EPA-600/2-76-269
October 1976
Environmental Protection Technology Series
SOURCE ASSESSMENT:
GLASS CONTAINER MANUFACTURING PLANTS
industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. Environmental Protection Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socibeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed
to develop and demonstrate instrumentation, equipment and
methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U. S. Environmental Protection
Agency, and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
This document is available to the public through the National
Technical Information Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-269
October 1976
SOURCE ASSESSMENT:
GLASS CONTAINER
MANUFACTURING PLANTS
by
J.R. Schorr, Diane T. Hooie,
Philip R. Sticksel, and Clifford Brockway
Battelie-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
Contract No. 68-02-1323. Task 37
ROAP No. 21AFA-013
Program Element No. 1AB015
EPA Task Officer: Edward J. Wooldridge
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
PREFACE
The Industrial Environmental Research Laboratory (IERL) of EPA has
the responsibility for insuring that air pollution control technology is
available for stationary sources. If control technology is unavailable,
inadequate, uneconomical or socially unacceptable, then development of the
needed control techniques is conducted by IERL. Approaches considered
include: process modifications, feedstock modifications, add on control
devices, and complete process substitution. The scale of control technology
programs range from bench to full scale demonstration plants.
The Chemical Processes Section of IERL has the responsibility for
developing control technology for a large number (> 500) of operations in
the chemical and related industries. As in any technical program, the first
step is to identify the unsolved problems.
Each of the industries is to be examined in detail to determine if there
is sufficient potential environmental risk to justify the development of
control technology by IERL. This report contains the data necessary to make
that decision for glass container manufacturing plants.
Battelle's Columbus Laboratories was contracted with EPA to investigate
the environmental impact of the glass container industry, which represents a
source of emissions in accordance with EPA's responsibility as outlined above.
Dr. J. Richard Schorr served as Program Manager for this study. Mr. Edward J.
Wooldridge served as EPA Project Monitor. The study was completed by IERL-RTP.
Project responsibility was transferred to the Industrial Pollution Control
Division of lERL-Cincinnati on October 15, 1975.
iii
-------
iv
-------
TABLE OF CONTENTS
Page
INTRODUCTION . 1
SUMMARY A
DESCRIPTION OF GLASS-CONTAINER INDUSTRY 11
General Process Description. .... 11
Plants and Locations 13
Shipment Volume and Weight . „ 13
Process Details 18
Batch Handling „ 18
Melting and Fining 23
Forming 29
Postforming » 31
EMISSIONS 33
Raw-Materials Preparation and Handling .... 33
Glass Melting. 36
Nitrogen Oxides ........... 39
Sulfur Oxides 40
Particulates 42
Carbon Monoxide 44
Hydrocarbons 45
Selenium 45
Other Emissions 46
Forming and Finishing 46
Forming 46
Surface Treatment 47
v
-------
TART.K OF CONTENTS
(Continued)
Page
Annealing • 48
Decorating. . . . . 48
Emission Characteristics . . . 50
Raw-Materials Preparation „ • 50
Glass Melting . . .... . . . . . 50
Forming and Finishing 50
Ground-Level Concentrations 52
Affected Population „ „ . . . 61
CONTROL TECHNOLOGY. . 64
Raw-Materials Preparation. ..... . 65
Emissions ........ .... 65
Raw-Materials-Control Technology 66
Glass-Melting Operation. 69
Emissions 69
Glass-Melting-Control Technology 71
Efficiency of Equipment 79
Forming and Finishing. . 80
Forming and Finishing Control Technology. ........ 81
Surface Treatment ...... . . 81
Surf ace-Treatment-Control Technology 82
Decorating. ................. 82
Decorating-Control Technology ........ 82
FUTURE GLASS-CONTAINER PRODUCTION 83
UNUSUAL RESULTS 85
REFERENCES 86
vi
-------
TABLE OF CONTENTS
(Continued)
Page
APPENDIX A
Geographical Listing of the 122 Container Glass Plants A-l
APPENDIX B
Emissions Data B-l
APPENDIX C
Stack Heights from the Various Phases of Glassmaking C-l
APPENDIX D
State Listing of Total Emissions as of 1972 D-l
APPENDIX E
Conversion Factors E-l
APPENDIX F
Glossary of Terms F-l
APPENDIX G
Letters of Comment G-l
LIST OF TABLES
TABLE 1. GLASS-INDUSTRY STATISTICS.
TABLE 2. AVERAGE EMISSIONS OF MAJOR SPECIES FROM GLASS-
CONTAINER PLANT 6
TABLE 3. SOURCE SEVERITY FOR GLASS-CONTAINER EMISSIONS 9
TABLE 4. MAJOR GLASS-CONTAINER MANUFACTURERS IN THE
UNITED STATES 14
TABLE 5. PRODUCT SHIPMENTS OF THE GLASS-CONTAINER INDUSTRY 16
TABLE 6. MINOR CONSTITUENTS OF CONTAINER GLASS 23
TABLE 7. PARTICIPATE EMISSIONS DURING RAW-MATERIAL
PREPARATION AND HANDLING FOR CONTAINER GLASS 37
vii
-------
TABLE OF CONTENTS
(Continued)
Page
TABLE 8. EMISSIONS FROM FLINT AND AMBER CONTAINER GLASS-
MELTING-FURNACE-OPERATIONS 38
TABLE 9. EMISSIONS FROM THE ANNEALING OF SODA/LIME
CONTAINER GLASSES.
49
TABLE 10. GENERAL SPECIFICATION LIMITS FOR RAW MATERIALS
USED IN CONTAINER GLASS MANUFACTURE 51
TABLE 11. PARAMETERS FOR GLASS-MELTING FURNACES OF A
REPRESENTATIVE PLANT IN THE GLASS-CONTAINER
INDUSTRY AS USED IN ATMOSPHERIC-DISPERSION
CALCULATIONS 53
TABLE 12. RELATIVE FREQUENCY OF ATMOSPHERIC STABILITIES 57
TABLE 13. MAXIMUM POLLUTANT CONCENTRATIONS AND SOURCE
SEVERITY FOR EMISSIONS FROM THE MELTING FURNACES
FOR A REPRESENTATIVE GLASS CONTAINER PLANT 58
TABLE 14. MAXIMUM AVERAGE GROUND-LEVEL CONCENTRATIONS (X ) OF
AIR POLLUTANTS FROM CONTAINER GLASS-PLANT SOURSf§ BE-
SIDES THE MELTING FURNACE 60
TABLE 15. GLASS-GRADE PARTICLE-SIZE SPECIFICATIONS FOR
SAND, LIMESTONE, AND 10- AND 20-MESH DOLOMITE 6/
TABLE 16. MAXIMUM USE TEMPERATURE FOR VARIOUS FABRIC-
FILTER MATERIALS 75
TABLE 17. GLASS CONTAINER PRODUCTION STATISTICS 84
LIST OF FIGURES
FIGURE 1. PROCESS-FLOW DIAGRAM FOR GLASS CONTAINERS (SIC-3221) . . 12
FIGURE 2. LOCATION OF GLASS-CONTAINER OPERATIONS 15
FIGURE 3. APPROXIMATE BREAKDOWN OF THE TYPES OF CONTAINER GLASS
(1974) 17
FIGURE 4. GLASS-CONTAINER MANUFACTURE 19
FIGURE 5. PROCESS FLOW DIAGRAM OF A TYPICAL BATCH PLANT 20
FIGURE 6. ILLUSTRATIVE SKETCH OF A SIDE-PORT REGENERATIVE-
MELTING FURNACE 25
vlii
-------
TABLE OF CONTENTS
(Continued)
FIGURE 7. LONGITUDINAL SECTION OF A SIDE-PORT REGENERATIVE
FIGURE 8.
FIGURE 9.
FIGURE 10.
FIGURE 11.
FIGURE 12.
FIGURE 13.
FIGURE 14.
FURNACE
TRANSVERSE CROSS SECTION OF FOREHEARTH
CROSS SECTION OF SIDE-PORT REGENERATIVE FURNACE .....
ILLUSTRATION OF THE I.S. BLOW -AND -BLOW PROCESS
FOR FORMING NARROW-NECK CONTAINERS
TYPICAL POINTS OF PARTICULATE EMISSION FROM
RAW-MATERIALS HANDLING
PARTICULATE EMISSIONS SHOWN ARE LINEAR WITH THE
RECIPROCAL OF BRIDGEWALL TEMPERATURE. ....
ILLUSTRATION DEPICTING CALCULATION OF AREA
USE OF COMMERCIAL-COLLECTION EQUIPMENT FOR EMISSION
CONTROL ON BOTH AMBER AND SODA/LIME GLASS FURNACES. . . .
. . 26
. . 26
. . 28
. . 30
. . 34
. . 43
. . 62
. . 74
ix
-------
-------
LIST OF SYMBOLS
Symbol Definition
AAQS Ambient air quality standard
A, B, C, D, E, F Atmospheric stability classes
a, b, c, d, e, f Constants in dispersion equations
A_ The ratio Q/acmi
? 2
BB The ratio -Hz/2c
K
CI Confidence Interval
D. Inside stack diameter
e Natural logarithm base
H Effective stack height
h Physical stack height
AH Plume rise
k "Student t" test variable
m Number of samples
N. Sample value
p Atmospheric pressure
Q Mass emission rate
R Downwind dispersion distance from source
of emission release
s Sample standard deviation
S Source severity, ratio X /AAQS
max
T Ambient temperature
a
T Stack gas temperature
s
t Instantaneous averaging time of 3 minutes
o
t Averaging time
TLV Threshold limit value
u National average wind speed
v Stack gas exit velocity
s
R Horizontal distance from centerline of
dispersion
xi
-------
LIST OF SYMBOLS (Continued)
Symbol
V
IT
CT
X
X
max
max
Definition
Sample mean
3.14
Standard deviation
Downwind ground level concentration at
reference coordinate x and y with emission
height of H
Time average ground level concentration of
an emission
Instantaneous maximum ground level
concentration
Time average maximum ground level
concentration
xii
-------
SECTION I
INTRODUCTION
Air emissions released in the manufacture of glass containers were
examined in this study. This report describes the glass-container industry,
the nature of air emissions and their environmental impact, the control
technology employed, and the future growth of this industry.
Glass containers represents the largest of three segments of
the glass industry (glass containers, flat glass, and pressed-and-blown
glassware). Each segment is defined by a Standard Industrial Classification
(SIC) number, as used by the Department of Commerce. Glass containers
is designated by SIC 3221 and includes the manufacture of glass containers
for food, beverages, medicines, toiletries, and cosmetics. This includes
both narrow-neck and wide-mouth containers. Shipments in this segment have
grown at an average rate of about 3.5 percent since 1971. Industry shipments
in 1973 had a value of $2.3 billion, or about 51 percent of the glass industry
total.
Flat glass is designated by SIC 3211. This includes both the
manufacture of flat glass and some fabrication of flat glass into a tempered-
or laminated-glass product. Flat-glass products include: window glass,
plate glass, wire glass, tempered glass, and laminated glass. These products
are consumed primarily by the automotive and construction industries. Value
of shipments in 1973 was $1.1 billion, which was 24 percent of the whole glass
industry.
Pressed-and-blown glassware is designated by SIC 3229 and includes
all glassware not classified under SIC 3221 or SIC 3211. This industry
segment is very diversified and includes products such as:
• Table, kitchen, art, and novelty glassware
• Lighting and electronic glassware
• Scientific, technical, and other glassware
• Textile fibers.
Industry shipments in 1973 had a value of $1.3 billion, which was about
25 percent of the total industry.
-------
Table 1 gives some 1973 statistics on the three segments of the
glass industry. It shows that over 154,000 people produced merchandise
valued at over $4.5 billion. Some 70 percent of the glass products are made
by the glass-container segment.
Separate Source Assessment Documents were prepared for the flat-
glass and pressed-and-blown glassware segments. This report deals only with
glass containers; however, some of the same emissions and control technology
will also be found in the other glass industry segments. The report
delineates the various emission points, identifies the type and quantity
of emissions, and describes the characteristics of the air pollutants found.
The mass emissions of criteria pollutants (particulates, NO , SO , and CO
X X
and hydrocarbons) from glass container plants are compared with national
emissions from all stationary sources. The maximum average ground-level
concentrations of emissions from a glass-container plant are compared to
the corresponding ambient air-quality standards. Control technology which
is being used or could be applicable to the manufacture of glass containers
is also discussed.
The manufacturing operations for glass containers is grouped into
three categories:
• Preparation of raw materials
• Glass melting
• Forming and finishing.
Emissions and control technology for each of these three areas is
presented.
-------
TABLE 1. GLASS-INDUSTRY STATISTICS
(a)
1973 1973
Employees, Value of Shipments,
SIC Industry Segment (103) ($ Million)
3221 Glass Containers 77.8 2,316
3211 Flat Glass 26.3 1,118
fc)
3229 Pressed -and -Blown Glass 50.0 1,120
Total 154.1 4,554
, 1973 ,,.
Shipments *• \
(Metric Tons x 10 )
11.32
3.12
1.57
16.01
(a) Source: Department of Commerce and References 1 and 15.
(b) Metric Ton = 1012 grams.
(c) Excludes textile fibers.
-------
SECTION II
SUMMARY
This document describes a study of air emissions released during
the production of glass containers. The industry is defined by Standard
Industrial Classification (SIC) No. 3221. It encompasses the preparation
of raw materials at the plant site, the production of molten glass in a
furnace, the forming of glass containers, and certain post-forming operations
required to manufacture these products.
The glass-container industry in the United States produced 11.005 Tg
(12.133 million ton) of salable product in 1974. Of that total, about
15 percent (1.651 Tg) was amber glass; the remainder (green and clear
glass) is classified as flint glass. In 1974, the domestic glass-container
industry consisted of 25 manufacturers operating 120 plants. Geographically,
these plants are located near the local markets they serve, with the largest
concentration being in the East, North-Central, and Middle Atlantic regions
of the country. Glass-container plants are located in 29 states; California, Illinoi
Indiana, Ohio, Pennsylvania, and New Jersey are states with the largest
number of manufacturers. The average county population density at a plant
2
site is estimated to be 356 people/km .
Manufacturing Technology
In a glass-manufacturing process, raw materials (e.g., sand, soda
ash, limestone) are uniformly mixed and these loose materials transported
to a furnace where they are melted at elevated temperatures (>1500 C) into
a homogeneous mass. More than 95 percent of the glass is made in fossil
fuel-fired furnaces where energy is predominately transferred to the glass
by radiation from a flame or reradiation from the refractory chamber containing
molten glass. Molten glass is kept at elevated temperatures until it is
of a quality (bubble-free) sufficient for making containers.
12
* Tg = 10 gram. Metric prefixes and other conversion
factors are given in Appendix F, Page F-l.
-------
The glass is then cooled to approximately 1300 C, removed from the furnace, and
cut into "gobs". The gobs are fed to a machine and formed into containers.
Approximately 30 percent of these containers then undergo surface treatment
and about 3 percent are decorated. The exact operations used in a specific
plant depend upon the type of product desired. All containers go through a
gas-fired annealing furnace for removal of residual stresses. Highest
temperatures during annealing range from about 590 to 650 C.
Emissions
Emissions were examined from three areas within the glass-container
manufacturing plant: (1) raw-materials preparation and handling, (2) glass
melting, and (3) forming and finishing. The largest emissions occur from
the glass-melting operation.
Manufacturing Plant
Table 2 summarizes the average emissions of major species from a
glass-container manufacturing plant determined by this study. While emissions
for amber and flint glass are discussed separately in the report, significant
differences were not found, and the data in Table 2 is for a plant producing
flint glass. The emissions listed are for an average plant production
capacity of 319 Mg/day (352 ton/day). Annual production was 105 Gg (90 percent
of capacity). This table shows that over 97 percent of the plant emissions
come from the glass-melting furnace. The major species (over 90 percent) are
NO , SO , and particulates. Furnace stack heights average 26 m (65 ft) when
X X
ejection air is used and 45 m (147 ft) for natural draft.
Total Industry
Nitrogen oxides have the highest emission factor (3.07 g/kg + 47%). This
includes both flint and amber glass as shown later in Table 8. Total annual
emissions are 38.9 Gg. Accordingly, NO contributes the greatest amount to the
X
national NO emissions from all stationary sources (0.34 percent).
x
-------
(a)
TABLE 2. AVERAGE EMISSIONS OF MAJOR SPECIES FROM GLASS-CONTAINER PLANTk
Emission Species
NOX
sox
Particulates CO
Hydrocarbon
Selenium
HC1
EMISSION FACTOR, g/kg
(1) Raw Materials
(2) Glass Melting
(3) Form and Finish
0 0 0.03 + 100% 0 0
3.40 + 43% 1.84 + 36% 0.71 + 30% 0.06 + 170% 0.08 + 100% 0.002 + 100%
0.02 + 100% 0 0.05 + 100% 0.002 + 100% 4.43 ± 100% 0.02 + 100%
TOTAL ANNUAL EMISSIONS . Mg
(1)
(2)
(3)
Raw Materials
Glass Melting
Form and Finish
0
358
2
0
193
0
4
86
2
(a) Production 319 Mg/day (352 ton/day), assuming 85 percent pack rate of flint glass or 104.8 Gg/yr.
(2)
(b) Represents primarily surface treatment emissions for 30 percent of production.
(c) Represents decorating emissions for 3 percent of production (42%) and emissions from annealing lehrs (29%)
and forming operations (29%).
(d) Emission factors are only for flint glass, which constitutes an estimated 85 percent of glass production.
Emission factors for both flint and amber are given later in Table 8.
-------
Sulfur oxides have the next highest emission factor which is
1.70 + 47%. This includes both flint and amber as given in Table 8.
Total annual emissions are 21.6 Gg. This amounts to 0.31 percent of the
national emissions from all stationary sources.
Particulates for both flint and amber glass were determined to have
an emission factor of 0.68 g/kg + 36%, with total annual emissions of 8.6 Gg.
This would be 0.007 percent of national emissions from all stationary sources.
Carbon monoxide has an emission factor of 0.07 g/kg + 143% with
total annual emissions of 0.9 Gg. The emission factor for hydrocarbons
was 0,08 g/kg + 100% with total annual emissions of 0.7 Gg, These contribute
0.005 and 0.003 percent respectively to the national emissions from all
stationary sources. Finally, the emission factor for selenium was
2 mg ± 100% with total annual emissions of 0.02 Gg.
Emissions from raw-materials preparation and handling do give
rise to some particulate emissions, primarily from dust generated-'during
discharging, conveying, crushing, and mixing operations. Composition of the
emissions is the same as that of the raw materials (i.e., sand, limestone,
soda ash). The average emission factor is 0.03 g/kg Hh 100%. Total annual
emissions for the glass container industry were estimated to be only
459 Mg or 0.0003 percent of the national particulate emissions from all
stationary sources. Over 90 percent of the industry employs controls (primarily
filter bags) in dust generating areas.
Many different processes can be used in the forming and finishing
operations. Emissions of the major species from forming and finishing are relatively
low, as can be seen in Table 2. These emissions consist of hydrocarbons
emitted from the forming operation (0.03 g/kg), tin oxide and
hydrated tin chloride particulates and HC1 emitted from a surface treatment
operation performed on about 30 percent of the containers produced. Combustion
products are emitted from gas-fired annealing and decorating lehrs, and
hydrocarbons are emitted from decorating operations used with about 3 percent
of the glass container production. The emissions from all of these areas have
been combined in Table 2, because relatively speaking they constitute less
than 2 percent of the emissions from a-normal manufacturing plant. They
are broken out by emission source and type in the body of the report.
-------
Control Technology
Emissions from glass container melting furnaces are generally not
controlled by add-on equipment. This is not the case in every State.
Frequently emission standards can be met with proper operating conditions.
Baghouses and electrostatic precipitators are used by a few manufacturers
to control submicron particulate emissions.
Source Severity
Impacts of these emissions are directly related to the ambient
concentrations the emissions create at ground level. Atmospheric dispersion
calculations were made to calculate maximum average ground-level concentrations
(X ) for the emissions from an average plant producing 105 Gg/year. Results
in 3.x
of these calculations are presented in Table 3.
Source severity factor, S, has also been used to describe the impact
of the emissions. For those pollutants which have an ambient air-quality
standard (AAQS), S is the ratio of X to the primary AAQS. In cases where
max
no AAQS has been established, S is based upon the Threshold Limit Value (TLV)
through the following equation which includes a factor for correcting the
TLV to a 24-hour day (8/24) and a safety factor (1/100).
X
max
TLV (8/24) (1/100)
Results of the source severity factor calculations also appear in Table 3.
The highest severity factor (0.38) is produced by nitrogen oxide emissions
from the melting furnace and hydrocarbon emissions from decorating lehrs (0.15).
Other severity factors were less than 0.05.
A third measure of the impact of the plant emissions is given by
the affected population. This measure is defined as the population around
the representative plant who are exposed to a source severity factor greater
than 1.0. Computations of the affected population showed that all sources
were less than 1.0. Nitrogen oxides would be the highest if computed on
a more restrictive basis such as a source severity factor of 0.1. In this
case 11,700 persons would be affected.
-------
TABLE 3. SOURCE SEVERITY FOR GLASS-CONTAINER EMISSIONS
Source —
Pollutant
Melting
Furnace —
NOX
SOX
Particulates
CO
Hydrocarbons
Selenium
Materials
Handling--
Particulates
Container
Decorating —
Hydrocarbons
Surface Treatment -
HCL
Primary Ambient
Air-Quality Standard
Averaging
Hg/m3 Time, hr
100
365
260 ,
4 x 104
160
Trace
260
160
(b)
Trace
24
24
24
8
3
24
24
3
24
Maximum Average
Ground - Le ve 1_
Concentration, max
M-g/m3
37.9
20.5
9.2
0.8
°'9 /^
Trace (b)
0,02
23
(b)
Trace
Severity
Factor
0.38
0.056
0.035
2 x 10"^
5.9,x 10
c
9.2 x 10
0.15
(a) Other severity factors, including those for selenium, tin particulates, and
HC1 were quite low.
(b) Trace <10 yg/m3.
-------
10
Future Growth
Historically, the glass container industry's growth has fluctuated
considerably in the past 8 years, but shipments have steadily increased since
1967 at an average annual rate of 6 percent. A large portion of this growth
is attributable to the popularity of the nonreturnable beverage bottle. In more
recent years, growth in shipments has been less. Future growth may well be
tied to legislation restricting use of nonreturnable containers. It is likely
that 1980 production levels will be 20 percent higher than for 1974. Emissions
would increase proportionally and possibly at an increased rate, since
the industry is moving away from the use of natural gas to the use of oil.
The actual effect of conversion from gas to oil firing on emission rates
is not known.
-------
11
SECTION III
DESCRIPTION OF GLASS-CONTAINER INDUSTRY
This section describes in general terms the process steps used
in the manufacture of glass containers and presents certain statistical
information pertinent to the glass-container industry described by the
Department of Commerce for Standard Industrial Classification (SIC) 3221.
General Process Description
Figure 1 is a process-flow diagram which generally depicts the
flow of materials through a glass container manufacturing plant. The process
is categorized into four steps: batch handling, melting and fining, forming,
and postforming. These specific steps are discussed in more detail later
in this section of the report.
Basically, the manufacture of glass containers consists of melting
(and reacting) a mixture of raw materials (consisting primarily of silica,
soda, and lime) which have been prepared in the batch-handling step so
as to minimize segregation and impurities in the batch. Cullet (scrap
glass) is also added at this stage. In the glass melter, the materials are
melted down, the molten glass is fined (i.e., residual trapped gases are
removed), and then the temperature of the glass is lowered so that it can be
handled in the forming step. The glass passes from the melter to the forming
equipment via the forehearth, a relatively shallow, narrow refractory
channel having a refractory roof and individual heating and cooling systems
for controlling glass temperature. Individual gobs of glass are fed to the
forming machines where the molten glass is transformed into a product by one
of two methods: blow and blow, or press and blow. The formed container
may now go through a series of postforming steps, depending upon the product
desired, but which always includes annealing, where stresses are removed
through a controlled, uniform-cooling cycle. Finally, the containers may
undergo various additional steps, such as surface treatment, inspection, testing,
decoration, and plastic coating.
-------
12
Raw
materials
receipt
and
storage
Batch
weighing
and
mixing
Crushed
glass
cullett
Batch
charg-
ing
Scratch
resistant
surface
treatments
Lubricity
surface
treatments
Inspection
and
testing
Decoration
and/or
plastic
coating
Pack
Ll
o
m
CD
O>
1
o
TO
FIGURE 1. PROCESS-FLOW DIAGRAM FOR GLASS CONTAINERS (SIC-3221)
-------
13
Plants and Locations
According to information gathered from the Department of
(1)*
Commerce 1972 Census of Manufacturers and from the Glass Containers
(2)
Manufacturers Institute (GCMI)v ', there were 120 establishments manu-
facturing glass containers in the United States. These 120 plants (see
Appendix A) are operated by 25 manufacturers as shown by Table 4.
Statistics obtained from 1975 glass-industry directories and from industry
(2-5)
sources indicate that approximately half of these plants are operated
by the five largest companies.
Geographically, glass-container plants are located near the local
markets they serve. As such, plants are found throughout the United States,
but a large number are concentrated in the East North Central and Middle
Atlantic portion of the United States. The states containing the largest
number of manufacturers are California, Illinois, Indiana, Ohio, Pennsylvania,
and New Jersey. The regional distribution of the ma^or plants is shown in
Figure 2.
Shigment Volume and Weight
Table 5 provides estimated 1974 output data for the glass-container
industry (SIC 3221). The volume of shipments for 1974 was 276,382 thousand
gross as compared to 267,732 thousand gross in 1972, or only 3.2 percent
growth for the 2-year period. The weight of glass containers shipped
increased from 10,772 Gg (23,748 million pounds) in 1972 to 11,005 Gg
(24,266 million pounds) in 1974, or an increase of 2.1 percent. Three general
types of container glass are produced: amber, green, and clear. For this
report, green and clear glass are considered together as a single category
designated as flint. The major difference between these two is the iron
oxide additions. Emission data was not found to be separately reoorted.
Figure 3 illustrates the estimated breakdown between amber and flint.
* References are listed on page 86.
