EPA-600/2-76-269
October 1976
Environmental Protection Technology  Series
                               SOURCE  ASSESSMENT:
         GLASS  CONTAINER  MANUFACTURING PLANTS

                                   industrial Environmental Research Laboratory
                                        Office of Research and Development
                                       U.S. Environmental Protection Agency
                                 Research Triangle Park, North Carolina 27711

-------
                  RESEARCH REPORTING  SERIES


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

          1.  Environmental Health Effects Research
          2.  Environmental Protection Technology
          3.  Ecological Research
          4.  Environmental Monitoring
          5.  Socibeconomic Environmental Studies

This report has been assigned to the  ENVIRONMENTAL PROTECTION
TECHNOLOGY series.  This series describes research performed
to develop and demonstrate instrumentation, equipment and
methodology to repair or prevent environmental  degradation from
point and non-point sources of pollution.  This work provides the
new or improved technology required for  the control and treatment
of pollution sources to meet environmental quality standards.

                      EPA REVIEW NOTICE

This report has been reviewed by the U. S. Environmental Protection
Agency, and approved for publication.  Approval does not signify that
the contents necessarily reflect the views and policies of the Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
This document  is  available  to  the public through the National
Technical  Information  Service,  Springfield,  Virginia  22161.

-------
                                   EPA-600/2-76-269

                                   October 1976
        SOURCE  ASSESSMENT:

          GLASS  CONTAINER

      MANUFACTURING PLANTS
                     by

        J.R. Schorr, Diane T. Hooie,
   Philip R. Sticksel, and Clifford Brockway

       Battelie-Columbus Laboratories
               505 King Avenue
            Columbus, Ohio  43201
       Contract No. 68-02-1323. Task 37
            ROAP No. 21AFA-013
         Program  Element No. 1AB015
   EPA Task Officer:  Edward J. Wooldridge

 Industrial Environmental Research Laboratory
   Office of Energy, Minerals, and Industry
      Research Triangle Park, NC  27711
                Prepared for

U.S. ENVIRONMENTAL PROTECTION AGENCY
      Office of Research and Development
            Washington, DC 20460

-------
                                   PREFACE
     The Industrial Environmental Research Laboratory (IERL) of EPA has
the responsibility for insuring that air pollution control technology is
available for stationary sources.  If control technology is unavailable,
inadequate, uneconomical or socially unacceptable, then development of the
needed control techniques is conducted by IERL.  Approaches considered
include:  process modifications, feedstock modifications, add on control
devices, and complete process substitution.  The scale of control technology
programs range from bench to full scale demonstration plants.
     The Chemical Processes Section of IERL has the responsibility for
developing control technology for a large number (> 500) of operations in
the chemical and related industries.  As in any technical program, the first
step is to identify the unsolved problems.
     Each of the industries is to be examined in detail to determine if there
is sufficient potential environmental risk to justify the development of
control technology by IERL.  This report contains the data necessary to make
that decision for glass container manufacturing plants.
     Battelle's Columbus Laboratories was contracted with EPA to investigate
the environmental impact of the glass container industry, which represents a
source of emissions in accordance with EPA's responsibility as outlined above.
Dr. J. Richard Schorr served as Program Manager for this study.  Mr. Edward J.
Wooldridge served as EPA Project Monitor.  The study was completed by IERL-RTP.
Project responsibility was transferred to the Industrial Pollution Control
Division of lERL-Cincinnati on October 15, 1975.
                                     iii

-------
iv

-------
                           TABLE OF CONTENTS






                                                                      Page




INTRODUCTION	.	     1




SUMMARY 	     A




DESCRIPTION OF GLASS-CONTAINER INDUSTRY 	    11




     General Process Description. .... 	    11




     Plants and Locations	    13




     Shipment Volume and Weight . „	    13




     Process Details	    18




          Batch Handling	„	    18




          Melting and Fining	    23




          Forming		    29




          Postforming	»	    31




EMISSIONS	    33




     Raw-Materials Preparation and Handling .... 	    33




     Glass Melting.	    36




          Nitrogen Oxides 	 ...........  	    39




          Sulfur Oxides	    40




          Particulates	    42




          Carbon Monoxide	    44




          Hydrocarbons	    45




          Selenium	    45




          Other Emissions	    46




     Forming and Finishing	    46




          Forming	    46




          Surface Treatment 	    47
                                     v

-------
                            TART.K OF CONTENTS
                               (Continued)


                                                                       Page

           Annealing	•	    48

           Decorating.  .  .  . .	    48

      Emission Characteristics  .  .  .	    50

           Raw-Materials  Preparation  „	•	    50

           Glass Melting  .  . ....  .  .  .  .  .	    50

           Forming and  Finishing  	    50

      Ground-Level Concentrations	    52

           Affected Population  „  „  .  .  .	    61

 CONTROL TECHNOLOGY. .	    64

      Raw-Materials Preparation.  .....  	  .  	    65

           Emissions ........  	  .... 	    65

           Raw-Materials-Control  Technology	    66

      Glass-Melting Operation.	    69

           Emissions	    69

           Glass-Melting-Control  Technology	    71

           Efficiency of  Equipment	    79

      Forming  and Finishing. .	    80

           Forming and  Finishing  Control  Technology.  ........    81

           Surface Treatment ......  	  . . 	    81

           Surf ace-Treatment-Control Technology	    82

           Decorating.  ................. 	    82

           Decorating-Control Technology  ........ 	    82


FUTURE GLASS-CONTAINER PRODUCTION	    83

UNUSUAL RESULTS  	    85

REFERENCES	    86

                                    vi

-------
                               TABLE OF CONTENTS
                                  (Continued)

                                                                             Page
                                  APPENDIX A

 Geographical Listing of the 122 Container Glass Plants	A-l


                                  APPENDIX B

 Emissions Data	B-l

                                  APPENDIX C

 Stack Heights from the Various Phases of Glassmaking	C-l

                                  APPENDIX D

 State Listing of Total Emissions as of 1972	D-l

                                  APPENDIX E

 Conversion Factors	E-l

    	                           APPENDIX F

 Glossary of Terms	F-l

                                  APPENDIX G

 Letters of Comment	G-l


                               LIST OF TABLES
TABLE  1.    GLASS-INDUSTRY  STATISTICS.
TABLE  2.   AVERAGE  EMISSIONS OF MAJOR SPECIES FROM GLASS-
           CONTAINER PLANT	   6

TABLE  3.   SOURCE SEVERITY FOR GLASS-CONTAINER EMISSIONS	   9

TABLE  4.   MAJOR GLASS-CONTAINER MANUFACTURERS IN THE
           UNITED STATES	14

TABLE  5.   PRODUCT  SHIPMENTS OF THE GLASS-CONTAINER INDUSTRY	16

TABLE  6.   MINOR CONSTITUENTS OF CONTAINER GLASS	23

TABLE  7.   PARTICIPATE EMISSIONS DURING RAW-MATERIAL
           PREPARATION AND HANDLING FOR CONTAINER GLASS 	  37
                                        vii

-------
                             TABLE OF CONTENTS
                                  (Continued)

                                                                      Page

  TABLE 8.    EMISSIONS  FROM FLINT AND AMBER CONTAINER GLASS-
             MELTING-FURNACE-OPERATIONS  	    38

  TABLE 9.    EMISSIONS  FROM THE ANNEALING OF SODA/LIME
             CONTAINER GLASSES.
                                                                       49
  TABLE 10.   GENERAL  SPECIFICATION LIMITS FOR RAW MATERIALS
             USED  IN  CONTAINER GLASS MANUFACTURE	   51

  TABLE 11.   PARAMETERS FOR GLASS-MELTING FURNACES OF A
             REPRESENTATIVE PLANT IN THE GLASS-CONTAINER
             INDUSTRY AS USED IN ATMOSPHERIC-DISPERSION
             CALCULATIONS  	   53

  TABLE 12.   RELATIVE FREQUENCY OF ATMOSPHERIC STABILITIES	   57

  TABLE 13.   MAXIMUM  POLLUTANT CONCENTRATIONS AND SOURCE
             SEVERITY FOR  EMISSIONS FROM THE MELTING FURNACES
             FOR A REPRESENTATIVE GLASS CONTAINER PLANT 	   58

  TABLE 14.   MAXIMUM  AVERAGE GROUND-LEVEL CONCENTRATIONS  (X   ) OF
             AIR POLLUTANTS FROM CONTAINER GLASS-PLANT SOURSf§ BE-
             SIDES THE MELTING FURNACE	   60

  TABLE 15.   GLASS-GRADE PARTICLE-SIZE SPECIFICATIONS FOR
             SAND, LIMESTONE, AND 10- AND 20-MESH DOLOMITE	   6/

  TABLE 16.  MAXIMUM USE TEMPERATURE FOR VARIOUS FABRIC-
            FILTER MATERIALS	   75

  TABLE  17.  GLASS CONTAINER PRODUCTION STATISTICS	   84


                               LIST OF FIGURES


 FIGURE 1.  PROCESS-FLOW DIAGRAM FOR GLASS CONTAINERS (SIC-3221)  .  .   12

 FIGURE 2.  LOCATION  OF  GLASS-CONTAINER OPERATIONS 	   15

 FIGURE 3.  APPROXIMATE BREAKDOWN OF THE TYPES OF CONTAINER GLASS
            (1974)	   17

 FIGURE 4.   GLASS-CONTAINER MANUFACTURE	   19

FIGURE  5.  PROCESS FLOW DIAGRAM  OF  A TYPICAL BATCH PLANT	   20

FIGURE 6.  ILLUSTRATIVE SKETCH OF A SIDE-PORT REGENERATIVE-
           MELTING FURNACE	   25

                                      vlii

-------
                              TABLE OF  CONTENTS
                                  (Continued)
FIGURE 7.   LONGITUDINAL  SECTION  OF A  SIDE-PORT  REGENERATIVE

FIGURE 8.
FIGURE 9.
FIGURE 10.
FIGURE 11.
FIGURE 12.
FIGURE 13.
FIGURE 14.
FURNACE 	
TRANSVERSE CROSS SECTION OF FOREHEARTH 	
CROSS SECTION OF SIDE-PORT REGENERATIVE FURNACE .....
ILLUSTRATION OF THE I.S. BLOW -AND -BLOW PROCESS
FOR FORMING NARROW-NECK CONTAINERS 	
TYPICAL POINTS OF PARTICULATE EMISSION FROM
RAW-MATERIALS HANDLING 	
PARTICULATE EMISSIONS SHOWN ARE LINEAR WITH THE
RECIPROCAL OF BRIDGEWALL TEMPERATURE. .... 	
ILLUSTRATION DEPICTING CALCULATION OF AREA
USE OF COMMERCIAL-COLLECTION EQUIPMENT FOR EMISSION
CONTROL ON BOTH AMBER AND SODA/LIME GLASS FURNACES. . . .
. . 26
. . 26
. . 28
. . 30
. . 34
. . 43
. . 62
. . 74
                                         ix

-------

-------
                           LIST OF SYMBOLS
     Symbol                              Definition
      AAQS                   Ambient air quality standard
A, B, C, D, E, F             Atmospheric stability classes
a, b, c, d, e, f             Constants in dispersion equations
       A_                    The ratio Q/acmi
                                         ?   2
       BB                    The ratio -Hz/2c
        K
       CI                    Confidence Interval
       D.                    Inside stack diameter
       e                     Natural logarithm base
       H                     Effective stack height
       h                     Physical stack height
       AH                    Plume rise
       k                     "Student t" test variable
       m                     Number of samples
       N.                    Sample value
       p                     Atmospheric pressure
       Q                     Mass emission rate
       R                     Downwind dispersion distance from source
                             of emission release
       s                     Sample standard deviation
       S                     Source severity, ratio X   /AAQS
                                                     max
       T                     Ambient temperature
        a
       T                     Stack gas temperature
        s
       t                     Instantaneous averaging time of 3 minutes
        o
       t                     Averaging time
       TLV                   Threshold limit value
       u                     National average wind speed
       v                     Stack gas exit velocity
        s
       R                     Horizontal distance from centerline of
                             dispersion
                                      xi

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

  V

  IT

  CT

  X
X
 max
 max
           Definition

Sample mean

3.14

Standard deviation

Downwind ground level concentration at
reference coordinate x and y with emission
height of H

Time average ground level concentration of
an emission

Instantaneous maximum ground level
concentration

Time average maximum ground level
concentration
                             xii

-------
                                  SECTION  I
                                  INTRODUCTION

          Air emissions released  in  the manufacture of glass containers were
examined in this  study.   This report describes  the glass-container industry,
the nature of air emissions  and their environmental impact, the control
technology employed,  and  the future  growth of this industry.
          Glass containers represents the largest of three segments of
the glass industry  (glass containers, flat glass, and pressed-and-blown
glassware).  Each segment is defined by a Standard Industrial Classification
(SIC) number, as  used by  the Department of Commerce.  Glass containers
is designated by  SIC  3221 and includes the manufacture of glass containers
for food, beverages,  medicines, toiletries,  and cosmetics.  This includes
both narrow-neck  and  wide-mouth containers.  Shipments in this segment have
grown at an average rate  of  about 3.5 percent since 1971.  Industry shipments
in 1973 had a value of $2.3  billion,  or about 51 percent of the glass industry
total.
          Flat glass  is designated by SIC 3211.  This includes both the
manufacture of flat glass and some fabrication  of flat glass into a tempered-
or laminated-glass  product.  Flat-glass products include:  window glass,
plate glass, wire glass,  tempered glass, and laminated glass.  These products
are consumed primarily by the automotive and construction industries.  Value
of shipments in 1973  was  $1.1 billion, which was 24 percent of the whole glass
industry.
          Pressed-and-blown  glassware is designated by SIC 3229 and includes
all glassware not classified under SIC 3221  or  SIC 3211.  This industry
segment is very diversified  and includes products such as:

          •  Table, kitchen, art,  and novelty glassware
          •  Lighting and electronic glassware
          •  Scientific,  technical,  and other glassware
          •  Textile  fibers.

