United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA 600 2 79 101
April 1979
Research and Development
Summary Report on
Emissions from the
Glass Manufacturing
Industry
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental 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.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161. ,
-------
EPA-600/2-79-101
April 1979
SUMMARY REPORT ON EMISSIONS FROM THE
GLASS MANUFACTURING INDUSTRY
by
E. D. Spinosa, D. T. Hooie, and R. B. Bennett
Battelle Columbus Laboratories
Columbus, Ohio 43201
Contract No. 68-01-3159
Contract No. 68-01-4431
Project Officer
Charles H. Darvin
Industrial Pollution Control Division
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed and approved for publication by Battelle's
Columbus Laboratories and the U.S. Environmental Protection Agency. Approval
does not signify that the contents necessarily reflect the views and policies
of the U.S. Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and
improved methodologies that will meet these needs both efficiently and
economically.
The purpose of this report is to present data and conclusions from a
comprehensive testing program on glass melting furnaces. Previous EPA
and industry reports on this subject have relied primarily upon literature
and randomly collected data without regard to similitude of testing
procedures and experimental design. Extreme caution, therefore, was taken
on this program to ensure that each test was conducted in a similar manner
and that each site tested reflected a typical operating situation in the
industry. Wherever statistically possible, conclusions were developed to
support or refute past conclusion on industry emissions.
The results of the report, therefore, will be extremely valuable to
R&D programs concerned with emissions from the glass manufacturing industry.
For further information concerning this subject, contact the Industrial
Pollution Control Division, Metals and Inorganic Chemicals Branch.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
iii
-------
ABSTRACT
The objective of the research described in this report was to deter-
mine the character of emissions from glass manufacturing operations. To
accomplish that objective, standardized testing procedures were conducted
at several glass plants. These plants included the three industry segments
(glass containers, SIC 3221; flat glass, SIC 3229; pressed and blown glass,
SIC 3229) that produce glass articles through melting of raw materials and
scrap glass. The products of purchased glass, (SIC 3231), and mineral wool,
(SIC 3296), segments were specifically excluded because the former had
no melting operations and the latter had melting operations vastly different
from the other three. Also, textile fiber operations were specifically
excluded.
This report presents a comprehensive review of the emission testing data.
For the purposes of this report, the glass manufacturing process has been
divided into five subprocesses: batching and mixing, melting, forming, post
forming, and product packaging. This report presents new information about the
size distribution of fine particulate matter emitted from the glass melting
subprocess, especially with regard to the distribution of particulates
considered to be respirable (i.e., £ 3 urn diameter). This report also presents
important new information about the type of emissions issuing'from some of
the forming and post-forming process steps. Finally, the pertinent data from
the field testing are compared to similar data that has been reported in the
Source Assessment Documents for the three glass making segments.
This report was submitted in fulfillment of Contracts 68-01-3159 and
68-01-4431 by Battelle's Columbus Laboratories under the sponsorship of
the U.S. Environmental Protection Agency. It covers the period from
June 4, 1976 to September 1, 1978.
IV
-------
CONTENTS
Foreword ill
Abstract iv
Figures vi
Tables vii
1. Introduction 1
2. Conclusions and Recommendations 3
3. Industry Description 6
4. Test Methods 20
5. Results and Discussion 26
References 48
-------
FIGURES
Number
Regional Distribution of Glass Container
Manufacturers
2 Regional Distribution of Pressed and
Blown Glassware Manufacturers 12
3 Regional Distribution of Flat Glass
Manufacturers 17
4 Adsorbent Sampling System 23
5 Typical Particle Size Distributions 40
6 Sampling Locations as viewed from the
Ventilator 44
vi
-------
TABLES
Number Page
1 Summary of Emission Information Available
in Source Assessment Documents 2
2 Estimated Shipments, Employees, and Plant
Numbers for the Glass Industry, 1971-73 6
3 Major Glass Container Manufacturers in the
United States 8
4 Major Pressed and Blown Glassware
Manufacturers in the United States 11
5 Proportion of Industry Output Accounted
for by the Consumer 13
6 Typical Glass Compositions 15
7 Percentage Distribution of Flat Glass
Capacity by Seven Major Companies 18
8 Source Emission Test Procedures for Glass
Melting Furnaces 22
9 Organic Fractions Measured with Adsorbent
Sampler 24
10 Furance Emissions Reported in the SAD
Appendices 27
11 Summary of Onsite Emissions Testing Data 28
12 Summary of Emissions Testing Data Supplied
by Manufacturers 30
13 Comparison of Data for SOX Emissions 31
14 Comparison of Data for NOX Emissions 33
15 Comparison of Data for Total Particulate
Emissions 35
vii
-------
TABLES
Number Page
16 Percentage of Particulate Loading In the
Sampling Train 37
17 Most Prevalent Chemical Species in
Particulate Train 38
18 Particulate Size Data 41
19 Percent by Weight of Captured Organic
Emissions from Forming and Postforming
Operations 43
20 Separable Forming and Postforming
Organic Emissions 45
21 Uncontrolled Fumes from Acid Etching 46
viii
-------
SECTION 1
INTRODUCTION
BACKGROUND AND OBJECTIVES
The Clean Air Act of 1977 requires the U.S. Environmental Protection
Agency (EPA) to project the need to control emissions from a given industry.
Source Assessment Documents (SAD's) are viewed as one means of making such
projections. In the glass manufacturing industry, SAD's have been written
for its three major .Standard Industrial Classification segments—glass
containers (SIC-3221),1 flat glass (SIC-3211),2 and pressed and blown glass,
nee (SIC-3229).^ These documents review the existing emissions information
and are primarily based on data found in the National Environmental Data
System (NEDS)^ data bank.
This report evaluates information presented in the SAD's (Table 1) by
using new data collected by onsite testing. The immediate object of this
project is to measure the emissions from all glass manufacturing operations
in all three major segments of the industry within the limits of available
time and funding. Emission measurements have been taken in thirteen stacks
at seven plants.
A number of problems were encountered during the course of the study.
Chief among them was the fuel crisis brought about by the severe winter of
1975-76. Many companies that had indicated willingness to participate were
unable to do so because of fuel curtailments. Others were prevented from
participation by plant shutdown or manufacturing problems that could not be
solved within the time available.
TEST METHODS
For the purposes of this research program, the glass manufacturing
operation was divided into five processes: batching and mixing, melting,
forming, postforming, and product packaging. Because of the low levels of
emissions for batching and mixing and product packaging, these process steps
were tested. Standard, EPA-recommended testing procedures were
used to establish the emission rates of criteria pollutants from the melting
process, and a porous-polymer adsorbent trap has been used for Level-I
testing for organic emissions from the forming and postforming process steps.
-------
TABLE 1. SUMMARY OF EMISSION INFORMATION AVAILABLE
IN SOURCE ASSESSMENT DOCUMENTS
Type Of Emission
No. Of Cases
Percent of
Total Furnaces
Glass containers (325 furnaces):
Particulates:
Loading
SO
NO
i
HC
x
Pressed and blown glass
(383 furnaces):
Particulates:
Loading
SO
x
NO
x
CO
HC
Flat glass (29 furnaces):
Particulates:
Loading
SO
89
49
24
36
x
NO
77
59
63
55
55
5
3
3
27
15
7
11
20
15
16
14
14
17
10
10
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
Some of the test data from this research program disagrees with those of
the SAD's. The magnitude of the disagreement varies for each criteria
pollutant, and no particular trends are evident for either pollutants or
industry segments. (See Section 5)
Several interesting results have been established for the melting
emissions, over a range of compositions including soda-lime, borosilicate,
and lead glass. They also span process rates from 811 to 15,900 kg/hour..
Consequently, these results are expected to be generally applicable to all .
three industry segments. First, statistically significant, empirical
(regression) relationships have been established for (1) SOX emission rates
as a function of process rate and fuel use rate, (2) for NOX emission rates
as a function of fuel use rates, and (3) for total particulate emission rates
as a function of fuel -use rates. For the tested plants, these relationships
are not strongly influenced by furnace size, type of glass melted, or type of
fuel used. Also, more than 80 percent of the total particulate matter is
respirable (i.e., median diameter <_ 3ym), and approximately 15 percent of
that total particulate matter is formed by condensation or reaction with the
impingers in the EPA-5 sampling train. Finally, emissions of unburned
hydrocarbons and carbon monoxide are negligible.
The Level-I testing for organic emissions has revealed that 47 percent
of the emissions from the forming and postforming process steps are composed
of the following species:
Aliphatic hydrocarbons
Aromatic hydrocarbons
Polyorganic matter (POM)
Polychlorinated biphenyls (PCB)
Esters
Ethers
Nitrocompounds
Halides
An additional 39 percent of the emissions may be composed of one or all of
the following:
• Sulfonates
• Sulfoxides
• Sulfonic acids.
-------
The emissions from the latter list may be preferentially produced from the
postforming process. However, caution must be exercised in interpreting
these results because they are not definitive. They indicate only a
potential problem area and not the severity of the potential problem.