-------
14
TABLE 4. MAJOR GLASS-CONTAINER MANUFACTURERS IN THE UNITED STATES
(a)
Manufacturer
No.
of Plants
Manufacturer
No.
of Plants
Anchor Hocking
Arkansas Glass
Ball Corporation
Bartlett Collins
Brockway Glass
Chattanooga Glass
Columbia Gas
Diamond Glass
Gallo Glass
Glass Containers Corp.
Glenshaw Glass
Hillsboro Glass
Indian Head
Industrial Glass
Kerr Glass
9
1
4
1
14
7
1
1
1
12
2
1
7
1
7
Latchford Glass
Leone Industries
Liberty Glass
Madera Glass Co.
Metro Glass
Midland Glass
National Bottle Corp,
National Can
Owens"Illinois
Thatcher Glass
Underwood Glass
Wheaton Glass
Total
1
2
1
1
4
4
4
4
20
6
1
120
(a) Source: Material provided from GCMI,
dated 10/24/75.
-------
FIGURE 2. LOCATION OF GLASS-CONTAINER OPERATIONS(6)
-------
16
TABLE 5. PRODUCT SHIPMENTS OF THE GLASS-CONTAINER INDUSTRY
(Department of Commerce Classification SIC-3221)
(a)
Shipments of All Types of Glass Containers
1974 1973 1972
Number(b) Weight. Gg Number (b_) Weight.Gg Number Q*). Weight,Gg
276,382 11,000 276,328 11,326 267,732 10,748
(a) Sources: U.S. Department of Commerce, Bureau
of the Census, Series M32G(74)-13 and
Statistics Glass Containers 1975 pub-
lished by The Glass Container Manufac-
turers Institute.
(b) Number of containers in thousands of gross.
-------
17
15 % Amber
1,650 Gg
FIGURE 3. APPROXIMATE BREAKDOWN OF THE TYPES OF CONTAINER GLASS (1974)^15^
-------
18
Process Details
While specific equipment will vary, depending upon the manufacturer
.and the product being made, the basic manufacturing process is essentially
the same for all glass-container manufacturers. Portions of raw materials
are mixed with each other and with cullet (scrap glass), conveyed to a melting
furnace, melted, fined, conditioned, and fed into forming machines. The
formed containers are then taken through a variety of postforming and product-
handling steps.
The subsequent paragraphs describe the process operations and
the raw material ingredients; thus indicating the potential sources of
materials that can be emitted into the air as pollutants. Figure 4
schematically shows the overall manufacturing process.
Batch Handling
The function of the batch-handling operation is to prepare and
feed to the melting furnace a batch which is both chemically and physically
uniform in composition. Control of the composition, impurity level, and
the size and moisture of the raw materials is important. Cullet is crushed
and either mixed with the raw materials or added later. Each of the raw
materials is carefully weighed, mixed together, and then conveyed to the
batch chargers. Care must be taken so that segregation of a uniformly
mixed batch does not occur.
A large plant manufacturing container glass usually houses the
raw material mixing and conveying equipment in a structure termed a "batch
house", A flow diagram of a typical batch house is shown in Figure 5. In
most batch houses, the storage bins are located on top, with the weigh hoppers
and mixers located below them to make use of gravity flow.
Major raw materials and cullet (broken scrap glass) are conveyed
from railroad hopper cars or hopper trucks by a combination of screw
conveyors, belt conveyors, and bucket elevators, or by pneumatic conveyors
to the elevated storage bins. Minor ingredients are usually delivered to the
plant in paper bags or cardboard drums and transferred by hand to small bins.
-------
19
Glass sand
Si02 5 99*
to yield Si02
crushed, washed
and screened
to ~ 20 -100
mesh
Soda ash
Na2C03
to yield Na20
~ 20-120 mesh
or granular
Limestone
or burnt lime
to yield CaO.
Usually some
MgO also results
~ 20-120 mesh
Feldspar
R2O.AI203.6SI02
to yield
AI20,,Si02
Na20 and K20
pulverized or
granular
Other additions
for K20, MgO,
ZnO, BaO, PbO,
etc and those for
fining, oxidizing,
coloring, and
decolorizing
Side-port
continuous tank,
looking down
through top
Packing, warehousing,
and shipping
FIGURE 4. GLASS-CONTAINER MANUFACTURE
C7)
-------
20
GULLET
RAW MATERIALS
RECEIVING
HOPPER
V
SCREW
CONVEYOR
FILTER
VENTS
STORAGE BINS
MAJOR RAW MATERIALS
MINOR
INGREDIENT
STORAGE
BINS
TO
ATM
BELT CONVEYOR
(I
BATCH
STORAGE
BIN
FURNACE
FEEDER
FIGURE 5. PROCESS FLOW DIAGRAM OF A TYPICAL BATCH PLANT
-------
21
Materials are gravity fed from the storage bins into weigh hoppers
and then transported by transverse belts or bucket elevators into a mixer.
Gullet is crushed to a desired size (usually between 0.5 and 2.0 cm). After
mixing, the glass batch is transferred to a charging bin located next to
the glass-melting furnace or into a batch-storage bin, depending upon the
design of the batch-handling system. Positive displacement or vibrator
feeders at the bottom of the bins feed the materials to the glass-melting
furnace chargers. Gullet may be added to the batch in the mixer, while the
fO\
batch is being transferred, or charged separately to the melting furnace .
Batch is fed into the melter in either a dry or moist state.
Many companies add two to four percent water to the dry batch to help
prevent segregation during transport of the batch, to minimize dust problems
in the melter, and to avoid carryover of dust into the regenerators.
The various handling and mixing operations are a source of
particulate emissions which are similar (same materials, same processes)
as those in other industries ' . Because of environmental and economic
incentives, most large manufacturers practice dust control, usually by
means of cloth filters and baghouses .
Batch Composition. The basic raw materials for soda/lime container
glass are silica sand, soda ash (Na^CO.,) and limestone (primarily CaCO_, plus
some MgCO., in dolomitic limestones). Feldspar minerals are also utilized as
a source of alumina and alkali. Minor amounts of other oxides are introduced
as impurities and additional minor ingredients are added for specific
purposes discussed later.
Glass sand must be of high purity (& 99 percent Si02). Primary
impurities are Fe203 and A1203 which together will be less than 1 percent.
The chief sources are unconsolidated bank sand from New Jersey and the
standstones of the Alleghenies and the Mississippi valley.
The U.S. supply of soda ash, Na CO , has changed from predominantly
synthetic to natural in recent years. Prior to 1973, more than 60 percent of
the domestic soda ash was produced from NaCl by the Solvay process. Now
more than half comes from the natural deposits of trona ores (sodium
sesquicarbonate, Na2C03NaHC03.2H20). The primary deposits are in Wyoming,
but it is also found near Searles Lake and Owens Lake, California.
-------
22
Limestone is the source of calcium (and magnesium) oxides in
the glass batch. This rock has widespread occurrence as either a high-calcium
limestone consisting essentially of calcite, CaC03, or as a dolomitic
limestone which is a mixture of dolomite (CaCO^MgCO.^ and calcite. Good
limestones contain less than 0.1 percent ?&2°3 and about * percent of
silica and alumina. Calcite limestone deposits occur in the central,
southern, and eastern U.S. Large dolomitic deposits occur in the central-
midwestern parts of the U.S.
Feldspars are anhydrous aluminosilicates containing potassium,
sodium, and calcium in varying ratios. They are present in virtually all
igneous rocks; but most production comes from pegmatites which are coarsely
crystalline rocks formed in the later stages of crystallization of a magma.
The principal accessory minerals are quartz, mica, and other silicates.
The most critical requirement for glass feldspar is low iron content.
(2)
A typical glass-batch composition is :
Silica sand 909 Kg (2000 lb) (55.6%)
Soda ash 306 Kg ( 674 lb) (18.7%)
Feldspar 118 Kg ( 260 lb) (7.2%)
Limestone 294 Kg ( 648 lb) (18.0%)
Salt cake (Na2S04) 6.8 Kg ( 15 lb) (0.4%)
Total 1634 Kg (3597 lb) (99.9%)
Typically, the above ingredients would melt down to 1370 Kg
(3020 lb) of glass and give off 259 Kg (569 lb) of gases, primarily (> 99%)
C02< The batch volume of 1.2 m (42 cu ft) would produce 0.6 m3 (21 cu ft) of
fluid glass and 858 m (30,300 cu ft) of gaseous products (measured at the
C2\
furnace temperature of 1500 C)v '.
The batch will contain minor ingredients such as salt cake (sodium
sulfate) and various fining, coloring, or decolorizing agents. These
compounds rarely exceed 5 percent and are often less than 0.1 percent of
the total glass composition. Table 6 lists minor ingredients and their
effect on the glass.
-------
23
TABLE 6. MINOR CONSTITUENTS OF CONTAINER GLASS(2'7)
Purpose Effected By
Amber color FeS (pyrites)
X
Green color Cr2°3» Fe2°3» Cu°
Blue color CoO, FeO, CuO
2_
Decolorization, i.e., MnO, Se , NiO, Co,0, CeO,
Mask Fe203 color J 4 "
2_
Fining SO. , C
Gullet (scrap glass) collected from the plant, purchased on the
open market or from recycling centers, is crushed to below 2 cm in size.
It is blended with the raw batch in varying amounts. Normally, however,
only about 15 to 20 percent of the batch going into a furnace is cullet.
If a plant is producing both flint (clear) and colored containers, the cullet
must be individually collected and stored to prevent undesirable color
fluctuations and to avoid glass foaming conditions caused by mixing amber
and flint glass having different oxidation states.
Melting and Fining
The mixed batch is fed into a large continuous-melting furnace
where the batch is melted, fined, and conditioned. The melting furnace
consists of three distinct regions: the melting end, refiner (conditioning
chamber), and forehearth. At the doghouse (batch feeding end of the melting
compartment), the raw materials are fed onto a molten mass of glass having
a temperature near 1500 C. The batch materials react, melt, and disappear
into the liquid glass after floating about one-third to one-half of the length
of the melter. As the molten glass moves on through the furnace tank, it is
fined, trapped gases (bubbles) are removed, and the melt homogenized.
-------
24
The glass is essentially free from bubbles when it reaches the end of
the melting chamber. Then the glass passes through a submerged refractory
throat into a conditioning chamber, popularly called the refiner,
where it is cooled to approximately 1300 c< A refractory bridge wall above
the throat prevents any glass surface scum from passing into the refiner and
also acts as a heat barrier. In the refiner, the glass is cooled to
increase its viscosity to the proper working level and to dissolve any
remaining tiny bubbles or gaseous inclusions. Then the glass flows through
shallow, refractory-lined channels, called forehearths, to the forming machines.
In the forehearth, a uniform temperature of the molten glass is maintained.
This will be near 1100 C and is adjusted by individual heating and cooling
systems. At the end of the forehearth, a stream of molten glass is cut into
individual gobs of glass and fed to the forming machine.
Figure 6 is a simplified illustration of a side-port regenerative
melting furnace and forehearth. A longitudinal section through the melter
and refiner of such a furnace is shown in Figure 7. A transverse section
through a forehearth showing the design capability for adjusting glass
temperature as it moves to the feeder is shown in Figure 8.
Characteristic dimensions of container glass melting furnaces
fall within the following ranges: length 6 to 18 meters (20 to 60 ft),
width 3 to 8 meters (10 to 25 feet), and depth 0.6 to 2 meters (2 to 6 feet).
The molten glass holding capacity ranges from 36 to 454 Mg (40 to 500 tons),
and output 27 to 317 Mg/day (30 to 350 tons/day).
Melting Energy Sources. The glass-container industry predominantly
uses regenerative furnaces of side-port or endrport design burning a fossil
fuel. In the side-port design commonly used for larger furnaces (> 175 ton/day),
combustion products and flames pass in one direction across the top of the
molten glass and exhaust through ports on the opposite side of the furnace.
In the end-port configuration, combustion products, and flame travel in a
horizontal U-shaped path across the surface of the glass within the melter. Fuel
and air mix and ignite at one port and discharge through a second port adjacent
-------
25
RAW MATERIALS
REGENERATOR
FEEDER
FORMING
ANNEALING
FIGURE 6. ILLUSTRATIVE SKETCH OF A SIDE-PORT REGENERATIVE-MELTING FURNACE
-------
26
Refiner End
Melting End
Bridge Wall Crown
Port
Doghouse
Glass
0
Throat
Bottom
FIGURE 7. LONGITUDINAL SECTION OF A SIDE-PORT REGENERATIVE FURNACE
(315)
Cooling Air
Burner-
Refractory
Cooling Air
Burner
FIGURE 8. TRANSVERSE CROSS SECTION OF FOREHEARTH^15^
-------
27
to the first on the same end wall of the furnace. To conserve fuel, the
regenerative-firing system is used which employs dual chambers filled with
brick checkerwork. While the products of combustion from the melter pass
through and heat one chamber, combustion air is preheated in the opposite
chamber. The functions of each chamber are interchanged periodically.
Reversals occur about every 15 to 20 minutes as required for maximum
conservation of heat. Figure 9 is a cross section of a side-port regenerative
furnace. The regenerators are about two stores tall and are positioned on
each side of "side-port" furnaces or at one end of "end-port" furnaces.
The glass container industry has historically used natural gas
as the primary energy source. In 1971, over 83 percent of the energy was
derived from the combustion of natural gas . In more recent years, the
limited availability of natural gas has resulted in a shift to greater use
of fuel oil. In 1975, the-energy supplied by natural gas was between 70
(2)
and 75 percent .
Since molten glass is an electrolyte, it can also be heated
electrically. Such heating is used primarily to supplement fossil-fuel
heating, rather than replace it and is referred to as electric boosting. One
estimate indicated that approximately 40 percent of all furnaces now
have electric boosters. Less than 5 percent of container glass furnaces in
the U.S. were heated entirely by electricity in 1974
Future energy sources could vary considerably from the present-day
pattern depending upon the availability and cost of fossil fuels.
Me11ing-Furnace Emissions. Exhaust gases from the melting furnace
are major sources of air pollutants from the glass container manufacturing
process. Emissions primarily include NO . SO , and particulates. Nitrogen
X X
oxides (predominately NO) are formed by the reaction of atmospheric nitrogen and
oxygen under the higher temperature conditions of the furnace. Sulfur oxides
(predominately S0?) come from the decomposition of sulfates in the melt and,
where fuel oil is used for heating, from the combustion of sulfur in the fuel.
Particulate emissions may arise from: (1) volatilization of materials in the
melt which subsequently condense in the checkers or stack, or (2) to some extent by
-------
28
Crown
Port
FIGURE 9. CROSS SECTION OF SIDE-PORT REGENERATIVE FURNACE
-------
29
physical etitrainment of batch dust in combustion gases^ . Batch constituents
which can contribute to volatilization are predominately alkali sulfates and
other minor constituents such as selenium.
Forming
The forming of glass containers by industry's highly mechanized
equipment involves several steps.
• The molten glass is cut into gobs by a set of shears
as the glass leaves the forehearth.
• Delivery equipment directs the gobs into blank molds.
• In the blank mold, the gob is partially shaped into a
parison. This task is performed predominantly by an air-
blowing process for narrow-neck containers, but it may
also be performed by pressing. For wide-neck containers,
it is performed by pressing.
• The parison is then inverted and transferred into a
blow mold.
• The parison is blown into final shape with compressed air.
The process of forming narrow-neck containers in which the parison
is formed by blowing steps is illustrated in Figure 10.
Most glass containers are formed on individual section (I.S.)
machines which may be designed with up to 10 sections per forming machine
and may have maximum speed capabilities exceeding 200 containers per minute.
In the operation of the forming equipment, coolants and lubricants
are employed. Water-based sprays are used to cool the shears which cut the
glass gobs. Delivery equipment which directs glass gobs into the blank molds
may be sprayed with lubricating emulsions of hydrocarbon- or silicone-based
materials, and metal molds are commonly lubricated with mixes of graphite,
greases, oils, etc.
-------
30
Delivery
Settle blow
Counter blow
/
n
/
^
/
*
,
?
\
| risn^ZaH
Transfer from blank mold to blow mold
Reheat
Final blow
Takeout
FIGURE 10. ILLUSTRATION OF THE I.S. BLOW-AM)-BLOW PROCESS
FOR FORMING NARROW-NECK CONTAINERSW
-------
31
Postforming
This step can consist of many operations, depending on the product
requirements. These include surface treatments, annealing, decorating,
and coating.
Surface treatment with titanium or tin chloride may be applied
to hot containers as they are transported along conveyors between the forming
machine and annealing oven. The metal chloride reacts to yield titanium or
tin oxide on the glass surface and releases HC1. The surface oxide greatly
increases the container scratch resistance and thus produces containers of
higher service strength.
The primary step, through which all newly formed containers pass,
is annealing. Annealing is necessary to remove stresses that will weaken
the glass or cause it to fail. In this step, the entire piece of glassware
is brought to a uniform temperature high enough to permit the release of
internal stresses (590 to 650 C) and then cooled at a uniform rate to prevent
new stresses from developing. Annealing is done in long continuous furnaces
called lehrs.
Following annealing, but while the containers are still hot (about
100 to 150 C) and on the lehr conveying belt, they may receive additional
surface treatment. Water-based emulsions of polymer or organic materials are
most commonly sprayed onto the hot bottles. The latent heat of the bottles
evaporates the water and bonds the organic which yields a lubricious container
surface. This surface lubricity minimizes or prevents jamming of bottles
as they move along in line conveyors and reduces surface damage.
After annealing (and surface treating, if applied), containers are
visually and optically/mechanically inspected. In addition, cylindrical
bottles to contain pressurized products may be "proof tested" by applying an
internal air pressure, and/or passing them between rubber belts or rollers
which squeeze the bottle. Carbonated beverage bottles are also statistically
sampled and subjected to destructive-burst tests.
Decorations or labels may also be applied to containers in the
production plant. For beverage bottles, screen-printing processes may be
employed for some bottles to apply organic resins or vitrifiable glass
colors, although other techniques and materials may be used. Only a small
portion, about 3 percent, are decorated.(2)
-------
32
Containers, especially for carbonated beverages, may also be coated
with plastic. Such coatings or sheaths may serve one or more functions.
They may reduce bottle-to-bottle handling damage, reduce filling-line noise,
serve as labels, or provide containment in the event of fracture of a
pressurized beverage container. These coatings reportedly may be applied by:
wrapping and heat-shrink techniques, electrostatic-powder applications, dip-coating,
(12)
or other methods . Only a small percentage (< 5%) of containers are
currently plastic coated, but this percentage may increase in the future.
-------
33
SECTION IV
EMISSIONS
Emissions from glass-container plants are categorized according to
three operations within the manufacturing process
• Particulate emissions from the raw-materials handling,
preparation, and transfer.
• Gaseous and particulate emissions from the glass-melting
furnace. These contain SO , NO , submicron sodium sulfate
_ • XX
condensates, hydrocarbons, CO, and other minor emissions
such as selenium.
• Gaseous and particulate emissions from a variety of forming
and postforming operations. These result from annealing,
decorating, surface treatment, and coating operations and can
include particulates (e.g., Sn09), hydrocarbons, NO , and SO .
fc> X X
This section describes the various emissions, characteristics, their levels,
total quantities, and environmental effects. The information is organized
according to the three sources within the manufacturing process.
• Raw-materials preparation and handling
• Glass melting
• Forming and finishing.
Raw-Materials Preparation and Handling
Typical points of particulate emissions during raw-materials
preparation and handling are shown in Figure 11. These points include
• Unloading and conveying
• Crushing of cullet (scrap glass)
• Filling and emptying of storage bins
• Weighing and mixing of batch
• Feeding of batch to glass furnace (batch charging).
-------
34
Conveying to storage
bins by
Acme
Minerals
• Screw conveyors
• Belt conveyors
• Bucket elevators
• Pneumatic conveyors
Gravity-fed into
weight hoppers
Transported by
• Transverse belts
« Bucket elevators
Transferred to
charging bin
Batch charging
Transferred
to charging
bin
Gullet is ground
Unloading of cullet
*
Transferred to storage bin
\
Storage bin
t
Transferred to grinder
FIGURE 11. TYPICAL POINTS OF PARTICULATE EMISSION
FROM RAW-MATERIALS HANDLING
-------
35
All of these are potential sources of particulate emissions;
however, those participates which remain in the manufacturing plant may
constitute an OSHA health and safety consideration distinct from plant
emissions. For the purposes of this study, fugitive-dust emission
was defined as particulate. emissions that result from industrial-related
operations, and which escape to the atmosphere through windows, doors, vents,
etc., but not through a primary exhaust system, such as a stack flue, or
control system. This definition is derived in part from a paper presented by
Lillis and Young of the U. S. Environmental Protection Agency .
Information obtained from ambient sampling up-wind and down-wind of the
manufacturing facility is the preferred source of data. If actual data
from high-volume samples are not available, engineering estimates based on
the particle size of raw material which can contribute to dust emissions are
desired. For the purposes of this study, the particle size range of the
raw materials which will be considered as contributing to fugitive dust emissions
was 100 microns or less in diameter. This definition of dust is not as broad
as the technical definition given by Stern , but this particle-size
distribution seems reasonable for glass-manufacturing processes. The settling
velocity of a 100-micron-diameter sphere, with a specific gravity of 2.0 g/cc,
is approximately 50 cm/sec in still air at 25 C and 1 atm. Such a settling
rate is sufficiently slow that the emission of dust from a tall source,
such as the raw materials storage bins, would probably contribute to the
total air emissions.
Fugitive dust is usually emitted during unloading and conveying
operations. To minimize dust emissions, these operations are generally
enclosed and the vents on storage bins and mixers exhausted through fabric
filters. Batch wetting, or the addition of water to the batch during the
mixing operation, is another practice commonly used which minimizes particulate
emissions. As a result, limited data on particulate emissions from stacks
are available and no data were available on fugitive dust. Particulate-
(14)
emission data from point-source measurements were reported in NEDS
(National Emission Data System) and are given in Appendix B. Although these
-------
36
particulate emission data were not broken down into specific ingredients, these
data enable the calculation of overall average emission factors for raw-
materials handling and preparation on a worst-case basis. The overall
emission rate is determined to be 29 mg/kg + 100 percent. Total annual
particulate emissions for raw-materials handling and manufacturing are 459 Mg
+ 100 percent. This is based on 12.7 Tg of glass produced, and is equivalent
to 0.0003 percent of the national particulate emissions from stationary
sources.
Table 7 shows a breakdown of raw-materials handling for the various
(14)
points of emission. This was determined primarily by NEDS data , and
confirmed by observations made during plant visits.
Most materials used to make glass containers have specified particle
size limits greater than 100 microns (150 mesh). Therefore, the amount of
material emitted from the plant site due to inertial forces will be minimal.
Composition of the particulates is given in a later section.
Glass Melting
Container glass is predominantly melted in fossil-fuel-fired
furnaces in the United States. Emissions from these furnaces are by far the
largest source of pollutant from a glass plant. Primary pollutants are
categorized as NO , SO , and particulates. The NO is composed predominately
x x x
of NO and the S0x composed predominately of SO . Emissions of CO, hydrocarbons,
selenium, and other materials can and do occur. The emission rate does
depend to some degree on glass type (i.e., flint or amber glass).
The overall emission rates and total emissions for furnaces melting
flint and amber glasses are given in Table 8 as well as in Appendix B. The
emission factors are based upon data reported in NEDS and derived from
(13 18—19^
various literature sourcesv ' . Data referred to as source measurements,
reported in NEDS, were obtained by actual point source test measurements. The
emission rates depend a great deal upon the operating conditions of the
glass-melting furnace. For instance, N0x emissions factors are reported to range
from 0.58 g/kg to 6.29 g/kg, S0x from 0.21 g/kg to 8.35 g/kg, and particulates
from 0.13 g/kg to 1.95 g/kg. Each type of emission is discussed £n greater
-------
37
TABLE 7. PARTICUIATE EMISSIONS DURING RAW-MATERIAL
PREPARATION AND HANDLING FOR CONTAINER GLASS
Process Step
Emission Factor,
rag/kg
Total Annual
Emissions(a>,
Mg
Handling (unloading,
conveying
Glass crushing
Storage bins
Mixing and weighing
Batch charging
TOTAL
22 + 100%
1 + 100%
1 + 100%
5 + 100%
Negligible (b)
29 + 100%
348
16
16
79
459
(a) Based on 15.8 Tg of raw materials processed to melt
12.7 Tg of glass.
(b) <0.1
-------
TABLE 8. EMISSIONS FROM FLINT AND AMBER CONTAINER
GLASS-MELTING-FURNACE OPERATIONS
Emission Factor,
g/kg
Species
NO
X
S0x
Particulates
CO
Hydrocarbons
Selenium
Flint
3.40 +
1.84 +
0.71 +
0.06 +
0.08 +
0.002 +
43%
36%
30%
166%
100%
100%
Amber
1.22
0.93
0.48
0.11
0.05
+ 100%
+ 100%
+ 100%
+ 100%
+ 100%
0
Total
3.07 + 47%
1.70 + 42%
0.68 + 36%
0.07 + 143%
0.08 + 100%
0.002 + 100%
Total
Annual Emissions
Based on Glass
Manufactured^3) . Gg
Flint
36.57
19.79
7.64
0.65
0.65
0.02
Amber
2.32
1.76
0,91
0.21
0.09
0
Total
38.89
21.55
8.55
0.86
0.74
0.02
Percent of National
Emissions from
all Stationary
Sources
Flint
0.3179
0.3073
0.0062
0.0036
0.0024
—
Amber
0.0199
0.0028
0.0006
0.0011
0.0003
—
Total
0.3378(a)
0.3101
0.0068
0.0047
0.0027
—
(a) 10.758 TG for flint, 1.898 TG for amber, 12.656 TG total, assuming a pack rate of 85 percent.
(b) Standard for NO-.
00
-------
39
detail for both flint and amber glasses. As was shown previously (Figure 3),
flint and amber glass comprises approximately 85 percent and 15 percent,
respectively, of the glass produced in the United States. Green glass is
included in flint. No differences in emissions are expected between green
and clear glasses.
