. Memo from James Eddinger, U.S. EPA, to Ken Durkee, U.S. EPA, dated
9. Letter from Eric J. Schmidt, Container Corporation of America, to
Don Goodwin, U.S. Environmental Protection Agency, dated July 7, 1982.
10. Letter from Ken Berryman, Shasta County Department of Health Services,
to James A. Eddinger, U.S. EPA, dated June 11, 1982.
11. Memo from Jay M. Sinnott, EPA Region VIII, Montana Office, to
Jim Eddinger, U.S. EPA, dated June 22, 1982.
6-7
-------
7. ENFORCEMENT ASPECTS
EPA regional offices, State agencies, the NCASI, and companies subject
to the NSPS were contacted to determine any problems with either enforcing
the NSPS or complying with the NSPS.
Discussions with EPA offices and State agencies indicated that there
are no problems in enforcing the NSPS. The Tennessee State Agency did
express concerns about kraft mills not being required to continuously
monitor TRS emissions because the EPA has not yet promulgated the TRS
monitoring specifications.1 These specifications were proposed on July 20,
1981, (46 FR 3787) and are scheduled for promulgation in 1983. Once
promulgated, all recovery furnaces and lime kilns subject to the NSPS will
be required to continuously monitor TRS emissions.
The Alabama State Agency stated that more investigation is needed
concerning incineration of the exhaust gases from brown stock washers in
kraft recovery furnaces.2 The one mill which has received an exemption
from EPA for their washer system is located in Alabama. This issue was
investigated and the findings are discussed in Chapter 8.
The EPA Region I office suggested that the aging factor of ESP's be
included in the review.3 This comment was made because a mill in the
region has overdesigned their ESP in anticipation of performance deteriora-
tion. The possible deterioration of ESP's performance was commented on fay
the industry during the NSPS development and was investigated at that time
As discussed in Chapter 4, this issue has been reinvestigated and a problem
does not appear to exist in an ESP's capability in achieving the riSPS level
over the long term.
The Washington State Agency was asked about the need to develop
standards for water treatment ponds. The Agency commented that they have
no problems with odors from treatment ponds, since they have received no
complaints. However, agency personnel commented that the retention times
of the ponds at mills in Washington are not as long as ponds at mills
elsewhere in the country. Information from industry personnel indicate
that most new pulping installations have, and would in the future, install
condensate stripping systems to either reduce the BOD load to the treatment
7-1
-------
facility, control odors from the mill's recycle water, or control odor from
the ponds.4'5,5
The NCASI indicated three areas to consider in the review of the
NSPS.7 The first area pertained to a smelt dissolving tank which has not
been able to meet the TRS standard when using the control technology upon
which the standard is based. This was discussed in Chapter 5. The second
area pertained to the ability of a lime kiln equipped with an ESP to achieve
the TRS standard on a routine basis, since there is no opportunity to add
caustic to the scrubbing water for additional TRS control. The NSPS for
TRS from lime kilns is based on process controls, good lime mud washing,
and caustic scrubbing. This is discussed in more detail in Chapter 8.
The third area recommended for consideration is the use of a wet bottom ESP
using unoxidized black liquor. ESP design could promote contact between
the flue gas and the liquor in the bottom of the ESP resulting in an
increase in TRS concentrations as the furnace flue gas passes through the
ESP. These items were investigated and analyzed, and the findings are
discussed in Chapter 8.
Mill personnel contacted during this review indicated that, except for
the cases discussed above, they have no problems in complying with the
NSPS. As previously discussed in Chapter 5, some companies had problems in
initially achieving the NSPS during startup. These problems were design or
installation problems and were successfully corrected with modifications.
The companies also commented that they have not experienced any problems in
interpreting the NSPS or in degradation of control equipment performance in
terms of emissions or operability since initial installation. These units
subject to the NSPS have generally been in operation less than 3 years.
7-2
-------
7.1 REFERENCES
1. Letter from Greg Forte, Tennessee Department of Public Health, to
James Eddinger, U.S. EPA, dated March 16, 1982.
2. Letter from Glen Golson, Alabama Air Pollution Control Commission, to
James A. Eddinger, U.S. EPA, dated March 23, 1982.
3. Memo from Frank W. Lilley, EPA Region I, to Stanley T. Cuffe, U.S EPA
dated March 19, 1982.
4. Trip Report - Champion International Corporation, Courtland, Alabama
May 20, 1982.
5. Trip Report - Longview Fibre Company, Longview, Washington, May 26, 1982
6. Letter from John E. Pinkerton, NCASI, to James A. Eddinqer, U S EPA
dated April 26, 1982.