Industry shipments  in 1973 had a value of $1.3  billion, which was about
25 percent of the total industry.

-------
           Table  1  gives  some  1973 statistics on the three segments  of  the
 glass  industry.  It  shows  that over 154,000 people produced merchandise
 valued at  over $4.5  billion.  Some 70 percent of the glass products are made
 by the glass-container segment.
           Separate Source  Assessment Documents were prepared for  the flat-
 glass  and  pressed-and-blown glassware segments.  This report deals  only with
 glass  containers;  however, some of the same emissions and control  technology
 will also  be  found in the  other glass industry segments.  The report
 delineates the various emission points, identifies the type and quantity
 of emissions, and  describes the characteristics of the air pollutants  found.
 The mass emissions of criteria pollutants  (particulates, NO , SO  ,  and CO
                                                           X    X
 and hydrocarbons)  from glass  container plants are compared with national
 emissions  from all stationary sources.  The maximum average ground-level
 concentrations of  emissions from a glass-container plant are compared  to
 the corresponding  ambient  air-quality standards.  Control technology which
 is being used or could be  applicable to the manufacture of glass  containers
 is also discussed.
           The manufacturing operations for glass containers is grouped into
 three  categories:

           •  Preparation of raw materials
           •  Glass melting
           •  Forming and finishing.

Emissions  and control technology for each of these three areas is
presented.

-------
                     TABLE 1.  GLASS-INDUSTRY STATISTICS
                                                        (a)
1973 1973
Employees, Value of Shipments,
SIC Industry Segment (103) ($ Million)
3221 Glass Containers 77.8 2,316
3211 Flat Glass 26.3 1,118
fc)
3229 Pressed -and -Blown Glass 50.0 1,120
Total 154.1 4,554
, 1973 ,,.
Shipments *• \
(Metric Tons x 10 )
11.32
3.12
1.57
16.01
(a)   Source:  Department of Commerce and References 1 and 15.
(b)   Metric Ton = 1012 grams.
(c)   Excludes textile fibers.

-------
                                  SECTION II

                                   SUMMARY

           This document describes a study of air  emissions released during
 the production of glass containers.   The industry is  defined by Standard
 Industrial Classification (SIC)  No.  3221. It encompasses  the preparation
 of raw materials at the plant site,  the production of molten glass in a
 furnace, the forming of glass containers, and certain post-forming operations
 required to manufacture these products.
           The glass-container industry in the United  States produced 11.005 Tg
 (12.133 million ton) of salable  product in 1974.   Of  that  total, about
 15 percent (1.651 Tg) was amber  glass; the remainder  (green and clear
 glass) is classified as flint glass.   In 1974,  the domestic glass-container
 industry consisted of 25 manufacturers operating  120  plants.   Geographically,
 these plants are located near the local markets they  serve, with the largest
 concentration being in the East,  North-Central, and Middle Atlantic regions
 of the country.   Glass-container plants are  located in 29  states;  California, Illinoi
 Indiana, Ohio, Pennsylvania,  and New Jersey  are states with the largest
 number of manufacturers.   The average county population density at a plant
                                      2
 site is estimated to be 356 people/km .

                         Manufacturing Technology

           In a glass-manufacturing process,  raw materials  (e.g., sand, soda
 ash,  limestone)  are uniformly mixed and these loose materials transported
 to a furnace where they are melted at elevated temperatures (>1500 C) into
 a  homogeneous mass.   More than 95 percent of the  glass is  made in fossil
 fuel-fired furnaces  where energy is  predominately transferred to the glass
 by radiation from a  flame or  reradiation from the refractory chamber containing
 molten glass.  Molten glass is kept  at elevated temperatures until it is
 of a quality  (bubble-free)  sufficient for making  containers.
          12
*  Tg = 10   gram.  Metric prefixes and other  conversion
   factors are given in Appendix F, Page  F-l.

-------
The  glass  is  then cooled to approximately 1300 C,  removed  from the  furnace,  and
cut  into "gobs".   The gobs are fed to a machine and  formed into containers.
Approximately 30  percent of these containers then  undergo  surface treatment
and  about  3 percent are decorated.  The exact operations used  in a  specific
plant  depend  upon the type of product desired.   All  containers go through a
gas-fired  annealing furnace for removal of residual  stresses.   Highest
temperatures  during annealing range from about 590 to  650  C.

                                   Emissions

         Emissions were examined from three areas  within the glass-container
manufacturing plant:  (1) raw-materials preparation  and handling, (2) glass
melting, and  (3)  forming and finishing.  The largest emissions occur from
the  glass-melting operation.

Manufacturing Plant

         Table 2  summarizes the average emissions  of major species  from a
glass-container manufacturing plant determined by  this study.   While emissions
for  amber  and flint glass are discussed separately in  the  report, significant
differences were  not found, and the data in Table  2  is for a plant  producing
flint  glass.   The emissions listed are for an average plant production
capacity of 319 Mg/day (352 ton/day).  Annual production was 105 Gg (90 percent
of capacity).   This table shows that over 97 percent of the plant emissions
come from  the glass-melting furnace.   The major species (over  90 percent) are
NO , SO  ,  and particulates.  Furnace stack heights average 26  m (65 ft) when
  X    X
ejection air  is used and 45 m (147 ft) for natural draft.

Total  Industry

         Nitrogen oxides have the highest emission factor  (3.07 g/kg + 47%).  This
includes both flint and amber glass as shown later in Table 8.   Total annual
emissions  are  38.9  Gg.   Accordingly,  NO  contributes the greatest amount to  the
                                        X
national NO   emissions  from all stationary sources (0.34 percent).
           x

-------
                                                                                                 (a)
                           TABLE  2.  AVERAGE EMISSIONS  OF MAJOR SPECIES FROM GLASS-CONTAINER PLANTk
Emission Species
NOX
sox
Particulates CO
Hydrocarbon
Selenium
HC1
EMISSION FACTOR,  g/kg


     (1)  Raw Materials


     (2)  Glass Melting


     (3)  Form and Finish
     0            0       0.03 + 100%         0             0


3.40 + 43%   1.84 + 36%   0.71 + 30%     0.06 +  170%   0.08 +  100% 0.002 +  100%


0.02 + 100%       0       0.05 + 100%   0.002 +  100%   4.43 ±  100%                0.02 + 100%
TOTAL ANNUAL EMISSIONS . Mg
(1)
(2)
(3)
Raw Materials
Glass Melting
Form and Finish
0
358
2
0
193
0
4
86
2
(a)  Production 319 Mg/day  (352  ton/day), assuming 85 percent pack rate of flint glass or 104.8 Gg/yr.

                                                                                    (2)
(b)  Represents primarily surface treatment emissions for 30 percent of production.


(c)  Represents decorating  emissions  for 3 percent of production  (42%) and emissions  from annealing  lehrs  (29%)

     and forming operations  (29%).


(d)  Emission factors are only for flint glass, which constitutes an estimated  85 percent of  glass production.

     Emission factors for both flint and amber are given later  in Table 8.

-------
          Sulfur oxides have the next highest emission factor which is
 1.70 + 47%.   This includes both flint and amber as given in Table  8.
 Total annual emissions are 21.6 Gg.   This amounts  to  0.31 percent  of  the
 national emissions from all stationary sources.
          Particulates for both flint and amber  glass  were determined  to have
 an emission  factor of 0.68 g/kg + 36%, with total  annual emissions of 8.6  Gg.
 This would be 0.007 percent of national emissions  from all stationary sources.
           Carbon monoxide has  an emission factor of 0.07 g/kg + 143% with
 total annual emissions of 0.9  Gg.  The emission  factor  for  hydrocarbons
 was  0,08 g/kg + 100% with total annual emissions of 0.7  Gg,  These  contribute
 0.005 and 0.003 percent respectively to the  national  emissions from all
 stationary sources.   Finally,  the emission factor  for selenium was
 2  mg ± 100%  with total annual  emissions of 0.02 Gg.
           Emissions from raw-materials preparation and handling do  give
 rise to some particulate emissions,  primarily from dust  generated-'during
 discharging,  conveying, crushing,  and mixing operations.  Composition of the
 emissions is the same as that  of the raw materials (i.e., sand, limestone,
 soda ash).   The average emission factor is 0.03 g/kg Hh 100%.  Total annual
 emissions for the glass container industry were estimated to be only
 459  Mg or 0.0003 percent of the national particulate  emissions from all
 stationary sources.   Over 90 percent of the  industry  employs controls  (primarily
 filter bags)  in dust generating areas.
           Many different processes can be used  in  the forming and  finishing
 operations.   Emissions of the  major  species  from forming and finishing are relatively
 low,  as can  be seen in Table 2.   These emissions consist  of hydrocarbons
 emitted from the forming operation (0.03  g/kg), tin oxide and
 hydrated tin chloride particulates and HC1 emitted from  a  surface  treatment
 operation performed on about 30 percent of the  containers produced.   Combustion
 products  are  emitted from gas-fired  annealing and  decorating lehrs, and
 hydrocarbons  are  emitted from  decorating operations used with about 3 percent
of the  glass  container  production.   The emissions  from all  of these areas  have
been  combined  in Table  2, because  relatively speaking they  constitute  less
than  2 percent  of  the  emissions  from a-normal manufacturing plant.  They
are broken out  by  emission  source  and type in the  body of the report.

-------
                               Control Technology

             Emissions  from glass  container melting furnaces are generally not
    controlled by add-on equipment.  This is not the case in every State.
    Frequently emission standards  can be met with proper operating conditions.
    Baghouses and electrostatic precipitators are used by a few manufacturers
    to  control submicron particulate emissions.

                                Source Severity

              Impacts of these emissions are directly related to the ambient
    concentrations the  emissions create at ground level.  Atmospheric dispersion
    calculations were made to calculate maximum average ground-level concentrations
    (X   ) for the emissions from an average plant producing 105 Gg/year.  Results
     in 3.x
    of these  calculations are presented in Table 3.
              Source severity factor, S, has also been used to describe the impact
    of the emissions.   For those pollutants which have an ambient air-quality
    standard  (AAQS), S  is the ratio of X    to the primary AAQS.  In cases where
                                        max
    no AAQS has been established, S is based upon the Threshold Limit Value (TLV)
    through the following equation which includes a factor for correcting the
    TLV to a  24-hour day (8/24) and a safety factor (1/100).
                                      X
                                       max
                               TLV (8/24) (1/100)
Results of  the  source  severity  factor calculations also appear  in Table  3.
The highest severity factor  (0.38)  is produced by nitrogen oxide emissions
from the melting  furnace and hydrocarbon emissions from decorating  lehrs (0.15).
Other severity  factors were  less  than 0.05.
          A third measure  of the  impact of  the plant  emissions  is given  by
the affected population.   This  measure is defined as  the  population around
the representative plant who are  exposed to a source  severity factor greater
than 1.0.   Computations of the  affected population showed that  all  sources
were less than  1.0.  Nitrogen oxides would  be the highest if computed on
a more restrictive basis such as  a  source severity factor of 0.1.   In this
case 11,700 persons would  be affected.

-------
             TABLE 3.   SOURCE SEVERITY FOR GLASS-CONTAINER EMISSIONS
Source —
Pollutant
Melting
Furnace —
NOX
SOX
Particulates
CO
Hydrocarbons
Selenium
Materials
Handling--
Particulates
Container
Decorating —
Hydrocarbons
Surface Treatment -
HCL
Primary Ambient
Air-Quality Standard
Averaging
Hg/m3 Time, hr


100
365
260 ,
4 x 104
160
Trace


260


160
(b)
Trace


24
24
24
8
3
24


24


3

24
Maximum Average
Ground - Le ve 1_
Concentration, max
M-g/m3


37.9
20.5
9.2
0.8
°'9 /^
Trace (b)


0,02


23
(b)
Trace
Severity
Factor


0.38
0.056
0.035
2 x 10"^
5.9,x 10


c
9.2 x 10


0.15


(a)  Other severity factors,  including those for selenium,  tin particulates, and
     HC1 were quite low.

(b)  Trace <10 yg/m3.

-------
                                     10
                                Future Growth

         Historically, the glass container industry's growth has fluctuated
considerably in the past 8 years, but shipments have steadily increased since
1967 at an average annual rate of 6 percent.  A large portion of this growth
is attributable to the popularity of the nonreturnable beverage bottle.  In more
recent years, growth in shipments has been less.  Future growth may well be
tied to legislation restricting use of nonreturnable containers.  It is likely
that 1980 production levels will be 20 percent higher than for 1974.  Emissions
would increase proportionally and possibly at an increased rate, since
the industry is moving away from the use of natural gas to the use of oil.
The actual effect of conversion from gas to oil firing on emission rates
is not known.

-------
                                       11
                                 SECTION III

                    DESCRIPTION OF GLASS-CONTAINER INDUSTRY

          This section describes in general terms the process steps used
in the manufacture of glass containers and presents certain statistical
information pertinent to the glass-container industry described by the
Department of Commerce for Standard Industrial  Classification (SIC)  3221.

                         General Process Description

          Figure 1 is a process-flow diagram which generally depicts the
flow of materials through a glass  container manufacturing plant.  The process
is categorized into four steps:  batch handling, melting and fining, forming,
and postforming.  These specific steps are discussed in more detail later
in this section of the report.
          Basically, the manufacture of glass containers consists of melting
(and reacting) a mixture of raw materials  (consisting primarily of silica,
soda, and lime) which have been prepared in the batch-handling step so
as to minimize segregation and impurities  in the batch.  Cullet (scrap
glass) is also added at this stage.  In the glass melter, the materials are
melted down, the molten glass is fined (i.e., residual trapped gases are
removed), and then the temperature of the  glass is lowered so that it can be
handled in the forming step.  The  glass passes from the melter to the forming
equipment via the forehearth, a relatively shallow, narrow refractory
channel having a refractory roof and individual heating and  cooling systems
for controlling glass temperature.  Individual gobs of glass are fed to the
forming machines where the molten  glass is transformed into  a product by one
of two methods:  blow and blow, or press and blow.  The formed  container
may now go through a series of postforming steps, depending  upon the product
desired, but which always includes annealing, where stresses are removed
through a controlled, uniform-cooling cycle.  Finally, the containers may
undergo various additional steps,  such as  surface treatment, inspection,  testing,
decoration, and plastic coating.

