EPA-600/2-76-032b
March 1976
Environmental Protection Technology Series
                                     SOURCE  ASSESSMENT:
                     FLAT GLASS 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.    Socioeconomic 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 policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                               EPA-600/2-76-032b
                               March 1976
             SOURCE ASSESSMENT:

      FLAT  GLASS MANUFACTURING PLANTS
                     by

             Richard  B.  Reznik
       Monsanto  Research  Corporation
             Dayton  Laboratory
            Dayton,  Ohio  45407
          Contract  No.  68-02-1874
             ROAP No. 21AXM-071
         Program Element No. 1AB015
     EPA Project  Officer:  Dale A. Denny

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 con-
sidered include: process modifications, feedstock modifica-
tions, add on control devices, and complete process substi-
tution.  The scale of control technology programs range
from bench to full scale demonstration plants.

The Chemical Processes Branch 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 flat glass manufacturing plants.

Monsanto Research Corporation has contracted with EPA to
investigate the environmental impact of various industries
which represent sources of emissions in accordance with EPA's
responsibility as outlined above.  Dr. Robert C. Binning
serves as Program Manager in this overall program entitled,
"Source Assessment," which includes the investigation of
sources in each of four categories:  combustion, organic
materials, inorganic materials and open sources.  In this
study of flat glass manufacturing plants, Mr. Edward J.
Wooldridge served as EPA Project Leader.

                             iii

-------
This study was completed by IERL-RTP.  Project responsibility
was transferred to the Industrial Pollution Control Division
of lERL-Cincinnati on October 15, 1975.
                               IV

-------
                          CONTENTS
Section                                                Page
I    Introduction                                        1
II   Summary                                             3
ill  Description of the Flat Glass Industry              9
     A.   Types of Glass Plants                          9
     B.   Types of Glass                                13
     C.   The Glass Manufacturing Process               18
          1.   Preparation of Raw Materials             19
          2.   The Glass Melting Furnace                22
          3.   Flat Glass Forming and Finishing         26
               Operations
IV   Emissions                                          37
     A.   Selected Pollutants                           37
          1.   Raw Materials Preparation                37
          2.   Glass Melting Furnace                    39
          3.   Forming and Finishing Operations         46
     B.   Emission Characteristics                      47
          1.   Raw Materials Preparation                47
          2.   Glass Melting Furnace                    48
          3.   Forming and Finishing Operations         48
     C.   Environmental Effects                         50
          1.   Total Emissions                          50
          2.   Ground Level Concentrations              50
          3.   Affected Population                      53
V    Control Technology                                 55
     A.   Preparation of Raw Materials                  55
     B.   The Glass Melting Furnace                     59
          1. '  Process Modification                     61
          2.   Pollution Control Devices                65
     C.   Forming and Finishing Operations              74
VI   Growth and Nature of the Flat Glass Industry       75
VII  Unusual Results                                    77
                               v

-------
Section                                                Page
VIII Appendixes                                         79
     A.   Calculation of Glass Production on a          81
          Tonnage Basis
     B.   Flat Glass Plant Listing                      85
     C.   Emissions Data                                89
     D.   Total Flat Glass Emissions                   111
     E.   Plume Rise Correction                        119
     F.   Derivation of Source Severity Equations      121
IX   Glossary of Terms                                 137
X    Conversion Factors                                139
XI   References                                        141
                              VI

-------
                       LIST OF FIGURES
Figure                                                 Page
   1      Locations of Flat Glass Plants                14
   2      Flow Diagram of Flat Glass Manufacturing      20
   3      Process Flow Diagram of a Batch Plant         21
   4      Regenerative Side Port Glass-melting Furnace  23
   5      Flow Diagram of Forming and Finishing         27
          Operations
   6      Flat Glass Manufacturing:  Pilkington Float   28
          Process
   7      Flat Glass Manufacturing:  Fourcault Process  31
   8      Flat Glass Manufacturing:  PPG Pennvernon     33
          Process
   9      Flat Glass Manufacturing:  LOF-Colburn        34
          Process
  10      Flat Glass Manufacturing:  Rolling Process    35
  11      Points of Emission During the Preparation     38
          of Raw Materials
                                VII

-------
                       LIST OF TABLES
Table                                                  Page
  1       Emissions from the Glass Melting Furnace       4
  2       Source Severity for Flat Glass Emissions       6
  3       Affected Population                            7
  4       U. S. Glass Industry,  1972                    10
  5       Plant Listing for Flat Glass Industry         11
  6       Compositions by Weight Percent of             16
          Commercial Glasses
  7       Particulate Emissions  During the              40
          Preparation of Raw Materials
  8       Emissions from the Glass Melting Furnace      41
  9       Emissions from Gas-fired Annealing Lehrs      47
 10       Characteristics of Emissions from Glass       49
          Melting Furnace
 11       Total Emissions                               50
 12       Stack Heights of Flat  Glass Furnaces           51
 13       Source Severity for Flat Glass Emissions      53
 14       Affected Population                           54
 15       Particulate Controls for the Preparation      56
          of Raw Materials
 16       Acceptable Mesh Specification for Glass       58
          Sand
 17       Acceptable Mesh Specification for Glass       60
          Grade Dolomite
 18       Acceptable Mesh Specification for Glass       60
          Grade Limestone
 19       Time for NO Formation  in a Gas Containing     65
          75% Nitrogen and 3% Oxygen
 20       Maximum Use Temperature for Fabric Filters    '69
 21       Control Techniques for Glass Melting           72
          Furnaces
 22       Flat Glass Production  Statistics              76
 B-l      Listing of Flat Glass  Plants                  86
 C-l      Summary of Source Test Data from NEDS -       90
          Materials Handling and Mixing
 C-2      Particulate Emissions  from the Preparation     91
          of Raw Materials

                              viii

-------
Table                                                  Page
 C-3      Stack Heights                                 93
 C-4      NO  Emissions from Glass Melting Furnaces     94
            A
 C-5      Source Test Data from NEDS - SO  Emissions    96
          from Glass Melting Furnaces
 C-6      Source Test Data from NEDS - Particulate     100
          Emissions from Glass Melting Furnaces
 C-7      CO Emissions from Glass Melting Furnaces     106
 C-8      Hydrocarbon Emissions from Glass Melting     106
          Furnaces
 C-9      Emissions from Gas-fired Burners             109
C-10      Emission Factors for Annealing Lehrs         109
 D-l      State by State Listing of Emissions          112
 D-2      NEDS Emission Summary by State               113
 D-3      State Listing of Emissions as of July 2,     115
          1975
 E-l      Plume Rise for Flat Glass Furnaces           120
 F-l      Pollutant Severity Equations                 121
 F-2      Values for a for the Computation of a        123
 F-3      Values of the Constants Used to Estimate     124
          Vertical Dispersions
 F-4      Summary of National Ambient Air Quality      128
          Standards
                               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/aciru
      H.
     B_,             The ratio -H2/2c2
      l\
     D.             Inside stack diameter
      i
      e             Natural logarithm base
      H             Effective stack height
      h             Physical stack height
     AH             Plume rise
      k             "Student t" test variable
      n             Number of samples
      p             Atmospheric pressure
      Q             Mass emission rate
      s             Sample standard deviation
      S             Source severity, ratio X   /AAQS
                                            HiaX
     T              Ambient temperature
      a
     T              Stack gas temperature
     t0             Instantaneous averaging time of 3 minutes
      t             Averaging time
     TLV            Threshold limit value
      u             National average wind speed
     V              Stack gas exit velocity
      x             Downwind dispersion distance from source
                    of emission release
                              XI

-------
              LIST OF SYMBOLS (Continued)
Symbol                     Definition

   y        Horizontal distance from centerline of
            dispersion

   y        Sample mean

   TT        3.14

  a         Standard deviation of horizontal dispersion

  a         Standard deviation of vertical dispersion
   Z
   X        Downwind ground level concentration at
            reference coordinate x and y with emission
            height of H

   x"        Time average ground level concentration of an
            emission

 X          Instantaneous maximum ground level concentration

 X^          Time average maximum ground level concentration
                            xn

-------
                          SECTION I
                        INTRODUCTION

Flat glass is manufactured at 29 plants in 14 states, and the
production of finished products amounted to 2.9 x 106 metric
tons  in 1972.  However, total production was approximately
4 x 106 metric tons, the difference being lost to breakage,
edge loss, cutting, and off-quality glass.

This report discusses air emissions released during the
manufacture of flat glass.  The different emission points
within the manufacturing process are identified, the types
and quantities of emissions from each point are delineated,
and the characteristics of air pollutants are listed.  State
and national emissions of criteria pollutants (particulates,
NO , SO , CO, and hydrocarbons) from the flat glass industry
  2C    X
are compared to total state and national emissions from all
stationary sources.  The maximum average ground level
concentrations of emissions from a typical flat glass plant
are compared to the corresponding ambient air quality
standards.  The effect of present and emerging control
technology is also discussed.

The manufacturing process is divided into three phases:
(1) preparation of raw materials, (2) glass melting, and
 1 metric ton = 10  grams = 2,205 pounds - 1.1 short tons
 (short tons are designated "tons" in this document); other
 conversion factors and metric system prefixes are presented
 in Section X.

-------
(3) forming and finishing operations.  The first two oper-
ations are common to all flat glass products while the latter
has four variations:

   •  Manufacture of float glass (77% of total capacity).
   •  Manufacture of sheet glass (17% of total capacity).
   •  Manufacture of rolled glass (4% of total capacity).
   •  Manufacture of plate glass (2% of total capacity).

During the past 10 years, the float process has replaced
almost all of the older plate glass capacity, and only one
plate glass plant is still in operation.  The future may
see a changeover of some sheet glass production to thin
float glass.

-------
                         SECTION II
                           SUMMARY

This document describes a study of air emissions released
during the production of flat glass, Standard Industrial
Classification No. 3211.  It encompasses the preparation of
raw materials (sand, limestone, and soda ash) at the plant
site, the production of molten glass in the melting furnace,
and the forming of flat glass products (sheet glass, float
glass, and rolled glass).  It does not cover the production
of plate glass,  which is now practiced commercially at only
one plant in the United States.

The estimated 1972 production of finished flat glass products
was 2.9 x 106 metric tons (3.2 million tons) ± 10%.  The
amount of glass actually made was higher (M x 106 metric tons)
because of breakage and off-quality glass.   The quantity of
raw materials handled was 4.5 x 106 metric tons, of which
5 x 105 metric tons volatilized (over 90% to C02) during
melting.

Flat glass is manufactured at 29 plants in 14 states, with
Tennessee, Pennsylvania, and Ohio accounting for 46% of total
capacity.  The average county population density at a plant
site is 248  persons/km2.  The majority of flat glass capacity
is float glass (77%) while sheet glass accounts for 17%,
rolled glass for 4%, and plate glass for 2% of the total.

-------
A typical flat glass plant has a capacity of  500 metric  tons/
day  (550 tons/day) and a yearly production  (75%  capacity)  of
1.4 x 105 metric tons.  Over  99% of  the  plant emissions  arise
from the glass melting furnace, with NO  , SO  , and  particu-
                                       X   X
lates the major  (>99%) pollutants.   The  particulates  are
alkali sulfates of submicron  particle  size.   Furnace  stack
heights average 30 m  (100 ft) when ejection air  is  used  and
60 m (200 ft) for natural draft.

Emissions from the melting furnace are listed in Table 1
along with emission factors and total annual  emissions.
Nitrogen oxides have the highest emission factor  (4 g/kg)
and annual emissions  (1.6 x lO4 metric tons).  When national
emissions of each pollutant from the flat glass  industry are
compared to the corresponding national emissions from all
stationary sources, NO  also contributes the  greatest percent
                      X
(0.07%).   The greatest contributions on  a statewide basis  are
for NO  emissions in Tennessee (0.6%) and Oklahoma  (0.5%).
      J^

         Table 1.  EMISSIONS FROM THE GLASS MELTING  FURNACE
Species
NO
x
SO
x
Particulates
CO
Hydrocarbons
Emission
factor ,
g/kg (Ib/ton)
4 ± 30%
(8)
1.5 ± 27%
(3)
1 ± 60%
(2)
0.02 ± 100%
(0.04)
0.04 ± 100%
(0.08)
Total annual
emissions (based on
4.0 x 106 metric tons
glass manufactured) ,
10 3 metric tons (tons)
16.0 ± 4.8
(17,600)
6.0 ± 1.6
(6,600)
4.0 ± 2.4
(4,400)
0.08 ± 0.08
(88)
0.16 ± 0.16
(180)
Percent of
national
emissions from all
stationary sources
0.07
0.02
0.02
<0.01
<0.01

-------
The preparation of raw materials gives rise to particulate
emissions (90 metric tons/yr ± 100%)  from handling operations
 (unloading,  conveying, storage bin vents, glass crushers, and
raw material .mixers.  The composition of the emissions is the
same as that of the raw materials (i.e., sand, soda ash, lime).
Dusting is controlled by enclosing the handling operations and ...
filtering the exhaust air from storage bins, crushers, and
mixers.  In addition, the glass batch ingredients are purchased
with size specifications that limit the amount of <44 ym  (<325-
mesh) particles.  Over 90% of the industry employs controls,
and the average emission factor for this phase of production
is 0.02 g/kg  (0.04 Ib/ton) ±  100% of material processed.

The only atmospheric emissions from forming and finishing
operations are combustion products from gas-fired annealing
lehrs.  The amount of these emissions is so small (national
emissions <100 metric tons/yr) that they are not controlled.

Emissions from flat glass melting furnaces are not controlled
with add-on equipment because the industry is able to meet
emission standards with proper operating conditions.  Sub-
micron particulates in the stack gas may cause opacity limits
to be exceeded.  These emissions have been controlled in other
sectors of the glass industry by baghouses and electrostatic
precipitators.  Although scrubbers were found to control SO
                                                           X
emissions, operational problems were experienced and perfor-
mance on particulates (<90% efficiency) was not satisfactory.
No controls have been developed for NO  emissions.
                                      J\,

One measure of the potential environmental effect of melting
furnace emissions is their maximum average ground level con-
centrations  (x^   ) .  These have been calculated for a typical
              in 3.x
flat glass (500 metric tons/day) plant and appear in Table 2.
A Gaussian plume dispersion model was used to calculate
                          <&
values of x"    for stack heights of 30 m and 60 m.
           IflclX

-------
          Table 2.  SOURCE SEVERITY FOR FLAT GLASS EMISSIONS



Emission
NO
X
SO
X


Particulates



CO



Hydrocarbons



Primary ambient
air quality standard,
mg/m3
0.100

0.365



0.260



40.0



0.160
Averaging time
Annual arith-
metic mean
24 hr; not to
be exceeded
more than once
per year
24 hr; not to
be exceeded
more than once
per year
1 hr; not to
be exceeded
more than once
per year
3 hr; 6-9 AM
X /
Amax
yg/m3
Stack
height
=30ma
130

44



29



1.0



1.7
Stack
height
=60ma
57

20



13



0.45



0.74


S
Stack
height
=30ma
1.3

0.12



0.11



<0.01



0.01
Stack
height
=60ma
0.57

0.05



0.05



<0.01



<0.01
                                                           %max
 Effective stack height is 30.m higher due to plume rise.

The source severity,  S,  has  been  defined as the ratio of
to the primary ambient air quality standard (AAQS) for crite-
ria pollutants.  Values  for  S  and AAQS are also listed in
Table 2.  The  largest severity factors are for NO  emissions
from a 30 m stack  (S  =  1.3)  and a 60 m stack (S =  0.57).   The
severities for SO   and particulates lie between 1.0 and  0.05,
while those for CO  and hydrocarbons are < 0.01.
The affected population  has  been defined as the population
around a typical plant who are  exposed to a x (average ground
level concentration) value which is >0.1 or 1.0 of the cor-
responding AAQS.  This value is given for each emission
species in Table 3.  The largest value is for NO  emissions
                                                 3C
from a 30 m stack  (14,600 persons)..

-------
                Table 3.  AFFECTED POPULATION



Emission
N0x
SO
X
Particulates
CO
Hydrocarbons

N0x
sox
Particulates
CO
Hydrocarbons
X/AAQS > 0.1
Affected area, km2
Stack
height. =
30 m°
59.0
3.0

2.6
0
0
Stack
height.=
60 m°
55.8
0

0
0
0
Affected population3
Stack
height. =
30 mD
14,600
750

640
0
0
Stack
height.=
60 mD
13,800
0

0
0
0
X/AAQS > 1.0
2.8
0
0
0
0
0
0
0
0
0
703
0
0
0
0
0
0
0
0
0
 Based on an average population density of 248 persons/km2.
 Effective stack height is 30 m higher due to plume rise.

Over the past 10 years the flat glass industry has experienced
a profound technological change as the old plate glass manu-
facturing process has been replaced by the new float process.
In the future float glass may also supplant much of the
sheet glass market.
Flat glass production experienced an annual growth rate of
10% from 1967 to 1973.  If growth recovers after the 1974
economic slump, 1978 production will be 46% above that of
1973.  Industry emissions will also increase by this amount
without new developments in control technology.  Two factors
which cannot be evaluated are the decrease in emissions due
to better furnace operations, and the increase in emissions
due to a conversion from gas to oil firing.  Gas firing is
preferred whenever possible because it causes less wear on
refractories in the melting furnace.
                              7

-------
                         SECTION III

           DESCRIPTION OF THE  FLAT GLASS  INDUSTRY


A.   TYPES OF GLASS PLANTS


The manufacture of glass is an important American industry

with 1972 sales of $5.5 billion and glass production of

1.86 x 107 metric tons.  The industry has been subdivided by

the U.S. Department of Commerce into four categories1  (Table

4), of which only flat glass is considered in this report.'

Manufacturing data in each group are compiled by the depart-

ment and published in various government reports,  the most

comprehensive being the 1972 Census of Manufactures.2"5
Standard Industrial Classification Manual, 1972 Edition.
 Washington.  Superintendent of Documents, 1972.  p. 136-138.

Preliminary Report, 1972 Census of Manufactures, Industry
 Series, Flat Glass, SIC 3211.  U.S. Department of Commerce.
 Washington.  MC 72(P)-32A-1.  January 1974.  7 p.

Preliminary Report, 1972 Census of Manufactures, Industry
 Series, Glass Containers, SIC 3221.  U.S. Department of ,
 Commerce.  Washington.  MC 72(P)-32A-2.  December 1973.  6 p.

^Preliminary Report, 1972 Census of Manufactures, Industry
 Series, Pressed and Blown Glass, SIC 3229.  U.S. Department
 of Commerce.  Washington.  MC 72(P)-32A-3.  February 1974.
 7 P.                                              ...

Preliminary Report, 1972 Census of Manufactures, Industry.
 Series, Products of Purchased Glass, SIC 3231.  U.S.
 Department of Commerce.  Washington.  MC 72(P)-32A-4.
 February 1974.  7 p.                     .'.,..:

-------
           Table 4.  U.S. GLASS INDUSTRY, 19722~5



Standard Industrial
Classification (SIC)
3211 - Flat glass
3221 - Glass containers
3229 - Pressed and blown
glass, N.E.C.
(not elsewhere
classified)
3231 - Products of
purchased glass
TOTAL
Production
of glass,
106 metric
tons (details
in Appendix A)
2.9
11.7
4.0



oa

18.6


Percent
of total
production
15.6
62.9
21.5



0

100.0


Number
of
plants
31b
115
254



860

1,260
 Plants in this industry use glass made at other sites.
 Two plants have closed since 1972.

The flat glass industry is composed of 29 plants in 14
states.  These are listed by company in Table 5, along with
the type of glass they produce:  float glass, plate glass,
sheet glass, or rolled glass.6'7
The four flat glass products differ in the way they are
formed.  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 the old plate glass which required grinding and
6Directory Issue.  The Glass Industry.  54 (10);1-178,
 September 1973.
71974 Glass Factory Directory Issue.  American Glass Review.
 9£(8A):1-204, February 1974.
                              10

-------
    Table 5.  PLANT LISTING FOR FLAT GLASS INDUSTRY6'7

      (A more detailed listing is given in Appendix B.)
Company and location
PPG Industries
Carlisle, PA
Cumberland, MD
Crystal City, MO -
Fresno, CA
Meadville, PA
Mt. Vernon, OH
Mt. Zion, IL
Wichita Falls, TX
Total est. capacity
Libbey-Owens-Ford Co.
Charleston, WV
Lathrop, CA
Laurinburg, NC
Ottawa, IL
Rossford, OH
Toledo, OH
Total est. capacity
Ford Motor Co., Glass Div.
Dearborn, MI
Nashville, TN
Tulsa, OK
Total est. capacity
ASG Industries, Inc.
Greenland, TN
Jeanette, PA
Kingsport, TN
Okmulgee, OK
Total est. capacity
Type and estimated
plant capacity,9
metric tons/day
Float

900
400
360

730


900
3,290


400
680
360
900
400
2,740

360
1,360
900
2,620

400



400
Plate
























320



320
Sheet




360

600
360

1,320

360





360







245

160,
.405
Rolled


*























295

295
Data on plant capacities were estimated based on communica-
tion with flat glass manufacturers; data on ASG, Fourco and
CE Glass are contained in References 6 and, 7.
                             11

-------
 Table 5  (continued).  PLANT LISTING FOR FLAT GLASS INDUSTRY
Company and location
CE Glass, Inc.
Cinnaminson, NJ
Erwin , TN
Floreffe, PA
Fuller ton, CA
St. Louis, MO
Total est. capacity
Fourco Glass Co.
Clarksburg, WV
Fort Smith, AR
Total est. capacity
Guardian Industries
Carleton, MI
Total est. capacity
TOTAL VALUES
Capacity, metric tons
Percentage
Number of plants
Number of melting furnaces
Type and estimated
plant capacity,9
metric tons/day
Float
450
360
810



820
820

10,680
77
17
25
Plate







320
2
1
1
Sheet


180
200
380




2,460
17
8
14
Rolled
64
64
177
305






600
4
4
9
TOTAL CAPACITY
  Average plant capacity:
  Average float plant capacity:
  Average sheet plant capacity:
  Average rolled plant capacity:
14,060 metric tons/day
   500 metric tons/day
   600 metric tons/day
   300 metric tons/day
   150 metric tons/day
aData on plant capacities were estimated based on communica-
 tions with flat glass manufacturers; data on ASG, Fourco
 and CE Glass are contained in References 6 and 7.
                              12

-------
polishing to produce a smooth surface.  It is used for
automobile windows and large picture windows.  Average
thicknesses range from 3.2 mm to 6.4 mm.

Sheet glass is made by drawing molten glass upward from the
melt.  It is thinner than float glass (1.6 mm 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 U.S.

Table 5 also lists the capacity of all 29 flat glass plants,
and gives total capacities for each flat glass product.
Average plant capacity is 500 metric tons/day.  The product
breakdown is:  float glass, 77%; plate glass, 2%; sheet
glass, 17%; and rolled glass, 4%.

The distribution of plants across the country is shown in
Figure 1.  A detailed listing by state,  county and Air
Quality Control Region (AQCR) appears in Appendix B.  The
average county population density of the counties containing
plants is 248 persons/km2.