A comparison of forming and postforming organic emissions measured in
this test program and the values reported in the SAD's is inappropriate
because a Level-I test does not define emission rate. Also, the data in
the SAD's are reported as engineering estimates from mass balance calculations
and may not reflect actual conditions. Therefore, any comparison of data
from these two sources can be erroneous. The Level-I testing has revealed
a list of toxic substances that may potentially be emitted during the manu-
facture of glass articles. The rate, severity, and exact character of these
emissions has not been established. More testing is therefore required before
the need for forming and postforming emission control can be established. If
control of these emissions is required, then the only economically and
technically feasible solution is replacement of the oils and sprays that
generate these emissions. Such a replacement will require a considerable
technical development effort.
Other sources of emission from the forming and postforming process
steps are acid fumes and metal (particularly tin) vapors. Actual testing
for these species was not conducted during this program because permission
was unobtainable for testing sites that could potentially emit those species.
However, data that have been supplied by the participating manufacturers
indicate that the information reported in the SAD's adequately estimates
these emissions.
The control technology available for the melting operation seems to
control particulate emissions adequately within current regulations. But,
these devices may not be adequate as the regulations become more stringent.
The fine particulate emissions are especially important when reviewed with
general information about the efficiency of control devices as follows:
GENERAL EFFICIENCIES OF CONTROL DEVICES
FOR PARTICULATE MATTER 5
Efficiency by Median Particle Size
Device 0.5 ym 1 ym 5 ym
Baghouse 90 percent 95 percent > 95 percent
Venturi Scrubber 80 percent 95 percent 95 percent
ESP <7Q percent 75 percent 90 percent
-------
The electrostatic precipitator (ESP) is emerging as"the preferred control
device in the industry for several technical reasons. However, it may be
ill-suited for control of the fine particulate matter issuing from glass
melting furnaces. The NO and SO emissions present difficult control
problems because proven control devices for these species are unavailable.
Electric melting may offer a means of reducing these emissions, but that
technology has been demonstrated only for smaller melting units (100 to 110
Mg/day) and a development effort would be required to raise the capacity of
electric melting to the nominal rate of 220 Mg/day. Implementation of electric
melting in the flat glass industry poses an even larger problem because the
technology must be scaled to units that produce glass at a rate of 600 to
800 Mg/day. Other technologies such as batch preheating and oxygen enrich-
ment may offer potential solutions to these problems, but they require con-
siderable development and evaluation before they can become commercially
available. Reasonable control over HC and CO emissions can be accomplished
by adequate combustion control. Equipment and combustion monitoring devices
are commercially available for accomplishing that control.
-------
SECTION 3
INDUSTRY DESCRIPTION
The glass industry consists of five segments: flat glass, glass
containers, pressed and blown glass, mineral wool and products of purchased
glass (SIC's 3211, 3221, 3220, 3296, 3231). Table 2 gives the value of
shipments, employees, and plant numbers estimated for 1971 and 1973 for four
of these segments. In 1973, some 1,335 plants employing more than 200,000
persons produced merchandise valued at $6.3 million. The glass container
segment is the largest in terms of shipments and employees. Approximately
57 percent of all glass melted is produced by the glass container segment.
The remaining glass melting is roughly divided between the flat glass (24
percent) and pressed and blown glass (19 percent) segments. The products of
purchased glass segment has no melting operations (the primary emissions
source in glass manufacturing), thus that industry has not been included in
this study. The melting process for the mineral wool industry is a cupola
process which is vastly different from the traditional, fossil fuel glass
melter, consequently, this industry segment has been excluded from this study.
Textile fibers are included in SIC-3229, but the contract for this project
specifically excluded that portion of SIC-3229 from testing.
TABLE 2. ESTIMATED SHIPMENTS, EMPLOYEES, AND PLANT
NUMBERS FOR THE GLASS INDUSTRY, 1971 AND 1973
Value of Shipments
SIC No.
3211
3221
3229
3231
Industry segment
Flat glass
Glass containers
Pressed and blown glass
Products of purchased
glass
Total
Average
1971
(Millions
811
1,944
1,108
1,157
5,020
1
2
1
]_
6
1973
of $)
,118.
,316
,423
,440
,297
Change
1971
Employees
1973
Change
(%) (Thousands) (%)
37
19
28
24
25
.8
.1
.4
. 4
.4
24.
74.
53.
29.
131.
0
6
4
8
8
26.3
77.8
61.9
35.5
201.5
9.6
4.3
15.9
19.1
10.8
Number of
plants
75
140
260
860
1,335
(a) Source: Department of Commerce
-------
GLASS CONTAINERS (SIC-3221)
Glass containers (SIC-3221) is the largest of the three major segments
of the glass industry. It includes the manufacture of narrow-neck and wide-
mouth glass containers for foods, beverages, medicines, toiletries, and
cosmetics. Shipments in this segment have grown at an average rate of about
3.5 percent since 1971. Three general types of container glass are produced:
amber, green, and clear. For this report, green and clear glass are
considered together in a single category designated "flint". The major
difference between amber and flint is the iron-oxide additions. Amber
accounted for approximately 15 percent of glass production in 1976 and flint
represented about 85 percent.
According to information gathered from the Department of Commerce 1972
Census of Manufacturers and from the Glass Packaging Institute (GPI), 119
establishments manufacture glass containers in the United States. These
119 plants are operated by 27 manufacturers (Table 3). Statistics obtained
from 1976 glass industry directories and from industry 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. Plants are found throughout the United States, but a
large number are concentrated in the East, North-Central, and Middle Atlantic
regions. The regional distribution of the major plants is shown in Figure 1.
The volume of shipments for 1976 was 293 million gross compared to 268
million gross in 1972, a 9-percent increase for the 4 year period. The
weight of glass containers shipped increased from 10,772 Gg in 1972 to
approximately 11,809 Gg in 1976, an increase of 10 percent.
The basic raw materials for soda/lime container glass are silica sand,
soda ash (Na^CX^), and limestone (primarily CaCO^, plus some MgC03 in
dolomitic limestones). Feldspathic minerals (anhydrous aluminosilicates
containing potassium, sodium, and calcium in varying ratios) are also
utilized as sources of alumina and alkali. Minor amounts of other oxides
are introduced as impurities, and additional minor ingredients are added for
specific purposes. A typical soda-lime glass-batch composition is:
Silica sand 909
Soda ash 306
Feldspar 118
Limestone 294
Salt cake (Na2SO^) 7
Total 1,634 kg
-------
TABLE 3. MAJOR GLASS CONTAINER MANUFACTURERS
IN THE UNITED STATES3
Manufacturer
No.
Of Plants
Anchor Hocking 9
Arkansas Glass 1
Ball Corporation 4
Bartlett Collins 1
Brockway Glass 14
Chattanooga Glass 7
Columbia Gas 1
Diamond Glass 1
Gallo Glass 1
Glass Containers Corp. 12
Glenshaw Glass 2
Hillsboro Glass 1
Indian Head 7
Industrial Glass 1
Manufacturer
No.
Of Plants
Kerr Glass 7
Latchford Glass 1
Leone Industries 2
Liberty Glass 1
Madera Glass Co. 1
Metro Glass 4
Midland Glass 4
National Bottle Corp. 4
National Can 4
Owens-Illinois 20
Thatcher Glass 6
Underwood Glass 1
Wheaton Glass 2
Total 119
a) Source: Material provided from
dated October 24, 1975.
GPI,
-------
West South
Central
East
Figure 1. Regional distribution of glass container manufacturers.
-------
Typically, these ingredients will produce 1,370 kg of glass and give off 259
kg of gases, primarily (>99 percent) COo. The batch volume of 1.2 m^ will
produce 0.6 nr of fluid glass and 858 nP of gaseous products (measured at
the furnace temperature of 1,500 C). The minor ingredients such as salt
cake (sodium sulfate) and various fining, coloring, or decolorizing agents,
rarely exceed 5 percent and are often less than 0.1 percent of the total
glass composition.
Historically, growth of the glass container industry has fluctuated
considerably, 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 slower, and future growth may be tied to
legislation restricting use of nonreturnable containers. Production levels
for 1980 are likely to be 20 percent higher than levels for 1974. Emissions
are expected to increase proportionally, and possibly at an increased rate,
since the industry is moving from the use of natural gas to the use of fuel
oil. The actual effect of this conversion on emission rates is not clear.
PRESSED AND BLOWN GLASS, NEC (SIC-3229)
The pressed and blown glassware industry, as represented by SIC-3229,
essentially includes all industrial establishments primarily engaged in
manufacturing glass and glassware that is pressed, blown, or shaped from
glass produced in the same establishment. It consists of every category of
glass or glassware except flat glass (SIC-3211) and glass containers
(SIC-3221). Establishments include those manufacturing textile glass fibers;
lighting, electronic, and technical ware; and machine-made and handmade
table, kitchen, and art-ware glass products. The pressed and blown glass
industry consists of approximately 155 manufacturing establishments (Table 4).
Approximately 40 plants produce hand pressed and blown glassware almost
exclusively. The industry is concentrated in or about the North Central
region of the United States—primarily New York, Pennsylvania, West Virginia,
Ohio, Indiana, and Illinois. Plants are also located in 22 other states
(Figure 2).
Each of three product types makes up a significant portion of the pressed
and blown glass industry (Table 5). The value of shipments for the industry
in 1976'is estimated to be $1.6 billion, compared with $1.1 billion in 1971—
an average, annual increase of 9 percent. Shipment weights from the various
industry segments are difficult to estimate because of the many product
types and the different methods of reporting. But, the pressed and blown
glass segment is estimated to have shipped 1.5 Tg of glass in 1976.
