United States
Environmental Protection
Agency
Industrial Environmental Research  EPA-600/7-79-141
Laboratory         August 1979
Cincinnati OH 45268
Research and Development
Environmental
Considerations of
Selected Energy-
Conserving
Manufacturing
Process Options

Volume XVI.
Sulfur Oxides
Summary Report

Interagency
Energy/Environment
R&D Program Report

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology.  Elimination of traditional grouping  was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.   Environmental  Health Effects Research
      2.   Environmental  Protection Technology
      3.   Ecological Research
      4.   Environmental  Monitoring
      5.   Socioeconomic Environmental Studies
      6   Scientific and Technical Assessment Reports (STAR)
      7   Interagency  Energy-Environment Research and Development
      8.   "Special" Reports
      9.   Miscellaneous Reports

This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded  under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments  of,  and development of, control technologies for energy
systems; and  integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                EPA-600/7-79-141
                                                August 1979
   ENVIRONMENTAL  CONSIDERATIONS OF SELECTED
ENERGY-CONSERVING MANUFACTURING PROCESS  OPTIONS

   Volume XVI.  Sulfur Oxides Summary Report
                       by

            Arthur  D.  Little, Inc.
        Cambridge,  Massachusetts 02140
             Contract No.  68-03-2198
                 Project Officer

             Herbert S. Skovronek
   Power Tec-hnology and Conservation Branch
 Industrial Environmental Research Laboratory
            Cincinnati, Ohio 45268
   INDUSTRIAL ENVIRONMENTAL  RESEARCH LABORATORY
        OFFICE OF RESEARCH AND DEVELOPMENT
      U.S. ENVIRONMENTAL PROTECTION AGENCY
             CINCINNATI, OHIO 45268

        For sale by the Superintendent of Documents, U.S. Government
              Printing Office, Washington, D.C. 20402

-------
                                 DISCLAIMER
     This report has been reviewed by the Industrial Environmental Research
Laboratory-Cincinnati, U.S. Environmental Protection Agency, and approved for
publication.  Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
                                     ii

-------
                               FOREWORD
     When energy and material resources are extracted, processed, con-
verted, and used, the related pollutional impacts on our environment,
and even on our health, often require that new and increasingly more
efficient pollution control methods be used.  The Industrial Environ-
mental Research Laboratory - Cincinnati (lERL-Ci) assists in developing
and demonstrating new and improved methodologies that will meet these
needs both efficiently and economically.

     This report summarizes information on sulfur oxides from a study
of 13 energy-intensive industries.  If implemented over the coming 10
to 15 years, these processes and practices could result in more effec-
tive utilization of energy resources.  The study was carried out to
assess the potential environmental/energy impacts of such changes and
the adequacy of existing control technology in order to identify po-
tential conflicts with environmental regulations and to alert the
Agency to areas where its activities and policies could influence the
future choice of alternatives.

     The results will be used by the EPA's Office of Research and De-
velopment to define those areas where existing pollution control tech-
nology suffices, where current and anticipated programs adequately ad-
dress the areas identified by the contractor, and where selected pro-
gram reorientation seems necessary.

     Specific data will also be of considerable value to individual
researchers as industry background and in decision-making concerning
project selection and direction.

     The Power Technology and Conservation Branch of the Energy Sys-
tems-Environmental Control Division should be contacted for additional
information on the program.
                                        David G. Stephan
                                            Director
                          Industrial Environmental Research Laboratory
                                           Cincinnati
                                  iii

-------
                            ABSTRACT
     Under EPA Contract No. 68-03-2198, Arthur D. Little, Inc.
undertook a study of the "Environmental Considerations of
Selected Energy-Conserving Manufacturing Process Options."
Some 80 industrial process options were examined in 13 industrial
sectors.  Results were published in 15 volumes, including a
summary, industry prioritization report, and 13 industry
oriented reports (EPA-600/7-76-034 a through o).

     This present report summarizes the information regarding
particulates in the 13 industry reports.  Four parallel reports
treat sulfur oxides, nitrogen oxides, solid residues, and toxics/
organics.  All of these pollutant-oriented reports are intended
to be closely used with the original 15 reports.
                               iv

-------
                               CONTENTS
Foreword                                                        ill
Abstract                                                         iv
Tables                                                           vi
English-Metric (SI) Conversion Factors                         viii
1.     INTRODUCTION                                               1

       BACKGROUND AND PURPOSE                                     1
       APPROACH                                                   2

2.     FINDINGS AND R&D OVERVIEW                                  5

       FINDINGS                                                   5
       R&D AREAS                                                 21

3.     PROCESSES AND SO  EMISSIONS                               25

       BASES OF CALCULATIONS                                     25
       SOX CONTROL METHODS                                       26
       PETROLEUM REFINING                                        28
       CEMENT                                                    37
       OLEFINS                                                   40
       AMMONIA                                                   46
       ALUMINA AND ALUMINUM                                      50
       PULP AND PAPER                                            58
       GLASS                                                     62
       COPPER                                                    69
       CHLOR-ALKALI                                              76
       IRON AND STEEL                                            81
       PHOSPHORUS/PHOSPHORIC ACID                                85
       FERTILIZERS                                               89
       TEXTILES                                                  95

REFERENCES                                                       99

-------
                                   TABLES
Number
    1   Projected U.S. Production in Industries Studies                  3

    2   Summary of Estimated Annual SOX Emissions                        6

    3   SOX Nationwide Emission Estimates (1974)                        15

    4   Estimated Increase in Controlled SOX Emissions 1989-1974
             Assuming Industry Expands Using Process Types Indicated    17

    5   Estimated Range in Controlled SOX Emissions in 1989 for
             New Processes Likely to be Implemented                     19

    6   Estimated SO  Emission Factors - Petroleum Refining Industry    29

    7   Estimated SO  Emissions - Petroleum Refining Industry           32

    8   Estimated SO  Emissions - Cement Industry                       41

    9   Estimated SO  Emissions - Cement Production                     42
                    x

   10   Sulfur Distribution and Estimated SO  Emission Factors -
             Olefin Industry                                            44

   11   Estimated SO  Emissions - Olefins Industry                      47

   12   Estimated SO  Emissions - Ammonia Industry                      49
                    X

   13   Estimated SO  Emission Factors - Alumina and Aluminum
             Industry                                                   55

   14   Estimated SO  Emissions - Alumina and Aluminum Industry         56

   15   Uncontrolled Emissions of TRS and Proposed Emission Standards
             for New Kraft Pulp Mills                                   59

   16   Estimated SO  Emission Factors - Pulp and Paper Industry        63

   17   Estimated SO  Emissions - Pulp and Paper Industry               64
                    X

   18   Estimated SO  Emission Factors - Glass Industry                 70
                    x
                                     VI

-------
                                 TABLES (cont.)



Number



   19   Estimated SO  Emissions - Glass Industry                        71



   20   Estimated SO  Emission Factors - Primary Copper Industry        77
                    X


   21   Estimated SO  Emissions - Copper Smelting                       78
                    X


   22   Estimated SO  Emissions - Chlor-Alkali Industry                 82



   23   Estimated SO  Emissions - Iron and Steel Industry               86



   24   Estimated SO  Emissions - Phosphorus/Phosphoric Acid

             Production                                                 90



   25   Estimated SO  Emissions - Fertilizer Industry                   93



   26   Estimated SO  Emissions - Textile Industry                      98
                                     vii

-------
                  ENGLISH-METRIC (SI)  CONVERSION FACTORS
 To Convert  From
To
Metre2
Pascal
Metre
Joule
Pascal-second
Degree Celsius
Degree Kelvin
Metre
3
Metre /sec
Metre3
2
Metre
Metre/sec
2
Metre /sec
3
Metre
Watt
Watt
Watt
Metre
Joule
Metre3
Metre
Metre
Metre
Pascal-second
Newton
Kilogram
Kilogram
Kilogram
Kilogram
Kilogram
Kilogram
Multiply By
4,046
101,325
0.1589
1,055
0.001
t° = (t° -32)/1.8
0.3048
0.0004719
0.02831
0.09290
0.3048
0.00002580
0.003785
745.7
746.0
735.5
0.02540
3.60 x 106
1.000 x 10"3
1.000 x 10~6
0.00002540
1,609
0.1000
4.448
0.4536
0.02916
1,016
1,000
907.1
1,000
Acre
Atmosphere  (normal)
Barrel  (42  gal)
British Thermal  Unit
Centipoise
Degree  Fahrenheit
Degree  Rankine
Foot
    3
Foot /minute
Foot
    2
Foot
Foot/sec
Foot2/hr
Gallon  (U.S. liquid)
Horsepower  (550  ft-lbf/sec)
Horsepower  (electric)
Horsepower  (metric)
Inch
Kilowatt-hour
Litre
Micron
Mil
Mile (U.S.  statute)
Poise
Pound force (avdp)
Pound mass  (avdp)
Ton (Assay)
Ton (long)
Ton (metric)
Ton (short)
Tonne
Source:  American National Standards Institute, "Standard Metric Practice
         Guide," March 15, 1973,   (ANS72101-1973)  (ASTM Designation E380-72)
                                     viii

-------
                              SECTION 1

                             INTRODUCTION


BACKGROUND AND PURPOSE

     During 1975 and the first half of 1976, under EPA Contract No.
68-03-2198, Arthur D. Little, Inc., undertook a study of the "Environ-
mental Considerations of Selected Energy-Conserving Manufacturing Pro-
cess Options" in 13 energy-intensive industry sectors for the U.S. En-
vironmental Protection Agency (EPA).  The results of these studies were
published in the following reports:

     •    Volume I — Industry Summary Report (EPA-600/7-76-034a)

     •    Volume II — Industry Priority Report (EPA-600/7-76-034b)

     •    Volume III — Iron and Steel Industry (EPA-600/7-76-034c)

     •    Volume IV — Petroleum Refining Industry (EPA-600/7-76-034d)

     •    Volume V — Pulp and Paper Industry (EPA-600/7-76-034e)

     •    Volume VI — Olefins Industry (EPA-600/7-76-034f)

     •    Volume VII — Ammonia Industry (EPA-600/7-76-034g)

     •    Volume VIII — Alumina/Aluminum Industry (EPA-600/7-76-043h)

     •    Volume IX — Textiles Industry (EPA-600/7-76-034i)

     •    Volume X — Cement Industry  (EPA-600/7-76-034j)

     •    Volume XI — Glass Industry  (EPA-600/7-76-034k)

     •    Volume XII — Chlor-Alkali Industry (EPA-600/7-76-0341)

     •    Volume XIII — Phosphorus/Phosphoric Acid Industry
                         (EPA-600/7-76-034m)

     •    Volume XIV — Copper Industry (EPA-600/7-76-034n)

     •    Volume XV — Fertilizer  Industry  (EPA-600/7-76-034o)

-------
     In the course of this study, some 80 industrial process options
were examined, focusing on:

     •    Identification of any major sources and amounts of pollutants
          (air, water and solid) expected from the processes,

     •    Development of estimated capital and operating costs for both
          production and pollution control aspects of the processes,

     •    Estimation of the types and amounts of energy used in both
          production and pollution control for the processes,

     •    Assessment of the economic viability and likelihood of im-
          plementation of those alternative process options being
          studied,

     •    Identification of areas where EPA's activities and policies
          could influence the future choice of alternatives, and

     •    Identification of research and development areas in both
          process and pollution control technology.

     Because of the industry orientation of the study (encompassing
15 volumes and some 1,700 pages), it was felt that pollutant-specific
information across all the 13 sectors studied should be summarized.
Five such pollutants were identified to be of particular interest:

     •    Nitrogen oxide (NO ) emissions,
                            X

     •    Sulfur oxide (SO ) emissions,
                          X

     •    Fine particulate emissions,

     •    Solid residues, and

     •    Organic, and/or toxic pollutants.

     A summary pollutant report in each of these areas has been pre-
pared.  Although we did attempt some estimates and extrapolations on
pollutants where information was readily available, in general, we did
not attempt to go beyond the contents of the 15 original reports.

APPROACH

     These summary pollutant reports are intended to be used closely
with the original 15 reports.  Generally, information, such as detailed
descriptions of the processes, has not been duplicated in these pol-
lutant reports.  Sections of the previous 15 reports in which this in-
formation can be found have been extensively referenced by volume num-
ber and page number (e.g., Volume VII, page 20, refers to page 20 of
the Ammonia Industry report).

-------
      TABLE 1.    PROJECTED U.S.  PRODUCTION IN INDUSTRIES STUDIED
Commodity
Alumina
Aluminum
Ammonia
Cement
Chlorine
Coke
Copper
Fertilizers (HNO )
Glass (flat)
Iron
Olefins (ethylene)
Petroleum
Pulp (kraft)
Pulp (newsprint)
Phosphoric Acid
(detergent grade)
Phosphoric Acid
(wet acid grade)
Steel
Textiles (knit)
Textiles (woven)
Total U.S.
production
in 1974
(106 tons*)
7.7
5.0
9.2
79.0
11.0
62.0
1.6
8.2
29.0
100.0
13.0
740.0**
16.0
3.9

1.4

9.0
133.0
0.32
2.1
Projected
rate
of growth
(%/yr)
6.0
6.0
6.0
2.0
5.0
2.5
3.5
4.0
2.5
2.5
8.0
1.5
5.0
2.5

2.5

2.5
2.5
2.2
2.2
Increase in
Total annual
projected production
production in 1989 over
in 1989 that of 1974
(106 tons) (106 tons)
18.5
12.0
22.0
106.3
22.9
89.8
2.7
14.8
42.0
144.8
41.2
925.0***
33.3
5.6

2.03

13.0
193.0
0.44
2.91
10.75
7.0
12.8
27.3
11.9
27.8
1.1
6.6
13.0
44.8
28.2
185.0****
17.3
1.7

0.63

4.0
60.0
0.12
0.81
   *A11 tons referred to in these reports are net tons, unless otherwise
    indicated.

  **Approximate equivalent of 30 quads (1 quad is equal to 10   Btu).

 ***Approximate equivalent of 37.5 quads.

****Approximate equivalent of 7.5 quads.

-------
     In Section 2 of this report (Findings and R&D Overview), summary
information on generic, cross-industry problems that emerge and suggest-
ions for research and development work in the areas of both pollution
control technology and process technology are presented.  In Section 3
of this report, availability and applicability of SO  pollution control
technology is presented, and SO  emissions and controls reported in our
previous study are summarized.  Unless otherwise noted, the SO  in gas
streams is largely S0? with only 1 to 3 percent of the sulfur values
reported as SO .   All emissions are estimated unless specifically ref-
erenced, since we believe that actual data do not exist for many of the
processes described, which are frequently still under development.

     To give the reader a sense of the size of the industries for which
the pollution problems covered in these summary pollutant reports are
considered, Table 1 lists these industries, their total production in
1974 (the baseline year for the study), and their projected incremental
production in 1989 — 15 years hence.

-------
                               SECTION 2

                      FINDINGS AND R&D OVERVIEW

FINDINGS

     Eight of the 13 industries addressed in the original study have
process-related SO  emissions that are affected by potential process
changes and are reviewed in this report.  These industries are petro-
leum refining, cement, olefins, ammonia, alumina/aluminum, pulp and
paper, glass, and copper.  In addition, process changes in the re-
maining five industries show changes for steam and/or electric energy
requirements and thus SO  emissions are affected at the generating
source.  Table 2-1 listsXthe estimated SO  emissions from processes,
steam, and power boilers for both the base case and alternative pro-
cesses and process alternatives for these industries.

     The SO  emission shown in the table include both process and power
boiler emissions.  Emissions from the power boiler were generally not
considered in the original study, but are included here because emis-
sions from the base case and alternative processes are compared in
this report.  The power boilers include on-site power boilers for steam
and electricity generation.  The emissions represent controlled emis-
sions from the process as further described in Section 3 where it is
seen that the emission factors vary from less than 1 Ib of SO /ton of
product to more than 1400 Ib of SO /ton of product as found in the
copper industry.  The emission factors from the conventional smelter
in the copper industry are high because of both high sulfur content in
the ore and an absence of economical control technology for reverbera-
tory furnace off-gas.  The emissions for the aluminum industry range
from an estimated 138 Ib of SO /ton to 200 Ib of SO /ton.  These high
emissions result mostly from power boilers, because the production of
aluminum requires large quantities of electricity which we assume will
be generated largely from coal to meet incremental aluminum production
requirements.

     As shown in Table 2, the petroleum industry was responsible for
the greatest volume of emissions of sulfur compounds from the base case
process in the year 1974, followed by copper, aluminum, kraft, pulp,
cement, and so forth.  Table 2 also shows estimated incremental SO
emissions (1989 - 1974) based on the increase in annual production from
1974 to 1989.  Incremental emissions are shown from both the base case
and alternative processes, assuming dedication of 100% of the incre-
mental production rather than a fraction to each process  (base case and
alternative).  Thus incremental SO  emissions (1989 - 1974) are cal-
culated by multiplying the emission factors (Ib SO  per ton of product
(e.g., copper) by the increase in production of product between 1974 to

-------
          TABLE 2.   SUMMARY OF ESTIMATED ANNUAL SOV EMISSIONS
Commodity Process
Controlled
SOX emissions
in 1974 from
base case process
(106 Ib/yr)
Controlled
SOX emissions
(1989-1974)**
(106 Ib/yr)
Change in SOX
emissions from
base case
(in 1989)***
(106 Ib/yr)
Petroleum Base case:
     East Coast refinery

  •  Direct combustion of
     asphalt in process
     heaters and boilers

  •  Flexicoking

Base case:
     Gulf Coast refinery

  •  On-site electric power
     by combustion of
     vacuum bottoms

Base case:
     West Coast refinery

  •  Hydrocracking of heavy
     bottoms

  •  High-purity hydrogen
     via partial oxidation
     of asphalt
3,030****
1,230****
2,580****
758****



840****

803****


308****



308****


545****


728****



653****
                                        +82****

                                        +45****
                                        +83****
                                        +8****

-------
                                         TABLE  2.    (continued)
Commodity
Cement





Olefins



Controlled
SO emissions
in 1974 from
Process base case process
(106 lb/vr)
Base case :
Long dry kiln 553
• Suspension preheater —
• Flash calciner
• Fluidized bed
• Coal as fuel instead
of gas in long dry kiln
Base Case :
Ethane- Propane Process 28.6
• Naphtha process —
• Gas-oil process
Controlled
SO emissions
(1989-1974)**
(106 Ib/yr)
191
183
183
134

516

62
102
361
Change in SO
emissions from
base case
(in 1989)***
(1()6 Ib/yr)
—
-8
-8
-57

+325

—
+40
+299
Ammonia
Base case :
     Ammonia via natural gas

  •  Ammonia via coal
     gasification

  •  Ammonia via heavy
     fuel oil
                                                 5.5
7.7
                                                                  30.7
                                                                  20.5
                     +23
                     +12.8

-------
                                              TABLE 2.    (continued)
00
Commodity
Alumina



Aluminum





Controlled
SOX emissions
in 1974 from
Process Base Case Process
(106 Ib/yr)
Base case:
Bayer process 25.4
• Hydrochloric acid
ion exchange —
• Nitric acid
ion exchange —
• Toth alumina —
Base case:
Hall-Heroult 1,045
(current practice, C.P.)
• Hall-Heroult \.new)
• Alcoa chloride —
• Refractory hard
metal cathode —
Base case:
Bayer with
Hall-Heroult (C.P.) 1,070
• Clay Chlorination (Toth
Controlled
SOX emissions
(1989-1974)**
(106 Ib/yr)
36
18
663
153
1,463
1,113
966
1,183
1,505
1,043*****
Change in SOX
emissions from
base case
(in 1989)***
(106 Ib/yr)
—
-18
+627
4-117
—
-350
-497
-280

-462 *****
                       alumina) and Alcoa
                       chloride

-------
                                         TABLE 2.    (continued)



Commodity







Newsprint
pulp



Process

Base case:
Kraft pulping
• Alkaline oxygen
pulping
• Rapson effluent-
free Kraft pulping

Base case:
Controlled
SOX emissions
in 1974 from
base case process
(106 Ib/yr)

766
—

—




Controlled
SOX emissions
(1989-1974)**
(106 Ib/yr)

829
59

844



Change in SOY
A
emissions from
base case
(in 1989)***
(106 lb/vr)

0
-770

+15



Flat glass
      Refiner mechanical       96
      pulp  (RMP)

   •  Thermo-uiechanical        —
      p'ulp  (TMP)

   •  De-inking of old
      news for newsprint
      iaanufacture

Base case:
      Regenerative furnace     27.3

   •  Coal gasification

   •  Direct coal firing
                     Coal-fired hot gas
                     generation
                                                                       42
                                                                       34
                                                                       12
12.2

13.1

43.4

16.1
                   -8
                  -30
  0

 +0.9

+31.2

 +3.9

-------
                    TABLE 2.    (continued)
Commodity
Flat glass
(cont. )

Copper




Chlorine ,
NaOH



Controlled
SOX emissions
in 1974 from
Process base case process
(106 Ib/yr)
• Electric melting
• Batch preheat with
natural gas firing
Base case:
Conventional smelting 2,250
• Outokumpu flash —
smelting
• Noranda
• Mitsubishi
• Arbiter
Base case :
Graphite-anode 454
diaphragm cell
• Dimensionally stable —
anodes
• Expandable DSA
• Polymer modified
Controlled
SOX emissions
(1989-1974)**
Q0b Ib/yr)
129
10.4
1,550
152
152
152
68
491
472
424
455
Change in SOX
emissions from
base case
(in 1989)***
(106 Ib/yr)
+117
-1.8
0
-1,398
-1,398
-1,398
-1,482
—
-19.0
-67.0
-36.0
asbestor

-------
                                        TABLE 2.   (continued)
Commodity
Chlorine ,
NaOH (cont.)


Steel


Controlled
SOX emissions
in 1974 from
Process base case process
(106 Ib/yr)
• Polymer membrane
• Ion exchange
membrane
• Mercury cell
Base case:
No off-gas recovery
• Off-gas recovery
—
"
—

16.4
—
Controlled
SOX emissions
(1989-1974)**
(106 Ib/yr)
455
447
557

12.0
6.0
Change in SOX
emissions from
base case
(in 1989)***
(106 Ib/yr)
-36.0
-44.0
+ 66

—
-6.0
Blast furnace
hot metal
Coke
Base case:
      Blast furnace

   •  Blast furnace with
      external desulfur-
      ization

Base case:
      Wet quenching of
      coke

   •  Dry quenching of
      coke
30
                                             N.A.
13.4

13.4
                        N.A.
                                                                     -30.6 ******
                                                                                            0.0
                                           -30.6******

-------
TABLE 2.   (continued)
Commodity
Steel
(integrated)

Phosphoric
acid
(detergent
grade)

Fertilizers
(nitric *cid)




Controlled
SOX emissions
in 1974 from
Process base case process
(106 Ib/yr)
Base case:
Steelmaking coke 189
oven, blast furnace
BOP route
• Direct reduction, —
EAF route
Base case:
Electric furnace 100
• Chemical cleanup of —
wet-process acid
• Solvent extraction —
of wet-process acid
Base case:
No NOX control -3.3
• Catalytic reduction
• Molecular sieve —
• Grand paroisse —
• CDL/Vitak
Controlled
SOX emissions
(1989-1974)**
(106 Ib/yr)
138
540
45
8.4
5.9
-2.6
-4.0
75.2
0.66
209.0
Change in SOX
emissions from
base case
(in 1989)***
(106 Ib/yr)
—
+402.0
-36.6
-39.1
—
-1.4
+77.8
+3.26
+212.0

-------
                                         TABLE 2.   (continued)
                                           Controlled                                 Change in SOX
                                          SO  emissions            Controlled        emissions from
                                          in 1974 from            SOX emissions         base case
 Commodity            Process           base case process          (1989-1974)**        (in 1989)***
	(106 Ib/yr)	(106 Ib/yr)	(1Q6 Ib/yr)

 Fertilizers       •  Masar                    —                    383.0               +386.0
 (cont.)