7. Letter from John E. Pinkerton, NCASI, to James A. Eddinqer, U S EPA
dated August 19, 1982. '
7-3
-------
8. ANALYSIS OF POSSIBLE REVISIONS TO THE STANDARDS
EPA regional offices, State agencies, and industry organizations were
contacted to determine the number and location of new and modified facilities
subject to the NSPS. Available NSPS compliance test data and opinions of
control agency and industry personnel regarding all facets of the NSPS were
solicited. As shown in Chapter 5, data on facilities subject to the NSPS
indicate that, except for two recovery furnaces and one smelt dissolving
tank, all facilities are in compliance with the NSPS. No new control
technology has emerged since the development of the NSPS. Some start-up
difficulties in complying with the NSPS for TRS emissions from recovery
furnaces, lime kilns, and smelt dissolving tanks have been reported. No
difficulties or problems have been reported in terms of enforcing the
NSPS. The only criticism voiced of the NSPS was excessive paperwork due
to recordkeeping requirements.
This chapter will analyze possible revisions to the standard for each
affected facility.
8.1 POSSIBLE REVISIONS TO NSPS
8.1.1 Recovery Furnace Systems
All recovery furnace systems subject to the NSPS, and for which compliance
test results are available, are in compliance with the particulate and TRS
standards, except for two recovery furnaces which have not achieved the TRS
standard. Although there has been improvement in the ESP's used, the
particulate test results do not provide sufficient justification for revision
of the particulate standard. There have been no improvements that would
provide justification for revising the TRS standard. TRS emissions are
controlled by maintaining proper process conditions and use of either a
non-contact furnace, or a direct contact furnace with black liquor oxidation.
Method 9 opacity data were not supplied with the compliance test
results for particulate matter. The particulate and Method 9 opacity data
obtained during the NSPS development were reexamined, and provided no
justification for revising the present opacity standard.
8-1
-------
The NCASI indicated that a potential problem area might exist in
complying with the NSPS for TRS emissions if a wet-bottom ESP with unoxidized
black liquor is used. Compliance test results obtained on six wet-bottom
ESP's installed on recovery furnaces subject to the NSPS using unoxidized
black liquor indicated that all but two are in compliance with the TRS
standard. One mill initially had a problem during startup with the
original design of the wet-bottom ESP which resulted in an increase in
TRS emissions. Modifications were made to the ductwork and internal
baffling to decrease contact of the unoxidized liquor with the flue gas.
This recovery furnace is now in compliance. The two recovery furnaces not
in compliance plus two additional new installations with wet-bottom ESP's
have not achieved the TRS standards due to an increase in TRS emissions
through the ESP. These installations are in the startup phase. Similar
modifications have been made to these ESP's, and although the TRS emissions
have been reduced, they have not met the NSPS. Additional baffling as
well as other alternatives to achieve compliance are being investigated
by the companies and vendors.
Presently, the mills with wet-bottom ESP's using unoxidized black
liquor are not required to continuously monitor the recovery furance system
until the continuous monitoring system specifications are promulgated.
Therefore, continuous monitoring data will not be available to show that
the NSPS can be continuously met under normal operating conditions at these
installations.
The types of ESP's tested by the EPA (in 1972) during the NSPS
development were dry-bottom ESP's or wet-bottom ESP's installed on direct-
contact recovery furnaces (using oxidized black liquor). Wet-bottom
ESP's are extensively used on recovery furnaces equipped with direct-
contact evaporators.1 Oxidizing the black liquor prior to its introduction
into the direct-contact evaporator and, therefore, prior to the wet-bottom
ESP, prevents stripping of the TRS during contact with the flue gas.