10
-------
TABLE 4. MAJOR PRESSED AND BLOWN GLASSWARE
MANUFACTURERS IN THE UNITED STATES
Manufacturer
No. of
Plants
Manufacturer
No. of
Plants
American Optical 3
Anchor Hocking 4
Arrowhead Puritas 1
Atlantic Optical Moulding 1
B and J Optical 1
Bartlett-Collins 1
Bausch and Lomb 1
Beaumont Company 1
Big Pin Glass Company 1
Blenko Glass Company 1
Brock Glass Company 1
Brockway Glass 2
E. 0. Brody Company 1
Cambridge Glass 1
Canton Glass 1
Cataphote 1
Champion Agate Company 1
Chemglass, Inc. 1
Corning Glass Works 14
Crescent Glass Company 1
Crystal Art Glass 1
Davis-Lynch Glass Company 1
EMC Glass Corporation 1
Eas tman Kodak.Company 1
Emerson and Cuming 1
Elite Company, Inc. 1
Erie Glass Mfg. Company 1
Erskine Glass and Manufacturing 1
Federal Glass Company 1
Fenton Art Glass 1
Fischer and Porter Company 2
Fostoria Art Glass 1
Friedrich and Dimmock, Inc. 1
GTE Sylvania, Inc. 4
General Electric 9
Gentile Glass Company 1
Gillinder Brothers 1
Gladding-Vitro-Agate Company 1
Glass Works,.Inc. 1
Guernsey Glass Company 1
Haley Glass Company 1
Hamon Handcrafted Glass 1
Harvey Ind. 1
Holophane 1
Houze Glass Corp. 1
Imperial Glass Corp. 1
Indiana Glass Corp. 1
Industrial Glass and Plastics 1
Interpace Corp. 1
Jeanette Corporation 1
Jeanette Shade and Novelty Co. 1
Johnson Glass 1
Kanawa Glass Company 1
Kapp Glass, Inc. 1
Kaufman Glass Company 1
Kessler, Inc. 1
King Seeley Thermos Company 2
Labino Glass Laboratories 1
Lancaster Glass Corporation 1
Lenox Crystal 1
Lewis County Glass 1
Louie Glass Company 1
Marble King 1
Masden Ind. 1
Mayflower Glass 1
Mid-Atlantic 1
J. H. Millstein 1
Minners 1
Multicolor Glass 1
Nuclear-Pacific 1
Overmyer-Perram 1
Owens-Illinois 10
Pacific Glass Works 1
Peltier Glass. Company 1
Penbarthy Electromelt Intl., Inc. 1
Pennsboro Glass Company 1
Phoenix Glass Company 1
Pikes Peak Glass Company 1
Pilgrim Glass Corp. 1
Pittsburgh Corning Corp. 2
Pape Scientific, Inc. 1
Potter Industry, Inc. 4
RCA Corp. 1
Rainbow Art Glass 1
Ray-Lite Glass, Inc. 1
Reha Glass Company 1
Rocky Mountain 1
Rodifer Gleason Glass Corp. 1
St. Clair Glass Works 1
Scandia Glass Works I
Schott Optical Glass Company 1
Scott Depot Glass Company 1
Sellstrom Mfg. 1
Earl Shelly Glass Company 1
Sinclair Glass 1
L. E. Smith Glass Company 1
Super Glass Corp, 1
Swift Glass Corp. 1
Techniglass, Inc. 1
Thermal American Fused Quartz 1
Thomas Ind., Inc. I
Variety Glass, Inc. 1
Venejian Art Glass 1
Victory Glass Company 1
Viking Glass Company 1
West Virginia Glass Specialty Co. 1
Westinghouse Electric Corp. 1
Westmoreland Glass Company 1
Wheaten,Ind. 1
The Paul Wissmach Co., Inc. 1
11,
-------
Pacific
West Norton
Central
West South Central
Figure 2. Regional distribution of pressed and blown
glassware manufacturers.
-------
TABLE 5. PROPORTION OF INDUSTRY OUTPUT ACCOUNTED FOR
BY THE CONSUMER, SCIENTIFIC, TECHNICAL, AND
INDUSTRIAL GLASSWARE SEGMENTS OF SIC-3229(a'
Process
and Major Products
Table, Kitchen, and Art
Ware
Machine Made
Hand Made
Lighting and Electronic
Scientific and Industrial
Total Industry (Percent)
(Actual)
% of Industry
SIC 1972
(32291) 33
8
(32292) 35
(32294) 24
100
$1.1 billion
Shipment (value )
1976
39
6
32
23
100
$1,6 billion
(a) Source: Current Industrial Reports, MA-32E
for shipment value in SIC-32291,
-32292, and -32294 — 1972 and 1976„
13
-------
Compositionally, the four important categories of glass manufactured by
the pressed and blown glass industry are as follows:
Estimated percent of
Glass category total production
Soda/lime 77
Borosilicate 11
Lead silicate 5
Opal 7
Soda/lime glasses are overwhelmingly the most important type of glass
in terms of variety of use as well as tonnage melted. The combination of
silica sand, soda ash, and limestone produces a glass that is easily melted
and shaped and that has good chemical durability. In the soda-lime-silica
system, an optimum glass with respect to cost, durability, and ease of manu-
facture is typically composed as shown in Table 6. Magnesia is used primarily
to reduce cost by the substitution of dolomitic limestone for limestone as
a raw material. The alumina improves chemical durability and decreases the
problem of crystallization during forming. Primary pressed and blown
products employing this type of glass are incandescent lamps, tubing, and
tableware.
Borosilicate glasses are basically a combination of silica sand, boric
oxide, and soda ash. The compositional ranges of typical commercial
borosilicates are shown in Table 6. The borosilicate glasses have excellent
chemical durability and electrical properties, and their low thermal expansion
yields a product with high resistance to thermal shock. These combined
properties make them ideal for demanding industrial and domestic uses such
as chemical laboratory ware, cookware, pharmaceutical ware, and some lens
reflectors and lamp envelopes. Pyrex , produced by Corning Glass Works, and
Kimax^, produced by the Kimble Division of Owens-Illinois, Inc., are examples
of products made from borosilicate glasses.
Lead silicate glasses are composed of silica, lead oxide, and significant
amounts of alkali oxide. The compositional range of typical commercial lead
glasses is shown in Table 6. The lead glasses are characterized by high
electrical resistivity, high refractive index, and slow rate of increase in
viscosity with decreasing temperature. This viscosity characteristic makes
them particularly well suited to hand fabrication. Lead glasses are used
in high-quality art and tableware, for special electrical applications,
optical glasses, fluorescent lamp envelopes, and X-ray, gamma-ray, and
neutron radiation shielding windows.
Opal glasses are translucent and may be colored. Commercial products
of opal glass include lighting globes, ointment jars, dinnerware, and wall
paneling. The translucency or opacity of opal glasses is produced by multiple
scattering of light inside the glass. This scattering is achieved by the
precipitation of crystals (or an immiscible amorphous phase) with an index
of refraction different from that of the base glass. Commercial opal glasses
commonly employ fluorine additions to yield opacifying crystals of sodium or
calcium fluoride. A typical commercial opal glass jar composition is given
in Table 6 along with an opal illumination glass composition. \
14
-------
TABLE 6. TYPICAL GLASS COMPOSITIONS
Major
Components
Soda/lime glasses :
Silica
Soda
Lime and
Magnesia
Alumina
Other Oxides
Borosilicate
glasses :
Silica
Alumina
Boron Oxide
Sodium Oxide
Magnesia
plus calcia
Potassium Oxide
Lead glasses :
Silica
Lead Oxide
Sodium Oxide
Potassium Oxide
Alumina
Percent
By Weight
»
i
72
15
10
2
1
60-81
1-17
5-2U
1-15
h-n
1- 8
35-70
12-60
4-8
5-10
0.5-2.0
Major
Components
Opal glass :
Si02
A1203
CaO
Na20
K20
F2
Opal Illumination
glass :
Si02
A1203
CaO
MgO
Na20
F2
ZnO
PbO
Percent
By Weight
71.2
7.3
4.8
12.2
2.0
4.2
59.0
8.9
4.6
2.0
7.5
5.0
12.0
3.0
15
-------
The pressed and blown glass industry is very diverse, and estimates of
future growth trends are difficult to make. The diversity of the industry
tends to exempt it from fluctuations caused by changes in other glass industry
segments. An annual growth rate of 5 to 8 percent is not an unreasonable
expectation for this segment.
FLAT GLASS (SIC-3211)
This industry contains establishments primarily engaged in manufacturing
flat glass as well as some laminated and tempered glass. The major products
shipped by the flat glass industry are: window glass, plate and float glass,
rolled and wire glass, tempered glass, and laminated glass. Tempered and
laminated glass production is also reported under glass products made of
purchased glass (SIC 3231).