 Fertilizers (mixed):   Converting fertilizer dryers  (with baghouses) from natural gas to oil.

                Base case:
                      Natural gas             Nil                    Nil                  Nil

                   •  Better equipment         —                      0.29                 0.29
                      technique with
                      fuel  oil

                   •  Installing scrubbers     —                      0.22                 0.22

 Textiles       Knit fabric:
                   •  Base  case—               2.9                    1.1
                      conventional aqueous

                   •  Advanced aqueous         —                      1.0                 -0.1

                   •  Solvent processing       —                      0.5                 -0.6

                Woven  fabric:
                   •  Base  case                50.0                   19.3

                   •  Advanced aqueous         —                      8.8                -10.5

-------
                                           TABLE 2.    (continued)
     "Except where denoted by superscript ****.


    '""Based on incremental production from the year 1974 to 1989 derived from anticipated  growth  rates
      (Table 1).


   ^""Assumes no retirement of existing facilities by 1989.

    ""National emission rates calculated as though the alternative process applies  to  all  national
      oil refinery production.

 *** /\ >\ /\ &
 "    Bayer process, plus Hall-Heroult (new plant) orocess, is used for comparative  analysis.
!L JL J- JL JU *!.
      Change estimated from base case.

     N.A. not available.

-------
1989.  Total actual emissions from each process for the year 1989 would
be the sum of 1974 emissions from the base case process assuming no re-
tirement of existing facilities and the actual incremental emissions per
year from the base case and new processes installed between 1974 and 1989.

     The estimated increase or decrease in emissions from the alterna-
tive processes in relation to the base case process is also shown in
Table 2 for the year 1989.  This number shows the potential for emis-
sion reduction in 1989 using new process technology.  It is seen that
the greatest potential for emission reduction by new processes exists
in the copper industry, followed by the kraft pulp and aluminum indus-
tries.

     In the copper industry, alternative processes using smelters are
aimed at generating gas that has a sufficiently high concentration of
S0x> so that the SO  can be controlled by installing sulfuric acid plants.
Thus, these processes would have lower SO  emissions.
                                         X
     In the kraft pulp industry, compounds containing sulfur are not
used in alkaline oxygen pulping, which eliminates emissions that con-
tain sulfur compounds.

     The alternative processes in the aluminum industry are aimed at
reducing power consumption, which results in lower SO  emissions at
the power boilers.

     Some of the processes studied are aimed at changing the form value
of fuel from natural gas or oil to coal or asphalt.  These process
changes would generally result in increased SO  emissions because of
higher sulfur content in the fuel.  This is seen in the petroleum, ce-
ment, alumina, and glass industries.  The effect of changing from low-
sulfur feedstock to high-sulfur feedstock is seen in the olefin  indus-
try and results in increased SO  emissions.
                               x
     To give some perspective to the magnitude of SO  emissions,
Table 3 shows the industrial processes category to be the second larg-
est emitter of SO  after electric utilities.
                 x
          TABLE 3    SOx NATIONWIDE EMISSION ESTIMATES  (1974)

Emission source
Stationary fuel combustion
- Electric utilities
- Other
Industrial Processes
Transportation
Solid waste
Miscellaneous
TOTAL

106 Ib/yr

42,200
11,400
12,600
1,600
200
200
68,200

% of total

61.9
16.7
18.5
2.3
0.3
0.3
100.0

 Source:   National  Air  Quality  and  Emissions  Trend  Report,  1975,
          EPA,  Research Triangle Park,  N.C.;  NTIS PB-263  922

                                   15

-------
     Should U.S. industry expand using current (base case) technology,
Table 4 shows that estimated controlled SO  emissions in 1989 would
increase by 5.6 x 10  Ib compared to Sf>r emissions of 12.6 x 10  Ib
from industrial processes and 42.2 x 10 ^Ib from utilities in 1974
(Table 3).  However, if all U.S. industry expanded by implementing the
technologies considered here that emitted the largest amounts of SO ,
Table 4 shows that the increase in SO  emissions in 1989 would be 8.0
x 10  Ib, or some 43% higher than using conventional technology.  On
the other hand, if industry expanded by implementing the least SO -
emitting technology, SO  emissions in 1989 are calculated to increase
by 2.6 x 10  Ib or somex54% less than by conventional technology.  If
the national production growth rate projections shown in Table  1 are
accurate, then  calculated emissions will probably lie somewhere be-
tween these extremes, with energy  (Btu) saving processes with lower
SO  emissions somewhat balancing the switch from natural and fuel oil
to  coal.  Thus, incentives for  the implementation of energy conserving
technology can  have  a significant effect on future SO  emissions in
the industrial  sector.                               x

     Table 5 shows our estimate of the  types of processes likely to be
installed in the time period up to 1989 with related SO  emissions
from new plants calculated for  the year 1989 assuming no retirement
of  existing facilities.  For example, a reading of the industry reports
shows that in the copper sector incremental copper smelting capacity
will be effected by  one of the newer oxygen or flash smelting processes
(e.g., Outokumpu 4,  Noranda, Mitsubishi). If such capacitygis installed,
anticipated annual SO  emissions in 1989 would be 152 x 10 Ib SO /yr
(Table 5) compared toxthe 1,550 x 10  Ib S0-/yr, if conventional rever-
beratory furnace technology were employed.  Similar judgments were made
in  other sectors to  arrive at total calculated annual emissions of
2.4 to 5.6 x 109 Ib  SO  emitted in 1989 from new plant capacity in-
stalled in the  period 1974 - 1989.

     Examination of  the last column in  Table 2 shows that greatest
reduction compared to the base case process in Ib of SO  emissions per
year can be achieved by selective implementation of new processes in -

     •    copper    (oxygen or flash processes)

     •    aluminum (Alcoa chloride, refractory hard metal cathodes)

     •    pulp  (alkaline-oxygen)

     Process changes in some of the industries shown in Table 2 may be
implemented because  of feed stock shortages (e.g., the manufacture of
olefins from low-sulfur naphtha rather  than higher sulfur gas oil),
and fuel switching  (e.g., use of coal in cement making).  In other
cases, processes may be developed for other reasons, such as develop-
ment of a domestic alumina industry based on indigenous kaolin  clays.
Such a process, based on using coal to  the extent possible, would
                                   16

-------
             TABLE  4.   ESTIMATED  INCREASE  IN  CONTROLLED
                           SOX  EMISSIONS
                             1989-1974
      ASSUMING INDUSTRY EXPANDS USING PROCESS TYPES  INDICATED
                           (106 lbS02/yr)

Commodity (vol tio)
Steel (III)
Petroleum (IV)
Kraft pulp (V)
Newsprint pulp (V)
Olefins (VI)
Ammonia (VII)
Alumina (VIII)
Aluminum (VIII)
Textiles-knit (IX)
Textiles-woven (IX)
Cement (X)
Flat glass (XI)
Chlorine, NaOH (XII)
Phosphoric acid (XIII)
Copper (XIV)
Fertilizers (HN03) (XV)
TOTAL
Base case
process
138
758**
829
42
62
8
36
1,463
1
19
191
12
491
45
1,550
-3****
5,642
Using process
with largest
potential S0x
emissions
540
840
844
42
361
31
663
1,463
1
19
516
129
557
45
1,550
383
7,984
Using process
with smallest
potential SOX
emissions
101***
758
59
12
62
8
18
966
< 1
9
134
12
424
6
68
-4
2,634
*Volume Number of Industry Report
                                 17

-------
                        TABLE 4 (continued)
  **Assumes East Coast Refinery model applied nationally

 ***Includes credit of 37 x 106 Ib S02 attributed to energy
    saved in dry quenching and BOP off gas collection.

****Credit for steam raised:  see coverage on fertilizers in
    Section 3.
                                18

-------
    TABLE 5.  ESTIMATED RANGE IN CONTROLLED SOX EMISSIONS IN 1989
              FOR NEW PROCESSES LIKELY TO BE IMPLEMENTED
Commodity
(vol no)*
     Likely types of
     processes to be
     implemented in
       new plants
 Calculated range
  in annual SOX
 emissions for new
plant capacity-1989
    (106Ib/vr)
Steel (III)

Petroleum  (IV)

Kraft pulp (V)

Newsprint
pulp (V)

Olefins (VI)

Ammonia (VII)

Alumina**
(VIII)

Aluminum**
(VIII)

Textiles,
knit(IX)

Textiles,
woven (IX)

Cement  (X)

Flat glass (XI)
Chlor-alkali
(XII)
Coke oven, blast furnace, BOP****,***    132-138

Hydrocracking, flexicoking, etc.         308-803*****

Kraft, Rapson, alkaline-oxygen            59-844
RMP, TMP, De-inking

Naphtha, gas oil

Heavy fuel oil, coal


Bayer, leaching domes tic  clays


Hall-Heroult, aluminum chloride


Advanced aqueous, solvent


Advanced aqueous

Preheaters, coal firing, etc.

Regenerative furnaces, preheaters,
electric furnaces
Dimensionally stable anodes, new
membranes
Phosphoric acid!
detergent grade
(XIII)
      12-42

     102-361

      21-31


      36-663


     966-1050


       1-1


       9-9

     134-516

      12-129



     424-472
Wet acid cleanup
       8-8
                                   19

-------
                          TABLE 5  (continued)
Commodity
(vol no)*
Likely types of
processes to be
implemented in
  new plants
 Calculated range
  in annual SOX
 emissions for new
plant capacity-1989
    (106 Ib/yr)
Copper (XIV)     Oxygen or flash processes***          152-152

Fertilizers-
Nitric Acid (XV) Various NO  control technologies      (4)-383
                           X
       TOTAL                                       2,372-5,602


*Volume Number of Industry Report

   **A significant fraction of the incremental U.S. demand is
     expected to be imported.

  ***In addition, electric furnaces are expected to be installed
     based at least partially on scrap.

 ****With collection of CO from BOF's.

*****Base Case Gulf Coast Refinery model applied nationally.
                                  20

-------
result in significantly higher SO  emissions than a Bayer plant based
on natural gas, as described in tne Alumina/Aluminum Industry report.
Although the estimated change in emissions listed in Table 2 was based
on incremental capacity from 1974 to 1989 only, in some cases an alter-
native process, or a process modification, may replace existing capac-
ity.   For example, in the aluminum industry, refractory hard metal cat-
hodes may be installed in existing Hall Heroult cells.  The application
of alternative processes to existing plant capacity will increase the
potential effect on SO  emissions, compared to the estimated effect
shown in Table 2.  Further perspectives in each of these industry sec-
tors with descriptions of the processes can be obtained from the indi-
vidual industry sector reports (Volumes III through XV).

R&D AREAS

     The following areas of potential R&D for industry, government, or
other institutions have been identified with regard to SO  emissions
from the new technologies investigated:
SOYRelated
          Better definition of the environmental, health, and ecologi-
          cal impacts of SO  emissions with respect to obtaining more
          quantitative knowledge for establishing appropriate emission
          regulations.

          In going from gas to oil to coal, increased SO  emissions may
          be expected.  Therefore, it would be desirable to promote
          coal-cleaning methods to remove SO  , coal gasification with
          SO  control, and flue gas desulfurlzation techniques which
          would help to minimize the dispersal of SO  and mitigate the
          SO  emission problems.
            x
          Promote process alternatives aimed  at reducing energy re-
          quirements.  For a fuel containing  a fixed amount of sulfur,
          SO  emissions are in direct proportion to the energy con-
          sumption and, therefore, reduction  in energy consumption
          will result in lower SO  emissions.  Examples of such poten-
          tial energy-conserving processes include:

               a)  Rapson, alkaline/oxygen and de-inking technologies
                   in the pulp sector;

               b)  advanced aqueous or solvent processes in textiles;

               c)  CO collection from basic oxygen furnaces;

               d)  aluminum production by refractory metal cathodes
                   or by the aluminum chloride  (Alcoa) electrolysis
                   route;
                                   21

-------
               e)  phosphoric acid (wet acid grade) by the strong
                   acid process;

               f)  detergent grade phosphoric acid by the chemical
                   clean-up of wet acid; and

               g)  flotation of copper values from copper smelting
                   slags.

In addition, similar energy conservation and SO  reduction are achieved
by the generic technologies of solid preheating (e.g., in glassmaking,
or flash calcining in cement) or use of oxygen (e.g., flash smelting
in copper, or the Mitsubishi and Noranda oxygen-based processes).

Process-Related
Cement—
          Develop and implement a comparative test program at a number
          of cement plants with clinkering facilities employing long-
          rotary-kiln, suspension preheater, or flash-calciner process-
          es in which coal is burned as the fuel.   Coal of various sul-
          fur levels should be tested to determine the effect on oper-
          ation of the level and nature of sulfur in gas, dust, and
          clinker.  The benefits which derive from the physical and/
          or chemical cleaning of coal to reduce pyritic sulfur levels
          in coal for cement manufacturing should also be quantified.

          Develop and implement a commercial-scale test program on one
          or more flash-calciner-equipped rotary-kiln, cement-making
          facilities to characterize SO  emissions.  Of particular
          interest would be the emissions from operating with a bypass
          of a considerable amount of the combustion gases to eliminate
          alkalies.
Olefins—
          The olefins industry can benefit from additional research
          on the removal of sulfur from the cracked gas stream.   This
          stream contains hydrogen sulfide, some carbonyl sulfide, and
          varying percentages of diolefins and other reactive compounds
          which tend to foul the acid-gas-removal system.  This  problem
          is now being handled by depropanizing the cracked-gas  stream
          before acid gas is removed by scrubbing with diethanolamine.
          A method for removing the sulfur compounds and acid gases
          from the cracked-gas stream in the presence of diolefins
          (i.e., before the depropanizer) would be of significant eco-
          nomic benefit to the olefin producers.

          Naphtha and atmospheric gas oil feedstocks produce signifi-
          cant quantities of byproduct pyrolysis fuel oil.  If the
                                  22

-------
          feedstock material to the olefins plant has a sulfur content
          above a certain concentration,  the byproduct pyrolysis fuel
          oil has sulfur levels too high  for its environmentally accept-
          able use as a fuel without flue gas desulfurization.  These
          byproduct fuel oils also contain substantial amounts of un-
          saturates as well as other reactive materials which tend to
          polymerize and form gums on handling.   These present problems
          when attempting to desulfurize  the oils.  It would be desir-
          able to develop an economically attractive process for desul-
          furizing the pyrolysis fuel oil to a level where it would be
          environmentally acceptable as a fuel.   At present, most ole-
          fin producers limit the sulfur  content of their feedstock to
          circumvent this problem.  However, this limitation severely
          restricts their choice of feedstocks.
Aluminum—
          Consider materials research in the field of producing tita-
          nium diboride cathodes suitable in quality to permit long
          operating life in the Hall Heroult cell.  This development
          would have an effect on energy savings in the existing al-
          uminum plants, thus reducing SO  emissions at the power
          boiler.                        x
Pulp and Paper—
Glass—
          The alkaline oxygen process would alleviate the SO  pollution
          problems presently associated with the major alternative man-
          ufacturing method—namely, kraft pulping and bleaching.  Eval'-
          uation of this process and process emissions is desirable.

          The de-inking of old news for the manufacture of newsprint
          presents an opportunity to save energy and reduce SO emissions
          accompanying the pulping of an equivalent amount of virgin
          feedstock.  Broader commercial application should be support-
          ed, because it also could reduce the amount of municipal sol-
          id waste.
          A proven system to control glass furnace emissions of SO  and
          particulates must be developed if cost-effective pollution
          control is to be obtained.  Such a system is required if any
          of the coal-related processes is to be utilized.  While sev-
          eral systems have been developed, such as a coated fabric
          filter built by Teller Environmental Systems,  (Worcester,
          Mass.), none has been generally accepted by the glass indus-
          try  as a proven cost-effective technology.
                                   23

-------
Copper—
          Methods ehould be developed for removing impurities (e.g.,  Bi)
          from blister copper via modified fire-refining procedures.
          If impurities can be removed from copper, one-step smelting
          can be used.  This would significantly decrease fugitive SO
          emissions from smelters.
                                 24

-------
                              SECTION 3

                     PROCESSES AND SO  EMISSIONS
                                     X

BASES OF CALCULATIONS

     In Volume II (P19) describing the methodology used in this study,
we indicated that selected State air emission regulations along with
the Federal Government's stationary source performance standards and
effluent limitation guidelines were surveyed to:

     •  establish the most probable limits of air emissions, and

     •  obtain a perspective of the types of pollution control systems
        to be considered.

While there are a large number of different regulations for airborne
emissions at the State regulatory level, we found that approximately
the same type of air pollution control systems would be required,
regardless of the State or Federal regulations to be met.  Generally,
these air Pollution control systems included baghouses, venturi scrubbers,
inertial impact devices, and electrostatic precipitators for particulates
and chemical based systems for sulfur removal, such as alkaline-based
aqueous scrubbing for SO
                        X •

     In this Section, SO  emissions from the process industries
(Volumes III through XV) and specific control methods  are reviewed.
There are two general sources of emissions:  the processes themselves
and power boilers used for generation of steam and/or  electricity.
While emissions from the power boiler generally were not considered in
the original study, they are included in the present analysis to show
the net change in emissions resulting from a process change.  Both on-
site power boilers and electric utility power boilers  are included.
     It was found to be important to include power boilers in order to
     meaningful compar:
alternative processes.
make meaningful comparisons on SO- emissions between the base case and
     Power boilers are assumed to burn coal containing 3.5% sulfur, and
the uncontrolled SO  emissions are estimated at 5.54 Ib of SO /10° Btu
(12,000 Btu/lb of coal), or 0.058 Ib of SO /kWh (based on an average
requirement of 10,500 Btu to generate 1 kwft).  The uncontrolled  SO
emissions assume 95% of the sulfur in the coal appearing as SO   in the
flue gas largely as S0~ with only 1-3% as S0_.  The controlled emissions
from new plants are assumed to meet the new source performance standard
(NSPS) of 1.2 Ib of SO /106 Btu, equivalent to about 0.0126 Ib of SO  /kWh.
The SO  emissions from the base case process and process alternatives
      X
                                  25

-------
are described and summarized for each industry.  The SO  emission
factors are estimated, and the effect of these factors on the incre-
mental production from 1974 to 1989 is calculated.

SO  CONTROL METHODS
  x
     Control methods are limited to point sources only and fugitive
emissions are not discussed.  Since SO  control methods have been well
discussed in the literature, such methods are not described in detail
here but are only listed.

     Uncontrolled SO  emissions result from the sulfur present in the
fuel and in the feedstock which upon combustion results in SO  air
                                                             •v
emissions from high-temperature processes.  The control technology
for reducing emissions containing sulfur compounds are:

     1)   reduction of sulfur in the fuel and feedstock;

     2)   process modifications to prevent escape of sulfur compounds
          to the atmosphere; and

     3)   removal of sulfur compounds in the tail gas before it is
          exhausted to the atmosphere.

Other methods include use of tall stacks and controlling the sulfur
compounds concentration in the vicinity of the source.  These methods
do not reduce the emissions of sulfur compounds to the atmosphere, but
are aimed at dispersing the emissions to the atmosphere and at using
production curtailment to control ground-level pollutant concentrations.
The production curtailment method, used when adverse weather conditions
prevail, has been referred to as a "supplementary control system (SCS),"
when it is based on the monitoring of sulfur compound concentration at
ground level at various sites in the areas surrounding the source,
knowledge of which is used to control the production rate.

Reduction of Sulfur Compounds in Feed Materials and Fuel

     Sulfur emissions from a process may be reduced by using feed mat-
erials and fuel having lower sulfur or content, or by reducing sulfur
content in such materials before they are used.  The first alternative
is not available in all cases, because of the limited supply of mater-
ials.  Reducing the sulfur content of feed materials and fuel is widely
practiced for oil and, to limited extent, for coal.

     The applicability of desulfurization of oil in the petroleum in-
dustry is determined principally by the characteristics of the oils to
be processed and the end-product sulfur specification.  Sulfur removal
of 90+ % is technically feasible for most oil stocks.  The important
feedstock properties with respect to influence on the desulfurization
of oil are:  organometallic compounds (principally nickel and vanadium)
                                  26

-------
and the asphaltene content of the feedstock.  Lower organometallic com-
pounds and asphaltenes are desirable for desulfurization.

     The desulfurization of coal is practiced to a lesser extent.  Sul-
fur removal has been in the range of 10-50%.  Part of the fuel value of
the coal is lost in the process and the loss increases with increased
sulfur removal.  However, the heating value per ton of cleaned coal may
actually rise, especially if some inert materials are simultaneously
removed.

Process Modifications to Prevent Escape of Sulfur Compounds in Air
Emissions

     This method of emission reduction has been applied, to a limited
extent, to coal combustion.  An alkali such as limestone is added with
the coal to the boiler.  The limestone reacts with sulfur to form cal-
cium-sulfur salts, which are removed along with the bottom ash.  Lime-
stone may also be used to remove sulfur compounds in the generation of
gas from coal.  A detailed description of this process is given in the
discussion of the glass industry found later in this Section.

     In some cases, alkali materials, present in the process, help to
reduce SO  emissions.  For example, in cement kilns only a fraction of
the sulfur present in the feedstock and fuel appears in the flue gas,
since the remainder is removed by alkaline particles in the kiln feed.

Removal of Sulfur Compounds from Process Gaseous Emissions

     The sulfur compounds in tail gas are present as reduced sulfur
compounds or as sulfur oxides.  Reduced sulfur compounds, such as hy-
drogen sulfide, usually have associated odors and, therefore, their e-
missions are more obnoxious than those of the sulfur oxides.  The SO
                                                                    X
in the exhaust gas is generally present as S0_, except in some cases
(such as in glass manufacture) where sulfur emissions are generated
from sulfates  in the raw materials and result in SO  emissions.  SO
also forms because of oxidation of SO  and, similarily,  S0« may  form
because of reduction of  SO  .

     The SO  emissions are difficult to control because  the SO   is gen-
erally present in the form of  submicronic mist at the operating  temper-
ature of the pollution control equipment.  As a result,  a very efficient
de-mister is required for control.

     In the pulp and paper mills, the sulfur compounds are present in
reduced form in combination with hydrogen or organic compounds.  The
control method used involves incineration of the sulfur  compounds to
convert to sulfur oxides often followed by a pollution control device
used to remove the SO  .
                     x
     In the petroleum industry, the sulfur compounds are present as COS,
H_S, etc.  The gases are removed in an acid gas treatment system for


                                   27

-------
process requirments.  The acid gases are removed by amine absorption or
caustic scrubbing.  The concentrated gases from the regeneration step
of systems, such as amine absorption, are treated in the Glaus plants,
followed by a tail gas cleanup system, such as the Stretford process.
In this way, the sulfur compounds are converted to elemental sulfur.
If the volume of the gas stream is small, the Stretford process may be
used directly to remove sulfur compounds after the acid gas has been
removed.