B.   TYPES OF GLASS

Glass is normally defined as an inorganic product of fusion
which has cooled to a rigid condition without crystallizing.
The chemical composition and corresponding properties may
                              13

-------
Figure 1.  Locations of flat glass plants

-------
vary over a wide range.  Over 90% of the glass made, and
100% of the flat glass produced, is called soda-lime glass.
The name is derived from its basic ingredients of sand  (Si02),
soda ash (Na2CC>3) , and limestone (basically CaCO$ plus some
MgCOs).  Other types of glass used for special purposes are
borosilicate glass (heat resistance), lead glass (crystal
artware, TV tubes) and opal glass  (tableware).  The composi-
tions of these commercial glasses are given in Table 6.8~10

Glass raw materials, indigenous to many parts of the country,
are inexpensive.  Glass sand is the source of Si02 and it
should be of high purity (>99% Si02).  The impurities in
glass sand are A1203 and Fe203.  Although iron is an objec-
tionable impurity in other types of glass because of its
greenish tinge, it is often added in small amounts  (0.1%) to
flat glass (e.g., automobile window glass).

Soda ash is 98% to 100% Na2C03, with NaCl as the major
impurity in soda ash made by the Solvay process.9/10
Naturally derived soda ash contains ^0.025% NaCl.

Limestone is available as high-calcium limestone consisting
essentially of calcite, CaC03  (95%), and dolomitic limestone
which is a mixture of dolomite  (CaCOs-MgCOs) and calcite.
8Dietz, E. D.  Glass.  In:  Chemical and Process Technology
 Encyclopedia, Considine, D. M.   (ed.).  New York, McGraw-
 Hill Book Co., 1974.  p. 552-561.
9Hutchins, J. R., and R. V. Harrington.  Glass.  In:
 Kirk-Othmer Encyclopedia of Chemical Technology, 2nd
 Edition, Vol. 10, Standen, A.  (ed.).  New York, Interscience
 Publishers, Divn. of John Wiley & Sons, Inc., 1966.
 p. 533-604.
10Shreve, R. N.  Chemical Process Industries, 3rd Edition.
  New York, McGraw-Hill Book Co., 1967.  p. 190-210
                              15

-------
     Table 6.   COMPOSITIONS BY WEIGHT PERCENT OF -COMMERCIAL GLASSES8"10
Component
Si02
A1203
B203
Na20
K20
CaO
MgO
BaO
PbO
Soda-lime glasses
Containers
70 to 74
1.5 to 2.5
0
13 to 16
10 to 14
0
0
Flat glass
71 to 74
0 to 2
0
12 to 15
8 to 12
0
0
Tableware
71 to 74
0.5 to 2
0
13 to 15
5.5 to 7.5
4.0 to 6.5
0
0
Borosilicate
glass3
70 to 82
2 to 7.5
9 to 14
3 to 8
0.1 to 1.2
0 to 2.5
0
Lead
glass9
35 to 70
0.5 to 2.0
0
4 to 8
5 to 10
0
0
12 to 60
Specialty glassware.

-------
Quality limestones contain less than 0.1% Fe203 and about
1% silica arid alumina.9

Gullet, which is scrap glass that is to be recycled, makes
up from 30% to 50% of the total charge to a flat glass
melting furnace.   Use of cullet facilitates higher melting
rates and utilizes a waste material.9»1°
In addition to these major ingredients, feldspar
(R2O-A1203 • 6SiC>2 where R = Na or K) is added to sheet glass
as a source of aluminum, while salt cake  (Na2SOit) is added
as a fluxing agent.  Fluxes promote the melting process by
reacting with silica to form lower melting silicates.  A
typical glass batch recipe is:11

          Silica sand           910 kg     (55.6%)
          Soda ash              302 kg     (18.5%)
          Feldspar5              80 kg     ( 4.9%)
          Limestone             309 kg     (18.9%)
          Salt cake (Na2SOi,)     36 kg     ( 2.1%)
               TOTAL           1637 kg     (100.0%)

These ingredients melt down to 1,341 kg of glass  and give
off 295 kg of gases, primarily (>90%) C02.  The batch volume
of 1.27 m3 produces 0.57 m3 of fluid glass and 708 m3 of
gaseous products  (measured at the furnace temperature of
1,500°C).11
 Based on information from two flat glass manufacturers.
 Sheet glass only; not used in float glass.

^Holscher, H. H.  The Glass Primer.  New York, Magazines
  for Industry, Inc., 1972.  58 p.

                              17

-------
Although many minor ingredients  (<5% 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% of total produc-
tion) of colored float glass is made, but this has not been
considered in this document.

The only major variation in the production of other types of
flat glass is that iron oxide is not added.  One sheet glass
manufacturer still employs arsenic as a fining agent.
Fining is the process of removing gas bubbles from the melted
glass.  Fining agents react chemically in the melt and re-
lease gases that cause existing bubbles to increase in size
and rise to the surface.9  However, over half of all sheet
glass (and at least 90% of all flat glass) is made without
the use of arsenic,  and emissions of arsenic have not been
studied for this report.  The only compound that could be
considered to act as a fining agent, that is still used in flat
glass production, is sodium sulfate.

C.   THE GLASS MANUFACTURING PROCESS

The manufacture of flat glass can be broken down into three
basic steps:

   • Preparation and handling of raw materials (includes
     mixing)
   • Glass melting and refining
   • Forming and finishing operations  (includes annealing)
aInformation supplied by.two flat glass manufacturers,
                              18

-------
A flow diagram of the overall process is shown in Figure 2.

1.   Preparation of Raw Materials
                                   i

A typical plant  (500 metric tons/day) manufacturing float
glass houses the raw material mixing and conveying equipment
in a structure termed a "batch plant."  Figure 3 is a flow
diagram of a typical batch plant.  The storage bins for
major raw materials are elevated with the weigh hoppers and
mixers located below them to make use of gravity flow.

Sand, soda ash, lime, 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.  Cullet from within the plant travels by belt conveyors
to the cullet bin.  Powdered coal and iron oxide are stored
in small bins.

Ingredients comprising a batch of glass are dropped by
gravity from the storage bins into weigh hoppers and then
released to fall into the mixer.  Cullet is ground and then
mixed with the dry ingredients in the mixer.  Ground cullet
may also bypass the mixer and be mixed instead with the
other blended materials in the bottom of a bucket elevator;
unground cullet may be fed directly to the furnace.

Raw materials are blended in large (4 metric tons capacity)
mixers.   After 3 to 5 minutes, the mix is conveyed to a
charge bin located alongside the melting furnace.  At the
bottom of the charge bins, rotary valves feed the blended
materials into reciprocating- or screw-type furnace feeders
which force the blended raw materials into one end of the
glass melting furnace.
                              19

-------
N)
o
            RAW MATERIAL

             SAND (Si02)
          SODA ASH (N32C03)
            LIMESTONE OR
          DOLOMITE (CaO/MgO)
             ADDITIVES
            BATCH MIXER
 CRUSHED
  GULLET
 STORAGE
MELTING FURNACE
                                         REGENERATOR
  MOLTEN GLASS
    @1,500°C
                                         REGENERATOR
 GULLET
CRUSHING
      FORMING
                                                                     FLOAT BATH
      DRAWING
   (SHEET GLASS)
                                ROLLING
                              (ROLLED AND
                              PLATE GLASS)
                       PACKAGING,
                        SHIPPING
                                                                                        FINISHING,
                                                                                        INSPECTION
ANNEALING
                          Figure 2.   Flow diagram  of flat  glass  manufacturing

-------
             CULLET
       RAW MATERIALS
       RECEIVING
       HOPPER
            V
               SCREW
               CONVEYOR
                                                   FILTER
                                                   VENTS
STORAGE BINS
MAJOR RAW MATERIALS
                                                 MINOR
                                                 INGREDIENT
                                                 STORAGE
                                                 BINS
                   BELT CONVEYOR
                                                                         (I
BATCH
STORAGE
BIN
                                                                      r
                                                                       FURNACE
                                                                       FEEDER
                                                     GLASS-
                                                     MELTING
                                                     FURNACE
                   Figure  3.   Process  flow diagram  of a  batch plant
                                                                            12
12Danielson,  J. A.   Air  Pollution Engineering Manual,  2nd Ed.   Environmental
  Protection  Agency.  Research  Triangle Park.  Publication No.  AP-40.  May 1973,
  p.  765-782.

-------
The various handling and mixing operations are a source of
particulate emissions which are similar  (same materials,
same processes) to those in other industries.12'13  Because
of environmental and economic incentives, large manufacturers
practice dust control by means of cloth filters and bag-
houses.  Only limited data are available on particulate
emissions from the preparation of raw materials, but they do
indicate the generally low level of emissions (<0.05 g/kg).14

2.   The Glass Melting Furnace

In the glass furnace, the various raw materials are melted
together to form a homogeneous viscous liquid.  A typical
float glass melting tank is 46 m to 61 m long, 7 m to 9 m
wide, and holds molten glass 1.2 m to 1.5 m deep (Figure 4).

It contains 1,100 metric tons of molten glass and has a
capacity of 400 metric tons per day.  Sheet glass tanks
average about half this capacity.9'15"21
 13A Screening Study to Develop Background Information to
  Determine the Significance of Glass Manufacturing.
  Prepared by The Research Triangle Park Institute for the
  Environmental Protection Agency.  Research Triangle Park.
  Contract 68-02-0607, Task 3.  December 1972.

 ll+Point Source Listing for Glass, SCC 3-05-014, National
  Emission Data System.  Environmental Protection Agency.
  Research Triangle Park.  May 1974.

 15Svec, J. J.  LOF Operates World's Largest Glass Furnace.
  Ceramic Industry.  103 (2) -.30-32, August 1974.

 16Svec, J. J.  Double Float Glass Line Produces 300 Million
  Square Feet.  Ceramic Industry.  100:66-69, April 1973.

 17Ford Motor Controls Glass Batch by Chemical Wetting.
  Ceramic Industry.  102;28-30, March 1974.

 18Svec, J. J.  Float Plant a Showcase at Pilkington.
  Ceramic Industry.  101(6):34-36, December 1973.

 19Allen, A.C. New Canadian Plant Draws 14 Miles'of Sheet Glass
  Per Day. Ceramic Industry.  9^(6):52-54, December 1968.

 20Allen, A. C.  Canada Builds First Float Glass Plant.
  Ceramic Industry.  §9J6) : 43-45, December 1967.
 21Allen, A. C.  One of the World's Largest Glass Tanks on
  Stream.  Ceramic Industry.  8£(4):50-51, October 1967.

                              22

-------
U)
                           RAW MATERIAL
                               FEED
 GLASS
SURFACE
INMELTER
                                                   MELTER
                                                 SIDE WALL
                                                           MELTER BOTTOM
                                 COMBUSTION
                                 AIR BLOWER
                                                                                      PORT
                                                           BURNER
                      Figure 4.   Regenerative side  port glass-melting  furnace'
       After  Reference  14.

-------
 In contrast to container glass furnaces, there is no throat
 (a refractory barrier with a submerged opening) dividing a
 flat glass tank into melting and conditioning sections.
 Rather, additional tank length (container glass tanks are
 only 15 m long) is needed to melt the glass.9  Large melting
 tanks are also necessary in the float process so that the
 large quantity of molten glass can provide part of,the
 heating energy for the tin float bath.22  Sheet glass tanks
 may employ a fire-clay floater or water-cooled skimmer to
 separate the melting and working areas in the tank.

f Raw materials are fed into one end of the furnace and layered
 on top of the molten glass.  Temperatures in the melting sec-
 tion of the tank average 1,500°C.  As the molten glass flows
 toward the other end of the tank, sand grains melt completely,
 gas bubbles rise to the surface,  and convection currents
 produce a homogeneous melt.  The temperature within the fur-
 nace gradually falls to 1,100°C as the glass leaves the tank.

 Flat glass furnaces in the U.S. have been traditionally heated
 with natural gas, but gas shortages have led to an increased
 usage of oil.  Gas is preferred because it causes less wear
 on the furnace and it is low in sulfur.  Producers in the
 eastern part of the country may operate on oil for several
 months of the year.  Electrical boosting is not used as a
 source of furnace heat.  Although no data are available on the
 exact usage ratio of gas to oil,  a 1973 breakdown of total
 energy used by the flat glass industry was as follows:23
 22Svec, J. J.  Pilkington Manufacturers 2.3 mm Float Glass.
   Ceramic Industry.  103 (1):36-37, July 1974.
 23Schorr, J. R., and G. A, Anderson.  Final Report on
   Industrial Energy Study of the Glass Industry.  Prepared
   by Battelle for the Federal Energy Administration and the
   Department of Commerce.  Washington.  Contract
   14-01-0001-1667.  December 1974.  p. 13-16,36.

                               24

-------
          Energy Form     Percent of Total Energy Used
          Fuel oil                     6.5.
          Natural gas                 77.5
          Electricity                 14.9
          Coal                         1.1

After subtraction of the electricity and coal (which are not
used to heat the furnace), the gas/oil ratio is 92% to 8%.
Since 75% of the total energy is consumed by the melting fur-
nace, this ratio should be close to the actual amount used in
melting.  One manufacturer reported using oil from 10% to 15%
of the time.

The efficiency of glass furnaces is between 10% and 30%, with
newer designs being more efficient.  The theoretical heat for
melting is ^1.85 MJ/kg.

Burners are placed on the sides of the tank furnace above the
molten glass surface, and the burner flames are directed across
the top of the melten glass.  This side port configuration is
used exclusively in flat glass furnaces since placing the bur-
ners at the feed end (end port design) is not feasible in a
large long melting tank.  To conserve fuel the typical furnace
uses a regenerative firing system for heat recovery.  Such a
system consists of dual chambers filled with brick checkerwork
on each side of the tank.  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
the chambers are interchanged during the reverse cycle.  Re-
versals occur every 15 to 20 minutes as required for maximum
conservation of heat.9"12

Emissions from the glass melting furnace include NO , SO ,
                                                   X    X
particulates, CO and hydrocarbons.  Nitrogen oxides are
formed by the reaction of atmospheric nitrogen at the high
                              25

-------
temperature conditions of the furnace.  Sulfur oxides come
from the volatilization 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 in two ways:
(1) by the physical entrainment of dust by combustion gases
when a batch is added to the furnace, and  (2) by volatiliza-
tion of materials in the melt which subsequently condense in
the checkers or stack.12'24'25  Carbon monoxide and hydro-
carbons arise from incomplete fuel combustion and from coal
in the glass batch.

3.   Flat Glass Forming and Finishing Operations

Forming and finishing operations within the glass industry
are very diverse and depend on the type of product being made,
Flat glass products may be divided into four categories:
float glass, sheet glass, rolled glass and plate glass.  All
are made by continuous processes from the initial drawing to
cutting.  The manufacturing steps used in the production of
flat glass are outlined below.  Figure 5 gives an overall
flow diagram of the forming and finishing operations.

a.   Float Process - The float glass process is a radical
departure from all previous flat glass forming operations.
The glass from the melting furnace may be rough rolled, pour
drawn, or formed through a nozzle directly to a combined
forming and finishing step.  In this step the glass enters a
sealed chamber containing a float bath of molten tin which
is maintained under a neutral nitrogen atmosphere (Figure 6).
24Ryder, R. J., and J. J. McMackin.  Some Factors Affecting
  Stack Emissions from a Glass Container Furnace.  The Glass
  Industry.  50:307-310, June 1969; 346-350, July 1969.
25Arrandale, R. S.  Pollution Control in Fuel-Fired Tanks.
  The Glass Industry.  55_: 12-13, 21, August 1974; 16-17,27,
  September 1974.
                              26

-------
                                                      FLOAT GLASS
to
-o
          MELTING
          FURNACE
                             FLOAT PROCESS
                                TIN BATH
                 ANNEALING
                    LEHR
                  CUTTING
               PACKAGING
                                                       SHEET GLASS
  DRAWING
 MACHINES

• FOURCAULT
• PENNVERNON
• COLBURN
ANNEALING
  LEHR
CUTTING
PACKAGING
                                                      ROLLED GLASS
                            ROLLING MACHINE
                                  WITH
                             PATTERNED ROLLS
                 ANNEALING
                    LEHR
                  CUTTING
               PACKAGING
                                                      PLATE GLASS
                ROLLING
                MACHINE
 ANNEALING
   LEHR
 CUTTING
GRINDING
                PACKAGING
                        Figure 5.   Flow  diagram of forming and finishing operations

-------
                                                        ANNEALING LEHR
00
                                                   o   o
oo
                           MOLTEN METAL FLOAT BATH
              Figure 6.  Flat glass manufacturing:  Pilkington float process9
              Reprinted by permissions of The Glass Industry and Pilkington Brothers, Ltd.

-------
The glass flows onto and is drawn across the tin at a pre-
determined rate.  The bath is temperature controlled in a
way that permits the glass to flow and form a perfectly flat
surface and then to harden before entering the annealing
lehr.9,15,16,18

The tin bath is 60 m long and 64 mm to 76 mm deep. The
temperature in the bath varies from 1,100°C at the hot end
to 600°C at the cool end.15'18

The float process is very flexible and offers a number of
outstanding advantages.  It provides a finely finished sur-
face to the glass without grinding and polishing.  The lower
surface of the glass is finished by the upper surface of the
liquid metal while the upper surface of the glass is formed
either by gravity or by a pool of molten metal located on the
upper surface of the glass.  The position of this pool remains
constant while the sheet of glass passes underneath it.  The
pool of molten material is maintained in place by ridges on
the edges of the glass sheet and by the speed of the glass
through the chamber.

As the glass sheet leaves the float bath the lower surface
may be sprayed with SC>2 to develop a protective coating of
sodium sulfate.  This prevents rollers in the annealing lehr
from marring the smooth finish on the glass.26

Annealing is the process of preventing or removing objec-
tionable stresses in glassware which result from too rapid
cooling.  It is carried out in long annealing lehrs which
26Development Document for Effluent Limitations Guidelines
  and New Source Performance Standards for the Flat Glass
  Segment of the Glass Manufacturing Point Source Category.
  U.S. Environmental Protection Agency.  Washington.
  EPA-44011-74-001-C.  January 1974.  p. 44.

                              29

-------
are precisely temperature controlled to give a predetermined
cooling schedule.  The lehr may be up to 150 m long and is
heated with gas or electricity.  Older lehrs are gas fired
while some newer units use electricity for better temperature
control.9'16'27'28  Most float glass lehrs are heated elec-
trically.

The glass sheet passes through the annealing lehr and is
then cut, inspected, and stored for shipment.

b.   Sheet Glass - Sheet glass is produced by continuous
drawing operations of which there are three types.

(1)  Fourcault process - The Fourcault process (Figure 7) is
a vertical draw process as are the two discussed below.  The
major feature of this drawing process is the debiteuse, a
rectangular refractory collar which is partially submerged in
the glass melt.  In the collar there is a slot 100 mm to
200 mm wide and up to 2.4 m long.  Molten glass is drawn up-
wards from the slot by powered rollers located in a vertical
annealing lehr 7.6 m high.  The speed at which the sheet is
drawn and the width of the slot in the debiteuse determine
the thickness of the glass sheet.  The annealing lehr provides
controlled cooling of the sheet to prevent cracking, stressing,
or visual distortion of the final product.8"11

After the sheet emerges from the top of the annealing lehr it
                                   j
is cut to size, inspected for flaws, and stacked for shipment.
27Fuller, R. A.  Recirculating Lehr for Annealing Glassware.
  Ceramic Bulletin.  4_8_: 1065-1068, November 1969.
28Roos, P. W.  Lehr Priority:  Design Concepts to Save Energy,
  The Glass Industry.  56_: 18-22, April 1975.

                              30

-------
 DRAW ING MACHINE AND
VERTICAL ANNEALING LEHR
        ROLLERS
     SHEET COOLERS
                           CO
                                      DEPRESSED
                                      DEBITEUSE


  Figure 7.  Flat glass manufacturing:  Fourcault process9

       Reprinted by permission of The Glass Industry.

                        31

-------
 (2) PPG Pennvernon process - The only difference between this
process and the Fourcault process is the method by which the
glass leaves the surface of the melt.  In place of a debiteuse
this process uses a totally submerged refractory drawbar
 (Figure 8).  The draw is made from the melt surface directly
over the bar and then through rollers in a vertical annealing
lehr.  The drawbar partially conditions the glass, determines
the point of origin of the glass sheet, and reduces irregular-
ities in the product resulting from convection currents in
the melt.8-11

 (3)  LOF-Colburn process - This process differs from the
others described above in two ways.  One difference consists
of the method by which the sheet is drawn from the melt.
Instead of a debiteuse or drawbar to control the size of the
draw, rollers pull the glass directly from the melt (Figure
9).  It is therefore necessary to place cooling edge rollers
close to the surface of the melt to help control the thickness
of the sheet and to prevent necking down the sheet.  The
second difference between this process and the two described
above is in the use of a horizontal annealing lehr which re-
quires that the sheet of glass be bent over rollers 1 m above
the surface of the melt.  This is the last possible moment at
which the glass can be bent without damage.  From this point
the sheet of glass enters the annealing lehr.8"11

The above three processes all have the same annealing,
cutting, inspection, and storage steps.

c.   Rolled Glass - In this process, glass is taken directly
from the melting furnace by a set of rollers which form the
glass into a sheet (Figure 10).   From these primary rollers
the glass sheet is taken into an annealing oven before it
becomes inflexible.  The design for patterned or frosted win-
dow glass is imprinted on one or both of the rollers drawing

                              32

-------
                                 DRAWING MACHINE AND
                                VERTICAL ANNEALING LEHR
                 CO
                                       ROLLERS
SHEET COOLERS
                       \  SUBMERGED
                           DRAW BAR
  Figure 8.  Flat glass manufacturing:  PPG Pennvernon process'


Reprinted by permissions of The Glass Industry and PPG Industries, Inc.

                           33

-------
                                                      MACHINE ROLLS
     \/
>£>
                                 BENDING ROLL
                        COOLERS
                                       / / / / / / /I / / / / X / / //////
                                             10  O  GTO  (J
                Figure 9.   Flat glass manufacturing:   LOF-Colburn process
          Reprinted by permissions "of Van Nostrand Reinhold Co- and Libbey-Owens-Ford Glass Co.

-------
OJ
Ul
        / / / / / XXX
                       /
                       /
                       /
                       /
                       /
                       /
                       /
ROLLING
MACHINE
 ANNEALING
   LEHR


TRAY ROLLS
                     11 n A n C* I'
                 TRAY "BARS
                         / / /// X / /XXI
          XXXXXXX
                                            / / / / x\x /
                                                         O   (D   O  OOOO
                                                                                       / 7
                     Figure 10.   Flat glass manufacturing:   rolling process'
                             Reprinted by permission of The Glass Industry.

-------
the glass from the furnace.  Wired safety glass is formed by
feeding a roll of wire mesh into the glass in front of the
primary rollers.  After annealing, the glass is cut and
packaged for shipment.8"11

d.   Plate Glass - Plate glass is manufactured by the
rolling process using smooth rollers.  Because the rollers
introduce surface distortions, plate glass is ground and
polished after annealing to produce a smooth surface.

e.   Emission Points - In general, the forming and finishing
operations in flat glass manufacture do not cause emissions
directly to the atmosphere.  The processes take place within
the plant building and any emissions are released inside the
building.  Combustion products from the annealing lehr are
an exception.
                              36

-------
                         SECTION IV
                          EMISSIONS

A.   SELECTED POLLUTANTS

The manufacture of glass gives rise to air emissions during
each phase of production.

   • The preparation of raw materials is a source of
     particulate emissions.
   • The glass melting furnace emits NO ,  SO ,  particulates,
     CO and hydrocarbons.
   • The only emissions in forming and finishing operations
     are combustion products from gas-fired annealing lehrs.