Flat glass shipments vacillated significantly in the latter 1960's and
early 1970's. The value of shipments during 1972 was $937.2 million—up
15.5 percent from the previous year. The large fluctuations were due to
domination of the industry demand by the construction and automotive
industries. The four largest companies in the flat-glass industry accounted
for more then 92 percent of the total value of industry shipments. In 1973,
the average value of shipments from each of the 31 establishments was
approximately $1.4 billion. Two areas in the U.S. accounted for slightly
more than 70 percent of the value of shipments made by all establishments
in the industry—the North Central (42 percent) and the South Central (29
percent) regions. Most flat-glass manufacturing plants are located in Penn-
sylvania, Ohio, Tennessee, Michigan, Illinois, Texas, and California (Figure 3).
The seven flat glass manufacturers are shown in Table 7, as well as the
approximate market share (by product type) for each of these companies.
Four flat glass products—float, sheet, rolled, and plate—are manufactured
in the United States. Of these float glass accounts for more than 90 percent
of the total flat glass production. Float glass is made by floating molten
glass from the melting furnace on a bath of molten tin until the glass
hardens. This glass, with its high optical quality, has replaced plate
glass, which required grinding and polishing to produce a smooth surface.
It is used for automobile windows and large picture windows. Average
thickness ranges from 3.2 to 6.4 mm. Sheet glass is made by drawing
molten glass upward from the melt. It is thinner than float glass (1.6 to
3.2 mm) and is used for windows in residential construction. Rolled or
patterned glass is formed by drawing molten glass through rollers with
patterns impressed on them. This decorative glass is used for special
purposes such as shower doors and partitions. Plate glass is made by
drawing molten glass through smooth rollers and then grinding and polishing
both glass surfaces to a smooth finish. Only one plate glass furnace is
still in operation in the United States.
16
-------
West South
Central
Ndrth East
Atlantic
Figure 3. Regional distribution of flat-glass manufacturers.
-------
TABLE 7. PERCENTAGE DISTRIBUTION OF FLAT GLASS
CAPACITY BY SEVEN MAJOR COMPANIES(a>
Manufacturer
PPG Industries, Inc. 45.1
Libbey-Owens-Ford Company 26.4
Ford Motor Company 13.1
ASG Industries, Inc. 6.3
Guardian Industries Corp. 3.0
C-E Glass, Inc. 1.5
Fourco Glass Company 4.6
100.0
18
-------
The composition of flat glass is exclusively soda-lime-silica with the
following typical recipe:
kg wt%
Silica sand 910 56
Soda ash 302 18
Feldspar 80 5
Limestone 309 19
Salt cake (Na S04) 36 2
Total 1,637 100
These ingredients melt to 1,341 kg of glass and give off 295 kg of gases,
primarily (>90 percent) C02- The batch volume of 1.27 nr produces 0.57 nr
of fluid glass and 708 m^ of gaseous products (measured at the furnace
temperature of 1,500 C). Although many minor ingredients (<5 percent of
the batch) can be added to the glass batch, very few are used in making flat
glass. The only other ingredients used in making clear float glass are
carbon in the form of powdered coal (used as a reducing agent for sulfates)
and iron oxide (to provide a greenish tint). No borates, fluorides,
selenium, or arsenic compounds are added. A small amount (<10 percent of
total production) of colored flat glass is made, but that production has
not been considered in this document.
The flat glass industry has sustained a compound annual growth for the
value of shipments of 6.7 percent between 1967 and 1976. A continued
growth rate averaging 5 percent is not unexpected. However, this industry
is strongly tied to the housing and automotive industries and the future
growth trends of those industries will greatly affect the flat glass industry.
19
-------
SECTION 4
TEST METHODS
For the purpose of this project, the glass manufacturing operation has
been divided into the following processes:
Batching and mixing
Melting
Forming
Postforming
Product packaging
These processes generally hold for both the glass container and pressed and
blown glass segments. In the flat glass industry, forming and postforming
are usually integrated into one continuous process. The primary function of
the postforming operation is to remove detrimental, residual stresses in the
glass product, and in that sense, forming and postforming operations in the
flat-glass manufacturing industry can be considered as being equivalent to
similar operations in the other two glass manufacturing segments.
The test procedures that have been used in this program have generally
been standard EPA procedures. Since this program is not a compliance testing
program, some of the tests have been modified to maximize the information
while minimizing the cost. The test procedures and modifications will be
described as they apply to each process.
BATCHING AND MIXING
The first process in the manufacture of glass articles is the weighing
and mixing of the raw materials—sand, alkali, alkaline-earth carbonates,
various mineral products, and cullet (scrap glass). Minor ingredient
additions to the glass recipe include carbon and sulfates that provide a
means for removing gaseous inclusions (refining), and various agents that
impart color to the final product. The raw materials and cullet are
accurately weighed, as specified by the particular glass recipe, and
intimately mixed before delivery to the melting unit.
The major emission encountered in this process is dust that results from
the transfer, handling, and weighing of the generally fine-grained raw
materials. These emissions are either controlled through the use of a bag-
house (in which case the control efficiency is >90 percent), or they are
uncontrolled and become fugitive emissions. Batch wetting is sometimes used
to reduce dusting during material transf.er. If the control device is in
20
-------
operation, there is no point in measuring the very low level of emission from
that device. If on the other hand the dust is fugitive, there is no reliable
means of measuring that emission. Therefore, no test data have been collected
for this process.
MELTING
The glass melting operation is a high-temperature combustion process and
is the major source of emissions in the glass manufacturing operation. The
emissions from this process include particulates, NO , SOX, CO, and hydro-
carbons. Both natural gas and fuel oil are used as primary fuels in this
step. Where possible, the sampling program includes testing for both fuels.
The test procedures for melting are listed in Table 8. In some cases, part
of the raw material feed includes a fluoride flux. When a fluoride flux is
used, the EPA-5 train is modified to include the impingers for an EPA-13
test. This modification allows an analysis of particulate fluorides. The
particle-size distribution is determined by using an Anderson cascade
impactor, which takes data at the stack temperature. This impactor is run
at an average flow point, which is determined during a traverse of the stack
for the particulate loading test. This average flow point is also used for
the testing of NOX, SO , and CO, and hydrocarbons. The integrated bag sample
is also taken from this average flow point, and that bag is used to perform
an Orsat analysis.
FORMING
Both pressed and blown glass and glass containers are formed in automated
machines. Various organic sprays and swabbing compounds are used in this
process to facilitate release of the formed product from the hot metal mold.
Volatilization of these compounds provides a source of emissions which exit
from the area through ventilators located in the roof of the building. A
Level-I test for the organic emissions is performed for this process step
using a porous polymer adsorbent trap^ (Figure 4). This test allows
.calculation of emissions concentration and determination of eight classes of
organics (Table 9).
The forming process that is used in the manufacture of flat glass is
considerably different from that used in pressed and blown glass and in glass
containers. In flat glass manufacturing, the molten glass floats onto a
bath of molten tin and is drawn into a very large flat sheet. The emissions
from this process step are expected to be very minimal and are calculated
on a worst-case basis using actual, annual use rates of tin.
POSTFORMING
The postforming process includes annealing to remove detrimental,
residual stresses in the glass products, lubricity coatings, and decorating.
The lubricity coatings and decorating media are potential sources of organic
emissions during postforming. These potential emissions are determined using
21
-------
TABLE 8. SOURCE EMISSION TEST PROCEDURES FOR GLASS MELTING FURNACES
No. of
samples
per source Sample
(runs) type
3 Particulates
3 Particle size
3 Gases
Sample
method
EPA-5 a
EPA-13e
Anderson
at one
traverse
point
EPA-10e
at one
traverse
point
Minimum
sample
time
(min)
60
~10d
60
Minimum
sample
vol ume
(scf)
30
~7d
1
Initial
analysis
b
Mass
1/2 Massb
1/2 Mass
Mass size
Orsatf
ii
Final
analysis
Trace metals, OES-AA
Trace metals, OES-AA
Fluorides
-
Gas
Gas
sox&so2
Grab
EPA-7
EPA-8
at one
traverse
point
-«15 sec 2-liter flasks
•^IS sec 2-liter flasks
^15 sec 2-liter flasks
«*15 sec 2-liter flasks
20 10
NOX
GCg
N0xh
sox&so2
a Twelve-point samples at 5 min/point. EPA-5 as in 6/8/76 Federal Register.
b Front and back half mass as per EPA-5 in 8/17/71 Federal Register.
c When flourides are present, use Method 5 train with EPA-13B solutions and
fluoride analyses. One run for mass and 2 runs for fluoride.
d Approximate. Depends on stack gas participate loading.
e Tedlar bag and flasks are black or covered and protected from light.
f Orsat or Fyrite; if CO is present, Qrsat is required.
g GC analysis through C-4 (one flask analyzed per source other backup).
h EPA-7 flask for phenoldisulfonic acid method (one flask analyzed per source
other backup).
22
-------
FLOW DIRECTION
RETAINING SPRING-i
ADSORBENT
GLASS FRITTED
DISC
FRITTED STAINLESS STEEL DISC-
15-MM SOLV-SEAL JOINT
Figure 4. Adsorbent sampling system.