     There are several methods available for removing SO  from flue gas.
At low concentration (less than 1-2% SO ) and at low gas-flow rates,
alkaline scrubbing may be used to control SO  emissions.  At low con-
centrations (less than 0.5% SO.) and at high gas-flow rates, such as
from power boilers, flue gas desulfurization methods (FGD) are used.
Depending on the product stream, these methods are classified as fol-
lows :

     •    Waste salt process,

     •    Concentrated SO  process,

     •    Direct sulfuric acid process, and

     •    Elemental sulfur process.

     For processes that generate concentrated emissions (more than 4-6%
     and large-volume gas streams (such as copper
acid plants are used to control the SO. emissions,
S0?) and large-volume gas streams (such as copper smelting), sulfuric
     The SO  emissions may also be removed by dry processes, as in bag
filters, if a coat of alkaline material is present.  In cement kilns,
bag filters are used to control particulate emissions and, because of
the alkaline nature of the particles, a significant fraction of the SO
is removed.  In some systems, such as in controlling the emissions from
glass furnaces, alkaline material is introduced upstream of the bag
filters to control SO  emissions.
                     x

PETROLEUM REFINING

Base Case Process - 1985 Refinery

     A description of the petroleum refining industry is presented in
Volume IV, page 9.  In assessing the impact of process changes in the
refining industry, a 1985 refinery configuration was used as a base
case.   The reasons for using this approach are given in Volume IV,
page 22.  Three configurations were selected to represent the local con-
ditions:  the East, Gulf, and West Coasts.  Results of our analyses are
given in Table 6.
                                   28

-------
                TABLE 6.   ESTIMATED SOX EMISSION FACTORS - PETROLEUM REFINING INDUSTRY


                           Emission factor* - no control  Emission factor* - with  control
                           (lb/10y Btu of refinery output) (lb/109 Btu"of refinery output)    Control
    Process                Process  Power boiler  Total   Process Power boiler  Total          0/iency
    ^_	&	
Base case:
     East Coast refinery      84        80         164      84        17         101         38

  •  Direct combustion of    360        87         447      93        19         112         75
     asphalt in process
     heaters and boilers

  •  Flexicoking             364        80         444      90        17         107         76

Base case:
     Gulf Coast refinery      31        46          77      31        10         41          47

  •  Onsite electric power    31        23          54      31        10         41          24
     by combustion of
     asphalt

Base case:
     West Coast refinery      74        55         129      74        12         86          33

  •  Hydrocracking of        167        63         230      83        14         97          58
     heavy bottoms

  •  High-purity hydrogen    119        59         178      74        13         87          51
     production via partial
     oxidation of asphalt
^Emissions reported as Ib SOo/lO  Btu of refinery output.

-------
 East  Coast  Refinery

 Base  Case Process—

      The base case East Coast  refinery is described in Volume IV,
 page  23.  The energy intake in this refinery (in terms of crude  oil,
 fuel  oil, steam, and electricity)  is 1.31 x 10   Btu/day and production
 is  equivalent to 1.19 x 101 Btu/day (Vol.  IV, page 35).

      The major airborne pollutants emitted by refineries have been pre-
 viously  identified in Volume IV, page 24.   The emission factors  sum-
 marized  in  EPA's "Compilation  of Air Emission Factors" (1973)  were used
 to  determine  SO emissions  from the base  case refinery.   The SO  emis-
 sions are:  45,200 Ib SO /day  from combustion sources  (heaters^oil-
 fired boilers); 52,800 1§ SO /day  from fluid catalytic crackers; and
 2,200 Ib  SO /day from the Glaus plant.  Uncontrolled emissions of
 100,280  Ib  §0 /day are equivalent  to an emission factor of  84  Ib SO /10^
 Btu of refinery output.   The emissions from the combustion  sources and
 fluid catalvtic crackers are not controlled.   Sulfur emissions from the
 Glaus plant"  are based on 99.5% recovery  of the sulfur in the  acid gas.

      Purchased  steam and electricity in the base case  refinery are e-
 quivalent to  8.2 x 109 Btu/day  and  8.9  x  10  Btu/day,  respectively.  The
 estimated emissions  from the associated coal-fired  power  boiler are
 94,700 Ib S0x/day,  based on no  control, and 20,500  Ib  SO  /day, assuming
 control  to  meet the  New Source  Performance  Standard  (NSP*-).  The esti-
 mated emissions are  equivalent  to  79.8  Ib  SOX/10^ Btu  of  refinery  output
 based on  no control,  and 17.3 Ib SOX/109 Btu  of  refinery  output,
 based on  control to  meet the NSPS  resulting in total emissions of
 101 Ib SO /109  Btu,  as shown in Table  6.
          X
 Process Option  1 - Direct Combustion of Asphalt  in  Process  Heaters and
 Boilers—

      A detailed description of  this  alternative  appears  in  Volume  IV,
 page  32.   Utilizating  asphalt for combustion  is  intended  primarily to
 upgrade the overall  form value  of refinery  products rather  than to act-
 ually  increase  the overall  thermal  efficiency  within the  refinery.   Part
 of  the refinery gas  (36.6%)  and all  of  the  fuel  oil are displaced
 by  asphalt.    The energy  balances for  the base  case process  and the  al-
 ternative are summarized  in Volume  IV,  page  35.   In the alternative
 process,   SO   emissions  are  increased because of  the increased  sulfur in
 the asphalt.

     The  emissions from  the  asphalt combustion are 15,395 Ib SO /hr
 (Volume IV,  page 38).  The  emissions from the  combustion  sources in
the base  case process are eliminated in this alternative, because  oil
is not used  for combustion.   Thus,  the  total refinery emissions are
425,000 Ib SO /day, equivalent  to about 360 Ib SO /109 Btu of  refinery
output.      X                                   x
*For discussion, see Vol. IV, p 152.

                                    30

-------
     The emissions from the asphalt combustion may be controlled by an
add-on FGD system.  If the SO  removal efficiency is 85% (Volume IV,
page 39), the emissions from the combustion sources will be reduced to
an estimated value of 55,400 Ib SO /day.  The total emissions are est-
imated at 110,400 Ib SO /day, equivalent to 93 Ib SO /10  Btu of refin-
ery output, as shown in Table 6.
                                                 9                     9
     Purchased steam and electricity are 9.7 x 10  Btu/day and 8.9 x 10
Btu/day, respectively (Volume IV, page 35).  Estimated emissions from
the power boiler are 87 Ib SO /10  Btu of refinery output, based on no
control, and 19 Ib SO /10  Btu of refinery output, based on control to
meet the NSPS.  Thus, combining these figures results in about an 11%
increase in SO  emissions compared with the base case, as shown in Ta-
ble 6 and 7 .

Process Option 2 - Flexicoking —

     Flexicoking is the combination of fluid coking with coke gasifi-
cation.  Although fluid coking is a commercially available technology,
there are no commercially operating flexicokers.  A detailed process
description is given in Volume IV, page 52, with emissions discussed on
page 59.

     The major air pollution problem associated with the Flexicoking
process is in controlling sulfur in the streams of fuel gas  (light
hydrocarbons containing l^S) and flexigas  (a low-Btu fuel gas contain-
ing N2» CO, C02» H2, and sulfur, and its compounds).

     The sulfur in the fuel  gas is removed using an amine scrubbing
system, and the exhaust of that scrubbing  system is sent  to  the refin-
ery Glaus  plant.  The hydrogen sulfide  in  the low-Btu flexigas  is  too
low in  concentration to be economically scrubbed out, and, therefore,
this process comes with an integral Stretford unit  for sulfur  removal.
The sulfur concentration in  the product flexigas is approximately
170 ppm, which is within allowable standards for combustion  without
sulfur  control.

     The emissions having sulfur compounds are  2,920  Ib SO /day from
the combustion of flexigas and  159,580  Ib  sulfur /day, in  tfie acid  gas
stream.  The  total emissions from  the refinery  are  equivalent  to  364 Ib
SO /10   Btu of refinery output.  If the sulfur  emissions  in  the acid
gas stream are controlled, the  total emissions  from the refinery  are
 reduced  to  an  estimated  value  of  90  Ib  SO /10   Btu of refinery output,
 as  shown in Table  6.
 *Some  additional  background  on the Stretford process is found in
  Volume XI,  p 38, "Environmental Considerations of Selected Energy
  Conserving  Manufacturing Process  Options:   Petroleum Refining Industry
  Report"
                                    31

-------
                         TABLE 7.    ESTIMATED SOV EMISSIONS - PETROLEUM REFINING INDUSTRY
u>

Emission factor
SOV Emissions
(Ib SOX/109 stu of refinery output) Change in (106 Ib/yr)
Process Process Power boiler Total emission factor 1974 1989-1974*
Base case:
East Coast refinery 84 17 101
• Direct combustion of 93 19 112 +11
asphalt in process
heaters and boilers
• Flexicoking 90 17 107 +6
Base case:
Gulf Coast refinery** 31 10 41
• Onsite electric power 31 10 41 0
by combustion of
vacuum bottoms
Base case:
West Coast refinery** 74 12 86
• Hydrocracking of 83 14 97 +11
3,030** 758
840
803
1,230** 308
308
2,580** 645
728
   heavy bottoms

•  High purity hydrogen    74
   production Via partial
   oxidation of asphalt
                                                13
87
+1
653

-------
                                         TABLE 7 (continued)
 *Based on incremental national production from 1974 to 1989 equal to 7.5 quads (7.5 x 10  Btu)
  multiplied by emission factor asssuming no retirement of existing facilities.

**Estimated 1974 emissions, based on the total emission  factors and production of 30 quads are
  3,030 X 106 ibs, 1,230 x 106 Ibs and 2,580 x 106 Ibs for East Coast, Gulf Coast and West
  Coast models, respectively.

-------
     The emissions from the power boiler are comparable to those in the
base case process, i.e., about 80 Ib SO /10  Btu of refinery output,
based on no control, and 17 Ib SO /IO  §tu of refinery output, based on
control to meet the NSPS.  Total SO  emissions thus increase by about
6% as shown in Tables 6 and 7.

Gulf Coast Refinery

Base Case Process—

     The base case refinery configuration in the year 1985, located on
the Gulf Coast is described in Volume IV, page 23.  Total energy input
in the base case refinery is 1,358 x 10  Btu/day and production is
equivalent to 1,197 x 10  Btu/day.  SO  emissions from this base case
process were developed in a manner similar to that for the East Coast
refinery using oil-fired boilers.

     The emissions include 27,600 Ib SO /day from combustion sources;
8,400 Ib SO /day from fluid catalytic crackers and 680 Ib SO /day from
the Glaus pfant.  The total emissions of 36,680 Ib SO /day (Volume IV,
page 25) are equivalent to 31 Ib SO /IO  Btu of refinery output.  Sul-
fur emissions from the Claus plant are based on 99.5% recovery of the
sulfur in the acid gas.
                                                            Q
     Based on electricity consumption equivalent to 9.2 x 10  Btu/day
and steam consumption equivalent to 0.7 x 10" Btu/day (Volume IV,
page 70), the estimated emissions from the power boiler are 46 Ib SO /
10" Btu, based on no control, and 10 Ib SO /IO" Btu, based on control!
to meet the NSPS, resulting in total discharges of 41 Ib SO /IO9 Btu.

Process Option - On-site Electric Power by Combustion of Vacuum Bottoms—
     In this process alternative, electric power is generated within
the refinery rather than purchased from a local electric utility.  The
fuel is assumed to be asphalt.  A simplified flow sheet for this alter-
native is shown in Volume IV, page 69.

     Generation of electric power within the refinery neither conserves
energy nor consumes more energy than when power is purchased, assuming
that the internal and external power plants would operate at the same
efficiencies.  In effect, the form value of the asphalt is upgraded to
a higher form value of electric power for refinery use.  The design
capacity of the generator, as shown in Volume IV, page 77, is 36.5 me-
gawatts.

     This process alternative will result in increased on-site SOX
emissions.  SO  emissions from the on-site power boiler are 27,456 Ib
SO /day (Volume IV, page 73), equivalent to 23 Ib SO /IO9 Btu of refin-
ery products.  The costs of sulfur control for on-site electric power
generation is shown in Volume IV, page 76.  If the emissions are


                                 34

-------
controlled to meet the NSPS, the resulting SO  emissions will be about
10 Ib SO /109 Btu of refinery output.        X
        X
West Coast Refinery

Base Case Process—

      The base case refinery configuration in the year 1985, located on
the West Coast, is described in Volume IV, page 23.  Total energy input
in the base case refinery is 1,009 x 10  Btu/day and production is
equivalent to 950 x 10^ Btu/day.  SOX emissions from the base case pro-
cess were developed in a manner similar to that for the East Coast
refinery using oil-fired boilers.

     The emissions include 58,800 Ib SO /day from combustion sources;
7,600 Ib SO /day from fluid catalytic crackers; and 3,680 Ib S0x/day
from the Claus plant.  The total emissions of 70,080 Ib SO /day (Volume
IV, page 26) are equivalent to 74 Ib SO /109 Btu of refinery output.
Sulfur emissions from the Claus plant are based on 99.5% recovery of
the sulfur in the acid gas.
                                                            g
     Based on electricity consumption equivalent to 9.5 x 10  Btu/day
(Volume IV, gage 46), the estimated  emissions from the power boiler are
55 Ib SO /109 Btu of refinery output, based on no control, and  12 Ib
SO /109 8tu based on control, to meet the NSPS, resulting in total dis-
charges of 86 Ib SO /109 Btu.

Process Option 1 - Hydrocracking of  Heavy Bottoms—

     In this option, asphalt is used as a feedstock for a hydrocracking
process such as H-Oil or Isomax in which  the heavy bottoms are  convert-
ed to lighter  fuel oils and gaseous  products.  The H-Oil process was
chosen to exemplify a heavy-ends hydrocracking process.  A detailed des-
cription of  the process is  given in  Volume  IV, page 43.  The pollutants
of concern in  this option are t^S, generated  in  the gaseous  products,
and  S02> generated from the combustion of a  small  amount of  asphalt.
The  gases from the H-Oil reactor will pass  through an  amine  scrubbing
system to remove H£S and C02-   These are  sent  to a Claus plant  for  sul-
fur  control.   Since  a  refinery  normally would have a  Claus  plant,
the  cost of  control  for this option  is only  the  incremental  cost  re-
quired to expand  the Claus  plant to  handle  the additional sulfur  load
(Volume  IV,  page  152).

      In  addition  to  the sulfur  control costs  for hydrotreating, the
equivalent of  900  BPD  of asphalt also has to  be  burned in the  process
heaters  to  free  up  refinery gas which can be  used  to  generate  the  ad-
ditional hydrogen  needed for  the hydrotreater.   Since  the asphalt  has
a sulfur content  higher than  the allowed  maximum uncontrolled  sulfur
content  of a fuel,  flue gas desulfurization will be required.   If  all
of  the  asphalt were  burned  in a single process heater, the  size would
be  approximately  240  x 106  Btu/hr.
                                   35

-------
     The incremental uncontrolled emissions include 715 Ib SO., /day from
purge gas, 54,530 Ib S/day from acid gas stream, and 12,748 Ib SOX per day
from asphalt combustion (Volume IV, page 48).  Total emissions fiom the
refinery are about 192,600 Ib SO /day, equivalent to 167 Ib SO /109 Btu
of refinery output.

     If the sulfur emissions from the tail gas are controlled by the
Glaus plant and the emissions from the asphalt combustion are controlled
by the wet scrubber, the incremental emissions will be reduced to an
estimated value of 8,170 Ib SO /day (to meet EPA standards).  Therefore,
the total estimated controlled emissions from the refinery are about
83 Ib SO /l(r Btu of refinery output.
        X
     Based on electricity consumption of 10.8 x 10  Btu/day (Volume IV,
page 46), the estimated emissions from the power boiler are 63 Ib SO /Q
10  Btu of refinery output if no control is provided, and 14 Ib SO /$0
Btu of refinery output if the emissions are controlled to meet the^SPS.
The total SO  emissions will then be 83 plus 14 or 97 lb/10  Btu of out-
            v
put; an increase of 13%.

Process Option 2 - High-Purity Hydrogen Production via Partial
Oxidation of Asphalt—

     This alternative is based on the production of high-purity hydro-
gen for hydrotreating from vacuum bottoms, using a partial oxidation
process.  The feedstock freed up by this approach would then be avail-
able for sale outside the refinery in the form of pipeline gas or
naphtha.  A detailed process description is given in Volume IV, page 78.

     The major emission associated with this process change is the sul-
fur removed from the raw syngas in the form of I^S.  This is removed
from the gas using an amine scrubber system.  The exhaust from the amine
regenerator, which contains the sulfur as well as C02» is sent to a
Glaus plant for sulfur recovery and air pollution control.  Several
states have emission standards regulating tailgas for sulfur recovery
plants so that tailgas cleanup will be required to limit emissions from
the plant to 250 ppm or less.  The incremental uncontrolled emissions
are 21,200 Ib of S/hr, equivalent to 45 Ib of SO /109 Btu of refinery
output.  If the emissions are controlled, the additional emissions are
estimated to be equivalent to 0.2 Ib SOX/109 Btu of refinery output.

      The  emissions  from the power boiler (coal-fired)  are estimated at
 59 Ib  SO  /10   Btu of  refinery  output,  based on no  control,  and 13 Ib
 SO /10y Btu of refinery output,  based  on control  to  meet  the  NSPS.   Ta-
 bles  6 and  7  show that  SO   emissions  are only  1% higher  than  those  for
 the base  case  process.

 Summary

     Table  6  shows  that  uncontrolled  emissions  are increased  in  all
 alternatives,  except  for the case of  on-site generation of  electric

                                   36

-------
power by combustion of asphalt.   The sulfur content in the asphalt is
lower than that in the coal burned in the power boiler.

     The emission control methods include removal of reduced sulfur in
acid gas cleaning followed by the Glaus and Stratford plants, and flue
gas desulfurization.  As shown in Table 6, overall control efficiency
varies from 24 to 76 percent.  Also, the controlled emissions in the
alternative processes are comparable or higher than those in the cor-
responding base case process.  Based on the assumption that each proc-
ess option is considered individually for application in 100% of refin-
ery expansion, the controlled SO  emissions from refinery industry are
summarized in Table 7.  Examination of Table 2 shows that even small
percentage changes in the petroleum industry can have significant nat-
ional impact.

CEMENT

Base Case Process - Long Rotary Kiln

     The production of finished cement from raw materials involves four
steps:  crushing, grinding, clinkering, and finish grinding.  From a
materials viewpoint hydraulic cement is a powder made by heating lime,
silica, alumina, iron oxide, and magnesia together in a kiln and then
pulverizing the product.  The base case process selected for the cement
industry was the long rotary kiln with dry grinding.  In the dry cement
process, grinding is performed dry, very much like wet grinding, except
that no water is added and the material is ground at 1% moisture con-
tent or less.  A detailed description of  the process is given in Volume
X, page 84.

     The base case  process is based on natural gas or oil as fuel.  How-
ever, the emissions are estimated using natural gas as a fuel in the
cement kiln since more natural gas  than oil was used in 1974 in  the
cement industry  (Volume X, page 92).  Very little sulfur is  present  in
the natural gas, so all the  emissions result from sulfur present in  the
raw materials.  The uncontrolled emission factor is 10.2 Ib  SO /ton  of
cement as reported  in "Compilation  of Air Emission Factors"  (EPA,  1973).
In addition, gaseous emissions from the combustion of high  sulfur  fuel
in the kiln are usually not  sufficient  to create significant air pol-
lution problems.  Most of  the sulfur dioxide formed from the sulfur  in
the fuel  is recovered as  it  combines with the  alkalies and  also  with
the lime when  the alkali  fume is low.

     Usually bag filters  are used to control particulate emissions from
the cement kilns  (Volume  X,  page 101).  With bag filters, approximately
50% of the SO   is removed  by passage through the alkaline particulate
filter cake.   Other control  methods include reduction of  sulfur  in the
feed and  fuel  to the  kiln.

      In addition to the process  emissions, electricity consumed  in the
                                   37

-------
plant will result in emissions at the power boiler.  The electricity
consumed in the long kiln is equivalent to 1.61 x 10" Btu/ton of cement
resulting in estimated uncontrolled emissions of 8.9 lb SO /ton of cem-
ent.  If the emissions are controlled to meet the NSPS, the emissions
will be reduced to an estimated value of 1.9 lb SO /ton of cement.  Con-
sequently, the total controlled SO  emissions willxbe 7.0 Ib/ton.
                                  X

Process Option 1 - Suspension Preheater

     The suspension preheater is a modification of, or addition to, the
cement rotary kiln.  It is attached to the raw feed inlet end of the
kiln, totally replacing the preheating zone.  The suspension preheater
preheats the raw material and also accomplishes a considerable amount
of raw material calcination.  Typical suspension preheaters heat cold
raw feed to approximately 1,400 F and accomplish 30-40% of the total
calcination or thermal decomposition of the calcium carbonate, the main
component of the raw feed.  Consequently, with the suspension preheater,
the rotary kiln receives hot and partially calcined raw material.  There
are several variations on the suspension preheater (described in Volume
X, page 19).

     Raw material which has been partially calcined is highly reactive
with sulfur dioxide and oxygen, forming calcium sulfite and sulfate.
Any SO  which might form in the combustion gases in a rotary kiln, even
using nigh-sulfur coal as fuel, contacts the raw material so intimately
that the use of a suspension preheater system should not present any
sulfur dioxide emission problems.   However, no quantitative data are
available, so emissions cannot be estimated.

     The electricity consumption is reduced to 1.32 x 10  Btu/ton of
cement, so the estimated emissions from the power boiler are 7.3 lb SO /
ton of cement, based on no control and 1.6 lb SO /ton of cement based x
on control to meet the NSPS.  Overall, some reduction on SO  emissions
can be expected.

Process Option 2 - Flash Calciner

     Although designs of flash calcining systems vary, the main feature
which characterizes the flash calciner rotary kiln is the flash cal-
cining vessel added between the rotary kiln and the suspension pre-
heater.  A detailed process description is given in Volume X, page 34,

     The flash calciner arrangement (described in Volume X, page 23)
requires a considerable amount of excess combustion air in burning the
fuel in the rotary kiln so that enough air is present in the combustion
gases leaving the kiln to permit combustion of the fuel in the flash
calcining vessel.

     By the time the combustion gases exit to the atmosphere, essential-
ly all of the SO  should be absorbed and reacted with the raw feed.
                X
                                  38

-------
However, any gases which might be bypassed could be different in SO
content from suspension preheater bypass gases because the raw feed
entering the flash calciner kiln is almost completely calcined, and
also has a significantly lower kiln residence than in a suspension pre-
heater.  However, no quantitative data are available, so the emissions
cannot be estimated but are thought to be lower than the long kiln.

     The electricity consumption is comparable to that in the suspen-
sion preheater alternative and therefore the estimated emissions from
the power boiler are 7.3 Ib SO /ton of cement based on no control, and
1.6 Ib SO /ton of cement basedxon control to meet the NSPS.
         x
Process Option 3 - Fluidized-Bed Cement Process

     The only difference between the fluidized-bed cement process and
the conventional process is the high-temperature clinkering step.  All
of the other steps are essentially identical.  A detailed description
of the process is given in Volume X, page 40.  Comparing the energy
consumption in the fluidized bed and in the long kiln  (Volume X, page
46) , one finds the fuel consumption in the fluidized bed to be about
10% higher.  No quantitative data regarding SC^ emissions are available
so emissions cannot be estimated but are thought to be lower than in
the long kiln.