Direct-contact recovery furnaces subject to the NSPS and equipped with
wet-bottom ESP's employed black liquor oxidation and are in compliance
with the NSPS.2>3
8-2
-------
The non-contact recovery furnace was first introduced in the late
1960's. The ESP's installed on the first several non-contact recovery
furnaces had dry-bottom designs. However, many companies favored the wet-
bottom design because of increased maintenance requirement and salt cake
handling problems of the dry-bottom units. The first wet-bottom ESP
using unoxidized black liquor was installed on a non-contact recovery
furnace around 1971. An additional five wet-bottom ESP's using unoxidized
black liquor began operation between 1974 and 1978. The question of
stripping TRS from unoxidized black liquor at the ESP wet-bottom by the
flue gas was not a priority concern by the industry until the NSPS was
promulgated and the use of non-contact furnaces and wet-bottom ESP's
increased.1 In 1978, the NCASI reviewed the situation and tested a wet-
bottom ESP. The study indicated that there was little, if any, TRS
contribution from the flue gas contact with the unoxidized black liquor.4
Therefore, companies elected to install wet-bottom ESP's because information
available indicated that the TRS standards could be achieved if TRS levels
from the recovery furnace were low.5
As stated previously, some companies are having problems in achieving
the NSPS due to a pickup of TRS emissions in the wet-bottom ESP because
of contact between the flue gas and the unoxidized black liquor. These
companies have, however, indicated that the TRS standard is being achieved
from the recovery furnace prior to the ESP.6 The NCASI is currently
gathering information on these wet-bottom ESP installations in an effort
to identify the factors influencing the concentrations of TRS at the ESP
outlet. Data on internal ESP design, air flow within the ESP, flue gas
characteristics, and black liquor characteristics are being obtained from
the mills. This study should be completed in early 1984 and will
hopefully identify modifications that can be made to bring these units
into compliance.
Based on the information available at this time, there is no
justification for revising the NSPS for recovery furnace systems because:
(1) some wet-bottom units are capable of achieving the NSPS, (2) the
problem is now known and companies have indicated that dry-bottom units
will be installed until the problem with wet-bottom units is solved6,
8-3
-------
(3) the companies and vendors are still investigating ways to achieve the
NSPS for those out of compliance, and (4) the NCASI is conducting a study
to determine the factors that influence TRS emissions from a wet-bottom
ESP.
3.1.2 Smelt Dissolving Tanks
The NSPS is based on the use of a low energy scrubber and the use of
water that is not highly contaminated with dissolved sulfides. The
emission control techniques being used are the same as those upon which
the NSPS is based. All smelt dissolving tanks subject to the NSPS and
for which compliance test data are available are in compliance with the
NSPS, except for one mill which has not been able to meet the TRS standard.
The mill has attempted to reduce the TRS emissions using various scrubbing
liquids, but has not found a solution to continuously achieve the TRS
standard. A continuous monitor was installed to help analyze the problem.
Results of the monitoring indicate that the TRS standard can be achieved,
but not on a continuous basis. Data obtained during the NSPS development
also indicated that there is a large range of variation in emissions from
even well-controlled facilities. The level of the NSPS was set to allow
for some degree of variation in the emissions. However, no continuous
monitoring data were available at the time of the NSPS development to
indicate the variation in emissions from an individual smelt tank. The
data available were from short-term source tests which indicated only the
variation in emissions between well-control!ed smelt dissolving tanks.
Therefore, there is sufficient justification to revise the TRS standard
for smelt dissolving tanks based upon the analysis of the monitoring data.
The impact of this revision will be minimal, since it will not result in a
change in the basis for the standard (i.e., control technology used).
The promulgated standard requires that compliance be determined by
summation of the mass emission rate of each TRS compound from the concentra-
tion results of Method 16. This provision [FR 60.285(d)(3)] was added
during promulgation of the NSPS to enable a more accurate determination of
the mass emission rate. Since promulgation, the EPA has proposed (46 FR 31904;
Method 16A (Impinger Technique) as an equivalent method to Method 16 for
8-4
-------
determining TRS concentrations. Method 16A is simple and much cheaper (60
to 30 percent lower costs) to operate, and involves fewer and less complicated
components that reduce chances of measurement error. Method 16A results in
a single value for all reduced sulfur compounds and does not measure each
individual compound. Therefore, as presently written, the regulation would
prevent the use of Method 16A (when promulgated) to determine compliance
for smelt dissolving tanks.
The mass TRS values reported in the NSPS background information document
and upon which the NSPS is based are in terms of equivalent H2S. That is,
the total TRS concentration measured by Method 16 was assumed to be h^S
when calculating the mass emission rate. When the provision for calcu-
lating the mass rate was modified at promulgation, no changes were deemed
necessary to the numerical standards, because the difference in the calculated
mass values were not significant. Therefore, since Method 16A is equivalent
to and less expensive than the existing method, sufficient justification
exists to revise the standard to allow the industry to use Method 16A
in determining compliance for smelt dissolving tanks. The revision
would not change the numerical standards for reasons previously discussed.
The standard could be revised to be identical to the lll(d) guideline
level which defines the mass level in terms of TRS as HgS. The impact of
this revision would be negligible, since it would not result in a change
in the basis for the standards (i.e., control technology used), but only
in how the mass value is calculated and reported.
3.1.3 Lime Kiln
Lime kilns subject to the NSPS and for which compliance test results
are available are in compliance with both the particulate and TRS standards.