23
-------
TABLE 9. ORGANIC FRACTIONS MEASURED
WITH ADSORBENT SAMPLER
Approximate Infrared
Fraction Compound type sensitivity (pg)
1 Aliphatic hydrocarbons 1810
2 Aromatic hydrocarbons 1.10
POM
PCB
Halides
3 Esters 0.1-1
Ethers
Nitro compound s
Epoxides
4 Phenols 0.1-1
Esters
Ketones
Aldehydes
Phthalates
5 Phenols 0.1-1
Alcohols
Phthalates
Amines
6 Amides 0.1-1
Sulfonates
Aliphatic acids
Carboxylic acid salts
7 Sulfonates 0.1-1
Sulfoxides
Sulfonic acids
8 Sulfonic acids 0.1-1
24
-------
a porous polymer absorbent trap. Often these emissions cannot be separated
from the forming emissions because a common ventilator is used for both
areas. A separate postforming test was performed in those few plants which
had separate ventilators for this process step.
The manufacture of primary flat glass products does not involve the
application of decorative sprays or lubricity coatings. The major post-
forming process step is that of annealing to remove detrimental, residual
stresses. In the flat glass industry, this step utilizes natural gas as a
fuel, and combustion is expected to be near perfect. Thus, because emissions
are expected to be minimal and because no direct gas stream (e.g., stack) is
available, no postforming tests were attempted in flat glass plants.
PRODUCT PACKAGING
Emissions from the product packaging step are expected to be non-
existent or negligible. If emissions are present, they are in the form of
paper dust from the packaging materials, which is a fugitive emission.
Therefore, no testing has been done for this process.
25
-------
SECTION 5
RESULTS AND DISCUSSION
The data obtained from the field sampling are presented by the process
subdivisions from which they have been generated, and when appropriate are
given both for the glass industry as a whole and by the industry segment.
When possible, the field data are further subdivided by glass composition
and fuel type. The field data are treated statistically to compare results
among various segments of the j^lass manufacturing industry. The primary
statistics are: sample mean (X), sample variance (o), and confidence interval
(taken at 10 percent significance) as percent of the mean. Also, regression
analyses are used when they are appropriate. They include statistical
measures of correlation and confidence of a real relationship (i.e, nonzero
slope). The results of these statistical treatments are used to compare the
field sampling to the information that has been reported in the SAD's
(Table 10).
BATCHING AND MIXING
As discussed in Section 4, no emissions testing was conducted for
batching and mixing. However, observations made by testing personnel while
onsite indicated that the dusting from batch preparation and conveying was
generally confined within the buildings. All of the plants in which these
observations were made were equipped with baghouses, and no visible emission
was seen from the exit side of this control device. Therefore the conclusion
reached by the SAD's, that only fugitive emissions escape from this process
step, appears correct.
MELTING
The glass melting operation is the largest source of emissions for the
glass manufacturing industry, and consequently, most efforts in this research
program have been directed toward measuring the emission rates of various
criteria pollutants that are produced from this high-temperature process.
Two sources of data have been used to gather information for this process
step: actual testing under EPA contract, and data voluntarily provided by some
participating manufacturers. Most of the discussions in this report will be
directed toward the data gathered by onsite testing. These data are
presented in Table 11 for the 13 manufacturing operations that voluntarily
26
-------
TABLE 10. FURNACE EMMISSIONS^ REPORTED IN THE SAD APPENDICES
to
Source
Container
Glass
Pressed and
Blown Glass
Flat/float
Glass
Production
Qg/yr
52.744
(±21.11)
25.05
(±26.48)
162.000
(b)
Particulates
Mg/yr
37.294
(±27.72)
67.520
(±23.52)
131
(b)
sox
Mg/yr
91.691
(±21.41)
51.52
(b)
349
(b)
NOX
Mg/yr
163.578
(±26.80)
41.071
(±48.61)
620
(b)
CO
Mg/yr
3.278
(±10.94)
3.000
(b)
-
HC
Mg/yr
4.07
(±2.51)
4.0
(b)
—
a Table entries are:
x
(90% confidence interval as percent of x)
Insufficient data for calculating confidence interval
-------
TABLE 11. SUMMARY OF ONSITE EMISSIONS TESTING DATA
Emissions
Data
Source
Container
A
D
E
H
I
J
L
M
Process
Rate,
kg/hr
(Ib/hr)
5,
(12
8,
(18
7,
(16
6,
(13
5,
(11
5,
(11
9,
(21
11
(24
470
,058)
391
,499)
274
,039)
227
,728)
162
,380)
384
,871)
873
,767)
,204
,701)
Fuel
No. 6 Oil
No. 6 Oil
No. 2 Oil
No. 6 Oil
No. 6 Oil
No. 6 Oil
Gas
Gas
Total
Particulate,
kg/hr
(Ib/hr)
8.
11
4.
6.
6.
15
11
9.
27
.5
40
12
62
.0
.98
58
(18.2)
(25.3)
(9.71)
(13.5)
(14.6)
(33.0)
(26.7)
(21.1)
16
24
11
23
53
54
25
39
kg/hr
(Ib/hr)
.21 (35.7)
.3 (53.6)
.3 (25.0)
.8 (52.3)
.35 (117.6)
.10 (119.3)
.21 (55.57)
.22 (86.46)
NOX,
kg/hr
(Ib/hr)
15.74
22.3
15.8
13.6
12.28
13.49
48.00
29.50
(34.7)
(49.1)
(34.9)
(29.9)
(27.1)
(29.74)
(106)
(65)
CO, HC,
ppm ppm
Trace3 Trace3
Trace3 Trace3
5 0
Trace3 Trace3
<1 56.2
47.85 6.96
4.44 7.5
17.1 9.15
Pressed and
blown glass
C (Soda/lime) 1,657
(3,653)
F (Borosilicate) 931
(2,053)
G (Borosilicate) 950
(2,094)
Gas 1.8 (3.9)
Gas 3.0 (6.5)
Ho. 2 Oil 3.4 (7.6)
0.46 (1.013) 8.94 (19.7) 30 <5
Trace3 1.0 (2.3) Trace3 13
2.2 (4.8) 6.0 (13.3) Trace3 7
K (Lead)
Flat
B
X
S
N
±
811
(1,789)
15,876
(35,000)
6,093
(13,434)
4,521
13
37
Gas 3.14 (6.93)
Gas 40.0 (88.2)
9.60 (21.12)
9.98
13
52
0.10 (0.21)
68.20 (150.3)
24.49 (53.88)
23.00
13
46
10.0 (22.06) 2.2
86.60 (191) 19.57
21.79 (47.94)
22.79
13
52
9.7
1.46
-
-
-
-
a Trace <0.1.
b 90 percent confidence band about the mean as % x.
- Dash indicates insufficient data for calculation.
28
-------
participated in this testing program. The sources are distributed as follows:
eight glass container plants, four pressed and blown glass plants, and one
flat glass plant. Data voluntarily supplied by various manufacturers are
summarized in Table 12. These voluntarily supplied data are distributed
among the three major glass manufacturing segments in the following manner:
one glass container furnace, thirteen pressed and blown glass furnaces, and
four flat glass furnaces. Specific results and conclusions drawn from the
data in Tables 11 and 12 are presented individually for each criteria
pollutant sampled.
Oxides of Sulfur
The criteria pollutants with the largest emission rates are the oxides
of sulfur, reported as S0x because of the various gaseous stages in which
they can exist. The summaries in Table 11 indicate that SOX is the pollutant
with the largest emission rate in the container industry, and it is slightly
more prevalent than NOX for the entire glass industry as a whole. For the
glass container and pressed and blown glass segments, the information in
Table 11 can be used to calculate the range bracketing expected annual
emission rates. For the glass container industry, this range is 200 to 342
Mg/yr. A similar calculation for the pressed and blown industry data
indicates that the emission rate for SOX ranges from 0 to 13.4 Mg/yr. These
data are compared to those of the SAD's and manufacturers in Table 13.
Clearly, the SO data from the three sources are at variance. The source of
these differences cannot be definitively established. However, the non-
standard testing procedures that have been used in the data from which the
SAD's have been generated can certainly be expected to contribute to the
descrepancies. The lack of manufacturer's data also indicates inconsistancy
in obtaining SO data. Since the current testing program used standard-
ized procedures, its data is expected to be the most reliable.
Plotting of the SOX emission rate data indicates that a nonlinear
relationship exists between the SO emission rate, the process rate, and the
fuel-use rate (expressed as equivalent kilowatts). A stepwise regression
using the logarithms of these three variables indicates that the logarithm
of the SOX emission rate is a function of the logarithm of the ratio of the
process rate to the equivalent kilowatt usage. A linear regression analysis
involving these two logarithms gave the following statistical results:
1. Correlation — 93 percent
2. Unexplained variation — 17 percent
3. Confidence of nonzero slope — 99.95 percent
The conclusion that can be drawn from this regression analysis, particularly
in light of the statistics it yields, is that most of the S0x emission rate
is related to the amount of material being processed through the melting unit
and to the rate at which the fuel is used. The high confidence of nonzero
slope clearly indicates that the process rate and fuel use rate are the
controlling variables for SOX emissions. The fact that the regression line
explains 75 percent of the variation in the data indicates that the difference
in the melting units among the three major manufacturing segments are not as
large as indicated by references in the SAD's. The unexplained variation
29
-------
TABLE 12.