     There  is a net recovery of energy in the  form of  electricity  in
the fluidized bed process  (0.15 x 10^ Btu/ton  of cement; Volume X, page
46) resulting in a credit  (i.e., negative emissions) at the power  boil-
er  estimated at 0.83 Ib SO /ton of cement, based on  no control; and
0.18 Ib  SO  /ton of cement.^Sased on control  to meet  the NSPS.
          X
Process  Option 4 - Conversion  to Coal Fuel from Natural Gas

     The process description  using coal as a  fuel  in the  cement kiln  is
described in Volume X,  page 60.  The  use  of  coal as  fuel  in  the cement
kiln will increase the  SO  emissions  because  there  is  more  sulfur  in
the coal than in  the gas.  The reported emission factor is 6.8 Ib  SOX/
ton of cement per  percent  sulfur in the coal  (EPA,  1973).   Therefore,
for a  3.5%  sulfur  coal, emissions  are estimated at  23.8 Ib  SO  /ton of
cement.  The  total emissions  from  the cement  kiln,  due to sulrur  in
both  the feed materials and the fuel, are 34  Ib SO  /ton of  cement.
Again, if bag  filters  are  used to  control particulate emissions,  the  SO
emissions are  reduced  by  50%  to 17 Ib/ton of  cement  (EPA,  1973).   The  X
emissions  from  the power  boiler are the  same as those from the base
case  process,  giving a total  controlled  SO  emission of 18.9  Ib/ton  of
cement,  a  significant  170% increase.

 Summary

      SO  control devices  are  generally  not used in the cement  industry.
However* bag filters used for particulate control can reduce SO
                                                                x
                                   39

-------
 emissions by  50%  because  of  the  alkaline  filter  cake.   The controlled
 emission factors  are  shown in  Tables  8  and  9  for the base case and al-
 ternative processes.

     Controlled emissions from the  use  of coal as fuel  (instead of nat-
 ural gas) in  the  long-dry kiln alternative  are highest  because of the
 higher  sulfur content of  the coal.  The sulfur in the feed in other
 processes is  lower.   Emissions from the fluidized bed should be lowest
 because of higher SO   removal  in the  fluidized bed.
                    x

 OLEFINS

 Ethylene Production

 Base Case Process - Ethylene Production Based on Ethane and Propane
 Cracking—

     The base case technology  selected  for  the assessment of the domes-
 tic olefin industry was ethane and  propane  (E-P)  cracking.  A detailed
 description of the process of  producing ethylene from ethane and pro-
 pane is given in  Volume VI,  page 17.

     As a basis for calculation,  a  feedstock containing 10 ppm sulfur
 is used.  During  the  ethylene  cracking  process,  sulfur  is produced in
 the form of I^S and COS.  These  gases and C02 (the acid gases) are re-
 moved from the compressed cracked gases by  a caustic (sodium hydroxide)
 scrubber (Volume  VI,  page 103).   The  liquid effluent from the caustic
 scrubber is contacted  with naphtha  to absorb entrained  hydrocarbons.
 The naphtha solution  is decanted and  then used as  a fuel.  The water
 effluent from the naphtha wash is neutralized with sulfuric acid, re-
 sulting in the following  effects:

     •    Sulfides, such  as  Na2S or NaHS, are replaced  by Na2SO^.  This
          sulfate can  be discharged from  the plant as dissolved solids
          in  water effluents unless local conditions prohibit;

     •    The  acid gases  are regenerated  and must be incinerated to
          convert H^S  to  SOo before venting to the atmosphere;

     •    In  the  base  case process, a feed  sulfur concentration of 10
          ppm in  E-P results in  an S02  exhaust of about 0.04 Ib/ton of
          ethylene (Volume VI, page 43).

     The only  other sulfur emission from an E-P  cracker occurs during
 the decoking operation.  The quantity of sulfur  emitted is very small.

     Some SO   emissions are expected from the fuel used in the power
boiler.   The electricity consumed is equivalent  to 900 Btu/lb of ethyl-
ene.   The estimated uncontrolled emissions are 9.98 Ib  SO /ton of ethyl-
ene and the emissions controlled to meet the NSPS are estimated at 2.13
                                   40

-------
                  TABLE 8.   ESTIMATED SO  EMISSION FACTORS - CEMENT  INDUSTRY
Emission factor - no control
(Ib/ton of cement)
Process Process Power boiler Total
Base case:
Long kiln (natural gas) 10.2
• Suspension preheater <10.2
8.9 19.1
7.3 <17.5
Emission factor - with control
(Ib/ton of cement)
Process Power Total
5.1
< 5.1
1.9 7.0
1.6 <6.7
Control
Efficiency
%
63
62
   (natural gas)

•  Flash calciner          <10.2       7.3
   (natural gas)

•  Fluidized bed          < 10.2      -0.8
   (natural gas)

•  Coal as fuel instead     34         8.9
   of gas in long kiln
 <17.5     < 5.1
<  9.4
5.1
  42.9     17
           1.6
-0.2
           1.9
           < 6.7
< 4.9
           18.9
           62
           56

-------
                  TABLE 9.    ESTIMATED CONTROLLED SOV EMISSIONS - CEMENT PRODUCTION
                                                    X
Process
Base case:
Long kiln (natural gas)
• Suspension preheater
(natural gas)
• Flash calciner
(natural gas)
• Fluidized bed
(natural gas)
• Coal as fuel instead of
gas in long kiln
SOjx^Emission factor
for
(Ib/ton of cement)
Process Boiler Total
5.1 1.9
< 5.1 1.6 <
< 5.1 1.6 <
< 5.1? -0.2 <
7.0
6.7
6.7
4.9
17 1.9 18.9
cement production
Change in
emission factor
—
-0.3
-0.3
,-2.1
+11.9
SOX emissions
(10b Ib/yr)
1974** 1989-1974*
553 191
183
183
VL34
516
 *Based on the incremental production of cement from 1974 to 1989 of 27.3 x 106 tons (Table 1)
  multiplied by total SOX emission factors;  this assumes no retirement of existing facilitites.
**The SOX emissions in 1974 are based on the emission factor of 7.0 Ib S02/ton  of cement are
  7.0 x 79 x 106 = 553 x 106 Ibs.

-------
Ib SO /ton of ethylene.   The total SO  emissions are then about 2.2 lb/
     x                               x
ton.

     The process and boiler emissions are summarized in Table 10.  Emis-
sions from alternative processes are also included in the table.

Process Option 1 - Ethylene from the Pyrolysis of Naphtha—

     Pyrolysis of naphtha already accounts for 7% of domestic ethylene
production and is the predominant technology used in Europe and Japan.
This process alternative is described in Volume VI, page 22.

     A sulfur concentration of 500 ppm in the naphtha feed was used as
a basis for calculation.  This is equivalent to 6.1 lb SO /ton of ethyl-
ene.  Sulfur in the feed is later found in the pyrolysis gas (typically
80% of the total sulfur is found in the hydrocarbon fraction containing
four carbon molecules or less, e.g., C'^s and lighter, equivalent to
4.9 lb SO /ton of ethylene); in the pyrolysis gasoline (8% of the sul-
fur equivalent to 0.45 lb SO /ton of ethylene); and in the fuel oil
(12% of the sulfur equivalent to 0.75 lb SO /ton of ethylene).
                                           X
     Two methods of l^S removal are available:  simple caustic  scrub-
ging, and regenerative amine scrubbing followed by a caustic wash  to
remove the final traces of sulfur and C02  .  In the Petroleum Refining
Industry Report  (Vol. IV) a simple caustic scrubbing system was  chosen
since such a system normally is used until sulfur levels  in the €4  frac-
tion of the  pyrolysis gases exceed 600 ppm.

     All  the fuel oil produced  in the olefin plant  is  consumed  in  heat-
ers, so sulfur  in the fuel oil  will  result in an  estimated  uncontrolled
emission  of  0.75 lb SO  /ton of  ethylene;  these  are  not currently con-
trolled.

      SO emissions are expected  from  the  fuel burned in the  power boiler.
The  electricity  consumed  is equivalent to  1,100 Btu/lb of ethylene.
The  estimated  uncontrolled  emissions are  12.2 lb  SO /ton of  ethylene
and  the emissions controlled  to meet the  NSPS are estimated  at 2.6 lb
SO /ton of  ethylene.  The total controlled SOX  emissions are  then 3.6
Ibfton of ethylene; an  increase of  64% over  the base case.

Process Option 2 -  Ethylene From  the Pyrolysis  of Gas-Oil —

      Several plants now being constructed will  use gas-oil as feed.
The design of  such  plants is  well established at the commercial level
and the  practice is clearly going to become  common as ethylene produc-
 ers move  to assure  themselves of some flexibility in their choice of
 feedstock.   A process description for ethylene  production based on gas-
oil is given in Volume  VI,  page 27.

      A sulfur concentration of 2,000 ppm in the feedstock (equivalent
 to 32.2  lb S02/ton  of ethylene) was used as  the basis for calculation.
                                    43

-------
   TABLE 10.   SULFUR DISTRIBUTION AND ESTIMATED S02 EMISSION FACTORS
                             OLEFIN INDUSTRY
                          (LB S02/ TON ETHYLENE)


                                 Base case
                               Ethane-Propane     Naphtha    Gas oil

  Feed*                             0.03            6.1        32.2

  Distribution

       Acid gas**                   0.03            4.9         8.7
       Fuel oil                      —             0.75       21.0
       Gasoline                      —             0.45        2.5

       Total                        0.03            6.1        32.2

  Emissions (uncontrolled)

       Acid gas                     0.03            4.9         8.7
       Heater exhaust***             —             0.75        8.82
       Total process                0.03            5.65       17.52
       Power boiler                 9.98           12.2        16.64
       Total                       10.0            17.9        34.2

  Emissions (controlled)
Acid gas removal
exhuaust
Heater exhaust***
Total process
Power boiler
Total
Control efficiency, %
0.03

—
0.03
2.13
2.2
78
0.26

0.75
1.01
2.6
3.6
80
0.39

8.82
9.21
3.54
12.8
63
  *Feed does not include fuel for power boiler.

 **Before acid gas recovery ,(sulfur is present as H2S and COS, but
   sulfur weight is calculated on S02 basis.)

*** From combustion of fuel oil - 100% of the fuel oil produced is
    consumed in the naphtha process and 42% of the fuel oil produced
    is consumed in the gas oil process.  These emissions are normally
    not controlled.
                                   44

-------
The S02 emission problems and control methods in this alternative are
identical to those in the ethylene from the pyrolysis of naphtha alter-
native.  The uncontrolled and controlled emissions are shown in Table
10.

     As in the base case process, the gas-oil cracker supplies its own
fuel needs.  This is accomplished by a recycling of all the residue gas
and 42% of the fuel oil produced.  Thus, estimated uncontrolled emis-
sions from the heaters are 8.82 Ib SO /ton of ethylene as shown in
Table 10.                            x

     Additional SO  emissions are expected from the fuel burned in the
power boiler.  The electricity consumed is equivalent to 1,500 Btu/lb
of ethylene  (or 3 million Btu/ton ethylene). The estimated uncontrolled
emissions are 16.6 Ib SO /ton of ethylene and the emissions controlled
to meet the NSPS are estimated at 3.54 Ib SO /ton of ethylene.  The
total controlled emissions are 12.8 Ib/ton or ethylene, a 480% increase
over the base case, largely due to heater exhaust emissions.

Long-Term Process Options

     The more active development programs in the area of olefin tech-
nologies which were not analyzed in depth in the original study include:

     Cracking technology, such as

     •    Coil cracking of vacuum gas oil,

     •    Hydropyrolysis,

     •    Autothermic pyrolysis, and

     •    Fluid bed cracking, as well as

     Coal-based technology, such as

     •    Plasma arc pyrolysis, and

     •    Clean coke process.

     Since most of these advanced technologies are being developed and
commercialized during a period when environmental regulations are in
effect, the developers recognize the need to comply with existing en-
vironmental codes and are taking appropriate measures while developing
the process.  Sulfur is an even more significant problem for these
advanced technologies than for the existing technologies, because of
the nature and sulfur content of the proposed feedstocks.  However, for
all, the gaseous sulfur is in the form of hydrogen sulfide, for which
an abundance of control technology is available, although some of this
technology may require modifications to be effective.  For example,
Union Carbide has had to do this to reduce the problem of butadiene


                                  45

-------
polymerization in amine scrubbing systems.  Again, the fuel oils and
pitch produced as byproducts will present internal use and marketing
problems if steps are not taken to reduce their sulfur contents.  How-
ever, the problems of sulfur content in the byproducts are generally
recognized by the developers.  Thus, to be acceptable, the commercial
versions of the processes must incorporate techniques for coping with
this problem.

Summary

     The emission factors for the base case process and alternative
processes in the olefin industry are summarized in Table 10.   The un-
controlled emissions from the alternative processes are significantly
higher than those from the base case process.  The higher emissions are
mainly due to the higher sulfur content in the feedstocks (naphtha and
gas oil) and partly due to increased consumption of electricity.  The
major fraction of the emissions is controlled by add-on control tech-
nology.  The control efficiency varies from 60 to 80%.  The estimated
SO  emissions after control from the olefin industry are summarized
inXTable 11.  Obviously, with the very high growth rate anticipated by
this industry (8%/yr) , careful attention will have to be given to all
forms of potential discharges, including sulfur, and to effective and
economical control.  As shown in Table 11, use of the heavier gas-oil
feedstocks will very dramatically increase the controlled SO  emissions.
                                                            X

AMMONIA

Base Case Process - Ammonia Production Based on Natural Gas

     Ammonia is made by the reaction of nitrogen with hydrogen.  All
processes manufacturing ammonia utilize air as the source of nitrogen.
Natural gas is generally used as the source of hydrogen.  The four maj-
or operations in manufacturing ammonia are:  gas preparation, carbon
monoxide conversion, gas purification, and ammonia synthesis.  A de-
tailed description of the process is given in Volume VII, page 25.

     There are no SO  emissions from the process.  The emissions from
a coal-fired power boiler are estimated at 2.6 Ib SO /ton of ammonia
with no control, and 0.6 Ib SO /ton of ammonia based on control to meet
the NSPS.  The emissions are based on electricity consumption of 45.5
kWh/ton of ammonia (Volume VII, page 29).

Process Option 1 - Ammonia Production Based on Coal Gasification

     Using coal as a feedstock for ammonia production involves freeing
the hydrogen that is present in the fuel, and reacting the carbon in
the fuel with water vapor to release more hydrogen.  A detailed process
description is given in Volume VII, page 37, for a hypothetical plant
in Southern Illinois using coal with 4.33% sulfur.
                                  46

-------
                  TABLE 11.  ESTIMATED CONTROLLED SOY EMISSIONS - OLEFINS INDUSTRY
Process
Base case:
E-P process
• Naphtha process
• Gas-oil process
SOX Emission factor for olefin production
(Ib/ton ethylene) Change in
Process Power boiler Total emission factor

0.03 2.13 2.2
1.01 2.6 3.6 +1.4
9.21 3.54 12.8 +10.6
SOV Emissions
(106 Ib/yr)
1974 1989-1974*

28.6** 62
102
361
 *Based on incremental production from 1974-1989 of 28.2 million tons (Table 1)  multiplied by
  total SOX emission factor.   This assumes no retirement of facilities existing  in 1974.

**Estimated 1974 emissions, based on the total emission factor of 2.2 Ib S0x/ton of ethylene,
  are 2.2 x 13 x 106 = 28.6 x 105 Ib.

-------
                                     o
     The effluent stream  from  the CO shift-conversion step contains hy-
drogen sulfide, carbonyl  sulfide, etc.  These gases must be removed for
process reasons, and a number  of different acid gas removal systems may
be used.  The  acid gas may be  removed  in a Rectisol process (Volume VII,
page 42) followed by a Glaus conversion plant.

     Coal consumption is  1.57  tons per ton of ammonia (Volume VII, page
45).  The uncontrolled emissions, based on 4.33% sulfur in Illinois
coal, are 136  Ib S/ton of ammonia.  The controlled emissions from a
tail gas cleanup system in the Glaus unit are estimated at 0.4 Ib SO /
ton of ammonia (Volume VII, page 61).                               x

     The estimated emissions from the power boiler are 9.4 Ib SO /ton
of ammonia with no control, and 2.0 Ib SO /ton of ammonia, basedxon
control to meet the NSPS, giving a total controlled emission of 2.4 Ib
SO /ton, which is 300% more than from  the base case process.
  X
Process Option 2 - Production  of Ammonia from Heavy Fuel Oil

     The production of ammonia from heavy fuel oil includes reacting
fuel oil with  oxygen in the presence of steam at a temperature of 2,000
2,500 F to produce syngas, followed by shift reaction, heat recovery,
acid gas removal, final gas purification, compression, and synthesis.
An air separation plant is required to produce oxygen and nitrogen.  A
detailed description of the process is given in Volume VII, page 67.

     The air pollution problems in this alternative are similar to those
in the production of ammonia from coal.  However, the uncontrolled emis-
sions are about two-thirds of  those from an alternative process based
on coal (Volume VII, page 78).  The estimated uncontrolled emissions
are 92 Ib S/ton of ammonia.  The controlled emissions from the tail gas
system in the  Glaus plant are  0.3 Ib SO /ton of ammonia.

     The estimated emissions from the power boiler are 6.0 Ib SO /ton
of ammonia with no control, and 1.3 Ib SO /ton of ammonia, basedxon
control to meet the NSPS.  Thus, the totaJ SO  discharges are 1.6 Ib/ton
ammonia or 67% more than  from  the base case process.

Summary

     The SO  emissions from the base case process and process alterna-
tives are summarized in Table  12.  The uncontrolled emissions are sig-
nificantly higher in the alternative process because of the increase in
the sulfur content of the feedstock and increased consumption of elec-
tricity.  The  sulfur emissions may be controlled by both Glaus plant and
FGD systems.  The controlled sulfur emissions from the alternative pro-
cesses are two to four times those from the base case process, but re-
main relatively small in comparison to other national problems.   The,
total SO  emissions for the industry in 1989 would be only 13.2 x 10
Ib, using the current process, or 36.2 Ib, using the base process, plus
48

-------
                  TABLE 12.  ESTIMATED CONTROLLED SOX EMISSIONS - AMMONIA INDUSTRY
Process
Ammonia production
based on
Base case:
Natural gas
• Coal gasification
• Heavy fuel oil
SOX emission factor for ammonia production
(Ib/ton ammonia) Change in
Process Power boiler Total emission factor


0 0.6 0.6
0.4 2.0 2.4 +1.8
0.3 1.3 1.6 +1.0
S0y Emissions
(10b Ib/yr)
1974 1989-1974*


5.5** 7.7
30.7
20.5
 *Based on increment in production from 1974-1989 — 12.8 x 106 ton/yr.

**Estimated 1974 emissions, based on the total emission factor of 0.6 Ib S0x/ton of ammonia,
  are 0.6 x 9.2 x 106 = 5.5 x 106 Ib/yr.

-------
the ammonia from coal process for all future expansion ("worst case").
Although this is small on a national basis when compared to current or
anticipated coal usage and SO  generation for power utilities, there
may nevertheless be significant regional impacts because of the region-
al character of the ammonia industry.

ALUMINA AND ALUMINUM

Alumina Production

Base Case Process - Bayer Process for Producing Alumina—

     The Bayer process for producing alumina is based on imported baux-
ite.  The process includes digestion of ground bauxite, removal of im-
purities, precipitation of aluminum trihydrate, treatment of spent li-
quor to regenerate the caustic, and calcination of aluminum trihydrate
to produce pot feed alumina.  A detailed description of the process is
given in Volume VIII, page 106, based on natural gas use in calcining.

     There are no SO  emissions from the process using natural gas.
Estimated SO  emissions from the power boiler (based on electricity
consumption of 275 kWh/ton of alumina) are 16 Ib SO /ton of alumina, if
not controlled, and 3.3 Ib SO /ton of alumina, if controlled to meet
the NSPS.                    X

Process Option 1 - Hydrochloric Acid Ion Exchange Process—

     This process includes dehydration of the raw clay, leaching with
hydrochloric acid, separation of residue, purification of the solution
by amine ion exchange, crystallization of aluminum chloride and decom-
position and calcination to obtain alumina.  A detailed description of
the process is given in Volume VIII, page 22.  No commercial plant em-
ploying this process has ever been built (Volume VIII, page 27); thus,
data for emissions are best estimates, based on natural gas for cal-
cining.

     As in the base case process, SO  emissions are not present in this
process.  Estimated SO  emissions from the power boiler are 8 Ib SO /ton
of alumina with no control and 1.7 Ib SO /ton of alumina, if controlled
to meet the NSPS — a 48% reduction.

Process Option 2 - Nitric Acid Ion Exchange Process—

     This process includes calcining the kaolin clay, leaching the cal-
cined clay with hot nitric acid, separating the clay insolubles, remov-
ing the iron and other impurities, recovering the alumina by hydrolysis,
recovering the nitric acid, and calcining to obtain alumina.

     In this process, coal and oil are used as fuel.  SO  is present in
several waste gas streams (Volume VIII, page 38) and in ihie flue gas
from the combustion processes.  Most of the coal is used for calcination,
                                   50

-------
Part of the SO  will be removed in the alkaline dust present in the
kiln.  It is expected that SO  emissions may be comparable to those
from the coal combustion in tne cement kiln.  If coal containing 3.5%
sulfur is used in the cement kiln, the SO  emissions will be 23.8 Ib
SO , based on the emission factor reported* in "Compilation of Air Pol-
lution Emission Factors" (EPA, 1973).  The fuel consumption in the ce-
ment kiln is about 4.6 x 10  Btu/ton of cement (Volume X, page 46).
The coal consumption in the nitric acid process is 24 x 10  Btu/ton of
alumina.  Therefore, the estimated uncontrolled SO  emissions are 120 Ib
SO /ton of alumina.                               x
  x

     In the operation of a cement kiln, about 50% of the SO  emissions
are removed in the particulate collection device.  Comparable collect-
ion efficiency may be expected in this alternative.  The estimated emis-
sions from the particulate collection device will be 60 Ib SO /ton of
alumina.

     The emissions from the power boiler are comparable to those in the
hydrochloric acid ion-exchange process alternative, about 8 Ib SO /ton
of alumina, if not controlled, and 1.7 Ib SO /ton of alumina, if con-
trolled to meet the NSPS.  Consequently, at §1.7 Ib SO /ton, this route
will be a significantly larger contributor of SO  than the conventional
bauxite process, largely because of coal use in fhese calculations.

Process Option 3 - Toth Alumina Process—

     This process involves the chlorination of alumina-containing raw
materials in the presence of carbon to produce aluminum chloride vapor
and other volatile chlorides.  These products are subsequently purified
to eliminate other metal chlorides and then oxidized to alumina and
chlorine for recycle.  The details of the process are given in Volume
VIII, page 39.