1.   Raw Materials Preparation

Points of emission in the preparation of raw materials are
identified in Figure 7.  These have been classified into five
general groups:

   • Handling of raw materials (unloading and conveying)
   • Crushing of scrap glass
   • Filling and emptying of storage bins
   • Batch mixing (and weighing)
   • Feeding of mixed glass batch to melting furnace
     (batch charger)

Although all of these operations are potential sources of
particulate emissions, at least 90% of .the industry has
installed dust controls.  Handling operations are enclosed
                              37

-------
       GULLET
                GLASS
               CRUSHER
                   CONVEYING
to
oo
   SAND  -
 LIMESTONE-

 SODA ASH -
  MINOR  __
INGREDIENTS
                                    STORAGE
                                     BINS
                    STORAGE
                      BINS
                   UNLOADING
                                       CONVEYING
   i
WEIGHING
                                        CONVEYING
                                         MIXING
                                                      CONVEYING
   1
                                                                               CONVEYING
 BATCH
CHARGER
                                                                                  I
                                                                           MELTING
                                                                           FURNACE
             Figure 11.  Points of emission during the preparation of raw materials
               (All of these steps are potential sources of particulate emissions.)

-------
to eliminate fugitive emissions and vents on storage bins and
mixers are exhausted through baghouses.  In addition, the
particle size of the raw materials is chosen to minimize
dusting, and water is added to the mixed batch.  Because of
this, there are few data available on particulate emissions
from stacks and no data on fugitive emissions.

Source test data that have been reported in NEDS (National
Emission Data System) are given in Appendix C.  The resulting
average emission factors are given in Table 7 along with
total industry emissions.  The zero emission factors for
handling and batch charging were determined during a personal
visit to a large float glass plant.  Handling operations were
either completely enclosed or performed within the plant
building, and no fugitive dust emissions were visible.  The
batch charger was partially enclosed and situated inside the
plant.  No dust could be seen in the charging area  (the
mixed batch had been moistened to alleviate dusting and to
prevent the segregation of raw materials in the batch).

The overall emission factor is 0.02 g/kg ±  100% and  total
particulate emissions are 90 metric tons ±  100%.  The composi-
tion of the particulate emissions will be that of the raw
materials (basically sand, lime, and soda ash) since no
chemical reactions take place.  As discussed in Section V on
control technology, glass sand is processed to remove <44-ym
(minus 325-mesh) material.  Consequently, there will be no
respirable quartz in the emissions.

2.   Glass Melting Furnace

The different species emitted from the glass melting furnace
are listed in Table 8 along with their average emission fac-
tors and total yearly emissions.  Emission  factors are based
on test results from container glass furnaces and pressed and
                              39

-------
Table 7.  PARTICULATE EMISSIONS DURING THE PREPARATION OF RAW MATERIALS
Process step
Handling (unloading,
conveying)
Glass crusher
Storage bins
Mixing (and weighing)
Batch charger
TOTAL (Independent rounding)
Emission factor,
g/kg
0
0.0123- ± 100%
0.0123 ± 100%
0.00158 ± 106%
0
0.02 ± 100%
Material processed,
106 metric tons
4.5
1.8
4.5
4.5
4.5
4.5
Total emissions,
metric tons
0
22 ± 22
55 ± 55
7 ± 7
0
90 ± 90

-------
          Table 8.  EMISSIONS FROM THE GLASS MELTING FURNACE
Species
NO
X
SO
X
Particulates
CO
Hydrocarbons
Emission factor,
g/kg
4 ± 30%
1.5 ± 27%
1 ± 60%
0.02 ± 100%
0.04 ± 100%
Glass produced,
10^ metric tons
4.0
4.0
4.0
4.0
4.0
Total emissions,
10 ^ metric tons
16.0 ± 4.8
6.0 ± 1.62
4.0 ± 2.4
0.080 ± 0.080
0.160 ± 0.160
blown glass furnaces, in addition to  flat glass  furnaces,
since the melting process is the same for all  soda-lime
glasses  (i.e., same melt temperatures and major  raw materials)
However, other emissions which may arise from  minor ingredi-
ents in the melt  (e.g., fluoride or borate)  are  not present
in stack gases from flat glass furnaces.  While  chlorine  is
present as a batch impurity  (^0.06 g/kg as NaCl),  it  has  not
been detected in the furnace exhaust.

The parameters that affect emissions  are considered in more
detail in Section V under Process Modifications.   In  general,
the furnace temperature and the batch raw materials are the
two key process variables.  Oil vs. gas firing is  also
important when oil heating is practiced  (^10%  of the  time).
Secondary factors relate to overall furnace  efficiency and
may include combustion conditions, age of furnace  and
checkerwork, type of refractories used in the  furnace, and
checkerwork design.

a.   Nitrogen Oxides - Nitrogen oxides represent the  largest
fraction by mass  (^61%) of glass furnace emissions.   They
are formed by the combination of atmospheric nitrogen and
oxygen at the high temperatures ( 1,500°C) within  the
melting furnace.  The reaction is very temperature sensitive,
                              41

-------
as indicated in one study where the NO  concentration  in the
                                      X
stack gas increased sixfold  (from 100 ppm to ^600 ppm) when
the production rate was doubled (from 90 to 180 metric tons/
day) and the furnace temperature increased from 1,460°C ,to
l,551°C.2lt'29

Source test measurements reported in NEDS give an average
emission factor of 4 g/kg, with individual values ranging
from 0.71 to 10.5 g/kg.ltf  The average is based on 27 values
representing ^9% of total U.S. glass production (see
Appendix C.2).  The average is accurate within ±1.2 g/kg at
a 95% confidence level.

One large flat glass manufacturer reported that furnaces are
presently operating at higher efficiencies than in the past.
Since the test data in Appendix A are primarily from
1970 - 1972, they may overestimate the average NO  emission
                                                 .A.
factor.

b.   Sulfur Oxides - Sulfur oxide emissions result from the
decomposition of sulfates in the melt and from the oxidation
of sulfur in the fuel.  Consequently, the emission factor
will depend on the sulfur content of the feed material and
fuel, and the furnace temperature.  Sodium sulfate is used
as a fluxing agent in the glass batch (~2% by weight) and
decomposes above ^1,000°C to Na20 and S03.  Above 1,200°C/
803 is unstable with respect to S02 and 02-  Not all the
sulfur oxides are emitted as such; some (^25%) combine with
29Control Techniques for Nitrogen Oxide Emissions from
  Stationary Sources.  U.S. Department of Health, Education
  and Welfare.  Washington.  NAPCA Publication No. AP-67
  (PB 190265).  March 1970.  115 p.
                              42

-------
sodium vapor [presumably in the form of sodium hydroxide
dimer, Na2(OH)2] to form Na2S(\ . 30

If sulfur is present in the fuel it will oxidize and appear
as SO  in the exhaust gas.  A fuel oil containing 1% sulfur
     J\,
by weight will give ^600 ppm S02 in the flue gas.31  The
emission factor can be calculated on the basis of a heating
value of 41.9 GJ/m3 of fuel oil and a 30% furnace efficiency
(About 1.85 MJ are needed to make a kilogram of glass.)9
The ^100 liters of fuel oil needed to melt a metric ton of
glass contain ^1.2 kg of sulfur.  If all the sulfur were
oxidized to S02/ the emission factor would be ^2.6 g/kg.  It
will actually be less than this because of conversion to
Source test measurements for SO  emissions range from 0.2 to
                               H
4.4 g/kg.  The average emission factor is 1.5 g/kg, based on
NEDS data from 12% of total U.S. production.  This value is
accurate to ±0.4 g/kg at a 95% confidence level.  Detailed
data are presented in Appendix C.2.  As previously mentioned,
the present day average emission factor may be lower than
1.5 g/kg because of improved furnace operation.

c.   Particulates - Particulate emissions in glass furnaces
     -                            ?                 {
result from (1) the physical entrainment of dust from the feed
material in hot combustion gases, and (2) the volatilization^
of sulfates in the melt which later condense as they leave
the furnace.  Data show that large sized particles (>1 ym)
30Davis, R. E., W. H. Manring, and W. C. Bauer.  Carryover
  Studies in Glass Furnaces.  In:  Collected Papers from
  the 34th Annual Conference on Glass Problems.  Dept. of
  Ceramic Engineering, University of Illinois, Urbana,
  November 1973.  p. 109-126.
31Reed, R. H.  Combustion Pollution in the Glass Industry.
  The Glass Industry.  5_4_:24,26, 38, April 1973.
                              43

-------
generated by the first mechanism are trapped in the regener-

ative checkerwork and do not go out the stack,12r2^> 25 as

evidenced by:


   • Analysis of dust collected in the checkers which shows
     that it resembles the feed material (high silica, e.g.).
     Dust going out the stack is primarily alkali
     sulfate.12/25,32

   • Dust trapped in the checkers is micron size, but dust
     leaving the stack is submicron size.12/33

   • The emission rate increases as the furnace temperature
     increases.  (The logarithm of the emission factor
     decreases in direct proportion to the reciprocal of
     the absolute temperature.)2 4,3 ^
Submicron particulates are formed from condensed vapors from

the glass melt in accordance with the second mechanism

indicated.13,25/30,35


Source test measurements in NEDS of particulate emissions

factors vary from 0.22 to 8.3 g/kg (see Appendix C.2), with

an average of 1 g/kg based on ^19% of total U.S. production.

This is identical to the emission factor given in the
32Mills, H. N., and J. Jasinski.  Evaluating Batch Changes.
  The Glass Industry.  5^:223-227, May 1970.

33Stockham, J. D.  The Composition of Glass Furnace
  Emissions.  Journal of the Air Pollution Control
  Association.  2_1:713-715, November 1971.

34Arrandale, R. S.  Air Pollution Control in Glass Melting.
  Symposium Sur La Fusion du Verre, Brussels.  October 1968.
  p. 619-644.                                    ,

35Custer, W. W.  Electrostatic Cleaning of Emissions from
  Lead, Borosilicate, and Soda-Lime Glass Furnaces.  United
  McGill Corp.  (Presented at the 35th Annual Conference on
  Glass Problems.  Ohio State University, Columbus,
  November 14-15, 1974.)  13 p.


                              44

-------
Compilation of Air Pollution Emission Factors (AP-42).36
The value of 1 g/kg is accurate within ±0.6 g/kg at a 95%
confidence level.  Because of improved furnace efficiency,
the current average factor may be lower than 1 g/kg.

The composition of particulate emissions from soda-lime glass
furnaces is primarily (^80%) sodium sulfate. l2 / 24 / 25 > 30 » 3 3
The other elements found in the particulate vary greatly and
reflect variations in the feed composition.  Since flat glass
does not contain ingredients such as boron, fluoride, and
arsenic, these are absent in the particulates.  One analysis
of flat glass particulates showed 98% Na2SOit.

d.   Carbon Monoxide - Carbon monoxide is emitted from two
sources in the glass melting furnace:   (1) incomplete
combustion of fuel, and  (2) reaction of coal used as a
reducing agent in the glass batch.  Powdered coal is added
to the batch to reduce sulfates to sulfites; the carbon is
oxidized to CO or C02.  The amount varies from 5% to 10% by
weight of the sulfate content of the batch,37 or about
1.5 g coal/kg raw materials for a mix containing 2.2%
The average emission factor based on two sets of measurements
is 0.02 g/kg ± 100%  (Appendix C.2).  The accuracy  is esti-
mated because assumptions had to be made in the calculations.
For comparison another emission factor was calculated based
Information supplied by a flat glass manufacturer.
3Compilation of Air Pollution Emission Factors, Second
  Edition.  Environmental Protection Agency.  Washington,
  Publication No. AP-42.  April 1973.  p. 8.13-1.
371974 Annual Raw Material Processing Handbook.  Ceramic
  Industry.  10_2_:97, January 1974.
                              45

-------
on CO emissions data from gas-fired burners.  This value is
0.066 g/kg ± 55%.  Details are given in Appendix C.2.
Emissions of both CO and hydrocarbons are strongly dependent
on combustion conditions, and improper mixing of air and
fuel or mixing without enough oxygen can raise the emission
rate by a factor of 10 or more.

e.   Hydrocarbons - Hydrocarbons form in glass furnaces in
the same way that CO does:  by incomplete fuel combustion
and by decomposition of powdered coal in the melt.  The
emission factor based on one set of measurements is
0.04 g/kg.± 100%.(Appendix C.2).  The accuracy was determined
as it was for the CO factor.  For comparison the emission
factor based on data from gas-fired burners is
0.042 g/kg ± 144%.  Details appear in Appendix C.2.

3.   Forming and Finishing Operations

The only atmospheric emissions from flat glass forming and
finishing operations are combustion products from gas-fired
annealing lehrs.  A plant visit established that other
processes are not emission sources.  The machinery that is
used does not bring the hot glass into contact with lubrica-
ted surfaces, as is true in the production of glass
containers, so no lubricants are vaporized.  The float glass
tin bath is totally enclosed except for openings at either
end where the glass sheet enters and leaves.  No emissions
are visible from the openings, and a material balance on the
tin bath shows that possible losses to the atmosphere are
less than 0.001 g of tin per kg glass melted.  As the glass
sheet leaves the float bath, the hot surface is sprayed with
S02 which reacts to form a protective coating of Na2SOtt.
There is no detectable odor of S02 in the vicinity of this
operation and it is concluded that all the S02 reacts to form
the sulfate.
                              46             .

-------
Annealing lehrs provide controlled cooling of glass products.
Thus their only emissions are combustion products.  Since
these have never been measured, emission factors were esti-
mated from other data on gas combustion  (see Appendix C.3).
The results are given in Table 9.  Although the majority of
plants use electrically heated lehrs, the exact percentage
of use is unknown.  Total national emissions were therefore
calculated on a worst case condition of  50% gas firing.
         Table 9.  EMISSIONS PROM GAS-FIRED ANNEALING LEHRS
Emissions
NO
X
SO
X
Particulates
CO
Hydrocarbons
Emission factor,
g/kg
0.016 ± 113%
0
0.0012 ± 196%
0.0022 ± 55%
0.0014 ± 144%
Glass annealed,
106 metric tons
2.0
2.0
2.0
2.0
2.0
Total emissions,
metric tons
32 ± 36
0
2 ± 5
4 ± 2
3 ± 4
B.
EMISSION CHARACTERISTICS
1.   Raw Materials Preparation

Emissions of particulates are composed of sand, soda ash,
limestone, sodium sulfate, and glass dust.  The exact
composition has never been measured.  The materials that
tend to dust most easily are the sodium compounds and
limestone since they are soft and crush readily.  Glass sand
does not cause dusting and it has been processed to remove
<44-ym  (minus 325-mesh) particles.

The primary ambient air standard for particulates is
260 yg/m3.  Materials for which threshold limit values
                              47

-------
 (TLVs®) have  been  established  include  limestone  (10 mg/m3)
 and glass dust  (10 mg/m3).38
Sand and glass dust are  stable compounds while  soda ash and
limestone, being  slightly basic, may react with SO  in the
                                                  A.
atmosphere.

2.   Glass Melting Furnace

Emissions from the melting furnace consist of, the criteria'
pollutants:  NO , SO  , particulates, CO, and hydrocarbons.
               X    X
These materials interact in the formation of photochemical
smog while NO  and SO  are irritating to the lungs.  The
             X       X
ambient air standards and TLV's for these compounds are
given in Table 10.

A question has recently arisen concerning the possible
health effects from particulate sulfate emissions.  Since
glass furnace particulates are primarily (98%)  sodium
sulfate, they may pose some health hazard.  It  is unclear at
this time,  however, whether the health effects  stem from the
sulfate ion or the associated metal ion.  No TLV has been
established for sodium sulfate.

3.   Forming and Finishing Operations

The emissions from forming and finishing operations are
combustion products from gas-fired annealing lehrs.  These
are the same criteria pollutants discussed in Section 2 above,
38TLVs® Threshold Limit Values for Chemical Substances and
  Physical Agents in the Workroom Environment with Intended
  Changes for 1975.   American Conference of Governmental
  Industrial Hygienists.  Cincinnati.  1975.  97 p.
                              48

-------
                        Table 10.  CHARACTERISTICS OF EMISSIONS FROM GLASS MELTING FURNACE
Compound
NO
X
SO
X
Particulate


CO
Hydrocarbons
TLV,
mg/m3
9 (for NO2)
13 (for S02)
10


55
(1000 ppm
for methane)
Ambient air
quality standard39
mg/m3
0.100
0.365
0.260


40
0.160
Averaging time
Annual arithmetic
mean
24 hr; not to be
exceeded more than
once per year
24 hr; not to be
exceeded more than
once per year
1 hr; not to be
exceeded more than
once per year
3 hr; 6-9 AM
Health effects
Dangerous irritant
to lungs
Dangerous irritant
to lungs
. Unknown


High concentrations
(1000 ppm) are
asphyxiating
Simple asphyxiant
Atmospheric stability
Contributes to photo-
chemical smog
Contributes to photo-
chemical smog; also
forms acid mist
Stable


Stable
Contributes to photo-;
chemical smog
VO
      39Code of Federal Regulations, Title 42 - Public Health, Chapter IV - Environmental Protection
        Agency, Part 410 - National Primary and Secondary Ambient Air Quality Standards, April 28, 1971.
        16 p.

-------
c.
ENVIRONMENTAL EFFECTS
1.   Total Emissions

Total emissions from flat glass manufacturing are shown in
Table 11.  The melting furnace contributes the largest
amount of emissions  (99%), and NO  , SO  , and particulates
are the only significant emissions  (>103 metric tons).
                 Table 11.  TOTAL EMISSIONS
                       (metric tons/yr)
Source
Raw materials
preparation
Melting furnace
Annealing lehr
TOTAL
NO
X
0
16,000
32
16,000
S0x
0
6,000
0
6,000
Parti-
culates
90
4,000
2
4,000
CO
0
80
4
80
Hydro-
carbons
0
° 160
3
160
When total flat glass emissions are compared to total
national emissions from all stationary sources, they are all
less than 0.1%  (NO  = 0.07%, SO  = 0.02%, and parti-
                  X            X
culates = 0.02%).  On a statewide basis the largest contri-
butions are from NO  emissions in Tennessee (0.6%) and
Oklahoma (0.5%).  A detailed breakdown can be found in
Appendix D.     ,

2.   Ground Level Concentrations

As an aid to evaluating the environmental effects of stack
emissions, a plume dispersion equation was used to calculate
the maximum average ground level concentration, x"   , of
                                                 TH3.X
each emission species around a representative flat glass
                              50

-------
plant.  The representative plant has a production rate of
500 metric tons/day and emission rates as shown in Table 8.
The plant makes float glass in a side port regenerative furnace
and uses natural gas for heating.  Only furnace emissions were
considered since they account for over 99% of plant emissions.
Two average stack heights  (30 m and 60 m) were used in the cal-
culations since a tabulation of stack heights  (Table 12) showed
a bimodal distribution.  The taller stacks are natural draft
while the shorter ones use an ejection air system.  (Some of
the tall stacks have been converted from natural draft to
ejection air.)

     Table 12.  STACK HEIGHTS OF FLAT GLASS FURNACES,14
                           (meters)
Stacks under 40 m
18
23
24
32
32
32
32
32
34
34
34
34






Average =30
Stacks over 40 m
46
46
48
49
52
56
61
61
61
61
63
63
64
65
67
76
84
84
Average = &0
                              51

-------
The values of x    were computed from the equation suggested
               in 3.x
by Turner:k°
                    xmax
                        *•« •
where    Q = emission rate, g/s
         H = effective stack height, m
         TT = 3.14
         e = 2.72
         u = wind speed, m/s
           = 4.5 m/s  (national average)
      X    = "instantaneous" (i.e., 3-minute average)
             maximum ground level concentration
        tg = 3 minutes
         t = averaging time, minutes

The averaging time chosen for each emission was the same as
for the corresponding ambient air quality standard  (AAQS) .
The effective stack height is equal to the physical stack
height plus the plume rise  (see Appendix E for a determina-
tion of the plume rise correction) .  A source severity, S,
was then defined as the ratio of x"    to the standard:
                           S =
                               AAQS
Equations for S are given in Appendix F.
 ^Turner,  D.  B.   Workbook  of  Atmospheric  Dispersion
   Estimates,  1970 Revision.   U.S.  Department of Health,
   Education  and  Welfare.   Cincinnati.   Public Health
   Service  Publication  No.  999-AP-26.   May 1970.   84 p.
                              52

-------
Values of  X-.,,.,,  and S for flat glass emissions  are given in
           in 3.x

Table 13.  The  highest values are for NO  emissions (S = 1.3
                                         J^

for a 30 m stack).   The severity for both CO and hydrocarbons


is <0.01.






           Table  13.  SOURCE SEVERITY FOR FLAT GLASS EMISSIONS

Emission
NO
X
SO
X
Particulates
CO
Hydrocarbons
Ambient air
quality standard,
mg/m3
0.100
0.365
0.260
40.0
0.160
"Vax'
yg/m3
Stack
height =
30 m3
130
44
29
1.0
1.7
Stack
height =
60 m
57
20
13
0.45
0.74
S
Stack
height =
30 m
1.3
0.12
0.11
<0.01
0.01
Stack
height =
60 m
0.57
0.05
0.05
<0.01
<0.01
 Effective stack height is 30 m higher due to plume rise.
3.   Affected Population
Dispersion equations predict that the average  ground level


concentration  ("x)  varies with .the distance, x,  away from a


source.  For elevated sources, x is zero at the source,


increases to some  maximum value, x   , as x increases,  and
                                ' Amax'

then falls back  to zero as x approaches infinity.   Therefore


a plot of )(/AAQS vs.  x will have the following appearance:
AAQS
01
. 1

J
/\
Xl

x2 	 	
                               53

-------
The affected population is defined as the population around
a representative plant exposed to a X/AAQS ratio > 0.1 or
1.0.  The mathematical derivation of the affected population
can be found in Appendix F.  Results for flat glass furnace
emissions are given in Table 14.  The largest population
affected is for NO  emissions from a 30 m stack  (14,600)
                  J\,
persons) .,

               Table 14.  AFFECTED POPULATION



Emission
NO
X
SO
X
Particulates
CO
Hydrocarbons

N0x
SO
X
Particulates
CO
Hydrocarbons
X/AAQS > 0.1
Affected area, km2
Stack
height. =
30 mD
59.0
3.0
2.6
0
0
Stack
height.=
60 m°
55.8
0
0
0
0
Affected population
Stack
height, =
30 mb
14,600
750
640
0
0
Stack
height, =
60 mD
13,800
0
0
0
0
X/AAQS > 1.0
2.8
0

0
0
0
0
0

0
0
0
703
0

0
0
0
0
0

0
0
0
 Based on an average population density of 248 persons/km2.
 Effective stack height is 30 m higher due to plume rise.
                              54

-------
                          SECTION V
                     CONTROL TECHNOLOGY

Economic and environmental considerations have motivated the
glass industry to develop methods for reducing air emissions
from the manufacturing process.  Two general approaches have
been employed:  (1) the use of air pollution control equip-
ment, and (2) changes in the production process.  An example
of the first approach is the use of baghbuses to control
particulate emissions from raw material mixing operations.
The other method is typified by the reduction of volatile
materials such as sodium sulfate in the glass batch, resulting
in lower particulate emissions from the melting furnace.

A.   PREPARATION OF RAW MATERIALS

The handling and mixing of raw materials is a source of
particulate emissions.  The problem is similar to that in
other industries using granular or powdery materials, and
standard control techniques are available.12  These are
summarized in Table 15.