Nitrogen Oxides
In a fossil-fuel-fired furnace, nitrogen oxides are formed by a
combination of atmospheric nitrogen and oxygen at the elevated temperatures
(> 1500 C) required for making glass. Because of the high temperature,
NO would be expected to be the primary oxide of nitrogen formed. For purposes
of this analysis, nitrogen oxides are designated as NO . In this study,
X
NO is compared against an NO air quality standard. The assumption that
X 2.
the NO- emission factor is equal to the NO emission factor is believed valid,
£" X
because once the plume has been diluted sufficiently with air (dispersion
calculations show that the plume is diluted approximately 1000 to 1 at the
point where it touches the ground), the photochemical conversion of NO to N0?
is quite rapid.
Nitrogen oxides represent the largest fraction by mass (^ 54%) of
emissions from the glass-melting furnace. The formation of NO in a glass-
X
melting furnace is extremely temperature sensitive. In one case, NO
X
concentration was increased some six times (from 100 ppm to MJOO ppm) as the
furnace temperature (measured at the bridgewall) increased from 1460 to 1550 C
(18)
and the flint glass production rate was doubled
The rate of NO formation depends upon factors such as peak flame
X
temperature, percent excess oxygen, and posttime/temperature history of the
flame. Consequently, considerable variation in the rate of NO emissions
Jb
can and does occur.
*
This report considers all nitrogen oxides as N0x and does not attempt
to determine the relative proportions of each. This differs from that method
preferred by the Glass Packaging Institute (GPI). Copies of correspondence from
GCMI (now GPI), EPA, and Battelle relating to this point as well as other points
are given in Appendix G.
-------
40.
Flint Glass. Source measurements reported in NEDS and taken
from the open literature give an average emission rate of 3.40 g of NO
J^
per kg of flint glass produced. This average is based on 21 measurements
(see Appendix B) and is calculated to be accurate to within + 43%
at a 95 percent confidence level. Individual values range from 0.58 to
6.29 g/kg. This amounts to approximately 36.6 Gg of NO emitted annually
X
from furnaces melting flint container glass. This is equivalent to 0.32 percent
(20)
of 1972 National N0~ emissions from all stationary sources
Amber Glass. Only three source measurements were found for NO
*W*MIHIIBW*» 80 percent) of sulfur oxide emissions are derived from sulfur
in the oil. Sulfur oxides from the batch generally combines with alkali
volatiles and exits as a particulate, while sulfur in the fossil fuel exits
predominately as SO .
X
Glass generally contains about 0.15 weight percent sulfate (added
usually as salt cake, gypsum, or blast-furnace slag). This is added for
melting and fining purposes and is a necessary ingredient for making container
glass. A range of values for sulfate (S0^=) in glass was reported in 1973 as
varying from 0.03 to 0.32 percent, with 82 percent of some 106 glasses
analyzed falling between 0.10 and 0.20 percent ^15\ The amount of mineral
sulfate added in the batch will, of course, be higher and usually falls within
-------
41
the 0.5 to 1.0 percent range. Sulfur oxide emissions from the batch materials
f-t n\
do occur and these depend primarily upon the quality of glass
melted.
Sulfur oxide emissions will be greatly influenced by any switch from
natural gas (the primary fossil fuel), which is essentially sulfur-free
to fuel oils or powdered coal containing sulfur. Such a trend does exist
primarily because of the reduced availability of natural gas in most sections
of the country. Sulfur in fossil fuels readily oxidizes in the glass-melting
furnace and appears as SO in the exhaust gases. For instance, a fuel oil
X
containing one weight percent sulfur emits approximately 600 ppm (calculated
(21)
as S02) in the flue gasv '.
Source measurements reported in NEDS or in the literature are
essentially for natural gas-fired furnaces. Hence, the emission are not
fully representative of an industry which is gradually switching
to fuels containing sulfur. However, such emissions will essentially
correspond directly to the sulfur found in the fuel oil.
Flint Glass. Source measurements for flint glass give an average
SO emission rate of 1.84 g/kg. This is based on 46 point source measurements
x
(see Appendix B) and is calculated to be accurate to within +0.36% at
a 95 percent confidence level. The values ranged from 0.21 to 8.35 g/kg. This
gives an estimated total annual emission of SO of 19.8 Gg, which is equivalent
X (20)
to 0.31 percent of 1972 National SO emissions from all stationary sources .
X
Amber Glass. Again, data were only available from three sources
for furnaces melting amber glass. These gave an average S0x emission rate
of 0.93 g/kg. The three values were 0.32, 2.06, and 0.41 g/kg. As can be
seen, these values are essentially equivalent to those for flint glass. Total
annual SO emissions from furnaces melting amber glass are estimated to be
X
1.76 Gg which is 0.003 percent of 1972 National S0x emissions from all
stationary sources .
-------
42
Particulates
Particulates from glass-melting chamber can originate both from
physical entrainment of batch materials being charged to the melting furnace
and from condensation of compounds, such as sodium sulfate (which forms through
the combination of sulfur oxides and volatilized sodium). Particulates exiting
with exhaust gases are essentially all (>95 percent) condensates, as indicated
by the fact that collected material is almost entirely water soluble. Studies do
show that batch materials are carried out of the melting chamber by the combustion
products; however, such materials do not show up in the stack-gas samplings;
therefore, it is assumed these coarser batch materials are retained in the
„ _ (22,23,24)
furnace-flue system.
Considerable opinion exists as to the mechanism by which condensate
particles are formed; however, analysis of these particulate emissions show them to
(9 18 22-24)
consist predominately (>75 percent) of submicron sodium sulfate ' ' .
The formation of these particulates depend upon batch composition,
temperatures in the melting furnace, production rate, surface area of molten
glass, and cullet ratio. Of these factors, production rate, temperature,
and surface area of molten glass are the most important factors affecting
the rate of particulate emissions. Since these variables are interrelated,
it is difficult to determine the relative influence of each, although it
would appear that temperature is the most significant variable. Data from one
furnace melting a soda lime glass showed that at zero production rate, the
particulate emissions were approximately 20 percent of that measured at its
f-t Q\
normal furnace capacity . Temperature was maintained at a constant value
(1450 C). Emissions ranged from 1.8 kg/hr at zero pull to 7.7 kg/hr at
(18)
normal pull of 211 Mg/day. Other data ' collected on soda/lime glass
during the study indicated that particulate emissions followed an Arrhenius
curve when plotted against the reciprocal of temperature; that is, a linear
relationship with the logarithm of the emission rate. This is shown in
Figure 12.
(24-25)
Stockham and others studied the emissions from furnaces
melting a flint and an amber glass and found the geometric mean particle
size of the particulate emissions to be 0.13 micron for the flint and 0.11
micron for the amber.
-------
100
50
43
(3D
"eo 10
-------
44
Flint Glass. Source measurements for particulate emissions from
flint glass-melting furnaces give an average emission rate of 0.71 g/kg. The
emission rate varied from 0.22 g/kg to 1.95 g/kg. Source measurements were
taken from 66 points and are calculated to be accurate within + 30%
at a 95 percent confidence level. This represents an estimated total annual
particulate emission of 7.64 Gg» or 0.006 percent of the 1972 National
particulate emissions from all stationary sources .
Amber Glass. Source measurements for particulate emissions from
furnaces melting amber glass give an average emission rate of 0.48 g/kg.
The emission rate varied between 0.13 g/kg and 0.83 g/kg. The average is
based on 23 source test measurements and is calculated to be accurate within
+ 41% at a 95 percent confidence level. This represents an estimated
total annual particulate emission of 0.91 Gg or 0.001 percent of the 1972
National particulate emissions from all stationary sources .
Carbon Monoxide
Carbon monoxide can be emitted through incomplete combustion
of the fossil fuel through the use of a luminous flame, or by reaction of a
powdered coal added to the glass batch to reduce sulfate compounds. The
emission rate varied between 0.05 and 0.13 g/kg. An estimated emission rate
for flint glass is 0.06 g/kg, based upon 19 reported measurements and accurate
to within + 167% at a 95 percent confidence level. This would
represent an annaul emission of 0.65 Gg of CO, or 0.004 percent of 1972
National CO emissions from all stationary sources . Only three source
measurements are available for amber glass. These show an emission rate of
0.12, 0.10, and 0.09 g/kg for an average of 0.11 g/kg. This would represent
an annual emission of 0.21 Gg for CO from furnaces melting amber glass.
-------
45
Hydrocarbons
Hydrocarbon emissions form in glass-melting furnaces primarily
through the incomplete combustion of a fossil fuel. The emission rate varied
from 0.01 to 0.53 g/kg. An estimated average emission rate for flint glass
is 0.08 g/kg, based upon 33 measurements calculated to be accurate within
+ 100%. Such an emission rate represents an annual emission of
0.65 Gg of hydrocarbons by manufacturers of flint glass, or 0.002 percent of
1972 National emissions for hydrocarbons from stationary sources . Only
three measurements were available for amber glass. These are 0.06, 0.03,
and 0.05 g/kg, which give an average emission rate of 0.05 g/kg or total annual
emission of 0.09 Gg.
Selenium
Selenium is used by flint glass manufacturers as a decolorizer to
neutralize the tint from transition metal oxide contaminants such as iron.
It is usually used in amounts of up to 0.001 percent. Test measurements on
selenium emissions were not found. Selenium volatilizes at rather low
(9 22)
temperatures (315 C for SeO, 685 C for Se) ' ; therefore, it can be expected
to be present in the waste gases. If the temperature of the waste gases is
dropped below 200 C, then selenium condensates are likely to be found.
The maximum amount of selenium release can be determined from the
amount of selenium consumed annually by the glass container industry. A
(32)
minimum of 60 percent is retained in the glass. Some 0.06 Gg of selenium is
(2)
consumed annually by the glass container industry (about one-sixth of U. S. usage).
Under these circumstances, which is believed to be a worst case, the emission
rate for selenium would be 0.002 g/kg. Total annual emissions would be 0.02 Gg.
-------
46
Other Emissions
Other minor emissions can include antimony and arsenic, which in
the past were added as fining and decolorizing agents. Both of these
materials have been virtually eliminated from use in recent years, Similarly,
chlorine was emitted in the past because of its association with soda ash
produced by the Solvay process. In recent years, most of the glass-container
industry has switched from soda ash produced synthetically to that manufactured
from a naturally occurring ore which does not contain Cl. By 1977, more than
90 percent of the industry will be using natural soda ash '
Forming and finishing
Molten glass, properly conditioned, leaves the forehearth of the
melting furnace, where it is cut into individual "gobs", which are then
transferred to the forming machine. The gob is formed into a container
by the blow-and-blow or press-and-blow method. After forming, a hot-end
coating or surface treatment may be applied, followed by the annealing
operation. Following this, the containers may then undergo a variety of
decorating or coating operations.
Emissions from the forming and finishing operations can include
hydrocarbons emitted during the forming operations, HCl and metal oxides
emitted during surface-treatment operations, emissions associated with combustion
gases produced during annealing, and organic fumes emitted from the coating
and decorating operations.
Very little emission data are available from the forming and
finishing operations. This section discusses emissions according to four
operations: forming, treatment, annealing, and decorating.
Forming
Gob shears, delivery chutes, and the forming molds for container
glass are lubricated with various solutions. These solutions can contain
grease, oils, graphite, and silicon-based emulsions. In the past decade,
-------
47
there has been a transition from grease and oil lubricants to the use of
silicone-emulsions and water-soluble oils (1 part silicone or oil to 90-150
(Q\
parts water) on gob shears and gob-delivery systems . Grease and oils are
still utilized on molds. Observing the forming operation, one can frequently
see a puff of white smoke occur when the molds are swabbed with a lubricating
solution. Although the smoke dissipates in a few seconds, hydrocarbon vapors
are probably released. These emissions are probably drawn through the large
ventilators on the roof of the plant. Emission data on hydrocarbon emissions
was not available; however, engineering calculations (Appendix B) indicate
that the maximum emission rate is low (0.03 g/kg). Total emissions are
estimated to be 0.44 Mg or less than 0.0001 percent of National Emissions.
Surface Treatment
Some glass containers receive a metal oxide (titanium or tin)
surface treatment to improve their resistance to scratching. Additionally,
this transparent treatment acts as a lubricant which can facilitate handling
and shipping operations. The oxide treatment is obtained by subjecting the hot
container (coming from the forming machine) to a vapor of metal chloride.
This is done within a hood. The metal chloride pyrolyzes to the metal oxide on the
container surface, leaving a metal-oxide film and releasing hydrogen chloride.
The bottles then go to the annealing oven.
Emissions from the surface-treatment operation will consist of HC1,
metal oxides, and hydrated-metal chlorides. Anhydrous tin chlorides which
do not react with the glass will decompose by the action of heat and moisture
within the exhaust ductwork to form metal oxides, hydrates metal chlorides,
and HC1. Estimations based upon available data ' indicate that approximately
60 percent of the total weight of the metal chloride input is released into
the atmosphere. Using tin tetrachloride as the input material, these
estimations reveal that of the total weight input, 14 percent is released
into the atmosphere as a metal oxide, 27 percent as hydrated tin chloride,
and 21 percent as HC1.
Emissions from the surface-treatment operation were determined by eng-
ineering calculation to be 0,02 g/kg of tin or titanium oxide, 0.03 g/kg of hydrated
tin or titanium chloride, and 0.02 g/kg of HC1. Total emissions are estimated to be
-------
48
0.08 Gg of metal oxide, 0.11 Gg of hydrated metal chloride, and 0.08 Gg of HC1
annually. This estimate is based on 30 percent of the containers being
treated.
Annealing
All glass containers undergo an annealing operation, where the
glass is brought to a temperature (approximately 550 C) to remove residual
stresses and then cooled uniformly to about 150 C where they are removed from
the annealing lehr (oven). All lehrs used by glass container manufacturers are
heated by natural gas, with propane as a possible alternate fuel.
The only emissions from annealing lehrs are combustion products.
Since natural gas is used exclusively, although lehrs can be heated
electrically and the temperatures are relatively low, emissions are low.
Measurement data are not available and emission rates were estimated on the
basis of emission factors for the combustion of natural gas. The results
are given in Table 9. Total emissions were calculated on the worst case
basis of all product being annealed in gas-fired lehrs.
Decorating
Glass containers are sometimes decorated with vitrifiable glass enamels
or organic materials. A wide variety of decorating techniques are employed.
Emissions occur predominately from organic solvents and binders used in
decorative coatings which are released during the curing of these compounds.
Data supplied by three glass container manufacturers would indicate that only
3 percent of all (both flint and amber) containers have decorative coatings.
This amounts to 330 Gg of container glassware decorated annually. Based upon
(14)
5 material balances given in NEDS , a hydrocarbon emission rate of 4.37 g/kg was
determined for container decorating. No point source measurements were available.
This amounts to 1.44 Gg of HC emitted annually. This represents 0.005 percent of
the National HC emissions from all stationary sources.
-------
49
TABLE 9, EMISSIONS FROM THE ANNEALING OF
SODA/LIME CONTAINER GLASSES
Species
NO
X
SO
X
Particulates
CO
Hydrocarbons
Total
Emission
8/kg
0.0015
0.025
0
0
0.035
Factor
(Ib/ton)
(0.003)
(0.05)
0
0
(0.07)
Total Annual
Emissions^ ,
Gg (ton)
0.019 ( 21)
0.316 (329)
-
-
0.443 (488)
Percent of
National
Emissions
from all
Stationary
Sources
0.0002
0.0005
-
-
0.0017
(a) Based on 12.656 Tg of glass processed.
-------
50
Emission Characteristics
Raw Materials Preparation
Emissions from this part of the manufacturing process will reflect
the raw materials used soda ash, limestone, feldspar, sand and sodium
sulfate), since no chemical reactions take place. Soft materials like
limestone and soda ash will be more easily crushed to dust.
Manufacturers generally specify particulate sizes ranging from
820 to 44 micron (-20 to +325 mesh). Table 10 illustrates general specification
limits for several raw materials used to make glass containers. Glass sand
would not be expected to cause significant dusting since only a small fraction
is below 100 micron and all particles are greater than 44 micron.
o
The primary ambient air standard for particulates is 260 ug/m .
Glass Melting
Emissions from the melting furnace consist of the criteria pollutants:
NO , SO , particulates, CO and hydrocarbons, as well as selenium. These
A X
emissions contribute to photochemical atmospheric reactions to produce smog and
can be irritating to the lungs.
Particulates consist predominately of sodium sulfate (>85 percent).
It is unclear as to whether these sulfate emissions pose a health hazard^ .
Forming and Finishing
Emissions from the forming and finishing operations consist of
the following:
(1) N0x, particulates, CO, and hydrocarbons emitted from
gas-fired annealing lehrs.
(2) Hydrocarbons produced by flash vaporization of mold lubricants
used in forming glass containers.
(3) HC1, tin or titanium oxide, and hydrated metal chlorides
exhausted from fume chambers during surface treatment operations.
-------
51
TABLE 10. GENERAL SPECIFICATION LIMITS FOR RAW
MATERIALS USED IN CONTAINER GLASS MANUFACTURE
Material Specifications Range
Mineral
Cerium oxide
Dolomite
Feldspar
Limestone
Sand
Soda ash
Sodium Nitrate
Chemical
Formula
Ce02
(Ca,Mg)C03
-
CaC03
Si02
Na2C03
NaN03
Amount ,
I
-
0.5
2
002
3
0
1
Mesh*
-
+16
+40
+20
+30
+20
+6
Amount ,
100
50
10
0.5
6.6
4.2
1.5
Mesh*
-60
-100
-200
-300
-100
-120
-100
* U. S. standard mesh size - see Table 15
for micron equivalents.
-------
52
(4) Hydrocarbons produced during the curing of decorative
coating used on some glass containers.
Total nationwide emission of the criteria pollutants produced in the
different stages of the container glass-manufacturing process were listed
previously in Table 2.
At a glass plant, the major amount of atmospheric emissions comes
from the melting furnaces. Calculations to portray the effect which a glass
plant has on its neighboring air environments have been made in the
following sections. Principal attention has been given to the pollutants issuing
from the melting furnace stacks.
Ground-Level Concentrations
Ground-level ambient concentrations of pollutants were used in
determining the environmental effects of the atmospheric emissions. These
were calculated for representative operations used in the manufacturing of
glass containers. A single plant having an annual production of 104.8 Gg
(115,000 ton) was selected as being representative, following the calcula-
tions made in earlier portions of this report.
Stack heights for container-glass furnaces, as given in the NEDS
listing, ranged from 5 to 50 meters with the predominant height being about
40 meters. Two other frequently occurring stack heights are in the vicinity
of 20 m and 45 m. There is no correlation between stack height and production
rate. In general, stacks taller than 30 m can be assumed as natural draft
stacks or natural draft stacks that have been converted to an ejection-air
system. A mean stack height of 38.2 m (125 ft) was selected for the melting
furnaces in a representative plant. Other stack parameters are chosen from the
NEDS data for their compatibility with a 38 m stack. The furnace-stack
emissions are derived from the emission factors given in Table 2 and are applied
to the 104.8 Gg annual production rate. All of the parameters for the melting
furnaces and stacks are listed in Table 11.
The maximum ground-level concentration was used to determine infor-
mation for the environmental effects criteria. This maximum concentration can
-------
53
TABLE 11. PARAMETERS FOR MELTING FURNACES OF A REPRESENTATIVE PLANT IN THE
GLASS-CONTAINER INDUSTRY AS USED IN ATMOSPHERIC-DISPERSION CALCULATIONS
Stack Parameters
Glass produced: 104.8 Gg/yr (115,000 T/yr)
Stack height: 38.2 m (125 ft)
Stack diameter: 1.53 m (5 ft)
Exit temperature: 353 C (650 F)
Gas flow rate: 910 m3/min (32,000 ACFM)
Exit velocity: 8.35 m/sec (27.2 ft/sec)
Meteorological Conditions
(a)
Wind speed: at 10 meters — 4.1 m/sec (9.2 mph)
at top of stack — 8.2 m/sec(b) (18.4 mph)
Ambient temperature at top of stack: 15 C (59 F)
Atmospheric pressure: 1000 millibars
Atmospheric stability: D(C)
Calculated Parameters
Plume rise: 5.72 m(d) (18.7 ft)
Effective stack height: 43.9 m (144 ft)
Estimated Parameter
Mean wind speed affecting the plume between the effective stack
height and the surface: 6 m/sec
Emissions (G)
NOX: H.25 g/sec (390 T/yr)
SO : 6.14 g/sec (213 T/yr)
Particulates: 2.73 g/sec (94 T/yr)
CO: 0.2 g/sec (6.9 T/yr)
Hydrocarbons: 0.2 g/sec (6.9 T/yr)
(a) Average of annual mean wind speeds measured at city airports
near 30 glass-container plant locations.
(b) Increase of wind with height in suburbs and level country as
given in Figure 1-3 of ASME Recommended Guide for the Pre-
diction of the Dispersion of Airborne Effluents. 1968.
(c) D stability is the predominant stability as determined from a
cross section of Star Program results (see Table 12).
(d) Plume rise was calculated from the Holland equation for neutral
stability.
-------
54
be obtained for substitution into an equation or from a nomograph. The
equation is
f
where
X = maximum concentration (gm/m )
max
Q = pollutant emission rate (gm/sec)
u = mean wind speed (m/sec) at the height of the stack
H = effective stack height (m) , the physical height of
the stack plus the plume rise
CTZ = vertical plume standard deviation (m)
a = horizontal plume standard deviation (m)
e = base of natural logarithms, 2.718
rr = 3.14.
CT
Z
For D stability, the ratio CT is on the order of 0.5 varying from
y (27>
0.57 to 0.24 between 0.1 km and 10 km downwind from a source . The ratio is
approximately 1.0 for C stability. The maximum concentration occurs at a
( 27^
x u/Q and the distance to the point of maximum concentration can be
distance where a = h//2. Turner has presented a monograph from which
Z
determined for any stability and effective stack height. When emission rate
and wind speed are known, the value of X can be calculated.
max
The environmental effect criteria are developed for 24-hr average
concentrations, while the dispersion predictions discussed above are for
short periods (3 to 10 min). For longer periods, one must consider that
variations in wind direction and wind speed would cause the average concen-
tration at a downwind monitor to be less than the concentration calculated
for a short-term wind blowing constantly from the source to the monitor.
Turner has given an equation by which the long-term average concentration
can be estimated when the short-term concentration is known:
X -I fi
f <5 V 1-
* S ^ t „
-------
55
vhere
X^ = concentration for the long period (t.)
Xg = concentration for the short period (t )
The value of the dimensionless exponent, b, is
between 0.17 and 0.2
t. = long-time period, min
t = short-time period, min.
s
While this equation is most applicable for X. = 2 hr or less, it can be
Xi
applied to a 24-hr period. Turner gives the conversion coefficient of 0.35 for
transforming a 3-min average into a 24-hr average. Other conversion coefficients
are 1 hr, 0.61, and 3-hr, 0.51.
Before calculating ambient pollutant concentrations, representative
meteorological parameters for the area need to be chosen. These parameters
are required for determining plume rise and dispersion.
Plume rise was calculated from the Holland equation
AH = -a- (l.5 + 2.68 x 10"
T
s
where
AH = rise of the plume above the stack, m
v = stack gas exit velocity, m/sec
S
d = inside diameter of stack, m
u = wind speed at top of stack, m/sec
p = atmospheric pressure, millibars
T - stack gas temperature, K
s
T - air temperature, K.
3.
Choices of the meteorological parameters were made after a review
of climatology in some of the areas of the .country where glass plants are
(29)
found. Account was also taken of the variations of meteorology between the
surface and the top of the furnace stack. The values selected for the melting
furnace calculations are listed in Table 11. Stability Type D (neutral class)
is the most frequently occurring stability throughout the United States as
-------
56
(27)
calculated by the Turner method which considers the surface wind speed and
the net radiation (Table 12). A surface wind speed of 4.1 m/sec was chosen
as representative of the conditions at the glass plants based on a survey of
the average annual wind speeds listed for the National Weather Service
( 28^
meteorological stations located at 30 cities which have container glass
plants. It should be noted that 4 meter/sec wind speeds in Turner's scheme
for determining stabilities can accompany stabilities varying from Type B to
Type E, depending on the solar radiation. Type D was chosen for the dispersion
calculations on the basis of its predominant frequency.
Wind speeds increase with altitude and this effect was taken into
account for the 38.2 meter effective mean furnace stack height for the represen-
tative plant. Wind speed in the layer in which the downward dispersion of the
plume would take place, 0-44 meters, was estimated to be 6 meters/sec. This
was an extrapolation from the standard wind-measurement height of 10 meters
(29)
following a guide giving examples of the variation of wind with height
over suburban and level county areas. For stack heights of 30 to 40 meters, the
wind speed is expected to be 1.5 (level terrain) to 3 (urban areas) times
stronger at the top of the stack than at 10 meters.
Table 13 (in its second column) presents the maximum pollutant
concentration predicted for ground level in the vicinity of the representative
glass melting plant. These concentrations are the contributions from only
the melting furnaces, and do not take into account other glass plant emissions
or emissions from sources other than the glass plant. Table 13 also presents
data for selenium, a minor pollutant emitted by a container glass furnace.
finissions from three other sources representative of air emissions
from a container glass manufacturing operations were also considered in
relation to their effect on ambient-air quality. These were:
(a) Particulates from a baghouse collecting the emissions
from materials handling
(b) Hydrocarbons from a container-decorating operation.
(c) Tin oxide and hydrated tin chloride particulates
(d) HC1 from a surface treatment operation.
To make the ambient-concentration estimates for these sources, emissions and
stack parameters were adapted from data given in the NEDS listing. Meterological
-------
57
TABLE 12. RELATIVE FREQUENCY OF ATMOSPHERIC STABILITIES
Station
Stability Class
D E and F
Milwaukee
0.001 0.031 0.094 0.636 0.238
St. Louis Q005 0.047 0.103 0.555 0.289
Peoria
0.003 0.042 0.102 0.577 0.276
Pittsburgh 0.001 0.022 0.083 0.567 0.306
Columbus, Oh. 0.010 0.058 0.100 0.500 0.331
Mobile
0.008 0.052 0.115 0.453 0.371
Los Angeles 0.001 0.041 0.148 0.482 0.329
Dallas
0.004 0.042 O.io? 0.586 0.262
* Based on Output from U.S. Department of Commerce National Climatic
Center Star Program for Five Years of Data.