     Coal is the only source of sulfur in this process.  Coal is used
in kilns, chlorinator, etc.  The  coal consumption in the process is
equivalent to 8.45 x 10  Btu/ton  of alumina.  If all the sulfur from
the  coal (containing 3.5% sulfur) appears in the flue gas, it will re-
sult in an estimated uncontrolled emission  of 49 Ib SO /ton of alumina.
                                                      x
     The major  sulfur source is the chlorinator.  After chlorine is re-
moved,  the gas  contains carbon monoxide, carbon  dioxide, hydrogen  chlo-
ride, chlorine, and  sulfur compounds, such  as H~S, COS, etc.  This gas
stream  is burned in  a CO boiler and  then caustic scrubbed  to remove SO  .
The  cost of  SO  control is shown  in Volume  VIII, page 46.  If 80%  of  x
the  SO  is removed,  the estimated emissions will be 10 Ib  SO /ton  of
      x                                                     x
alumina.

     The electricity consumption  in  this alternative is 333 kWh/ton of
alumina.  Estimated  emissions  from the power boiler are 19 Ib  SO  /ton
of  alumina,  based  on no control,  and 4.2 Ib SO  /ton of alumina,  D"ased
on  control to meet  the NSPS.   Combining all sources, the  total  control-
led  SO  emissions  are 14.2 Ib/ton.
      X
                                  51

-------
Aluminum Production

Base Case Process - Hall-Heroult Process—

     This process involves reduction of alumina to aluminum, using elec-
trolytic cells.  The existing plants use Soderberg electrodes and the
new plants use prebaked electrodes which consume less energy.  A de-
tailed description of the process is given in Volume VIII, page 52.

     Electrolytic cell operation produces particulate, sulfur, and hydro-
carbon emissions, as well as fluoride emissions.  The amount of emissions
depend upon the type of cell used.

Prebake Cells and Soderberg Cells—

     The electrolytic reduction of aluminum produces a CO exhaust at
the anode of the cell.  As the exhaust leaves the cell, it entrains
particulates including fluoride salts.  The exhaust also contains
noxious gases such as HF and traces of I^S.

     In a prebake plant the carbon anode, which is consumed as a part
of the reaction, is formed in a baking furnace.  The manufacturing
process is similar to coke-making in that a paste made of pitch and coal
is devolatilized forming a solid carbon anode.  This process emits large
amounts of hydrocarbons, sulfur compounds, and particulates.

     Plants which use Soderberg cells do not require anode furnaces
because the anode is formed from a coke-based paste within the elec-
trolytic cell itself.  In this case, the particulate, sulfur and hydro-
carbon emissions, common to the anode furnace of a prebake cell, will
be emitted in the electrolytic cell of the Soderberg process instead.

SOX Emissions--

     Total SOX emissions are estimated to be 60 Ib/ton of aluminum, but
of course dependent upon the sulfur content of the pitch and coke used
to manufacture the anodes (Volume VIII, page 134) and cell type.  Based
on 85% control efficiency using scrubbers, controlled emissions are esti-
mated to be 9 Ib SO,,/ton of aluminum.
                   A

     When current aluminum technology is used, major SOX emissions are
from the power boiler used to generate electricity and are estimated at
900 Ib S0x/ton of aluminum,  with no control, and 200 Ib S0x/ton of aluminum
based on control at the power plant to meet the NSPS.  The newer plants
consume less electrical energy (12,000 kWh/ton of aluminum for the newer
plants versus 15,600 kWh/ton of aluminum for older plants) and therefore
the emissions from the power boiler are reduced to an estimated value
of 700 Ib S0x/ton of aluminum based on no control and 150 Ib S0x/ton of
aluminum based on control to meet the NSPS.
                                      52

-------
Process Option 1 - Alcoa Chloride Process —

     This process starts with pot feed alumina from the Bayer process.
The alumina is converted into aluminum chloride by chlorination in the
presence of carbon to form volatile aluminum chloride.  This, in turn,
is purified and fed to the electrolytic cells to produce molten alumi-
num.  Calculations below are Arthur D. Little estimates based on limited
published data.

     Energy consumed by the process includes No. 6 fuel oil equivalent
to 24.8 x 106 Btu/ton of aluminum  (Volume VIII, page 69).  The estimated
emissions are 41 Ib S0x/ton of aluminum based on 1% sulfur in the
oil burned.  The burned off gases  from the coker must be treated to re-
move sulfur emissions.  Because of the extremely high sulfur loadings
expected, the stack scrubbing process used should be regenerative.  Most
of the scrubbing processes appropriate for this type of control were  de-
veloped for utility boilers.  Such systems include a scrubbing system
 (for gas-liquid contact) and an alkali handling system  (for regeneration
of caustic, for example).  These systems have been proven  effective on
pilot-scale systems, and several full-scale  systems are now in operat-
ion.
     The  SO   control costs are  shown  in  Volume  VIII,  page  67.   The es-
 timated  controlled emissions  are 8 Ib SO /ton of  aluminum  based  on 80%
 efficiency.                              x

     Based on the  electricity consumption  of 10,500 kWh/ton  of  aluminum
 (Volume  VIII,  page 69),  emissions  from the power  boiler are  estimated
 at  610 Ib SO  /ton  of aluminum,  based  on  no control, and 130  Ib  SO  ,
 based  on control  to meet  the  NSPS.  The  combined  controlled  emissions
 are  then  138  Ib SO /ton  of aluminum,  a slight  improvement, achieved by
 reduced  electricity consumption.

 Process  Option 2  - Refractory Hard Metal Cathodes Process—

     This process  is based on Bayer alumina  and uses  titanium diboride
 cathodes  instead  of conventional carbon cathodes.   The cathodes  are
 assumed  to be retrofitted  in  existing large  cells to  increase product-
 ion  and  reduce energy  consumption  (Volume  VIII, page  74).  A detailed
 process  description is  given  in Volume VIII, page 68.

     The  emission  loadings and  control methods  in this alternative are
 the  same  as  in the base  case  process.

     The electricity consumed in the  process is 12,480 kWh/ton  of  alu-
 minum  (Volume VIII, page 76).   Estimated emissions from the  power  boil-
 er  are 720  Ib SO  /ton  of aluminum, based on  no  control, and  160  Ib SO /
 ton  of aluminum Eased  on control to meet the NSPS.   Total  SO  emissions
 are  thus  reduced  by about  20%.                               x

 Process  Option 3  - Combination  of  the Clay Chlorination Process  and
 the  Alcoa Chloride'—

     The details  of the  combined process are given in Volume VIII,

                                    53

-------
page 79.  Since intermediate alumina production is eliminated, there is
some reduction in emissions.  The estimated emissions from the Toth pro-
cess include 49 Ib SO /ton of alumina.  For aluminum production, 1.93
tons of alumina/ton or aluminum are required (Volume VIII, pages 50 and
81).  Therefore, emissions are estimated to be 95 Ib SO /ton of alumi-
num.  The controlled emissions are estimated at 19 Ib s6 /ton of alumi-
num.

     The electricity consumed in the process is 10,637 kWh/ton of alu-
minum (Volume VIII, page 81), resulting in an estimated emissions of
620 Ib SO /ton of aluminum, based on no control, and 130 Ib of SO /ton
of aluminum based on control to meet the NSPS.   The total controlled
emissions would then be approximately 149 Ib SO /ton, which is not sig-
nificantly different from other anticipated technology.

Summary

     The estimated SO  emission factors for the alumina and aluminum
                     •y
industry are summarized in Table 13.

     The major difference in the SO  emissions in alumina production
results from the use of coal in the nitric acid ion exchange and Toth
processes.  Natural gas is assumed to be used as fuel in the remaining
alumina processes.  Control methods include removal of SO  in the bag
filters (because of the alkaline nature of the particulate cake) and
in wet scrubbers.

     In the aluminum industry, the majority of the emissions come from
the power boiler because of the consumption of large quantities of
electricity.  These emissions which are proportional to energy demand
may be controlled by flue gas desulfurization and coal cleaning methods.

     The SO  emissions in the alumina and aluminum industry are summar-
ized in Table 14.  The process alternatives in aluminum production are
aimed at reducing the consumption of electricity and therefore do re-
duce power boiler emissions concurrently.

     SO  emissions from the aluminum production are significantly high-
er (about 40 times) than those from the alumina production (Table 14).
While a significant portion of aluminum manufacturing now relies on
hydroelectric power, we expect that much of the incremental aluminum
capacity to be installed in the next 15 years will be based on fossil
fuels and, to a lesser extent, on nuclear power.  Thus we believe the
SO  emissions calculated here are not unrealistic if the alumina/alumi-
num processes discussed are actually implemented and U.S. capacity
growth projections are realized.
                                   54

-------
       TABLE 13-   ESTIMATED CONTROLLED SOX EMISSION FACTORS - ALUMINA AND ALUMINUM INDUSTRY
          Process
Emission factor - no  control   Emission  factor - with  control
      (Ib/ton product)                (Ib/ton product)          Control
Process  Power boiler Total   Process  Power boiler  Total  efficiency  (%)
Alumina

Base case:
     Bayer                    0

  •  Hydrochloric acid        0
     ion exchange

  •  Nitric acid            120
     ion exchange

  •  Clay chlorination       49
     (Toth)

Aluminum

Base case:
     Hall-Heroult            60
     (current practice)
             16

              8


              8


             19
 16

  8


128


 68
 0

 0


60


10
3.3

1.7


1.7


4.2
            900
960
          200
 3.3

 1.7


61.7


14.2
         209
79

79


52


79
                                                                  78
• Hall-Heroult (new)
• Alcoa chloride
• Refractory hard metal
cathode
• Combination of clay
chlorination (Toth)
and Aloca chloride
60
41
60

95


700
610
720

620


760
651
780

715


0,9
8
9

19


150
130
160

130


159
138
169

149


79
79
78

79



-------
TABLE 14.  ESTIMATED CONTROLLED SOX EMISSIONS - ALUMINA AND ALUMINUM INDUSTRY
Controlled SOY emission factor SOV emissions
(Ib/ton product)
Process Process Power boiler
Product : Alumina
Base case:
Bayer
• Hydrochloric acid ion exchange
• Nitric acid ion exchange
• Clay chlorination
„ (Toth alumina)
[V
Product : Aluminum
Base case:
Hall-Heroult (current practice)
• Hall-Heroult (new)
• Alcoa chloride
• Refractory hard metal cathode
Base case:
Bayer and Hall-Heroult (c.p.)
• Toth alumina and Alcoa chloride


0
0
60
10


9
9
8
9

9
19


3.3
1.7
1.7
4.2


200
150
130
160

206***
130
Change in (10 6 Ib/yr)
Total emission factor 1974 1989-1974*


3.3
1.7
61.7
14.2


209
159
138
169

215
149


25.4** 35.5
-1.6 — 18.3
+58.4 — 663
+10.9 -- 153


1,045 1,463
-50 — 1,113
-71 — 966
-40 — 1,183

1,070** 1,505
-66 — 1,043
                                                                                    (continued)

-------
                                               TABLE 14.(continued)
       *
        Based on multiplying emission factor by increments in production from 1974 to 1989 of 10.75
        million tons alumina/year and 7.0 million aluminum tons.  This assumes no retirement of
        1974 facilities.
      **
        Estimated 1974 emissions, based on the emission factor of 3.3 Ib S0x/ton of alumina, are
        3.3 x 7.7 x 10° = 25.4 x 10° Ib SO  from alumina production; and, based on the emission
        factor of 209 Ib SO /ton of aluminum, are 209 x 5 x 106 = 1,045 x 106 Ib SOX from aluminum
        production;  resulting in a  total of 1,070 x 10" Ib SOX from alumina and aluminum.
     ***
        Bayer process, plus Hall-Heroult process  (current practice), is used for comparative analysis
        based on 1.93 tons alumina per ton aluminum (Volume VIII, P. 116).
01       C.P. - current (1974) practice

-------
PULP AND PAPER

Chemical Pulp

Base Case Process - Kraft Pulping—

     The base case process selected for chemical pulp was the Kraft
process, which includes cooking of wood chips at elevated temperature
and pressure in a digester.  When cooking is completed, the contents of
the digester are forced into the blow tank where the spent cooking li-
quor is drained.  After unreacted chunks of wood are removed, the pulp
is washed, bleached, pressed, and dried into the finished product.  Fur-
ther details are found in Volume 5, page 55.

     Recovery of both the inorganic cooking chemicals and the heat con-
tent of the spent liquor, which is separated from the cooked pulp in
the blow tank, is economical.  Recovery is accomplished by first con-
centrating (in an evaporator) the liquor to a level that will support
combustion and then feeding it to a furnace (recovery boiler) where
heat recovery takes place.  This is followed by chemical recovery in a
smelt dissolving tank.

     The characteristic odor of the Kraft mill is partially caused by
hydrogen sulfide.  The major source is the direct contact evaporator in
which the sodium sulfide in the black liquor reacts with the carbon di-
oxide in the furnace exhaust.  In addition, the Kraft-process odor also
results from an assortment of organic sulfur compounds, all of which
have extremely low odor thresholds.  These compounds are emitted from
many points within a mill; however, the main sources are the digester/
blow tank systems and the direct contact evaporator.

     The lime kiln can also be a potential source as a similar reaction
occurs involving residual sodium sulfide in the lime mud.  Lesser amounts
of hydrogen sulfide are emitted with the non-condensible off-gases from
the digesters and multiple-effect evaporators.

     Sulfur dioxide emissions result mainly from oxidation of reduced
sulfur compounds in the recovery furnace.  Sulfur dioxide may also be
present in the power boiler effluent gas, depending on the sulfur con-
tent of the fuel used.

     The uncontrolled total reduced sulfur (TRS) emissions from the
Kraft process are shown in Table 15.  The emissions are expressed as
equivalent weights of sulfur.  The EPA is planning to establish air
pollution emission  standards for new Kraft pulp mills with proposed
standards shown in Table 15.   The control method for eliminating re-
duced sulfur compounds (TRS)  is oxidation to SO .  The uncontrolled TRS
emissions from the various process steps are 24 Ib TRS/ton of pulp
(Table 15) and will result in emissions of 48 Ib of SOX (as S02) per
ton of pulp.
                                  58

-------
      TABLE 15    UNCONTROLLED EMISSIONS OF TRS AND PROPOSED EMISSION
                    STANDARDS FOR NEW KRAFT PULP MILLS
                             Uncontrolled
                           emissions of TRS
                 Proposed standard for TRS
                            (Ib/ton of pulp*)    (ppm) (Ib/ton of pulp)*

Recovery boiler

Lime kiln

Smelt tank
*****
Brown stock washer
*****
Black liquor oxidation
******
Condensate stripping
******
Digester

15

Trace

0.1

1.0

N.A.

N.A.

8
**
5
***
5^
*** ****
5

5

5

5

5

0.15

0.025
****
0.025

0.01

0.01

0.01

0.01
    Multiple ejfg
    evaporator

    Bleaching

    Lime slaker

           TOTAL
   N.A.

Negligible

   Trace

 •>. 24
5          0.01

5          0.01
     Air-dried basis.
    **
     TRS corrected to 8% oxygen.
   ***
     TRS corrected  to 10% oxygen.
  ****
      Fresh water will ensure  compliance.
 *****
     Likely  control method  is utilization of  gas  stream  as  combustion
     air  in  the  recovery  furnace.
******
      Likely  control method  is utilization  of  gas  stream as  combustion
      air  in   the  lime kiln.
                                      59

-------
     Combination boilers  (bark and oil, or bark and coal) or bark boil-
ers are used as on-site power boilers in the pulp and paper industry.
The discharge from a bark boiler consists of gaseous products of com-
bustion containing negligible sulfur compounds.  The combination boil-
er  (coal and bark) will have SO  emissions in proportion to the con-
sumption of coal and the sulfur content of the-coal.  The estimated
emissions based on the consumption of 1.4 x 10  Btu fossil fuel (con-
taining 3.5% sulfur in coal) per ton of pulp are 7.8 Ib SO /ton of pulp
from the boiler.  If these emissions are controlled, the estimated con-
trolled emissions will be 1.7 Ib SO /ton of pulp.
                                   x

     The energy credits in the Kraft process include 140 kWh/ton of
pulp (Volume V, page 57).  Thus, the estimated reduction in emissions
at the power boiler are 8.1 Ib SO /ton of pulp, with no control, and
1.8 Ib SO /ton of pulp, based on §0  controlled to meet the NSPS.
Therefore* the net estimated reduction in emissions from the above two
power boilers is 0.3 Ib SO /ton of pulp, with no control, and 0.1 Ib/
ton of pulp, based on control to meet the NSPS.  Total SO  controlled
emissions are estimated at 47.9 Ib/ton.

Process Option 1 - Alkaline-Oxygen Pulping—

     The alkaline-oxygen (A-0) pulping process is receiving industry
interest because of its potential for a non-sulfur cooking step, which
would eliminate the air pollution due to sulfur compounds.  The steps
in the A-0 process include an alkaline treatment to soften the wood
chips, mechanical disintegration, and treatment with oxygen under al-
kaline conditions to remove most of the lignin.  This is followed by
the last three stages of the conventional multistage bleaching sequence:
chlorine dioxide, caustic extraction, and(chlorine dioxide). A detailed
description of the process is given in Volume V, page 84.

     Because sulfur is not used in the process, chemicals containing
sulfur are not present in the air emissions.  However, SO  may be pres-
ent in the air emissions from the lime kiln and the powerxboiler because
of the sulfur present in the fuel.  The emission factor for fuel com-
bustion in lime kilns is not available.  It is expected that a major
fraction of the SO  will be absorbed by the alkali dust in the kiln
and by the air pollution control device, provided wet scrubbers or bag
filters are used.  The estimated emissions from the on-site power boil-
er are 18.3 Ib SO /ton of pulp and the estimated controlled emissions
are 4.0 Ib SO /ton of pulp.  The emissions are based on consumption of
coal containing 3.5% sulfur at a rate equivalent to 3.3 x 106 Btu/ton
of pulp.   The energy credits in the A-0 process include 50 kWh/ton of
pulp (Volume V, page 90).  Thus, the estimated reduction in emissions
at the power boiler is 2.9 Ib S0x/ton of pulp and the estimated re-
duction in the controlled emissions is 0.63 Ib SO /ton of pulp.
                                                 X        r  r

     The net total emissions from the A-0 process are 15.4 Ib SO /ton
of pulp,  with no control, and the net total controlled emissionsxare
3.37 Ib SO /ton of pulp — a very significant improvement over the con-
ventional Kraft process.

                                  60

-------
Process Option 2 - Rapson Effluent-Free Kraft Process—

     A number of changes in the base case Kraft process have been made
to eliminate effluents in the Rapson process.  These changes are des-
cribed in detail in Volume V, page 95, and include closing the water
cycle, use of chlorine dioxide rather than €!„, countercurrent washing,
etc.

     The air emissions from the process are not affected by the above
changes and the emissions are 24 Ib TRS/ton of pulp (Volume V, page 104).
The emissions from the on-site power boiler are less because of reduced
consumption of purchased fuel (coal).  The total energy consumed in the
process, excluding that required for the lime burning, is 0.7 x 10  Btu/
ton of pulp (Volume V, page 101).  The estimated emissions are 3.88 Ib
SO /ton of pulp with no control, and the estimated controlled emissions
are 0.84 Ib SO /ton of pulp.  Electricity from the on-site power boiler
is sufficient tor the process requirements in this alternative.

Newsprint Pulp

Base Case Process - Refiner Mechanical Pulping (RMP) Process—

     RMP is a mechanical pulping process that is an improvement over
the conventional groundwood process.  Wood chips, sawdust, and shavings
from sawmills or plywood mills can be used as raw materials  for the RMP
process, but such materials cannot be used as raw materials  for the
groundwood process.  The wood particles are  reduced to  fibers  in a
pressurized disc refiner which consists of two circular metal  plates
that generally rotate in opposite directions.  The RMP  pulp  (80%),  to-
gether with Kraft pulp  (20%), is used in newsprint paper  production.  A
detailed description of the RMP process is given in Volume V,  page  60.

     There are no emissions containing sulfur compounds from the RMP
process, except those from  the power boiler.  The purchased  electricity
is  1,475 kWh/ton of pulp.   Therefore, the estimated emissions  from  the
power boiler are 85.6 Ib SO  /ton of  pulp, with no control,  and the  es-
timated controlled emissions  are 18.6 Ib  SO  /ton of pulp.   Of  course,
if  hydroelectric power  is used, no  SO emissions would  occur.
                                      X

     The newsprint pulp consists of  80% RMP  pulp and  20%  Kraft pulp
 (Volume V, page 112).   Therefore,  the estimated uncontrolled emissions
are 4.8 Ib TRS/ton of newsprint  pulp from the  process and 68.4 Ib  SO  /
ton of newsprint pulp from  the  on-site and utility  power  boilers.   Tn"e
estimated  controlled  emissions  are  9.6 Ib SO /ton of  newsprint pulp
 (TRS  converted  to  SO  )  from the  process  and  $4.9  Ib  SO /ton of news-
print  pulp  from the power boilers  for a  total  of  24.5 ¥b  SO /ton.

Process Option  1  - Thermo-Mechanical Pulping (IMP)--•

      The  TMP  process  is similar to  the base  case  process  (RMP process),
 except that  the wood  particles  are  preheated to  130 C for a short period
                                   61

-------
and then reduced to fibers in a pressurized disc refiner.  A detailed
process description is given in Volume V, page 104.

     The emissions from the TMP process are identical to those from the
base case process, except that the newsprint pulp can be composed of
95% TMP pulp and 5% Kraft pulp.  Therefore, the estimated uncontrolled
emissions are 1.2 Ib TRS/ton of newsprint pulp from the process and
81.3 Ib SO /ton of newsprint pulp from the on-site and utility power
boilers.  The estimated controlled emissions are 2.4 Ib SO /ton of news-
print pulp from the process (TRS converted to SO ) and 17.? Ib SO /ton
of newsprint pulp from the power boilers, totaling 20.1 Ib/ton, an 18%
reduction per ton of newsprint pulp.

Process Option 2 - De-inking of Old News for Newsprint Manufacture—

     The de-inking of old news for newsprint manufacture is a well-
established commercial practice and a detailed description of the process
is given in Volume V, page 113.  The only emissions from the process
are those generated in the power boilers (on-site and utility).  The
total energy required is 6.1 x 10  Btu/ton of newsprint pulp.  There-
fore, the estimated emissions are 33.8 Ib SO /ton of newsprint pulp,
with no control, and 7.32 Ib SO /ton of newsprint pulp based on control
to meet the NSPS.              X

Summary

     The emission factors and total emissions for the base case pro-
cesses and alternative processes are summarized in Tables 16 and 17.
The emissions in the alkaline-oxygen pulping alternative are signifi-
cantly lowered due to elimination of compounds containing sulfur from
the process.  The emissions from pulp and paper mills generally are not
controlled, except that the form of sulfur in the emissions is changed
from reduced sulfur to sulfur oxides.  Newsprint based on TMP and RMP
processes results in smaller emissions than the base case process, larg-
ely because of the smaller fraction of Kraft pulp.  De-inking benefits
from an absence of sulfur-containing compounds in the de-inking of old
news for newsprint manufacture.