-------
                                        12
                                                     Raw
                                                   materials
                                                    receipt
                                                     and
                                                    storage
 Batch
weighing
  and
 mixing
Crushed
 glass
 cullett
                                 Batch
                                 charg-
                                   ing
          Scratch
         resistant
          surface
        treatments
        Lubricity
         surface
        treatments
                               Inspection
                                  and
                                 testing
                                            Decoration
                                              and/or
                                              plastic
                                              coating
                                 Pack

                                      Ll
                                         o
                                         m
                                                                         CD
                                                                         O>

                                                                         1
                                                                         o
                                         TO
FIGURE 1. PROCESS-FLOW DIAGRAM FOR GLASS CONTAINERS (SIC-3221)

-------
                                        13
                              Plants and Locations

           According to information gathered from the Department of
                                      (1)*
 Commerce 1972 Census of Manufacturers     and from the Glass Containers
                               (2)
 Manufacturers Institute (GCMI)v ', there were 120 establishments manu-
 facturing glass containers in the United States.  These 120 plants (see
 Appendix A) are operated by 25 manufacturers as shown by Table 4.
 Statistics obtained from 1975 glass-industry directories and from industry
        (2-5)
 sources      indicate that approximately half of these plants are operated
 by the five largest companies.
           Geographically, glass-container plants are located near the  local
 markets they serve.  As such, plants are found throughout the United States,
 but a large number are concentrated in the East North Central and Middle
 Atlantic portion of the United States.  The states containing the largest
 number of manufacturers are California, Illinois, Indiana, Ohio, Pennsylvania,
 and New Jersey.  The regional distribution of the ma^or plants is shown in
 Figure 2.

                           Shigment Volume and Weight

           Table 5 provides estimated 1974 output data for the glass-container
 industry (SIC 3221).  The volume of shipments for 1974 was 276,382 thousand
 gross as compared to 267,732 thousand gross in 1972, or only 3.2 percent
 growth for the 2-year period.  The weight of glass containers shipped
 increased from 10,772 Gg (23,748 million pounds) in 1972 to 11,005 Gg
 (24,266 million pounds) in 1974, or an increase of 2.1 percent.   Three general
 types of container glass are produced:  amber, green, and clear.  For  this
 report, green and clear glass are considered together as a single category
 designated as flint.  The major difference between these two is the iron
 oxide additions.   Emission data was not found to be separately reoorted.
 Figure 3  illustrates the estimated breakdown between amber and flint.
*  References are listed on page  86.

-------
                                        14
     TABLE 4.  MAJOR GLASS-CONTAINER MANUFACTURERS IN THE UNITED STATES
                                                                        (a)
    Manufacturer
   No.
of Plants
                                                 Manufacturer
                             No.
                          of Plants
Anchor Hocking
Arkansas Glass
Ball Corporation
Bartlett Collins
Brockway Glass
Chattanooga Glass
Columbia Gas
Diamond Glass
Gallo Glass
Glass Containers Corp.
Glenshaw Glass
Hillsboro Glass
Indian Head
Industrial Glass
Kerr Glass
    9
    1
    4
    1
   14
    7
    1
    1
    1
   12
    2
    1
    7
    1
    7
Latchford Glass
Leone Industries
Liberty Glass
Madera Glass Co.
Metro Glass
Midland Glass
National Bottle Corp,
National Can
Owens"Illinois
Thatcher Glass
Underwood Glass
Wheaton Glass

            Total
  1
  2
  1
  1
  4
  4
  4
  4
 20
  6
  1
120
(a)  Source:  Material provided from GCMI,
              dated 10/24/75.

-------
FIGURE 2.  LOCATION OF GLASS-CONTAINER OPERATIONS(6)

-------
                                         16
           TABLE 5.  PRODUCT SHIPMENTS OF THE GLASS-CONTAINER INDUSTRY
                      (Department of  Commerce Classification SIC-3221)
                                                             (a)
 	Shipments of All Types of Glass  Containers   	

	1974	     	1973	     	1972	
Number(b)     Weight. Gg     Number (b_)     Weight.Gg     Number Q*).     Weight,Gg

 276,382       11,000        276,328       11,326        267,732        10,748


(a)  Sources:  U.S. Department of Commerce, Bureau
               of the Census, Series M32G(74)-13 and
               Statistics Glass Containers 1975 pub-
               lished by The Glass Container Manufac-
               turers Institute.

(b)  Number of containers in thousands of gross.

-------
                                17
              15 % Amber
              1,650 Gg
FIGURE 3.  APPROXIMATE BREAKDOWN OF THE TYPES OF CONTAINER GLASS  (1974)^15^

-------
                                        18
                                Process Details

           While specific equipment  will  vary, depending  upon the  manufacturer
.and the product being made,  the basic manufacturing  process  is  essentially
 the same for all glass-container manufacturers.  Portions  of raw  materials
 are mixed with each other and with  cullet  (scrap glass), conveyed to a melting
 furnace, melted, fined,  conditioned,  and fed  into  forming  machines.   The
 formed containers are then taken through a variety of  postforming and product-
 handling steps.
           The subsequent paragraphs describe  the process operations  and
 the raw material ingredients; thus  indicating the  potential  sources of
 materials that can be emitted into  the air as pollutants.  Figure 4
 schematically shows the  overall manufacturing process.

 Batch Handling

           The function of the batch-handling  operation is  to prepare and
 feed to the melting furnace  a batch which  is  both  chemically and  physically
 uniform in composition.   Control of the  composition, impurity level, and
 the size and moisture of the raw materials is important.   Cullet  is  crushed
 and either mixed with the raw materials  or added later.  Each of  the raw
 materials is carefully weighed,  mixed together, and  then conveyed to the
 batch chargers.   Care must be taken so that segregation  of a uniformly
 mixed batch does not  occur.
           A large plant  manufacturing container glass  usually houses the
 raw material mixing and  conveying equipment in a structure termed a  "batch
house",   A flow diagram  of a typical  batch house is  shown  in Figure  5.  In
most batch houses,  the storage bins are  located on top, with the  weigh hoppers
and mixers  located  below them to make use  of  gravity flow.
          Major  raw materials and cullet (broken scrap glass) are conveyed
from railroad hopper  cars  or hopper trucks  by a combination  of  screw
conveyors, belt  conveyors, and bucket elevators, or  by pneumatic  conveyors
to the elevated  storage bins.  Minor  ingredients are usually delivered to the
plant in paper bags or cardboard  drums and  transferred by  hand  to small bins.

-------
                                        19
Glass sand
Si02 5 99*
to yield Si02
crushed, washed
and screened
to ~ 20 -100
mesh
Soda ash
Na2C03
to yield Na20
~ 20-120 mesh
or granular


Limestone
or burnt lime
to yield CaO.
Usually some
MgO also results
~ 20-120 mesh

Feldspar
R2O.AI203.6SI02
to yield
AI20,,Si02
Na20 and K20
pulverized or
granular
                                                                    Other additions
                                                                     for K20, MgO,
                                                                    ZnO, BaO, PbO,
                                                                    etc and those for
                                                                    fining, oxidizing,
                                                                     coloring, and
                                                                      decolorizing
   Side-port
continuous tank,
 looking down
  through top
                             Packing, warehousing,
                                and shipping
          FIGURE 4.   GLASS-CONTAINER MANUFACTURE
                                                                 C7)

-------
                                                     20
        GULLET
RAW MATERIALS
RECEIVING
HOPPER
      V
          SCREW
          CONVEYOR
                                                        FILTER
                                                        VENTS
STORAGE BINS
MAJOR RAW MATERIALS
                                                     MINOR
                                                     INGREDIENT
                                                     STORAGE
                                                     BINS
                                    TO
                                    ATM
                        BELT CONVEYOR
                                                                                   (I
BATCH
STORAGE
BIN
                                                                                FURNACE
                                                                                FEEDER
                FIGURE 5.   PROCESS FLOW DIAGRAM OF A TYPICAL BATCH  PLANT

-------
                                        21
          Materials  are gravity fed from the storage bins  into weigh hoppers
 and  then  transported by transverse belts or bucket  elevators  into a mixer.
 Gullet  is crushed  to a desired size (usually between 0.5 and  2.0 cm).  After
 mixing, the  glass  batch is transferred to a charging bin located next to
 the  glass-melting  furnace or into a batch-storage bin,  depending upon the
 design  of the  batch-handling system.   Positive displacement or vibrator
 feeders at the bottom of the bins feed the materials to the glass-melting
 furnace chargers.  Gullet may be added to the batch in  the mixer, while the
                                                                        fO\
 batch is  being transferred, or charged separately to the melting furnace   .
          Batch is fed into the melter in either a  dry  or  moist state.
 Many companies add two to four percent water to the dry batch to help
 prevent segregation  during transport of the batch,  to minimize dust problems
 in the  melter, and to avoid carryover of dust into  the  regenerators.
          The  various handling and mixing operations are a source of
 particulate  emissions which are similar (same materials, same processes)
 as those  in  other  industries  '   .  Because of environmental and economic
 incentives,  most large manufacturers practice dust  control, usually by
 means of  cloth filters and baghouses    .

          Batch Composition.  The basic raw materials for  soda/lime container
 glass are silica sand, soda ash (Na^CO.,) and limestone  (primarily CaCO_, plus
 some MgCO., in  dolomitic limestones).   Feldspar minerals are also utilized as
 a source  of  alumina  and alkali.  Minor amounts of other oxides are introduced
 as impurities  and  additional minor ingredients are  added for  specific
 purposes  discussed later.
          Glass sand must be of high purity (& 99 percent  Si02).  Primary
 impurities are Fe203 and A1203 which together will  be less than 1 percent.
 The  chief sources  are unconsolidated bank sand from New Jersey and the
 standstones  of the Alleghenies and the Mississippi  valley.
          The  U.S. supply of soda ash, Na CO , has  changed from predominantly
 synthetic to natural in recent years.   Prior to 1973, more than 60 percent of
 the domestic soda  ash was produced from NaCl by the Solvay process.  Now
more  than half comes from the natural deposits of trona ores  (sodium
sesquicarbonate, Na2C03NaHC03.2H20).   The primary deposits are in Wyoming,
but it is also found near Searles Lake and Owens Lake,  California.

-------
                                       22
           Limestone  is  the  source of calcium  (and magnesium)  oxides  in
 the  glass  batch.   This  rock has widespread occurrence  as  either a high-calcium
 limestone  consisting essentially of calcite,  CaC03,  or as a dolomitic
 limestone  which  is a mixture  of dolomite  (CaCO^MgCO.^  and calcite.   Good
 limestones contain less than  0.1 percent  ?&2°3  and about  * percent of
 silica and alumina.   Calcite  limestone deposits occur  in  the  central,
 southern,  and  eastern U.S.  Large dolomitic deposits occur in the central-
 midwestern parts  of  the U.S.
           Feldspars  are anhydrous aluminosilicates containing potassium,
 sodium,  and calcium  in  varying ratios.  They  are present  in virtually all
 igneous rocks; but most production comes  from pegmatites  which are coarsely
 crystalline rocks formed in the later stages  of crystallization of a magma.
 The  principal  accessory minerals are quartz,  mica, and other  silicates.
 The  most critical requirement for glass feldspar is  low iron  content.
                                              (2)
           A typical  glass-batch composition is   :

           Silica  sand             909 Kg    (2000  lb)     (55.6%)
           Soda ash                306 Kg    ( 674  lb)     (18.7%)
           Feldspar                118 Kg    ( 260  lb)     (7.2%)
           Limestone                294 Kg    ( 648  lb)     (18.0%)
           Salt cake  (Na2S04)       6.8 Kg    ( 15  lb)     (0.4%)
                     Total         1634 Kg    (3597  lb)     (99.9%)

           Typically,  the above ingredients would melt  down to 1370 Kg
 (3020  lb)  of glass and  give off 259 Kg (569 lb) of gases, primarily  (> 99%)
 C02<   The  batch volume  of 1.2 m  (42 cu ft) would produce 0.6 m3 (21 cu ft) of
 fluid  glass and 858 m  (30,300 cu ft) of  gaseous products (measured  at the
                              C2\
 furnace  temperature of  1500 C)v '.
           The batch will contain minor ingredients such as salt cake (sodium
 sulfate) and various  fining,  coloring, or decolorizing agents.  These
 compounds  rarely exceed 5 percent and are often less than 0.1 percent of
 the total glass composition.  Table 6 lists minor ingredients and their
effect on the glass.

-------
                                         23
             TABLE  6.  MINOR CONSTITUENTS  OF  CONTAINER GLASS(2'7)
             Purpose                     Effected By

             Amber  color                 FeS   (pyrites)
                                           X
             Green  color                 Cr2°3»  Fe2°3» Cu°
             Blue color                 CoO, FeO, CuO
                                                2_
             Decolorization,  i.e.,       MnO, Se  , NiO,  Co,0, CeO,
               Mask Fe203  color                          J 4     "
                                           2_
             Fining                     SO.   ,  C
          Gullet  (scrap  glass)  collected from the  plant, purchased on the
open market or  from recycling centers,  is crushed  to below 2 cm in size.
It is blended with the raw batch in varying amounts.  Normally, however,
only about 15 to  20 percent of  the batch going into a furnace is cullet.
If a plant is producing  both flint (clear)  and colored  containers, the cullet
must be individually collected  and stored to prevent undesirable color
fluctuations and  to avoid  glass foaming conditions caused by mixing amber
and flint glass having different oxidation states.