One lime kiln which is in start-up has not been compliance tested, but
has reported difficulties in achieving the TRS standard. This lime kiln
is equipped with an ESP and does not have the opportunity to add caustic
to the scrubbing water for additional TRS control. The TRS standard is
based on caustic addition in conjunction with process controls and good
lime mud washing. Most lime kilns subject to the NSPS have demonstrated
compliance with the NSPS without caustic addition. These kilns are
8-5
-------
controlled for TRS by maintaining proper process conditions and good
mud washing. The results of these short-term compliance tests along with
short-term data obtained during the NSPS development indicate that the
level of the TRS standard can be met wtihout caustic addition. However,
the NSPS was based on the long-term monitoring data from a lime kiln
controlled by a caustic scrubbing system. Presently, the mills are not
required to continuously mom"tor the lime kiln system until the continuous
monitoring system specifications are promulgated. Until continuous
monitoring data are available, insufficient data exist for making any
definitive judgement about the ability to continuously achieve the TRS
standard without caustic addition.
The one lime kiln equipped with the ESP which has not yet met the
TRS limit is in the start-up phase. The mill is trying to improve the
lime mud quality in order to achieve the NSPS. If improvements in the
mud washing system are not sufficient to reduce TRS emissions to the MSPS
level, lime mud oxidation will likely be employed. Lime mud oxidation is
already employed at the mill for the other existing lime kilns. This
lime kiln also operates at a lower cold-end temperature than other kilns
subject to the NSPS. It would appear that raising the cold-end temperature
from 300°F to 450°F (as costed out in the NSPS) is also an alternative if
the mud washing proves insufficient.
Therefore, since the lime kiln equipped with the ESP is still in
startup, and there exist other control alternatives which other kilns are
already using, there is not sufficient justification at the present time
for revision to the NSPS. Information pertaining to this area will be
analyzed as it becomes available.
The particulate standard is based on venturi scrubbers because the TRS
standard is based on caustic addition, and during the original MSPS develop-
ment the industry contended that the use of an ESP could cause an explosion.
The industry postulated that the noncondensable gases added to the kiln
combustion air for incineration might explode in the ESP when flameout
occurred in the kiln. Only one lime kiln installation was controlled by an
ESP at the time of the NSPS development. Since the proposal of the NSPS,
ESP's have been installed on lime kilns at two mills. Both ESP systems are
equipped with automatic cutoff systems for preventing explosions.
8-6
-------
The NSPS is, nevertheless, based on a venturi scrubber with caustic
addition. A revision to the particulate standard based on the performance
of an ESP cannot be considered, unless it is determined that the TRS standard
can be continuously achieved without caustic scrubbing or a corresponding
revision to the TRS standard is made reflecting the level achievable with
process controls and good mud washing only. The latter option was considered
during the NSPS development program. However, no new data or control
systems have come available that would justify revising the NSPS.
As discussed in Chapter 4, research has been done on the use of a
baghouse for kraft lime kilns. However, full-scale systems using baghouses
have not been installed, so long-term performance data are not available,
and therefore, the baghouse cannot be used as a basis for the NSPS or a
revision.
8.1.4 Digester and Multiple-Effect Evaporator Systems
In all cases, the TRS emissions from digesters and multiple-effect
evaporators subject to the NSPS are controlled by incineration of the
noncondensable gases in either a lime kiln, power boiler, or separate
incinerator. Incineration of the gases in the lime kiln is the basis for
the NSPS. Therefore, there is no justiciation for revision to the present
NSPS for digesters and multiple-effect evaporators.
8.1.5 Brown Stock Washer Systems
TRS emissions from washer systems subject to the NSPS are generally
controlled by incineration of the vent gases. The NSPS states that a
washer system (or a black liquor oxidation system) can be exempted from
the standard if the owner or operator demonstrates that incineration of the
exhaust gases in an existing facility is technologically or economically
not feasible. One company has received an exemption from the standard for
an additional washer stage.
As discussed in Chapter 6, the cost of control of TRS emissions from
vacuum drum type washer systems is estimated to be about $2,400/ton of TRS
removed. This value is considerably higher than that for the other TRS
sources. Vacuum drum type washer systems were generally used at the time
of the NSPS development. However, the majority of washer systems subject
to the NSPS are the diffusion types which do not have the hooding require-
ments or the large gas volumes of the vacuum drum types. As shown in
8-7
-------
Chapter 6, data supplied by one company indicate that the capital cost of
controlling emissions from diffusion washers are about one-fifth the cost
of controlling emissions from vacuum drum washers. Therefore, the cost-
effectiveness when using a diffusion washer is estimated to be about
3900/ton of TRS removed and will be similar to the other TRS sources.