     On a national scale A-0 pulping would have the greatest impact in
reducing SO  emissions.
           x

GLASS

Base Case Process - Regenerative Furnace

     The base case process selected was the natural gas-fired furnace
with cold charge.  A detailed description of the furnace is given in
Volume XI, page 17.   Operating conditions for the base case process are:
                                   62

-------
                   TABLE 16.  ESTIMATED SOX EMISSION FACTORS - PULP AND PAPER INDUSTRY
Emission factor* - no control
(lb/ton of pulp)**
Process Process Boiler Total
Base case:
Kraft pulping 48 -0.3 47.7
• Alkaline-oxygen pulping 0 15.4 15.4
• Rapson effluent-free 48 3.88 51.9
Kraft pulping
Base case:
Newsprint - RMP 9.6 68.4 78.0
i
• Newsprint - TMP 2.4 81.3 83.7
• De-inking of old 0 33.8 33.8
news for newsprint
manufacture
Emission factor* - with control Control
(lb/ton of pulp)** Efficiency
Process Boiler Total (%)

48 -0.1 47.9
0 3.37 3.4 -78
48 0.84 48.8 -6

9.6 14.9 24.5 -69
2.4 17.7 20.1 -76
0 7.3 7.3 -78
 *Reported as Ib S02/ton of ADP.  The uncontrolled process emissions are present as TRS compounds.




**Air-dried basis.

-------
                TABLE 17 — ESTIMATED CONTROLLED S0v EMISSIONS:   PULP AND PAPER INDUSTRY
                                                   X
Process
Base case:
Kraft pulping
• Alkaline-Oxygen (A-0)

Emission
(Ib/ton of pulp)*
Process Boiler Total
48
-0.1 47.9
3.37 3.4
factor
Change in
emission factor
-44.5
SOX emissions
(10°
1974
766***
lb/yr> **
1989-1974
829
58.8
       pulping

       Rapson effluent-free
       Kraft pulping
                                48
0.84  48.8
+0.9
844
  Base case:
                                                                                   ***
Newsprint - RMP 9.6
• Newsprint - TMP 2.4
• De- inking of old news —
for newsprint manufacture
14.9
17.7
7.3

24.5
20.1
7.3

—
-4.4
-17.2

95.6 42
34
12

***
 Air-dried  basis.

 Based  on increment  in  production  from  1974  to  1989:   17.3 x  10^  tons of Kraft pulp and
 1.7  x  10"  tons  of newsprint pulp.
c
 Estimated  1974  emissions,  based on  the total emission factor of  47.9 Ib/ton of  Kraft  pulp are
 47.9 x 16  x 106 = 766  x 10° Ib SOX  /yr from the Kraft pulp process and, based on  the  total
 emission factor of  24.5 Ib/ton of newsprint pulp,  are 24.5 x 3.9 x 106 =  95.6 x 10^ Ib  S0x/yr
 from the newsprint  pulp process.

-------
     Furnace type:  - Side port,  regenerative
     Fuel:  - Natural gas
     Glass  type:  -  Soda lime
     Plant  location: - East North Central
     Pull rate:  - 200 tons per day
     Feed rate:  - 240 tons per day
     Efficiency:  -  90%, or 180 tons per day.
     The typical glass melting operation is described in detail in Vol-
ume XI, page 19.  A glass-melting furnace has both particulate and gas-
eous emissions which must be controlled.

     The sulfur oxide emissions arise from two sources:  sulfur in the
fuel and decomposition of mineral sulfates in the glass melt.  The
sulfur within the fuel (natural gas) is generally a minor source of
sulfur oxide emissions and generally results in the production of S0~.
The decomposition of mineral sulfates in the melt produces S0_.  Al-
though some SO. decomposes to S0? at temperatures above 2790 F, some
will remain in  the form of SO..  The emission factor for the glass fur-
nace is 3.0 Ib  SO /ton of glass (Volume XI, page 27).
                 X
     The emissions containing high concentration of SO- are expected to
increase the corrosion potential and make air pollution control more
difficult.  The control techniques  for  SO  and SO  consist of  contact-
ing the gases with caustic or lime  to convert the gaseous emissions  to
sulfate or sulfite salts of sodium or calcium.  Scrubbers for  this pur-
pose should utilize high-efficiency entrainment separators to  avoid  car-
ryover of sulfuric acid mist.

     Other add-on control systems include Dry Tesisorb  systems .  The
Tesisorb X  (chemical  composition not available), which  is added up-
stream of the bag filter, is capable of  collecting gaseous pollutants.
The reaction products are solid inorganic particles and are  removed  in
the bag filter.  The  bag filter is  operated  at 200°F.   The outlet SO
concentration is less than 50 ppm,  equivalent to 0.7 Ib SO /ton of   x
glass.                                                    x

     The estimated emissions from the power  boiler to  generate electri-
city are 1.1 Ib SO /ton of glass with no control, and  0.24 Ib  SO  /ton
of  glass based  on control to meet the NSPS.                     x

Process Option  1 - Coal  Gasification

     Coal gasification processes include,  with some variation, the  fol-
lowing steps:   coal  handling  and storage,  coal preparation,  gasifica-
tion,  oxidant  feed  facilities,  and  gas  cleaning.   The  gas produced  from
coal gasification  is used as  a  fuel source in the  glass furnace.   The
 *
  Marketed by Teller Environmental Systems,  Worcester,  Mass.
                                   65

-------
details of the process are given In Volume XI, page 32.

     When coal gasification is used to generate the gaseous fuel for
the glass melting furnace, the major environmental difference is not in
the glass-making process, but rather in the fuel-generating process.

     The fuel gas produced in the coal gasifier will contain sulfur,
mainly in the form of l^S.  So that excessive SC^ emissions do not oc-
cur, the sulfur must be controlled either before or after gas combus-
tion in the furnace.  Two control technologies could be considered:  a
sulfur-recovery process, such as the Stretford process, to remove lUS
from the fuel gas prior to burning the fuel in the furnace, or flue gas
desulfurization, such as the processes currently used on utility boil-
ers for removing SC^ from exhaust gases.  The latter case is the same
as the control for direct coal firing (as discussed below).

     The estimated emissions from burning of coal are 47.0 Ib S0x/ton
of glass based on no control.  Thus, the total uncontrolled emissions
are 50.0 Ib S0x/ton of glass (includes 3.0 Ib S0x/ton of glass from
mineral sources).  If the Stretford process (sulfur collection effi-
ciency of 90%) is used to collect sulfur compounds, the estimated
emissions from the glass furnace are 7.7 Ib S0x/ton of glass.  The esti-
mated capital and operating costs for the Stretford process are given
in Volume XI, page 39.

     The SO  can be further removed in the wet scrubber or Tesisorb
systems.   The estimated concentration of SOX in the tail gas will be
50 ppm, equivalent to 0.77 Ib S0x/ton of glass.  The emission factor
is based on an estimated 10% higher gas flow rate compared to that in
the base case process.

     The SOX emissions from the power boiler are comparable to those
from the base case process, and are about 1.1 Ib S0x/ton of glass with
no control, and about 0.24 Ib S0x/ton of glass based on control to meet
the NSPS.  Therefore, the total SOX emissions after control are only
slightly increased to 1.01 Ib/ton.

Process Option 2 - Direct-Coal-Firing

     In this option, pulverized coal is used directly in burners to
supply the energy.   The details of the process are in Volume XI, page 35.

     Coal is high-sulfur fuel (3.5% sulfur), so SO  will be present in
the flue gas.   The uncontrolled emissions in this alternative will be
comparable to those from a coal gasification process, and amount to
50 Ib S0x/ton of glass.

     The emissions may be controlled in a wet scrubber.  Calcium or
                                   66

-------
sodium-based alkali may be used in the scrubber.   The cost of alkali
will be significantly higher because of high concentrations of SO .
The SO  concentration in the outlet gas may be reduced to about 200  ppm.
The gas flow rate in this alternative is about 10% higher than that  from
the base case process.  Therefore, the estimated controlled emissions
are 3.1 Ib SO /ton of glass.
             X

     The SO  emissions from the power boiler in this alternative are
comparable to those from the power boiler in the base case process,
about 1.1 Ib SO /ton of glass with no control, and 0.24 Ib SO /ton of
glass based on control to meet the NSPS.  The total then is a signifi-
cantly higher 3.3 Ib/ton.

Process Option 3 - Coal-Fired Hot Gas Generation (COHOGG)

     This system generates a hot gas by separating  char and volatiles,
burning the char, and then mixing the products of combustion with the
volatiles and burning them together.  A pneumatic conveying system
feeds powdered coal to the pyrolyzer along with limestone or a mix of
limestone and sodium chloride.  The outlet temperature of the gases
leaving the afterburner is 3,000 F.  These hot gases are used in the
glass melting furnace to heat the glass batch.  The details of the pro-
cess are given in Volume XI, page 48.

     From one-third to one-half of the sulfur in the coal is expected
to come off in the pyrolyzer as hydrogen sulfide (H_S).  The limestone,
in turn, reacts with  the hydrogen sulfide, producing water and calcium
sulfide (CaS).  Unreacted limestone, char, and calcium sulfide go to
the char burner, which is a fluidized-bed combustor, where the remain-
ing sulfur  forms SO  .  There the remaining limestone from the pyrolyzer
step reacts with the  SO  , forming calcium sulfite  which  leaves  the bed
along with  the ash from  the char.

     The only difference between  the coal gasification alternative  and
that of heating the glass-melting furnace with a hot combustion  gas  is
that the sulfur control  in  the latter  option  is inherent within  the
process itself and is not required as  a part  of the pollution control
apparatus.  The gas volumes to the glass furnace are higher  with the
COHOGG process because of the efficiency losses inherent  in  the  system.
Except for  the slight size  difference,  however, the pollution control
system for  the glass  melting  furnace  itself will be  the  same as  for
coal gasification.   The  concentration  of SO   in the  tail  gas is  about
the same as that in  the  coal  gasification process with SO control
 (Stretford  process).  However, because of  the higher  gas  r"low rates
 (25% higher compared  to  the gas  flow  rate  from  the  coal  gasification
process;   see Volume  XI,  page 40  for  details),  the  emission  factor
with control  will  be  about  1.0 Ib SO  /ton  of  glass  for COHOGG.
                                    X
  Some sulfates are also formed.
                                   67

-------
     The SO  emissions from the power boiler in this alternative are
comparable £o those from the base case process, or about 1.1 Ib S0x/ton
of glass with no control, and about 0.24 Ib SO /ton of glass based on
control to meet the NSPS.  Total controlled emissions, at 1.24 Ib/ton,
are 32% larger than the base case.

Process Option 4 - All-Electric Melting Process

     Molten glass can be heated by the passage of an electric current.
Both the design and the operation of an all-electric, glass-melting
furnace differ greatly from those of the typical natural gas-fired, re-
generative furnace.  The.electric furnace without its regenerative
checker structure is a much simpler design.  The details of the electric
furnace are given in Volume XI, page 53.

     The option to heat glass-melting furnaces electrically results in
a shift in the environmental problems irom the furnace to the electric
power generating plant.  In this case, the only exhaust from the glass-
melting furnace is from the decomposition of carbonates, sulfates, ni-
trates, etc., in the glass batch.  The exhaust will be almost entirely
CO  with approximately 3 Ib SO /ton of glass (Volume XI, page 27).
  ^                           X

     The control system for this exhaust is identical to the one used
for the base case and for coal gasification, but the size of the system
is considerably smaller because of the greatly reduced exhaust volume.
If the SO  emissions are assumed to be controlled to 200 ppm, the emis-
sions will* be reduced to 0.1 Ib SO /ton of glass.
                                  X

     The electricity consumption is increased in this alternative to
780 kWh/ton of glass.  Therefore, the estimated emissions from the pow-
er boiler are 45.2 Ib SO /ton of glass, with no control, and 9.82 Ib
SO /ton of glass, based on control, to meet the NSPS.  Total controlled
emissions are significantly increased, at 9.82 Ib/ton.

Process Option 5 - Batch Agglomeration/Preheating

     The intent of preheating is to prereact the batch ingredients.
Batch preheating is an energy-conserving technology relating to a fur-
nace modification rather than a method of furnace heating.  Hence, this
technology is applicable to all of the previously discussed methods of
heating, except for electric heating.  The details of the process are
given in Volume XI, page 58.

     The uncontrolled SO  emissions from this process alternative are
comparable to those from the base case process, or about 3.0 Ib SO /ton
of glass.   The gas flow rate is about 20% lower than that from the^Sase
case process.

     The control methods used in the base case process may be used here.
If the SO  emissions are controlled to a concentration of 50 ppm, the
                                   68

-------
estimated emission rate is 0.56 Ib SO /ton of glass.
                                     X
     The SO  emissions from the power boiler in this  alternative are
comparable to those from the base case process, or about 1.1 Ib SO /ton
of glass, with no control, and 0.24 Ib SO /ton of glass, based on con-
trol to meet the NSPS.  Thus applying preheating, even to the conven-
tional furnace, can reduce SO  emissions by about 15%.
                             X

Summary

     The emission factors and total emissions for the flat glass indus-
try are summarized in Tables 18 and 19.

     The uncontrolled emissions from the base case process and process
alternative depend on the sulfur content of the fuel.  Processes based
on natural gas obsiously have significantly lower emissions compared
to those from the processes based on coal, as expected.

     The controlled emissions depend on the extent of control .  The con-
trolled emissions range from 0.8 to 10 Ib SO /ton of  glass.  Also, the
SO  emissions from the electric melting process alternative are the
highest because of emissions from the electric power  generating site.
Preheating appears to be an attractive alternative because both energy
is saved and SO   emissions are reduced.
               X

COPPER

Base Case Process - Conventional Copper Smelting

     Conventional smelting involves the smelting  of  sulfide  concentrates
in the  reverberatory  furnace either directly or  after roasting.   The
mixture of molten sulfides from  the reverb  is  converted to blister  cop-
per in  converters.  A detailed description  of  this process  is given in
Volume  XIV,  page  23.  The capacity of  the base case  smelter  is 100,000
tons of anode copper  per  year,

     With  over  30% sulfur, most  copper  concentrates  contain  more  sul-
fur than copper.  About 1-2% of  the  sulfur  entering  the smelter is  lost
in the  slag  and  perhaps 3-4% evolves  as  fugitive emissions.   The  remain-
ing sulfur  is  lost/discharged  in gaseous  emissions from roaster,  reverb,
and converters.   Typical  sulfur  distributions  from conventional smelt-
ing are:

      Source                 Calcine Smelting   Green  Charge Smelting
      Roaster                      20
      Reverb                       25                    40
      Converter                    50                    55
      Slag and Fugitives          _ 5                   _ 5_
           Total                  100                   100

                                     69

-------
                      TABLE 18.   ESTIMATED SOV EMISSION FACTORS - GLASS INDUSTRY
                                              X
Emission factor* - no control Emission factor* - with control
(Ib/ton of glass) (Ib/ton of glass) Control
Process Process Power boiler Total Process Power boiler Total Efficiency
 • 	
1.1 4.1 0.7 0.24 0.94 77
1.1 51.1 0.77 0.24 1.01 92
1.1 51.1 3.1 0.24 3.34 93
1.1 51.1 1.0 0.24 1.24 98

45.2 48.2 0.1 9.82 9.92 79
1.1 4.1 0.56 0.24 0.80 80

*Reported as Ib S0_/ton of glass.

-------
                   TABLE 19.   ESTIMATED CONTROLLED SOY EMISSIONS - GLASS  INDUSTRY
SOX emission factor for glass production
SOX emissions
(Ib/ton of glass) Change in (10b Ib/yr)
Process Process Boiler Total Emission factor 1974 1989-1974*
Base case :
Regenerative furnace** 0.7 0.24 0.94
• Coal gasification 0.77 0.24 1.01 +0.07
• Direct coal firing 3.1 0.24 3.34 +2.40
• Coal-fired hot 1.0 0.24 1.24 +0.30
gas generation
• Electric melting 0.1 9.82 9.92 +8.98
• Batch preheat with 0.56 0.24 0.80 -0.14
natural gas firing
27.3** 12.2
13.1
43.4
16.1

129
10.4

 *Based on increment in production from 1974-1989 = 13 x 106 ton/yr. (see Table 1)

**Estimated 1974 emissions,  based on the total emission factor of 0.94 Ib S0x/ton of glass,
  are 0.94 x 29 x 106 = 27.3 x 106 Ib/yr.

-------
     The estimated total uncontrolled emissions based on the 28.6% cop-
per and 33.4% sulfur in the copper sulfide concentrate are 2.31 tons of
S0_/ton of copper.
     SO  emissions from smelters are controlled by the New Source Per-
formance Standards (NSPS) as follows:

     •    Emissions from streams containing high concentrations of SO-
          (converter and roaster gases) have to be controlled by tech-
          nology such as sulfuric acid plants.  Double absorption plants
          (or equivalent) are mandatory for new sources.  This combina-
          tion of uncontrolled reverbs but controlled roasters and con-
          verters can recover about 50-70% of the sulfur in the feed in
          the form of sulfuric acid.

     •    New Source Performance Standards would allow the use of re-
          verbs only for the smelting of "impure" concentrates (con-
          centrates containing As, Sb, Bi, etc.), but reverb emissions
          would not have to be treated for S0_ removal.  About 30% of
          the feed sulfur is emitted from the reverbs.

     •    Federal Ambient Air Quality Standards  (AAQS's) which define
          permissible ground level concentrations of S0_ have to be
          met at all times by a combination of permanent controls (e.g.,
          acid plants) and production curtailment.  The issue — when
          should production curtailment be used — has not been resolved.
          The degree of permanent controls necessary to meet Ambient
          Air Quality Standards varies greatly with location.  The con-
          trol of roaster or converter emissions is adequate for meet-
          ing Ambient Air Quality Standards only in certain locations.

     At present, there are three control methods in use at copper smel-
ters for reducing the sulfur dioxide concentrations in the vicinity of
a smelter:  (1) the use of a tall stack to disperse dilute gas streams;
(2) the production of sulfuric acid by the contact process from con-
centrated gas streams to achieve a degree of reduction in emissions;
and (3) production curtailment.

     The contact sulfuric acid process is well established for treating
SO -containing off-gases from metallurgical plants.  Modern contact
acid plants require at least 4.5-5% sulfur dioxide in the feed gas to
operate autogenously (i.e., without external heat).  For handling lower
concentrations of SO., an additional fuel input  is required.  Acid plant
size is primarily a function of the volume of gas handled.  Thus, for
a constant acid output, an acid plant operating on more dilute gases is
much larger (and more expensive) than an acid plant operating on more
concentrated gases.  With the currently used vanadium pentoxide cata-
lysts, the upper level of SO^ concentration in the feed gas to an acid
plant is between 7% and 9%.  More concentrated gas streams require di-
lution.
                                   72

-------
     Conventional smelters are located in geographical areas where acid
markets are unavailable and where all SO^-containing gas streams are
vented to the atmosphere (after particulate control, if necessary).
Streams from the roaster and converter, handled to minimize air leakage,
contain SCL concentrations greater than 4-5%, which is adequate for
autogenous sulfuric acid manufacture—the most cost-effective control
technology for removing SO  from such streams.  The reverb gases are
a high-volume (up to 100,000 scfm) and low-concentration (0.5-2% SO )
stream, not amenable to autogenous sulfuric acid manufacture and have
to be discharged via tall stacks.  In short, the emission sources of
SO  from a controlled conventional smelter are:
  x
     •    Reverb gas - 82,000 scfm with 0.5-2% S02

     •    Acid plant tail gas - 38,800 scfm with 0.2% SO

     •    Anode furnace gas - data not available, contains some S0_.

     The above controlled emissions are equivalent to about 1,400 Ib
S0x/ton of copper, representing a 70% control efficiency.

     The electricity consumed in  the  copper  smelting  varies  from  347
to 441 kWh/ton of copper.   The estimated uncontrolled emissions at  the
power boiler are  20.0-25.6  Ib SO  /ton of copper and the  estimated con-
trolled  emissions are  4.4-5.6 lbXSO  /ton of  copper.
                                   X

Process  Option 1  - Outokumpu Flash Smelting

     The  flash smelting alternative  combines the  separate  roasting and
smelting  operations of conventional  copper extraction into one com-
bined  roasting/smelting process.  The major  advantages  of  the  method
are  a  reduction  in fuel used  for  smelting  and production of a  gas stream
having a concentration of  SO  which  is suitable  for sulfuric acid manu-
facture.   A detailed description  of  the process  is  given in Volume XIV,
page 41.

     The concentration of  SO   in flash smelter gas  is high,  containing
up  to  13% S0_.   Conventional  reverberatory furnace  gas,  on the other
hand,  typically  contains  0.5-2.0% SO-.  The high-strength gases are
most suitable  for the  manufacture of sulfuric acid.   The variable
 strength/variable volume  S0_  gas stream from converters can be mixed
with the steady  stream of flash smelter gas to provide a stream high
 enough in SO,,  for acid manufacture.   At high matte grades, a large
 amount  of sulfur is  eliminated  in the flash furnace.   This improves acid
 plant  performance,  because the volume and strength of the input stream
 are more constant.

      The emissions from the acid plant tail gas represent the major
 source of pollutants from the Outokumpu smelter (55,000 scfm,  0.05% SO )
 The sulfur loss in various process streams includes 0.07% in drying,
                                   73

-------
1% each in smelting and converting, 0.1% in the anode furnace, and 1.2%
in slag.  The total sulfur emissions to the atmosphere are 2.8% of the
sulfur in the feed (Volume XIV, page 216).

     The estimated uncontrolled emissions from the copper smelter are
4,620 Ib S09/ton of copper, and the controlled emissions are 133 Ib
S0_/ton of copper.

     The emissions from the power boiler are estimated at 21.1 Ib SO /
ton of copper with no control, and 4.6 Ib SO /ton of copper based onx
control to meet the NSPS.   The emissions areXbased on the electricity
consumption of 366 kWh/ton of copper (Volume XIV, page 52).  Obviously,
the use of flash smelting, facilitating and recovery, allows the dual
benefits of energy saving and SO  reduction.
                                X

Process Option 2 - Noranda Process

     The Noranda process combines the three operations of roasting,
smelting, and converting of copper concentrates in a single reactor.
The heat losses suffered during the transfer of concentrate from the
roaster to the reverberatory furnace, are suppressed, as well as those
occurring during the transfer of the matte from the reverberatory fur-
nace to the converter.  A detailed description of the process is given
in Volume XIV, page 55.

     The sulfur dioxide concentration in the reactor atmosphere is
around 23% on a dry basis.  Because of air infiltration around the hood,
the gas stream entering the acid plant contains 10-13% SO-.  This stream
is only interrupted 5% of the time during tapping, and can be mixed with
the off-gases of other reactors and Fierce-Smith converters.  Air in-
leakage (thereby preventing atmospheric emissions) is high, but does not
affect the subsequent l^SO^ plant, because its operation is at or below
the 10-13% SO  concentration.  This steady, high-S02, gas-generation
level is a significant advantage over the conventional reverberatory
process.

     After dry-gas cleaning, wet-gas cleaning equipment is required to
scrub out the remaining fine particulates.   The gas can then be treated
in a double-contact acid plant.  Total sulfur recovery is as high as
that achieved with the Outokumpu flash smelting process.

     The SO  emissions from the power boiler in this alternative are
the same asxin the Outokumpu flash smelter process, 21.1 Ib SO /ton of
copper, with no control, and 4.6 Ib SO /ton of copper, based on control
to meet the NSPS.                     X

Process Option 3 - Mitsubishi Process Alternative

     The Mitsubishi process consists of three metallurgical stages,
each of which is carried out in a separate furnace.  Thus, there is a
smelting furnace for concentrates, a converting furnace to oxidize iron


                                   74

-------
in the matte and make blister copper, and a slag cleaning furnace.   In-
termediate products in the molten state move continuously among the res-
pective furnaces, which are thus functionally connected with each other.
A detailed description of the Mitsubishi process is given in Volume XIV,
page 71.