Railroad hopper cars and hopper bottom trucks can be
connected to sealed receiving hoppers by fabric sleeves so
that dust generated in the hoppers during the loading
operation is either filtered through the sleeves or exhausted
through a baghouse.  Dust control equipment can be installed

                              55

-------
           Table 15.   PARTICULATE CONTROLS FOR THE
                PREPARATION OF RAW MATERIALS
     Control
    technique
Efficiency
  Advantages
Disadvantages
Control
  equipment

  Enclosing all
  handling
  operations
  Baghouse on
  exhausts
Process changes
  Eliminate
  minus 325-mesh
  material in
  feed

  Wet the
  mixed batch
   100%
    99%
 100% for
respirable
   dust

   100%
Eliminates
all dusting


Good control,
proven
technology
Removes
respirable dust
Controls dust
from batch
charging;
provides better
furnace melting
Added cost
for the plant
operation

Added cost
for the plant
operation
                              56

-------
on conveying systems that use open conveyor belts.  A
reduction in the size of the dust control equipment can be
realized by totally enclosing all conveying equipment and
sealing all covers and access openings with gaskets of poly-
urethane foam.  In fact, by totally enclosing all conveying
equipment, exhaust systems become unnecessary and relatively
small filter vents or dust cabinets can be attached directly
to the conveying equipment and storage bins.12

Weigh hoppers and mixers require ventilation because of
surges in material which result in large air flows.  The
exhaust can be satisfactorily filtered of particulates in a
baghouse.  Seals of polyvinylchloride should be installed
between the rotating body of the mixer and its frame to
reduce air leaks.

The problem of dusting has also been alleviated by severely
limiting the amount of fine particles (<325-mesh) in the
feed material.  Experience has shown that <44 ym  (minus
325-mesh) particles cause severe dusting.12'41  Manufacturers
specify that glass sand should be all <0.82 mm (minus 20-mesh)
and all >44 ym (plus 325-mesh).  The specifications for an
acceptable glass sand appear in Table 16.  The table also
gives the particle sizes corresponding to given mesh sizes.
Although there are no mesh requirements spelled out for the
sizes less than 0.41 mm (40-mesh) nor greater than 0.105 mm
(140-mesh), "the distribution should be reasonably uniform on
the intermediate screens.

In a similar way glass grade limestone and dolomite are
processed to remove fine material.  Since these materials
MBrown, C. J.  Selection Criteria for Sand, Dolomite, and
  Limestone in the Flat Glass Industry.  In:  Collected
  Papers from the 32nd Annual Conference on Glass Problems,
  Dept. of Ceramic Engineering, University of Illinois,
  Urbana, November 1971.  p. 163-171.
                              57

-------
 Table 16.  ACCEPTABLE MESH SPECIFICATION FOR GLASS SAND41
Requirement
Cum.
Cum.
Cum.
Cum.
Cum.
retained on
retained on
retained on
retained on
retained on
U.S. standard
mesh size
20
40
140
200
325
Particle
size
820 ym
410 ym
105 ym
74 ym
44 ym
Limits
0.0%
12.0% Max
92.0% Min.
99.5% Min
100.0% Min
         Mesh Size and Corresponding Particle Size
U.S. standard mesh size
8
16
20
40
100
140
200
325
Approximate particle size
2.3
1.3
820
410
150
105
74
44
mm
mm
ym
ym
ym
ym
ym
ym
Cum. = cumulative.
                              58

-------
crush to dust more easily than sand they are sold as a
coarser mixture.  Specifications that minimize dusting are
given in Tables 17 and 18.  Dolomite is sold as either a 2 mm
(10-mesh) stone or a 0.82 mm (20-mesh) stone.

Another process modification that controls dusting is the
addition of a small amount (3%) of water to the mixed batch.
The water contains a surfactant to aid in wetting the raw
materials.  The moisture eliminates dusting during the batch
charging operation, prevents the segregation of batch ingre-
dients, and permits the use of less salt cake in the batch.17
Pelletizing the batch materials is another technique which
accomplishes the same results.

All of these control techniques are now standard practice in
the flat glass industry.

B.   THE GLASS MELTING FURNACE

The major effort in controlling emissions from glass furnaces
has been through process modifications rather than pollution
control devices.  Typical approaches include controlling raw
materials to reduce the amount of volatilizable materials
(e.g., sulfates, fluorides, borates), changing the furnace
design and operation to give greater fuel efficiency, and
increasing the checkerwork volume for better heat savings.

A number of these process changes have been made primarily
to reduce the cost of glass making, but they also contribute
indirectly to a reduction in air emissions.  This indirect
effort comes about because a decrease in fuel requirements
will be accompanied by a decrease in combustion products, a
decrease in dust entrainment by hot combustion gases passing
over the molten gas, and in some cases by a decrease in the
furnace temperature.

                              59

-------
  Table  17.  ACCEPTABLE  MESH  SPECIFICATION FOR GLASS GRADE DOLOMITE
Requirement
Cum.
Cum.
Cum.
Cum.
Cum.
Cum.
retained on
retained on
retained on
retained on
retained on
retained on
U.S.
standard
mesh size
8
16
20
100
140
200
Particle
size
2.3 mm
1.3 mm
820 ym
150 ym
105 ym
74 ym
Limits
2 mm
(10-mesh)
0.0%
15.0% Max

90.0% Min,,

97.0% .Min,.,
0.82 mm
(20-mesh)
0.0%

2.0% Max
80.0% Min
95.0% Min
96.0% Min
Cum. = cumulative.
 Table 18.  ACCEPTABLE MESH SPECIFICATION FOR GLASS GRADE LIMESTONE41
Requirement
Cum.
Cum.
Cum.
Cum.
Cum.
retained on
retained on
retained on
retained on
retained on
U.S. standard
mesh size
8
16
20
140
x 200
Particle
size
2.3 mm
1.3 mm
820 ym
105 ym
74 ym
Limits
0.0%
2.0% Max.
10.0% Max
85.0% Min
94.0% Min
Cum. = cumulative.
                                   6.0

-------
In the same way factors that increase furnace life will
decrease emissions by improving the operating efficiency.
Better refractories are a case in point, both in the furnace
and in the checkerworks.  An eroded and partially plugged
checkerwork loses efficiency as a heat regenerator.

Unfortunately it is not possible to quantify the effects of
specific process changes in lowering emission rates because
a number of variables in the furnace operation are usually
changed simultaneously.  As an example, supplemental electric
heating will, in itself, result in a lower furnace tempera-
ture and lower particulate emissions.  However, the produc-
tion rate is generally increased at the same time so that
the furnace temperature remains the same.

Another factor which should be noted is the continual
improvement in furnace operations.  The 1970-1972 data from
which the emission factors in Section IV were determined do
not reflect the actual performance in 1975.  Nor would 1975
data give a true picture of 1978.

1.   Process Modification

The effects of process modifications can best be discussed
in terms of specific air emissions.  Particulates, for
instance, are generated by the entrainment of dust in
combustion gases and by the volatilization of materials in
the melt.  The elimination of <44 ym (minus 325-mesh)
particles in the feed material and the addition of water to
the glass batch, as discussed in Section V.A, will minimize
dust entrainment.

Volatilization of the melt can be reduced by controlling the
feed material, by proper furnace design, by lowering the
                              61

-------
furnace temperature, and by electric melting.12'42  Raw
materials which vaporize in the melt include sulfates,
borates, nitrates, fluorides, chlorides, arsenic, selenium,
antimony, and lead.  Of these only sodium sulfate is still
used as a standard ingredient in flat glass.  None of the
others is used by more than 10% of the industry.  Sodium
sulfate  (salt cake) or another sulfate is a necessary flux
that prevents scum formation in the melting furnace and aids
in the melting process.  Manufacturers reduce the salt cake
in the glass batch as much as possible consistent with good
glass making.  Exact details are considered proprietary.

Emissions from fluxing agents can be lowered by improving
overall furnace efficiency.  The following methods that
improve this efficiency are being practiced:*2/^3

   •  Applied instrumentation to regulate air/fuel mixtures,
     monitor furnace temperature and stack gas composition,
     automatically charge the batch into the furnace,  and
     reverse the air flow through the regenerative checkers.
   •  Combustion control to produce large luminous flames
     that eliminate hot spots in the furnace and provide
     better heat transfer to the melt.
   •  Increased checker volume for better heat recovery  (the
     ratio of checker volume to melter area is about
     2.74 m3/m2 today).
   •  Improved refractories for corrosion resistance and
     better insulation.
42Simon, H.,  and J. E. Williamson.  Control of Fine
  Particulates from Continuous Melting Regenerative
  Container Glass Furnaces.  Los Angeles County Air
  Pollution Control District.  (Presented at the 68th
  Annual Meeting of the Air Pollution Control Association,
  Boston.  June 15-20, 1975.)  12 p.
43Hamilton, J. C.  Applied Research in Glass Melting.
  Industrial and Engineering Chemistry.  62 ;16-21,
  February 1970.
                              62

-------
As a result of these measures the average amount of fuel needed
to melt a metric ton of glass decreased from about 10.5 GJ in
1936 to 7.1 GJ in 1964  (theoretical amount is 1.9 GJ/metric
ton).9'1*2

The increased checker volumes not only reduce fuel consumption
but also present a trap for dust particles.  Although dust
collects within checkers by the mechanisms of impingement and
settling (gaseous materials also condense out on the checkers),
the relationship among various factors influencing collection
is unknown.  These factors include gas velocity and temperature,
brick size and composition, flue spacing, and brick setting.
It is known that micron-size particles are trapped in the
checkers while submicron-size particles escape and go out the
stack (see Section IV).12/25

As mentioned in Section IV the furnace temperature has a
profound influence on the particulate emission rate.  Two
studies have found the emissions rate to increase exponen-
tially with temperature.24'31*  The furnace temperature can
be lowered by improving furnace efficiencies, by decreasing
the production rate, and by using supplemental electric
heating.

Glass conducts electricity at high temperatures and it can
be melted by passing an electric current through it.9  The
use of electrical energy to assist in a fuel-fired furnace
is called electrical boosting.  Although boosting is not
practiced in flat glass furnaces, it is widespread in the
container glass industry.  Boosting permits a furnace to
operate at a lower temperature with the same production rate
because heat is being introduced near the bottom of the
molten glass as well as at the surface.1*2  (All-electric
furnaces are available, but they are unsuited for the large-
scale flat glass production.)
                              63

-------
The results of all these process changes are illustrated by
two glass container furnaces in California that used the
options outlined above to meet state and county (Los Angeles)
emission standards.  They were tested 13 times from 1972 to
1975 and had an average particulate emission factor of
0.29 ± 0.048 g/kg  (vs. 1 g/kg for the industry average in
Table 8).  However, three other furnaces were unable to meet
the regulations by these means and pollution control equip-
ment had to be installed.42

All the methods discussed for reducing volatilization can also
be used to control gaseous emissions.  Sulfur oxides are formed
from sulfates that decompose in the melt, and they can be
controlled by limiting the amount of sulfate in the feed
material and by improving the furnace efficiency.   One study
found that the SO  emission rate increased directly with an
increase in production rate on a furnace melting soda-lime
glass.24  The authors stated that the increase was a result
of higher furnace temperatures (1,552°C vs. 1,460°C) at the
increased production rate (180 metric tons/day vs. 90 metric
tons/day).

A fuel oil containing 1% sulfur produces about 600 ppm SO  in
the flue gas.31  In such a case a change from oil to natural
gas will lower emissions, but this is no longer a viable op-
tion with the deteriorating natural gas supply.  Natural gas
firing is preferred by glass producers since it causes less
wear on the furnace refractories.

Nitrogen oxides form in the melting furnace by a combination
of atmospheric oxygen and nitrogen.  Emission levels can be
lowered by increasing the furnace efficiency and decreasing
the furnace temperature.  Other factors being equal, a 10%
decrease in fuel consumption should be reflected in a 10%
decrease in NO  emissions.

                              64

-------
The reaction between N2 and 02 is strongly temperature depen-
dent  (see Table 19), so that careful control of combustion
and elimination of hot spots in the furnace will reduce the
formation of NO . 29  Doubling the production rate from 90 to
               X
180 metric tons/day  (which was accompanied by a higher furnace
temperature, 1,460°C to 1,552°C) resulted in a sixfold in-
crease in the NO  emission rate (from 100 ppm to 600 ppm).21+
          Table 19.  TIME FOR NO FORMATION IN A GAS
           CONTAINING 75% NITROGEN AND 3% OXYGEN29
Temperature , ° C
1,360
1,538
1,760
1,982
Time to form
500 ppm NO, sb
1,370
16.2
1.10
0.117
NO concentration at
equilibrium, ppm
550
1,380
2,600
4,150
JThe glass melt temperature is ^1,500°C.
 A large  (500 metric tons/day) furnace would have, an air flow
 of ^94 m3/s and an air space in the furnace of ^2,830 m3,
 for a residence time of ^30 s.
2.
Pollution Control Devices
Add-on control devices have not been used on flat glass
furnaces because process modifications are able to meet
state and federal emission standards.  The only problem area
is opacity regulations which may be exceeded because of the
submicron size of the particulates.

Other sections of the glass industry, particularly the Glass
Container Manufacturers Institute, have studied the applica-
bility of different controls to particulate collection.
Problems encountered in adapting equipment to glass furnace
exhausts include the submicron particle size, the corrosive
                              65

-------
nature of exhaust gases, and high stack gas temperatures
(sometimes over 550° C)..  Systems that have undergone testing

are scrubbers, baghouses, electrostatic precipitators, and
two new methods proposed by Teller Environmental Systems,
Incorporated  (TESI).l2,^-50


a.   Scrubbers - A low-pressure, wet, centrifugal scrubber
that was used by the Thatcher Glass Co. in Saugus, California,
showed an overall particulate collection efficiency of 52%.12
The low efficiency demonstrates the inherent inability of
such scrubbers to collect particulates of submicron size.
^Edmondson,. J. N., L. Reitz, R. L. Weise, and J. Fraas.
  Design, Installation, and Operation of Equipment to Cool
  and Filter Particulate Matter from Flue Gas from a
  Regenerative Furnace.  In:  Collected Papers from the
  32nd Annual Conference on Glass Problems.  Dept. of
  Ceramic Engineering, University of Illinois, Urbana,
  November 1971.  p. 39-54.

45Teller, A. J.  Control of Emissions from Glass Manu-
  facture.  Ceramic Bulletin.  5^:637-640, August 1972.

tf6Frantz, C. N., D. L. Miser, H. N. Troy, and E. D. Stobbe.
  Glass Furnace Particulate Emission Control Equipment.  In:
  Collected Papers from the 32nd Annual Conference on Glass
  Problems.  Dept. of Ceramic Engineering, University of
  Illinois, Urbana, November 1971.  p. 25-38.

l+7Keller, G.  Scrubber System Lightens Load of Glass Furnace
  Emissions.  Chemical Processing.  3_8_: 9, January 1975.

^8Symposium on Pollution, Stratford-Upon-Avon, 30 May -
  1 June, 1973.  In:  Glass Technology.  14J6):140-144,
  December 1973.

49Tank Emissions "Bagged."  The Glass Industry.   55:18,
  July 1974.

50Moyer, T., S. Reigel, and C. Doyle.  Gas-Assisted
  Atomizers Help End High Heat Problem in Collector.
  Maintenance Engineering.  22:28-29, June 1972.
                              66

-------
Owens-Illinois tested an adjustable throat, high-energy
venturi scrubber with a system that collects and recycles
the scrubbing liquor back to the venturi throat.  Although
the scrubber performed adequately  (collection efficiencies
of 85% to 95%. and stack gas effluent within state standards) ,
the cost of a water treatment system for the scrubber liquor
blowdown was excessive.46   (Owens-Illinois did not report
actual cost data.)

The FMC Corporation has developed a new scrubber with a
packed bed preconditioning chamber.  Hot gases  (540°C)
containing volatilized sodium compounds enter the chamber
and the vapors condense out on the packing material.  This
material, which is wet by the scrubbing solution, provides a
large surface area for condensation.  A standard venturi
type scrubber completes the system.  The scrubber is
installed on a 150 metric ton/day container glass furnace in
Vernon, California, and reduces the particulate loading by
^70%.  Thus far the system has suffered from many malfunc-
tions and breakdowns."42/47

b.   Fabric Filters - At least four different fabric filters
have been used in baghouses to control glass furnace parti-
culate emissions  (Nomex, Dacron, Teflon, and fiber
glass) .12,42,44,>+6,18-so  Owens-Illinois conducted a
feasibility study in 1969 using Nomex bags on a borosilicate
glass furnace.  The collection efficiency was greater than
99%, but serious fabric plugging occurred with attendant
high pressure losses when the outlet gas temperature dropped
much below 150°C in the pilot unit, probably due to conden-
sation on the colder bags.46  Nomex was also used success-
fully on a multifurnace plant operated by General Electric.44
Some problems were encountered when the bags came apart at
the seams, perhaps because the stitching was abraded by the
dust.

                              67

-------
Owens-Illinois also tested a Teflon bag filter that had a
life of 15 months.48  No operating data were reported.

Dacron filter bags were used on a 130 metric ton/day container
glass furnace in Los Angeles.  Particulate collection effi-
ciencies ranged from 64% to 82% while the opacity reading was
0%.

Glass fiber bags have been used on several furnaces but their
lifetimes were only mentioned once. I2f l+9» 50  One baghouse
operating at 205°C alternately vents two small (820-kg and
2,300-kg) regenerative furnaces.12  Another operates on a
180-metric ton furnace in which exhaust gas at 650°C is cooled
by a water spray to ^260°C.13'50  A collection efficiency of
99+% was reported, but subsequent tests indicated that some
of the volatilized particulates were not collected because of
the high baghouse temperature.  The addition of cooling air
made the filtering velocity too high for sustained use.  A
third system utilized four collectors at 250°C.  Only one
bag failed in 18 months of test operations.1*9

Temperature control is very critical for the proper func-
tioning of a baghouse.  Since the stack gas from a glass
furnace is at 350°C to 650°C, the gas must be cooled to a
temperature compatible with the bag material.  Maximum
operating temperatures for some typical fabric filters are
given in Table 20.^

Temperatures must also be controlled at the low end since
803 and H2O in the gas stream can condense and foul the
bags.12  The furnace effluent can be cooled by several
methods, either alone or in combination.12'414  These methods
are:  (1) air dilution, (2) radiation cooling columns,
(3) air-gas heat exchangers, and  (4) water spray chambers.
                              68

-------
   Table 20.  MAXIMUM USE TEMPERATURE FOR FABRIC FILTERS
            Fabric
        Cotton sateen
        Standard nylon
        Wool
        Dacron
        Orion
        Nomex
        Teflon
        Fiber glass
Maximum temperature,
         °C
          88
          93
         107
         135
         135
         204
         232
         288
Each cooling method has its advantages and disadvantages.
Dilution of off-gases with air is the simplest and most
troublefree way to reduce temperature but requires the
largest baghouse.  Air-to-gas heat exchangers and radiation
and convection ductwork are subject to rapid fouling from
dust in the effluent.  A water spray increases the humidity
and requires careful temperature control to avoid condensa-
tion, but it does permit the use of a smaller baghouse.

The fabrics used in the baghouse must be chosen not only for
their heat resistance but also for their resistance to
corrosion and abrasion.  Experience has shown that cotton,
Orion, and Dacron are deteriorated by SO3 in the flue gas.12
It is likely that nylon and wool would also be attacked by
acidic gases.

c.   Electrostatic Precipitators - Tests with electrostatic
precipitators have shown that their collection efficiency is
from 80% to 98% for particulates from glass melting
                              69

-------
furnaces.l*2'46 'lt8 '   Owens-Illinois conducted a pilot-scale
study on a borosilicate furnace and achieved between 40% and
95% efficiency.  A full-scale unit operated at 80% to 90%
efficiency.  Factors that prevented better performance were
the size and resistivity of the particles and the plate
cleaning procedures.  Operating temperature,of the full-scale
unit was ^370°C and it was felt that a hotter operating
temperature (430°C), which would reduce the particle resis-
tivity by a factor of ]0, should improve the collection
efficiency.46

Another electrostatic precipitator was reported to achieve
80% removal only with difficulty, even though it was working
well within capacity.1*8

An electrostatic precipitator is installed on a glass container
furnace in California (^180 metric tons/day capacity).  The
two-chamber unit has an efficiency of 83% to 89% and functions
with equal efficiency whether one or two chambers are used.42

The Japanese firm of NAFCO Engineering, Ltd., has developed
an entirely new type of electrostatic precipitator.  It was
designed to clean dust emissions from glass furnaces to meet
Japanese emission standards.  In contrast to conventional
units in which hanging wires discharge electricity, the
NAFCO ESP  (electrostatic precipitator) uses thousands of
stainless steel needles affixed to the leading and trailing
edges of positively charged electrode plates.  The Japanese
have some 35 systems in operation, 60% on lead glass
furnaces, 15% on borosilicate glass, and 25% on soda-lime
glass.36   .
 An efficiency of 98% was reported by a flat glass
 manufacturer.on a pilot-scale unit..

                              70

-------
The licensed U.S. distributor, United McGill Corporation,
has installed the new ESP on 12 glass furnaces  (five boro-
silicate glass, four lead glass, three soda-lime glass), and
30 units are scheduled to be in operation by mid-1975  (10
borosilicate glass, 10 lead glass, 10 soda-lime glass).  All
of the systems in operation have an efficiency of at least
85%, based on a minimum uncontrolled emission factor of
1 g/kg glass and an air flow of 3 m3/kg glass.  The Japanese
ESP is designed so that additional sections can be added to
attain higher efficiencies  (stated to be >99%) if desired.36

d.   TEST - Teller Environmental Systems, Inc., (TESI) offers
a dry and a wet system for emissions control.  The wet system
uses a nucleation scrubber to effect collection of submicron
particulates and acid gases (SO ).  In the dry system a solid
                               X
absorbent is injected into the gas stream to react with nox-
ious gases.  The absorbent is separated from the gas along
with the particulates in a fabric filter.  Pilot-scale studies
on a slipstream (180 m3/min) from a fiber glass furnace showed
collection efficiencies of 96% for particulate and 99% for
SO .^  An attempt to use the system on a flat glass furnace
  X
failed because of engineering problems.

e.   Summary - Table 21 summarizes the advantages and dis-
advantages of the different control techniques.  Although
baghouses and electrostatic precipitators remove only
particulates, it has been suggested that spraying the stack
gas with an alkaline solution would cause acid gases  (SO )
                                                        X
to react and form particulates.36  These could then be
collected by the control device.