SUMMARY OF EMISSIONS TESTING DATA SUPPLIED
BY GLASS MANUFACTURERS
CO
O
Data source
Container glass:
MAA
Process rate,
kg/hr (Ib/hr) Fuel
10,932 (24,080)
Total particulate,
kg/hr (Ib/hr)
3.46 (7.63)
SOX,
kg/hr (Ib/hr)
56 ppm
Emissions '
kg/hr (Ib/hr)
60.8 (134)
CO, Gaseous F ,
ppm kg/hr (Ib/hr)
-
Pressed and Blown glass:
MA (Borosilicate)
MB (Soda/lime)
MC (Soda/lime)
MD (Soda/lime)
ME (Lead)
MF (Borosilicate)
MG (Borosilicate)
MH (Fluoride)
MI (Fluoride opal)
MJ (Lead)
ML (Borosilicate)
M} (Lead)
MY (Lead)
Flat /Float glass:
MR
MS
MW
MZ
Total Industry
X
s
N
± c
950 (2,094)
-
No.
8,184 (18,027)
1,678 (3,700)
-
2,703 (5,958)
2,326 (5,128)
-
-
3,579 (7,801) No.
-
~
-
-
-
-
4,336 (9,567)
0.955
7
0.701
Gas
Gas + boost
2 oil + boost
Gas
Gas
No. 2 Oil
Gas + boost
Gas
Gas
Gas
5 Oil + boost
Gas
Gas
Gas
No. 2 oil
Gas
Gas
0.06 (0.13)
6.80 (15)
11.98 (26.4)
2.00 (4.62)
10.16 (20.36)
2.68 (5.9)
0.84 (1.85)
3.12 (6.89)
12.79 (28.2)
6.35 (14)
1.83 (4.03)
9.27 (20.43)
9.27 (20.43)
21.55 (47.5)
16.10 (35.5)
14.92 (32.9)
8.71 (19.2)
7.88 (17.34)
1,455
18
0.597
-
10.48 (23.1)
52.94 (116.7)
-
-
5.49 (12.1)
-
-
-
-
-
2 ppm
0.05 ppm
-
-
-
-
25.68 (5.65)
3.250
3
5.479
1.000 ppm
.
<300 ppm
<2000 ppm
-
Trace b - -
Trace
21.64 (47.7)
-
Trace
543 ppm 0.1 ppm
533 ppm 0.1 ppm
-
-
-
-
547 ppm 0.1 ppm 21.64 (47.7)
0.779
8 21
0.522
a No HC data reported.
b Trace <0.1.
c 90 percent confidence band about the mean as
- Insufficient data
X.
-------
TABLE 13. COMPARISON OF DATA FOR SO EMISSION
x
Source(a)
SAD's
Site testing
Container glass,
Mg/yr
72-111
200-342
Pressed and
blown glass,
Mg/yr
51
0-31
Flat/float
glass,
Mg/yr
349
597
(a) Manufacturer's data was not supplied.
31
-------
(i.e., 25 percent) can probably be attributed to differences in the
rebuilding cycles among the tested tanks, operational differences among the
different manufacturers, and raw-material differences that occur because of
the different glass compositions present in the data. However, none of these
differences seriously detract from the large correlation that exists among
glass melting tanks using a variety of fuels and melting a variety of glasses.
Oxides of Nitrogen
The oxides of nitrogen created from the high-temperature combustion can
take many forms, such as NO?, ^0, NO, etc. and they are therefore listed in
Table 11 as NO . These oxides of nitrogen comprise the criteria pollutant
that has the second largest rate of emission from the glass melting furnaces.
Table 11 can be used to determine the annual emission rate for comparison
with the SAD's. For the glass container industry, the annual emission rate
from the tested furnaces is 133 to 241 Mg/yr. A similar calculation for the
data from the pressed and blown industry indicates that the annual emission
rate for NOX is 28 to 86 Mg/yr. Since only one flat glass furnace has been
tested under this program, a range cannot be calculated. The comparison of
these ranges to the data from the SAD's is shown in Table 14. The agreement
between these two data sources indicates that the existing NOX data is
adequate. The lack of manufacturer's data can probably be attributed to a
low priority for NO testing.
X
Since NOX production is the result of a high-temperature combustion
process, a regression between the NO emission rate and fuel use rate
(expressed as equivalent kilowatts) is not unreasonable. The results of such
a linear regression are as follows:
1. Correlation — 98 percent
2. Unexplained variation — 4 percent
3. Confidence of nonzero slope — 99.95 percent
The data that have been used to generate this regression line span a large
range of glass compositions, through-put rates, and melter sizes. Yet these
statistics strongly indicate that a real, positive, linear relationship
exists between the NOX emission range and the fuel use rate (i.e., an
increase in fuel use increases the NO emission rate). Therefore, in light
of the wide range of conditions that have been used to generate this line,
the differences among the various industries for NO emissions are not large.
Moreover, the regression indicates that the rate of NOX emissions is most
strongly dependent on the use of combustion air (proportional to fuel-use
rate) and is little influenced by the temperature of the operation. Such a
conclusion is not unexpected, because most glass furnaces generally operate
in the same high flame temperature realm (above 1400 C). This conclusion
is somewhat at variance with the brief NO discussions in the SAD's. Those
discussions, however, are based on an increase in furnace superstructure
temperature which may not necessarily correspond to an increase in flame
temperature. Also, the increased superstructure temperature is accomplished
by increased flows of fuel and combustion air. Thermodynamic calculations
32
-------
TABLE 14. COMPARISON OF DATA FOR NO EMISSION
x
Source
SAD's
Site testing
Container Glass,
Mg/yr
120-208
133-241
Pressed and
blown glass,
Mg/yr
21-41
28-86
Flat/float
glass,
Mg/yr
620
759
(a) Manufacturer's Data not supplied.
33
-------
show that at 1700 K (1427 C) the free energies of formation range from -11730
to -16830 cal/mole for N20 (g), NO (g), or N02 (g). These free energies are
clearly large enough to result in NOX formation. As larger concentrations of
reactants (N2 and 02) are supplied, larger concentrations of NOX can be
expected. The results of the regression analysis are completely consistent
with these thermodynamic considerations. Therefore, the cursory treatment in
the SAD's is insufficient to adequately explain the controlling variables in
NOX formation.
Total Particulate Matter
Particulate emissions from glass manufacturing furnaces have been
routinely reported by the manufacturers, as is evidenced by the data in
Table 12. The particulate emission data listed in Table 11 are the total
measured—the particulate catch from the probe, the filter, and the impingers.
The summaries in Table 11 can be used to calculate the range of total
particulate emission that can be expected from the various industries, and
these values can be compared to those reported in the SAD's. For the glass
container industry, the total particulate emission rate per furnace should
range from 65 to 96 Mg/yr. For the pressed and blown glass industry, the
total particulate emission rate per furnace should range between 20 and 30
Mg/yr. Again, firm comparison for the flat glass industry cannot be made
because only one flat glass furnace has been included in this program. The
comparison among the onsite data, the SAD's, and manufacturer's data is
shown in Table 15. For both the container glass and flat glass emission,
the onsite data from this program is nearly double that in the SAD's and
manufacturer's data. This circumstance is the result of reporting only the
front-half catch of the particulate train in the SAD and manufacturer's
data. The lower value for the onsite data range in the pressed and blown
industry may reflect an improved effort toward reducing the particulate
emission rate, or it may reflect a production rate for the plants tested
in this program that is lower than the industry norm. The data collected
are not adequate for statistically establishing the real source of the
lowered emission rate in the pressed and blown industry segment.
Several literature references in the SAD's for the glass industry
indicate that a relationship may exist between the logarithm of the
particulate emission rate and the process rate. The field-test data have
been put through a linear regression in those two variables with the
following result:
1. Correlation — 92 percent
2. Unexplained variation — 17 percent
3. Confidence of a nonzero slope — 99.95 percent.
The robustness of these statistics for the regression indicate that for a
particulate emission rate, the differences among the three manufacturing
segments of the glass manufacturing industry are not as large as expected.
The positive, nonzero intercept for the regression line indicates that a
minimum particulate rate exists when the process rate is zero. That rate is
attributable to vapor phase formation of particulates from the
34
-------
TABLE 15. COMPARISON OF DATA FOR PARTICIPATE EMISSION
Container glass
Source Mg/yr
SAD's 28-47
(a)
Site testing v ' 65-96
Manufacturer ' s
Data 29
Pressed and Flat/float
blown glass glass
Mg/yr Mg/yr
52-84 131
20-30 350
47-52 126
(a) Total catch — front half plus back half
35
-------
combination of combustion products and species evaporated from the molten bath
while the melting unit is in a soaking status.
Particulate emissions from the glass melting furnace are composed of
three parts—that which exists at stack temperature, that which is formed on
cooling to the filter temperature (120 C), and that which is formed from
either cooling to the impinger temperature (ambient), or by reaction with
impinger media. Table 16 provides a summary of the fraction of particulates
encountered in each of these three portions of the sampling train. Quite
clearly, most of the particulates already exist by the time the gas stream
has cooled to the stack temperature, and more than 80 percent of the partic-
ulate matter has been formed by the time the gas stream has cooled to the
filter temperature. These results are tssed to provide a justification for
requiring only the front half of the sample train (probe plus filter) to be
considered when reporting particulate emission data. However, the remaining
material that condenses in the impingers does account for 15 to 18 percent
of the total particulate emission rate, a nontrivial amount of matter.