In summary, there is a trend towards use of diffusion washers that are
less costly to control. Therefore, there is no justification for revision
to the standard for brown stock washer systems.
8.1.6 Black Liquor Oxidation Systems
Presently, only two black liquor oxidation (BLO) systems are subject
to the NSPS. TRS emissions from one system are controlled by incineration
of the vent gases in a power boiler, and the other system uses molecular
oxygen as the oxidizing medium so there are no gases vented. These two
control techniques were evaluated during the NSPS development.
As presented in Table 6-1, the cost effectiveness of incinerating the
BLO gases is over $9,000 per ton of TRS removed. As discussed in Chapter 6,
the high cost effectiveness is due to the small amount (16 tons per year
from a 1,000 ton per day mill) of TRS controlled. Although the industry
has demonstrated the control technology, the cost-effectiveness of control
would provide justification for revision to the standard for BLO systems.
The impact of rescinding the standard for BLO systems would depend
on the number of direct contact recovery furnaces installed. As previously
discussed, the majority of recovery furnaces being installed are the non-
contact type furnaces which do not require BLO systems to achieve the NSPS.
This trend towards non-contact furnaces existed before promulgation of the
NSPS and is expected to continue because of the economics of operating a
non-contact furnace system in comparison with a direct-contact furnace
system. Therefore, rescinding the BLO standard would have a minimal impact
on TRS emissions.
8.1.7 Condensate Stripping Systems
TRS emissions from condensate stripping systems subject to the NSPS
are controlled by incineration of the vent gases in either a lime kiln or
separate incinerator. Incineration is the basis for the NSPS. There is no
justification for revision to the present NSPS for condensate stripping
systems.
8-8
-------
8.1.8 Exemption for Brown Stock Washers and Black Liquor Oxidation Systems
Presently, the NSPS includes an exemption for new brown stock washer
systems or black liquor oxidation systems where combustion of the vent
gases in an existing facility is not feasible from a safety or economic
standpoint. This provision [Section 60 .283(a)(1)( IV)] was included because
the EPA'agreed with the industry's comment that older recovery furnaces do
not have the capability to accept large volumes of gases, and the costs
associated with altering these recovery furnaces could be prohibitive.
Information obtained during the NSPS review indicates that power boilers
instead of recovery furnaces are bein) used as the incin'Vation device for
the high volume gas streams. Discussions with mill personnel and an equipment
vendor indicate no operating problems with using the power boiler as an
incineration device.2,7 jne costs associated with using the power boiler
would be similar to the costs presented in Table 6.1 which are based on
using a new recovery furnace designed to handle the gases. However, the
one mill which received an exemption from the NSPS only added a single
additional washing stage. This was not an entirely new washer installation.
A diffusion washer stage is not likely to be added to an existing vacuum
drum washing line.8 In addition, the power boilers at many existing
mills may be physically located at a great distance from the washers.
Therefore, there is not sufficient justification for removing this exemption.
8.2 POSSIBLE REVISIONS TO MONITORING REQUIREMENTS
The existing NSPS for kraft pulp mills require that continuous monitoring
systems be installed, calibrated, maintained, and operated to measure:
1. Opacity and TRS emissions from the recovery furnace system,
2. TRS emissions from the lime kiln system, and
3. Combustion temperature at the point of incineration of gases which
are emitted from digesters, multiple-effect evaporators, brown stock washers,
black liquor oxidation systems, and condensate stripping systems.
The industry has commented that monitoring of combustion temperature
is unnecessary when a power boiler, recovery furnace, or lime kiln is used
as the incineration device, because the flame temperatures are typically
1600°F or higher. We agree that if compliance with the minimum temperature
of 1200°F for at least 0.5 second has been demonstrated, there is no necessity
8-9
-------
to monitor combustion temperature, since these devices will normally operate
at temperatures higher than required by the NSPS. Therefore, there is
sufficient justification to revise the NSPS to require monitoring of the
combustion temperature only when the gases are combusted in an incinerator.
The volume concentration of TRS emissions can be monitored by use of
monitoring systems. There are no process or control device parameters that
are indicators of concentration of TRS emissions generated from recovery
furnaces and lime kilns. As discussed in Chapter 2, TRS emissions from
these sources are affected by many operating parameters. Therefore, since
the gas stream TRS monitoring system is the only method of monitoring
concentrations of TRS emissions from these affected facilities, there is no
justification for revising the NSPS requirement for TRS monitoring. The
continuous monitoring system specifications for TRS monitors are expected
to be promulgated in the near future.