Control development has focused on particulate removal to
meet emission standards for stack opacity.  As a result,
there is no proven technology for NO  removal from stack
                                    X
gases, and the effect on NO  emission levels by process
                           A.
                              71

-------
Table 21.  CONTROL TECHNIQUES FOR GLASS MELTING FURNACES
Control
Process
modifications
Remove
volatile
materials
from batch

Remove minus
325-mesh
materials;
wet the mixed
batch
Reduce sulfate
in batch
Improve
furnace
efficiency

Electrical
boosting

Emission
controlled


Borates
Fluorides
Arsenic
Selenium
Chlorides
Particulates

Particulates
SO
X
Particulates
SO
NOX
. cox
Hydrocarbons
Particulates
SO
NOX
cox
Hydrocarbons
Efficiency


100%
100%
100%
100%
100%
Primarily controls
dust carried into
checkerwork rather
than stack emissions

Depends on % sulfate
reduction
Function of furnace
efficiency

Dependent on the
amount of boosting

Advantages


Most economical ,
complete control


Controls dust
entrainment; stops
segregation of raw
materials

Inexpensive way to
reduce particulates
and SO
X
Actually saves money

Reduces all
emissions at same
production rate

Di sadvantage s


May affect glass
quality




Careful operation is
required to maintain
glass quality
Even at 100% energy
utilization there
are still emissions

Not developed for
flat glass


-------
Table 21 (continued).   CONTROL TECHNIQUES FOR GLASS MELTING FURNACES
Control
Control devices
Scrubbers
Fabric filters
. Conventional
ESP
. Japanese ESP
TESI
Emission
controlled

Particulates
SO
NOX
X
Particulates
Particulates
Particulates
Particulates
SO
NOX
X
Efficiency

Up to 95%
Up to 90%
Unknown
. Up to 99%
Up to 98%
Up to 99%
96%
99%
Unknown
Advantages

Controls both
particulate and SO
Good particulate
control
Good particulate
control
Good control;
proven technology
Controls both
particulate and SO
Disadvantages

Low efficiency
for submicron
particulates; high
energy requirement ;
wastewater
Only controls
particulates ;
temperature
control is critical,
fabric corrosion
Only controls
particulates
Only controls
particulates
System has failed to
operate on more than
a pilot scale

-------
modifications is unknown.  The wide variation in reported
NO  emission factors (Table C-4 in Appendix C) suggests a
best case condition of 1 g/kg  (vs. an average of 4 g/kg).
However, it is doubtful whether other furnaces could reach
this level without lowering their production level.

C.   FORMING AND FINISHING OPERATIONS

Because atmospheric emissions from these operations are low
(<0.1 g/kg) or nonexistent, control devices are unnecessary.
Annealing lehrs of increased efficiency do produce fewer
combustion products per metric ton of glass annealed, but that
is not why they are used.  Rather, they are desirable because
they consume less energy.  The new recirculating air type
lehrs can save 40% to 60% of the fuel cost over the older,
nonrecirculating models.27/28
                              74

-------
                          SECTION VI
         GROWTH AND  NATURE OF  THE FLAT  GLASS  INDUSTRY

The past 10 years have witnessed a revolution  in  the  flat
glass  industry as the float process has displaced the older
plate  glass process and  even  made inroads on sheet glass pro-
duction.  There  is  presently  only one  plate  glass plant still
in operation and two sheet glass plants have been shut down.
Future trends in the industry will see sheet glass being
replaced by thin float glass.22/26'51

The production statistics in  Table 22  indicate that float
production in 1973  was more than twice the 1967 level of
float  and plate, while sheet  production remained  constant.
Since  plate glass weighs twice  as much as sheet glass, the
overall increase was 81%,  or  about 10% per year.  In  1974
the slump in the economy caused a slight decline  in pro-
duction.  If growth resumes in  1975, the 1978  production
level  will be 46% above  that  for 1973.  Total  national
emissions will also increase  by this amount  unless there
are new developments in  control technology.

Other  development trends in the flat glass industry include
.the increasing use  of automation15'21'51 and a new, more
 51Child, F. S.  The  Impact of Flat Glass  Imports.   The
  Glass  Industry.  52^166-169, May 1971.

                              75

-------
       Table 22.  FLAT GLASS PRODUCTION STATISTICS2'52
Year
1974
1973
1972
1967
Sheet glass
production,
km2
92.4
110.1
117.5
101.7
Plate, float, rolled and
wire glass production,
km2
(only 5% is rolled and
wire glass)
170.8
173.4
141.4
77.2




efficient design for annealing lehrs.  The new lehrs are
expected to replace the older nonrecirculating type lehrs.27
New techniques are also available for making tinted and
reflective glass for architectural applications.
52
  Ceramic Industry Newsletter.  Ceramic Industry,
  May 1975.
104:9,
                              76

-------
                         SECTION VII
                       UNUSUAL RESULTS

Along with many other industries the flat glass industry is
experiencing a natural gas shortage.  This has a direct
effect on air emissions because the oil that is used as a
fuel substitute contains sulfur.  As a general guideline, a
changeover from gas to oil will double the emission factors
for SO  and particulates.
      jrL

A factor tending to counter this effect is the improved
furnace efficiency being achieved by process modifications
that act to reduce stack emissions.  There is no way to
predict the quantitative impact of these two factors on
future air emissions.
                              77

-------
                        SECTION VIII

                         APPENDIXES
A.   Calculation of Glass Production on a Tonnage Basis
B.   Flat Glas.s Plant Listing
C.   Emissions Data
D.   Total Flat Glass Emissions
E.   Plume Rise Correction
F.   Derivation of Source Severity Equations
                              79

-------
                         APPENDIX A
     CALCULATION OF GLASS PRODUCTION ON A TONNAGE BASIS

The production of glass products is not reported on a
tonnage basis; instead the statistics list square feet of
flat glass or number of glass containers.  Consequently the
U.S. production was estimated from the amount of raw
materials consumed in each category in 1972.2~k

1.   FLAT GLASS
      Raw materials
     consumed in 1972      metric tons          (tons)
     Sodium carbonate         494,000           (544,500)
     Glass sand             1,840,000         (2,028,000)
     Sodium sulfate            60,000            (66,700)
     Gullet (glass scrap)     374,000           (412,450)

The total weight of these raw materials is 2.768 x 106 metric
tons (3,051,650 tons).  Two facts should be noted:   (1) a
number of raw materials  (e.g., lime) are not listed individu-
ally (the total cost for all other materials is reported
instead); and (2) sodium carbonate and lime decompose in the
glass furnace and give off C02-  The consumption of lime can
be estimated since it muast be approximately equal to the
consumption of sodium carbonate to give the proper batch
composition, i.e., 4.94 x 105 metric tons (544,500 tons).
                              81

-------
Data from the Bureau of Mines indicate that consumption of
aluminum-bearing minerals  (feldspar, aplite, and nepheline
syenite) in the glass industry was about 9.10 x 105 metric
tons (1 million tons) in 1972.53  This has been broken down
as follows:  9.1 x lQk metric tons  (100,000 tons) in flat
glass,  2.73 x 105 metric tons (300,000 tons) in pressed and
blown glass, and 5.44 x 105 metric tons (600,000 tons) in
glass containers.  The low value for flat glass reflects the
fact that these materials are not used in float glass
production.

The total weight of these ingredients is 3.353 x 106 metric
tons (3,696,150 tons).  After the decomposition of sodium
carbonate and limestone to give off C02 and sodium sulfate to
give off 803 there will be a weight loss of 4.56 x 105 metric
tons (503,200 tons).  The final production is then 2.9 x 106
metric tons (3.2 million tons), which compares favorably with
a 1972 transportation survey showing flat glass shipments of
2.8 x 106 metric tons (3.1 million tons).54

As noted previously, there is a high (30% to 50%) ratio of
cullet used in the glass batch.  The quantity given in the
data (11%) represents only purchased cullet material, not
off-quality and broken glass that is recycled.  Therefore
the amount of glass which is actually made is higher than
reported or approximately 4 x 106 metric tons (4.4 million
tons).   Finally, the total amount of raw materials handled
(taking into account volatilization losses) is 4.5 x 106
metric tons (4.9 million, tons).  The accuracy of these numbers
53Wells, J. R.  Feldspar, Nepheline Syenite, and Aplite.
  In:  Minerals Yearbook 1972, Volume I:  Metals, Minerals
  and Fuels.  Bureau of Mines, Washington.  1974.  p. 515-523,
54Ceramic Industry Newsletter.  Ceramic Industry.  104:7,
  February 1975.

                              82

-------
is on the order of ±10% because of the variation in the cullet
ratio.  These figures are summarized below:
      Flat Glass - 1972  106 metric tons/year    (tons/year)
     Finished products
     Total flat glass
       made
     Raw materials
       handled
   2.9
   4.0

   4.5
(3.2  million)
(4.4  million)

(4.9  million)
For comparison the current  (mid-1975) flat glass capacity is
^14,290 metric tons/day.  An 80% capacity for a full year
would give a production level of 4.2 x 106 metric tons
(4.6 million tons).
2.   GLASS CONTAINERS
     Raw materials
    consumed in 1972
     Sodium carbonate
     Glass sand
     Cullet
     Aluminum-bearing
       minerals (est.)
metric tons
 2,380,000
 7,379,000
 1,459,000
   545,000
    (tons)
  (2,617,800)
  (8,116,300)
  (1,604,800)
    (600,000)
The flat glass calculations show that the amount of lime is
close (8%) to the volatilization loss.  With this assumption
the estimated production of glass containers would be
11.7 x 106 metric tons  (12,938,900 tons).  This compares well
with the reported "net weight of machine-made glass containers
packed" of 11.5 x 106 metric tons (12,:663,950 tons).3
                              83

-------
3.   PRESSED AND BLOWN GLASS, N.E.C.
      Raw materials
     consumed in 1972          metric tons           (tons)
     Sodium carbonate             230,000            (252,500)
     Glass sand                 3,309,000          (3,639,400)
     Gullet                       350,000            (385,350)
     Aluminum-bearing             273,000            (300,000)
       minerals  (est.)

Estimated production of finished products is 4.062 x 106
metric tons (4,477,250 tons), or ^4.1 x 106 metric tons
(^4.5 million tons).

4.   TOTAL PRODUCTION

The total production of finished glass products is then
1.86 x 107 metric tons.
                              84

-------
                         APPENDIX B
                  FLAT GLASS PLANT LISTING
Flat glass plants are listed in Table B-l by state.  Data
presented are plant capacities, number of melting furnaces
(designated "tanks"), location by city and county, county
population density, and AQCR.  Plant capacities are based
on data in References 6, 7, 14 and 26 and communications
with industry representatives.  Actual capacities were not
always known and had to be estimated from production data.
                              85

-------
                                   Table  B-l.   LISTING OF  FLAT GLASS PLANTS
CO

Plant
Fourco Glass Co.
C.E. Glass, Inc.
Libbey-Owens-Ford
PPG Industries
Libbey-Owens-Ford
PPG Industries
PPG Industries
Ford Motor Co.
Guardian Ind.
Corp.
C.E. Glass, Inc.
PPG Industries
C.E. Glass, Inc.
(new plant)
Libbey-Owens-Ford
Libbey-Owens-Ford
-Libbey-Owens-Ford

Product
Sheet
Rolled
Float
Sheet
Float
Sheet
Float
Float
Float
Rolled
Float
Float
Float
Float
Float

State
Arkansas
California
California
California
• Illinois
Illinois
Maryland
Michigan
Michigan
Missouri
Missouri
New Jersey
North
Carolina
Ohio
Ohio
Location
City
Fort Smith
Fullerton
Lathrop
Fresno
Ottawa
Mt. Zion
Cumberland
Dearborn
Carleton
St. Louis
Crystal
City
Cinnamin-
son
Laurinburg
Rossford
Toledo

County
Sebastion
Orange
San Joaquin
Fresno
LaSalle
Macon
Allegany
Wayne
Monroe
(Independent
City)
Jefferson
Burlington
Scotland
Wood
Lucas

AQCR
17
24
31
31
71
75
113
123
124
70
70
45
169
124
124
County population
density.
(persons/mi2 )
57
(147)
696
(1802)
78
(201)
26
(68)
37
(96)
83
(215)
74
(192)
1792
(4638)
82
(211)
1083a •
(2803)
59
(153)
153
(396)
32
(83)
55
(144)
539
(1396)

capacities
2 tanks;
200 metric tons
(225 tons)
1 tank;
64 metric tons
(70 tons)
1 tank;
^400 metric tons
(-v-450 tons)
1 tank;
^360 metric tons
(i400 tons)
1 tank;
^360 metric tons
(•v400 tons)
1 tank;
^400 metric tons
(•v450 tons)
1 tank;
•\-360 metric tons
(i400 tons)
1 tank;
•v.360 metric tons
(i400 tons)
2 tanks;
820 metric tons
(900 tons)
2 tanks;
177 metric tons
(195 tons)
1 tank;
^360 metric tons
(•MOO tons)
1 tank;
•v.450 metric tons
(^500 tons)
1 tank;
•\-680 metric tons
(•>-750 tons)
2 tanks;
A.900 metric tons
(^1000 tons)
1 tank;
•v400 metric tons
(•v450 tons)
                 'Population density is for the city of St. Louis plus the surrounding county of St. Louis.

-------
                     Table B-l  (Continued).   LISTING OF FLAT  GLASS PLANTS
oo
•-J
Plant
PPG Industries
ASG Industries
Ford Motor Co.
(new plant)
ASG Industries
C.E. Glass, Inc.
PPG Industries
PPG Industries
ASG Industries
ASG Industries
C.E. Glass, Inc.
Ford Motor Co.
PPG Industries
(new plant)
Fourco Glass
Libbey-Owens-Ford
Product
Sheet
Sheet
Float
Sheet
Float
Float
Float
Float6
Rolled
Rolled
Float
Float
Sheet
Sheet
Location
State
Ohio
Oklahoma
Oklahoma
Pennsylvania
Pennsylvania
Pennsylvania
Pennsylvania
Tennessee
Tennessee
Tennessee
Tennessee
Texas
West
Virginia
West
Virginia
City
Mt. Vernon
Okmulgee
Tulsa
Jeanette
Floreffe
Carlisle
Meadville
Greenland
Kingsport
Erwin
Nashville
Wichita
Falls
Clarksburg
Charleston
County
Knox
Okmulgee
Tulsa
Westmoreland
Allegheny
Cumberland
Crawford
Sullivan
Sullivan
Unicol
Davidson
Wichita
Harrison
Kanawha
AQCR
175
186
186
197
197
196
178
207
207
207
208
210
235
234
County population
density,
persons/km2
(persons/mi2)
31
(79)
19
. (49)
268
(694)
142
(369)
842
(2180)
109
(283)
30
(77)
116
(301)
116
(301)
31
(81)
402
(1041)
75
(194)
66
(171)
95
(247)
Furnaces and
capacities
3 tanks;
••\-600 metric tons
(1-660 tons)
2 tanks;
160 metric tons
(180 tons)
2 tanks;
1-900 metric tons
(•vlOOO tons)
1 tank;
245 metric tons
(270 tons)
1 tank;
.1*360 metric tons
(1.400 tons)
2 tanks;
1-900 metric tons
(i-lOOO tons)
2 tanks ;
1-730 metric tons
(i-SOO tons)
2 tanks;
1-720 metric tons
(800 tons)
5 tanks;
295 metric tons
(325 tons)
1 tank;
1-64 metric tons
(i-70 tons)
3 tanks :
1-1360 metric tons
(1-1500 tons)
2 tanks;
1-900 metric tons
(i-lOOO tons)
2 tanks;
180 metric tons
(200 tons)
2 tanks;
1-360 metric tons
(1-400 tons)
bl float, 1 polished plate (color only) .

-------
                         APPENDIX C
                       EMISSIONS DATA
1.   RAW MATERIALS PREPARATION

The preparation of raw materials has been divided into five
operations:   (1) handling of raw materials  (unloading,
conveying);  (2) crushing of scrap glass;  (3) filling and
emptying storage bins; (4) batch mixing  (and weighing); and
(5) feeding of mixed glass batch to melting furnace  (batch
charger).  Source test data are summarized in Table C-l;
additional data based on material balances and engineering
estimates appear in Table C-2.  The data base includes not
only flat glass plants but container glass and pressed and
blown glass.  Since the preparation of raw materials is the
same for all three categories, this provides a broader base.

As mentioned in the text, emissions from handling and batch
charging were estimated to be zero based on observation at
a large float glass plant.  The low level of emissions from
other points was also confirmed since there were no visible
particulate emissions from the batch house.

The average emission factor for batch mixing, based on the
six source test measurements, is 1.58 mg/kg ± 106% at a 95%
confidence level.  For raw materials storage it is
12.3 mg/kg, based on one source test and seven engineering

                               89

-------
Table C-l.  SUMMARY OF SOURCE TEST DATA FROM NEDS - MATERIALS HANDLING AND MIXING

Plant
X
B

R
W


MM
Material
processed,
metric tons/yr
(tons/year)
258,000 (284,000)
755,000 (830,000)

46,000 ( 50,600)
277,000 (305,000)
135,000 (148,000)
277,000 (305,000)
291,000 (320,000)
Particulate
emissions, .
metric tons/yi
(tons/year)
0.9 (1)
3.6 (4)

0 (0)
0.9 (1)
0 (0)
0 (0)
0.9 (1)
Emission
factor ,
mgAg
(Ib/ton)
3.5 (0.007)
5 (0.010)

0 (0)
3 (0.006)
0 (0)
0 (0)
3 (0.006)

Control
equipment
Fabric filter
Fabric filter

Wet scrubber
Fabric filter
Fabric filter
Fabric filter
None listed

Process
Batch mixing
Raw material
handling and storage
Batch mixing
Batch mixing
•Batch mixing
Batch mixing
Batch mixing

-------
                    Table  C-2.   PARTICULATE EMISSIONS FROM THE PREPARATION OF RAW MATERIALS14
                              (Based on material balance and engineering knowledge)
Plant
R
SS
SS
SS
SS
TT
LL
LL
W
W
Material processed,
metric tons/yr
(tons/yr)
46,000 ( 50,600)
88,000 ( 96,800)
122,000 (134,000)
232,000 (255,000)
232,000 (255,000)
165,000 (182,000
515,000 (567,000)
158,000 (174,000)
57,000 ( 63,000)
12,800 ( 14,100)
• Emissions,
metric tons/
year
(tons/yr)
0 (0)
(5 points)
0.9 (1)
0 (0)
0.9 (1)
4.5 (5)
3.6 (4)
12.7 (14)
9.1 (10)
0.9 (1)
0 (0)
Emission factor,
mg/kg
(Ib/ton)
0 (0)
10.5 (0.021)
0 (0)
4 (0.008)
20 (0.039)
22 (0.044)
25 (0.049)
55 (0.11)
16 (0.032)
0 (0)
Control type
Fabric filter
Fabric filter
Fabric filter
Fabric filter
Fabric filter
Fabric filter
Fabric filter
Centrifugal
collector
Fabric filter
Fabric filter
Operation
Batch mix - cullet
Batch house
Batch house sand
Batch mixing
Batch delivery
Batch silo conveyors
Raw material
Receiving and storage
Glass crush
Lime silo filtering
Slag silo filtering
vo

-------
estimates.  Only one value was reported for glass crushing
(55 mg/kg)> for a crusher equipped with a centrifugal
collector instead of a baghouse for dust control.  Since the
typical control practice is a baghouse, it was decided to
use the same value for the crusher as for raw materials
storage (i.e., 12.3 mg/kg).

The accuracy was found by using the t-test for source test
data.  Thus with batch mixing the sample mean, y, was
1.58 mg/kg and the sample standard deviation, s, was 1.59.
The confidence limits on y are then ±ks//n where k is the
"Student's t" variable for n-1 degrees of freedom
(t = 2.571), and n is the number of samples.  The accuracy
of the engineering estimates was assumed to be ±100%.

Stack heights for the various operations are listed in
Table C-3.  They range from 1.5 m (5 ft) to 44.2 m (145 ft).
Since all the emissions are particulates an overall average
stack height of 21.5 m (70.6 ft) and a total emission factor.
of  0.02 g/kg were used to calculate x    and S.  The value
                   _                  max
of S is 0.02 while Xm=v equals 5.0 yg/m3.
                    nicLX                         *

2.   GLASS MELTING FURNACE

a.   Nitrogen Oxides

Source test measurements of NO  emissions as reported in
                              . X
NEDS are tabulated in Table C-4.  Emission factors vary from
0.71 to 10.05 g/kg (1.42 to 20.1 Ib/ton), and reflect the
different operating conditions  (especially temperature)
found in glass furnaces.   The average emission factor of
3.94 g/kg (7.88 Ib/ton) was found by dividing the total
emissions by total production.  Another average can be found
by adding the emission rates together and dividing by the
number of values.  This value is 4.37 g/kg (8.37 Ib/ton),
                            .•  92

-------
                 Table C-3.  STACK HEIGHTS14
                            (ft)
Batch mixing
15
15
15
120
16
5
80
117
75
43
80
75
Storage
72
72
40
120
120
120
75
145




Glass crushing
62











      Ave =54.7
Ave =95.5
Ave = 62
     Overall average = 70.6 ft or 21.5 m

indicating that small furnaces have higher emission factors.
However, the difference is not significant because the
standard deviation is ±3 g/kg  (±6 Ib/ton), and the 95%
confidence limit is 1.2 g/kg  (2.4 Ib/ton).  In addition, the
individual measurements are accurate to only ±25%.  An
average emission factor of 4 g/kg (8 Ib/ton) is used in the
main report.

b.   Sulfur Oxides

Source test data are presented in Table C-5.  Emission factors
vary from 0.20 to 4.43 g/kg (0.41 to 8.85 Ib/ton), with a
number average of 1.35 g/kg (2.71 Ib/ton) and a weight
average of 1.62 g/kg (3.23 Ib/ton).  Since the standard
                                 93

-------
Table C-4.  NO  EMISSIONS FROM GLASS MELTING FURNACESlk
              x
Plant .
A '.''• '

C
C -. '
c " ;
c ••'.
D V
E
E ''.'
E .'"•
F .
F
G
H
H
H
I

Production,
metric tons/year
(tons/year)
46,200

32,000
11,800
34,800
48,000
143,000
36,600
41,400
99,100
85,800
47,700
56,000
39,700
34,000
39,000
70,300

( 50,800)

( 35,100)
( 13,000)
( 38,300)
( 52,700)
(157,000)
( 40,300)
( 45,500)
(109,000)
( 44,100)
( 52,500)
( 62,000)
( 43,700)
( 37,400)
( 42,900)
( 77,300)

Emissions,
metric tons/yr
(tons/year)
51 ( 56)

168 (185)
63 ( 69)
185 (203)
254 (279)
1,490 (1,640)
227 (250)
118 (130)
382 (420)
52 ( 57)
300 (330)
152 (167)
222 (244)
295 (325)
393 (432)
350 (385)

Emission factor,
g/kg
(Ib/ton)
1.1 ( 2.20)

5.3 (10.5)
5.3 (10.6)
5.3 (10.6)
5.3 (10.6)
10.5 (20.9)
6.2 (12.4)
2.85 ( 5.71)
3.85 ( 7.71)
1.3 ( 2.59)
6.3 (12.6)
2.7 ( 5.39)
5.6 (11.2)
8.7 (17.4)
10.5 (20.1)
5.0 ( 9.96)

Product
Glass containers "(one
stack, two furnaces)
Glass containers
Glass containers
Glass containers
Glass containers
Flat glass
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
(two furnaces)

-------
                       Table C-4 (continued) .  NO  EMISSIONS  FROM GLASS MELTING FURNACES11*
                                                 x
Plant
K
K
.K
P

S
S
Y
LL
MM
ZZ
ZZ
TOTALS
Production,
metric tons/year
(tons/year)
27,500 ( 30,300)
34,800 ( 38,300)
71,200 ( 78,300)
179,000 (197,000)

58,800 ( 64,700)
38,700 ( 42,600)
20,800 ( 22,900)
171,000 (188,000)
173,000 (190,000)
23,300 ( 25,600)
4,430 ( 4,870)
1,622,000 (1,784,170)
Emissions,
metric tons/yr
(tons/year)
27.3 ( 30)
49.5 ( 55)
78 ( 86)
735 (809)