     On leaving any of the three furnaces, the mixed off-gases are ex-
pected to average over 10% SO  when the smelting furnace is operated
with air enriched to 25% oxygen.  This steady, high-SO  gas generation
is a significant advantage over the conventional reveroeratory process,
as sulfur can be readily recovered as sulfuric acid.  Since the molten
liquids flow continuously over very short distances, minimum air pol-
lution is generated in transfer operations, and "converter aisle losses"
typical of conventional operations are avoided.  Thus, fugitive emis-
sions are expected to be lower than for conventional or for Outokumpu
and Noranda (matte) processes.

     After cooling and dry cleaning in electrostatic precipitators  (or
fabric filters operated above the dew point), the collected off-gases
usually require a wet cleaning stage to remove any  fine particulates
and cool the gas to remove excess moisture.  The cleaned gases are
then admitted to a double-contact acid plant for ^SO^ manufacture.
Total sulfur recovery is over 90%, as with all these advanced pyrometal-
lurgical processes.

     The acid plant tail gas flow rate and SO  concentrations in the
gas are comparable to those from the Outokumpu smelter.  The emissions
from the power boiler are the same as in  the Outokumpu  flash smelter
process, or about  21.1 Ib SO /ton of copper, with no control, and  4.6  Ib
SO /ton of copper  based on control to meet the NSPS.

Process Option 4 - The Use of Oxygen in  Smelting Process Alternative

     Copper smelting can be conducted with pure oxygen  or  by using oxy-
gen-enriched air.  An increase  in oxygen  concentration  will result in
higher  process  temperatures.  A detailed  description is given  in Volume
XIV, page 82.   The specific example  selected  for examination  is Outokum-
pu flash  smelting.  Oxygen enrichment  results  in the reduction of  ef-
fluent  volume;  however,  the operating  temperatures  are  generally  in-
creased.

     Use  of oxygen will  result  in reduced effluent  volumes.  The  total
uncontrolled SO emissions from the  smelter will remain the same.   If
 the  emissions  are  controlled  to a constant  exit  concentration,  the con-
 trolled emissions  will be  reduced in direct  proportion to  the  gas  vol-
ume.   The interplay between  air preheat,  oxygen  enrichment, and  fuel
use  is  extremely complex and  is described in  more  detail in Volume XIV,
page 84.   This  report  concluded that use of  oxygen does not  lead  to any
 significant energy savings in  the copper sector.
                                   75

-------
Process Option 5 - Metal Recovery from Slag Process Alternative

     In conventional copper smelting, converter slag is recycled to the
reverb and all the slag tapped from the revert is discarded.  The cop-
per contained in the discarded slag is lost.  The amount of copper lost
in the slag is significant, at about 1.5 to 3% or more of the copper in
the feed materials.  The processes for recovering metal from slag are
described in Volume XIV, page 89.  Since there is little sulfur in the
slag, the process SO  emissions are minor.  The major source is the pow-
er boiler, but it is impossible to quantify these without a detailed
analysis.  In the copper sector these SO  emissions would be relatively
small.

Process Option 6 - The Arbiter Process Alternative

     The Arbiter process is a hydrometallurgical process.  A process
description is given in Volume XIV, page 95.  The Arbiter process causes
little air pollution.  The process is energy-intensive and uses large
quantities of electricity and steam.  As reported in Volume XIV, page
104, steam consumption is 20,000 Ib/ton of copper, and electricity con-
sumption is 3,000 kWh/ton of copper.  The estimated emissions for the
power plant are 284 Ib SO /ton of copper with no control, and 62 Ib SO /
ton of copper based on control to meet the NSPS.                      x

Summary

     The emission factors and total emissions in the copper industry
are summarized in Tables 20 and 21.  The uncontrolled emissions from
the pyrometallurgical processes are essentially the same.  There are no
SO  emissions from the Arbiter process.  The emissions from the power
boiler in the Arbiter process are significantly higher than other pro-
cesses.  However, the total uncontrolled emissions from the Arbiter
process are about 5% of those from other processes.

     The emissions from the pyrometallurgical processes are controlled
by add-on control technology, mainly acid plants.  The control efficien-
cy is only 70% in the conventional smelting process and over 90% in the
other newer technologies proposed in the options, which also offer en-
ergy advantages.  Therefore, on these two bases, it can be expected
that the newer approaches will see rapid implementation as plants are
replaced or additional capacity installed.

CHLOR-ALKALI

     Some process changes have little direct affect on SO  emissions
and thus received little or no attention in the Industry Reports (Vo-
lumes I-XV).  However, indirect SO  emissions by differing electric
energy or steam requirements between base case and alternative techno-
logy may arise.  Generally, such impacts are of a much smaller magni-
tude than in the industries mentioned above.  Examples of such potential
changes in SO  emissions, depending on the process implemented, are
found in the chlor-alkali industry and others subsequently discussed
in this report.
                                  76

-------
                    TABLE 20.   ESTIMATED S0x EMISSION FACTORS - PRIMARY COPPER INDUSTRY

Emission
factor* - no control
(Ib/ton of copper)
Process

Base case:
Conventional smelting
• Outokumpu flash smelting
• Noranda
• Mitsubishi
• Arbiter
j
j
Process


4,620
4,620
4,620
4,620
— »••
Power boiler


23
21
21
21
284
Total


4,643
4,641
4,641
4,641
284
Emission
factor*
(Ib/ton of
Process


1,400
133
133
133
^^
-with control
copper)
Power boiler Total


5
5
5
5
62


1,405
138
138
138
62

Control
Efficiency
7.

70
97
97
97
78
*Reported as Ib SO^/ton of anode copper.

-------
                      TABLE 21.   ESTIMATED CONTROLLED SOV EMISSIONS - COPPER SMELTING
                                                         x
—4
00


Process
Base case:
Conventional smelting
• Outokumpu flash smelting
• Noranda process
• Mitsubishi process
• Arbiter
Controlled
(Ib/ton of
Si
Ox emission factor
anode copper)
Process Power

1,400
133
133
133


5
5
5
5
62

Change in
boiler Total Emission factor

1,405
138
138
138
62

—
-1,267
-1,267
-1,267
-1,343
SOX emissions
(10° Ib/yr)
1974 1989-1974*

2,250** 1,550
152
152
152
68
      *Based on increment in production from 1974 to 1989 of 1.1 million  ton/yr.  (Table 1)
     **Estimated 1974 emissions, based on the emission factor of  1,405  Ib  S0x/ton  of  anode copper,
       are 1,405 x 1.6 x 106 = 2,250 x 106 Ib/yr.

-------
Base Case Process - Graphite Anode Diaphragm Cell

     The graphite anode diaphragm cell was selected as a basis for judg-
ing the energy and environmental effects resulting from the process
changes studied, i.e., conversion to modified anodes and conversion to
modified diaphragms, both changes being aimed at energy conservation.

     Chlorine and caustic soda are produced by the electrolysis of brine.
In the diaphragm cell, chlorine is formed at the graphite anode, while
sodium ions migrate through the cell diaphragm to the cathode where a
dilute solution of NaOH is produced.  A more complete description of
this process is presented in Volume XII, pages 21-25.

     The electrolysis of brine is an energy-intensive process, requir-
ing about 3,274 kWh/ton of chlorine.  In addition, the plant requires
7,368 Ib of low-pressure steam/ton of chlorine for caustic evaporation
and brine heating.  Sulfur dioxide emissions from power generation are
estimated at 41.3 Ib SO /ton of chlorine if the NSPS for coal-fired
units is met.  Without rlue gas desulfurization, emissions are estimat-
ed at 190 Ib/ton of chlorine.  It was assumed  (Volume XII, page 26) that
the steam would be generated with byproduct hydrogen supplemented with
natural gas.  Hence, sulfur dioxide emissions  from steam generation may
be assumed to be nil.

Process Option 1 - Dimensionally Stable Anode  (PSA)

     The dimensionally stable  anode  (DSA) constructed of titanium and
coated with precious metal/rare earth oxides offers  numerous  advantages
over the graphite anode which  result  in power  savings of up  to  20%.  The
anode area and anode-cathode spacing  of the DSA  remains  constant  through
out use, thereby preventing  increased voltage  requirements  over  time.
Additional characteristics  of  the DSA are presented  in  Volume XII,  pages
44-47.   The DSA  diaphragm cell process  requires  3,151 kWh  and 6,402 Ib
of  steam/ton  of  chlorine.   Steam  is  generated  by the byproduct  hydrogen
and natural gas; hence,  SO   emissions are assumed to be nil.
                          X
Sulfur  Dioxide  emissions  from  power  generation are estimated at 39.7  Ib
SO  /ton of chlorine,  if  the NSPS  for coal-fired  units is met and 183  Ib
  -tr                 *
SO  /ton of chlorine,  if  the emissions are uncontrolled,  a  slight improve-
ment over  the base  case.

Process Option  2 -  Expandable  DSA

      Cell  power consumption can be  reduced  further by decreasing the
gap between  the anode and the  cathode.   With the rigid  DSA, a "working"
 space  must be allowed to assemble the cell.   The expanded  DSA is con-
 structed so  that the electrodes can be  moved inward after  the cell is
 assembled.   This reduced spacing results in a reduction of about 325
 kWh/ton of chlorine compared to the rigid DSA configuration.  If emis-
 sions  are controlled to meet the NSPS,  then this power savings reduces
                                    79

-------
sulfur dioxide emissions to about 35.6 Ib SO /ton of chlorine from the
emissions estimated for the rigid DSA cells, a 14% reduction when com-
pared with the conventional graphite anode process.  If emissions are
uncontrolled, then this power savings reduces emissions by 19 Ib/ton of
chlorine to 164 Ib SO /ton of chlorine.
                     x
Process Option 3 - Polymer-Modified Asbestors

     By replacing the conventional asbestos diaphragm by one which is
polymer-treated and is baked into place on the cathode, power consump-
tion can be reduced because diaphragm swelling does not occur.  Elec-
trical consumption may be reduced by as much as 280 kWh/ton if an extra
wide anode is used.  Thus, if flue gas is treated at the power source
to meet the NSPS, then SO  emissions are estimated to be 38.2 Ib/ton of
                         "V
chlorine, or 3.1 Ib/ton less than the emissions estimated for the rigid
DSA cells with standard diaphragms.  Thus, compared to the base case,
the controlled SO  emissions are reduced by 12%.
                 X
Process Option 4 - Polymer Membranes

     Microporous Teflon-type polymer membranes which would replace the
asbestos diaphragm entirely are being developed.  These would give an
energy performance equivalent to the  polymer modified asbestor with
the "extra wide" anode.  Hence, the controlled and uncontrolled emis-
sions would be identical to those of Process Option 3.

Process Option 5 - Ion Exchange Membranes

     These membranes would separate the anode and cathode compartments
of the cell and would allow the diffusion of sodium ions to the cathode
but would not allow the diffusion of hydroxyl ions to the anode.  Thus,
the ion exchange cell is capable of producing a 25 to 40% caustic so-
lution, whereas the standard cell produces a 10% caustic solution.

     Energy use for a DSA with an ion exchange cell producing 40% NaOH
is 2,980 kWh and 1,466 Ib steam/ton of chlorine.  Significantly less
steam is required since a rather concentrated NaOH solution is produced
directly from the cell.

     Sulfur dioxide emissions from power generation are estimated to be
37.6 Ib/ton of chlorine if the NSPS for coal-fired boilers is met.  If
flue gas is untreated, then emissions would be about 173 Ib/ton of chlo-
rine.

Process Option 6 - Mercury Cells

     Chlorine and caustic can also be produced in a mercury cell.  In a
mercury cell, brine flows through a slightly sloped trough.  At the di-
mensionally stable anodes, located at the top cover of the trough, chlo-
rine is produced.  A dilute sodium amalgam is produced at the cathode
(a thin layer of mercury which flows along the bottom of the trough).


                                  80

-------
a 50% caustic solution is produced from the amalgam;  and the mercury is
recycled to the cell.   Energy requirements for the mercury cell include
3,714 kWh of power and 550 Ib of steam/ton of chlorine.

     Sulfur dioxide emissions produced by power generation are estimat-
ed at 46.8 Ib/ton of chlorine if the NSPS for coal-fired boilers is met.
If flue gas is untreated, then emissions would be about 215 Ib/ton of
chlorine.

Summary

     The emissions factor and total emissions for the base case and al-
ternative processes are presented in Table 22.  The mercury cell is es-
timated to result in a 13% yearly increase from the base case because
of the relatively larger power requirements.  The modified anodes and
modified diaphragm options offer a 4 to 14% yearly reduction in SO
emissions  (up to 67 million Ib/hr) from the base case process.  On a
national basis, these changes are small compared to those examined in
the aluminum sector which also depends, to a large extent, on electric
energy.

IRON AND STEEL

Recovery of Carbon Monoxide from BOP Vessels

Base Case  Process - Complete Combustion System—

     The base  case process is a complete  combustion  system.   The  gases
issuing  from  the mouth of the furnace  are collected  in  a hood with con-
siderable  infiltration of air, burned  in  the  hood, and  cooled  and cleanec
of particulates before being released  to  the  atmosphere.   A detailed
description  of  this system is presented  in Volume  III,  page 18.

     The  combustion system consumes  14 kWh of electricity per ton of
steel;  this  electricity  is required  for  the operation of the heat scrub-
ber  system.   Sulfur dioxide  emissions  from the  power  boiler are estimat-
ed at  0.8  Ib/ton  of steel, if the  emissions are uncontrolled,  and 0.2
Ib/ton of  steel,  if the  emissions  are  controlled to  meet the NSPS.

Alternative  Process - Non-Combustion BOP Off-Gas Recovery System~

      In the  alternative  process,  carbon monoxide is  collected and re-
 covered from the  BOP  off-gas.   Two prominent systems, the OG process
 and  the IRSID-CAFL process,  are quite  similar and are discussed in Vol-
 ume  III,  pages 19-22.

      The off-gas  (OG) process  consumes about 8 kWh of electricity per
 ton  of steel.   Energy consumptions are less than those for the combus-
 tion system as a result  of  the  lower gas volumes handled in the non-
 combustion systems.   Sulfur  dioxide emissions from the power boiler are
 estimated at 0.5  Ib/ton  of  steel,  if the emissions are uncontrolled,


                                   31

-------
                     TABLE 22.   ESTIMATED CONTROLLED SOV EMISSIONS - CHLOR-ALKALI INDUSTRY
oo
ro
     **

Process
Base case:
Graphite-anode^
diaphragm cell
• Dimensionally
stable anodes
• Expandable DSA
• Polymer-modified
asbestos
• Polymer membrane
• Ion exchange
memb rane
• Modern mercury
cell
Controlled SOV
emission
(Ib/ton of chlorine)
Process Power boiler Total
0 41.3
0 39.7

0 35.6
0 38.2

0 38.2
0 37.6
0 46.8
41.3
39.7

35.6
38.2

38.2
37.6
46.8
factor SOX. emission
Change in (10° Ib/yr) ^
emission factor 1974 1989-1974
**
454 491
-1.6 — 472

-5.7 — 424
-3.1 — 455

-3.1 — 455
-3.7 — 447
+5.5 — 557

Based on increment in
production from 1974 to
1989 of
*
Estimated 1974 emissions, based on the total emission
are 41.3 x 11.0 x 106 = 454 x 106 Ib.
11.9 million ton/yr. (Table 1)
factor of 41.3 Ib S0x/ton of chlorine,

-------
and 0.1 Ib/ton of steel, if the emissions are controlled to meet the
NSPS.

Blast Furnace

Base Case Process - Low Sulfur Blast Furnace Hot Metal—

     The base case system is considered to include a blast furnace, cy-
clone, and venturi scrubber.  The blast furnace sulfur content is com-
pletely controlled by adding limestone to form the sulfur-bearing slag
and by limiting the sulfur content of the metallurgical coke.  This
process is further discussed in Volume III, page 29.

     Most of the sulfur leaves the blast furnace in the liquid slag and
hot metal; the off-gases contain only a negligible portion of the sul-
fur.  Hence, sulfur dioxide emissions from the process are estimated to
be nil.

     Electricity consumed in the base case process is 0.25 kWh per  ton
of hot metal.  Estimated emissions from the power boiler are 1.5 Ib SO/
ton of hot metal if the emissions are uncontrolled, and 0.3 Ib S0x/ton
if the emissions are controlled to meet the NSPS.

Process Alternative - Blast Furnace with External Desulfurization—

     Addition of an external desulfurization  step is  an alternative
method of controlling the  sulfur content of blast furnace  hot metal.  Ex-
ternal desulfurization  is  achieved by injecting  sulfur-reacting  reagents
 (e.g., calcium or magnesium compounds carried in an inert  gas such as
nitrogen) into high-sulfur hot metal from  a blast furnace.   These  com-
pounds form  a sulfide slag that is skimmed off prior  to charging the
hot  metal to the BOP.   Use of  external desulfurization  either permits
limestone and coke ratios  to be reduced, or allows  the  sulfur content
in the coke  to be  increased without  increasing the  limestone charge to
the  furnace. A  detailed description of  this  process  is given in Volume
III,  page 31.

      Sulfur  dioxide  emissions  from  the blast  furnace  off-gas are assumed
 to be nil.   Electricity consumption  in this  alternative process is the
 same as  the  base case process,  so power  boiler emissions  estimates are
also the  same.   In neither case,  is  the  contribution  significant.

 Quenching of Coke

 Base Case Process  -  Wet Quenching —

      In the base case  process, the  hot  coke is pushed from the  oven into
 a coke car  where it  is  quenched with water.   The steam which is produced
 is vented to the atmosphere.   Excess water is allowed to  drain  and is
 often recirculated.   Electricity consumed by the wet quenching process
 is not estimated in the analysis presented in Volume III, but it can be
                                    83

-------
assumed to be quite low.

Process Alternative Option - Dry Quenching—

     In dry quenching of coke, the hot coke, pushed from the ovens, is
cooled in a closed system.  "Inert" gases extract heat from incandescent
coke by direct contact.  The heat is then recovered from the inert gas-
es in a waste heat boiler or by other techniques.

     The incremental electrical requirements are estimated at 8.4 kWh/
ton of coke.  Incremental emissions from the power boiler are estimated
at 0.5 Ib SO /ton of coke, if the emissions are uncontrolled, and 0.1
Ib SO /ton or coke, if the emissions are controlled to meet the NSPS.
     x
     In addition, if the recovered heat, estimated at 1.1 x 10  Btu/ton
of coke, is used in the plant, then an air emissions 'credit' may
be realized.  This credit is estimated to be 6.1 Ib S0x/ton of coke,
if the emissions from the power boiler remain uncontrolled.  The credit
would be 1.2 Ib SO /ton of coke, if the emissions were controlled to
meet the NSPS.    x

     Thus, the net incremental emissions credit is estimated to be 5.6
Ib SO /ton of coke, if the emissions are uncontrolled, and 1.1 Ib SO /
ton or" coke, if the emissions are controlled to meet the NSPS.      x

Steelmaking

Base Case Process - Coke Oven, Blast Furnace (BF), and Basic Oxygen
Furnace (BOF) Route for Steelmaking—

     The base case process is the conventional process for Steelmaking.
It is assumed, for our analysis, that 30% scrap metal is used in the
BOP.  A detailed description of the process is given in Volume III, page
54.

     Electricity consumed by the coke oven, BF, BOF, and pollution con-
trol equipment is estimated at 120 kWh/ton of steel (Volume III, pages
79, 82-84).  SO  emissions from the power boiler are estimated at 7.0
Ib/ton of steelx if the flue gas is untreated, and 1.5 Ib/ton of steel
if the NSPS is met.

     Steam, consumed by the coke-making facilities, is required at about
670 Ib/ton of steel.  If this steam is generated with coal, then SO
emissions are estimated at 3.7 Ib SO /ton of steel if the  flue gas
emissions are uncontrolled, and 0.8 lx SO /ton of steel, if the flue gas
is treated to meet the NSPS.            x

Process Option - Direct Reduction Route for Steelmaking—

     Iron oxide pellets, ore lumps, and the like, can be partially re-
duced in the solid state by reaction with a reducing gas mixture.  These


                                  84

-------
prereduced materials can partially or entirely replace purchased scrap
in the steelmaking electric arc furnaces.  In this alternative process,
it was also assumed that 30% scrap is used with 70% prereduced pellets
to make steel.  A detailed process description is given in Volume III,
page 54.

     Electricity consumed by the sponge iron facilities, the electric
furnace shop, and the pollution control equipment is estimated at 712
kWh/ton of steel.  Sulfur dioxide emissions from the power boilers are
estimated at 41.3 Ib/ton of steel, if the emissions are uncontrolled,
and 9.0 Ib/ton of steel if the emissions are controlled to meet the
NSPS.

Summary

     The emissions factor and total emissions for the base case pro-
cesses and the alternative processes are summarized in Table 23.

     In the production of iron, it is seen that external desulfurization
offers no reduction in SO  emissions from the base case process, large-
ly because most of the sulfur values leave the blast  furnace area in
the slag.

     In the production of steel, recovery of  the  low-sulfur off-gas from
the BOF offers a 50% reduction in emissions because of  its waste heat
value, which  can potentially reduce need for  high-sulfur fuels.  How-
ever, steelmaking by the direct reduction method  will result in nearly
a  three-fold  increase in emissions over  the conventional route  because
of the high power requirements of this alternative route.

PHOSPHORUS/PHOSPHORIC ACID

Base Case Process - Electric Furnace Production  of Phosphorus  and
Conversion of Phosphorus  to Phosphoric Acid

      In this  process, phosphate  rock is  reduced  to  elemental  phosphorus
by coke in an electric  furnace.   Phosphorus vapor and carbon  monoxide
are  produced.  The  phosphorus  is  condensed and subsequently converted
to pure phosphoric  acid.    The details of  this process  are presented
in Volume XIII,  pages 23-24.

      Electric furnace production of  phosphorus is a very  energy-inten-
sive operation.   Electrical energy requirements  are estimated at 13,000
kWh  per ton  of phosphorus (P,).   More  than 90% of this energy is re-
quired  by  the furnace;  the remainder is  required for  pumping  operations.
In addition,  1.9 tons of  coke/ton of P,  are  charged  to the  furnace,
where carbon monoxide  is  produced.   The  carbon monoxide is  recovered
and  fired  in the rotary kiln where  the furnace feed materials are pre-
pared.  A small  amount  of natural gas  (6 million Btu/ton of P.) is re-
quired to supplement  the  carbon  monoxide gas.
                                   85

-------
                TABLE 23.  ESTIMATED CONTROLLED SOV EMISSIONS - IRON AND STEEL INDUSTRY
                                                  X
Process
                                                SOX emission factor
                                    	     S0.y emissions
     (Ib/ton of product)            Change in         (106 Ib/yr)
Process   Power boiler   Total   emission factor    1974   1989-1974*
     Base case:

          BOP with no
          off-gas recovery

       •  BOP with off-gas
          recovery

     Base case:

oo         Blast furnace

       •  Blast furnace with
          external desulfurization

     Base case:

          Wet quenching of coke

       •  Dry quenching of coke

     Base case:

          Coke oven, blast
          furnace, BOP route

       •  Direct reduction,
          EAF route
                                     0.8
                                                 0.2


                                                 0.1



                                                 0.3

                                                 0.3
               1.5
                                                 9.0
                          0.2


                          0.1



                          0.3

                          0.3
              N.A.       N.A.

  -1.2***      0.1***    -1
2.3
                          9.0
             -0.1
              0.0
                                                                         -1.1
              6.7
                          16.4**    12.0
           6.0
                          30**
          13.4

          13.4
N.A.      N.A.

         -30.6***



189**    138


Small    540
                                                                                            (continued)

-------
                                              TABLE  23 (continued)
       Based on  increment in production  from  1974  to  1989  -  27.8  x  106  tons  of  coke,  44.8  x 106
       tons of iron, and 60 x 10  tons of steel.