-------
58
TABLE 13. MAXIMUM POLLUTANT CONCENTRATIONS AND SOURCE SEVERITY FOR
EMISSIONS FROM THE MELTING FURNACES FOR A REPRESENTATIVE
GLASS CONTAINER PLANT
Pollutant
NO
X
SO
X
Particulates
CO
Hydrocarbons
Selenium
Quality Standard
100 (a)
365
0.79(f>
0.95
0.023
Severity,
S
0.38
0.056
0.035
2.0xlO"5
5.9x!0"3
3.4xlO~2
(a) Annual arithmetic mean assumed here as 24-hr standard for N0_.
(b) 24-hr standard.
(c) 8-hr standard.
(d) 3-hr standard.
(e) Threshold Limit Value for 8 hrs.
(f) 3-min Xmax adjusted to match sampling time of the standard using
X = X
max max-
std
(g) 3-min X adjusted to 24-hr sampling time.
-------
59
conditions similar to those used in the glass-furnace emission-dispersion
calculations were used for these other sources with adjustments for differing
stack heights. Information regarding two of these calculations is given in
Table 14 m The maximum amounts of coating used for surface treating as listed
in the NEDS is 6 tons. Calculation of tin and HC1 emissions from the indicated
values are on the order of 50 ^g/sec which are considered to be negligible.
For each of the maximum ambient concentrations that were calculated,
a source severity, S, was also determined. Source severity for criteria
pollutants (particulates, sulfur oxides, nitrogen oxides, carbon monoxide, and
hydrocarbons) is determined from the following equation:
X
o _ max
AAQS
where
X — maximum average ground-level concentration of
max
the pollutant for the time period of the
3
standard (|J,g/m )
3
AAQS = ambient air-quality std (p,g/m ).
For noncriteria pollutants, the source-severity equation uses the threshold
limit value instead of the ambient air-quality standard with a correction for
a 24-hour period and a safety factor:
S = max
TLV (8/24) (1/100)
where
TLV = Threshold Limit Values for each species
8/24 = correcting factor for the 8-hr work day which is
the basis for the TLV
1/100 = safety factor.
A review of the source-severity factors in Tables 13 and 14 shows
the highest value to be that produced by nitrogen oxides emitted from the
melting furnace stacks. For the conditions used to portray a representative
-------
60
TABLE 14. MAXIMUM AVERAGE GROUND-LEVEL CONCENTRATIONS (X ) OF AIR
POLLUTANTS FROM CONTAINER GLASS-PLANT SOURCES BESIDES THE
MELTING FURNACE
Source 1. Baghouse Controlling Materials-Handling Emissions
Emissions: 0.004 grams/sec (0.1 tons/yr) of particulates
Emission Point -- Stack
height - 35 m; diameter = 0.3 m;
exit temperature = 10 C; exit velocity = 17.3 m/s
Xmax(3min)' Xmax' ^S/I"3' Ambient Severity
Species u,g/nr (specified time) Std Factor
Particulates 0.06 0.02 (24 hr) 260 |0,g/m3 9.2 x 10
Source 2. Container Decorating
Emissions: 0.44 grams/sec (15 tons/yr) of hydrocarbons
Emission Point -- Stack
height - 12.2 m; diameter = 0.6 m;
exit temperature = 149 C; exit velocity = 11.4 m/s
-5
Species
Hydrocarbons
Xmaxj3Jin)
47
Xmax' M
-------
61
glass container plant, the source severity factor is 0.38, a low value.
All other severity factors, including those from sources other than the
melting furnace are also quite low.
Affected Population
As a consequence of the dispersion of pollutants, the severity
starts at zero near the stack, increases downwind, reaches a maximum, and
then decreases to zero again (see Figure 13)„ The affected population is
defined as the population around a representative plant exposed to a severity
greater than 1.0. To determine the downwind distances enclosing the affected
population, the standard dispersion equation for the centerline concentration
from an elevated source is used.
c-ol
where
-3
X = pollutant concentration at surface (gm )
u «= average wind speed through the dispersion
layer (m sec )
Other parameters are the same as in the earlier
dispersion equation.
It is assumed that winds from all directions are equally likely.
By rearranging, this equation becomes
XTTU 1 1 / H \2~1
~ " yT exp ~^JJ-
The value of X is specified by the requirement for S = 1.0 and then it is
corrected to the three-minute average concentration which the dispersion
equation gives. Substituting values of a and a from Turner's graphs of
dispersion coefficient as a function of distance downwind into the right-hand
-------
62
X
*»
c
o
0)
u
o
o
c
o
o
QL
o
o
0)
>
0)
Crt
0>
o
1_
I
1.0
•max.
Distance Downwind,X
R2-
R , -
Outer radius
Inner radius
FIGURE 13. ILLUSTRATION DEPICTING CALCULATION OF AREA
WHICH CONTAINS THE AFFECTED POPULATION
-------
63
side of the equation will produce a short table for values of the right-hand
side of the equation versus downwind distance. These are plotted in a
fashion similar to Figure 13 and the values of 1^ and R^ are determined.
These values form the inner and outer radii of an annulus enclosing the
affected population.
In calculating the number of people affected, it has been assumed
that the population density around the representative glass plant is 248
people per square kilometer.
Since none of the pollutant sources produced severity factors
greater than 1.0, all affected population values for the container glass
sources are zero.
As an illustration calculations were made for nitrogen oxides from
the melting furnace using a severity factor of 0.1 instead of the standard 1.0.
Results of this calculation is presented below.
Source/Emission
Concentration (3 min)
Downwind
Value, Distance,
|xg/m3
m
Affected Population *
Radii
Inner, Outer, Persons
m m Affected
Melting Furnace/NO
x
105
880
420
3900
11,700
* Based on the more restrictive source severity factor of 0.1 instead of
the accepted 1.0. There were no glass container sources which produced
a source severity factor greater than 0.4.
-------
64
SECTION V
CONTROL TECHNOLOGY
Control of emissions in the glass-container industry varies
considerably, depending on the type, source, and amount. Control technology
has evolved for both economic and environmental reasons. Various methods
are utilized to reduce air emissions from the different portions of the glass-
manufacturing process. These include: (a) development of process modifica-
tions, (b) new furnace designs, and (c) application of control equipment.
For example
(1) Arsenic is no longer used as a fining agent
(2) Many fossil fuel-fired furnaces are equipped with
electric boosting which can increase output, thus
reducing the amount of effluent per unit of output
(3) Fabric filters, electrostatic precipitators, and scrubbers
are being used or have been examined for removing dust
i
particulates. In addition, several commercial equipment
manufacturers are attempting to develop methods for
removal of SO and NO emissions at the same time
X X
particulates are removed.
This section discusses the control technology currently being used or that
which might be considered for use by the glass-container industry. It was
not the intent of the study to consider the economics associated with the
control technology or even to verify the technology itself. Rather, the
purpose of this section is to identify control technology reportedly applicable
to the control of emissions from glass manufacturing plants. The discussion
is organized in a manner similar to the emission section.
-------
65
Raw-Materials Preparation
The handling and mixing of raw materials is a source of particulate
emissions from a container glass plant. Raw materials are usually conveyed
from hopper railroad cars or trucks (by screw conveyors, belt conveyors,
bucket elevators, or pneumatic conveyors) to elevated storage bins, as was
shown previously in Figure 5. A few minor glass-batch ingredients are
delivered to the plant in paper bags or cardboard drums. These are later
transferred by hand to smaller storage bins or fed directly from the storage
device.
Materials are gravity fed from the storage bins into weigh"hoppers
and then transported by transverse belts or bucket elevators into a mixer.
Gullet is crushed to a desired size. After mixing, the glass batch is
transferred to a charging bin located next to the glass-melting furnace
or into a batch storage bin, depending upon the design of the batch-handling
system. Positive displacement or vibratory feeders at the bottom of the bins
feed the materials to the chargers, where it is fed into the glass-melting
furnace. Gullet may be added to the batch in the mixer, while the batch
(8)
is being transferred, or charged separately to the melting furnace .
Emissions
Little information is available regarding plant emissions due to
dusting during the raw-materials handling stages of the process. As
discussed in the previous section, the fraction of the dust generated which
leaves the plant site will consist of particles smaller than 100 microns in
diameter.
As described in the previous section on emissions, the particulate
emission rate for raw-materials handling is estimated to be 22 mg/kg.
Based upon the total glass batch handled by the container industry, particulates
exhausted annually would average 348 Mg.
-------
66
Information on the composition of these participate emissions is
not available, but they will consist essentially of the same raw materials
being handled (soda ash, silica, limestone, etc.), since no chemical
reactions occur during this portion of the manufacturing process. Limestone
and soda ash can be expected to predominate because of their relative
softness. Glass manufacturers generally specify raw materials which are
coarser than 100 micron, as shown in Table 15. Consequently, the amount of
raw material emitted from the plant site due to inertial forces alone would
(14)
be relatively small. This is in line with the reported measurements
Raw-Materials-Control Technology
Process Modif icajijlonji ^ Container glass
manufacturers have minimized dusting problems in batch-handling operations by
limiting the amount of fine particles (<100 microns) in the batch material,
as can be seen in Table 15. Specifications for glass-grade raw materials
will generally require removal of the finer sizes of material, especially
with softer materials that crushed to dust easier than sand.
Another batch-preparation method that is used to control dusting
during handling is the addition of water to the raw batch (batch wetting).
Trials have also been conducted where the batch is wet with a liquid caustic-
(32)
soda solution that is substituted for soda ash . Water is presently added
in amounts up to 4 percent to the mixed batch materials. The substitution of
a caustic-soda solution for a soda ash is not generally practiced by the
(32-33)
glass-container industry
Efficiency of Control Equipment. Transport of raw materials in
railroad hopper cars and hopper-bottom trucks (dump trucks) is still
practiced. During unloading of these trucks or railroad cars, materials
dumped onto conveyor belts can result in some dust being dispersed into the
air. Generally, the hopper cars or trucks are connected to sealed receiving
hoppers with fabric sleeves and the dust generated during the unloading operation
is filtered through the sleeves or exhausted through a baghouse^ .
-------
TABLE 15. GLASS-GRADE PARTICLE-SIZE SPECIFICATIONS FOR
SAND, LIMESTONE, AND 10- AND 20-MESH DOLOMITE
Approximate
Particle
Size
2.3 mm
1.3 mm
820 \i
410 ji
150 jj,
105 p.
74 p,
44 y,
U.S. Standard
Mesh Size
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Glass- Glass -Grade
Grade Glass-Grade Dolomite, %
Sand, % Limestone, % 10-Mesh
8 - 0.0 0.0
16 - 2.0 max 15.0 max
20 0.0 10.0 max
40 12.0 max
100 - - 90.0 min
140 92.0 min 85.0 min
200 99.5 min 94.0 min 97.0 min
325 100.0 min
20-Mesh
0.0
-
2.0 max
-
80.0 min
95.0 min
96.0 min
-
a\
-------
68
Enclosing the loading area with a suitable fabric structure and sealing
all covers and access openings with gaskets is effective in reducing dust
during this operation. This results in an inward-air velocity across the open
(35)
mouth of the bag that prevents an eruption of dust into the atmosphere
Trapped air and fine dust can then be filtered by a conventional fabric
filter and the cleaned air exhausted into the atmosphere.
Weigh hoppers and mixers require ventilation because of surges in
material from the large air flows. In older mixers, polyvinylchloride seals
are installed between the rotating body of the mixer and its frame to reduce
air leaks. In newer mixers, the body does not rotate. The exhaust gases
are usually filtered of particulates greater than submicron size by the use
of fabric filters.
The use of fabric filters for separation of particulates from air
has been practiced for a number of years in the glass industry. The earliest
fabric filters were known as "baghouses", since these were large free-standing
units for exposed fiber bags. By passing the exhaust air through layers of
a woven fabric, the particulates were collected. Unfortunately, as the
thickness of the collected layer of particulates increased, the pressure
differential required for continued air flow also increased. Thus, the
collected dust must be periodically removed by manual or mechanical shaking.
Almost all container glass plants use fabric filters to remove entrained
dust particles ' . The fabric filters used today are totally enclosed,
and most have a continuous removal operation for the trapped particulates.
The traditional woven and synthetic fabrics are used. Today,
fabric filters are generally made of low-temeprature materials such as
f o / o c \
Nomex, nylon, terylene, or Orion
Fabric filters are used to collect particulates from the raw-
materials and handling operations for several reasons. First, they have an
efficiency of greater than 99 percent and they can be used to collect submicron-
size particulates. In addition,the trapped particulates can later be recovered
(9 34-35)
for reuse or recycle ' . One manufacturer had from 2 to 6 baghouses
with a stack height less than 50 feet at a plant manufacturing 72.6 Gg
(80,000 ton) of container glass per year( ' '. These used nylon-fabric
filters operating at 98 percent efficiency and collecting about 72.6 Mg
(80 ton) of dust per year.
-------
69
_Glass-Meltlng Operation
In a glass-melting furnace, raw materials are heated until a
homogeneous viscous liquid, free of gas inclusions, is formed. Temperatures
in the melter will generally fall in the 1500 to 1600 C range (2730 to 2912 F)(36)
Natural gas and fuel oil are the principle types of fuel, .with natural gas
predominating (^70 percent)(15>37). Over 90 percent of glass-melting furnaces
have regenerative-firing systems for purposes of heat recovery and fuel
(14) J
conservation .
In order to increase melting capacity, many furnaces now have
electric-boosting systems. These systems consist of several water-cooled
electrodes equally spaced along the sides or bottom of the melter, below the
surface of the glass. Additionally, all-electric melting is used by a few
manufacturers.
Emissions
Important air emissions from a glass-melting furnace consist of
NO , SO , and particulates. Other emissions can include CO, hydrocarbons,
X X
and selenium.
Nitrogen oxides represent the largest fraction by mass, about
54 percent of glass-furnace emissions ' '
As was described earlier, the source test measurements of NO
X
emission rates vary from 0.58 to 6.29 g/kg of glass produced. For additional
information, see Appendix B. Based on an average emission rate of 3.07 g/kg,
glass-melting furnaces with a total production rate of 12.656 Tg would emit
38.8 Gg of NOX yearly(14).
SO , on the other hand, depends both on the sulfur content of the
fuel and on the sulfur content of the batch material. Sulfur present in the
fuel oil will oxidize and appear as S0x in the exhaust gas. A fuel oil
containing 1 percent sulfur by weight emits « 600 ppm S02 in the flue gas<21>.
Sulfur is also present in the batch materials, usually as Na2S04> During heatup,
the sulfate decomposes and sulfur dioxide is formed, some of which is
chemically encorporated into the glass (as SO^) and some released within the
furnace. An average emission rate of S0x is 1.70 g/kg. Thus, plants
-------
70
producing 12.656 Tg of glass annually would emit approximately 21.5 Gg
of SO yearly.
x
Particulate emissions from a glass-melting furnace result primarily
from volatilization of materials in the melt that combine with gases such as
SO to form condensates in the flue system. Particulate emissions from a
glass-container furnace consist of approximately 80 percent sodium
sulfate^9'18'22"2 . These particulates form from the condensed vapors in
the melt and are submicron sized^ ' ~ . The median particle diameter from
(24-25)
a flint glass furnace was found to be 0.13 y ,
Source-test measurements for particulate emission rates vary from
0.13 to 1.95 g/kg of glass produced. This averages to a particulate
emission rate of 0.68 g/kg.
Other emissions exhausted from glass-melting furnaces include CO,
hydrocarbons, and selenium.
Carbon monoxide is probably exhausted from the glass-melting
furnace as a result of incomplete fuel combustion. Source-test measurements
have reported emission rates from 0.05 to 0.13 g/kg. An estimated average
emission rate is 0.07 g/kg.
Hydrocarbons are also formed in the glass-melting furnace as a
result of incomplete fuel combustion. Source-test measurements have
reported emission rates from 0.01 to 0.53 g/kg. The calculated average emission
rate was 0.08 g/kg. Actual emission rates are a function of firing conditions
(extent of fuel/air mixing, excess air, firing temperature).
Selenium is generally used in amounts of 0.001 weight percent or
less in the batch as a decolorizer to neutralize the green tint in container
(9 10 22)
glasses caused by iron impurities. ' ' ' No test measurements on actual
selenium emissions have been reported, but it likely leaves the stack
as selenium vapor, because of its low vaporization temperature (315 C for SeO
(9 22)
and 685 C for Se) ' . A worst case emission rate was calculated to be
0.002 g/kg, with total annual emissions of 0.02 Gg.
-------
71
Glass-Melting-Control Technology
Control of emissions from the glass-melting furnace have occurred
primarily because of environmental considerations. Four general approaches
have been employed:
(1) Modification of feed material
(2) Modification of furnace design
(3) Increase of checker volume
(4) Adoption of commercial-control apparatus.
Modification of feed material and furnace design have been primarily used to
control gaseous emissions, while the other two methods are used for control
of particulate emissions.
Modification of Feed Material. Raw materials have a tendency to
vaporize or decompose in the glass-melting furnace. Several raw materials
that readily vaporize include nitrates and selenium. By minimizing the amount
of these or other ingredients used or by substitutions of materials, the
amount of gaseous emissions exhausted from the glass-melting furnace is
reduced. For example, arsenic has been essentially eliminated as a fining
agent. Cerium is used to partially replace selenium as a decolorizer. In
addition to reducing the selenium in the batch (by about 40 percent), this
modification leads to the elimination of arsenic in the batch, since cerium
and arsenic are not compatible. Cerium is especially appealing because it
tends to form high-melting compounds which do not readily vaporize.
Modification of Furnace Design. Increasing the fuel efficiency
of the glass-melting furnace can in turn lead to a decrease in combustion
products, a decrease in dust entrainment by hot combustion gases passing
over the melting glass batch, and possibly a decrease in furnace temperature.
In addition, emissions from low melting and easily vaporized fluxing or fining
agents can be lowered. Several methods currently in practice to improve
furnace efficiency are:
-------
72
(1) Better instrumentation for regulating air/fuel mixtures
and monitoring furnace temperature and stack gas composition.
(2) Combustion control to produce long luminous flames that
eliminate spurious hot spots in the furnace and provide
better heat transfer to the melt
(3) Improved refractories to increase corrosion resistance,
which permits furnaces to be more fully insulated
(4) Use of electric melting to reduce incomplete combustion
and volatilization losses
(5) Use of electric boosting to increase furnace capacity,
increase furnace efficiency, and lower temperatures above
the molten glass.
All of these methods have been employed to control gaseous emissions. Sulfur
oxides that form can be controlled by both limiting the sulphate in the feed
(18)
material and by the improvement of furnace efficiency. Ryder and McMackin
found that the SO emission rate increased directly with an increase'in
X
production rate on a sideport furnace melting soda/lime glass. This increase
comes about because of the larger quantities of sulphate being added to
the furnace when the production rate doubled.
NO emissions can be also lowered when the furnace efficiency is
X
increased if the furnace temperature also drops. A 10 percent decrease
in fuel consumption will result in a 10 percent decrease in NO emissions .
X.
Electric boosting is commonly used on fossil-fuel fired furnaces
in the container glass industry, primarily to increase output. About 40 percent
of these glass-melting furnaces are equipped with boosters. Boosting will
normally result in a reduction in emissions per unit of output. '
Electric melting furnaces essentially eliminate both particulate
and gaseous emissions from the glass-melting operation. In 1975, less
than 3 percent of container glass manufacturers used electric melting .
-------
73
Adoption of Commercial-Control Apparai-i.a. Participates can be
cleaned from the glass furnace exhaust by scrubbers, fabric filters, or
electrostatic precipitators (ESP). Figure 14 shows a breakdown of commercial-
control apparatus presently in use. Scrubbers can also be used to collect
SO emissions, while fabric filters and ESP's only remove particulates
(51)
Teller v suggested spraying the stack gas with an alkaline solution. This
would cause the acidic gases (S0x, HF, or HCL) to react and form particulates
that could then be collected by the control device.
Scrubbers . One type of particulate scrubbing is a two-step process.
Initially, particles in the exhaust gases are "contacted" or wet by a
scrubbing fluid that draws the particles into agglomerates. These agglomerates
are then separated from the gas stream by an inertial mist-elimination process.
A low-pressure (<10-in. water) centrifugal scrubber used by one
(9)
glass company in California had two separate contacting sections within
a single casing. Separate 50.7 metric horsepower (50 horsepower) circulating
fans forced dirty gas through each section containing two to three impingement
elements similar to fixed blades of a turbine.
One reference mentions a scrubber that uses a packed-bed
preconditioning chamber. Hot gases (538 C) containing volatilized sodium
compounds enter the chamber, and the vapors condense out onto the packing
material. This material is wet by a scrubbing solution and provides a large
surface area for condensation. A standard Venturi-type scrubber completes
the system. This scrubber is presently installed on a 0.181 Gg/day (200 ton/day)
container glass furnace and it reduced particulate loading from more than
0.23 to less than 0.046 g/sdm3 (from more than 0.10 to less than 0.02 g/sdcf)
One flint-glass manufacturer installed a tower scrubber (2.9-meter
diameter) on a 44.8 meter2 (482 ft2) melter. Hot effluent from the furnace is
initially quenched and saturated with a caustic solution passing through the
exhaust gas at 900 gal/min. The gas then passes into a 300 gal/min variable
throat Venturi operating at 30 in. of water. This scrubber has been plagued by
malfunctions and breakdowns. A highly visible steam plume is exhausted when
it is not working.
-------
3.6% Electrostatic Precipitator
0.07% Centrifugal Collector
2% Fabric Filters
2% Scrubber
7% Gravity Collector
-(No collection equipment used)
Amber
(No collection equipment used)
Soda/Lime
FIGURE 14. USE OF COMMERCIAL-COLLECTION EQUIPMENT FOR EMISSION, .
CONTROL ON BOTH AMBER AND SODA/LIME GLASS FURNACESV '
-------
75
Fabric Filters. Fabric filters, also known as "baghouses"(36),
collect particulates by filtering exhaust gas from container glass-melting
furnaces through closely woven natural or synthetic fabric filters that are
capable of trapping submicron particulates. Unlike wet scrubbers, fabric
filters are unaffected by variations in the gas flow rate. Temperature control,
however, is critical for proper functioning and the type of fabric filter
selected is dependent upon the temperature of the gases exhausted. Fabric
filters are generally made of cotton sateen, standard nylon, wool, dacron,
/ o ^_O^"S
orlon, NOMEX, teflon, and fiberglass . Maximum operating temperatures
for these fabrics are given in Table 16.
TABLE 16. MAXIMUM USE TEMPERATURE FOR
VARIOUS FABRIC-FILTER MATERIALS
Fabric
Cotton Sateen
Standard Nylon
Wool
Dacron
Orion
Nomex
Teflon
Fiber Glass
Maximum Temperature, C
99
93
107
135
135
204
232
288
Since stack gas from a glass container melting furnace is at 316 to 645 C
(600 to 1200 FK4 , the gas must be cooled to a temperature compatible
with the fabric filter bag. This can be accomplished by using the following
methods, either alone or in combination with each other
(1) Air dilution
(2) Radiation-cooling columns
(3) Air/gas heat exchangers
(4) Water-spray chambers.
-------
76
Dilution of off gases with air is the simplest and most trouble-free method
of reducing temperature, but requires the largest baghouse because of the
increased volume of gases. Air-to-gas heat exchangers, and radiation and
convection ductwork are subject to fouling from dust in the effluent. A
water-spray increases humidity and requires careful temperature control to
avoid condensation, but does permit use of smaller baghouses. Care must
be taken, with all of these methods, to avoid cooling the gas to the
temperature where SO and H.,0 would condense out and foul or react with the
fabric filters.
In addition to being selected for their thermal compatibility,
fabricfilter ba,gs must also be corrosion and abrasion resistant. Cotton,
( 26}
orlon, and dacron can deteriorate from the S03 in the flue gas
A fabric-filter air-pollution-control system was installed in 1974
? 9 (1T^
on a 41.8 m (450 ft ) melter producing amber glass v '. The 482 C (900 F)
effluent from the furnace was initially cooled to 177 C (350 F) . A fine
powder aluminate precoat was then introduced into the air stream at 18.1 kg/hr
(40 Ib/hr) along with ambient air. This further reduced the gas temperature
to 121 C (250 F). The baghouse contained 1200 m2 (12,915 ft2) of dacron-
filter cloth divided into six compartments, each containing 900 filter bags.
During normal operation, the air-to-cloth ratio was 1.55, but this would
increase to 1.86 during the cleaning cycle. The pressure drop ranged from
3.5 to 4.5 in. of water across the bags. An exhaust blower had to develop
16 to 18 in. of water pressure to overcome the resistance of the checkers,
heat exchanger, baghouse, and about 46 meters of duct. Initially, the heat
exchanger required maintenance about 15 percent of the time due to plugging
with material condensing from the gas stream. By blocking off about 40 percent
of the tubes, the temperature increased slightly, but this permitted a
normal maintenance schedule. Discharge of particulates from the baghouse
outlet was typically 1.1 kg/hr (2-3 Ib/hr). Tests using a Brinks Impactor
showed these particulates to be <0.75 micron.
ElectrostaticPrecipitator (ESP). In an electrostatic precipitator
(ESP), a voltage source creates a negatively charged area, usually by hanging
wires in the gas flow path. Grounded collecting plates composed the sides of
the ESP. A powerful electric field is created by the high potential difference
-------
77
between these grounding plates and the discharging wires. As the gas stream
passes through the field, the particles become electrically charged and are
drawn to the collecting plates. Periodically, accumulated particles are
removed from these plates by vibration, rapping, or rinsing. Thus, by applying
the collecting force only to the particles to be collected, a much lower power
input is required (i.e., 200 watts per 0.5 m /s)^ .
NAFCO Engineering, Ltd. (a Japanese firm) has developed a new type
of ESP. In contrast to the conventional units, the NAFCO ESP uses thousands
of stainless steel needles affixed to the leading and trailing edges of
positively charged electrode plates. Thirty five of these systems are now in
operation in Japan, with nine of them being used on soda/lime glass-melting
furnaces
United McGill Corporation, who is the licensed U.S. distributor
for the NAFCO ESP, has installed the unit on 10 soda-lime glass furnaces
(25)
to date . All of these systems are designed to have an outlet loading of
<0.046 g/std m3 (0.02 g/scfd).
2 2
An 84.4 m (908 ft ) melting furance, used for producing flint glass,
had an ESP installed in early 1974^ . It consisted of dual chambers, where
the air flow could be directed to either chamber or divided between them.