Melting and Fining

          The mixed batch  is fed into a large continuous-melting furnace
where the batch is melted,  fined,  and conditioned. The melting furnace
consists of three  distinct  regions:   the melting end, refiner (conditioning
chamber), and forehearth.   At the  doghouse (batch  feeding end of the melting
compartment), the  raw materials are fed onto a molten mass of glass having
a temperature near 1500  C.   The batch materials react,  melt, and disappear
into the liquid glass after floating about  one-third to one-half of the length
of the melter.   As  the molten glass moves on through the furnace tank, it is
fined,  trapped  gases  (bubbles)  are removed,  and the melt homogenized.

-------
                                       24
The  glass is essentially free from bubbles when it reaches the end of
the  melting chamber.  Then the glass passes through a submerged refractory
throat  into a conditioning chamber, popularly called the refiner,
where it is cooled to approximately 1300 c<  A refractory bridge wall above
the  throat prevents any glass surface scum from passing into the refiner and
also acts as a heat barrier.  In the refiner, the glass is cooled to
increase its viscosity to the proper working level and to dissolve any
remaining tiny bubbles or gaseous inclusions.  Then the glass flows through
shallow, refractory-lined channels, called forehearths, to the forming machines.
In the  forehearth, a uniform temperature of the molten glass is maintained.
This will be near 1100 C and is adjusted by individual heating and cooling
systems.  At the end of the forehearth, a stream of molten glass is cut into
individual gobs of glass and fed to the forming machine.
          Figure 6  is a simplified illustration of a side-port regenerative
melting furnace and forehearth.  A longitudinal section through the melter
and  refiner of such a furnace is shown in Figure 7.  A transverse section
through a forehearth showing the design capability for adjusting glass
temperature as it moves to the feeder is shown in Figure 8.
          Characteristic dimensions of container glass melting furnaces
fall within the following ranges:  length 6 to 18 meters (20 to 60 ft),
width 3 to 8 meters (10 to 25 feet), and depth 0.6 to 2 meters (2 to 6 feet).
The  molten glass holding capacity ranges from 36 to 454 Mg (40 to 500 tons),
and  output 27 to 317 Mg/day (30 to 350 tons/day).

          Melting Energy Sources.  The glass-container industry predominantly
uses regenerative furnaces of side-port or endrport design burning a fossil
fuel.   In the side-port design commonly used for larger furnaces  (> 175  ton/day),
combustion products and flames pass in one direction across the top of the
molten  glass and exhaust through ports on the opposite side of the  furnace.
In the  end-port configuration, combustion products, and flame travel in  a
horizontal U-shaped path across the surface of the glass within the melter.   Fuel
and air mix and ignite at one port and discharge through a second port adjacent

-------
                                            25
                        RAW MATERIALS
                                                REGENERATOR
                                                                    FEEDER
                                                                FORMING
                                                   ANNEALING
FIGURE 6.  ILLUSTRATIVE SKETCH  OF A SIDE-PORT REGENERATIVE-MELTING FURNACE

-------
                                  26
     Refiner End
   Melting End
               Bridge Wall    Crown
Port
                                                          Doghouse
                                   Glass
               0
               Throat
    Bottom
FIGURE 7.  LONGITUDINAL SECTION OF A SIDE-PORT REGENERATIVE FURNACE
                                                                    (315)
           Cooling Air
     Burner-
     Refractory
    Cooling Air
             Burner
     FIGURE 8.  TRANSVERSE CROSS  SECTION OF FOREHEARTH^15^

-------
                                     27
 to  the first on the same end wall of the furnace.   To conserve fuel,  the
 regenerative-firing system is used which employs dual chambers filled with
 brick checkerwork.   While the products of combustion from the melter  pass
 through and heat one chamber, combustion air is preheated in the opposite
 chamber.  The functions of each chamber are interchanged periodically.
 Reversals occur about every 15 to 20 minutes as required for maximum
 conservation of heat.  Figure 9 is a cross section of a side-port regenerative
 furnace.   The regenerators are about two stores tall and are positioned on
 each side of "side-port" furnaces or at one end of "end-port" furnaces.
           The glass container industry has historically used natural  gas
 as  the primary energy source.  In 1971, over 83 percent of the energy was
 derived from the combustion of natural gas    .  In more recent years, the
 limited availability of natural gas has resulted in a shift to greater use
 of  fuel oil.   In 1975, the-energy supplied by natural gas was between 70
               (2)
 and 75 percent   .
           Since molten glass is an electrolyte, it can also be heated
 electrically.   Such heating is used primarily to supplement fossil-fuel
 heating,  rather than replace it and is referred to as electric boosting.  One
 estimate indicated  that approximately 40 percent     of all furnaces now
 have electric boosters.  Less than 5 percent of container glass furnaces in
 the U.S.  were heated entirely by electricity in 1974
           Future energy sources could vary considerably from the present-day
 pattern depending upon the availability and cost of fossil fuels.

           Me11ing-Furnace Emissions.  Exhaust gases from the melting  furnace
 are major sources of air pollutants from the glass container manufacturing
 process.   Emissions primarily include NO . SO , and particulates.  Nitrogen
                                         X    X
 oxides  (predominately NO) are formed by the reaction of atmospheric nitrogen and
 oxygen  under  the higher temperature conditions of  the furnace.   Sulfur oxides
 (predominately S0?)  come from the decomposition of sulfates in the melt and,
where  fuel  oil is used for heating, from the combustion of sulfur in  the fuel.
Particulate emissions may arise from:   (1) volatilization of materials in the
melt which  subsequently condense in the checkers or stack, or (2)  to  some extent  by

-------
                            28
                 Crown
Port
FIGURE 9.   CROSS  SECTION OF SIDE-PORT REGENERATIVE FURNACE

-------
                                        29
 physical  etitrainment of batch dust in combustion gases^    .   Batch  constituents
 which  can contribute to volatilization are predominately  alkali  sulfates and
 other  minor constituents such as selenium.

 Forming

           The  forming of glass containers  by industry's highly mechanized
 equipment involves  several steps.

           • The  molten glass is cut into  gobs  by a  set of shears
             as the glass leaves the forehearth.
           • Delivery equipment directs the gobs  into blank molds.
           • In the blank mold, the gob is partially shaped into a
             parison.  This task is performed predominantly by an air-
             blowing process for narrow-neck containers,  but  it may
             also be performed by pressing.   For  wide-neck containers,
             it is  performed by pressing.
           • The  parison is then inverted  and transferred into a
             blow mold.
           • The  parison is blown into final shape with compressed  air.

           The  process of forming narrow-neck containers in which the parison
 is formed by blowing steps is illustrated  in Figure  10.
           Most glass containers are formed on individual  section (I.S.)
 machines  which may  be designed with up to  10 sections per forming machine
 and may have maximum speed capabilities exceeding 200 containers per minute.
           In the  operation of the forming  equipment, coolants and lubricants
 are employed.  Water-based sprays  are used to cool the shears which cut the
 glass  gobs.  Delivery equipment which directs glass  gobs  into the blank molds
may be sprayed with lubricating emulsions  of hydrocarbon- or  silicone-based
materials,  and metal molds are commonly lubricated with mixes of graphite,
greases, oils, etc.

-------
                             30
      Delivery
Settle blow
                     Counter blow
/
n
/
^
/

*


,

?
\

| risn^ZaH
                    Transfer from blank mold to blow mold
      Reheat


                           Final blow
                      Takeout
FIGURE 10.   ILLUSTRATION OF  THE  I.S. BLOW-AM)-BLOW PROCESS
              FOR FORMING NARROW-NECK CONTAINERSW

-------
                                        31
 Postforming

           This step can consist of many operations,  depending on the  product
 requirements.   These include surface treatments, annealing,  decorating,
 and coating.
           Surface treatment with titanium or tin chloride may be applied
 to hot containers as they are transported along conveyors between the forming
 machine and annealing oven.  The metal chloride reacts to yield  titanium or
 tin oxide on the glass surface and releases HC1.  The surface oxide greatly
 increases the  container scratch resistance and thus  produces containers of
 higher service strength.
           The  primary step, through which all newly  formed containers pass,
 is annealing.   Annealing is necessary to remove stresses  that will weaken
 the glass or cause it to fail.  In this step, the entire  piece of glassware
 is brought to  a uniform temperature high enough to permit the release of
 internal stresses (590 to 650 C) and then cooled at  a uniform rate to prevent
 new stresses  from developing.  Annealing is done in  long  continuous furnaces
 called lehrs.
           Following annealing, but while the containers are  still hot (about
 100 to 150 C)  and on the lehr conveying belt, they may receive additional
 surface treatment.   Water-based emulsions of polymer or organic  materials are
 most  commonly  sprayed onto the hot bottles.  The latent heat of  the bottles
 evaporates the water and bonds the organic which yields a lubricious  container
 surface.   This surface lubricity minimizes or prevents jamming of bottles
 as they move along  in line conveyors and reduces surface  damage.
           After annealing (and surface treating, if  applied), containers are
 visually and optically/mechanically inspected.   In addition, cylindrical
 bottles to contain  pressurized products may be "proof tested" by applying an
 internal air pressure,  and/or passing them between rubber belts  or rollers
which  squeeze  the bottle.   Carbonated beverage bottles are also  statistically
 sampled and subjected to  destructive-burst tests.
           Decorations or  labels may also be applied  to containers in  the
production plant.   For  beverage bottles, screen-printing  processes may be
employed  for some bottles  to  apply organic resins  or vitrifiable  glass
colors, although other  techniques  and materials may  be used.  Only a  small
portion,  about  3 percent,  are decorated.(2)

-------
                                       32
          Containers, especially for carbonated beverages, may also be coated
with plastic.  Such coatings or sheaths may serve one or more functions.
They may reduce bottle-to-bottle handling damage, reduce filling-line noise,
serve as labels, or provide containment in the event of fracture of a
pressurized beverage container.  These coatings reportedly may be applied by:
wrapping and heat-shrink techniques, electrostatic-powder applications, dip-coating,
                (12)
or other methods    .  Only a small percentage (< 5%) of containers are
currently plastic coated, but this percentage may increase in the future.

-------
                                       33
                                   SECTION IV
                                   EMISSIONS

          Emissions  from glass-container  plants  are  categorized according to
 three operations within  the manufacturing process

          •  Particulate emissions from the  raw-materials handling,
             preparation,  and  transfer.
          •  Gaseous and particulate  emissions from  the glass-melting
             furnace.  These contain  SO , NO , submicron sodium sulfate
    _            •                       XX
             condensates, hydrocarbons, CO,  and  other minor  emissions
             such as selenium.
          •  Gaseous and particulate  emissions from  a variety of forming
             and postforming operations.   These  result from annealing,
             decorating,  surface treatment,  and  coating operations and can
             include particulates  (e.g.,  Sn09),  hydrocarbons, NO , and SO .
                                             fc>                   X        X

This section describes the various emissions, characteristics, their levels,
total quantities, and environmental effects.  The information is organized
according to the three sources within the manufacturing process.

          •  Raw-materials preparation  and handling
          •  Glass melting
          •  Forming and  finishing.

                     Raw-Materials  Preparation and Handling

          Typical points of particulate emissions during raw-materials
preparation and handling are shown in Figure 11.   These points include

          •  Unloading and conveying
          •  Crushing of cullet (scrap glass)
          •  Filling and emptying  of  storage bins
          •  Weighing and mixing of batch
          •  Feeding of batch  to glass  furnace (batch charging).

-------
                                           34
                                                               Conveying to storage
                                                               bins by
Acme
Minerals
                                                                  • Screw conveyors
                                                                  • Belt conveyors
                                                                  • Bucket elevators
                                                                  • Pneumatic conveyors
                                     Gravity-fed into
                                     weight hoppers
                                    Transported by
                                      • Transverse belts
                                      « Bucket elevators
                                    Transferred to
                                    charging bin
Batch charging
            Transferred
            to charging
            bin
                                      Gullet is ground
   Unloading of cullet
          *
Transferred to storage  bin
          \
      Storage bin
          t
  Transferred to grinder
               FIGURE 11.   TYPICAL POINTS OF PARTICULATE  EMISSION
                            FROM RAW-MATERIALS HANDLING

-------
                                       35
          All of these are potential sources of particulate emissions;
however, those participates which remain in the manufacturing plant may
constitute an OSHA health and safety consideration distinct from plant
emissions.   For the purposes of this  study, fugitive-dust emission
was defined  as particulate. emissions that  result  from industrial-related
operations,  and which escape to the atmosphere through windows, doors, vents,
etc., but not through a primary exhaust system, such  as a stack flue, or
control system.  This definition is derived in part from a paper presented by
Lillis and Young of the U. S. Environmental Protection Agency    .
Information  obtained from ambient sampling up-wind and down-wind of the
manufacturing facility is the preferred source of data.  If actual data
from high-volume samples are not available, engineering estimates based on
the particle size of raw material which can contribute to dust emissions are
desired.  For the purposes of this study,  the particle size range of the
raw materials which will be considered as  contributing to fugitive dust emissions
was 100 microns or less in diameter.   This definition of dust is not as broad
as the technical definition given by Stern   , but this particle-size
distribution seems reasonable for glass-manufacturing processes.  The settling
velocity of  a 100-micron-diameter sphere,  with a specific gravity of 2.0 g/cc,
is approximately 50 cm/sec in still air at 25 C and 1 atm.  Such a settling
rate is sufficiently slow that the emission of dust from a tall source,
such as the  raw materials storage bins, would probably contribute to the
total air emissions.
          Fugitive dust is usually emitted during unloading and conveying
operations.  To minimize dust emissions, these operations are generally
enclosed and the vents on storage bins and mixers exhausted through fabric
filters.  Batch wetting, or the addition of water to  the batch during the
mixing operation, is another practice  commonly used which minimizes particulate
emissions.  As a result, limited data  on particulate  emissions from stacks
are available and no data were available on fugitive  dust.  Particulate-
                                                                  (14)
emission data from point-source measurements were reported in NEDS
(National Emission Data System) and are given in Appendix B.  Although these

-------
                                       36
particulate emission data were not broken down into specific ingredients, these
data enable the calculation of overall average emission factors for raw-
materials handling and preparation on a worst-case basis.  The overall
emission rate is determined to be 29  mg/kg + 100 percent.  Total annual
particulate emissions for raw-materials handling and manufacturing are 459 Mg
+ 100 percent.  This is based on 12.7 Tg of glass produced, and is equivalent
to 0.0003 percent of the national particulate emissions from stationary
sources.
          Table 7 shows a breakdown of raw-materials handling for the various
                                                               (14)
points of emission.  This was determined primarily by NEDS data    , and
confirmed by observations made during plant visits.
          Most materials used to make glass containers have specified particle
size limits greater than 100 microns  (150 mesh).  Therefore, the amount of
material emitted from the plant site due to inertial forces will be minimal.
Composition of the particulates is given in a later section.