The chemical composition of the particulate matter captured in the three
portions of the sampling train is partially listed in Table 17, which presents
the three most prevalent elements (in excess of 1,000 ppm), determined by
optical emissions spectroscopy. These data indicate that the captured
particulates are a complex mixture of the oxides, or perhaps sulfates, of the
constituents that are most easily volatilized from the surface of the molten
glass. The presence of aluminum (probably as Al^Oo) indicates that some of
the particulate matter may result from carryover of very fine dust in the
combustion products, since alumina does not easily volatilize from the molten
glass. The high fraction of aluminum in the impingers further indicates that
the particulate size for that species is too small to be trapped in either
the probe or the filter and must contact the liquid media before it can be
removed from the gas stream. This information does not exclude the possi-
bility that some of the species captured in the impingers result from the
chemical reaction between a vapor species from the glass melting furnace
and the vapor species from other processes in the immediate geographical
vicinity of those furnaces. The summaries in Table 17 also indicate that
boron, lead, and arsenic are not particularly large contributors in total
weight to the emissions from glass melting furnaces. Nevertheless, these
species can be particularly harmful in the immediate vicinity of those
furnaces which emit them. Petrographic analyses of some electrostatic
precipitator dusts that have been sent by the participating manufacturers
indicate that all of the particulate matter is quite small. These dusts are
primarily from furnaces that produce borosilicate and lead glasses. For the
borosilicate dust, most of the particulate is found to be optically
anisotropic and is probably either B-O- or H BO-. The rest of the
particulate dust is isotropic and is probably very fine glass . Also,
presurvey vis its to a plant that melts lead glass have indicated that the
stack is lined with a very fine yellow dust that is probably a compound of
lead. This observation indicates that lead compounds such PbO and PbSO,
form a significant portion of the particulate matter from furnaces that
36
-------
TABLE 16. PERCENTAGE OF PARTICULATE LOAD-
ING IN THE SAMPLING TRAIN
Source
A
B
C
D
E
F
G
H
I
J
K
L
M
X
s
Na
± a
Stack
76.8
17.3
57.6
78.0
40.6
89.4
77.3
80.2
59.8
30.8
58.0
55.7
50.0
59.4
21.1
13.0
13.0
Filter + Probe
3.6
24.7
0.0
6.1
52.0
8.3
18.7
10.4
21.6
57.0
33.0
27.0
33.9
22.8
17.8
13.0
29.0
Impingers
19.6
58.0
42.4
15.9
7.4
2.3
4.0
9.4
18.6
12.2
9.0
17.3
16.1
17.9
15.7
13.0
33.0
90% confidence band about
the mean as % X.
37
-------
TABLE 17. MOST PREVALENT CHEMICAL SPECIES'
IN PARTICUIATE TRAINb
Source
A
B
C
D
E
F
G
H
I
J
K
L
M
1st
Si
Ca
Ca
(c)
Na
Na
Na
(c)
Na
Na
Fe
Na
Na
Probe
2nd
Na
Na,Si
Kg
(c)
K
B
B
(c)
Ca
K
Pb
Ca
Ca
Filter
3rd
Al,Ca
Mg
Na
(c)
Ca,Si
K
K
(c)
K
V
K
K,Cr
Mg
1st
Na
K
Ca
(c)
Na
Na
Na
(c)
Na
Na
Pb
Na
Na
2nd
Al
-
Na
(c)
K
B,K
K
(c)
K
K
K
K
K
3rd
K
-
As
(c)
Ca
As
B
(c)
Ca
Ca
Na
Ca
Ca
Imoinsers
1st
Al
Si
Ca,Si
(c)
Na
Si
Ca
(c)
Sn
Sn
Ca
Sn
Sn
2nd
-
-
Na
(c)
Ca,Si
Ca,Al
Al
(c)
-
Si
Na
Mg
M
3rd
-
-
Mg.Fe
(c)
K
K
Si,Na
(c)
-
-
Sn
-
""
a Determined by optical emission spectro-
scopy to be above 1,000 ppm.
b EPA Method 5.
c No data available.
38
-------
melt lead glasses. Unfortunately, this plant could not be scheduled as
one of the tested sites, and no quantitative information is available from it.
The size distribution of the particulate matter was determined with
an Anderson cascade impactor. Representative plots of particle sizes
captured are shown in Figure 5. These graphs indicate that there is a
slight difference among the various melting operations, but that all of the
particulate matter' is very fine, with the median diameter resting somewhere
between 0.26 and 0.94 urn. Table 18 provides a summary of the size distribu-
tion data that have been gathered from the tested plants. That summary
indicates that the median diameter is between 0.5 and 0.83 vim (90-percent
confidence band at about the mean value) and that 79 to 88 percent of the
collected particulates are less than 3 urn in diameter. In other words, the
vast majority of particulate matter emitted from glass melting furnaces is
within the respirable size range. No relationship between particle size and
stack temperature can be established. Therefore, the size distribution of the
particulate matter captured from glass melting furnaces is established by
the time the gas stream has cooled to the stack temperature.
Other Emissions
The remaining criteria pollutants that can be expected from the glass
melting operation are unburned hydrocarbons and carbon monoxide that result
from incomplete combustion of the fossil fuels. Both Tables 11 and 12 present
data for these two species, but data on all potential emissions are not
available. Clearly, those instances that do not yield HC and CO are the ones
in which combustion is complete. The data in both of these Tables indicate
that for both criteria pollutants, trace amounts to amounts that are less
than 16 ppm can be generated from the glass melting furnaces and these
values agree with those extracted from the SAD's (Table 10).
FORMING AND POSTFORMING
The forming and postforming emissions can consist of organics, acid fumes,
and tin vapors. These emissions are the most difficult to characterize
because they are usually released inside the building and escape to the
atmosphere through roof-top ventilators. The organic emissions from various
participating manufacturers have been characterized using the previously
described test procedures. Tin vapors from the production of flat glass, and
acid fumes from etching and frosting operations have been surveyed from
data sent by manufacturers in those industries.
Organic Emissions
Organic compounds are released to the atmosphere from sprays and
lubricants used in the forming area to effect release of the finished product
from the mold, and in the postforming area from lubricity spraying. In
general, the emissions from these two operations are not separable because
of the construction of the roof-top ventilators. However, the manufacturing
facility at one of the tested plants has separate roof-top ventilators over
the forming and postforming areas, and separate measurements have been taken
39
-------
10.0,—
5.0
w
d
1.0
- 0.94
- 0.26 \i
.1
I—*
3
0.01
J L
Legend
CONT - CONTAINER GLASS
P&B - PRESSED AND BLOWN GLASS
FLAT- FLAT GLASS
10 50
CUMULATIVE PERCENT
90
99
99.99
Figure 5. Typical particle size distribution.
40
-------
TABLE 18. PARTICULATE SIZE DATA
Source
A
B
C
D
E
F ,
G
H
I
J
K
L
M
X
s
N
±b
Median
diameter, |J.
0.68
0.60
0.90
0.55
0.25
0.36
0.33
0.72
1.20
1.30
0.49
1.04
0.82
0.71
0.33
13.0
17.0
Respirable3
fraction ., °L
80.5
96.0
67.0
87.1
87.5
96.0
98.4
85.8
64.0
64.0
89.8
82.0
83.1
83.2
11.7
13.0
5.0
Stack
temperature, C
158
293
456
218
276
269
257
220
432
449
904
244
249
341
194
13
38
a Cumulative fraction <
b 90% confidence band as % X.
41
-------
at these two locations.
Table 19 lists the total loading and the weight fractions of the eight
classes of organic compounds that can be detected with the porous polymer
adsorbent trap. Statistics for the loading (expressed as Mg/nm , or grains/
DSCF) show a large variability indicative of the widely-ranging practices in
the application of the various release and lubricity compounds. The
statistics for the eight organic fractions are also quite variable, but
Table 19 does indicate that most of the organics can be found in Fractions
1 through 3 and 7 through 8.
One of the tested manufacturing facilities has separate ventilators over
the forming and postforming areas, and this construction allows separate
analyses to be conducted in these ventilators. A schematic plan view of that
facility is given in Figure 6 and the results of the sampling of this facility
are shown in Table 20. As shown in this table, the forming emissions tend to
generate the organic materials that are captured as the first three fractions
in the porous polymer adsorbent traps. The postforming emissions tend to
favor the species that can be captured as the final three fractions of the
adsorbent trap. Care must be exercised in interpreting Table 20, because
it represents data from only one plant.
Emissions from the forming operation present a particularly difficult
control problem because no specific route for the gas stream is provided—it
merely exits through a ventilator in the roof. Some glass manufacturers are
replacing some of the oils and sprays used for lubrication with water
soluble silicons and are thereby reducing the output of organic emissions from
this process step. This replacement is slow, however, because of the large
financial penalty that can be expected if production is lowered by unforeseen
problems. Moreover, much more research is needed before the severity of the
emission problem for this process step can be properly established. If that
research indicates that a real emission problem does exist, then the most
reasonable means of reducing that problem is to find replacements for the oils
and sprays currently in use. These replacements can only be established
through further research that defines the specific functions and mechanisms
involved in the release aid provided by these dopants.
Comparison of data in Table 19 and 20 to those contained in the SAD's is
inappropriate because the latter list the data as estimated emission factors
or weight of emission per weight of product. The data gathered onsite for
organic emissions cannot be converted into such units because the volume flow
through the roof-top ventilators is not calculable. However, these onsite
data are more reliable than the mass balance estimates in the SAD's.