51 ( 56)
38 ( 42)
78 ( 86)
•247 (272)
123 (135)
233 (256)
32 ( 35)
6,395 (7,034)
Emission factor,
g/kg
(Ib/ton)
0.99 ( 1.98)
1.44 ( 2.87)
1.1 ( 2.2)
4.11 ( 8.21)

0.86 ( 1.73)
0.99 ( 1.97)
3.75 ( 7.51)
1.45 ( 2.89)
0.71 ( 1.42)
10.0 (20.0)
7.2 (14.4)
Product
Glass containers
Glass containers
Glass containers
Glass containers
(five furnaces)'
Glass containers
Glass containers
Glass containers
Flat glass
Flat glass
Pressed and blown glass
Pressed and blown glass
3.94 ( 7.88) (average)
VD
Ul
      Total Glass Containers
                 1,108,000 (1,218,700)
      Total Flat Glass
                   486,000   (535,000)
      Total Pressed and Blown Glass
                    27,700   ( 30,470)
4,269 (4,696)

1,861 (2,047)

  265    (291)
3.85 ( 7.71) (average)
3.82 ( 7.65) (average)
9.55 (19.10) (average)

-------
Table C-5.
                           SOURCE TEST DATA FROM NEDS - SO  EMISSIONS FROM GLASS MELTING  FURNACES
                                                          x
Plant
A

A
C
C
C
C
D
F
F
L
L
K
K
K
N
P
Production,
metric tons/year
(tons/year)
46,200 ( 50,800)

15,900 ( 17,500)
31,900 ( 35,100)
11,800 ( 13,000)
34,800 ( 38,300)
47,900 ( 52,700)
143,000 (157,000)
40,100 ( 44,100)
47,700 ( 52,500)
41,100 ( 45,200)
30,100 ( 33,100)
27,500 ( 30,300)
34,800 ( 38,300)
71,200 ( 78,300)
81,800 ( 90,000)
179,000 (197,000)
Emissions,
metric tons/year
(tons/year)
27.3 . ( 30)

8.2 ( 9)
16.4 ( 18)
6.4 ( 7)
17.3 ( 19)
24.6 ( 27)
450 (495)
94.6 (104)
12.3 (135)
10.9 ( 12)
9.1 ( 10)
10.9 ( 12)
10.9 ( 12)
52.7 ( 58)
177 (195)
111 (122)
End ssion factor,
g/kg
(Ib/ton)
0.59 (1.18)

0.52 (1.03)
0.52 (1.03)
0.54 (1.08)
0.49 (0.99)
0.52 (1.03)
3.15 (6.31)
2.36 (4.72)
2.57 (5.14)
0.27 (0.53)
0.30 (0.60)
0.40 (0.79)
0.32 (0.63)
0.74 (1.48)
2.17 (4.33)
0.62 (1.24)
Product
Glass containers (two
furnaces
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Flat glass
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers (three
furnaces)
Glass containers (five
furnaces)
vo

CTi

-------
        Table C-5  (continued).   SOURCE TEST DATA FROM NEDS - SO  EMISSIONS PROM GLASS MELTING FURNACESJ
Plant
S
S
W
X
X
X
X
Y
Y
Z
Z
Z
Z
DD
GG
JJ '
JJ
JJ
Production,
metric tons/year
(tons/year)
58,800 ( 64,700)
38,700 ( 42,600)
105,000 (115,000)
60,500 ( 66,500)
57,000 ( 63,000)
45,000 ( 49,000)
51,000 ( 56,000)
40,300 ( 44,300)
20,800 ( 22,900)
34,600 ( 38,000)
33,600 ( 37,000)
52,000 ( 57,000)
14,400 ( 15,800)
126,000 (139,000)
37,000 ( 41,000)
73,400 ( 80,700)
99,000 (109,000)
35,400 ( 38,900)
Emissions,
metric tons/year
(tons/year)
27.3 ( 30)
50.0 ( 55)
209 (230)
50 ( 55)
103 (113)
69.1 ( 76)
72.7 ( 80)
8.2 ( 9)
43.6 ( 48)
145 (160)
64.6 ( 71)
96.4 (106)
39.1 ( 43)
37.3 ( 41)
38.2 ( 42)
53.6 ( 59)
45.5 ( 50)
10.2 ( 11)
Emission factor,
g/kg
(Ib/ton)
0.47 (0.93)
1.29 (2.58)
2.0 ( )
0.82 (1.65)
1.80 (3.59)
1.55 (3.10)
1.43 (2.86)
0.21 (0.41)
2.20 (4.19)
4.21 (8.42)
1.92 (3.84)
1.86 (3.72)
2.72 (5.44)
0.30 (0.60)
1.03 (2.05)
0.73 (1.46)
0.46 (0.92)
0.29 (0.57)
Product
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Pressed and blown glass-
ware (nine furnaces)
Glass containers
Glass containers
Glass containers
Glass containers
vo
-j

-------
        Table  C-5  (continued).   SOURCE TEST DATA FROM NEDS - SO  EMISSIONS FROM GLASS MELTING FURNACES14
                                                               X
       Plant
                Production,
             metric tons/year
                (tons/year)
  Emissions,
metric tons/year
   (tons/year)
Emission factor,
      g/kg
    (Ib/ton)
         Product
        LL


        MM

     TOTALS
             171,000    (188,000)


             173,000    (190,000)

           2,106,021  (2,316,600)
     Glass Containers
oo    Flat  Glass
   756      (832)


   539      (593)

 3,399    (3,739)
                 1,493,288 (1,642,600)      1,616   (1,778)
             486,400    (535,000)     1,745    (1,920)

Pressed and Blown Glass
                   126,400   (139,000)
                                         37.3    ( 41)
   4.43 (8.85)
   3.12 (6.24)
Flat glass  (three
  furnaces)
Flat glass
   1.62  (3.23)  (average)
                                                             1.08  (2.16)  (average)
                                                                  3.59 (7.18)  (average)
                        0.30 (0.60)  (average)

-------
deviation is ±1 g/kg  (±2 Ib/ton) and the 95% confidence
limit is ±0.37 g/kg  (0.75 Ib/ton), the difference between
the averages is not significant.  An average value of
1.5 g/kg ± 0.4 g/kg  (3 Ib/ton) is used in the main text.

The averages at the end of Table C-5 for different types of
glass indicate that flat glass has a higher S0x emission
factor  (3.59 g/kg vs 1.08 for glass containers).   This is
because in the past flat glass has been made with twice the
amount of salt cake  (Na2S04) as container glass, and the
data in Table C-5 are primarily from 1972.  However, manu-
facturers have reduced the amount of sulfate in the batch
in order to lower emissions, and one producer now uses a
salt cake content close to that for container glass.

In a recent sampling test on a typical float glass furnace,
the SO  emission factor was found to be 0.5 g/kg.  Another
      H
producer (who apparently uses more salt cake) reported
emission factors of 1.35 g/kg to 7.1 g/kg.  It appears from
these data, therefore, that an average emission factor of
1.5 g/kg is justified.

c.   Particulates

Source test measurements of particulate emissions are given
in Table C-6.   Emission factors vary from 0.22 to 12.55 g/kg
(0.44 to 25.1 Ib/ton), with an average of 1.13 g/kg
(2.27 Ib/ton).   However, a number of high values come from
pressed and blown establishments which are known to make
specialty glasses.   (Plant V makes borosilicate glass and
plants AA and EE make lead glass.)   As a result, the average
emission factor for soda-lime glass has been calculated by
using data from flat glass plants and glass container
                               99

-------
                                    Table C-6.  SOURCE TEST DATA FROM NEDS  -

                               PARTICULATE EMISSIONS FROM GLASS MELTING FURNACES14
Plant
A
A
B
C
C
C
C
D
F
F
N
J
J
K
K
K
Production -
metric tons/year
(tons/year)
46,200
15,900
755,000
31,900
11,800
34,800
47,900
143,000
40,100
47,700
82,000
68,100
65,500
27,500
34,800
71,200
( 50,800)
( 17,500)
(830,000)
( 35,100)
( 13,000)
( 38,399)
( 52,700)
(157,000)
( 44,100)
( 52,500)
( 90,000)
( 74,900)
( 72,000)
( 30,300)
( 38,300)
( 78,300)
Emissions,
metric tons/year
(tons/year)
35.5
5.5
679
23.6
9.1
26.4
36.4
105
23.6
26.4
303
43.6
48.2
19.1
33.6
61.8
( 39)
( 6)
(747)
( 26)
( 10)
( 29)
( 40)
(116)
( 26)
(29)
(333)
( 48)
( 53)
( 21)
( 37)
( 68)
Emission factor,
g/kg
(Ib/ton)
0.77 (1.54)
0.35 (0.69)
0.90 (1.80)
0.74 (1.48)
0.77 (1.54)
0.75 (1.51)
0.76 (1.52)
0.74 (1.48)
0.59 (1.18)
0.55 (1.10)
3.70 (7.40)
0.64 (1.28)
0.73 (1.47)
0.70 (1.39)
0.97 (1.93)
0.87 (1.74)
Product
Glass containers
(two furnaces)
Glass containers
Glass containers
(four furnaces)
Glass containers
Glass containers
Glass containers
Glass containers
Flat glass
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
o
o

-------
Table C-6 (continued).  SOURCE TEST DATA FROM NEDS -
 PARTICULATE EMISSIONS FROM GLASS MELTING FURNACES14
Plant
L
L
M
P
R

S
u
u
u
u
V

V
V
Production,
metric tons/year
(tons/year)
41,100 ( 45,200)
30,100 ( 33,100)
66,200 ( 72,800)
179,000 (197,000)
78,400 ( 86,200)

58,800 ( 64,700)
74,600 ( 82,000)
66,400 ( 73,000)
70,000 ( 77,000)
76,400 ( 84,000)
7,950 ( 8,750)

7,950 ( 8,750)
9,550 ( 10,500)
Emissions,
metric tons/year
(tons/year)
35.5
22.7
36.4
179
38.2

35.5
47.3
45.5
90
107
50

99
75
( 39)
( 25)
( 40)
(197)
( 42)

( 39)
( 52)
( 50)
( 99)
(118)
( 55)

(110)
( 82)
Emission factor,
g/kg
(Ib/ton)
0.87 (1.73)
0.76 (1.51)
0.55 (1.10)
1.00 (2.00)
0.49 (0.97)

0.60 (1.21)
0.64 (1.27)
0.69 (1.37)
1.28 (2.57)
1.41 (2.81)
6.30 (12.60)

12.55 (25.10)
7.80 (15.60)
Product
Glass containers
Glass containers
Glass containers
(two furnaces)
Glass containers
(five furnaces)
Pressed and blown
glassware
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Pressed and blown
technical glass
Pressed and blown
technical glass
Pressed and blown
technical glass

-------
Table C-6 (continued).  SOURCE TEST DATA FROM NEDS  -
 PARTICULATE EMISSIONS FROM GLASS MELTING FURNACES11*
Plant
V
V
V
X
X
X
X
Y
Y
z
z
z
z
z
AA
Production ,
metric tons/year
(tons/year)
7,950 ( 8,750) /
15,900 ( 17,500)
3,980 ( 4,380)
60,500 ( 66,500)
57,300 ( 63,000)
45,000 ( 49,000)
50,900 ( 56,000)
40,300 ( 44,300).
20,800 ( 22,900)
73,000 ( 80,000)
33,600 ( 37,000)
51,800 ( 57,000)
14,400 ( 15,800)
34,600 ( 38,000)
17,600 ( 19,400)
Emissions,
metric tons/year
(tons/year)
93 (102)
32 ( 35)
50 ( 55)
17 ( 19)
22 ( 24)
16 ( 18)
27 ( 30)
27 ( 30)
11 ( 12)
57 ( 63)
8 ( 9)
25 ( 28)
14 ( 15)
26 ( 29)
14 ( 16)
Emission factor,
g/kg
(Ib/ton)
11.50 (23.00)
2.0 (4.00)
12.55 (25.10)
0.29 (0.57)
0.38 (0.76)
0.36 (0.73)
0.53 (1.07)
0.68 (1.35)
0.52 (1.05)
0.79 (1.58)
0.25 (0.49)
0.49 (0.98)
0.95 (1.90)
0.76 (1.53)
0.82 (1.64)
Product
Pressed and blown
technical glass
Pressed and blown
technical glass
Pressed and blown
technical glass
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Pressed and blown
glass

-------
                             Table C-6  (continued).   SOURCE  TEST  DATA FROM NEDS  -

                              PARTICULATE EMISSIONS FROM GLASS  MELTING FURNACES14
Plant
AA
BB
CC
EE
EE
FF
GG
HH
II
II
II
II
LL
Production,
• metric tons/year
(tons/year)
4,630 ( 5,090)
29,900 ( 32,900)
14,900 ( 16,400)
29,100 ( 32,000)
12,700 ( 14,000)
109,000 (120,000)
39,300 ( 41,000)
47,700 ( 52,500)
16,500 ( 18,100)
8,950 ( 9,850)
17,900 ( 19,700)
19,400 ( 21,300)
171,000 (188,000)
Emissions,
metric tons/year
(tons/year)
21 ( 23)
63 ( 69)
8 ( 9)
44 ( 49)
159 (175)
120 (132)
35 ( 38)
72 ( 79)
3.6 ( 4)
7 ( 8)
149 (164)
11 ( 12)
154 (169)
Emission factor,
g/kg
(Ib/ton)
4.52 (9.04)
2.10 (4.19)
0.55 (1.10)
1.53 (3.06)
12.50 (25.00)
1.1 (2.2)
0.93 (1.85)
1.51 (3.01)
0.22 (0.44)
0.81 (1.62)
8.3 (16.6)
0.57 (1.13)
0.90 (1.80)
Product
Pressed and blown
glass, lead glass
Pressed and blown
glass
Pressed and blown
glass
Pressed and blown
glass
Pressed and blown
glass
Glass containers
(four furnaces)
Glass containers
Pressed and blown
glass
Glass containers
Glass containers
Glass containers
Glass containers
Flat glass
(three furnaces)
o
U)

-------
                        Table C-6  (continued).  SOURCE  TEST DATA FROM NEDS -
                         PARTICULATE EMISSIONS FROM  GLASS  MELTING FURNACES14
Plant
JJ
JJ
JJ
KK
KK
MM
TOTAL
Production ,
metric tons/year
(tons/year)
73,400 ( 80,700)
99,100 (109,000)
35,400 ( 38,900)
64,900 ( 71,400)
56,000 ( 61,600)
173,000 (190,000)
3,932,400 (4,325,570)
Emissions,
metric tons/year
(tons/year)
34 ( 37)
25 ( 28)
22 ( 24)
112 (123)
. 105 (115)
398 (438)
4,470 (4,912)
Emission factor,
g/kg
(Ib/ton)
0.46 (0.92)
0.26 (0.51)
0.62 (1.23)
1.73 (3.45)
1.87 (3.73)
2.31 (4.61)
Product
Glass containers
Glass containers
Glass containers
Glass containers
Glass containers
Flat glass
1.14 (2.27) (average)
Glass Containers
             3,157,700  (3,473,450)
Flat Glass
               486,400    (535,000)

Pressed and Blown Glass
               288,300    (317,120)
2,330   (2,563)
1,315   (1,446)
  821     (903)
0.74  (1.48)  (average)
1.35  (2.70)  (average)
2.85  (5.70)  (average)

-------
plants.  For these plants both weight and number average
emission factors are 1.0 ± 0.6 g/kg (2.0 ± 1.2 Ib/ton) at a
95% confidence limit.  The standard deviation is ±2.3 g/kg
(±4.5 Ib/ton).

d.   Carbon Monoxide

Source test data on CO emissions are scarce because this is
not a major glass furnace emission.  The high combustion
temperature and the presence of excess air do not favor its
formation.

Data are reported in ppm and cannot be converted directly to
grams of CO per kilogram of glass because air flow rates are
not given.  A survey of NEDS showed that air flow rates
varied from 1.4 to 8.2 standard m3/kg  (50,000 to
300,000 scf/ton) of glass.  The wide range is due to  such
factors as furnace efficiency, electric boosting, percent
excess air, and use of ejector air in  the stack.  An  average
value of 5.5 standard m3 air/kg  (200,000 scf air/ton) glass
was used in the calculations because the furnaces tested
were using ejector air.

Two test reports were not used in finding the average CO
emission factor.  One was a furnace melting borosilicate
glass.47  The other had CO concentrations of 40 ppm and
375 ppm in the stack gas,33 and these values are 10 and
100 times higher than other data.  A value of 375 ppm is
equivalent to ^5 g/kg, which is inconsistent with good
combustion conditions.

The emissions data are summarized in Table C-7.  Because of
the unknown air flow rates the accuracies are only estimates.
(In addition, Reference 24 provided only a range of values
for several tests.)  A sample calculation follows:

                               105

-------
              Table C-7.  CO EMISSIONS FROM GLASS MELTING FURNACES
Reference
24
55
Average :
CO emissions
In stack gas,
ppm
0-5
0 (11 runs)
5 ( 7 runs)
10 ( 1 run )
2 . 5 ppm
Glass produced,
g/kg (Ib/ton)
0 - 0.04 (0 - 0.08)
0 (0)
0.04 (0.08)
0.08 (0.16)
0.02 g/kg ± 100%
Comments
Series of 16 tests on furnace
melting 105 - 213 metric tons/day
with and without carbon in the
batch (@^0.5 g/kg).
Tests on glass furnace
melting ^227 metric tons/day.
-
          Table C-8.  HYDROCARBON EMISSIONS FROM GLASS MELTING FURNACES
 Reference
                     Hydrocarbon emissions
  In stack gas,
       ppm
  Glass produced,
   g/kg (Ib/ton)
          Comments
    24
   0-5
0, - 2 (As hexane)
Average;
0 - 0.02 (0 - 0.04)
0 - 0.05 (0 - 0.1 )
                    0.04 g/kg ± 100%
Series of 16 tests on flint
glass furnace melting
105 - 213 metric tons/day
with and without carbon in the
batch (@^0.5 g/kg).
55Bartz, D. R., K. W. Arledge, J. E. Gabrielson, L. G. Hays, and S. C. Hunter.
  Control of Oxides of Nitrogen from Stationary Sources in the South Coast Air
  Basin (of California).  Prepared by KVB Engineering, Inc., for the Air Resources
  Board, Sacramento.  Report No. ARB-R-2-1471-74-31 (PB 237688).  September 1974.
  p. A-24.

-------
     For 5 ppm CO and air flow of 200,000 scf/ton:  this
     equals 1 scf CO/ton of glass.  Since 1 scf =. 28.32 liters
     and one gram mole of gas occupies 22.4 liters, this is
     equal to 1.264 gram moles of CO or 35.4 g or 0.08 Ib.
                   !
The average emission factor for CO is 0.02 g/kg ± 100%.  For
comparison an emission factor was also computed from other
data on gas-fired burners (see Table C-9).  The CO emission
factor is 9.45 ng/J (0.022 Ib/million Btu ± 55%).  Assuming a
fuel consumption of 7 MJ/kg (6 million Btu/ton) of glass
melted, this gives a factor of 0.066 g/kg ± 55%.

e.   Hydrocarbons

There has been only one report of testing for hydrocarbons
in the stack gas from the melting furnace (Table C-8).
Since the data are in ppm the same assumptions were made as
in the CO computations to find the emissions factor in g/kg.
Examples are:

     For 5 ppm CH^ and air flow of 6.24 standard m3/kg
     (200,000 scf/ton):  this equals 1 scf CH^/ton of glass.
     Since 1 scf = 28.32 liter and one gram mole of gas
     occupies 22.4 liters, this is equal to 1.264 gram moles
     of CH4 or 20.2 g  (0.044 Ib).

     For 2 ppm CgHltt and air flow of 6.24 standard m3/kg
     (200,000 scf/ton):  this equals 0.4 scf hexane/ton of
     glass.  Since 1 scf = 28.34 liters and one gram mole
     of gas occupies 2.2.4 liters, this is equal to 0.5056
     gram moles of hexane or 43.5 g/or 0.096 Ib.

The average emission factor for hydrocarbons is  0.04 g/kg  ±
100%.  For comparison  an emission factor was also  found
                              107

-------
 using  data  from gas-fired burners  (Table C-9).   The average
 hydrocarbon emission factor  from these is 6.0  ng/J (0.014
Ib/million  Btu  ± 144%).   Based on a fuel  usage of  7 MJ/kg
 (6  million  Btu/ton)  of glass melted,  the hydrocarbon
 emission factor will be 0.042 g/kg ±  144%.

 3.   FORMING AND FINISHING OPERATIONS - ANNEALING  LEHRS

 Since  no emissions measurements  have  been made on  gas-fired
 annealing lehrs, emission factors were estimated from other
 data on  gas combustion.   A modern recirculating  air type
                                 \,
 lehr consumes 11 to  17 m3/hr (400 cfh to 600 cfh)  when
 annealing 91 metric  tons (100 tons) of glass per day.  Lehrs
 of  the older design  consume  34 to 57  m3/hr (1,200  cfh to
 2,000  cfh).28  Using a worst case of  57 m3/hr  (2,000  cfh)
 gives  a  usage rate of 0.0062 m3/kg  (200 cubic  ft/ton)  of
 glass.   For a typical flat glass plant making  500  metric tons
 (550 tons)  of glass  per day  this amounts to 130  m3/hr
 (4,583 cfh).   With a heating value of 37.3 MJ/m3
 (1000  Btu/cubic ft)  this can be  converted to  1.34  MJ/sec
 (4,583,000  Btu/hr) or 0.23 MJ/kg (200,000 Btu/ton) of glass.

 A series of tests on four gas-fired burners gave the emission
 data shown  in Table  C-9.56  Converting these  on  a  basis of
 0.23 MJ/kg  (200,000  Btu/ton) of  glass annealed yields the
 emission factors in  Table C-10.
56Hangebrauck,  R.  P.,  D.  J.  Von Lehmden, and J. E. Meeker.
  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.                               —
                               108

-------
      Table C-9.   EMISSIONS FROM GAS-FIRED BURNERS 56
Test
15
17
18
19
Average and
95% confi-
dence limits
Standard
deviation
Emissions, ng/J (Ib per million Btu)
N0x
60.2
(0.14)
150
(0.35)
38.7
(0.09)
25.8
(0.06)
69
(0.16)
±113%
±47
(±0.11)
S°x
-
0
(0)
0
(0)
—
0
(0)
0
(0)
Partic-
ulates
9.0
(0.021)
2.58
(0.006)
3.0
(0.007)
11.2
(0.026)
5.2
(0.012)
±196%
±5.2
(±0.012)
CO
5.6
(0.013)
8.6
(0.020)
11.2
(0.026)
12.9
(0.030)
9.. 5
(0.022)
±55%
±2.75
(±0.0064)
Hydro-
carbons
i:29
(0.003)
—
9.5
(0.022)
6.9
(0.016)
6.0
(0.014)
±144%
±3.4
(±0.0079)
It is thought that the particulates are present in the
inlet air and not formed during the combustion process.
     Table C-10.  EMISSION FACTORS FOR ANNEALING LEHRS
Emission
NO
S0x
Particulates
CO
Hydrocarbons
Emission
mg/kg
16 ± 113%
0
1.2 ± 196%
2.2 ± 55%
1.4 ± 144%
factor,
(Ib/ton)
(0.032)
(0)
(0.0024)
(0.0044)
(0.0028)
                             109

-------
                         APPENDIX D
                 TOTAL FLAT GLASS EMISSIONS
Total flat glass emissions were compared on a state and
national basis to emissions from all stationary sources.
State emissions were calculated by assuming that state
production was proportional to the state capacities listed
in Appendix B.  The state production percentage was then
multiplied by the national emissions in Table 11 to give
state emissions for NO , SO , and particulates.  Values for
                      J\.    J\.
CO and hydrocarbons were not calculated because they were
all <0.01.