                                Glass Melting

          Container glass is predominantly melted in fossil-fuel-fired
furnaces in the United States.  Emissions from these furnaces are by far the
largest source of pollutant from a glass plant.  Primary pollutants are
categorized as NO , SO , and particulates.  The NO  is composed predominately
                 x    x                           x
of NO and the S0x composed predominately of SO .  Emissions of CO, hydrocarbons,
selenium, and other materials can and do occur.  The emission rate does
depend to some degree on glass type (i.e., flint or amber glass).
          The overall emission rates and total emissions for furnaces melting
flint and amber glasses are given in Table 8  as well as in Appendix B.  The
emission factors are based upon data reported in NEDS     and derived from
                          (13  18—19^
various literature sourcesv  '        .  Data referred to as source measurements,
reported in NEDS, were obtained by actual point source test measurements.   The
emission rates depend a great deal upon the operating conditions of the
glass-melting furnace.  For instance, N0x emissions factors are reported to range
from 0.58 g/kg to 6.29 g/kg, S0x from 0.21 g/kg to 8.35 g/kg, and particulates
from 0.13 g/kg to 1.95 g/kg.  Each type of emission is discussed £n greater

-------
                             37
     TABLE 7.   PARTICUIATE EMISSIONS DURING RAW-MATERIAL
               PREPARATION AND HANDLING FOR CONTAINER GLASS
    Process Step
Emission Factor,
     rag/kg
 Total Annual
Emissions(a>,
     Mg
Handling (unloading,
conveying
Glass crushing
Storage bins
Mixing and weighing
Batch charging
TOTAL
22 + 100%
1 + 100%
1 + 100%
5 + 100%
Negligible (b)
29 + 100%
348
16
16
79

459
(a)   Based on 15.8 Tg of raw materials processed to melt
     12.7 Tg of glass.

(b)   <0.1

-------
                                      TABLE 8.  EMISSIONS FROM FLINT AND AMBER CONTAINER
                                                GLASS-MELTING-FURNACE OPERATIONS






Emission Factor,
g/kg
Species
NO
X
S0x
Particulates
CO
Hydrocarbons
Selenium
Flint
3.40 +
1.84 +
0.71 +
0.06 +
0.08 +
0.002 +
43%
36%
30%
166%
100%
100%
Amber
1.22
0.93
0.48
0.11
0.05

+ 100%
+ 100%
+ 100%
+ 100%
+ 100%
0
Total
3.07 + 47%
1.70 + 42%
0.68 + 36%
0.07 + 143%
0.08 + 100%
0.002 + 100%
Total
Annual Emissions
Based on Glass
Manufactured^3) . Gg
Flint
36.57
19.79
7.64
0.65
0.65
0.02
Amber
2.32
1.76
0,91
0.21
0.09
0
Total
38.89
21.55
8.55
0.86
0.74
0.02
Percent of National
Emissions from
all Stationary
Sources
Flint
0.3179
0.3073
0.0062
0.0036
0.0024
—
Amber
0.0199
0.0028
0.0006
0.0011
0.0003
—
Total
0.3378(a)
0.3101
0.0068
0.0047
0.0027
—
(a)  10.758 TG for  flint,  1.898 TG for amber,  12.656 TG total, assuming a pack rate of 85 percent.

(b)  Standard for NO-.
                                                                                                                            00

-------
                                       39
 detail  for both  flint  and  amber  glasses.   As was  shown previously  (Figure 3),
 flint and amber  glass  comprises  approximately  85  percent  and  15 percent,
 respectively,  of the glass produced in the United States.   Green glass is
 included in  flint.  No differences  in emissions are  expected  between  green
 and  clear glasses.

 Nitrogen Oxides

          In a fossil-fuel-fired furnace,  nitrogen oxides  are formed  by a
 combination  of atmospheric nitrogen and oxygen at the  elevated temperatures
 (> 1500 C) required for making glass.   Because of the  high temperature,
 NO would be  expected to be the primary oxide of nitrogen  formed.   For purposes
 of this analysis, nitrogen oxides are designated  as  NO .   In  this  study,
                                                       X
 NO   is  compared  against an NO air  quality standard.   The  assumption  that
  X                           2.
 the  NO- emission factor is equal to the NO emission factor is believed valid,
      £"                                   X
 because once the plume has been  diluted sufficiently with  air (dispersion
 calculations show that the plume is diluted approximately  1000 to  1 at the
 point where  it touches the ground), the photochemical  conversion of NO to N0?
 is quite rapid.
          Nitrogen oxides  represent the largest fraction by mass (^ 54%) of
 emissions from the glass-melting furnace.   The formation of NO in a  glass-
                                                               X
 melting furnace  is extremely  temperature sensitive.  In one case,  NO
                                                                    X
 concentration  was increased some six times (from  100 ppm to MJOO ppm) as the
 furnace temperature (measured at the bridgewall)  increased from 1460  to 1550 C
                                               (18)
 and  the flint  glass production rate was doubled
          The  rate of  NO   formation depends upon  factors such as peak flame
                        X
 temperature, percent excess oxygen,  and posttime/temperature  history  of the
 flame.  Consequently,  considerable  variation in the  rate of NO emissions
                                                               Jb
 can and does occur.
*
          This report  considers  all nitrogen oxides  as N0x and does not attempt
 to determine the relative  proportions of each.  This differs  from  that method
preferred by the Glass Packaging Institute (GPI).  Copies  of  correspondence  from
GCMI (now GPI),  EPA, and Battelle relating to  this point  as well as other points
are given in Appendix  G.

-------
                                       40.
          Flint Glass.  Source measurements reported in NEDS and taken
 from the open literature give an average emission rate of 3.40 g of NO
                                                                      J^
 per kg of flint glass produced.  This average is based on 21 measurements
 (see Appendix B) and is calculated to be accurate to within + 43%
 at a 95 percent confidence level.  Individual values range from 0.58 to
 6.29 g/kg.  This amounts to approximately 36.6 Gg of NO  emitted annually
                                                       X
 from furnaces melting flint container glass.  This is equivalent to 0.32 percent
                                                          (20)
 of 1972 National N0~ emissions from all stationary sources

          Amber Glass.  Only three source measurements were found for NO
          *W*MIHIIBW*» 80 percent) of sulfur oxide emissions are derived from sulfur
in the oil.     Sulfur oxides from the batch generally combines with alkali
volatiles and exits as a particulate, while sulfur in the fossil  fuel  exits
predominately as SO .
                   X
          Glass generally contains about 0.15 weight percent sulfate  (added
usually as salt cake, gypsum, or blast-furnace slag).  This is  added  for
melting and fining purposes and is a necessary ingredient for making  container
glass.  A range of values for sulfate  (S0^=) in glass was reported in  1973  as
varying from 0.03 to 0.32 percent, with 82 percent of some  106  glasses
analyzed falling between 0.10 and 0.20 percent ^15\  The amount of mineral
sulfate added in the batch will, of course, be higher and usually falls within

-------
                                       41
the 0.5 to 1.0 percent range.  Sulfur oxide  emissions from the batch materials
        f-t n\
do occur     and these depend primarily upon the quality of glass
melted.
          Sulfur oxide emissions will be  greatly influenced by any switch from
natural gas  (the primary fossil fuel), which is essentially sulfur-free
to fuel oils or powdered coal containing  sulfur.   Such a trend does exist
primarily because of the reduced availability of natural gas in most sections
of the country.  Sulfur in  fossil  fuels readily oxidizes in the glass-melting
furnace and appears as SO   in the  exhaust gases.   For instance, a fuel oil
                         X
containing one weight percent sulfur emits approximately 600 ppm (calculated
                       (21)
as S02) in the flue gasv  '.
          Source measurements reported in NEDS or  in the literature are
essentially for natural gas-fired  furnaces.   Hence, the emission are not
fully representative of an  industry which is gradually switching
to fuels containing sulfur.  However, such emissions will essentially
correspond directly to the  sulfur  found in the fuel oil.

          Flint Glass.  Source measurements  for flint glass give an average
SO  emission rate of 1.84 g/kg.  This is  based on  46 point source measurements
  x
(see Appendix B) and is calculated to be  accurate  to within +0.36% at
a 95 percent confidence level.  The values ranged  from 0.21 to 8.35 g/kg.  This
gives an estimated total annual emission  of  SO of 19.8 Gg, which is equivalent
                                               X                           (20)
to 0.31 percent of 1972 National SO  emissions from all stationary sources    .
                                   X

          Amber Glass.  Again, data were  only available from three sources
for furnaces melting amber  glass.  These  gave an average S0x emission rate
of 0.93 g/kg.  The three values were 0.32, 2.06, and 0.41 g/kg.  As can be
seen, these values are essentially equivalent to those for flint glass.  Total
annual SO  emissions from furnaces melting amber glass are estimated to be
         X
1.76 Gg which is 0.003 percent of  1972 National S0x emissions from all
stationary sources    .

-------
                                        42
 Particulates

           Particulates from glass-melting  chamber  can  originate  both  from
 physical entrainment of batch materials  being  charged  to  the melting  furnace
 and from condensation of compounds,  such as  sodium sulfate  (which  forms  through
 the combination of  sulfur oxides  and volatilized sodium).   Particulates  exiting
 with exhaust gases  are essentially all  (>95  percent) condensates,  as  indicated
 by the fact that  collected material  is almost  entirely water soluble.  Studies do
 show that batch materials are carried out  of the melting  chamber by the  combustion
 products; however,  such materials do not show  up in the stack-gas  samplings;
 therefore, it is  assumed these coarser batch materials are  retained in the
         „     _   (22,23,24)
 furnace-flue system.
           Considerable opinion exists as to  the mechanism by which condensate
 particles are formed; however, analysis  of these particulate emissions show them to
                                                                (9  18  22-24)
 consist predominately (>75 percent)  of submicron sodium sulfate   '  '       .
           The formation of these  particulates  depend upon batch  composition,
 temperatures in the melting furnace, production rate,  surface  area of molten
 glass, and cullet ratio.  Of these factors,  production rate, temperature,
 and surface area  of molten glass  are the most  important factors  affecting
 the rate of particulate emissions.  Since  these variables are  interrelated,
 it is difficult to  determine the  relative  influence of each, although it
 would appear that temperature is  the most  significant  variable.   Data from one
 furnace melting a soda lime glass showed that  at zero  production rate, the
 particulate emissions were approximately 20  percent of that measured  at  its
                        f-t Q\
 normal furnace capacity    .   Temperature  was  maintained  at a  constant value
 (1450 C).   Emissions ranged from  1.8 kg/hr at  zero pull to  7.7 kg/hr  at
                                       (18)
 normal pull of 211  Mg/day.   Other data   ' collected on soda/lime glass
 during the study  indicated that particulate  emissions  followed an Arrhenius
 curve when plotted  against the reciprocal  of temperature; that is, a linear
 relationship  with the logarithm of the emission rate.  This is shown in
 Figure 12.
                              (24-25)
           Stockham  and others        studied the emissions  from furnaces
 melting  a  flint and an amber glass and found the geometric  mean particle
 size  of  the particulate emissions to be  0.13 micron for  the flint and 0.11
micron for the amber.

-------
    100
     50
                                43
 (3D


"eo  10
 
-------
                                     44
           Flint  Glass.   Source measurements for particulate emissions  from
 flint glass-melting furnaces  give  an  average emission rate of  0.71  g/kg.  The
 emission rate varied from 0.22 g/kg to  1.95 g/kg.   Source measurements were
 taken from 66 points and are  calculated to be accurate within  + 30%
 at a 95 percent  confidence level.  This  represents an estimated total annual
 particulate emission of  7.64  Gg» or 0.006 percent of the 1972  National
 particulate emissions from all stationary sources    .

           Amber  Glass.   Source measurements for particulate emissions  from
 furnaces melting amber glass  give  an average emission rate of 0.48 g/kg.
 The emission rate varied between 0.13 g/kg and 0.83 g/kg.  The average is
 based on 23 source test  measurements  and is calculated to be accurate  within
 + 41% at a 95 percent confidence level. This represents an estimated
 total annual particulate emission  of  0.91 Gg or 0.001 percent  of the 1972
 National particulate emissions from all stationary  sources     .