Each chamber had three electrical fields connected in series. Designed for
12.9-sec treatment time at 0.67 m/s (2.2 fps) velocity through the treater,
one chamber was found to be as effective as two, the conclusion being that
the system was overdesigned.
Technological Advances. Collector systems previously discussed
are primarily useful for collecting particulates and for decreasing opacity
of gaseous emissions. One company now offers dry and wet systems to
control both particulate and gaseous emissions. A nucleation scrubber is
used on their wet system to effect collection of submicron particulates and
acidic gases (HF and SO ). A solid absorbent, on the other hand, is injected
into the gas stream to react with the noxious gases in their dry system.
The absorbant is then separated from the gas along with particulates in a
fabric filter.
-------
78
A patent (U.S. 3,789,628) was issued for a scrubber where an
aqueous solution of sodium silicate is sprayed into the gases as they are
exhausted in the furnace stack. Water from the solution evaporated in the
gas stream and the sodium silicate forms a small sticky sphere which the
patent claims to react chemically with NO and SO , and physically with
X X
particulates. These spheres can then be collected and recycled into the
(43)
glass batcb> .
The quantity of NO from a glass-melting tank was studied by Kitayama,
fc.\ ^
et al , in order to evaluate methods for reducing fuel consumption under
photochemical smog warnings. A glass-melting furnace (of unknown glass
composition) with a 154.2 Gg/day (170 ton/day) capacity using preheated air
at 1100 C, emits 850-1000 ppm of NO . By varying the damper opening and
2^
reducing the excess air by 10 percent, the NO emissions were reduced to
X
480 ppm. When the excess air was reduced 20 percent, the NO emissions were
X
reduced to 45 ppm.
(44)
Takasaki reports on a method for removing NO from flue gases by
X
wet oxidation and absorption. This technique claims to eliminate more than
90 percent of the NO from the flue gas of a glass-melting furnace. By
X 3
using activated carbon and chlorine acid soda, a pilot plant with 51 kg m /hr
reduced its NO emissions by 95 percent. This system consists of a special
X
liquid-gas contact tower that utilizes a chlorine dioxide and chlorine oxidizing
agent. NO is converted into N0_, which is absorbed by the liquid and
stabilized. The existing gas contains no NO, <10 ppm N09, <5 ppm SO-, no
chlorine oxide, chlorine, or hydrogen chloride, 13 percent C0_, 3.5 percent
2
0?, and 0.29 mg/kg m Of dusts. Other details were not reported.
(45) 1
Kanenatsu reports on scrubbers handling 377, 7,1, and 28.6 kg m /hr
m/hr of SO in the flue gas. By using a wet or dry desulfurization method,
X
the sulfur oxides are absorbed by NaOH solutions and oxidized in air, and
the SO recovered as tnirabilite.
f I £ \
Kanematsu suggested use of low sulfur fuels, high stacks, and
stack gas-desulfurization systems as methods for controlling SO emissions.
-------
79
Efficiency of Equipment
Least effective and least expensive of the air-pollution control
devices is the wet scrubber . In addition to having numerous malfunctions
and breakdown, they have been found to exhibit particulate-collection
efficiencies as low as 66 percent^ to as high as 90 percent^33'* (if grain
loadings were low). By fitting the column with impingement plates, efficiency
can range up to 95 percent with particles as small as 5 microns^3 . A major
advantage of this system is its ability to remove acidic gases.
Baghouses have a reputation for high efficiency and dependability.
Fabric filters are capable of >99 percent efficiencies and can collect
(13 33)
particulates down to below 0.75 micron ' . Major disadvantages are that
exhaust gases must be pretreated to remove gaseous emissions and must be
cooled before they contact the low-temperature fabrics.
Electrostatic-precipitator performance is highly sensitive to
temperature and volume fluctuations. Electrical characteristics of particulates,
which affect collection efficiency, vary with temperature, humidity, S02
content, and the type of particulate. Conventional ESP's have been shown to
have efficiencies up to 95 percent and collect particulates down to submicron
size. The NAPCO ESP, on the other hand, has a reported outlet loading of less
than 0.046 g/std m3 (0.02 grains/scf) . For an uncontrolled emission rate
of 1 kg of particulate/Gg (2 Ib of particulate/ton) of glass and an air flow
of 3119 std m3/Gg (100,000 scf/ton), the efficiency is reported to be
85 percent. For an emission rate of 10 kg/Gg (20 Ib/ton), the efficiency
is reported to be greater than 98 percent. This ESP was designed so
additional sections could be added so efficiencies greater than 99 percent
could be obtained(33»47).
Wet or dry desulfurization methods, presently in use by a glass
company, in Japan, has shown efficiencies of better than 97 percent for
the wet and 80 to 90 percent for the dry for S0x removal.
-------
80
Forming and Finishing
As the glass leaves the forehearth of the furnace, the molten
glass is cut into "gobs" by a pair of mechanical shears. Chutes direct the
gobs from the feeder into blank molds where it is formed by one or two
methods.
With the "blow-and-blow" technique, the gob is settled with compressed
air and preformed into a parison with a counter blow. The parison is inverted
and transferred into the blow mold where it is blown into its final shape.
Wide-mouth containers, on the other hand, are formed by a "press-
and-blow" technique. The gob is settled by pressing with a plunger and
"puffed" with a counter blow. The parison is inverted and transferred for
final blow forming.
The surface of approximately 30 percent of glass containers are
treated in an operation where hot bottles from the forming machine pass
through a fume chamber containing vapors of tin or titanium tetrachloride.
A surface layer of the metal oxide forms on the container. Unreacted
hydrated metal chlorides are exhausted into the atmosphere. The containers
are then annealed at 593 to 649 C (1100 to 1200 F) and uniformly cooled
in gas-fired, continuous ovens called lehrs.
A polymer coating may then be applied by spraying the containers
with an aqueous dispersion of coating material. The heat from the containers
evaporate the water and fuze the polymer into a uniform surface coating.
Decorative coatings are applied to about 3 percent of the glass
containers. Vitrifiable glass enamels or organic resins are applied by
brush, stencils, banding machines, rubber stamps, offset processes,
electrostatic printing, and silk screen printing. Metallic decorating materials
(liquid bright metals, such as gold, platinum, palladium, and silver, which
leave a mirror-like coating when fired on the glass) are also applied in
the same manner. If a container is to be glazed, a water suspension of glass-
forming ingredients is applied by spraying or dipping. These decorative
coatings are then cured in annealing ovens at approximately 600 C.
-------
81
Forming Emissions. Molds on forming machines, gob shears, and
delivery chutes are lubricated with solutions ranging from grease and oils to
graphite and silicone-based emulsions. During the past decade, silicone
emulsions and water-soluble oils have replaced grease and oil lubricants on
(9)
gob shears and gob delivery systemsv '. Grease and oils are still used on
molds and causes white smoke emissions during flash vaporization of the swab.
Although the smoke dissipates in a few seconds, hydrocarbon vapors are released.
These emissions are released inside the plant but probably drawn outside
through the large ventilators above the melting furnace and are minor.
Forming and Finishing Control Technology
Efforts to control the hydrocarbons emissions have centered on
finding lubricants capable of withstanding high temperature (900 C) and not
volatilize. Use of silicone emulsions are water-soluble oils (90 to 150 parts
of water to 1 part oil or silicone) can eliminate these emissions. Unfortunately,
(9)
they have not performed well as mold-release compounds . Emissions from
the forming machinery are dispersed within the plant and exhausted by the
room ventilating systems. No companies were identified which used any control
device for these emissions.
Surface Treatment
Emissions. Emissions from the surface treatment of glass containers
with tin or titanium tetrachloride include metal oxides, hydrated metal
chlorides, and HC1 that are released into the atmosphere. The emission rate
is estimated to be 0.02 g/kg for metal oxides, 0.03 g/kg for hydrated metal
chlorides, and 0.02 g/kg for HC1.
Particulates exhausted are generally composed of submicron size
metal chloride and oxide. A calculated particulate emissions rate is
0-05 g/kg.
-------
82
Surface Treatment Control Technology
A U.S. Patent 3,789,109 describes^ an apparatus to be used
for cleaning solid, liquid, and gaseous pollutants from a hot-end surface
treatment station of a container glass manufacturing plant. In this apparatus,
the air discharging from the hood is heated until the metal chlorides in
the air disassociate to metallic oxides and hydrogen chloride gas. Exhaust
gases are then sprayed with fresh water to cool the stream. Water reacts
with the hydrogen chloride to form hydrochloric acid. Exhaust air passes
into a scrubber where the pollutants are removed , and then conveyed to a gas
scrubber where metal oxides are removed.
Decorating
Emissions. Hydrocarbon emissions from organic solvents and binders
used in coatings on containers are released when decorative coatings are
cured in annealing lehrs. A calculated emission rate for these hydrocarbons
is 4.37 g/kg. Only 3 percent of the containers are decorated, giving a total
(2)
annual emission of 1.44 Gg.
Decorating-Control Technology
Process modifications are difficult to accomplish without harming
(49)
the quality of the coating . In addition, they do not completely eliminate
hydrocarbon emissions. Several such changes involve the substitution of
solvents and a reduction of solvent concentration in the coating.
Hydrocarbon emissions can be controlled by incineration, absorption
(activated charcoal or silica gel), or condensation .
-------
83
SECTION VI
FUTURE GLASS-CONTAINER PRODUCTION
The future production levels for glass containers is highly
unpredictable because of three major factors: (1) shortages in natural gas
and substitute energy sources, (2) potential large-volume penetration of the
beverage-container market by plastics plus continuing penetration by cans, and
(3) possible legislation restricting or outlawing nonreturnable containers.
The shortage of natural gas and the allocation of petroleum
products places a constraint on the container-glass industry. In recent
years, most of the industry has used petroleum products primarily as a
reserve or standby fuel and, therefore, does not have a base period of any
significant usage. At the same time, the industry has incurred reductions in the
use of natural gas, its primary fuel. Since oil is the only normal replacement
fuel, allocations based on historical demand would indeed constrain the
production on container glass in the United States.
The primary substitute fuel oil used by the glass industry is a
distillate, such as No. 2 fuel oil. However, both distillates and residuals
are used by the industry in the melting operation with properly designed fuel-
handling and burner systems.
Oil cannot normally be substituted for natural gas in other nonmelting
operations, such as in the refiner, forehearth, and annealing lehrs, for
reasons of glass quality. In such areas, propane is the only substitute. If
a limited amount of natural gas is available in the glass plant, it is
usually reserved for use in the nonmelting areas, assuming oil is available for
substitution in the melting area.
The extent to which plastic containers will penetrate the beverage
market is still highly speculative. But, extensive research to resolve
limiting factors such as cost, gas permeability, and creep has led to test-
market introduction of plastic containers by the two largest soft-drink
companies Coca-Cola Company and Pepsico, Inc. The initial plastic soft-
drink containers have concentrated on the 32-oz capacity or larger sizes of
which they are most cost competitive. Also, a large beer manufacturer has
announced intent to package in metal and plastic in the future rather than
-------
84
glass. If energy costs continue to escalate, the competitive position of
plastic containers may be improved both in manufacturing and transportation
(distribution) costs.
Metal cans have been claiming an increasing percentage of the
smaller capacity soft-drink containers and may be expected to increase these
inroads barring any legislative effects.
Federal legislative action relative to returnable versus nonreturnable
beverage containers has been under consideration for some time. If such
legislation does become law, it could have a major impact on both the total
numbers of glass containers manufactured and the competitive position of
glass, plastics, and metal. The details of such legislation would determine
how it would effect glass-container production.
The future of glass container production will also depend upon the
success of the industry's research on developing very light-weight beverage
containers. If gross reductions in glass weight are achieved through
improved forming techniques, strengthening processes, plastic coatings, etc.,
then the competitive position with metal and plastic containers could lead to
improved growth potential.
Historically, the shipments by the industry have grown an average
of 6 percent annually since 1967, as can be seen in Table 17.
TABLE 17. GLASS CONTAINER PRODUCTION STATISTICS^
Year
1974
1973
1972
1967
Tg Production, 10 tons
11.00
11.32
10.77
8.39
12.13
12.48
11.87
9.25
106 Bottles
39,800
39,790
38,550
33,271
This growth is attributable primarily to the increased popularity of the
nonreturnable bottle. In recent years, the growth has been less. It is likely
that 1980 production will be 20 percent higher than for 1974. Total National
emission will also increase by this amount without changes in control technology.
-------
85
SECTION VII
UNUSUAL RESULTS
As mentioned in the previous section, the use of natural gas is
declining as a primary energy source by glass container manufacturers. Its
usage has dropped from approximately 83 percent in 1971 to about 70 percent
C2)
in 1975 . Oil usage is increasing and this will have a direct effect on
SO and possibly particulate air emissions. Tending to counter this effect
X
is improved furnace efficiency achieved by process modifications that act to
reduce stack emissions. It is not possible at this time to predict the
quantitative impact of these changes on future air emissions.
-------
86
REFERENCES
(1) Current Industrial Reports: Glass Containers Summary for 1974. Series
M-32G (74)-13 (May 1975).
(2) Personal communication with Glass Containers Manufacturer Institute
(GCMI).
(3) "Directory Issue", The Glass Industry, .56 (10), 1975.
(4) "Glass Factory Directory Issue", American Glass Review, .95 (8A), Febru-
ary 28, 1975.
(5) Private communication with anonymous manufacturer.
(6) Kitayama, Hiroshi, Hideo Hayashi, Sataro Iwasaki, Tadashi Fujimura,
Tomohiko Mujano, Hideaki Murayama, Tomihiro Myuhata, "Effect of Com-
bustion Conditions on Nitrogen Oxides Formation of Furnaces", presented
at Japan,Soc. Air Pollution, 14th annual meeting, Fukushima, Japan,
(Jov. 1973).
(7) Hutchins, J. R. Ill, and Harrington, R. V., "Glass" from the Encyclopedia of
of Chemical Technology, 2nd Edition, 10, John Wiley & Sons., Inc. (1966)
533-604.
(8) Bauer, W. C., Tooley, F. V., and Manring, W. H., "Batch Materials Hand-
ling and Preparation", The Handbook of Glass Manufacture. 1, 57-94
(1974).
(9) Danielson, J. A., Air Pollution Engineering Manual, 2nd Edition, EPA
Publication No. AP-40 (May 1970).
(10) Anon, "A Screening Study to Develop Background Information to Determine
the Significance of Glass Manufacturing", prepared by Research Triangle
Park Institute for EPA, Contract No. 68-02-0607-Task 3 (December 1972).
(11) Private communication with E. Stable, Owens-Illinois, Inc., Toledo, Ohio.
(12) Chemical and Process Tech. Encyclopedia, Ed. Considine, 551-561 (1968)
(13) Simon, Herbert and Wiliamson, E., "Control of Fine Particulates from
Continuous Melting Regenerative Glass Furnaces", presented at the 68th
Annual Meeting of the APCA, Boston, Massachusetts (June 15-20, 1975).
(14) Anon, National Emission Data System, Environmental Protection Agency
Research Triangle Park, North Carolina (1974).
(15) Schorr, J. R. and Anderson, G. A., Final Report on Industrial Energy
Study of the Glass Industry to FEA and DoC", Battelle Columbus Labora-
tories, Contract No. 14-01-0001-1667, pages 80-142 (1974).
-------
87
(16) Lillis, E. J., and Young, D. "EPA Looks at 'Fugitive Emissions'", J. Air
Pollution Control Assoc., 2J5 (10), 1015-18 (1975).
(17)
Air Pollution, Vol. 1, Edited by A. C. Stern, 2nd Edition, Academic Press,
N. Y. (1968), "Nonviable Particles in the Air1, (M. Corn). 49-52.
(18) Ryder, R. J. and McMackin, J. J., "some Factors Affecting Stack Emissions
from a Glass Container Furnace", The Glass Industry, 50, 307-11, 346-350
.(June 1969).
(19) Arrandale, R. S., "Air Pollution Control in Glass Melting", Symposium
Sur La Fusion du Verre, Brussels (October 1968), 619-644.
(20) Anon, State-by-State Listing of Source Types that Exceed the Third Decision
Criteria, Special Project Report, Monsanto Research Corp., Contract 68-02-
1874, 1-3 (1975).
(21) Reed, R. J., "Combustion Pollution in the Glass Industry", The Glass Industry,
54 (4), 24-26, 36 (1973).
(22) Arrandale, R. S., "Pollution Control in Fuel Fired Tanks", The Glass Industry,
55 (12), 12ff (August & November 1974).
(23) Davis, R. E., Manring, W. H., and Bauer, W. C., "Carryover Studies in Glass
Furnaces", presented at the 34th Annual Conference on Glass Problems, 109-
126, U. of 111. (November 1973).
(24) Stockham, John D., "The Composition of Glass Furnace Emissions", Journal
of the Air Pollution Control Assoc., 21 (11), 713-715 (1971).
(25) Custer, W. W., "Electrostatic Cleaning of Emissions from Lead, Borosilicate,
and Soda/Lime Glass Furnaces", presented at the 35th Annual Conference on
Glass Proglems, Ohio State University (Nov. 14-15, 1974).
(26) Frantz, C. J., Miser, D. L., Troy, H. N., and Stabbe, E. D., collected papers
from the 32nd Annual Conference on Glass Problems, Dept. of Ceramic Engineer-
ing, University of Illinois, 25-38 (1971).
(27) Turner, D. B., Workbook of Atmospheric Dispersion Estimates. EPA Publication
No. AP-26 (1970), Figure 3-9.
(28) Climatic Atlas of the United States. U.S. Dept. of Commerce (1968).
(29) Recommended Guide for the Prediction of the Dispersion of Airborne Effluents.
Edited by M. Smith, ASME (1968).
(30) Reznik, R. B., Source Assessment; Flat Glass Manufacturing Plants. EPA En-
vironmental Protection Technology Series, Monsanto Research Corporation,
Dayton (October 1975).
-------
88
(31) Mills, H. N., and Jasinski, J., "Evaluating Batch Changes", The Glass Ind.,
51 (5), 223-227 (1970).
(32) Tooley, F. V., "Raw Materials", Handbook of Glass Manufacture. Vol. 1,
Books for Industry, New York (1974), Chap. 2.
.(33) Rymarz, Ted M., and Lipstein, David H., "Removing Particulates from Gases",
Chemical Engineering Deskbook. 82 (21), The McGraw-Hill Publishing Company,
New York 113-129 OctOber 1975).
(34) Swift, P., "Dust Control Related to the Bulk Delivery of Particulate
Materials", The Chemical Engineer, 143-150 (March 1975).
(35) Edmundson, J. N., Rietz, L., Weise, R. L., and Fraas, J., Collected papers
from the 32nd Annual Conference on Glass Problems, Dept. of Ceramic Engineer-
ing, University of Illinois, 39-54 (1971).
(36) Arrandale, R. S., "Furnaces, Furnace Design, and Related Topics", Handbook
of Glass Manufacture. Vol. 1, Books for Industry, New York (1974), Section
5, 249-387.
(37) Hibscher, William, Stertz, R., The U.S. Glass Industry's Challenge in These
Energy Critical Times", presented at the 35th Annual Conference on Glass
Problems, The Ohio State University, 85-101 (November 1974).
(38) Bartz, D. V., KVB Engineering, Inc., Control of Oxides of Nitrogen from
Stationary Sources in the South Coast Air Basin of California, California
State Air Resources Board (1974).
(39) Anon, "Symposium on Pollution, Stratford-Upon-Avon, 30 May-1 June 1973",
Glass Technology, J. (6), 140-144.
(40) Teller, A. J., "Control of Emissions from Glass Furnaces", Ceramic Bulletin
51, 637-640 (1972).
(41) Keller, G., "Scrubber System Lightens Load of Glass Furnace Emissions",
Chemical Processing, 38. 9 (Jan. 1975).
(42) Teller, A. J., "Control of Emissions from Glass or Ceramic Manufacture",
presented to ACS, St. Louis, Missouri, September 1971.
(43) Mahoney, W. P., "Method for Controlling Furnace Emissions", U.S. Pat.
3,789,628 (1974).
(44) Takasaki, Shoichi, "Flue Gas Denitration by West Oxidation and Absorption",
Heat Management Pollution Control, 26 (1), 57-62 (Jan. 1974).
(45) Kanematsu, Jado, "Air Pollution Control in Glass Industry", Seramikk
(Ceramics), .9 (1), 49-55 (Jan. 1974).
-------
89
(46) Kanematsu, Jado, "Countermeasures for Preventing Air Pollution Caused by
Glass Industry", Seramikksu (Ceramics), j) (1), 15-21 (1974).
(47) Wright, R. W., "Application of Electrostatic Precipitators for the Control
of Container Glass Emissions", IEEE Trans, on Industry Applications. 1A-11
(No. 4), 447-456 (July 1975).
(48) Lyon, R. S. and Lyon, R. L., "Method for Cleaning a Gas", U.S. Pat. 3,789,104
(1971).
(49) Troy, H. N. and Kalter, P. A.,"Pollution Control and Glass Decurating", The
Glass Ind., 52 (3), 102-105 (March 1971).
(50) LeMaire, W. H., "Why Bill Coors Wants A Plastic Bottle", Packaging Engineer-
ing. January, 1976.
(51) Teller, A. J., "Control of Glass Furnace Emissions", Glass Industry, 5_7 (2),
15-19, 22 (February, 1976).
(52) Roos, P. W., "Lehr Priority: Design Concepts to Save Energy", The Glass
Industry, 56, 18-22 (April, 1975).
(53) Hangebrauck, R. P., Von Lehmden, D. J., and Meeker, J« E., "Emissions of
Polynuclear Hydrocarbons and Other Pollutants from Heat-Generation and
Incineration Processes", Journal of the Air Pollution Control Association,
14, 267-278 (July, 1964).
-------
APPENDIX A
GEOGRAPHICAL LISTING OF THE
122 CONTAINER GLASS PLANTS
-------
TABLE A-l. GEOGRAPHICAL LISTING OF THE 122 CONTAINER GLASS PIANTS
County
Population
Density,
persons/km^
State
Alabama
Arkansas
California
Colorado
Connecticut
Florida
Plant
Brockway Glass Co., Inc.
Arkansas Glass Containers Corp.
Anchor Hocking Corp.
Anchor Hocking Corp.
Ball Corp.
Brockway Glass Co., Inc.
Brockway Glass Co., Inc.
Gallo Glass Co.
Glass Containers Corp.
Glass Containers Corp.
Glass Containers Corp.
Kerr Glass Mfg. Corp.
Latchford Glass Co.
Madera Glass Co.
Owens -Illinois
Owens-Illinois
Owens-Illinois
Thatcher Glass Mfg. Co.
Columbine Glass Co.
Glass Containers Corp.
Anchor Hocking Corp.
Industrial Glass Co., Inc.
Owens-Illinois
Thatcher Glass Mfg. Co.
City
Montgomery
Jonesboro
Los Angeles
San Leonadro
El Monte
Oakland
Pomona
Modesto
Antioch
Hayward
Vernon
Santa Ana
Los Angeles
Madera
Los Angeles
Oakland
Tracy
Saugus
Wheat Ridge
Day vi lie
Jacksonville
Bradenton
Lake land
Tampa
County
Montgomery
Craighead
Los Angeles
Alameda
Los Angeles
Alameda
Los Angeles
Stanislaus
Contra Costa
Alameda
Los Angeles
Orange
Los Angeles
Madera
Los Angeles
Alameda
San Joaquin
Los Angeles
Arapahoe
Windham
Duval
Manatee
Polk
Hillsborough
AQCR
2
20
24
30
24
30
24
31
30
30
24
24
24
31
24
30
31
24
36
41
49
52
52
52
(persons /mi2)
80 (206)
28 (73)
662 (1714)
558 (1445)
662 (1714)
558 (1445)
662 (1714)
50 (127)
290 (752)
558 (1445)
662 (1714)
696 (1802)
662 (1714)
7 (19)
662 (1714)
558 (1445)
78 (201)
662 (1714)
76 (196)
61 (159)
259 (670.3)
49 (127)
46 (119)
180 (467)
Furnaces
3
3
3
1
2
2
1
4
1
1
2
1
4
2
3
2
4
3
2
2
-------
TABLE A-l. (Continued)
State
Georgia
Illinois
Indiana
Louisiana.
Plant
Glass Containers Corp.
Midland Glass Co.
Owens - 11 linois
Anchor Hocking Corp.
Ball Corp.
Hillsboro Glass Co.
Kerr Glass Mfg. Co.
Metro Containers
Obear-Nester Glass Co.
Obear-Nester Glass Co.
Owens-Illinois
Owens-Illinois
Owens -11 linois
Thatcher Glass Mfg. Co.
Universal Glass Products Co.
Anchor Hocking Corp.
Brockway Glass Co., Inc.
Foster-Forbes Glass Co.
Glass Containers Corp,
Glass Containers Corp.
Kerr Glass Mfg. Co.
Midland Glass Co.
Owens-Illinois
Thatcher Glass Mfg. Co.
Laurens Glass Co.
Owens -Illinois
Underwood Glass Co.
City
Atlanta
Warner Robin
At lanta
Gurnee
Mundelein
Hillsboro
Plainfield
Dolton
East St. Louis
Lincoln
Alton
Chicago Heights
Streator
Streator
Joliet
Winchester
Lapel
Marion
Indianapolis
Gas City
Dunkirk
Terre Haute
Gas City
Lawrenceburg
Rust on
New Orleans
New Orleans
County
Fulton
Houston
Fulton
Lake
Lake
Montgomery
Will
Cook
St. Clair
Logan
Madison
Cook
LaSalle
LaSalle
Will
Randolph
Madison
Grant
Marion
Grant
Jay
Vigo
Grant
Dearborn
Lincoln
Orleans
Orleans
AQCR
56
54
56
67
67
75
67
67
70
75
70
67
71
37
67
76
76
76
80
76
76
84
76
79
22
106
106
County
Population
Density
persons /knr
(persons/mi^)
433 (1122)
63 (164)
433 (1122)
317 (821)
317 (821)
16 (42.3)
275 (712)
2197 (5689)
161 (417)
20 (52.9)
130 (337)
2197 (5689)
37 (96)
37 (96)
275 (712)
24 (63)
117 (304)
77 (199)
758 (1963)
77 (199)
23 (60.3)
105 (272)
77 (199)
37 (94.9)
27 (69.2)
1102 (2854)
1102 (2854)
Furnaces
2
1
4
1
1
2
3
3
2
8
9
2
1
3
2
5
2
3
3
3
2
4
2
5
I
l-o
-------
TABLE A-l. (Continued)
State
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
New Jersey
Plant
Carr-Lowrey Glass Co.