The only criticism voiced of the NSPS was excessive paperwork due to
recordkeeping requirements. These recordkeeping requirements are specified
in Section 60.7(b) and (c) of the general provisions for new source performance
standards. Section 60.7(b) requires the operator of any source subject to
a standard to maintain records of the occurrence and duration of any startup,
shutdown, or malfunction in the operation of an affected facility, any
malfunction of the air pollution control equipment, or any periods during
which a continuous monitoring system or monitoring device is inoperative.
Section 60.7(b) also requires any operator to maintain a file of all measure-
ments, including continuous monitoring system, all continuous monitoring
system performance evaluations, all continuous monitoring system calibration
checks, and adjustments and maintenance performed on these systems. Section
60.7(c) requires the operator to submit a written report of excess emissions
to the Administrator for every calendar quarter.
8.3 EXTENSION TO OTHER SOURCES
During the NSPS proposal (41 FR 42013), it was indicated that all
process gas streams identified as major sources of TRS and particulates at
kraft pulp mills were covered by the NSPS, except for power boilers and
water treatment ponds. Power boilers are a source of particulate emissions
8-10
-------
at kraft pulp mills. They were not included in the NSPS because the EPA
intended to develop standards for industrial boilers which would include
power boilers at kraft mills. The standard for industrial boilers is in
the process of being proposed.
Water treatment ponds were considered to be potentially significant
sources of TRS emissions at some kraft pulp mills. Standards for treatment
ponds were not included in the NSPS, because methods of measuring TRS
emissions from the ponds were not available, and further investigation of
emission control techniques were needed. This rationale still applies; 'a
proven sampling method is not yet available. It was felt that emissions
from ponds could be controlled by treating the process condensate stream in
a condensate stripper system prior to discharge to the ponds. However,
bacterial action could generate additional TRS in some ponds. Most new
pulping installations have installed condensate stripping systems for
controlling either BOD, odor from recycle water, or odor from ponds.
Therefore, there is presently no justification to revise the MSPS to
include treatment ponds.
A development program for sampling hydrocarbon emissions from a pond
is planned for another source category. Once the sampling method is demon-
strated, it could be adapted for sampling TRS emissions. A testing program
could then be performed to determine the magnitude of the TRS problem for
analysis during the next review of the NSPS.
8.4 EXTENSION TO OTHER EMISSIONS
Kraft pulp mills are also sources of sulfur dioxide (S02), nitrogen
oxides (NOX), and carbon monoxide (CO) emissions. The recovery furnace,
lime kiln, and power boiler have been identified as sources of S02- Power
boilers, as stated in Section 8.2, are being covered under a separate
industry category. EPA tests conducted during the NSPS development on two
recovery furnaces and three lime kilns show S02 emission levels of about
2.0 g/kg ADP (about 70 ppm) and 0.15 g/kg ADP (about 30 ppm), respectively.
Revisions to the NSPS to include S02 emissions from recovery furnaces and
lime kilns are not justified at the present time because no demonstrated
control techniques, considering costs, have been identified for these
8-11
-------
facilities. A study should be performed to investigate the magnitude of
S02 emissions from recovery furnaces, the factors that affect emissions,
the emission control methods available, and to determine whether regulation
is appropriate.
Recovery furnaces and lime kilns are also sources of CO and NOX. CO
emissions were measured by the EPA during the NSPS development on two
recovery furnaces and showed levels of about 1.3 g/kg ADP (about 100 ppm).
CO emissions from lime kilns average about 5 g/kg ADP. EPA tests on two
recovery furnaces showed NOX levels of about 1.0 g/kg ADP (about 50 ppm).
NOX emissions from lime kilns at kraft pulp mills are estimated to be about
0.6 g/kg ADP (150 ppm).9 Revisions to the NSPS to include CO and NOX
emissions are not justified, since no control techniques have been demonstrated
in the kraft pulping industry for these facilities.
8-12
-------
8.5 REFERENCES
1. Letter and enclosure from John E. Pinkerton, NCASI, to James A.
Eddinger, U.S. EPA, dated April 8, 1983.
2. Trip Report - Champion International Corporation, Courtland Alabama
May 20, 1982.
3. Memo from Jam's Forehand, EPA Region VI, to Jim Eddinger, EPA, dated
June 29, 1982.
4« The Effect of a Wet-Bottom Precipitator upon Recovery Furnace TRS
Emissions, NCASI Atmospheric quality Improvement Technical R»I let-in Mn. 98
September 1978.