 Carbon Monoxide
           Carbon monoxide  can be emitted through incomplete  combustion
of the  fossil  fuel  through the use of a luminous flame, or by  reaction of  a
powdered coal  added to  the glass batch to reduce sulfate  compounds.   The
emission rate  varied between 0.05 and 0.13 g/kg.  An estimated emission rate
for flint  glass is  0.06 g/kg, based upon 19 reported measurements  and accurate
to within  +  167%  at a  95  percent  confidence  level.   This  would
represent  an annaul emission of 0.65 Gg of CO, or 0.004 percent of 1972
National CO emissions from all stationary sources     .  Only three source
measurements are available  for amber glass.  These  show an emission rate of
0.12,  0.10, and 0.09 g/kg  for an average of 0.11 g/kg.  This would represent
an annual emission  of 0.21  Gg for CO from furnaces  melting amber glass.

-------
                                      45
 Hydrocarbons

           Hydrocarbon emissions form in glass-melting furnaces primarily
 through the incomplete combustion of a fossil fuel.  The emission rate varied
 from 0.01 to 0.53 g/kg.  An estimated average emission rate for flint glass
 is 0.08 g/kg, based upon 33 measurements calculated to be accurate within
 + 100%.   Such an emission rate represents an annual emission of
 0.65 Gg of hydrocarbons by manufacturers of flint glass, or 0.002 percent of
 1972 National emissions for hydrocarbons from stationary sources    .   Only
 three measurements were available for amber glass.  These are 0.06, 0.03,
 and 0.05 g/kg, which give an average emission rate of 0.05 g/kg or total annual
 emission of 0.09 Gg.

 Selenium

           Selenium is used by flint glass manufacturers as a decolorizer to
 neutralize the tint from transition metal oxide contaminants such as iron.
 It is usually used in amounts of up to 0.001 percent. Test measurements on
selenium emissions were not found.   Selenium volatilizes at rather low
                                           (9 22)
temperatures  (315 C  for SeO, 685 C for Se)   '    ;  therefore, it can be expected
to be present in the waste gases.  If the  temperature of the waste gases is
dropped below 200 C, then selenium condensates are likely to be found.
          The maximum amount of selenium release can be determined from the
amount of selenium consumed annually by the  glass  container industry.  A
                                               (32)
minimum of 60 percent is retained in the glass.      Some 0.06 Gg of selenium is
                                                                                   (2)
consumed annually by the glass container industry  (about one-sixth of U. S. usage).
Under these circumstances, which is  believed to be a worst case, the emission
rate for selenium would be 0.002 g/kg.  Total annual emissions would be 0.02 Gg.

-------
                                     46
Other Emissions

          Other minor emissions can include antimony and arsenic, which in
the past were added as fining and decolorizing agents.  Both of these
materials have been virtually eliminated from use in recent years,  Similarly,
chlorine was emitted in the past because of its association with soda ash
produced by the Solvay process. In recent years, most of the glass-container
industry has switched from soda ash produced synthetically to that manufactured
from a naturally occurring ore which does not contain Cl.  By 1977, more than
90 percent of the industry will be using natural soda ash  '

                             Forming and finishing

          Molten glass, properly conditioned, leaves the forehearth of the
melting furnace, where it is cut into individual "gobs", which are then
transferred to the forming machine.  The gob is formed into a container
by the blow-and-blow or press-and-blow method.  After forming, a hot-end
coating or surface treatment may be applied, followed by the annealing
operation.  Following this, the containers may then undergo a variety of
decorating or coating operations.
          Emissions from the forming and finishing operations can include
hydrocarbons emitted during the forming operations, HCl and metal oxides
emitted during surface-treatment operations, emissions associated with combustion
gases produced during annealing, and organic fumes emitted from the coating
and decorating operations.
          Very little emission data are available from the forming and
finishing operations.  This section discusses emissions according to four
operations:  forming, treatment, annealing, and decorating.

Forming

          Gob shears, delivery chutes, and the forming molds  for  container
glass are lubricated with various solutions.  These solutions  can contain
grease,  oils,  graphite, and silicon-based emulsions.  In the past decade,

-------
                                     47

there has been a transition from grease and oil lubricants to the use of
silicone-emulsions and water-soluble oils  (1 part silicone or oil to 90-150
                                                    (Q\
parts water) on gob shears and gob-delivery systems   .  Grease and oils are
still utilized on molds.  Observing the forming operation, one can frequently
see a puff of white smoke occur when the molds are  swabbed with a lubricating
solution.  Although the smoke dissipates in a few seconds, hydrocarbon vapors
are probably released.  These emissions are probably drawn through the large
ventilators on the roof of the plant.  Emission data on hydrocarbon emissions
was not available; however, engineering calculations (Appendix B) indicate
that the maximum emission rate is low  (0.03 g/kg).  Total emissions are
estimated to be 0.44 Mg or less than 0.0001 percent of National Emissions.

Surface Treatment

          Some glass containers receive a metal oxide (titanium or tin)
surface treatment to improve their resistance to scratching.   Additionally,
this transparent treatment acts as a lubricant which can facilitate handling
and shipping operations.  The oxide treatment is obtained by subjecting the hot
container (coming from the forming machine) to a vapor of metal chloride.
This is done within a hood.  The metal chloride pyrolyzes to the metal oxide on the
container surface, leaving a metal-oxide film and releasing hydrogen chloride.
The bottles then go to the annealing oven.
          Emissions from the surface-treatment operation will consist of HC1,
metal oxides, and hydrated-metal chlorides.  Anhydrous tin chlorides which
do not react with the glass will decompose by the action of heat and moisture
within the exhaust ductwork to form metal oxides, hydrates metal chlorides,
and HC1.  Estimations based upon available data  '    indicate that approximately
60 percent of the total weight of the metal chloride input is released into
the atmosphere.  Using tin tetrachloride as the input material, these
estimations reveal that of the total weight input, 14 percent is released
into the atmosphere as a metal oxide, 27 percent as hydrated tin chloride,
and 21 percent as HC1.
          Emissions from the surface-treatment operation were determined by  eng-
ineering calculation to be 0,02 g/kg of tin or titanium oxide, 0.03 g/kg of hydrated
tin or titanium chloride, and 0.02 g/kg of HC1.  Total emissions are  estimated to  be

-------
                                     48
 0.08 Gg of metal oxide, 0.11 Gg of hydrated metal chloride, and 0.08 Gg of HC1
 annually.   This estimate  is based on  30  percent  of  the  containers being
 treated.

 Annealing

           All glass containers undergo an  annealing operation, where  the
 glass is brought to a temperature  (approximately 550 C)  to remove residual
 stresses and then cooled  uniformly to about  150  C where  they  are  removed  from
 the annealing lehr (oven).  All lehrs used by glass container manufacturers  are
 heated by natural gas, with propane as a possible alternate fuel.
           The only emissions  from annealing  lehrs are combustion  products.
 Since natural gas  is used exclusively, although  lehrs can be  heated
 electrically and the temperatures are relatively low, emissions are low.
 Measurement  data are not  available and emission  rates were estimated  on  the
 basis of emission factors for  the combustion of  natural  gas.  The results
 are given in Table 9.  Total  emissions were  calculated  on the worst case
 basis of all product being annealed in gas-fired lehrs.

 Decorating

          Glass  containers are sometimes decorated  with  vitrifiable glass enamels
 or  organic materials.   A wide variety of decorating techniques  are employed.
 Emissions occur predominately from organic solvents and  binders used  in
 decorative coatings which are released during the curing of these compounds.
 Data supplied by three glass container manufacturers would indicate  that only
 3 percent of all (both flint and amber)  containers  have  decorative coatings.
This amounts to 330 Gg of container glassware decorated  annually.  Based  upon
                                 (14)
5 material balances given in NEDS     , a hydrocarbon emission rate of 4.37 g/kg was
determined for container decorating.  No point source measurements were  available.
This amounts to 1.44 Gg of HC emitted annually.   This represents  0.005 percent  of
the National HC emissions from all stationary sources.

-------
                                     49
                 TABLE  9,   EMISSIONS FROM THE ANNEALING OF
                            SODA/LIME CONTAINER GLASSES




Species
NO
X
SO
X
Particulates
CO
Hydrocarbons
Total



Emission
8/kg
0.0015

0.025
0
0
0.035




Factor
(Ib/ton)
(0.003)

(0.05)
0
0
(0.07)



Total Annual
Emissions^ ,
Gg (ton)
0.019 ( 21)

0.316 (329)
-
-
0.443 (488)

Percent of
National
Emissions
from all
Stationary
Sources
0.0002

0.0005
-
-
0.0017

(a)  Based on 12.656 Tg of glass  processed.

-------
                                     50
                            Emission Characteristics

 Raw Materials Preparation

          Emissions from this part of the manufacturing process will reflect
 the raw materials used  soda ash, limestone, feldspar, sand and sodium
 sulfate), since no chemical reactions take place.  Soft materials like
 limestone and soda ash will be more easily crushed to dust.
          Manufacturers generally specify particulate sizes ranging from
 820 to  44 micron (-20 to +325 mesh).   Table 10 illustrates general specification
 limits  for several raw materials used to make glass containers.  Glass sand
 would not be expected to cause significant dusting since only a small fraction
 is below 100 micron and all particles are greater than 44 micron.
                                                                       o
          The primary ambient air standard for particulates is 260 ug/m .

 Glass Melting

          Emissions from the melting furnace consist of the criteria pollutants:
 NO  , SO , particulates, CO and hydrocarbons, as well as selenium.  These
  A    X
 emissions contribute to photochemical atmospheric reactions to produce smog and
 can be irritating to the lungs.
          Particulates consist predominately of sodium sulfate (>85 percent).
 It is unclear as to whether these sulfate emissions pose a health hazard^   .

Forming and Finishing

          Emissions from the forming and finishing operations consist of
the following:

          (1)   N0x,  particulates, CO, and hydrocarbons emitted from
               gas-fired annealing lehrs.
          (2)   Hydrocarbons produced by flash vaporization of mold lubricants
               used  in forming  glass  containers.
          (3)   HC1,  tin or  titanium oxide, and hydrated metal chlorides
               exhausted from fume chambers during surface treatment operations.

-------
                                    51
           TABLE 10.  GENERAL SPECIFICATION LIMITS FOR RAW
                      MATERIALS USED IN CONTAINER GLASS MANUFACTURE
Material Specifications Range
Mineral
Cerium oxide
Dolomite
Feldspar
Limestone
Sand
Soda ash
Sodium Nitrate
Chemical
Formula
Ce02
(Ca,Mg)C03
-
CaC03
Si02
Na2C03
NaN03
Amount ,
I
-
0.5
2
002
3
0
1
Mesh*
-
+16
+40
+20
+30
+20
+6
Amount ,
100
50
10
0.5
6.6
4.2
1.5
Mesh*
-60
-100
-200
-300
-100
-120
-100
*  U. S. standard mesh size - see Table 15
   for micron equivalents.

-------
                                     52
           (4)  Hydrocarbons produced during the curing of decorative
               coating used on some glass containers.

           Total  nationwide emission of the criteria pollutants produced in the
 different stages of  the  container glass-manufacturing process were listed
 previously in  Table  2.
           At a glass plant, the major amount of atmospheric emissions comes
 from the melting furnaces.  Calculations to portray the effect which a glass
 plant has on its neighboring  air environments have been made in the
 following sections.   Principal attention has been given to the pollutants issuing
 from the melting furnace stacks.

                         Ground-Level Concentrations

           Ground-level ambient concentrations of pollutants were used in
 determining the  environmental effects of the atmospheric emissions.  These
 were calculated  for  representative operations used in the manufacturing of
 glass containers.  A single plant having an annual production of 104.8 Gg
 (115,000 ton)  was  selected as being representative, following the calcula-
 tions made in  earlier portions of this report.
           Stack  heights  for container-glass furnaces, as given in the NEDS
 listing,  ranged  from 5 to 50  meters with the predominant height being about
 40 meters.   Two  other frequently occurring stack heights are in the vicinity
 of 20 m  and 45 m.  There is no correlation between stack height and production
 rate.  In general, stacks taller than 30 m can be assumed as natural draft
 stacks or natural  draft  stacks that have been converted to an ejection-air
 system.   A mean  stack height  of 38.2 m (125 ft) was selected for the melting
 furnaces  in a  representative  plant.  Other stack parameters are chosen from the
 NEDS data for  their  compatibility with a 38 m stack.  The furnace-stack
 emissions are  derived from the emission factors given in Table 2 and are applied
 to the 104.8 Gg  annual production rate.  All of the parameters for the melting
 furnaces  and stacks  are  listed in Table 11.
           The maximum ground-level concentration was used to determine infor-
mation for  the environmental  effects criteria.  This maximum concentration can

-------
                                      53
TABLE 11.  PARAMETERS FOR MELTING  FURNACES  OF A REPRESENTATIVE PLANT  IN  THE
           GLASS-CONTAINER  INDUSTRY AS USED IN  ATMOSPHERIC-DISPERSION CALCULATIONS
     Stack Parameters

          Glass produced:  104.8 Gg/yr  (115,000 T/yr)
          Stack height:  38.2 m  (125 ft)
          Stack  diameter:   1.53 m  (5 ft)
          Exit  temperature:  353 C (650 F)
          Gas flow rate: 910 m3/min (32,000 ACFM)
          Exit velocity:  8.35 m/sec (27.2 ft/sec)

     Meteorological Conditions

                                                (a)
          Wind speed:  at 10 meters  —  4.1 m/sec     (9.2 mph)
                       at top of stack  — 8.2 m/sec(b)  (18.4 mph)

          Ambient temperature at top of stack:  15 C  (59 F)
          Atmospheric pressure:  1000 millibars
          Atmospheric stability:  D(C)

     Calculated Parameters

          Plume rise:  5.72 m(d) (18.7  ft)
          Effective stack height:   43.9 m (144 ft)

     Estimated Parameter

          Mean wind speed affecting the plume between the effective stack
          height and the surface:   6 m/sec

     Emissions (G)

          NOX:  H.25  g/sec  (390  T/yr)
          SO  :  6.14 g/sec  (213 T/yr)
          Particulates:  2.73 g/sec  (94  T/yr)
          CO: 0.2 g/sec  (6.9 T/yr)
          Hydrocarbons:  0.2  g/sec  (6.9  T/yr)

(a)  Average of annual mean wind speeds measured at city airports
     near 30 glass-container plant  locations.
(b)  Increase of wind  with height in suburbs and level country as
     given in Figure 1-3 of ASME Recommended Guide for the Pre-
     diction of the  Dispersion of Airborne Effluents. 1968.
(c)  D  stability is  the predominant stability as determined from a
     cross section of  Star Program results (see Table 12).
(d)  Plume rise  was  calculated from the Holland equation for neutral
     stability.