Columbia Glass Co.
Maryland Glass Corp.
Foster-Forbes Glass Co.
Owens -Illinois
Owens-Illinois
Brockway Glass Co.
Midland Glass Co.
Chattanooga Glass Co.
Glass Containers Corp.
Underwood Glass Co.
Anchor Hocking Corp.
Brockway Glass Co.
Gayner Glass Works
Kerr Glass Mfg. Corp.
Leone Industries
Metro Containers
Metro Containers
Midland Glass Co.
Owens-Illinois
Owens-Illinois
Thatcher Glass Mfg. Co.
Wheaton Industries
City
Baltimore
Baltimore
Baltimore
Milford
Mansfield
Charlotte
Roseroount
Shakopee
Gulf port
Jackson
Mineral Wells
Salem
Freehold
Salem
Millville
Bridgeton
Jersey City
Carteret
Cliffwood
Bridgeton
N. Bergen
Wharton
Millville
County
Baltimore
Baltimore
Baltimore
Worchester
Bristol
Eaton
Dakota
Scott
Harrison
Hinds
Union
Salem
Monmouth
Salem
Cumberland
Cumberland
Hudson
Middlesex
Monmouth
Cumberland
Hudson
Morris
Cumberland
AQCR
115
115
115
120
120
125
131
131
5
5
135
45
43
45
150
150
43
43
43
150
43
43
150
County
Population
Density,
persons/km^
(persons/mi^)
976 (2527)
976 (2527)
976 (2527)
162 (419)
307 (796.3)
46 (120)
93 (241)
35 (91)
87 (224)
92 (239)
17 (44.2)
7 (17)
369 (955)
7 (17)
92 (238)
92 (238)
4907 (12703)
714 (1849)
369 (955)
92 (238)
4907 (12703)
314 (813)
92 (238)
Furnaces
3
4
1
2
3
2
2
4
1
8
3
2
3
5
1
2
2
4
8
4
2
23
>
-------
TABLE A-l. (Continued)
State
Plant
City
County
AQCR
County
Population
Density,
persons/km2
(persons/nd.2)
Furnaces
New York
North
Carolina
Ohio
Oklahoma
Oregon
Pennsylvania
Glenshaw Glass Co.
Leone Industries
Owens-Illinois
Thatcher Glass Mfg. Co.
Ball Corporation
Laurens Glass Co.
Owens ~I1 linois
Brockway Glass Co., Inc.
Chattanooga Glass Co.
Ball Corp.
Bartlett-Collins Co.
Brockway Glass Co.
Brockway Glass Co.
Kerr Glass Mfg. Corp.
Liberty Glass Co.
Owens -Illinois
Anchor Hocking Corp.
Brockway Glass Co.
Brockway Glass Co.
Brockway Glass Co.
Brockway Glass Co.
Diamond Glass Co.
Foster-Forbes Glass Co.
Glass Containers Corp.
Orangeburg
Rochester
Brockport
Elmira
Asheville
Henderson
Winston-Salem
Zanesville
Mt . Vernon
Okmulgee
Sapulpa
Ada
Muskogee
Sand Springs
Sapulpa
Portland
Connellsville
Brockway
Cr ens haw
Washington
Washington
Royersford
Oil City
Marienville
Rock land
Monroe
Monroe
Chemung
Buncombe
Vance
Forsyth
Muskingum
Know
Okmulgee
Creek
Pontotoc
Muskogee
Tulsa
Creek
Multnomah
Fayette
Jefferson
Washington
Washington
Montgomery
Venango
Forest
43
160
160
164
171
166
136
183
175
186
186
188
186
186
186
193
197
178
178
197
197
45
178
178
502 (1300)
404 (1047)
404 (1047
94 (243)
81 (210)
51 (131)
189 (490)
45 (116)
31 (79)
19 (49)
19 (48)
15 (38)
27 (70)
14 (37)
19 (48)
500 (1295)
75 (193)
25 (66)
25 (66)
94 (244)
94 (244)
490 (1268)
35 (90)
4 (10)
2
3
3
3
1
1
3
3
1
2
1
2
1
4
4
5
2
2
4
3
2
1
-------
TABLE A-l. (Continued)
State
Plant
City
County
AQCR
County
Population
Density,
persons/km?
(persons/nd.2)
Furnaces
Rhode Island
South
Carolina
Tennessee
Texas
Washington
West
Virginia
Glass Containers Corp.
Glass Containers Corp.
Glenshaw Glass Co.
Metro Containers
Owens-Illinois
Pennsylvania Glass Products Co.
Pierce Glass Co.
Star City Glass Co.
Laurens Glass Co.
Chattanooga Glass Co»
Anchor Hocking Corp.
Chattanooga Glass Co0
Glass Containers Corp.
Kerr Glass Mfg. Corp.
Owens-Illinois
Northwestern Glass
Erockway
Chattanooga Glass Co.
Kerr Glass Mfg. Corp.
Owens-Illinois
Owens-Illinois
Universal Glass Products Co.
Parker
KnoK
Glenshaw
Washington
Clarion
Pittsburgh
Port Allegheny
Coventry
Laurens
Chattanooga
Houston
Corsicana
Palestine
Waxahachie
Waco
Seattle
Clarksburg
Keyser
Huntington
Fairmont
Huntington
Parkersburg
Butler
Clarion
Allegheny
Washington
Clarion
Allegheny
McKean
Kent
Laurens
Hamilton
Harris
Navarro
Anderson
Ellis
McLennan
King
Harrison
Mineral
Cabell
Marion
Cabell
Wood
197
178
197
197
178
197
178
120
203
55
216
215
22
215
212
229
235
113
103
235
103
179
61 (157)
24 (63)
842 (2180)
94 (244)
24 (63)
842 (2180)
20 (51)
315 (817)
26 (68)
170 (441)
386 (1000)
11 (28)
10 (26)
19 (48)
54 (139)
207 (536)
66 (171)
27 (69)
144 (373)
75 (194)
144 (373)
88 (229)
1
2
4
1
2
3
2
9
5
4
3
2
1
4
6
6
1
1
5
5
1
Wisconsin
Foster-Forbes Glass Co.
Burlington
Kenosha
239
166 (429)
-------
APPENDIX B
EMISSIONS DATA
-------
APPENDIX B
EMISSIONS DATA
Raw Materials Preparation and Handling
Five typical points for participate emissions were considered for
the raw materials preparation and handling operations: (1) unloading and
conveying, (2) crushing of cullet (scrap glass), (3) filling and emptying of
storage bins, (4) weighing and mixing of batch, and (5) feeding of batch to
glass melting furnace (batch charging). Source test data are summarized in
Table B-l.
Emissions from the raw materials preparation and handling operation
consist entirely of particulates from raw batch materials. In practice, only
fugitive dust emissions should be considered, since particulate emissions
remaining within the plant may constitute an OSHA health and safety considera-
tion distinct from plant emissions. As discussed in the text, only particles
below 100 micron are considered as contributing to figitive dust emissions.
Actual measurements of plant emissions from these operations were not avail-
able; however, personal observation indicates that there are no visible
(14)
emissions from the batch house. Measurements were available from NEDS
of particulate emissions within a few plants. These were used to determine
particulate emissions on a worst-case basis.
The average emission factors for the various raw material prepara-
tion and handling operations were taken to be the following, calculated on
a worst-case basis.
rig/kg.
1. Handling 22 + 100%
2. Crushing 1 + 100%
3. Storage 1 + 100%
4. Mixing 5 + 100%
5. Charging <0.1
29 + 100%
-------
B-2
TABLE B-l. SUMMARY OF SOURCE TEST DATA FOR MATERIALS
PREPARATION AND HANDLING^
Particulate Emissions
Mg/yr
3.63
0.91
(Tons/yr)
(4.
(1.
0)
0)
trace (a)
0.91
4.54
trace <">
0.91
(1.
(5.
(1.
0)
0)
0)
Production
Gg/yr
7
87
121
231
231
12
276
.53
.8
.5
.2
.2
.8
.6
(Tons/yr)
(830,
( 96,
(134,
(255,
(255,
( 14,
(305,
000)
800)
000)
000)
000)
100)
000)
Rate
mg/kg
(Ib/ton)
5
(0.
10
(0.
0
4
(0.
19
(0.
0
3
.0
010)
.5
021)
.0
.0
008)
.5
039)
.0
.5
Control
fabric
fabric
fabric
fabric
fabric
fabric
fabric
Equipment
filters
filters
filters
filters
filters
filters
filters
Operation
hand and storage
batch house
crushing
mixing
delivery
storage
mixing
(0.007)
3.63
trace
trace
trace
(4.
(b)
(b)
(b)
0)
134
276
257
165
.2
.6
.6
.1
(148,
(305,
(284,
(182,
000)
000)
000)
000)
0
0
0
22
.0
.0
.0
.0
fabric
fabric
filters
filters
fabric filters
fabric
filters
storage
storage
mixing
conveying
(0.044)
(a) Source NEDS
(14)
(b) Trace < 1.0.
-------
B-3
Total annual emissions were based on 15.8 Tg of raw materials being processed
to melt 12.7 Tg of glass. This assumes that 85 percent of glass melted
produces a saleable container.
Stack heights for these and other plant operations are listed in
Appendix C. They range from 5 m (16 feet) to 44 m (144 feet).
The accuracy was only obtainable for batch mixing where the sample
mean was 4.5 mg/kg and the sample standard deviation was 2.0 mg/kg. The
95 percent confidence level was - 3.187 mg/kg. The accuracy of engineering
estimates was assumed to be - 100 percent.
Glass Melting
Nitrogen Oxides
Source test measurements of NO emissions from NEDS are
X
listed in Table B-2. Emission factors vary from 0.58 to 6.29 g/kg (1.60 to
12.40 Ib/ton), which clearly reflect the wide range of operating conditions
found in glass melting furnaces. The average emission factor of 3.07 g/kg
(6.14 Ib/ton) assumes 85 percent of the glass melted is flint (clear or
green) which has an average emission factor of 3.40 g/kg and 15 percent is
amber, which has an average emission factor of 1.22 g/kg. The average
emission factors for these two glass types was determined by adding the
average emission factors together and dividing by the number of values.
Alternatively, the average found by dividing the total emissions by total
production was 3.20 g/kg (6.40 Ib/ton). The difference is not significant
because the standard deviation is - 3.2 g/kg, and the 95 percent confidence
level is - 1.469 g/kg.
Standard deviations (a) were determined by the following method.
\ - .
|_m i
m
E
-------
B-4
TABLE B-2.
NO EMISSIONS FROM GLASS CONTAINER FURNACES
x
Production
Emissions
Emission Factor
Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Amber
22
23
24
Average*
Gg/yr
46.1
58.8
70.1
38.7
56.2
40.0
47.6
31.8
11.8
34.7
47.8
36.5
41.3
98.9
34.7
71.0
161.5
66.2
53.5
41.1
49.9
14.3
28.6
37.8
50.1
(tons/yr)
(50.800)
(64,800)
(77,300)
(42,700)
(62,000)
(44,100)
(52,500)
(35,100)
(13,000)
(38,300)
(52,700)
(40,300)
(45,500)
(109,000)
(38,300)
(78,300)
(178,000)
(73,000)
(59,000)
(45,300)
(55,000)
(15,800)
(31,500)
(41,700)
(55,200)
Mg/yr
5.08
50.8
349.3
38.1
151.5
51.7
299.4
167.8
62.6
184.2
253.1
226.8
117.9
381.0
49.9
78.0
733.9
198.7
198.7
198.7
29.9
17.2
35.4
46.2
153.8
(tons/yr)
(56)
(56)
(385)
(42)
(167)
(57)
(330)
(185)
(69)
(203)
(279)
(250)
(130)
(420)
(55)
(86)
(809)
(219)
(219)
(219)
(33)
(19)
(39)
(51)
(170)
§/kg
1.10
0.86
4.98
0.98
2.69
1.29
6.29
5.27
5.31
5.30
5.29
6.20
2.86
3.85
1.44
1.10
4.55
3.00
3.71
4.83
0.58
1.21
1.24
1.22
3.07
(Ib/ton)
(2.20)
(1.72)
(9.96)
(1.96)
(5.38)
(2.58)
(12.58)
(10.54)
(10.62)
(10.60)
(10.58)
(12.40)
(5.72)
(7.72)
(2.88)
(2.20)
(9.10)
(6.00)
(7.42)
(9.66)
(1.16)
(2.41)
(2.48)
(2.45)
(6.14)
Assumes 85% of glass is Flint and 15% Amber.
sum of averages for these two glass types.
Average is a weighted
-------
B-5
where: m = number of samples
N. = sample value
jj, = sample mean.
The confidence interval (CI) was determined by
ka
1/2
m
where: k - "Student's t" variable for m-1 degrees of freedom.
Sulfur Oxides
(14)
Source test measurements of SO emissions from NEDS are listed
X
in Table B-3. Emission factors vary from 0.21 to 8.35 g/kg (0.41 to 16.7 Ib/ton).
The average emission factor of 1.70 g/kg (3.4 Ib/ton) is based upon 85 percent
flint glass having an average emission factor of 1.84 g/kg and 15 percent amber
having an average emission factor of 0.93 g/kg. The standard deviation is
2.2 g/kg and the 95 percent confidence level is - 0.66 g/kg.
SO emissions are believed to come entirely from natural gas fired
X
glass melting furnaces. Increased use of oil would increase both the rate and
amoung of emissions.
Particulates
Source test measurements of particulate emissions obtained from
NEDS^ ' are listed in Table B-4. Emission factors vary from 0.13 to 1.95 g/kg
(0.27 to 3.90 Ib/ton). The average emission factor of 0.68 g/kg (1.35 Ib/ton)
is based on 85 percent flint glass having an emission factor of 0.71 g/kg and
15 percent amber glass having an average emission factor of 0.48 g/kg. The
standard deviation is 1.0 g/kg and the 95 percent confidence interval is - 0.25.
-------
B-6
TABLE B-3. SO EMISSIONS FROM GLASS CONTAINER FURNACES
x
Production
Emissions
Emission Factor
Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39 •
40
41
42
43
Gg/yr
46.1
49.9
15.9
56.6
40.0
73.9
47.6
55.3
31.8
58.8
11.8
38.7
34.8
45.6
47.8
80.6
34.8
57.4
71.0
54.5
41.0
57.9
30.0
64.3
81.7
103.9
81.5
44.5
81.5
57.2
161.5
50.8
66.2
60.3
53.5
72.6
41.1
33.6
51.7
14.3
14.3
34.5
37.2
(tons/yr)
(50,800)
(55,000)
(17,500)
(62,400)
(44,100)
(81,500)
(52,500)
(60,900)
(35,100)
(64,800)
(13,000)
(42,700)
(38,300)
(50,300)
(52,700)
(88,800)
(38,300)
(63,300
(78,300)
(60,100)
(45,200)
(63,800)
(33,100)
(70,900)
(90,000)
(115,000)
(90,000)
(49,000)
(90.000)
(63,000)
(178,000)
(56,000)
(73,000)
(66,500)
(59,000)
(80,000)
(45,300)
(37,000)
(57,000)
(15,800)
(15,800)
(38,000)
(41,000)
Mg/yr
27.2
36.3
8.2
79.8
94.4
104.3
122.5
435.5
16.3
27.2
6.4
49.9
17.2
381.0
24.5
184.2
10.9
208.7
52.6
381.0
10.9
208.7
9.1
381.0
65.3
208.7
39.0
69.0
70.0
11.8
110.7
72,6
79.8
49.9
59.9
108.9
59.9
64.4
96.2
39.0
39.0
145.2
38.1
(tons/yr)
(30)
(40)
(09)
(88)
(104)
(115)
(135)
(480)
(18)
(30)
(07)
(55)
(19)
(420)
(27)
(203)
(12)
(230)
(58)
(420)
(12)
(230)
(10)
(420)
(72)
(230)
(43)
(76)
(77)
(13)
(122)
(80)
(88)
(55)
(66)
(120)
(66)
(71)
(106)
(43)
(43)
(160)
(42)
g/kg
0.59
0.67
0.51
1.41
2.36
1.41
2.57
7.88
0.51
0.46
0.54
1.28
0.50
8.35
0.51
2.59
0.31
3.63
0.74
6.99
0.27
3.61
0.30
5.92
0.80
2.00
0.48
1.55
0.86
0.21
0.69
1.43
1.21
0.83
1.12
2.14
1.47
1.92
1.86
2.72
2.72
4.21
1.02
(Ib/ton)
(1.18)
(1.34)
(1.02)
(2.82)
(4.72)
(2.82)
(5.14)
(15.76)
(1.02)
(0.98)
(1.08)
(2.56)
(1.00)
(16.70)
(1.02)
(5.18)
(0.62)
(7.26)
(1.48)
(13.98)
(0.54)
(7.22)
(0.60)
(11.84)
(1.60)
(4.00)
(0.96)
(3.10)
(1.72)
(0.42)
(1.38)
(2.86)
(2.42)
(1.66)
(2.24)
(4.28)
(2.94)
(3.84)
(3.72)
(5.44)
(5.44)
(8.42)
(2.04)
-------
B-7
TABLE B-3. (Continued)
Production
44
45
46
Amber
47
48
49
Average*
Gg/yr
73.2
98.9
35.3
14.3
28.6
37.8
50.1
(tons/yr)
(80,700)
(109,000)
(38,900)
(15,800)
(31,500)
(41,700)
(55,200)
Emissions
Mg/yr
53.5
45.4
10.0
4.5
58.9
15.4
85.2
(tons/yr)
(59)
(50)
(11)
(05)
(65)
(17)
(94)
Emission Factor
g/kg
0.73
0.46
0.28
0.32
2.06
0.41
1.70
(Ib/ton)
(1.46)
(0.92)
(0.56)
(0.63)
(4.13)
(0.82)
(3.40)
* Assumes 85% of glass is Flint and 15% Amber. Average is a weighted
sum of averages for these two glass types.
-------
B-8
TABLE B-4. PARTICULATE EMISSIONS FROM GLASS CONTAINER FURNACES
Production
Emissions
Emission Factor
Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Gg/yr
46.1
81.7
15.9
161.5
40.0
66.2
47.6
53.5
31.8
41.1
11.8
49.9
34.7
60.3
47.8
56.6
752.8
73.9
34.8
55.3
71.0
58.8
65.4
38.7
68.0
45.6
65.3
80.6
41.0
57.4
30.0
54.5
66.0
57.9
81.7
64.3
81.7
104.3
44.5
57.2
50.8
60.3
" (tons/yr)
(50,800)
(90,000)
(17,500)
(178,000)
(44,100)
(73,000)
(52,500)
(59,000)
(35,100)
(45,300)
(13,000)
(55,000)
(38,300)
(66,500)
(52,700)
(62,400)
(830,000)
(81,500)
(38,300)
(60,900)
(78,300)
(64,800)
(72,100)
(42,700)
(74,900)
(50,300)
(72,000)
(88,800)
(45,200)
(63,300)
(33,100)
(60,100)
(72,800)
(63,800)
(90,000)
(70,900)
(90,000)
(115,000)
(49,000)
(63,000)
(56,000)
(66,500)
Mg/yr
35.4
39.0
5.4
161.5
23.6
43.6
26.3
39.9
23.6
36.3
9.1
11.8
26.3
17.2
26.3
20.9
677.7
28.1
33.6
31.8
61.7
35.4
64.4
27.2
43.6
16.3
48.1
39.9
35.4
16.3
22.7
16.3
36.3
39.9
69.9
16.3
49.9
39.9
11.8
21.8
27.2
17.3
(tons/yr)
(39)
(43)
(6)
(178)
(26)
(48)
(29)
(44)
(26)
(40)
(10)
(13)
(29)
(19)
(40)
(23)
(747)
(31)
(37)
(35)
(68)
(39)
(71)
(30)
(48)
(18)
(53)
(44)
(39)
(18)
(25)
(18)
(40)
(44)
(77)
(18)
(55)
(44)
(13)
(24)
(30)
(19)
g/kg
0.77
0.48
0.34
1.00
0.59
0.66
0.55
0.75
0.74
0.88
0.77
0.24
0.76
0.29
0.76
0.37
0.90
0.38
0.97
0.57
0.87
0.60
0.98
0.70
0.64
0.36
0.74
0.50
0.86
0.28
0.76
0.30
0.55
0.69
0.86
0.25
0.61
0.38
0.36
0.38
0.54
0.29
(Ib/ton)
(1.54)
(0.96)
(0.68)
(2.00)
(1.18)
(1.32)
(1.10)
(1.50)
(1.48)
(1.76)
(1.54)
(0.48)
(1.52)
(0.58)
(1.52)
(0.74)
(1.80)
(0.76)
(1.94)
(1.14)
(1.74)
(1.20)
(1.96)
(1.40)
(1.28)
(0.72)
(1.48)
(1.00)
(1.72)
(0.56)
(1.52)
(0.60)
(1.10)
(1.38)
(1.72)
(0.50)
(1.22)
(0.76)
(0.72)
(0.76)
(1.08)
(0.58)
-------
B-9
TABLE B-4. (Continued)
Production
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
Amber
67
68
69
70
71
72
73
74
75
76
/ w
77
78
79
80
81
82
83
Gg/yr
72.6
33.6
51.7
14.3
14.3
34.5
74.4
66.2
69.9
76.2
10.9
37.2
16.4
8.9
19.3
64.8
55.9
57.3
165.1
73.2
98.9
35.3
30.9
35.7
14.3
28.6
37.8
3.9
3.9
3.9
3.9
3.9
2.8
2.8
*•• 0 \x
2.8
1.7
1.1
0.98
0.98
0.98
0.98
(tons/yr)
(80,000)
(37,000)
(57,000)
(15,800)
(15,800)
(38,000)
(82,000)
(73,000)
(77,000)
(84,000)
(120,000)
(41,000)
(18,100)
(9,800)
(21,300)
(71,400)
(61,600)
(63,200)
(182,000)
(80,700)
(109,000)
(38,900)
(34,100)
(39,300)
(15,800)
(31,500)
(41,700)
(4,345)
(4,345)
(4,345)
(4,315)
(4,345)
(3,121)
\ y *
(3,121)
\ y *
(3,121)
(1,836)
(1,224)
(1,083)
(1,083)
(1,083)
(1,083)
Emissions
Mg/yr
57.2
8.2
25.4
13.6
13.6
26.3
47.2
45.4
89.8
107.1
119.8
34.5
3.6
7.3
10.9
111.6
104.3
111.6
99.8
33.6
25,4
21.8
14.5
20,0
6.4
10.0
8.2
2.7
2.6
2.9
2.6
2.6
1.2
1.3
1.2
1.3
0.91
0.64
0.54
0.64
0.73
(tons/yr)
(63)
(09)
(28)
(15)
(15)
(29)
(52)
(50)
(99)
(118)
(132)
(38)
(04)
(08)
(12)
(123)
(115)
(123)
(110)
(37)
(28)
(24)
(16)
(22)
(7)
(ID
(9)
(3)
(2.9)
(3.2)
(2.9)
(2.9)
(1.3)
(1.4)
(1.3)
(1.4)
(1.0)
(0.7)
(0.6)
(0.7)
(0.8)
Emission Factor
g/kg
1.13
0.24
0.49
0.95
0.95
0.76
0.63
0.68
1.29
1.40
1.10
0.93
0.22
0.81
0.56
1.72
1.87
1.95
1.61
0.46
0.26
0.62
0.47
0.56
0.44
0.35
0.22
0.69
0.67
0.74
0.67
0.67
0.43
0.46
0.43
0.76
0.83
0.65
0.55
0.65
0.74
(Ib/ton)
(2.26)
(0.48)
(0.98)
(1.90)
(1.90)
(1.52)
(1.26)
(1.36)
(2.58)
(2.80)
(2.20)
(1.86)
(0.44)
(1.62)
(1.12)
(3.44)
(3.74)
(3.90)
(3.22)
(0.92)
(0.52)
(1.24)
(0.94)
(1.12)
(0.88)
(0.70)
(0.44)
(1.38)
(1.34)
(1.48)
(1.34)
(1.34)
(0.86)
(0.92)
(0.86)
(1.42)
(1.66)
(1.30)
(1.10)
(1.30)
(1.48)
-------
B-10
TABLE B-4. (Continued)
Production
84
85
86
87
88
89
Average*
Gg/yr
47.9
43.3
43.3
29o3
44.7
44.6
58.1
(tons/yr)
(52,836)
(47,712)
(47,712)
(32,327)
(49,316)
(49,140)
(64,000)
Emissions
Mg/yr
2.1
6.6
12.0
3.9
12.4
8.2
39.3
(tons/yr)
(2.3)
(7.3)
(13.2)
(4.3)
(13.7)
(9.0)
(43.3)
Emission Factor
g/kg
0.04
0.15
0.28
0.13
0.28
0.18
0.68
(Ib/ton)
(0.08)
(0.30)
(0.56)
(0.27)
(0.56)
(0.36)
(1.35)
Assumes 85% of glass is Flint and 15% Amber. Average is a weighted
sum of averages for these two glass types.
-------
B-ll
Carbon Monoxide
Source test measurements of carbon monoxide emissions are limited
because this is not a major glass furnace emission. It can form because the
industry uses a long diffusion flame to effect uniform heat release. Properly
controlled, the emissions are negligible. The emissions are listed in
Table B-5. Emission factors vary from 0,05 to 0.13 g/kg (0.09 to 0.25
Ib/ton). The average emission factor of 0.7 g/kg. (0.13 Ib/ton) is based on
85 percent flint glass having an emission factor of 0.06 g/kg and 15 percent
amber glass having an emission factor of 0.11 g/kg. The standard deviation
was 0.2 g/kg and the 95 percent confidence level was 0.10 g/kg.
Hydrocarbons
Source test measurements of hydrocarbon emissions are also limited.