    **
       Estimated 1974 emissions based on multiplying  emission  factor by 1974 production:   62 million
       tons coke, 100 million tons iron  and 82 million  tons  BOP steel.


       Emission  factor is based on incremental requirements  and credits of  the  dry-quench  process
       as it compares to the wet-quench process.
00
^     N.A.:  Not available  - see footnote *** for dry  quenching  of coke.

-------
     Sulfur dioxide emissions result primarily from the coal-fired pow-
er boiler.  It is estimated that 71.5 Ib SO /ton of P205 are emitted if
the NSPS for coal-fired units is met.  Without flue gas desulfurization,
emissions are estimated at 329 Ib/ton of ^2^5"

     Sulfur dioxide emissions from the furnace coke are negligible since
the sulfur in the coke would be removed with the slag by the calcium
from the phosphate ore.

Process Option 1 - Chemical Cleanup of Wet-Process Phosphoric Acid

     In the wet process, phosphate ore is reacted with sulfuric acid.
Phosphoric acid (32% solution) is recovered from the undigested ore and
the gypsum byproduct.  This acid is concentrated by evaporation to the
54% P2°5 Pr°duct.  The wet process is described in more detail in Vol-
ume XIII, pages 35-37.  However, there are a number of impurities in
the regular wet-process phosphoric acid which make it unsuitable for
use in certain applications.  Wet acid is purified to sodium tripoly-
phosphate by a two-stage neutralization process (Volume XIII, page 49).

     Steam is required to concentrate the acid solution.  If the phos-
phoric acid plant is integrated with a sulfuric acid plant, then this
steam is generated when the sulfur is oxidized, yielding 13 x 10^ Btu/ton
of P20cj and controlled SC^ emissions of 10 Ib/ton of P205-  Approximately
half of the steam heat is used in the sulfuric acid plant and the other
half is used by the phosphoric acid plant.

     Electricity (250 kWh/ton of P2°5^ is recluired to drive pumps, ag-
itators, and filters in the wet-acid process.  An additional 16 kWh/ton
of ?2®5 is required for the chemical cleanup of the acid.  Sulfur di-
oxide emissions from power generation is estimated at 3.4 Ib SO /ton
]?2^5' if the steam is generated from a coal-fired boiler which complied
with the NSPS.  Without sulfur dioxide control, emissions would be
about 15.4 Ib/ton of P 0  .

Process Option 2 - Solvent Extraction Cleanup of Wet-Process Phosphoric
Acid

     Cleanup of wet-process phosphoric acid is based on the fact that
phosphoric acid can be transferred from solution in an aqueous phase to
solution in an organic phase and leave behind undesirable impurities,
such as calcium chloride, in the aqueous layer.  The organic phase can
then be contacted in a separate unit with fresh water to yield a pure
solution of phosphoric acid.  A detailed process description is given
in Volume XIII, page 60.

     The solvent-extraction process requires about 10,000 Ib steam/ton
of ?2®5 f°r concentration of the acid from 15% as it is produced in
the extraction section to a concentration of about 60% Po^S'  ^ t'ie
steam is generated with low-sulfur oil so that flue gas desulfurization
is not required, then emissions are estimated at 5.5 Ib SO /ton of P00,..
                                                          x          z j

                                  88

-------
     Electricity requirements will be about 300 kWh/ton of ?2®5'
estimated SO  emissions from power boilers are 3.8 Ib/ton of P2°
no control, and about 17.4 Ib/ton of P205 with control to meet the NSPS.

Summary

     The emissions factor and total emissions for the base case process
and alternative processes are summarized in Table 24.  Both of the pro-
cess options for "pure" phosphoric acid offer a significant reduction
in SO  emissions over the base case process.  This reduction is pri-
marily a result of the lower power requirements for both the chemical
cleanup and solvent extraction process options.

FERTILIZERS

Nitric Acid Production

     The manufacture of nitric acid, described in Volume XV, pages 25-28,
generates  significant emissions of nitrogen oxide and this  area of pri-
mary interest  is discussed in more detail  in  "Volume XV, NO Summary
Report".   Adoption of air pollution control is a  recent practice  in the
industry.  The process change considered  in the nitric acid production
is the application of alternative NO  abatement systems.   These process-
es are described in Volume XV, pages 33-42.

Base Case  Process - Nitric Acid Production without NO  Emission Control—
                                                     X

     Nitric acid is produced  by the  oxidation of  ammonia,  usually under
high pressure  and temperature over  a platinum catalyst.   Waste heat re-
covered  from  the product  gases is used  to generate steam.   The cooled
gases  are  subsequently  sent  to an absorption  tower to  form the acid
product.

     Fuel  requirements  are  satisfied by natural  gas.   Steam generated
by  the waste  heat  is  included as  an energy credit.   An analogous  credit
 for  SO  emissions was  assumed, since the steam generated by the waste
heat would otherwise  be generated  by low-sulfur  oil  so that flue  gas
desulfurization would  not be required.   This  emissions credit is  es-
 timated  to be 0.4  Ib  SO /ton of  nitric  acid.
                        x
 Process  Option 1 -  Catalytic Reduction—

      In the catalytic reduction  process, tail gas from the absorber
 passes through a combustor where the nitrogen oxides are reduced to N»
 and 0~.   Natural gas is used as  a fuel  in the combustor.   Steam is gen-
 erated by waste heat recovered  from the product gases.

      Energy requirements for the process include natural gas and elec-
 tricity.  Although catalytic reduction is an energy-intensive process,
 there is an SO  emissions credit because steam, which would otherwise
               X
                                   89

-------
       TABLE 24.  ESTIMATED CONTROLLED SO  EMISSIONS - PHOSPHORUS/PHOSPHORIC ACID PRODUCTION
                                         X
          Process
                                       SOX emission factor
      (Ib/ton
Process  Boiler  Total
   Change in
emission factor
                                                                                    emissions
                                                (10  Ib/yr)   ~
                                             1974   1989-1974
 Base case:
                      **
      Electric furnace

   •  Chemical cleanup of
      wet-process acid

   •  Solvent extraction of
      wet-process acid
 0       71.5    71.5

10.0      3.4    13.4


 5.5      3.8     9.3
                              -58.1
                              -62.2
                                               100**    45.0
                                                         8.4
                                5.9
  Based on the incremental production detergent grade phosphoric acid from 1974 to 1989 of
  0.63  x 10° Ib/yr multiplied by total emission factor.
**
  The SOX emissions based on the emission factor of 71.5 Ib/ton of detergent-grade phosphoric
  acid in the year 1974 are 71.5 x 1.4 x 106 = 100 x 106 pounds.

-------
be generated by oil, is generated with waste heat from the process.  The
net SO  emissions credit is estimated to be 0.6 Ib SO /ton of nitric
  . ,  x                                              x
acid.

Process Option 2 - Molecular Sieve Method—

     This method is based on the principles of adsorption, oxidation,
and regeneration of the molecular sieve.  An oil-fired heater is used
to provide heat for regeneration of the sieve.  The process has high
efficiency for removal of NO  gases.  A detailed description of the
method is given in Volume XV, page 36.

     Energy requirements include fuel oil, steam, and electricity.  High
power requirements  (26 kWh/ton of acid) result from the added compres-
sion requirements and the need to regenerate the sieve.  SO  emissions
from power boilers without flue gas desulfurization are estimated  to
be 1.5 Ib SO /ton of acid, and 0.3 Ib SO /ton of acid from power boil-
ers with flue gas desulfurization required to meet the NSPS.  SO   emis-
sions from the fuel oil  (0.5% sulfur) used in the sieve regeneration
and steam production are estimated at 11.1 Ib/ton of acid, for a total
of 11.4 Ib.

Process Option 3 -  Grand Paroisse or Extended Water Absorption—

     In this process, tail gas from the existing absorption  tower  is
delivered to a second absorption tower  for "extended absorption of ni-
trogen oxides by water".

     The energy requirements for this process  are small — only 7.2 kWh/
ton  of nitric acid  product are used.  Power  plant emissions  are estimat-
ed at 0.1 Ib SO /ton of  acid, if the NSPS  for  coal-fired  units  is  met,
and  0.4 Ib  SO /con  of acid for plants without  flue gas desulfurization.
             X

CDL/Vitok Process—

     In this process,  the  tail gas  is  scrubbed  with  nitric acid under
conditions  which reduce  the nitrogen  oxides  to  the desired level.   En-
ergy requirements  include  steam  and  electricity.  The  total  steam re-
quirements  will result  in  SO  emissions of 31.5 Ib/ton of acid, if low-
sulfur  fuel oil is  used, so  that  flue  gas  desulfurization is not  re-
quired.   Electrical requirements  (24  kWh/ton of acid) will result in
SO   emissions of 0.3  Ib/ton  of acid,  if the  flue gas is desulfurized  to
meet NSPS  for coal-fired,  and 1.4  Ib/ton,  if no desulfurization is per-
formed .

The  Masar  Process—

      In  this process,  the  tail  gas  is  chilled and then scrubbed with an
urea-containing  solution.  As nitric  acid  is produced,  urea  hydrolyzes
and  forms  ammonium nitrate.
                                   91

-------
     Steam is consumed at 105,000 Ib/ton of acid, and electricity, at
1.5 kWh/ton of acid for this process.   Steam generation by low-sulfur
fuel oil will result in 57.6 Ib SO /ton of acid.  The electrical re-
quirement contributes essentially no SO  emissions, whether or not the
power boilers meet the NSPS for coal-fired units.

Conversion to Fuel Oil in Mixed Fertilizer Plants Equipped with Bag
Filters

Base Case Process—

     The base case operation is an ammoniation granulation plant which
uses a natural gas-fired dryer and is equipped with a baghouse filter
to control fertilizer dust produced during drying.  Only 20 percent of
the estimated 200 plants is equipped with baghouses and would there-
fore be affected by this process change.  Sulfur dioxide emissions from
the base case operation are essentially nil.

Installation of Scrubber on Baghouse-Equipped Plants when Converting
from Natural Gas to Fuel Oil—

     When fertilizer dryers were converted from natural gas to oil, op-
erational problems sometimes resulted from the plugging of the baghouse
filters with ash from the oil.  If these problems cannot be resolved by
the modification of the combustion process, then wet scrubbers will be
required for particulate control.

     Incremental SO  emissions result from conversion to fuel oil as-
sumed to contain 0.5% sulfur.  If 20% of the sulfur dioxide is removed
in the wet scrubber, then SO  emissions are estimated to be 0.14 Ib/ton
of fertilizer; no further specific control for SO  is anticipated.
                                                 X
Continued Operation of the Baghouse Filter after Conversion from
Natural Gas to Fuel Oil—

     If filter clogging problems can be alleviated by proper design and
operation of the fuel oil-fired dryer, the baghouses will not have to
be replaced by scrubbers.  If the fuel oil contains 0.5% sulfur, and
since 0.3 million Btu of fuel are required to dry each ton of fertilizer
(Volume IV, page 54), then 0.18 Ib SO  will be emitted per ton of pro-
duct by conversion to fuel oil.      X

Summary

     The emission factors and total emission for the base case and al-
ternative processes for NO  abatement and fuel conversion in fertilizer
drying are presented in Table 25.  For NO  abatement in nitric acid pro-
duction, only the catalytic reduction process option offers a reduction
of SO  emissions from the base case process.  This is primarily a result
of the large power requirements for the NO  control systems.  On the
contrary, the molecular sieve, CDL/Vitok, and Masar processes for NO
                                                                    X

                                   92

-------
TABLE 25.  ESTIMATED CONTROLLED SOY EMISSIONS  -  FERTILIZER  INDUSTRY
                                  A.
SOX emission factor SOX emissions
(Ib/ton of nitric acid) Change in (106 Ib/yr) ^
Process Process Power boiler Total emission factor 1974 1989-1974
Nitric acid production - With various processes to control NOX emissions
Base

•
•
•
•
•
case:
No NOX control
Catalytic reduction
Molecular sieve
Grand Paroisse
CDL/Vitok
Masar
Converting fertilizer dryers
Base


•



•
case :
Natural gas

Better equipment/
technique with
fuel oil

Installing scrubbers

-.4
-.7
11.1
0.0
31.5
57.6
(with

Nil

0.18



0.14

0.0
.1
.3
0.1
0.3
0.0
baghouses)

Nil

0.0



0.0

-.4 — -3.3** -2.6
-.6 -0.2 — -4.0
11.4 11.8 — 75.2
0.1 0.5 — 0.66
31.8 32.2 — 209.0
57.6 58.0 — 383.0
from natural gas to oil

Nil — Nil Nil
***
0.18 0.18 — 0.29


***
0.14 0.14 — 0.22

                                                                (continued)

-------
                                      TABLE 25 (continued)
*
   Calculated using 4% growth rate for 15 years based on 10 million tons/year in 1974 and rising
   to 18 million tons/year in 1989; thus potential growth of 18 x 106 tons.

**
   Estimated 1974 emissions based on total emission factor (credit) of -0.4 Ib S0x/ton of nitric
   acid are 0.4 x 8.2 x 10" = -3.3 x 10^ ton/yr (SOX emissions credit since energy is recovered
   in this process).

***
   Based on 4% growth rate for 15 years based on 2 million tons/yr in 1974 and rising to 3.6
   million tons in 1989; this potential growth of 1.6 million tons is multiplied by the
   emission factor.

-------
control may significantly increase estimated SO  emissions.
                                               X
     Fuel conversion for fertilizer dryers equipped with baghouse fil-
ters has a small impact on the total amount of sulfur dioxide emissions
on an industry-wide basis since only 20% of the fertilizer plants are
equipped with these air filters.

TEXTILES

     Two textile mills, an integrated knitting mill and an integrated
weaving mill, were examined.  The mill operation includes the knitting
of the greige yarn, or the weaving of greige fabric, and the subsequent
dyeing and finishing of the fabric.

Knit Fabrics Production

Base Case Process—

     In the knitting mills, yarn is knitted into fabric in the greige
mill.  The greige fabric is next scoured  to remove knitting oil, and
is then dyed, washed, and  spin-dried to remove as much water as  pos-
sible before hot-air drying.  A finish  (softener/lubricant) is then ap-
plied to the fabric, which is dried and heat-set.  Details of the knit-
ting mill operation are presented  in Volume IX, page 33.

     It is unlikely that natural gas, used  for the hot-air drying and
heat-set operations, will  be easily replaced, as it  is required  by all
of the equipment presently available for  fabric drying and heat-setting
operations.

     Steam is used  for  the heat input to  the  scouring, dyeing, and wash-
ing operations.  In the base case  process,  it was  assumed  that the steam
would be generated with low-sulfur oil, so  that  flue gas  desulfurization
would not be required.  Since 4 Ib of steam/lb of  fabric  are required,
then about 4.4  Ib  SO /ton  of fabric would be  emitted.  Because of the
relatively small energy requirements of textile mills,  it  seems  unlike-
ly that  coal will  be used  as a fuel for steam production  (Volume IX,
page 41).

     Electricity is required  to provide mechanical energy to transfer
fabric  from  the beginning  to  the  end of the process  line  and for knit-
ting the yarn  into the fabric.  Electrical  requirements of  0.18  kWh/lb
of  fabric will  result  in  SO emissions  of about  4.5  Ib  SO /ton of fab-
ric, if  the  NSPS is met.   Without SO  cleaning of  the  stack gas, SO
emissions would be 20.7  Ib/ton of fabric.

Advanced Aqueous Processing—

     The sequence  of operation is similar to  the base  case except that
 (1)  the  hot  air drier  is  replaced by an air/vacuum extractor,  thereby
reducing natural gas requirements and increasing electrical requirements,


                                    95

-------
and (2) the scouring, dyeing, and washing operations are modernized
with more efficient equipment, thereby reducing steam requirements.
The advanced process is described in detail in Volume IX, pages 35-37.

     Steam requirements of 1.9 Ib/lb of fabric will result in 2.1 Ib
SO /ton of fabric, if the steam is generated by low-sulfur oil so that
flue gas desulfurization is not required.  Electrical requirements of
0.25 kWh/lb of fabric will result in 6.3 Ib SO /ton of fabric, if the
NSPS are met.  Without flue gas desulfurization, SO  emissions would
be 29 Ib SO /ton of fabric.                        X
           x
Solvent Case—

     Solvent systems are assumed for the scouring, dyeing, and finishing
operations.  The fabric is steam-stripped to remove residual solvent.
Clean solvent is recovered by distillation with steam.  Details of a
solvent process are provided in Volume IX, pages 37-41.

     Steam requirements of 1.1 Ib/lb of fabric will result in 1.2 Ib
SO /ton of fabric, if low-sulfur oil is used to generate the steam and
flue gas desulfurization is not required.

     Electricity requirements of 0.12 kWh/lb of fabric will result in
3.0 Ib SO /ton of fabric, if the NSPS are met.  Without SO  removal
from the Boiler flue gas, emissions would be about 13.8 Ib SO /ton of
fabric.  By comparison with current technology, the solvent processing
route can reduce SO  emissions — after control — by some 53%.
                   x                                J

Woven Fabrics Production                            '

Base Case Process—

     The operations of woven fabric preparation, dyeing, and finishing,
involve a much longer processing sequence than knit fabrics.  These
steps are described in Volume IX, page 44.

     Natural gas is used in the drying, setting, and curing operations;
hence there are no SO  emissions in these steps.  Steam  (15 Ib/lb of
fabric) is used for process water heating in the dyeing, washing, and
finishing steps.  Steam generation will result in 16.5 Ib SO /ton of
fabric, if low-sulfur oil is burned and flue gas desulfurization is not
required.  Electrical energy requirements are estimated at 0.29 kWh/lb
of product.  This will result in SO  emissions of 7.3 Ib/ton of fabric,
if the NSPS are met.  Otherwise, uncontrolled emissions are estimated
at 33.6 Ib/ton of fabric.

Advanced Process—

     The advanced processing includes a polyvinyl alcohol (PVA) recov-
ery loop which recycles concentrated PVA solution back to sizing and
the hot water back to the desizing operation.  Details of the advanced

                                   96

-------
case sequence are presented in Volume IX, page 49.

     Steam requirements have been reduced to 6.4 Ib/lb of fabric by re-
duction in overall process water use and recycling of wash waters.  This
steam generation will produce 7.0 Ib SO /ton of fabric.  These emissions
are not controlled.  Electrical requirements have been reduced to 0.15
kWh/lb of steam, which will result in SO  emissions of 3.8 Ib/ton of
                                        X
fabric, if the NSPS are met.  Without control, sulfur dioxide emissions
are estimated at 17.5 Ib/ton of fabric.

Summary
                                   97

-------
                               TABLE 26.   ESTIMATED  SCI  EMISSIONS  -  TEXTILE INDUSTRY
                                                      X
CO


SOV emission factor
SO-^, emissions
(Ib/ton fabric) Change in (10° Ib/yr) A
Process Process Power boiler Total emission factor 1974 1989-1974
Knit fabric

Base case - aqueous 4.4
• Advanced aqueous 2 . 1
• Solvent processing 1.2
Woven fabric

Base case - aqueous 16.5
• Advanced aqueous 7.0


4.5 8.9
6.3 8.4 -.5
3.0 4.2 -4.7


7.3 23.8
3.8 10.8 -13.0

**
2.9 1.1
1.0
0.5

***
50 19 . 3
8.8

     **
     ***
Based on no retirement of existing facilities, incremental production from 1974 to 1989 of
1)  0.12 x 10° ton/yr (knit fabric) multiplied by total emission factor, and 2) 0.81 x 106 ton/yr
(woven fabric) multiplied by total emission factor.


Estimated 1974 emissions from knit fabric production based on the total emission factor of 8.9 Ib
S0x/ton of fabric are 8.9 x 0^32 x 106 = 2.85 x 106 Ib.

Estimated 1974 emissions from woven fabric production based on total emission factor of 23.8 Ib
S0x/ton of fabric are 23.8 x 2.1 x 106 = 50 x 106 Ib/yr.

-------
                              REFERENCES
U.S. Environmental Protection Agency, AP-42,  Compilation of Air Pollution
Emission Factors, Second Edition, March 1975.
                                 99

-------
                                   TECHNICAL REPORT DATA
                            (/'/case read fmumctions on the reverse before completing)
1. REPORT NO.
  EPA-600/7-79-141
                                                            3. RECIPIENT'S ACCESSI Of* NO.
4. TITLE ANDSUSTITLE
 Environmental Considerations  of Selected Energy-
 Conserving Manufacturing Process Options
 Volume  XVI Sulfur Oxides Summary Report	
                                                            5. REPORT DATE
                                                              August 1979  issuing date
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

 Arthur D.  Little, Inc.
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS

 Arthur D.  Little, Inc.
 20 Acorn  Park
 Cambridge,  MA  02140
             10. PROGRAM ELEMENT NO.

               INK 624R	
             11. CONTRACT/GRANT NO.

                 68-03-2198
12. SPONSORING AGENCY NAME AND ADDRESS
 Industrial  Environmental Research Laboratory
 Office  of Research and Development
 U.S. Environmental Protection  Agency
-Cincinnati.  Ohio  45268             	
             13. TYPE OF REPORT AND PERIOD COVERED
                 Final
             14. SPONSORING AGENCY CODE

                             EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
      Arthur  D.  Little, Inc.,  undertook a study of  the "Environmental  Considerations
 of Selected  Energy-Conserving Manufacturing Process  Options."  Some 80  industrial
 process  options were examined in 13 industrial sectors.   Results were published in 15
 volumes,  including a summary, industry prioritization report, and 13  industry oriented
 reports  (EPA-600/7-76-034 a through o).
      This present report summarizes the information  regarding sulfur  oxide pollutants
 in the 13 industry reports.   Four parallel reports treat nitrogen oxides,  particulates,
 solid residues, and toxics/organics.   All of these pollutant-oriented reports are
 intended to  be  closely used with the original 15 reports.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.IDENTIFIERS/OPEN ENDED TERMS
                             COSATI Field/Group
 Energy, Pollution,  Industrial Wastes
 Manufacturing Processes,
 Energy Conservation
                                                                               68D
13. DISTRIBUTION STATEMENT

 RELEASE  TO PUBLIC
19. SECURITY CLASS (This Reportj
 UNCLASSIFIED
21. NO. OF PAGES
     108
                                              20. SECURITY CLASS (Thispage)
                                               UNCLASSIFIED
                                                                          22. PRICE
EPA Form 2220-1 (9-73)
                                           100
                       ft US GOVEKHMENTniMTIIIGOFFICE: 1979-657-060/5404

-------