-------
                                     54
be obtained for substitution into an equation or from a nomograph.  The
equation is


                                        f

where
          X    = maximum concentration (gm/m )
           max
             Q = pollutant emission rate (gm/sec)
             u = mean wind speed (m/sec) at the height of the stack
             H = effective stack height (m) , the physical height of
                 the stack plus the plume rise
            CTZ = vertical plume standard deviation (m)
            a  = horizontal plume standard deviation (m)
             e = base of natural logarithms, 2.718
             rr = 3.14.
                                     CT
                                      Z
          For D stability, the ratio CT  is on the order of 0.5 varying from
                                      y                     (27>
0.57 to 0.24 between 0.1 km and 10 km downwind from a source    .   The ratio is
approximately 1.0 for C stability.  The maximum concentration occurs at a
                                 ( 27^
x    u/Q and the distance to the point of maximum concentration can be
distance where a  = h//2.  Turner     has presented a monograph from which
                Z
determined for any stability and effective stack height.  When emission rate
and wind speed are known, the value of X    can be calculated.
                                        max
          The environmental effect criteria are developed for 24-hr average
concentrations, while the dispersion predictions discussed above are for
short periods (3 to 10 min).  For longer periods, one must consider that
variations in wind direction and wind speed would cause the average concen-
tration at a downwind monitor to be less than the concentration calculated
for a short-term wind blowing constantly from the source to the monitor.
Turner has given an equation by which the long-term average concentration
can be estimated when the short-term concentration is known:

                           X  -I  fi
                            f    <5 V 1-
                            *    S ^ t „

-------
                                     55
vhere
          X^ =  concentration for  the  long period (t.)
          Xg =  concentration for  the  short period (t  )
          The value  of  the  dimensionless  exponent,  b, is
          between 0.17  and  0.2
          t. =  long-time  period,  min
          t  =  short-time period, min.
           s

While  this equation  is  most applicable  for X. =  2 hr or less, it can be
                                             Xi
applied  to a 24-hr period.   Turner  gives  the conversion coefficient of 0.35 for
transforming a  3-min average into a 24-hr average.  Other conversion coefficients
are  1  hr, 0.61, and  3-hr, 0.51.
          Before calculating ambient  pollutant concentrations, representative
meteorological  parameters for the area  need to be chosen.  These parameters
are  required for determining plume  rise and dispersion.
          Plume rise was  calculated from  the Holland equation
AH = -a-  (l.5 + 2.68 x 10"
                                                T
                                                 s
where
           AH = rise  of  the plume above the stack, m
           v  = stack gas  exit  velocity, m/sec
            S
            d = inside diameter of stack, m
            u = wind  speed at top of stack, m/sec
            p = atmospheric pressure, millibars
           T  - stack gas  temperature, K
            s
           T  - air temperature, K.
            3.

           Choices of the meteorological parameters were made after a review
of climatology  in some  of the  areas of the .country where glass plants are
      (29)
found.      Account was also taken of the variations of meteorology between the
surface and  the top  of  the furnace stack.  The values selected for the melting
furnace calculations are listed in Table 11.   Stability Type D (neutral class)
is the most frequently occurring stability throughout the United States as

-------
                                      56

                                (27)
 calculated  by  the  Turner method     which considers the surface wind speed and
 the net  radiation  (Table 12).  A surface wind speed of 4.1 m/sec was chosen
 as  representative  of  the conditions at the glass plants based on a survey of
 the average annual wind speeds  listed for the National Weather Service
                       ( 28^
 meteorological  stations     located at 30 cities which have container glass
 plants.   It should be noted that 4 meter/sec wind speeds in Turner's scheme
 for determining stabilities can accompany stabilities varying from Type B to
 Type E,  depending  on  the solar radiation.  Type D was chosen for the dispersion
 calculations on the basis of its predominant frequency.
          Wind  speeds increase with altitude and this effect was taken into
 account  for the 38.2  meter effective mean furnace stack height for the represen-
 tative plant.   Wind speed in the layer in which the downward dispersion of the
 plume would take place, 0-44 meters, was estimated to be 6 meters/sec.  This
 was an extrapolation  from the standard wind-measurement height of 10 meters
                 (29)
 following a guide     giving examples of the variation of wind with height
 over suburban and  level county areas.  For stack heights of 30 to 40 meters, the
 wind speed  is expected to be 1.5 (level terrain) to 3 (urban areas) times
 stronger at the top of the stack than at 10 meters.
          Table 13 (in its second column) presents the maximum pollutant
 concentration predicted for ground level in the vicinity of the representative
 glass melting plant.  These concentrations are the contributions from only
 the melting furnaces, and do not take into account other glass plant emissions
 or  emissions from sources other than the glass plant.  Table 13 also presents
 data for selenium, a minor pollutant emitted by a container glass furnace.
          finissions from three other sources representative of air emissions
 from a container glass manufacturing operations were also considered in
 relation to  their effect on ambient-air quality.  These were:
          (a)   Particulates from a baghouse collecting the emissions
                from materials handling
          (b)  Hydrocarbons from a container-decorating operation.
          (c)  Tin oxide and hydrated tin chloride particulates
          (d)  HC1 from a surface treatment operation.
To make the ambient-concentration estimates for these sources, emissions and
stack parameters were adapted from data given in the NEDS listing.  Meterological

-------
                                   57
      TABLE  12.  RELATIVE FREQUENCY OF ATMOSPHERIC STABILITIES
   Station
                                  Stability Class
                                 D       E and  F
Milwaukee
0.001      0.031    0.094     0.636       0.238
St. Louis         Q005      0.047    0.103     0.555      0.289
Peoria
0.003      0.042    0.102     0.577        0.276
Pittsburgh       0.001      0.022    0.083     0.567       0.306
Columbus, Oh.    0.010      0.058    0.100     0.500       0.331
Mobile
0.008      0.052    0.115     0.453       0.371
 Los  Angeles      0.001      0.041    0.148     0.482       0.329
Dallas
0.004      0.042    O.io?     0.586       0.262
*  Based on Output from U.S. Department of Commerce National Climatic
   Center Star Program for Five Years of Data.

-------
                                     58
TABLE 13.  MAXIMUM POLLUTANT CONCENTRATIONS AND SOURCE SEVERITY FOR
           EMISSIONS FROM THE MELTING FURNACES FOR A REPRESENTATIVE
           GLASS CONTAINER PLANT
Pollutant
NO
X
SO
X
Particulates
CO
Hydrocarbons
Selenium
Quality Standard
100 (a)

365
0.79(f>
0.95
0.023
Severity,
S
0.38

0.056

0.035
2.0xlO"5
5.9x!0"3
3.4xlO~2
(a)  Annual arithmetic mean assumed here as 24-hr standard for N0_.
(b)  24-hr standard.
(c)  8-hr standard.
(d)  3-hr standard.
(e)  Threshold Limit Value for 8 hrs.
(f)  3-min Xmax adjusted to match sampling time of the standard using
                      X    = X
                       max    max-
                                          std
(g)  3-min X    adjusted to 24-hr sampling time.

-------
                                      59
 conditions  similar to those used in the glass-furnace emission-dispersion
 calculations were used for these other sources with adjustments for differing
 stack heights.   Information regarding two of these calculations is  given in
 Table 14 m   The maximum amounts of coating used for surface treating as  listed
 in the NEDS is  6 tons.  Calculation of tin and HC1 emissions from the indicated
 values are  on the order of 50 ^g/sec which are considered to be negligible.
          For each of the maximum ambient concentrations that were  calculated,
 a source  severity, S, was also determined.  Source severity for criteria
 pollutants  (particulates, sulfur oxides, nitrogen oxides, carbon monoxide, and
 hydrocarbons) is determined from the following equation:
                                      X
                                  o _  max
                                      AAQS
 where
          X    — maximum average ground-level concentration of
            max
                  the pollutant for the time period of the
                                3
                  standard (|J,g/m )
                                               3
          AAQS = ambient air-quality std (p,g/m ).

 For noncriteria pollutants,  the source-severity equation uses the threshold
 limit value instead of the ambient air-quality standard with a  correction  for
 a 24-hour period and a safety factor:

                            S =        max    	
                                TLV (8/24) (1/100)
where
           TLV = Threshold Limit Values for each species
         8/24 = correcting factor for the 8-hr work day which is
                 the basis for the TLV
        1/100 = safety factor.

          A review  of the source-severity factors in Tables 13 and 14 shows
the highest value to be that produced by nitrogen oxides emitted from the
melting furnace stacks.   For the  conditions  used to portray a representative

-------
                                      60
 TABLE 14.  MAXIMUM AVERAGE GROUND-LEVEL CONCENTRATIONS (X   ) OF AIR
            POLLUTANTS FROM CONTAINER GLASS-PLANT SOURCES BESIDES THE
            MELTING FURNACE
 Source 1.  Baghouse Controlling Materials-Handling Emissions

      Emissions: 0.004 grams/sec (0.1 tons/yr)  of particulates

      Emission Point -- Stack
               height - 35 m; diameter = 0.3 m;
               exit temperature = 10 C; exit velocity = 17.3 m/s
                Xmax(3min)'       Xmax'  ^S/I"3'         Ambient    Severity
   Species          u,g/nr        (specified time)        Std       Factor

 Particulates       0.06           0.02  (24 hr)         260 |0,g/m3  9.2 x 10
 Source 2.   Container Decorating

      Emissions:  0.44 grams/sec (15  tons/yr)  of hydrocarbons
      Emission Point  --  Stack
               height -  12.2  m; diameter = 0.6 m;
               exit temperature = 149  C;  exit velocity = 11.4 m/s
-5
Species
Hydrocarbons
Xmaxj3Jin)
47
Xmax' M
-------
                                    61
glass container plant, the source severity factor  is 0.38, a low value.
All other severity  factors, including those from sources other than the
melting furnace are also quite low.

Affected Population

          As a consequence of the dispersion of pollutants, the severity
starts at zero near the  stack, increases downwind,  reaches a maximum, and
then decreases to zero again (see Figure 13)„  The affected population is
defined as the population around a representative  plant exposed to a severity
greater than 1.0.  To determine the downwind distances enclosing the affected
population, the standard dispersion equation for the centerline concentration
from an elevated source  is used.
                                          c-ol
where
                                                   -3
          X = pollutant concentration at surface  (gm )
          u «= average wind  speed through the dispersion
              layer (m sec   )
          Other parameters  are the same as in the earlier
          dispersion equation.
          It is assumed that winds from all directions are equally likely.
By rearranging, this equation becomes
                      XTTU       1          1 / H   \2~1
                      ~  "  yT   exp  ~^JJ-

The value of X is specified by the requirement for S = 1.0 and then it  is
corrected to the three-minute average concentration which the dispersion
equation gives.   Substituting values of a  and a   from Turner's graphs  of
dispersion coefficient as a function of distance downwind into the right-hand

-------
                       62
X
 *»

c
o
0)
u

o
o

c
o
o
QL
                    o
                    o
                    0)
                    >
                    0)
                    Crt

                    0>
                    o
                    1_


                    I
                                                    1.0
                   •max.
        Distance Downwind,X
R2-

R , -
Outer radius

Inner radius
FIGURE 13.   ILLUSTRATION DEPICTING CALCULATION OF AREA

            WHICH CONTAINS THE AFFECTED POPULATION

-------
                                     63


 side of the equation will produce a  short  table  for  values  of the  right-hand

 side of the equation versus downwind distance.   These are plotted  in a

 fashion similar to Figure 13 and the values of 1^ and R^ are  determined.

 These values form the inner and outer radii of an annulus enclosing the
 affected population.

          In calculating the number  of people affected, it has been assumed
 that the population density around the representative glass plant  is 248
people per square kilometer.

          Since none of the pollutant sources produced severity factors

greater than 1.0, all affected population values for the container glass
sources are zero.

          As an illustration calculations were made for nitrogen oxides from
the melting furnace using a severity factor of 0.1 instead of the standard 1.0.
Results of this calculation is presented below.
  Source/Emission
Concentration (3 min)
            Downwind
Value,      Distance,
|xg/m3
 m
                                                    Affected Population *
                                                       Radii
         Inner,   Outer,   Persons
           m       m     Affected
Melting Furnace/NO
                  x
 105
880
420
3900
11,700
*  Based on the more restrictive source severity factor of 0.1 instead of
   the accepted 1.0.  There were no glass container sources which produced
   a source severity factor greater than  0.4.

-------
                                     64
                                  SECTION V

                              CONTROL TECHNOLOGY

          Control of emissions in the glass-container industry varies
considerably, depending on the type,  source, and amount.  Control technology
has evolved for both economic and environmental reasons.  Various methods
are utilized to reduce air emissions  from the different portions of the glass-
manufacturing process.  These include:  (a) development of process modifica-
tions, (b) new furnace designs, and (c)  application of control equipment.
For example

          (1)  Arsenic is no longer used as a fining agent
          (2)  Many fossil fuel-fired furnaces are equipped with
               electric boosting which can increase output, thus
               reducing the amount of effluent per unit of output
          (3)  Fabric filters, electrostatic precipitators, and scrubbers
               are being used or have been examined for removing dust
                                    i
               particulates.  In addition, several commercial equipment
               manufacturers are attempting to develop methods for
               removal of SO  and NO  emissions at the same time
                            X       X
               particulates are removed.