Formation occurs for the same reasons as cited for carbon monoxide. These
emissions are listed in Table B-6. Emission factors vary from 0.01 to
0.53 g/kg (0.02 to 1.06 Ib/ton). The average emission factor of 0.08 g/kg
(0.150 Ib/ton) is based on 85 percent flint glass having an average emission
factor of 0.08 g/kg and 15 percent amber glass having an average emission
factor of 0.05 g/kg. The standard deviation is 0.5 g/kg, and the 95 percent
confidence interval is - 0.178 g/kg.
Selenium
No source test measurements are available for selenium emissions
(2)
from flint glass furances. Using data supplied by GCMIV and obtained from
the technical literature, a worst-case engineering calculation was made.
Selenium is used as a decolorizer to neutralize the tint from transition metal
oxide contaminants such as iron. Approximately 0.36 Gg (395 tons) of selenium
are consumed annually in the U.S., of which about one-sixth is used by glass
container industry (0.06 Gg) . ^ If it is assumed that 40 percent of the
selenium used volatilizes and is emitted from the melting furnace, then the
emission rate would be 0.002 g/kg. The accuracy of this calculation was taken
-------
B-12
TABLE B-5. CO EMISSIONS FROM GLASS CONTAINER FURNACES
Production
Emissions
Emission Factor
Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Amber
20
21
22
Average*
Gg/yr
65.3
65.3
53.5
41.1
49.9
60.3
56.6
73.9
55.3
22.9
16.4
18.6
102.5
74.4
66.2
69.9
76.2
30.9
35.7
14.3
28.6
37.8
50.3
(tons/yr)
(72,000)
(72,000)
(59,000)
(45,300)
(55,000)
(66,500)
(62,400)
(81,500)
(60,900)
(25,200)
(18,100)
(20,500)
(113,000)
(82,000)
(73,000)
(77,000)
(84,000)
(34,100)
(39,300)
(15,800)
(31,500)
(41,700)
(55,500)
Mg/yr
3.6
3.6
2.7
2.7
4.5
5.4
3.6
3.6
3.6
1.8
0.91
0.91
6.4
3.6
3.6
3.6
3.6
2.7
3.6
1.8
2.7
3.6
3.1
(tons/yr)
(4)
(4)
(3)
(3)
(5)
(6)
(4)
(4)
(4)
(2)
(1)
(1)
(7)
(4)
(4)
(4)
(4)
(3)
(4)
(2)
(3)
(4)
(3)
g/kg
0.06
0.06
0.05
0.07
0.09
0.09
0.06
0.05
0.07
0.08
0.06
0.05
0.06
0.05
0.05
0.05
0.05
0.09
0.10
0.12
0.09
0.10
0.07
(Ib/ton)
(0.11)
(0.11)
(0.10)
(0.13)
(0.18)
(0.18)
(0.12)
(0.10)
(0.13)
(0.16)
(0.11)
(0.10)
(0.12)
(0.10)
(0.10)
(0.10)
(0.10)
(0.18)
(0.20)
(0.25)
(0.19)
(0.19)
(0.13)
Assumes 85% of glass is Flint and 15% Amber.
sum of averages for these two glass types.
Average is a weighted
-------
B-13
TABLE B-6. HC EMISSIONS FROM GLASS CONTAINER FURNACES
Production
Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Amber
34
35
36
Average*
Gg/yr
32.7
34.8
71.0
65.4
67.9
65.3
95.2
105.2
65.3
65.3
53.5
41.1
49.9
60.3
56.6
73.9
55.2
22.9
102.5
72.6
33.6
51.7
14.3
14.3
34.5
74.4
66.2
69.9
76.2
64.4
56.3
30.9
35.7
14.3
28.6
37.8
53.4
(tons/yr)
(36,000)
(38,300)
(78,300)
(72,100)
(74,900)
(72,000)
(105,000)
(116,000)
(72,000)
(72,000)
(59,000)
(45,300)
(55,000)
(66,500)
(62,400)
(81,500)
(60,900)
(25,200)
(113,000)
(80,000)
(37,000)
(57,000)
(15,800)
(15,800)
(38,000)
(82,000)
(73,000)
(77,000)
(84,000)
(71,000)
(62,000)
(34,100)
(39,300)
(15,800)
(31,500)
(41,700)
(58,900)
Emissions
Mg/yr
17.2
5.4
7.3
0.8
7.3
10.0
12.7
12.7
0.9
0.9
0.9
0.9
2.7
2.7
0.9
1.8
1.8
0.9
0.9
6.4
3.6
3.6
1.8
1.8
2.7
1.8
1.8
1.8
1.8
9.1
7.3
0.9
1.8
0.9
0.9
1.8
4.0
(tons/yr)
(19)
(6)
(8)
U)
(8)
(11)
(14)
(14)
(1)
(1)
(1)
(1)
(3)
(3)
(1)
(2)
(2)
(1)
(1)
(7)
(4)
(4)
(2)
(2)
(3)
(2)
(2)
(2)
(2)
(10)
(8)
(1)
(2)
(1)
(1)
(2)
(5)
Emission Factor
g/kg
0.53
0.16
0.10
0.01
0.11
0.16
0.13
0.12
0.01
0.01
0.02
0.02
0.05
0.05
0.02
0.02
0.03
0.04
0.01
0.09
0.11
0.07
0.12
0.12
0.08
0.02
0.03
0.03
0.02
0.14
0.13
0.03
0.05
0.06
0.03
0.05
0.08
(Ib/ton)
(1.06)
(0.31)
(0.20)
(0.03)
(0.21)
(0.31)
(0.27)
(0.24)
(0.03)
(0.03)
(0.03)
(0.04)
(0.10)
(0.09)
(0.03)
(0.05)
(0.07)
(0.08)
(0.02)
(0.18)
(0.22)
(0.14)
(0.25)
(0.25)
(0.16)
(0.05)
(0.05)
(0.05)
(0.05)
(0.28)
(0.26)
(0.06)
(0.10)
(0.13)
(0.06)
(0.10)
(0.15)
* Assumes 85% of glass is Flint and 15% Amber.
of averages for these two glass types.
Average is a weighted sum
-------
B-14
as + 100 percent. This is determined by dividing the total selenium
(0.06 Gg) by the total glass melted (12.7 Tg) and multiplying by 0.4 (the
amount of selenium emitted) alternately, the selenium represents no more than
(32)
0.001 percent of the glass or 0.008 g/kg of glass produced. If 75 percent of
the production contains selenium, then the emission rate would also be 0.002 g/kg.
This is determined by dividing O.Oi g/kg (amount of selenium) by 1.25 (batch to
produce 1 unit of glass) and then multiplying by 0.40 (percent selenium emitted)
and 0.75 (percent glass using selenium).
Forming and Finishing
Emissions measurements from the forming and finishing operations
were not available; hence, engineering calculations, considering worst-case
situations, were used to determine the severity of emissions from this area.
Forming
During forming, an emulsion containing oil or silicone and water
is sprayed onto the molds. Approximately 1.4 g (1/20 ounce) of liquid is
sprayed per container produced. If one assumes that only oil is used as a
lubricant (silicone is also popular) and that the mixture of oil to water is
1:125, a normal situation. Then under these conditions, the emission rate
would be 0.035 g/kg, considering the production in 1974 of approximately 40
(2)
billion containers and 12.7 Tg. These were considered minor. The calculation
is shown below. Accuracy is taken to be +_ 100%,
Oil/container - 1.4 g * 125 » 0.0112 g
12
Container weight - 12.7 x 10 g * 40 x 1<
Emission rate - 0.0112 * 0.318 = 0,035 g/kg
Surface Treatment
Approximately 30 percent of containers produced receive a surface
treatment to improve resistance to scratching and to facilitate handling.
12 9
Container weight - 12.7 x 10 g * 40 x 10 containers = 0.318 kg/container
-------
B-15
Hot containers are subjected to a tin or titanium chloride vapor. Emissions
consist of metal oxide and hydrated metal chloride particulates and HC1.
Data received from glass container manufacturers showed the average
consumption of material in surface treatment operations was 0.12 g/kg (0.24 Ib/ton),
of which 60 percent or 0.07 g/kg (0.14 Ib/ton) escapes to the atmosphere. This is
composed of 22 percent metal oxide, 44 percent hydrated metal chloride, and 34
percent HC1. Emission rates were then calculated for tin compounds to be 0,02
g/kg (0.03 Ib/ton) of tin chloride, 0.03 g/kg (0.06 Ib/ton) of hydrated tin
chloride, and 0.02 g/kg (0.05 Ib/ton) of HC1. These are based on the tin
compounds comprising approximately 40 percent of the total weight input and
HC1 approximately 20 percent of the total weight input. Accuracy of the data
was taken as + 100 percent.
Annealing
No emission data were available for gas-fired annealing lehrs;
therefore, emission factors were estimated from other data on gas combustion.
A modern recirculating air type lehr consumes 11 to 17 m /hr (400 cfh to
600 cfh) when annealing 91 Gg (100 tons) of glass per day. Lehrs of older
design can consume 34 to 57 m3/hr (1200 cfh to 2000 cfh) ^ . On a worst case
3 3
basis (57 M /hr) would require 0.0062 m /kg of glass produced. For a plant
3
producing 319 Mg/day (352 ton/day) this would amount to 91 m /hr. With a
3
heating value of natural gas (1000 Btu/cf or 37.3 million joules/m ) this
amounts to 0.93 million joules per second or about 0.23 million joules per
kg of glass (200,000 Btu/ton).
(53)
Using tests on gas-fired burners , emission data was determined
as shown in Table B-7. Converting these on a basis of 0.24 million joules/kg
of glass gave the emission factors for annealing shown in Table B-8.
Decorating
Glass containers are sometimes decorated with vitrifiable glass
enamels or organic materials. Emissions are derived from organic solvents
and binders used in the coatings. Data supplied by glass manufacturers
-------
B-16
TABLE B-7. EMISSIONS FROM GAS-FIRED BURNERS
(53)
EmissionSj
mg/fcg
Partic- Hydro-
Test NO SO ulates CO caroons
X X
15 60.2 - 9.0
17 150 0 2.58
18 38.7 0 3.0
19 25.8 - 11.2
Average and 69 0 5.2
95% confi- -113% -196%
dence limits
Standard -47 0 -5.2
deviation
5.6 1.29
8.6
11.2 9.5
12.9 6.9
9.5 6.0
-55% -144%
-2.75 -3A
TABLE B-8. EMISSION FACTORS FOR ANNEALING LEHRS
Emission
Emission mg/kg
NO +
x 16 - 113%
SO 0
X
Participates 1.2 - 196%
CO 2.2 - 55%
Hydrocarbons 1.4 - 144%
factor
Clb/ton)
(0.032)
(0)
(0.0024)
(0.0044)
(0.0028)
-------
B-17
would indicate that only 3 percent of containers are coated. Point source
emission data were not available. Materials balances taken from NEDS
were used as the basis for estimating the emission rate for hydrocarbons from
decorating lehrs. Table B-9 lists the emission data. The hydrocarbon emission
rate was 4.37 g/kg (8.73 Ib/ton). The total hydrocarbons emitted annually from
all sources is estimated to be 1.44 Gg (1,590 tons). This was determined in the
following manner.
12
1974 container shipments = 11.005 x 10 g
Q
Decorated container (3%) = 330.1 x 10 g
9
Total Emissions = Shipments x Emission Rate = 1.44 x 10 g
-------
B-18
TABLE B-9. HC EMISSIONS FROM GLASS CONTAINER DECORATING
OPERATIONS
Production
1
2
3
4
5
Average
Gg/yr
21.2
20.2
15.7
15.7
20.2
18.6
(tons/yr)
(23,400)
(22,300)
(17,300)
(17,300)
(22,300)
(20,500)
Emissions
Mg/yr
80
80
80
80
80
80
(tons/yr)
(88)
(88)
(88)
(88)
(88)
(88)
Emission Factor
g/kg
3.76
3.95
5.09
5.09
3.95
4.37
(Ib/ton)
(7.52)
(7.89)
(10.17)
(10.17)
(7.89)
(8.73)
(141
* Material Balance Data Taken from NEDSV J
-------
APPENDIX C
STACK HEIGHTS FROM THE
VARIOUS PHASES OF GLASSMAKING
-------
TABLE C-l. TYPICAL STACK HEIGHTS OF BATCH HANDLING, TREATMENT
AND DECORATING OPERATIONS FOR SODA/LIME CONTAINER GLASS
Batch Handling
<40 m >40 m
No. of Stacks Height, m No. of Stacks Height, m
(3) 5 44
22
24
(4) 36
Total 9 Average 23 Total 1 Average 44
Median 24 Median 44
Treatment
No. of Stacks
(4)
(3)
(3)
(3)
Total 19
Height, m
8
12
13
14
15
17
20
23
25
38
Average 19
Median 15
Decoration
No. of Stacks Height, m
(3) 12
(2) 13
Total 5 Average 12
Median 12
o
-------
C-2
TABLE C-2. TYPICAL STACK HEIGHTS OF FLINT.AND
AMBER CONTAINER GLASS FURNACES11 '
<40
No. of Stacks
(9)
(2)
(7)
(5)
(15)
(2)
(3)
(3)
(7)
(4)
(6)
(25)
Total 99
Flint
Height, m
5
6
10
12
13
14
17
18
19
20
21
23
24
25
26
27
30
33
34
36
38
39
40
Average 27
Median 25
Glass
>40
No. of Stacks
(2)
(8)
(3)
(2)
(7)
(2)
(3)
Total 27
m Amber Glass
Height, m No. of Stacks Height, m
41 17
43 20
44 23
45
46
47
49
Average 45 Total 3 Average 2
Median 45 Median 2
-------
APPENDIX D
STATE LISTING OF TOTAL
EMISSIONS AS OF 1972
-------
D-l
TABLE D-l. STATE LISTING OF TOTAL EMISSIONS AS OF 1972
State
1 ALA&AfA
2 ALASKA
2 ARIZONA
4 ARKANSAS
5 CALIFORNIA
6 COLORADO
7 CONNECTICUT
8 DELAWARE
9 FLORIDA
10 GEORGIA
11- HAKAII
12 IDAHO
13 ILLINOIS
14 INDIANA
15 IOWA
16 KANSAS
17 KENTUCKY
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
2002000.0
1.53000
16340000.0
12.50000
3265000.0
2.49000
1619000.0
1.24000
5675000.0
4. 33000
3156000.0
2.1*1000
365600.0
0.27900
130200.0
0.09930
2430000.0
1.86000
2331000.0
1.78000
251200.0
0.19200
2430000.0
1.65000
3584000.0
2.74000
2202000,0
I.fe8000
2579000.0
1.97000
3358000.0
2.56000
1654000, D
1.42000
SO2
1226000.0
1.91000
222600.0
0. J4700
200200.0
0.31100
205400.0
0.31900
2557000.0
3.98000
473300.0
0.73600
1227000.0
1.91000
420700.0
0.65500
1755000.0
2.73000
1635000,0
2.54QOO
232000.0
0.36100
59140.0
0.09200
3714000.0
5.78000
3036000.0
4.72000
397400.0
0.61800
225000.0
0.35000
1627000.0
2.53000
NO
X
26160U.O
2.27000
31990.0
0.27700
75100.0
O.feblUO
77310.0
0.67UOO
796800.0
6.91000
116600.0
1.01000
152200.0
1.32000
45720.0
0.396UO
410300.0
3.56000
294200.0
2.55000
40790.0
0.35400
33220.0
0,28800
665100.0
5.77000
414400.0
3.59000
137700.0
1.19000
109900,0
0.95300
30200U.O
2.62000
Hydro-
carbons
342100.0
1.29000
iw&oo.o
0.53200
171100.0
0.64700
281700.0
1.07000
1914000.0
7.24000
294400.0
1.11000
259400.0
0,98100
77510.0
0.29300
536200.0
2,03000
526700.0
1,99000
62720.0
0.23700
163600.0
0.61900
1343000.0
5.08000
675100.0
2.55000
4Q0800.0
1.52000
742800.0
2.81000
274600. 0
1.0400B
CO
372600.0
2.04000
472200.0
2.50000
178300.0
0.97600
225800.0
1.24000
1987000.0
10.90000
105800.0
0.57900
92690.0
0.50700
24580.0
0.13500
3502000.0
19.20000
705400.0
3.86000
84750.0
0.46400
516300.0
2.84000
412500.0
2.26000
182100.0
0.99700
90720.0
0.49700
174600.0
0.95600
2193UO.O
1.20000
-------
D-2
TABLE D-l. (Continued)
State
16 LOUISIANA .
19 MAINE.
20 MARYLAND
Zl MASSACHUSETTS
Z2 MICHIGAN
23 MINNESOTA
24 MISSISSIPPI
25 MISSOURI
26 MONTANA
27 NEBRASKA
28 NEVADA
29 NEW HAMPSHIRE
30 NEW JERSEY
31 NEW MEXICO
v
42 NEW YORK
S3 N CAROLINA
31 N DAKOTA
35 OHIO
36 OKLAHOMA
37 OREGON
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
1651000.0
1.26000
1036000.0
0.79200
657300.0
0.50200
802700.0
0.61300
2801*000.0
2.14000
3056000.0
2.33000
m90000.0
1.11000
2839000,0
2.17000
4975000.0
3.60000
3049000.0
2.33000
3155000.0
2.41000
326500.0
0.24900
615800.0
0.62300
354SOOO.O
2.71000
2704000.0
2.0600U
2203000.0
1.66000
2654000.0
2.16000
3054000.0
2.33000
2276000.0
1.74000
2885000.0
2.20000
S02
585800.0
0.^1100
770700.0
1.20000
1352000.0
2.10000
3640000.0
5.97000
3513000.0
5,46000
846600.0
1.32000
280300.0
0.43600
1259000.0
1.96000
177000.0
0.27500
137100,0
0.21300
263100.0
0.40900
325600.0
0.50700
2922000.0
4.55000
441400.0
0.68700
5137000.0
7.99000
2298000.0
3.56000
326700.0
0'. 51100
4062UOO.O
6.32000
163400.0
0.25400
372500.0
0.57900
NO
X
219000.0
1.90000
54270.0
0.47000
2151UU.O
1.86000
322300.0 '
2.79.000
548000.0
4.75000
165000.0
1.60000
67010.0
0.75400
287500.0
2.49000
34650.0
0.30UOO
50940.0
0.44200
56500.0
0.50700
3606U.O
0.31300
323400.0
2.80000
109800.0
0.95200
721400.0
6.25000
336400.0
2.93000
61110.0
0.53000
735800.0
6.61000
130000.0
1.13000
62710.0
0.54400
Hydro-
carbons
1741000.0
6.58000
71970.0
0.27200
302300.0
1.14000
463100.0
1.75000
734000.0
2.78000
386000.0
1.47000
350200.0
1.32000
S88400.0
2.22000
174200.0
0.65600
255600.0
0.96600
36140.0
0.13700
44430.0
0.16800
786600.0
2.97000
310200.0
1.17000
1353000.0
5.11000
465100.0 .
1.76000
73930iO
0.28000
1244000.0
4.70000 '
674700.0
2.55000
204800.0
0.774UO
CO
139900.0
4.60000
61430.0
0.33600
163400,0
0.89400
190400.0
1.04000
299400.0
1.64000
150700.0
0.62500
228200.0
1.29000
268500.0
1.47000
230500.0
1.26000
59590.0
0.32600
28700.0 ^
0.15700
30200. 0>
0.16500
281400.0
1.54000
49460.0
0.27100
551600.0
3.02000
371500.0
2.03000
22340.0
0.12200
"482700.0
2.64000
200800.0
1.10000
304900.0
1.67000
-------
D-3
TABLE D-l. (Continued)
State
3s PENNSYLVANIA
39 RHOOi ISLAND
*0 S CAROLINA
HI S DAKOTA
H2 TENNLSSEE
43 TEXAS
44 UTAH
45 VERMONT
16 VIRGINIA
47 WASHINGTON
48 nl VIRGINIA
49 WISCONSIN
SO WYOMING
US TOTALS
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
3132000.0
2.39000
113200. U
0.06640
1209000.0
0.923UO
2861000.0
2.18000
1769000.0
1.37000
9302000. U
7.10000
2461000.0
1.68000
292100.0
0,22300
1607000.0
1.23000
2204000.0
1.66000
1261000.0
0.96200
2180000.0
1.66000
2851000.0
2.18000
131000000.0
SO2
5603000.0
0.72000
519900.0
0.80900
1076UOO.O
1.67000
69420.0
0.10800
1307000.0
2.03000
1817000.0
2.63000
285400.0
0.44400
112600.0
0.17500
1388000.0
2.16000
626400.0
0.97500
1455000.0
2.2600C
1216UOO.O
1.69000
513000.0
0.79600
64300000.0
NO
X
782200.0
*.7«OUO
3F.760.0
3.33600
146300.0
1.27000
16560.0
0.16100
261<100.0
2.29000
635500.0
6.03000
46410.0
0.42UOO
13710,0
0.11900
197600.0
1.71000
126300.0
> 1.09000
306500.0
2.66000
231300.0
2.00000
70570,0
0.61200
1150000U.O
Hydro-
carbons
1331000.0
S.0300U
93730.0
0.354UO
260500.0
0.985UO
91110.0
0.34400
340900.0
1.29000
4139000.0
lb. 60000
112800.0
0.42600
2S460.0
0.0963U
41S200.0
1.57000
361800.0
1.37000
172800.0
0.65300
362600.0
1.37000
275200.0
1.04000
26400000.0
CO
527000.0
2.68000
29390.0
3.16100
483900.0
2.65000
23480.0
0.12900
200300,0
1.10000
1501000.0
8.22000
46840.0
0.25600
14190.0
0.07770
235100.0
1.29000
425500.0
2.33000
435100.0
2.36000
161300.0
0.68300
20870.0
0.11400
16300000.0
-------
APPENDIX E
. rnwVKRSION FACTORS
-------
E-l
TABLE E-l. CONVERSION FACTORS
To Convert From
Btu
degree Fahrenheit (F)
foot (ft)
foot3 (ft3)
inch (in.)
mile2 (mi)
pound (mass, Ib)
ton (short)
Prefix Symbol
tera T
giga G
mega M
kilo k
milli m
micro M-
To
joule (J)
degree Celsium (C)
meter (m)
3 3
meter (m )
meter (m)
2 2
meter (m )
kilogram (kg)
gigagram (Gg)
PREFIXES
Multiplication
Factor
10
io12
IO9
io6
io3
io-3
io"6
Multiply By
1.055 x IO3
t°c • (t°p - 32)71.8
3.048 x IO"1
2.832 x IO"2
2.540 x IO"2
2.590 x IO6
4.536 x 10'1
9.072 x 10'4
Multiply By
,«12
1 Tg = 1 x 10 g
1 Gg = 1 x IO9 g
1 Mg = 1 x 10 g
3
1 km = 1 x 10 g
-3
1 mm = 1 x 10 m
1 pm = 1 x 10 m
-------
APPENDIX F
GLOSSARY OF TERMS
-------
F-l
APPENDIX F
GLOSSARY OF TERMS
ANNEALING - Controlled heating and cooling of glass to remove objectionable
stresses.
BATCH - Mixed glass raw materials.
BATCH HOUSE - Structure where raw materials are sorted, weighed, and mixed.
BOOSTING - Supplemental electrical heating in the glass furnace.
CHECKERS, CHECKERWORK - A network of refractory ducts on both sides of a
glass furnace, used as heat exchangers.
GULLET - Scrap glass that is to be recycled.
FINING - Process of removing gas bubbles from molten glass.
LEHR - A long oven for annealing glass continuously.
MELT - The molten glass in the glass furnace.
REFINING - Process of conditioning the molten glass to remove gas bubbles
and undissolved grains of sand.
TANK - That part of the glass melting furnace which holds the molten glass,
made of refractory material.
REGENERATORS - Chambers of refractory checkerwork on both sides of the
melting furnace. Hot exhaust gases from the furnace pass through one
regenerator and heat it while combustion air passes through the other
regenerator and is heated. At intervals of 15-20 minutes, the flow is
reversed.
-------
APPENDIX G
LETTERS OF COMMENT
-------
APPENDIX G
LETTERS OF COMMENT
This appendix contains letters from the Glass Packaging Institute
(GPI) (formerly GCMI), EPA, and Battelle that address certain points
raised by GPI during a review of the Preliminary Source Assessment Document.
Some of the comments and suggestions for revision were utilized. The
letters from EPA and Battelle are in response to questions raised by GPI,
and were included at the request of GPI and EPA.
-------
G-2
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
RESEARCH TRIANGLE PARK
NORTH CAROLINA 2771 1
May 28, 1976
Dr. John Turk
Glass Container Manufacturers Institute
1800 K Street, N.W.
Washington, D.C. 20060
Dear John:
I appreciate the opportunity for Dr. Schorr and myself to
get together with the Air Quality Task Force of GCMI to discuss
your comments on the PSAD for the-glass container industry. At
that meeting, several questions were left unresolved, and I
indicated that I would attempt to provide answers for or obtain
further clarifications as to EPA position regarding certain policy
issues. These issues or questions are outlined below along with
my response.
1) Question: What is the national primary ambient air quality
standard for nitrogen oxides? What reference method is used for
ambient sampling for nitrogen oxide and what species is measured?
What is the national primary ambient air quality standard for
sulfur oxides?
Answer: The national primary ambient air quality standard for
nitrogen oxides is nitrogen dioxide ,(N09). The standard for
NC^ is 100 micrograms/cubic meter-annual arithmetic mean. The
proposed reference method 'for measuring N0? in ambient conditions
is chemiluminesce. This method, however, has not been promulgated,
but is expected to be so in about 2 months. This method indirectly
measures N02 by first measuring NO in the sample and then reducing
the N02 to NO in a reducing chamber and remeasuring the total NO
concentration in the sample. The difference in the two measurements
represents the NO- concentration.
-------
G-3
the national primary ambient air quality standard for sulfur oxides
is sulfur oxides (S0x> measured as sulfur dioxide (SO ). The
standard for S(>x measured as S02 is 365 micrograms per cubic meter-
maximum 24 hours concentration.
Conforming with these ambient -air quality for these pollutants, the
PSAD will referenced these standards as N00 and SO .
2 x
2) Question: If the ambient air quality standard for nitrogen
oxide is for N02, how is EPA justified in comparing an NO emission
X
which is predominately NO species emitted from a glass melting
furnace to N0_ ambient air quality standard.