5. Letter from Alan M. Lindsey, International Paper Company, to Jack R.
Farmer, U.S. EPA, dated April 22, 1983.
6. Meeting Report - Industry Representatives and EPA Personnel, Durham
North Carolina, April 14, 1983.
7. Telephone Conversation between James Dickinson, Babcock & Wilcox
Company, and James Eddinger, U.S. EPA, on September 9, 1982.
8. Letter from Keith M. Bentley, Georgia-Pacific Corporation to Jack R
Farmer, U.S. EPA, dated April 22, 1983.
9. A Study of Nitrogen Oxides Emissions from Lime Kiln, NCASI Technical
Bulletin No. 107, April 1980.~~~
8-13
-------
9. CONCLUSIONS AND RECOMMENDATIONS
The primary objective of this report has been to assess the need for
revision of the existing NSPS for kraft pulp mills. Conclusions pertaining
to the particulate and TRS standards are reviewed below.
9.1 REVISION OF THE CURRENT STANDARDS
9.1.1 Conclusions Based on Control Technology
Since the standards were originally promulgated, no new control
•technology has emerged. Compliance test results for facilities subject to
the NSPS indicate that all are in compliance with the NSPS, except for two
recovery furnaces and one smelt dissolving tank which have not been able
to meet the TRS standard.
Four of eight recovery furnace installations with wet-bottom ESP's
using unoxidized black liquor have experienced an increase in TRS emissions
through the ESP resulting in TRS emissions above the NSPS. Modifications
made to one installation resulted in the unit achieving compliance.
Modifications made to the other four installations resulted in decreased
levels, but above the NSPS. Further modifications to these units are
being investigated by the companies.
Compliance test results indicate that there is a larger range of
variation in TRS emissions from well-controlled smelt dissolving tanks than
indicated by data obtained during the NSPS development.
Most lime kilns subject to the NSPS have demonstrated compliance
with the TRS standard without caustic addition. Presently, the mills are
not required to continuously monitor the TRS emissions. Until continuous
monitoring data are available, insufficient data exist for making any
judgement about the ability to continuously achieve the TRS standard
without caustic scrubbing.
9.1-2 Conclusions Based on Economic Considerations
There are 14 new recovery furnaces, 16 new smelt dissolving tanks,
and 19 new lime kilns in operation which are subject to the NSPS.
The growth rate averaged 3.5 percent per year between 1978 and 1981.
The forecast is for a 2.3 percent decline in overall industry production in
1982, with a 4.6 percent growth in 1983.
9-1
-------
0 It is cost effective to control participate sources since the
recovered participates are valuable materials.
0 The cost effectiveness of the various TRS controls ranges from zero
dollars per ton for the smelt dissolving tank to $9,250 per ton for the
black liquor oxidation system. The brown stock washer system with a
cost effectiveness of $2,500 per ton is the only other affected facility
with a cost effectiveness about $1,000 per ton. However, this cost was
based on the use of a vacuum drum type washing system, and newer mills
are installing diffusion washers with a cost-effectiveness of $900 per
ton of TRS controlled.
0 The high cost effectiveness of the BLO systems is due to the small
amount (16 tons per year) of TRS controlled.
9.1.3 Conclusions Based on Other Considerations
0 If compliance with minimum temperature of 1200°F for at least 0.5
second has been demonstrated, there is no necessity to monitor combustion
temperatures from devices other than an incinerator, since these devices
will normally operate at temperatures higher than required by the NSPS.
0 There are no process or control device parameters that are indicators
of concentration of TRS emissions generated from recovery furnaces and lime
kilns.
0 Method 16A is equivalent to and less expensive than Method 16.
0 The present NSPS prevents the use of Method 16A to determine compliance
for smelt dissolving tanks.
0 High volume, non-condensable gases from facilities subject to the
NSPS are generally incinerated in a power boiler instead of a recovery
furnace. Power boilers are used at all kraft pulp mills.
9.1.4 Recommendations on Revision of Current Standard
Based upon the technological, economic, and monitoring conclusions,
the following recommendations are made:
0 Revise the existing TRS standard for the smelt dissolving tank to
raise the standard to the level indicated by compliance data.
0 Revise the NSPS to rescind the requirement for controlling TRS
emissions from black liquor oxidation systems.
0 No revision of either the particulate or opacity standards should be
considered at the present time.
9-2
-------
No revision to the other TRS standards (other than for the smelt
dissolving tank and BLO system) should be considered at the present time.