Table D-l gives the breakdown of NO , SO , and particulates
                                   Jv    ^C
by state.  Total state emissions were taken from the NEDS
inventory,57 which is shown in Table D-2.  Emission ratios
appear in Table D-l.  Another comparison was made using an
emissions data base generated by Monsanto Research -
Corporation,58 and this is given in Table D-3.
571972 National Emissions Report.  Environmental Protection
  Agency.  Research Triangle Park.  Publication No.
  EPA-450/2-74-012.  June 1974.  422 p.
58State-by-State Listing of Source Types that Exceed the
  Third Decision Criteria.  Special Project Report
  Prepared by Monsanto Research Corporation for the
  Environmental Protection Agency, Research Triangle Park.
  Contract 68-02-1874.  July 1975.  p.  1-3.
                              Ill

-------
                        Table D-l.   STATE BY STATE LISTING OF  EMISSIONS


State
Arkansas
California , :•
Illinois
Maryland
Michigan
Missouri
New Jersey
North Carolina
Ohio
Oklahoma
Pennsylvania
Tennessee
Texas
West Virginia
United States

Capacity,
metric tons/
day
200
824
720
360
1,180
807
450
680
1,900
1,060
2,235
2,439
900
540
14,290
Annual
production
%
1.4
5.8
5.0
2.5
8.2
5.6
3.2
4.7
13.3
7.4
15.6
17.0
6.3
3.8
100
1Q3
metric
tons
57
232
201
101
328
226
126
189
533
298
624
680
252
151
4,000

Emissions, metric tons
NOX
224
928
800
400
1,312
" 896
512
752
2,128
1,184
2,496
2,720
1,008
608
16,000
S0x
84
348
300
150
492
336
192
282
798
444
936
1,020
378
228
6,000
Partic-
ulates
56
232
200
100
328
224
128
188
532
296
624
680
252
152
4,000
NEDS % of
state emissions
NOX
0.1
<0.1
0.1
0.5
<0.1
0.2
0.1
0.2
0.2
0.5
0.1
0.6
0.1
0.5
0.07
S0x
0.2
0.1
<0.1
<0.1
<0.1
<0.1
<0.1
0.1
<0.1
0.3
<0.1
0.1
<0.1
<0.1
0.02
Partic-
ulates
<0.1
<0.1
<0.1
<0.1
<0.1
0.1
0.1
<0.1
<0.1
0.3
<0.1
0.2
<0.1
0.1
0.02
MRC % of
state emissions
NOX
0.3
0.1
0.1
0.2
0.2
0.3
0.2
0.2
0.3
0.9
0.3
1.0
0.1
0.2
0.14
S0x
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
0.3
<0.1
0.1
<0.1
<0.1
0.009
Partic-
ulates
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
0.003
I-1
I-1
NJ

-------
Table D-2.
NEDS EMISSION SUMMARY BY STATE57
    (metric tons)

Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Dist. Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
Particulates
1,178,643
13,913
72,685
137,817
1,006,452
201,166
40,074
36,808
19,451
226,460
404,574
61,621
55,499
1,143,027
748,405
216,493
348,351
546,214
380,551
49,155
494,921
96,160
'705,921
266,230
168,355
202,435
272,688
95,338
94,040
14,920
SO
X
882,731
5,874
1,679,768
39,923
393,326
49,188
168,068
209,310
60,630
897,381
472,418
45,981
54,387
2,043,020
2,050,541
283,416
86,974
1,202,827
166,664
144,887
420,037
636,466
1,466,935
391,633
50,591
1,152,373
871,235
58,014
304,851
86,596
NO
X
397,068
32,757
123,871
168,989
1,663,139
147,496
155,832
58,407
46,824
644,794
369,817
44,221
48,552
974,372
1,371,233
242,524
233,987
419,142
442,817
76,741
265,204
334,379
2,222,438
311,834
172,519
448,300
148,405
101,948
88,933
67,309
HC
643,410
28,389
189,981
195,538
2,160,710
193,456
219,661
63,886
41,789
619,872
458,010
89,530
84,230
1,825,913
600,477
316,617
309,633
326,265
1,919,662
122,918
295,867
440,481
717,891
410,674
195,950
413,130
271,824
127,821
53,673
88,469
CO
1,885,657
167,357
815,454
843,204
8,237,667
875,781
897,580
204,227
190,834
2,695,817
2,036,010
275,566
343,720
6,412,718
2,933,780
1,440,621
1,002,375
1,189,932
5,633,827
376,196
1,261,804
1; 682, 218
3,243,526
1,760,749
829,094
1,854,901
611,061
569,522
215,751
256,380
                      113

-------
Table D-2  (Continued)
  NEDS
(metric
EMISSION
 tons)
SUMMARY  BY STATE57

New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
U.S. TOTALS
Particulates
151,768'
102,785
160,044
481,017
78,978
1,766,056
93,595
169,449
1,S810,598
13,073
198,767
52,336
409,704
549,399
71,692
14,587
477,494
161,934
213,715
411,558
75,427
16,762,000
SO
X
463,736
444,310
345,979
473,020
78,537
2,980,333
" 130,705
36,776
2,929,137
65,761
247,833
17,354
1,179,982
753,098
152,526
17,751
447,394
272,991
678,348
712,393
69,394
28,873,000
NO
X
489,216
199,181
572,451
412,599
85,708
1,101,470
222,687
135,748
3,017,345
46,921
521,544
49,490
426,454
1,303,801
80,998
24,286
329,308
187,923
229,598
408,525
72,572
21,722,000
HC
819,482
152,057
1,262,206
447,238
70,289
1,153,493
341,358
234,669
891,763
65,833
907,833
90,478
362,928
2,218,891
98,282
41,980
369,416
344,643
116,155
523,930
55,319
23,994,000
CO
2,877,319
504,249
4,881,922
1,734,398
318,679
5,205,719
1,456,627
929,247
3,729,830
283,650
4,222,168
387,356
1,469,253
6,897,748
402,527
150,510
1,548,031
1,659,117
494,214
1,582,869
303,297
91,782,000
                 ADJUSTMENTS TO GRAND TOTAL
The United States summary does not include certain source categories.
The following additions should be considered part of the United States
grand total for a more accurate picture of nationwide emissions.
New York
pt. sources
311,000

Forest wild fires 375,000
Agricultural
burning
272,000

Structural fires 52,000
Coal refuse fires 100,000
Total
U.S. Subtotal
(above)
U.S. Grand
Total
1,110,000
16,762,000

17,872,000'

993,000

0
15,000

0
128,000
1,076,000
28,873,000

29,949,000

382,000

88,000
29,000

6,000
31,000
536,000
21,722,000

22,258,000

127,000

529., 000
.272,000

61,000
62,000
1,051,000
23,994,000

25,045,000


44,000

3,089,000
1,451,000

200,000
308,000
5,086,000
91,782,000

96,868,000

                              114

-------
 Table D-3.  STATE LISTING OF
EMISSIONS AS OF JULY 2, 197558

State
i ALABAMA
2 ALASKA
3 ARIZONA
4 ARKANSAS
5 CALIFORNIA
6 COLORADO
7 CONNECTICUT
8 DELAWARE
9 FLORIDA
10 GEORGIA
11' HAWAII
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
1.52000
16340000.0
12.50000
3265000.0
2.49000
1619000.0
1.24000
5675000.0
4.33000
3156000.0
2.41000
365600.0
0.27900
130200.0
0.09930
2430000.0
1.86000
2331000.0
1.76000
251200.0
0.19200
2430000.0
1.85000
3584000.0
2.74000
2202000.0
1.68000
2579000.0
1.97000
3358000.0
2.56000
1B54000.0
1.42000
S02
1228UOO.O
1.91000
222600.0
0.44700
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
2.54000
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
261600.0
2.27000
31990.0
0.27700
75100.0
0.6blOO
77310.0
0.67UOO
796800.0
6.91000
116800.0
1.01000
152200.0
1.32000
45720.0
0.39600
410300.0
3.56000
294200.0
2.55000
40790.0
0.35400
33220.0
0.28600'
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
iVe&oo.o
0.53200
171109.0
0.6471)0
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
400800.0
1.52000
742800.0
2.81000
274600.0
1.0400U
CO
372600.0
2.04000
472200.0
2.58000
178300.0
0,97600
225800.0
1.24000
1987000.0
10.90000
105800.0
0.57900
92690.0
0.50700
24560.0
0.13500
3502000.0
19.20000
705400.0
3.86000
84750.0
0.46400
518300.0
2.84000
412500,0
2.26000
182100.0
0.99700
90720.0
0.49700
174600.0
0.95600
2193UO.O
1.20000
                 115

-------
Table D-3 (continued).  STATE LISTING OF
     EMISSIONS AS OF JULY 2, 197558



State
18 LOUISIANA .

19 MAINE.

20 MARYLAND

21 MASSACHUSETTS

22 MICHIGAN

23 MINNLSOTA

24 MISSISSIPPI

25 MISSOURI

26 MONTANA

27 NEBRASKA

28 NEVADA

29 NEW HAMPSHIRE

30 NEW JERSEY

31 NEW MEXICO

t.
32 NEW YORK

33 N CAROLINA

34 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
2801000.0
2.14000
3056000.0
2.33000
1490000.0
1,14000
2839000.0
2.17000
4975000.0
3.80000
3049000.0
2.33000
3155000.0
2.41000
326500.0
0.24900
815800.0
0.62300
3548000.0
2.71000

2704000.0
2.0600U
2203000.0
1.6SOOO
2854000.0
2.18000
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
846800.0
1.32000
280300.0
0.43600
1259000.0
1.96000
177000.0
0.27500
137100.0
0.21300
263100.0
0.40900
325800.0
0,50700
2922000.0
4.55000
441400.0
0.68700

5137000.0
7,99000
2298000.0
i. 58000
328700.0
D',51100
4062000.0
6.32000
163400.0
0.25400
372500.0
0.57900

NO
X
21900U.O
1.900UO
54270.0
0.47000
2151UO.O
1.86000
322300.0 '
2.79000
54800U.O
4.7bOOO
185000.0
1.60000
87010.0
0.75400
287500.0
2.49000
34650.0
0.30000
50940.0
0.44200
58500.0
0.50700
36060.0
0.31300
323400.0
2.80000
109800.0
0.95200

721400.0
6.25000
338400.0
2.93000
61110.0
0.53000
785800.0
6.81000
130000.0
1.J.3000
62710.0
0.54400
Hydro-
carbons
1741000.0
6.58000
71970.0
0.27200
302300.0
1.14000
463100.0
1.75000
734000.0
2.7SOOO
388000.0
1.47000
350200.0
1.32000
588400.0
2.22000
174200.0
0.65800
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
73930.0
0.28000
1244000.0
4.70000 '
674700.0
2.55000
204800.0
0.774UO

CO
339900.0
4.60000
61430.0
0,33600
163400.0
0.89400
190400.0
1,04000
299400.0
1,64000
150700.0
0.82500
228200.0
1.25000
268500.0
1.47000
230500.0
1.26000
59590.0
0.32600
28700.0
0.15700
30200.0V
0.16500
281400.0
1.54000
49460.0
0.27100

551600.0
3.02000
371500.0
2.03000
22320.0
0.12200
'482700,0
2.64000
200800.0
1.10000
304900.0
1.67000
                     116

-------
Table D-3 (continued).  STATE LISTING OF
     EMISSIONS AS OF JULY 2, 1975
                                 58
State
3a PENNSYLVANIA
39 KHOOL ISLAND
HO S CAROLINA
41 S DAKOTA
42 TENNESSEE
43 TEXAS
44 UTAH
45 VERMONT
46 . VIRGINIA
47 WASHINGTON
48 M VIRGINIA
49 WISCONSIN
50 WYOMING
US TOTALS
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
3132000.0
2.39000
113200.0
0.06610
1209000.0
0.92300
2861000.0
2.18000
1769000.0
1.37000
9302000. U
7.10000
2461000.0
i.aaooo
292100.0
0,22300
1607000.0
1.23000
2204000.0
1.66000
1261000.0
0.96200
2160000,0
1.66000
2851000.0
2.18000
131000000.0
SO2
S603000.0
H. 72000
519900.0
0.80900
1076000.0
1.67000
69420.0
0.10800
1307000.0
2.03000
1817000.0
2.83000
285*00.0
0.44400
112600.0
0,17500
1368000.0
2.16000
626*00. 0
0.97500
1455000.0
2.26000
1216000.0
1.89000
513000.0
0.79800
64300000.0
NO
X
762200.0
6.7BOOO
3S760.0
0.33600
146300.0
1.270UO
18560.0
0.16100
264100.0
2,29000
695500.0
6.03UOO
48410.0
0.42UOO
13710.0
0.11900
197800.0
1.71000
126300.0
i 1.09000
306500.0
2.66000
231300.0
2.000UO
70570.0
0.61200
11500000.0
Hydro-
carbons
1331000.0
b. 03000
93730. U
U.354UO
260500.0
0.985UO
91110.0
0.34400
340900.0
1.29000
4139000.0
Ib. 60000
112800.0
0.42600
25460.0
0.0963U
415200.0
1.57000
361800.0
1.37000
172800.0
0.65300
362600.0
1.37000
275200.0
1.04000
26400000.0
CO
527000.0
2.88000
29390.0
0.16100
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
2.38000
161300.0
0.88300
20870.0
0,11400
18300000.0
                     117

-------
                         APPENDIX E

                    PLUME RISE CORRECTION


The Gaussian plume equation that is used to predict ground
level concentrations contains a factor called the effective
stack height, H.  This is equal to the physical stack height
(h) plus the amount of plume rise (AH).

                         H = h + AH

An exhaust plume rises before dispersal due to its exit
velocity and temperature.  In the case of glass furnaces
this is a significant effect (AH/h > 50%).

Plume rise can be estimated from the Holland formula40

          AH = -S-i.  (1.5 + 2.68 x 10~3p
where   V  = stack gas exit velocity, m/sec
         S
           = inside stack diameter, m
         u = wind speed, m/sec

         p = atmospheric pressure, mb
        T  = stack gas temperature, °K
         S
        T  = ambient temperature, °K
         3.    .        .     • *
                              119

-------
Under C class stability conditions AH is increased by a
correction factor of 1.10.

Table E-l presents stack gas parameters for 12 flat glass
furnaces and the corresponding plume rise.  The variation
in stack gas temperature and exit velocity is partly caused
by the use of ejection air.  In the calculations the ambient
temperature was taken to be 294°K (70°F),  the wind speed
4.5 m/sec, and the pressure 1,013 mb (1 atmosphere).


     Table E-l.  PLUME RISE FOR FLAT GLASS FURNACESlk'3
Stack
1
2
3
4
5
6
7
8
9
10
11
12
Height,
m
83.8
51.8
64.9
64.3
64.3
76.2
34.4
67.1
61.3
61.0
28.8
27.4
Diameter,
m
2.44
2.74
2.44
2.44
2.44
2.44
2.74
2.13
2.29
2.59
2.54
1.78
V
m/sec
16.64
2.40
12.17
9.99
10.62
9.10
14.33
4.04
10.89
8.06
14.3
15.0
V
°K
942
589
858
778
797
753
450
781
783
756
438
598
Average
AH,
m
60.1
8.4
42.5
33.5
36.0
30.0
39.1
10.7
32.8
29.6
33.5
25.7
30
 Additional data supplied by a flat glass manufacturer.
                              120

-------
                         APPENDIX F

           DERIVATION OF SOURCE SEVERITY EQUATIONS
             (T. R. Blackwood and E. C. Eimutis)
1.
SUMMARY OF SEVERITY EQUATIONS
The maximum severity of pollutants may be calculated using
the mass emission rate, Q, the height of the emissions, H,
and the ambient air quality standard  (AAQS).  The equations
summarized in Table F-l are developed in detail in this
appendix.

          Table F-l.   POLLUTANT SEVERITY EQUATIONS
             Pollutant
       For elevated sources:
            Particulate

            SO
              x
            HC
            CO
                               Severity equation
                                     70 Q
                                      H2
                                     50 Q
                                      H2

                                     315 Q
                                     H2-l
                                     162 Q
                                       H2
                                     0.78 Q
                                      H2
                              121

-------
2.   DERIVATION OF x    FOR USE WITH U.S. AVERAGE CONDITIONS
                    max
                                                   /

The most widely accepted formula for predicting downwind

ground level concentrations from a point source is:'f°
where   x = downwind ground level concentration at reference
            coordinate x and y with emission height of
            H, g/m3
        Q = mass emission rate, g/s

       a  = standard deviation of horizontal dispersion, m

       a  = standard deviation of vertical dispersion, m
        £*
        u = wind speed, m/s
        y = horizontal distance from centerline of dis-
            persion, m
        H = height of emission release  (effective stack
            height), m

        x = downwind dispersion distance from source of
            emission release, m
        TT = 3.14
We assume that x     occurs when x»0 and y = 0.  For a given
                max
stability class,  standard deviations of horizontal and vertical
dispersion have often been expressed as a function of down-
wind distance by power law relationships as follows:59


        -                  gy = axb                      (F-2)


                        a  = cxd + f                    (F-3)
                         z
59Martin, D. 0., and J. A. Tikvart, A. General Atmospheric
  Diffusion Model for Estimating the Effects of Air Quality
  of One or More Sources.  (Presented at 61st Annual Meeting
  of the Air Pollution Control Association, for NAPCA, St.
  Paul, 1968.)  18 p.


                               122

-------
Values for a, b, c, d and  f  are  given  in Tables F-2 and F-3,
Substituting these general equations into Eq.  F-l yields:

                        Q
                    b+d ,_    f I
               ac-rrux    + airufx
 Assuming that x    occurs at x<100m or  the  stability class
                ITlclX
 is C, then f = 0 and Equation F-4 becomes:
                   v =    *=	 e>vn                      fF-R^
                   X        i^_i_,q exP    _  „ j,              \.r  31
                       acimx
 For convenience, let:
                     AR =  -Q_ and BR  =|
                           aciru          I 2c-
 so that Equation F-5 reduces to:
                    X = ARx-(b+d) exp|^H               (F-6)
                         K            i.i
     Table F-2.  VALUES OF a FOR THE COMPUTATION OF a a'60
                Stability class
                       A  ,
                       B
                       C
                       D
                       E
                       F
0.3658
0.2751
0.2089
0.1471
0.1046
0.0722
                 For the equation

                           a  = axb
                           ay

                where  x = downwind distance
                       b = 0.9031
 60Tadmor,  J.  and Y.  Gur.  Analytical Expressions  for  the
   Vertical and Lateral Dispersion Coefficients  in Atmospheric
   Diffusion.   Atmospheric Environment, _3_:688-689, 1969.


                              123

-------
         Table  F-3.  VALUES  OF  THE  CONSTANTS  USED TO
               ESTIMATE VERTICAL  DISPERSIONS3'59
Usable range
>1, 000 m






100-1,000 m





<100 m





Stability
class
A
B
C
D
E
F

A
B
C
D
E
F
A
B
C
D
E
F
Coefficient
ci
0.00024
0.055
0.113
1.26
6.73
18.05
C2
0.0015
0.028
0.113
0.222
0.211
0.086
0.192
0.156
0.116
0.079
0.063
0.053
a,
2.094
1.098
0.911
0.516
0.305
0.18
d2
1.941
1.149
0.911
0.725
0.678
0.74
0.936
0.922
0.905
0.881
0.871
0.814
fl
-9.6
2.0
0.0
-13
-34
-48.6
£2
9.27
3.3
0.0
-1.7
-1.3
-0.35






For the equation:
                          = ex
                             124

-------
Taking the first derivative of Equation F-6

           -b-d'/     F_  -2dl\ / ,,_  -2d-l
     »
   = A
      R
          X
                            l\  /
                            Jj  (-

                        + exp  [BRx-2d]   (-b-d)  x-"-*-1!   (F-7)
and setting this equal to  zero  (to  determine  the  roots which
give the minimum and maximum conditions  of  X  with respect
to x) yields:
         = o = A
      dx   ^   AR
                       -1 (exp[^BRx~2d]j  [-2dBRx~2d-b-d]    (F-8)
Since we define that x ^0 or °° at  x    /  the  following ex-
                                   max
pression must be equal to 0:

                     -2dBRx~2d-d-b =0                     (F-9)

or                    (b+d)x2d = -2dBR

or                 x2d = -^T. = ^±^_                 (F_ll)

                   c2(b+d)

                  x = /-AJi^\2d
                      \c2(b+d)/

Thus Equations F-2 and F-3 become:
                             ^c2(d+b)
                                            1
                        c2(b+d)/     \b+d
                               125

-------
 The  maximum will be determined for U.S.  average conditions
 of stability.   According to Slade61,  this is when a  = a .


 Since b = 0.9031,  and upon inspection of Table F-2 under
 U.S.  average conditions, a  = a ,  it  can be seen that
 0.881
-------
3.   DEVELOPMENT OF SOURCE SEVERITY EQUATIONS


The general source severity, S, relationship has  been  defined

as follows:

                            = xmax                      (5
                              AAQS


where  x"    = average maximum ground  level  concentration
        in 3.x
       AAQS = ambient air quality standard


As mentioned in the main text, values of X™^ are  found  from
the equation
                                  t,
                                      0.17
                     Xmax = Xmax  IT                    (F~22)
where t0 is the "instantaneous"  (i.e.  3 minute)  averaging
time and t is the averaging time used  for  the  ambient  air
quality standard.  These are given in  Table F-4.
(1)  CO Severity - The primary standard  for  CO  is  reported
for a 1-hr averaging time.  Therefore,


                         t = 60 min
                                       17
                                                        (F~23)
                      -  2 Q  ' J x                      (F-24)
                                               (0.6)     (F-25)
                        ireuH2


                                2 Q
                         (3.14) (2.72) (4.5) H2
                              127

-------
        Table F-4.  SUMMARY OF NATIONAL AMBIENT AIR
                    QUALITY STANDARDS62
Pollutant
Particulate
matter

Sulfur oxides


Carbon
monoxide
-
Nitrogen
dioxide
Photochemical
oxidants
Hydrocarbons
(nonme thane)
Averaging
time
Annual (Geometric
mean)
24-hourb
Annual (arith-
metic mean)
24-hourb
3-hourb
8-hour
l-hourb
Annual (arith-
metic mean)
l-hourb
3-hour
(6 to 9 a.m.)
Primary
standards
75 yg/m3
260 yg/m3
80 yg/m3
(0.03 ppm)
365 yg/m3
(0.14 ppm)
-
10 mg/m3
(9 ppm)
40 mg/m3
(35 ppm)
100 yg/m3
(0.05 ppm)
160 yg/m3
(0.08 ppm)
160 yg/m3
(0.24 ppm)
Secondary
standards
60a yg/m3
150 yg/m3
60 yg/m3
(0.02 ppm)
260° yg/m3
(0.1 ppm)
1300 yg/m3
(0.5 ppm)

(Same as
primary)
(Same as
primary)
(Same as
primary)
(Same as
primary)
aThe secondary annual, standard (60 yg/m3)  is a guide for
 assessing implementation plans to achieve the 24-hour
 secondary standard.