Answer: I have discussed this issue of comparing a predominately
NO emission species against a NO,, standard with our physics and chemistry
experts at RTF. In making this comparison, this assumes that the
NO- emission factor is equal to the NO emission factor. They feel
(and I concur) that this assumption is a fairly good one for the
following reason. Once the plume has been diluted sufficiently with
air (dispersion calculations show that the plume is diluted
approximately 1000 to 1 at the point where it touches ground),
the photochemical conversion of NO to NO- is quite rapid
particularly in a urban environment. Studies has shown that
the NO half life due to conversion to NO- is approximately one
hour. In a rural environment, it is approximately an order of
magnitude higher. Based on this evidence, our assumption of using
t
an NO emission factor to represent NO- emission factor appears
x c 2.
to be good. Unless the task force has ambient air sampling data
to the contrary, we feel justified in comparing the N0x emission
factor to the NO- standard.
Secondly, if we were to use NO as the reference instead of NO,,
for our source severity calculation, the risk factor (TLV X 8_ X _JL_ )
24 100
-------
G-4
would be the same as the ambient air quality standard for
3
N0_. The TLV for NO is 30 mg/m and when adjusted using our safety
3
factor of 1 (_8 x 1 ), the risk factor becomes 100 pg/m for the
source severity calculation which is the same as ambient air quality
standard for NO-. Therefore, in either case, the source severity
calculation would remain the same.
3) Question: The safety factor of 1 appears to be conservative
in the risk factor based on a comparison between TLV's and ambien
air quality standards for criteria pollutants.
Answer: I have reviewed the data comparing our risk factor for
source severity based on TLV's and then on an Ambient Air Quality
Standard for criteria pollutants. It would appear that we have
one to one correspondence between the two basis for determining
exposure which would cause an adverse health effect if for SO
and particulates, a safety factor of 10 is used; for N0_, a safety
factor of 33 is used; and for CO, a safety factor of 1000 is used.
I have discussed this issue with several people at RTF. Our
conclusion is to leave the safety factor at 100. The reasoning
is as follows: Expert judgement recommends a safety factor of 100
to account for exposure to a general population instead of a healtier
working population. Also, there is no reason to believe that all other
pollutants will have the same safety factor as evidenced by CO being
off by a factor of 1000. Additionally, the health effects are better
known and studied for the criteria pollutants for which ambient ao.r
quality standards have been"set, than for those pollutants for which
no standards have been set to date. For these reasons, we will
continue to use the safety factor of 100 in our source severity
calculations for non criteria pollutants.
-------
G-5
If my responses to these questions or issues are not satisfactory to
the Task Force, I will be glad to discuss these issues with them in
person at their convenience. However, I do hope they shed some
light as to our current thinking on these issues. Again, let me
thank you and the Task Force for your help in reviewing this document.
Sincerely,
E. J. Wooldridge
Chemical Processes Branch
cc:
Dr. Dale A. Denny
Dr. J. R. Schorr
-------
G-6
Glass Container Manufacturers Institute, Inc.
1800 K Street NW, Washington, D C 2i
(202)872-1280
TWX: 7108229337 GLASS VVSH
July 2, 1976
Mr. Ed Wooldridge
Reasearch Triangle Park
Environmental Protection Agency
Research Triangle Park, NC 27711
Dear Mr. Wooldridge:
Enclosed herein are limited comments from the GCMI Air Quality Task Group
relative to your June 1976 Draft Source Assessment Document. In view of the
time limitation imposed by your current schedule for finalizing this report,
we have had to hastily distribute copies of the draft, and have been unable
to give it enough attention for a comprehensive review. A full word by word
analysis of this report has not been attempted. However, several questions
have been raised where treatment was questionable, and the apparent impact
of a glass manufacturing operation was seriously in error. I would like to
point out that this draft still contains several errors, significant internal
contradictions, incomplete editing, and also-shows a lack of attention to
errors pointed out in earlier drafts for which there were agreed changes to
be included .
The incorporation of the attached suggestions will still not mean that we
are in accord with all of the remaining content of the report for the reasons
given above.
The Air Task Group is very much concerned at the rush to issue this report in
view of its current state, and the lack of time for comment allotted to the
GCMI Task Group, which has attempted the fullest cooperation possible in the
preparation of this document. Ue wish to call to your attention to earlier
communication between Battelle, EPA, and GCMI concerning the time available
for proper preparation of this source assessment. At our first meeting,
September 30, 1975, in Washington, you clearly stated that there were no time
constraints in publishing on this report. It was with this understanding that
there would be adequate time for constructive comment and corrective action
that the Committee agreed to cooperate in the source assessment study. We
regret that we consider further time necessary to complete this document so that
appropriate data is included and objective analysis is performed, as this
-------
G-7
process is time consuming and expensive to our membership, as well as to
Battelle and EPA. We sincerely request that the following conments be
considered objectively and that appropriate revisions to the source draft
be made and reviewed before any consideration of a printing of the final
document is considered.
Nitrogen Oxides from Glass Melting Furnaces
In earlier meetings, we have questioned several aspects of the treatment of
nitrogen oxides from glass melting in determining ambient impact of the plant.
First we questioned representation of total nitrogen oxides as the regulated
pollutant. It was clarified that nitrogen dioxide is the regulated pollutant.
Then we were informed that in converting a threshold limit value for nitric
oxide to a simulated ambient standard the value for nitrogen oxide would come
out equivalent for nitrogen dioxide. (This relation depending on the use of
1/100 "safety factor"). We had earlier commented on the questionable use of
the safety factor, and it is appropriate to further point out that were this
a recognized approach for setting a standard there would have been no justifi-
cation not to regulate total nitrogen oxides, rather than consider nitrogen
dioxide the pollutant.
We also questioned treatment of the furnace emissions as nitrogen dioxide, since
in nitrogen oxide emissions for combustion processes, glass melting not
exclusive, it is well established that approximately 95% of the total oxides
are emitted as nitric oxide (NO), We seriously questioned this treatment in
determination of a maximum ground level concentration only a few minutes
residence time from the point of discharge. The EPA response indicated the
treatment is based on "rapid" conversion of nitric oxide to nitrogen dioxide in
the atmosphere, with the half life on one hour cited. Our own reference
searches have not turned up a good documentation of atmospheric residence,
rather referring to a range of anywhere from a few hours to a few days. However,
we consider our question still valid, for reasons which will be shown later.
Referring to Table 12 of the June draft, we note the "representative" plant
contains a single glass melting furnace discharging through a single stack.'
We note the glass tonnage of this furnace has been increased from 80,000 tons
per year in the preceeding draft, to 115,000 tons per year in the current
draft. Using a typical production schedule of 350 operating days per year,
115,000 tons per year equates to 329 tons of glass per day, a very high
tonnage for a "representative" furnace. Since the nitrogen oxide emission
rate is directly correlated to production tonnage, this illustration would be
more representative of near the maximum level of nitrogen oxide emissions,
not a "representative" level.
We note in the same section of Table 12 gas flow rate and temperature are
shown. These are unchanged from the earlier draft. Considering a conservative
5.5 million BTU's per ton to melt glass, and a typical 8% stack oxygen level,
we can only account for the production of about 250 tons of glass per day at
this exhaust flow. This, together with a nitrogen oxide emission rate based
on the higher tonnage, produces an imbalance which shows nitrogen oxide
emission concentrations (which will be directly reflected in downwind level
concentrations) about 32% too high. (Supporting calculations are attached to
this letter).
-------
G-8
From Table 16, we find the maximum downwind concentration is located a
distance of 880 meters (2,887 feet) from the stack. From Table 12, we pick
the average wind speed used in the dispersion calculation of 6 meters per
second (19.8 feet per second). Relating these two, we can determine the
residence time of the nitrogen oxide from the point of stack discharge to the
point of maximum downwind concentration is 2.43 minutes.
Using the one hour half life for the oxidation of nitric oxide, as provided
by Mr. Wooldridge, we can calculate that in the 2.43 minutes atmospheric
residence time of the nitrogen oxides only 3% of the emitted nitric oxide
will have time to convert to nitrogen dioxide. Assuming that we start with
the well established 95% nitric oxide: 52 nitrogen dioxide ratio, at the end of
2.43 minutes in the atmosphere we calculate that a total of 7.85% of the nitrogen
oxide is nitrogen dioxide at the point of maximum downwind concentration.
If we concede (ignoring the 32% high value for nitrogen oxides) 11.25 grams
per second emmitted, only 1.35 grams per second of this can legitimately be
treated as nitrogen dioxide for the purpose of determining ambient dispersed
concentration at the maximum downwind concentration location. If this correction
is not made, we calculate that the nitrogen oxide value represented is 833% of
the correct value.
In referring to Table 14, we first note the regulated pollutant is again
represented as total oxides of nitrogen. The pollutant, of course, should be
nitrogen dioxide. We next note the represented quality standard is the annual
arithmetic mean of 100 micrograms per cubic meter, which, for the purpose of
this exercise is assumed here as a 24 hour standard. We do not believe this
is a legitimate assumption. In fact, we believe it introduces further error
in the analysis. We also believe other authors have related an annual mean
to a 24 hour standard by a ratio of about .3. Indeed, in reviewing the ratios
between annual and 24 hour for particulates (primary and secondary) and for
sulfur dioxide, pollutants where both annual and 24 hour standards exist, we
find the Federal EPA related them as ratios of .29, .40, and .22, respectively.
We submit that it would be consistant with EPA practice for a ratio of .3 to
be used in converting from a one year to a 24 hour standard. With reference to
Table 14, an appropriate 24 hour standard for nitrogen dioxide would, then, be
333 micrograms per cubic meter. Calculating in the other direction, converting
a 3 minute maximum to an annual mean, one would insert the multiple of (.36) x
(.3), or .11, as the scaling factor. We take note that Turner, who provided
the basis for the scaling factor from the three minute sample to a 24 hour
sample, had questioned using his own data for scaling for sampling times of
greater than two hours. This points out that even experts in the field of
ambient dispersion, consider these statistical treatments uncertain, and
certainly, in our view, adds question to the practice of weighting every
calculation to the most severe disadvantage of the source.
To summarize the corrections we feel are legitimate, we have prepared a new
Table 14 for your study.
If we use the new figures that we calculate for maximum downwind concentration
of nitrogen dioxide, we would expect to see an increase in background nitrogen
dioxide concentration of only 4.55 micrograms per cubic meter, which we doubt
-------
G-9
•is measurable with current analytical accuracy. To further illustrate the
low ambient nitrogen dioxide impact within the vicinity of the plant, we
are attaching a report summarizing ambient sampling of a members operating
factory. This was done under the indirect supervision of the San Joaquin
County Air Pollution Control District of California and EPA Region 9, and has
been given to Region 9 in its entirety. The ambient samplers were located
where results from EPA's UMAMAP computer program indicated a maximum ground
level concentrations would be found. Refer to Table 7, Figures 1 and Figures
6-9, for the nitrogen dioxide analyses. As is apparent in the data, the
plant contribution to the nitrogen dioxide background is not evident. It
should be specifically noted, that the measured total nitrogen dioxide
background is not as high as the predicted contribution from the furnace
alone as represented in the Battelle report.
Hydrocarbons from Forming and Finishing - (Principally Decorating)
Table 2 shows an emission factor of hydrocarbons from forming and finishing
as .36 grams per kilogram of glass produced (.72 pounds per ton). Based on
an annual decorated tonnage of 364,000 tons (3% of total annual production)
the industry emits from the forming - finishing operation a calculated 131
tons per year of hydrocarbons. While we believe the factor of .36 grns/Kg
is high, our real question of the treatment of hydrocarbons relates to the
assumptions used for the dispersion calculations.
From Table 15, Source 2, we see the single source used in the calculation
emits 79 tons of hydrocarbons per year, or an astonishing 60.3% of the total
annual emission for the industry.
Calculating another way with the Table 2 and Table 15 data, we can calculate
this sample source has to decorate 219,000 tons of glass a year, or 190* of
the glass produced in this sample plant (reference Table 12). Either of these
calculations illustrates that the emission assumptions used for dispersion
calculations are absurdly high.
Even assuming the hydrocarbon factor from Table 2 to be correct, and further
assuming the rest of the calculation is right, choosing the realistic
decorated tonnage we have given above and recalculating, would show a greatly
reduced severity factor.
Summary Comment
A strong motivation of the glass industry in cooperating in the development
of this source assessment study has been to see that a fair and impartial
treatment is given the glass container industry and its impact on the surrounding.
area and effected population. We are certainly not attempting to obscure or
cloud facts or potential problem areas.
While we would appreciate your incorporating the above comments into the
final report, we still would not be in a position to indicate to concurrence
with the balance of it since we have not had ample time to review it in detail.
-------
G-10
As noted previously, our comments simply point out some of the more obvious
errors. We note that this report still contains many errors and significant
internal contradiction.
Sincerely,
John G. Turk
Vice President - Technical
JGT/bb
Attachments )
Dr. Richard Schorr
-------
G-ll
eaneiie
Columbus Laboratories
">(!"> kini> Avt-nui.-
fi.lumbus, Ohio 4H201
A *. on im£
August 20, 1976
Dr. John G. Turk
Vice President - Technical
Glass Containers Manufacturers Institute
1800 K Street, N.W.
Washington, B.C. 20006
Dear Dr. Turk:
In reply to your letter of July 2, 1976, regarding changes in the draft
final of the source assessment document (SAD) on glass containers, we
have reviewed these comments along with representatives from EPA and would
like to offer the following consensus opinions.
Your letter refers to several serious errors, significant internal contra-
dictions, and several other faults, but we can only deal with the ones
which were actually called out; and this has been done.
Nitrogen Oxides
We do recognize your position regarding nitrogen oxide emissions but fail
to find any reasonable basis on which to change the report. Nitric oxide
is not the only oxide of nitrogen leaving the plant site, and little evidence
is available to document its relative proportion. However, the report does
point out the NO is the predominant specie. The report from Owens-Illinois
was most informative but does not really seem to conflict with the general
conclusions of the SAD. It is well demonstrated that NOX emissions can vary
widely depending upon the operating conditions of the melting furnace.
Mr. Wooldridge's letter of May 28, 1976, we believe clearly indicated that
the calculations for source severity were not changed by assuming NO to be the
pollutant instead of N02, because of the safety factor used for non-criteria
pollutants. Again, it is important to remember that the purpose of these
studies is not to set standards but rather to define the needs for control
technology development in the glass industry.
Regarding other specific comments in your letter, Table 12 represents the
emissions from the entire plant's melting furnaces and not a single furnace.
The title will be changed to more clearly reflect this. The treatment of the
data in Table 14 has been checked and while we agree that there are a variety
of opinions regarding the calculation of maximum pollutant concentrations and
source severity, the treatment which we used is generally consistent with that
done in source assessment documents for other industries.
-------
Dr. John Turk G~12 August 20, 1976
Hydrocarbons
The hydrocarbon emissions in Table 15 came from NEDS. We are currently
reevaluating why this number was so high. We believe the number to be in
error and that we perpetuated the error. Dropping this number would lower
the emission rate in Table 15 to 2.88 x 10~2 gm/sec (1 ton/year) and change
the values for XMAX (3 minutes) and Xj^ (3 hours) to 3.11 and 1.58 ym/m3,
respectively, and the severity factor to 9.9 x 10~3.
Summary
I would like to note that very little actual emission measurement data was
supplied to Battelle by GCMI for use in the conduct of this study. This
is contrary to what is implied by your letter of July 2, 1976.
We are most concerned that the report be consistent with other studies of
like nature and that it accurately reflect the data available to us. Your
letter contains many statements inferring just the opposite.
Should you have any additional questions regarding the report, you may call
me on Extension 3624. After making final changes, the report will be
duplicated and copies sent to you, as per Ed Wooldridge's requests.
Very truly yours
-------
G-13
Gloss
PockooinQ 180°K street NW' Washin9ton> D-c- 20005
(202)872-1280
TWX: 7108229337 GLASS WSH
October 18, 1976
Dr. Dale Denny
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dear Dr. Denny:
We appreciate the opportunity extended the Glass Packaging
Institute (formerly Glass Container Manufacturers Institute) Air
Quality Task Group to attach comment to your final report "Source
Assessment of the Glass Container Manufacturing Industry",
Contract Number 68-02-1323, Task 37.
As we understand from our phone conversation on or around
September 17, 1976 with you, your procedure does not allow us time
to see and comment on the final draft document before it is made
public. Our comment, therefore, is not a critique of the wording
and data of the final draft, as obviously we have no confirmation
:of what is to be in that draft.
Rather than submit further detailed comment, as we have done
in. reviewing the two earlier drafts of the Source Assessment
report, we wish to relate to you some general observations we have
about the Source Assessment project, the format of this project, and
.the handling of data. Specifically, we wish to comment on four
general points to which we attach considerable significance.
1. General Report Format and Project Perspective
The major concern we have with the format of this study, as it
was presented in the earlier report drafts, is that the summary
presentation does not allow us, or others, to develop a perspective
about the significance, or severity, of the emissions. Although we
understand it is not the responsibility of the contractor to draw
conclusions, we do believe the summary format could include the
criteria to be used by EPA in evaluating the sources, and in defining
the need for further programs.
-------
G-14
Dr. Dale Denny
October 18, 1976
Page -2-
2. Significance of Emission Data
The Glass Container Manufacturers, through the Glass Packaging
Institute, have attempted to provide assistance in collecting
emission data for this study. Information has been provided through
response to requests from the contractor for emission data, or by
way of comment concerning the draft reports. Referring to the draft
of June, 1976, the industry is satisfied that a fair representation
of the process emissions is reported. We question comments by the
authors, inferring inadequate data still exists.
Comparing the emission summary of Table 2 with comments in the
text, we note some contradictions. The emission summaries of Table 2
are reported in a manner indicative that statistically significant
data has been collected. However, for some minor process sources,
the authors have indicated relatively few individual measurements
were available to them. We question performing a large number of
expensive measurements on sources which initial data and process
analysis indicate to have negligible emissions and insignificant
ambient impact. We believe for those processes which have a detectable
ambient concentration, there is represented in the report an adequate
body of data from which to draw conclusions.
3. Calculation of Ambient Concentration in Source Severity
Determination __ . : .
The format provided the contractor for determining source
severity is the calculation of maximum downwind ambient concentration
(by common equations after Turner) and direct comparison with
ambient air quality standards or converted threshold limit values.
Where the emitted specie is chemically the same as the "regulated"
compound, the resultant source severity comparison can be reasonably
represented as a worst case example. However, where the emitted
compound is not a regulated compound, but is a precursor (that is,
by atmospheric conversion or reaction may become the regulated
compound), the question of reaction time and reaction rate cannot
be ignored in the projection of the maximum ambient concentration
of the regulated pollutant.
In the glass container source assessment, the treatment of
nitric oxide (NO), where it is assumed to react immediately to form
nitrogen dioxide (N02) is questionable. The reaction of nitric
oxide to form nitrogen dioxide is not instantaneous, but is more
on the order of hours to days, depending on atmospheric conditions.
The point of maximum downlevel concentration of the effluent from
a glass furnace is, timewise, only two to three minutes. Only a
small portion of the nitric oxide can be converted in this time
interval, (3 to 5%). The simplification then, in assuming
instantaneous reaction in this case, predicts the source impact
approximately ten times higher than it should. (Additional comments
about the mathematics of ambient disbursion calculations, and the
-------
G-15
Dr. Dale Denny
October 18, 1976
Page -3-
conversion units used in the Glass Industry Source Assessment are
attached to this letter.)
The treatment of nitric oxide emitted is of concern to us in
the glass industry. However, we wish to point out a generalized
treatment of this form will be a problem to any industry, so evaluated,
which emits a material considered precursor to any pollutant when
instantaneous reaction to form that pollutant is assumed rather than
a reaction time representative of the known reaction rate and
chemistry.
4. Treatment of Control Technology
A stated purpose of this project was to determine where
pollution control technology or process modification is needed,
Another purpose of this report was to determine where control or
process technology is currently unavailable or unsuitable. It is
our opinion that any useful discussion of emission reduction
technology must consider technical feasibility and economic
practicality considerations.
The treatment in the two draft reports has considered neither
technical nor economic factors but, instead, has been simply a
listing of claims and promotions from various sources, with no
screening of fact. We'believe such a treatment has no value in
enlightening one as to what technology is available, adequate or
economical, or what development needs exist.
Sincerely,
John 6. Turk
Vice President •*. Technical
OGT:drh
cc: E. Wooldridge - EPA
R. Schorr ~ Batten e
-------
G-16
APPENDIX
Ambient Source Severity Determination of Pollutant Precursor
The preponderant nitrogen oxide formed in a combustion process, such as
in a container glass melting furnace, is nitric oxide (NO). Measurements
on operating furnaces and literature references indicate no more than 5%
of the total oxides of nitrogen is nitrogen dioxide (N02), the regulated pollutanl
*
Nitric oxide is considered a precursor to nitrogen dioxide in atmospheric
reaction. References to the rate of atmospheric conversion of nitric oxide
to nitrogen dioxide are not specific. Rather, they refer to a range of a
few hours to a few days, depending on atmospheric conditions. EPA project
personnel, responding to our original comments about atmospheric reaction,
indicated the conversion to be "rapid," and cited a half-life for the
reaction of one hour. A half-life of one hour is used in calculations
presented in this correspondence.
Furnace ambient dispersion calculations by the contractor (June, 1976, draft
report) located the point of maximum downwind concentration of emissions
a distance of 880 meters (2,887 ft) from the furnace stack (Table 16). ' ""
The average ground wind speed used in the dispersion calculation was
6 meters (19.8 ft) per second (Table 12). Using these values, the calculated
residence time from the point of stack egress to the point of maximum down-
wind concentration would be 2.43 minutes.
Using the one hour half-life for the oxidation of nitric oxide, only 3% of
the emitted nitric oxide will be converted to nitrogen dioxide in the 2.43
minute atmospheric residence time. Assuming an emission of 95% nitric oxide:
5% nitrogen dioxide, only 7.85% of the total nitrogen oxides would exist as
nitrogen dioxide at the point of maximum downwind concentration.
-------
G-17
The ambient air quality standard for nitrogen dioxide is 100 micrograms
per cubic meter, annual arithmetic mean. The mathematical ambient dispersion
calculation used represents a 3 minute maximum sample. The question, then,
is conversion from a 3 minute maximum to an annual mean. A common scaling
factor for converting from a 3 minute to a 24 hour maximum is the multiple,
0.36.
The comparison of a 24 hour maximum and an annual arithmetic mean is a
point of disagreement. The format of the June, 1976, draft was to assume
«
the 24 hour and annual values to be the same. We disagree. Most authors
relate an annual arithmetic mean to a 24 hour standard by a ratio of
about 0.3. In reviewing the ratios between annual and 24 hour standards
for particulates (primary and secondary) and for sulfur dioxide pollutants,
we find the Federal EPA related them as ratios of 0.29, 0.40 and 0.22,
respectively. This is in good agreement with a general ratio of 0..3 for
the conversion. In converting', then, from the 3 minute maximum to an
annual mean, the multiple of (0.36) x (0.3), or 0.11, should be used as
the scaling factor.
A further point of disagreement is the ~ ratio used in the formula to
calculate maximum ground-level concentration (Page 55, June, 1976 draft).
For D (neutral) atmospheric stability, the stability class used, the authors
cite a value of 1.6 (and appear to use the reciprocal, 0.63, in the calcu-
lation). The actual ratio shown in the cited Turner reference is 0.46.
The following restatement of the NOX portion of Table 14 illustrates the
effect of correcting the scaling factor, correcting the ^Jy ratio in the
dispersion formula and considering only the portion of NOX calculated to
exist as the pollutant, N02, at the point of maximum ground-level concen-
tration.
-------
. REPORT NO.
EPA-600/2-76-269
,p. TECHNICAL REPORT DATA
(f lease read Instructions on the reverse before completing)
12.
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
SOURCE ASSESSMENT: GLASS CONTAINER
MANUFACTURING PLANTS
5. REPORT DATE
October 1976
6. PERFORMING ORGANIZATION CODE
J.R. Schorr, Diane T. Hooie, Philip R. Sticksel, and
Clifford Brockwav
8. PERFORMING ORGANIZATION REPORT NO.
4D ADDRESS
Battelle-Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
10. PROGRAM ELEMENT NO.
1AB015; RQAP 21AFA-013
11. CONTRACT/GRANT NOV
68-02-1323, Task 37
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 9/75-9/76
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTESTask officer for this report is E. J. Wooldridge, Mail Drop 62
919/549-8411 Ext 2547. Earlier related reports are: EPA-650/2-75-019a and the
EPA-600/2-76-032 series.
is. ABSTRACT
repOrj- summarizes results of a study to gather and analyze background
information and technical data related to air emissions from glass container manufac-
turing operations. It covers emissions from three plant areas: raw materials pre-
paration and handling, glass melting, and forming and finishing operations. Melting
furnace emissions account for over 95 percent of the total plant emissions. The major|
pollutants are NOx, SOx, and submicron particulates consisting mainly (over 90 per-
cent) of mineral sulfates. NOx has the largest emission factor (3 g/kg) with annual
emissions of 36. 5 x 10 to the 9th power g. Compared with national emissions from
stationary sources, NOx emissions from glass melting furnaces contribute 0.34 per-
cent of the total. Source severity factors determined by this study were 0. 38 for NOx,|
0. 56 for SOx, and 0. 035 for particulates , with others being less than 0. 01. Source
severity is a measure of the potential environmental effect of air emissions and is the
ratio of the maximum average ground level concentration to the primary ambient air
quality standard for criteria pollutants.
7.
DESCRIPTORS
Air Pollution Dust
Assessments Minerals
lass Industry Sulfates
ontainers
Nitrogen Oxides
Sulfur Oxides
s. DISTRIBUTION'STATEMENT
Unlimited
EPA Form 2220-1 (9-73)
KEY WORDS AND DOCUMENT ANALYSIS
b IDENTIFIERS/OPEN ENDED TERMS
__ —
Air Pollution Control
Source Assessment
Stationary Sources
Particulate
Mineral Sulfates
c. COSATI Field/Group
19. SECURITY CLASS (This Report)
Unclassified
20 SECURITY CLASS (Thispage)
Unclassified
11G
08G
13B
14B
11B
13D
07B
21. NO. OF PAGES
-------
|