0 Revise the units of the TRS standard for smelt dissolving tanks from
g/kg BLS to g/kg BLS as H2S to permit the use of Method 16A.
No revision to the exemption for brown stock washer systems where
combustion of these gases in an existing facility is not feasible from a
safety or economic standpoint should be considered at this time.
Revise the existing NSPS to rescind the requirement for monitoring
the combustion temperature when a device other than an incinerator is used
(i.e., the lime kiln, recovery furnace, or power boiler), if compliance
with a minimum temp-erature of 1200°F for at least 0.5 second has been
demonstrated.
No revision of the TRS monitoring requirements should be considered
at the present time.
9.2 EXTENSION OF STANDARDS
9.2.1 Conclusions Based on Control Technology
Cost-effective control of sulfur dioxide has not been demonstrated
in the kraft pulp industry.
Control of nitrogen oxides and carbon monoxide has not been
demonstrated in the kraft pulp industry.
Control of power boilers will be regulated under standards being
proposed for industrial boilers.
(Methods of measuring TRS emissions from water treatment ponds are
not available.
9-2.2 Recommendations on Extension of Standards
Standards for other processes and pollutants should not be considered
at the present time.
A study should be performed to investigate sulfur dioxide emission
control methods available and to determine whether regulation is appropriate.
9-3
-------
APPENDIX A
Kraft Pulp Mills Subject to NSPS
Company
Location
Champion International Court!and, Alabama
Union Camp
Container Corporation
Hammermill Paper
Nekoosa Paper
Georgia-Pacific
Simpson Paper
Buckeye Cellulose
Container Corporation
Georgia Kraft
Prattville, Alabama
Brewton, Alabama
Selma, Alabama
Ashdown, Arkansas
Crossett, Arkansas
Anderson, California
Perry, Florida
Fernandina Beach,
Florida
Rome, Georgia
A-l
Sources
Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
Digester
Black Liquor Oxidation System
Multiple-Effect Evaporators
Brown Stock Washer System
Condensate Stripper System
Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
Multiple-Effect Evaporators
Brown Stock Washer System
Digesters (2)
Digesters
Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
Multiple-Effect Evaporators
Digesters
Brown Stock Washer System
Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
Black Liquor Oxidation System
Digesters
Multiple-Effect Evaporators
Brown Stock Washer System
Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
Multiple-Effect Evaporators
Lime Kiln
Lime Calciner
Cross Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
-------
Company
Continental Forest
Buckeye Cellulose
Location
Port Wentworth,
Georgia
Oglethorpe, Georgia
Western Kraft
Continental Forest
International Paper
Hawesville, Kentucky
Hodge, Louisiana
Mansfield, Louisiana
Boise Cascade
Rumford, Maine
:ha
-------
Company
Continental Forest
Crown Zellerbach
Longview Fibre
Boise Cascade
Location
Hopewel1, Virginia
Camas, Washington
Longview, Washington
Wallula, Washington
Source
Recovery Furnace
Smelt Dissolving Tank
Lime Kiln
Lime Kiln
Recovery Furnace
Smelt 01ssolving Tank
L i me Kiln
A-3
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. 2.
4. TITLE AND SUBTITLE
Review of New Source Performance Standards
Pulp Mills
7 AUTHOR(S)
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standar
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27
12. SPONSORING AGENCY NAME AND ADDRESS
DAA for Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency, RTP,
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
r v1 "v ft Sep^^mb^r IQS^
1 UI l\( a 1 L 6 PERFORMING ORGANIZATION CODE
8 PERFORMING ORGANIZATION REPORT NO
10. PROGRAM ELEMENT NO.
dc
UJ 11 CONTRACT/GRANT NO.
711
13. TYPE OF REPORT AND PERIOD COVERED
14 SPONSORING AGENCY CODE
N.C. 27711 EPA 200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report reviews the current New Source Performance Standards for
Kraft Pulp Mills. It includes a summary of the current standards, the
status of current applicable control technology, and the ability of
mills to meet the current standards. Recommended changes to the
existing standards are discussed.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Air Pollution
Kraft Pulp mills
PartJculate matter
Total Reduced Sulfur
Standards of Performance
Pol lution Control
18. DISTRIBUTION STATEMENT ,. . -'
Release Unlimited ;. - • : ;,:. ., y
" '? ' ; " ' * '•','<
b. IDENTIFIERS/OPEN ENDED TERMS C. COSAT1 Field/Group
Air Pollution Control 13 B
19,;SECURITY CLASS { This Report) 21. NO. OF PAGES
'.Unclassified 82
20, SECURITY CLASS (This page I 22. PRICE
' Unclassified
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
------- |