Acid Fumes
Acid-fume emissions are generated in the pressed and blown glass industry
from etching and polishing, which are postforming operations. Uncontrolled
emissions from these processes have not been directly measured in this program.
However, some manufacturers did supply data from their own files (Table 21).
42
-------
TABLE 19. PERCENT BY WEIGHT OF CAPTURED ORGANIC EMISSIONS
FROM FORMING AND POSTFORMING OPERATIONS
Source
A
(Container)
C
Pressed and
blown
I
(Container)
J
(Container)
L
(Container)
M
(Container)
X
s
a
±
Loading
mg/m3
[grains/dscf]
7.482
[0.00326]
5.694
[0.00248]
0.979
[0.00426]
0.966
[0.00420]
8.046
[0.0035]
15.852
[0.0069]
6.503
[0.00283]
5.5
[0.00240]
51
(51)
Aliphatic
hydrocarbons
28.9
52.2
5.9
11.0
9.4
1.4
18.1
19.1
60.0
Aromatic
hydrocarbons,
POM,
PCB,
halides,
(2)
2.5
3.9
13.7
67.8
4.7
1.8
15.7
25.9
99.0
Esters
ethers,
nitro compounds,
halides
(3)
10.2
2.2
63.4
0.0
1.0
0.9
13.0
25.0
115.0
Phenols
esters,
ke tones,
aldehydes,
phthalates
(4)
26.3
4.4
7.5
0.0
0.0
0.1
6.4
10.2
96.0
Phenols ,
alcohols,
phthalates,
amines
(5)
8.0
2.0
0.0
2.4
1.1
3.6
2.8
2.8
60.0
Amides
sulfonates,
aliphatic acids, Sulfonates,
carboxylic, sulf oxides, Sulfonic,
acid salts, sulfonic acids acids
(6) (7) (8)
3.3 18.3 2.5
6.8 20.6 7.0
0.0 8.8 0.6
0.0 18.8 0.0
8.7 7.7 67.3
11.0 22.2 58.8
5.0 16.1 22.7
4.6 6.2 31.5
55.0 23.0 84.0
a 90 percent confidence band around the mean as % X.
-------
FORMING
A
A
POSTFORMING
LEHR
LEHR
LEHR
/'
j;
t\
LEHR A
•n
H
. f
if
PACKAGING
Figure 6. Sampling locations as viewed from the ventilator (arrows indicate sampling locations)
-------
TABLE 20,. SEPARABLE FORMING AND POSTFORMING
ORGANIC EMISSIONS
Captured species
Forming
Percent by weight captured
f?.0*
Postforming
(1) Aliphatic hydrocarbons 22.0
(2) Aromatic hydrocarbons
POM,
PCB,
halides
(3) Esters,
ethers,
nitro-compounds
11.3
1.6
4.3
10.0
3.3
0.0
0.0
0.5
0.6
0.3
0.5
(4) Phenols,
esters,
ketones,
aldehydes,
phthalates
(5) Phenols,
alcohols,
phthalates,
amines
0.0
0.0
0.2
2.2
0.0
1.9
0.0
4.2
(6)
(7)
(8)
Amides,
sulfonate,
aliphatic acids,
carboxylic acid salts
Sulfonates,
sulf oxides,
sulfonic acids
Sulfonic acids
Loading mg
3
m
Drains.,
( dS5P
15.6 3.6 0.0 9.0
2.7 76.2 0.0 4.8
46.7 0.0 97.6 80.6
3.035 2.157 1.962 9.848
(0.00132) (0.000939) (0.000854) (0.00249]
a Manufacturer's code
45
-------
TABLE 21. UNCONTROLLED FUMES FROM ACID ETCHING
Type of
Source glass HF , kg/hr (Ib/hr)
MQ TV 0.77 (1.702)
VS&1 Soda-lime 3.6 (7.9)
MB2 Soda-lime 3.1 (6.8)
MB3 Soda-lime 1.8 (4.0)
46
-------
Since acid-etching procedures have been used in the industry for several
years, these values are taken to be representative of the practices in the
industry. These values are within the range of information reported in the
SAD's (0.18 to 0.96 g of emission/kg of glass production).
Tin Vapors
The flat glass industry manufactures most of its products through the
float process. In this process, the molten glass passes over a bath of molten
tin to form the final product. This molten tin bath presents a potential
source of emissions from the flat glass manufacturing operations. One of the
participating manufacturers has supplied data that indicate that the tin
usage is approximately 50 Ib/yr, or 0.2 mg of tin (vaporized)/kg of glass
pro'duced. That value is nearly a factor of ten less severe than the value
of 1 mg/kg of glass, as has been reported in the SAD for the flat glass
industry. The actual emission factor for tin vapors is probably considerably
less than the estimated value of 0.2 mg/kg of glass because most of the tin
is lost through adhesion of a very thin tin film to the glass as it is
being drawn over the molten tin bath.
Other Emissions
Other potential emissions from the forming and postforming operations
include the products of combustion from the annealing layers, organic emissions
from the decorating media, and hydrochloric acid vapors from surface treat-
ment operations. The character of these emissions have not been discussed
because they either are not present in the tested plants or they are not
separable by the techniques used for this study. The largest potential
source for these other emissions is the combustion products from the annealing
operations. These operations generally use natural gas, and the precise
temperature control requires a good control over the combustion process.
Such carefully controlled combustion processes have a very low potential
for emitting unburned hydrocarbons or carbon monoxide. Therefore, the
estimates provided in the SAD's are not expected to be seriously in error with
actual practice.
None of the tested manufacturing operations produced articles that were
decorated, so no measure of emissions from this type of operation was made.
Some of the tested plants did apply lubricity coatings that could potentially
yield organic emissions and hydrochloric acid vapors. If organic emissions
were present, they were measured with the porous polymer adsorbent traps. On
the other hand, hydrochloric acid emissions, if present, were not measurable
by the techniques used in this study. Thus, SAD estimates for the emission
rates of organics from product decorating and for hydrochloric acid vapors
from lubricity spraying operations were not altered as a result of the present
study.
47
-------
REFERENCES
1. Schorr, J. R., Hooie, D. T., Sticksel, P. R., Brockway, M. C; Source
Assessment: Glass Container Manufacturing Plants EPA-600/2-76-269,
U. S. Environmental Protection Agency, Cincinnati, Ohio, 1976.
2. Reznik, R. B., Source Assessment: Flat Glass Manufacturing Plants
EPA 600-2/2-76-032b, U. S. Environmental Protection Agency,
Cincinnati, Ohio, 1976.
3. Schorr, J. R., Hooie, D. T., Brockway, M. C., Sticksel, P. R.,
Niesz, D. E., Source Assessment: Pressed and Blown Manufacturing
Plants EPA 600/2-77-005, U. S. Environmental Protection Agency,
Cincinnati, Ohio, 1976.
4. 1972 National Emissions Report EPA-450/2-74-012, U. S. Environmental
Protection Agency, Cincinnati, Ohio, 1972.
5. Cheremisinoff, P. N., and Young, R, A., "Control of Fine Particulate
Air Pollutants: Equipment Update Report", Pollution Engineering,
pp 22-29 (August, 1976).
6. Jones, P. K., Graffeo, A. P., Detrick, R., Clarke, P. A., Jakobsen,
R. J., Technical Manual for Analysis of Organic Materials in Process
Streams EPA 600/2-76-072, U. S. Environmental Protection Agency,
Cincinnati, Ohio, 1976.
7. Suta, B. E., Human Exposure to Atmospheric Arsenic, report to U. S.
EPA/ORD (Project Office A. P. Carlin, contract 68-01-4314) from
Center for Resource and Environmental System Studies, April, 1978.
48
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/2-79-101
3. RECIPIENT'S ACCESSION NO.
. TITLE AND SUBTITLE
.ummary Report on Emissions From the Glass Manufacturing
industry
5. REPORT DATE
April 1979 issuing date
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
5. D. Spinosa, D. T. Hooie, and R. B. Bennett
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Battelle Columbus Laboratories
Columbus, Ohio 43201
10. PROGRAM ELEMENT NO.
1AB604
11. CONTRACT/GRANT NO.
Contract No. 68-01-3159
Contract No. 68-01
- b. Laboratory
Office of Research and Development,
U.S. Environmental Protection Agency
Incinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
14/5PORSORING AGENCY CODE ~
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This project was undertaken to evaluate emissions rates from typical glass manufacturing
furnaces. The effort concentrated on the container segment of the industry, however,
tests were also conducted on the pressed blown, and flat glass segments of the industry.
The quantitative results of the test program were compared to earlier calculated
results derived in the source assessment documents for each segment of the industry.
Additional data collected during this test program included particle size distributions
of glass furnace emissions and trace metals analyses of glass furnace emissions. Other
sources within the typical glass manufacturing plant were also evaluated to determine
the types of pollutants that are generated from these sources.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Exhaust Emissions
Trace Elements
Particle Size Distribution
Glass Manufacturing
Glass Furnace
Pollution
99A
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
51
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
49
ft U.S. GOVERNMENT PRINTING OFFICE: 1979 -657-060/5320
------- |