 Not to be exceeded more than once per year.

CThe secondary annual standard (260 yg/m3)  is a guide  for
 assessing implementation plans to achieve the annual
 standard.


52Code of Federal  Regulations, Title 42 - Public Health,
  Chapter IV - Environmental Protection Agency,  Part 410
  National Primary and Secondary Ambient Air Quality
  Standards, April 28, 1971.  16 p.
                             128

-------
                           H2


                        _  (3.12 x

                    max         2
Substituting the primary standard for CO  (0.04 g/m3)  into
the equation for S then gives:
                 s = Xmax  =  (3.12 x 102)Q
                     AAQS       0.04 H2
or
                        S   - °-78
                        SCO	77
                                n
(2)  Hydrocarbon Severity - The primary standard  for hydro-
carbon is reported for a 3-hr averaging time.

                         t = 180 min

                                 3 \ °-17
                         (0.5) (0.052) Q                  (p_32)
                              H2

                             129

-------
For hydrocarbons, AAQS = 1.6 x lO"4 g/m3
and
                  S =       = °-026 Q	             (F-34)
                      AAQS    1.6 x IQ-^H2
or
                       SHC
                               H2
(3)  Particulate Severity - The primary  standard  for
particulate is reported for a  24-hr averaging  time.
                   -                    -17
                   *max = *max                          (F~36)
                     =  (0.052) Q  (0.35)                 (F-37)
                          H2
                     X
                     xmax
                          =  (0.0182) Q                  (F-38)
For particulates, AAQS = 2.6 x  lO"4 g/m
                                        3
                  s _ xmax  _    0.0182  Q                (F-39)
                      AAQS    2.6 x  10"1* H2
                         Sp =   -                       (F-40)
                          P    H2
(4)  SO   Severity - The primary standard  for  SO   is
     -™- X         r~                              X
reported for a 24-hr averaging time.
                              130

-------
                      max                              (F-4D
                      max       2
The primary standard is 3.65 x lO"4 g/m3.





and
                     xmax  _   (0.0182)  Q              (F-42)


                     AAQS    3.65 x 10"1* H2
or
                        sso  =                         (p_43)

                         S°x    H2
(5)   NO  Severity - Since NO  has a primary standard  with  a


1-yr averaging time,  the x    correction equation  cannot be
                          max

used.  As an alternative, the following equation was  selected:






                                                       (F-44)
                                f- i /
                                L  2U
A difficulty arises, however,  because a distance x,  from


emission point to receptor,  is included and hence,  the


following rationale is used:





The equation x    = 	

              max
is valid for neutral conditions or when a  =a .   This
                                         z   y

maximum occurs when
                          H  = /2"a
                                  z
                             131

-------
and since, under these conditions,
                          a  = ax
                           z
then the distance x    where the maximum concentration occurs


is:                                                       •
                        max
For class C conditions,
                          a = 0.113



                          b = 0.911
Simplifying Equation F-44
since
                    a  = 0.113 x   0.911
                     z          max
and
                     u = 4.5 m/sec
Letting  x = x    in Equation F-44,
                      max
                                   r_ i
                                   L  2

                                 . 098
                                0.16
                                                        (F-46)
                      x    = 7.5 H1-098
                       max
                                                        (F-47)
                              132

-------
and              -AQ- = 	i-S	             (F-48)
Therefore
Therefore:
              x
               max
1.911   (7,5 a1'098)1-911




  = 0^1_Qexr_ 1/JL\21           (F_49)

      H2- 1      L    \ az / J





    X  = 0.113X°-911                  (F-50)
     z
                  a  =  0.113 (7.5 H 1-1)0.911            (F-51)
                          a  = 0.71 H                    (F-52)
                           z
                       0.085 Q     I   1 /   H    \ 2-|
                X_  =——--*  exp - =•(	-}   \      (F-53)

                        H2-1       L    V 0.71 H
                      = 0^085_Q  ((K;m)                 (F_54)


                          H2.1
                      X... = 3-152V0"2 Q               (F-55)
                               -H2. 1
Since the NO  standard  is  1.0 x 10-It g/m3, the NO  severity
            x                          .           x

equation is:




                         -  (3.15 x IP"2)  Q                  ^
                    OMO   -  	               (F-56)

                      x    1  x 10-* H2-l
                         S     - 315 Q                    (v  z-7)
                         SNO   ~ "TT"                    (F~57)
                           x    &•• *•
                              133

-------
4.   AFFECTED POPULATION CALCULATION

Another form of the plume dispersion equation is needed to
calculate the affected population since the population is
assumed to be distributed uniformly around the source.  If
the wind directions are taken to 16 points and it is assumed
that the wind directions within each sector are distributed
randomly over a period of a month or a season, it can be
assumed that the effluent is uniformly distributed in the
horizontal within the sector.  The appropriate equation for
average concentration ()T) is then:40
To find the distances at which X/AAQS = 0.1, roots are
determined for the following equation:
             0 =
                   2.03 Q
AAQSa ux exp
- 0.1      (F-59)
keeping in mind that:
                       a  = a x   + c
where a, b, and c are functions of atmospheric stability
and are assumed to be for stability Class C.

Since equation F-59 is a transcendental equation the roots
are found by an iterative technique using the computer.

For a specified emission from a typical source, x/AAQS as
a function of distance might look as follows:
                              134

-------
The affected population is then in the  area
                               2     2
                       A = Tr(X2 - Xi
(F-60)
If the affected population density  is  D   then  the  total
                                       ir
affected population P is
                      P = D A  (persons)
(F-61)
                             135

-------
                         SECTION IX

                      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 stored,
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.

DEBITEUSE - A rectangular refractory collar with a slot,
placed in the melting tank.  Sheet glass is drawn through it,

FINING - Process of removing gas bubbles from molten glass.

FLOAT PROCESS - Process for making flat glass by floating
molten glass on a bath of molten tin until the glass hardens,

FLUX - Agent which promotes melting by reacting with silica
to form lower melting compounds.

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 20-30 min. the flow is reversed.

                               137

-------
                           SECTION X

                     CONVERSION FACTORS63
  To convert from
degree Celsius  (°C)
degree Kelvin  (°K)
joule  (J)
kilogram  (kg)

kilogram  (kg)

kilometer2  (km2)
meter  (m)
meter  (m)
meter2 (m2)
meter3 (m3)
metric ton
pascal (Pa)
pascal (Pa)
second (s)
         to
degree Fahrenheit
degree Celsius
British thermal unit
pound-mass  (Ib mass)
  avoirdupois)
ton (short, 2000
  Ib mass)
mile2
foot
mile
foot2
foot3
pound
bar
inch of water (60°F)
minute
   Multiply by
t°p = 1.8 t°c + 32
t°c = t°K - 273.15
9.479 x 10-4
2.204

1.102 x 10~3

2.591
3.281
6.215 x I0~k
1.076 x 101
3.531 x 101
2.205 x 103
1.000 x 10~5
4.019 x 10~3
1.667 x 10~2
63Metric Practices Guide.  American Society for Testing and
  and Materials.  Philadelphia.  ASTM Designation:  E380-74.
  November 1974.  34 p.
                               139

-------
                       METRIC PREFIXES
                     Multiplication
Prefix     Symbol        factor                Example
 giga        G             109          5 GJ = 5 x 109 joules
 mega        M         .    106          5 MJ = 5 x 106 joules
 kilo        k             103          5 kg = 5 x 103 grams
 milli       m             10~3         5 mb = 5 x 10~3 bar
 micro       y             10~6         5 ym = 5 x 10~6 meter
                              140

-------
                         SECTION XI

                         REFERENCES
1.    Standard Industrial Classification Manual, 1972
     Edition.  Washington.   Superintendent of Documents,
     1972.   p.  136-138.

2.    Preliminary Report, 1972 Census of Manufactures,
     Industry Series, Flat Glass, SIC 3211.  U.S. Department
     of Commerce.  Washington.  MC 72(P)-32A-1.  January
     1974.   7 p.

3.    Preliminary Report, 1972 Census of Manufactures,
     Industry Series, Glass Containers, SIC 3221.  U.S.
     Department of Commerce.  Washington.  MC 72(P)-32A-2.
     December 1973.  6 p.

4.    Preliminary Report, 1972 Census of Manufactures,
     Industry Series, Pressed and Blown Glass, SIC 3229.
     U.S. Department of Commerce.  Washington.  MC 72(P)-
     32A-3   February 1974.  7 p.

5.    Preliminary Report, 1972 Census of Manufactures,
     Industry Series, Products of Purchased Glass, SIC 3231,
     U.S. Department of Commerce.  Washington.  MC 72(P)-32A-4
     February 1974.  1 p.

6.    Directory Issue.  The Glass Industry.  54 (10);1-178,
     September 1973.

7.    1974 Glass Factory Directory Issue.  American Glass
     Review.  9£(8A):1-204, February 1974.

8.    Dietz, E.  D.  Glass.  In:  Chemical and Process
     Technology Encyclopedia, Considine, D. M. (ed.).
     New York,  McGraw-Hill Book Co., 1974.  p. 552-561.

9.    Hutchins,  J. R., and R. V. Harrington.  Glass.  In:
     Kirk-Othmer Encyclopedia of Chemical Technology, 2nd
     Edition, Vol. 10, Standen, A.  (ed.).  New York, Inter-
     science Publishers, Divn. of John Wiley & Sons, Inc.,
     1966.   p.  533-604.

                                141

-------
10.  Shreve, R. N.  Chemical Process  Industries,  3rd Edition.
     New York, McGraw-Hill Book Co.,  1967.  p.  190-210.

11.  Holscher, H. H.  The Glass Primer.  New York, Maga-
     zines for Industry, Inc., 1972.   58 p.

12.  Danielson, J. A.  Air Pollution  Engineering  Manual,
     2nd Edition.  Environmental Protection Agency.  Research
     Triangle Park.  Publication No.  AP-40.  May  1973.
     p. 765-782.

13.  A Screening Study to Develop Background Information  to
     Determine the Significance of Glass Manufacturing.
     Prepared by The Research Triangle Park Institute  for
     the Environmental Protection Agency.  Research Triangle
     Park.  Contract 68-02-0607, Task 3.  December 1972.

14.  Point Source Listing for Glass,  SCC 3-05-014, National
     Emission Data System.  Environmental Protection Agency.
     Research Triangle Park.  May 1974.

15.  Svec, J. J.  LOF Operates World's Largest  Glass Furnace.
     Ceramic industry.  103(2):30-32,  August 1974.

16.  Svec, J. J.  Double Float Glass  Line Produces 300 Million
     Square Feet.  Ceramic Industry.   100;66-69,  April 1973.

17.  Ford Motor Controls Glass Batch  by Chemical  Wetting.
     Ceramic Industry.  102;28-30, March 1974.

18.  Svec, J. J.  Float Plant a Showcase at Pilkington.
     Ceramic Industry.  101(6) :34-36,  Dec-ember  1-973.

19.  Allen, A. C.  New Canadian Plant Draws 14  Miles of
     Sheet Glass Per Day.  Ceramic Industry.  91(6);52-54,
     December 1968.

20.  Allen, A. C.  Canada Builds First Float Glass Plant.
     Ceramic Industry.  £9(6):43-45,  December 1967.

•21.  Allen, A. C.  One of the World's Largest Glass Tanks
     on Stream.  Ceramic Industry.  89_(4) : 50-51,  October  1967.

22.  Svec, J. J.  Pilkington Manufacturers 2.3  mm Float
     Glass.  Ceramic Industry.  103 (1);36-37, July 1974.

23.  Schorr, J. R., and G. A. Anderson.  Final  Report  on
     Industrial Energy Study of the Glass Industry.  Pre-
     pared by Battelle for the Federal Energy Administration
     and the Department of Commerce.   Washington.  Contract
     14-01-0001-1667.  December 1974.  p. 13-16,36,


                                142

-------
24.   Ryder, R.  J.,  and J.  J. McMackin.  Some Factors
     Affecting Stack Emissions from a Glass Container
     Furnace.   The Glass Industry.  5_0: 307-310, June 1969;
     346-350,  July 1969.

25.   Arrandale, R.  S.  Pollution Control in Fuel-Fired Tanks.
     The Glass Industry.  5_5_: 12-13,21, August 1974; 16-17,27,
     September 1974.

26.   Development Document for Effluent Limitations Guidelines
     and New Source Performance Standards for the Flat Glass
     Segment of the Glass Manufacturing Point Source Category.
     U.S. Environmental Protection Agency.  Washington.
     EPA-44011-74-001-C.  January 1974.  p. 44.

27.   Fuller, R. A.   Recirculating Lehr for Annealing Glass-
     ware.  Ceramic Bulletin.  £8:1065-1068, November 1969.

28.   Roos, P.  W.  Lehr Priority:  Design Concepts to Save
     Energy.  The Glass Industry.  5_6_: 18-22, April 1975.

29.   Control Techniques for Nitrogen Oxide Emissions from
     Stationary Sources.  U.S. Department of Health, Educa-
     tion and Welfare.  Washington.  NAPCA Publication No.
     AP-67  (PB 190265).  March 1970.  115 p.

30.   Davis, R. E.,  W. H. Manring, and W. C. Bauer.  Carryover
     Studies in Glass Furnaces.  In:  Collected Papers
     from the 34th Annual Conference on Glass Problems.
     Dept. of Ceramic Engineering, University of Illinois,
     Urbana, November 1973.  p. 109-126.

31.   Reed, R.  H.  Combustion Pollution in the Glass Industry.
     The Glass Industry.  5_£: 24 , 26 ,38 , April 1973.

32.   Mills, H. N.,  and J. Jasinski.  Evaluating Batch
     Changes.   The Glass Industry.  5_1: 223-227, May 1970.

33.   Stockham, J.  D.  The Composition of Glass Furnace
     Emissions.  Journal of the Air Pollution Control
     Association.   2^:713-715, November 1971.

34.   Arrandale, R.  S.  Air Pollution Control in Glass
     Melting.   Symposium Sur La Fusion du Verre, Brussels.
     October 1968.   p. 619-644.

35.   Custer, W. W.   Electrostatic Cleaning of Emissions
     from Lead, Borosilicate, and Soda-Lime Glass Furnaces.
     United McGill Corp.   (Presented at the 35th Annual
     Conference on Glass Problems.  Ohio State University,
     Columbus, November 14-15, 1974.)  13 p.


                               143

-------
36.  Compilation of Air Pollution Emission Factors, Second
     Edition.  Environmental Protection Agency.  Washington.
     Publication No. AP-42.  April 1973.  p. 8.13-1.

37.  1974 Annual Raw Material Processing Handbook.  Ceramic
     Industry.  102:97, January 1974.

38.  TLVs® Threshold Limit Values for Chemical Substances
     and Physical Agents in the Workroom Environment with
     Intended Changes for 1973.  American Conference of
     Governmental Industrial Hygienists.  Cincinnati.
     1973.  94 p.

39.  Code of Federal Regulations, Title 42 - Public Health,
     Chapter IV - Environmental Protection Agency, Part 410 •
     National Primary and Secondary Ambient Air Quality
     Standards, April 28, 1971.  16 p.

40.  Turner, D. B.  Workbook of Atmospheric Dispersion
     Estimates, 1970 Revision.  U.S. Department of Health,
     Education and Welfare.  Cincinnati.  Public Health
     Service Publication No. 999-AP-26.  May 1970.  84 p.

41.  Brown, C. J.  Selection Criteria for Sand, Dolomite,
     and Limestone in the Flat Glass Industry.  In:
     Collected Papers from the 32nd Annual Conference on
     Glass Problems.  Dept. of Ceramic Engineering, Univer-
     sity of Illinois, Urbana, November 1971.  p. 163-171.

42.  Simon, H., and J. E. Williamson.  Control of Fine
     Particulates from Continuous Melting Regenerative
     Container Glass Furnaces.  Los Angeles County Air
     Pollution Control District.  (Presented at the 68th
     Annual Meeting of the Air Pollution Control Associa-
     tion.  Boston.   June 15-20, 1975.)  12 p.

43.  Hamilton, J. C.  Applied Research in Glass Melting.
     Industrial and Engineering Chemistry.  62;16-21,
     February 1970.

44.  Edmondson, J. N., L. Reitz, R.  L. Weise, and J. Fraas.
     Design, Installation, and Operation of Equipment to
     Cool and Filter Particulate Matter from Flue Gas from
     a Regenerative Furnace.  In:  Collected Papers from the
     32nd Annual Conference on Glass Problems.  Dept. of
     Ceramic Engineering, University of Illinois, Urbana,
     November 1971.   p. 39-54.
                               144

-------
45.   Teller,  A.  J.   Control of Emissions from Glass Manu-
     facture.   Ceramic Bulletin.   5_1: 637-640, August 1972.

46.   Frantz,  C.  N.,  D. L.  Miser,  H.  N.  Troy,  and E. D.  Stobbe.
     Glass Furnace  Particulate Emission Control Equipment.   In:
     Collected Papers from the 32nd  Annual Conference on Glass
     Problems.  Dept. of Ceramic  Engineering, University of
     Illinois, Urbana, November 1971.   p.  25-38.

47.   Keller,  G.   Scrubber System  Lightens Load of Glass
     Furnace  Emissions.  Chemical Processing.  38;9,
     January  1975.

48.   Symposium on Pollution, Stratford-Upon-Avon, 30 May -
     1 June,  1973.   In:  Glass Technology.  14 (6);140-144,
     December 1973.

49.   Tank Emissions "Bagged."  The Glass Industry.   55:18,
     July 1974.

50.   Moyer, T.,  S.  Reigel, and C. Doyle.  Gas-Assisted
     Atomizers Help End High Heat Problem in Collector.
     Maintenance Engineering.  2^:28-29, June 1972.

51.   Child, F. S.  The Impact of  Flat Glass Imports.  The
     Glass Industry.  52^166-169, May 1971.

52.   Ceramic  Industry Newsletter.  Ceramic Industry.  104;9,
     May 1975.

53.   Wells, J. R.  Feldspar, Nepheline Syenite, and Aplite.
     In:  Minerals  Yearbook 1972, Volume I:  Metals, Minerals
     and Fuels.   Bureau of Mines, Washington.  1974.  p. 515-523,

54.   Ceramic  Industry Newsletter.  Ceramic Industry.  104:7,
     February 1975.

55.   Bartz, D. R.,  K. W. Arledge, J. E. Gabrielson, L.  G.
     Hays, and S. C. Hunter.  Control of Oxides of Nitrogen
     from Stationary Sources in the South Coast Air Basin
     (of California).  Prepared by KVB Engineering, Inc., for
     the Air  Resources Board, Sacramento.   Report No. ARB-R-
     2-1471-74-31 (PB 237688).  September 1974.  p. A-24.

56.   Hangebrauck, R. P., D. J. Von Lehmden, and J.  E. Meeker.
     Emissions of Polynuclear Hydrocarbons and other Pollu-
     tants from Heat-Generation and Incineration Processes.
     Journal  of the Air Pollution Control Association
     14:267-278, July 1964.
                                 145

-------
57.  1972 National Emissions Report.  Environmental Protection
     Agency.  Research Triangle Park.  Publication No.
     EPA-450/2-74-012.  June 1974.  422 p.

58.  State-by-State Listing of Source Types that Exceed the
     Third Decision Criteria.  Special Project Report
     Prepared by Monsanto Research Corporation for the
     Environmental Protection Agency, Research Triangle
     Park.  Contract 68-02-1874.  July 1975.  p. 1-3.

59.  Martin, D. 0., and J. A. Tikvart.  A General Atmospheric
     Diffusion Model for Estimating the Effects on Air Quality
     of One or More Sources.  (Presented at 61st Annual
     Meeting of the Air Pollution Control Association, for
     NAPCA, St. Paul, 1968.)  18 p.

60.  Tadmor, J. and Y. Gur.  Analytical Expressions for the
     Vertical and Lateral Dispersion Coefficients in Atmo-
     spheric Diffusion.  Atmospheric Environment.  3_: 688-689,
     1969.

61.  Gifford, F. A., Jr.  An Outline of Theories of Diffusion
     in the Lower Layers of the Atmosphere.  In:  Meteorology
     and Atomic Energy 1968, Chapter 3, Slade, D. A.  (ed.).
     Oak Ridge, Tennessee, U.S.  Atomic Energy Commission
     Technical Information Center.  Publication No. TID-24190.
     July 1968.  p. 113.

62.  Code of Federal Regulations, Title 42 - Public Health,
     Chapter IV - Environmental Protection Agency, Part 410
     National Primary and Secondary Ambient Air Quality
     Standards, April 28, 1971.   16 p.

63.  Metric Practice Guide, E 380-74.  American Society for
     Testing and Materials.  Philadelphia, November 1974.
     34 p.
                              146

-------
                                TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
 1. REPORT NO.
 EPA-600/2-76-032b
                           2.
                                  3. RECIPIENT'S ACCESSION1 NO.
4. TITLE A.NQSUBTITLE
Source Assessment: Flat Glass Manufacturing Plants
                                  5. REPORT DATE
                                   March 1976
                                                       6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

 Richard B. Reznik
                                  8. PERFORMING ORGANIZATION REPORT NO.

                                    MRC-DA-507
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Monsanto Research Corporation
 Dayton Laboratory
 Dayton, Ohio  45407
                                   10. PROGRAM ELEMENT NO.
                                   1AB015; ROAP 21AXM-071
                                   11. CONTRACT/GRANT NO.
                                   68-02-1874
 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 <
                                   Task Final; 7/74-12/75
                                                         COVERED
                                   14. SPONSORING AGENCY CODE
                                    EPA-ORD
 15. SUPPLEMENTARY NOTES E PA-650/2-75-019a is the first
 officer is D. A. Denny,  Mail Drop 62, Ext 2547.
                                report in this series. Project
 16. ABSTRACT
          The report describes air pollutants emitted during the production of flat
 glass, SIC No.  3211.  It covers raw materials preparation at the plant site, molten
 glass production in the melting furnace, and the forming of flat glass products.
 Melting  furnace emissions account for over 99% of the  total plant emissions; NOx,
 SOx, and particulates are the major (>99%) pollutants. The particulates are
 alkali sulfates of submicron size. NOx has the highest emission factor (4 g/kg) and
 annual emissions  (16,000 metric  tons).  When national emissions of each pollutant
 from this  industry are compared to the corresponding  national emissions from all
 stationary sources, NOx contributes 0.07% of the total.  Source severity is a measure
 of the potential environmental effect of air emissions from this industry: it is defined
 as the ratio of the  maximum  average ground level concentration compared to the
 primary ambient air quality standard for criteria pollutants.   The largest severity
 factors are for NOx emissions from a 30 m stack (S=1.3) and a 60  m stack (S=0. 57).
 Severities for SOx and particulates are  in the range 1.0-0.05.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                           b.IDENTIFIERS/OPEN ENDED TERMS
                                               c. COSATl Field/Group
Air Pollution
Assessments
Sheet Glass
Manufacturing
Nitrogen Oxides
Sulfur Oxides
Dust
Air Pollution Control
Stationary Sources
Particulate
Source Severity
13B
14B
11B
05C
07B
11G
18. DISTRIBUTION STATEMENT
 Unlimited
                                           19. SECURITY CLASS (This Report)
                                           Unclassified
                                               21. NO. OF PAGES
                                                 159
                      ?0. SECURITY ri ASK /Tlii.t
                       Unclassilied
                                               •>•>. PRICS
EPA Form 2220-1 (9-73)
                                          147

-------