This section discusses the control technology currently being used or that
which might be considered for use by the glass-container industry.  It was
not the intent of the study to consider the economics associated with the
control technology or even to verify the technology itself.  Rather, the
purpose of this section is to identify control technology reportedly applicable
to the control of emissions from glass manufacturing plants.  The discussion
is organized in a manner similar to the emission section.

-------
                                     65
                          Raw-Materials Preparation

          The handling and mixing  of  raw materials is a source of particulate
emissions from a container glass plant.  Raw materials are usually conveyed
from hopper railroad cars or  trucks  (by screw  conveyors, belt conveyors,
bucket elevators, or pneumatic  conveyors)  to elevated storage bins, as was
shown previously in Figure 5.   A few  minor glass-batch ingredients are
delivered to the plant in paper bags  or cardboard drums.  These are later
transferred by hand to smaller  storage bins or fed directly from the storage
device.
          Materials are  gravity fed from the storage bins into weigh"hoppers
and then transported by  transverse belts or bucket elevators into a mixer.
Gullet is crushed to a desired  size.  After mixing, the glass batch is
transferred to a charging bin located next to  the glass-melting furnace
or into a batch storage  bin,  depending upon the design of the batch-handling
system.  Positive displacement  or  vibratory feeders at the bottom of the bins
feed the materials to the chargers, where  it is fed into the glass-melting
furnace.  Gullet may be  added to the  batch in  the mixer, while the batch
                                                                  (8)
is being transferred, or charged separately to the melting furnace   .

Emissions

          Little information  is available  regarding plant emissions due to
dusting during the raw-materials handling  stages of the process.  As
discussed in the previous section, the fraction of the dust generated which
leaves the plant site will consist of particles smaller than 100 microns in
diameter.
          As described in the previous section on emissions, the particulate
emission rate for raw-materials handling is estimated to be 22 mg/kg.
Based upon the total glass batch handled by the container industry, particulates
exhausted annually would average 348  Mg.

-------
                                     66
           Information  on  the  composition of these participate  emissions  is
 not available,  but  they will  consist essentially of the same raw materials
 being handled (soda ash,  silica, limestone, etc.), since no chemical
 reactions occur during this portion of the manufacturing process.   Limestone
 and soda ash can be expected  to predominate because of their relative
 softness.  Glass manufacturers generally specify raw materials which are
 coarser than 100 micron,  as  shown  in  Table  15.   Consequently,  the  amount of
 raw material emitted from the plant site due to inertial forces alone would
                                                                     (14)
 be relatively small.   This is in line with the reported measurements

 Raw-Materials-Control  Technology

           Process Modif icajijlonji ^                     Container  glass
 manufacturers have  minimized  dusting problems in batch-handling operations by
 limiting the amount of fine particles  (<100 microns) in the batch  material,
 as can be seen in Table 15.   Specifications for glass-grade raw materials
 will generally require removal of  the finer sizes of material, especially
 with softer materials  that crushed to dust easier than sand.
           Another batch-preparation method that is used to control dusting
 during handling is  the addition of water to the raw batch  (batch wetting).
 Trials have also been  conducted where the batch is wet with a  liquid  caustic-
                                              (32)
 soda solution that  is  substituted  for soda ash    .  Water is  presently  added
 in amounts  up to 4  percent to the  mixed batch materials.  The  substitution  of
 a  caustic-soda  solution for a soda ash is not generally practiced  by  the
                        (32-33)
 glass-container  industry

          Efficiency of Control Equipment.  Transport of raw materials  in
railroad hopper  cars and hopper-bottom trucks (dump trucks) is still
practiced.  During unloading  of these trucks or railroad cars, materials
dumped onto conveyor belts can result in some dust being dispersed into the
air.  Generally,  the hopper cars or trucks are connected to sealed receiving
hoppers with  fabric sleeves and the dust generated during  the  unloading operation
is filtered through the sleeves or exhausted through a baghouse^       .

-------
TABLE 15.  GLASS-GRADE PARTICLE-SIZE SPECIFICATIONS FOR
           SAND, LIMESTONE, AND 10- AND 20-MESH DOLOMITE
Approximate
Particle
Size
2.3 mm
1.3 mm
820 \i
410 ji
150 jj,
105 p.
74 p,
44 y,
U.S. Standard
Mesh Size
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Cum retained on
Glass- Glass -Grade
Grade Glass-Grade Dolomite, %
Sand, % Limestone, % 10-Mesh
8 - 0.0 0.0
16 - 2.0 max 15.0 max
20 0.0 10.0 max
40 12.0 max
100 - - 90.0 min
140 92.0 min 85.0 min
200 99.5 min 94.0 min 97.0 min
325 100.0 min
20-Mesh
0.0
-
2.0 max
-
80.0 min
95.0 min
96.0 min
-
                                                                                 a\

-------
                                     68
 Enclosing the loading area with  a  suitable fabric structure  and  sealing
 all covers and access openings with gaskets is effective in  reducing  dust
 during this operation.   This  results in an inward-air velocity across the  open
                                                                       (35)
 mouth of the bag that prevents an  eruption of dust into the  atmosphere
 Trapped air and fine dust can then be filtered by a conventional fabric
 filter and the cleaned air exhausted into the atmosphere.
           Weigh hoppers and mixers require ventilation because of surges in
 material from the large air flows.  In older mixers, polyvinylchloride seals
 are installed between the rotating body of the mixer and its frame  to reduce
 air leaks.  In newer mixers,  the body does not rotate.  The  exhaust gases
 are usually filtered of particulates greater than submicron  size by the use
 of fabric filters.
           The use of fabric filters for separation of particulates  from air
 has been practiced for a number  of years in the glass industry.  The  earliest
 fabric filters were known as  "baghouses", since these were large free-standing
 units for exposed fiber bags.  By  passing the exhaust air through layers of
 a woven fabric,  the particulates were collected.  Unfortunately, as the
 thickness of the collected layer of particulates increased,  the pressure
 differential required for continued air flow also increased.  Thus, the
 collected dust must be  periodically removed by manual or mechanical shaking.
 Almost all container glass plants  use fabric filters to remove entrained
 dust  particles  '    .   The fabric  filters used today are totally enclosed,
 and most  have  a  continuous removal operation for the trapped particulates.
           The  traditional woven and synthetic fabrics are used.  Today,
 fabric filters are  generally made  of low-temeprature materials such as
                                 f o / o c \
 Nomex,  nylon,  terylene,  or Orion
           Fabric  filters are used  to collect particulates from the  raw-
materials  and  handling  operations  for several reasons.  First, they have an
efficiency of  greater than 99 percent and they can be used to collect submicron-
size particulates.   In  addition,the trapped particulates can later  be recovered
                     (9  34-35)
for reuse or recycle   '       .  One manufacturer had from 2  to 6 baghouses
with a  stack height less than 50 feet at a plant manufacturing 72.6 Gg
 (80,000 ton) of container glass per year( '  '.  These used  nylon-fabric
filters operating at 98 percent efficiency and collecting about 72.6  Mg
(80 ton) of dust per year.

-------
                                      69
                            _Glass-Meltlng Operation

           In  a glass-melting furnace, raw materials are heated until a
 homogeneous viscous  liquid, free of gas inclusions, is formed.  Temperatures
 in the melter will generally fall in the 1500 to 1600 C range (2730 to 2912 F)(36)
 Natural  gas and fuel oil are the principle types of fuel, .with natural gas
 predominating (^70 percent)(15>37).  Over 90 percent of glass-melting furnaces
 have regenerative-firing systems for purposes of heat recovery and fuel
            (14)                                              J
 conservation     .
           In  order to increase melting capacity, many furnaces now have
 electric-boosting  systems.   These systems consist of several water-cooled
 electrodes equally spaced along the sides or bottom of the  melter,  below the
 surface  of the glass.   Additionally, all-electric melting is used  by a few
 manufacturers.

 Emissions
          Important  air emissions from a glass-melting furnace  consist of
 NO  , SO  , and particulates.   Other emissions can include CO,  hydrocarbons,
  X   X
 and selenium.
          Nitrogen oxides  represent the largest fraction by mass,  about
 54  percent of glass-furnace  emissions   '   '
          As was described earlier, the source test measurements of NO
                                                                      X
 emission rates vary  from 0.58 to 6.29 g/kg of glass produced.   For additional
 information, see Appendix  B.   Based on an  average emission rate of 3.07 g/kg,
 glass-melting furnaces  with  a total production rate of 12.656 Tg would emit
 38.8  Gg of NOX yearly(14).
          SO , on the other  hand,  depends  both on the  sulfur  content  of the
 fuel and on the sulfur  content of  the batch material.   Sulfur present in the
 fuel oil will oxidize and  appear as S0x in the exhaust gas.   A  fuel oil
 containing 1 percent sulfur  by weight emits « 600 ppm  S02 in  the flue gas<21>.
 Sulfur is also present  in  the batch materials, usually as Na2S04>  During heatup,
 the sulfate decomposes  and sulfur  dioxide  is formed, some of  which is
 chemically encorporated  into  the glass (as SO^)  and some released within the
furnace.    An average emission rate of S0x is 1.70 g/kg.   Thus, plants

-------
                                     70
 producing 12.656  Tg  of  glass  annually would emit  approximately  21.5   Gg
 of SO  yearly.
      x
           Particulate emissions  from a glass-melting  furnace  result  primarily
 from volatilization  of  materials  in the melt  that combine with  gases such  as
 SO  to form condensates in  the flue system.   Particulate emissions from  a
 glass-container furnace consist  of approximately  80 percent sodium
 sulfate^9'18'22"2  .  These particulates form from the  condensed vapors  in
 the melt and are  submicron  sized^  '  ~    .   The median particle diameter  from
                                            (24-25)
 a flint glass furnace was found  to be 0.13 y       ,
           Source-test measurements for particulate emission rates vary from
 0.13 to 1.95 g/kg of glass  produced.  This averages to  a particulate
 emission rate of  0.68 g/kg.
           Other emissions exhausted from glass-melting  furnaces include  CO,
 hydrocarbons,  and selenium.
           Carbon  monoxide is  probably exhausted from  the glass-melting
 furnace as a result  of  incomplete fuel combustion.  Source-test measurements
 have reported emission  rates  from 0.05 to 0.13 g/kg.  An estimated average
 emission rate is  0.07 g/kg.
           Hydrocarbons  are  also  formed in the glass-melting furnace  as a
 result of incomplete fuel combustion.  Source-test measurements have
 reported emission rates from  0.01 to 0.53 g/kg.  The  calculated average  emission
 rate was 0.08 g/kg.  Actual emission rates are a  function of  firing  conditions
 (extent  of fuel/air mixing, excess air, firing temperature).
           Selenium is generally used in amounts of 0.001 weight percent  or
 less in  the  batch as a  decolorizer to neutralize  the  green tint in container
                                  (9 10 22)
 glasses  caused by  iron  impurities.  '  '  '  No test measurements on actual
 selenium emissions have been reported, but it likely leaves the stack
as selenium vapor, because of its low vaporization temperature  (315  C for  SeO
                  (9 22)
and 685 C for Se)   '     .  A worst case emission rate was calculated  to be
0.002 g/kg, with total annual emissions of 0.02 Gg.

-------
                                       71
 Glass-Melting-Control Technology

           Control of emissions from the glass-melting furnace have occurred
 primarily because of environmental considerations.  Four general approaches
 have been employed:

           (1)   Modification of feed material
           (2)   Modification of furnace design
           (3)   Increase of checker volume
           (4)   Adoption of commercial-control apparatus.

 Modification of feed material and furnace design have been primarily used to
 control  gaseous emissions, while the other two methods are used  for control
 of  particulate  emissions.

           Modification of  Feed Material.   Raw materials have  a tendency to
 vaporize or decompose in the glass-melting furnace.   Several  raw materials
 that  readily vaporize include nitrates and selenium.   By minimizing the amount
 of  these or other ingredients used or by  substitutions of materials, the
 amount of  gaseous emissions exhausted from the glass-melting  furnace is
 reduced.   For example,  arsenic has been essentially  eliminated as  a fining
 agent.   Cerium  is used to  partially replace selenium as a decolorizer.  In
 addition to reducing  the selenium in the  batch (by about 40 percent), this
 modification leads to the  elimination of  arsenic in  the batch, since cerium
 and arsenic are not  compatible.   Cerium is especially appealing  because it
 tends to form high-melting compounds which do not readily vaporize.

          Modification of  Furnace Design.   Increasing the fuel efficiency
 of the glass-melting  furnace can  in turn  lead to a decrease in combustion
products, a decrease  in dust entrainment  by hot  combustion gases passing
over the melting  glass  batch,  and possibly a decrease in furnace temperature.
In addition,  emissions  from low melting and easily vaporized  fluxing or fining
agents can be lowered.   Several methods currently in  practice to improve
furnace efficiency are:

-------
                                     72
           (1)   Better  instrumentation  for regulating  air/fuel mixtures
                and monitoring  furnace  temperature and stack gas  composition.
           (2)   Combustion  control to produce long luminous flames that
                eliminate spurious hot  spots in the furnace and provide
                better  heat transfer to the melt
           (3)   Improved refractories to increase corrosion resistance,
                which permits furnaces  to be more fully insulated
           (4)   Use of  electric melting to reduce incomplete combustion
                and volatilization losses
           (5)   Use of  electric boosting to increase furnace capacity,
                increase furnace efficiency, and lower temperatures above
                the molten  glass.

 All  of  these methods have  been employed to control gaseous emissions.   Sulfur
 oxides  that form can be controlled by both limiting the sulphate in the feed
                                                                           (18)
 material and by the improvement of furnace efficiency.  Ryder and McMackin
 found that the  SO  emission rate increased directly with an increase'in
                 X
 production rate on a sideport  furnace melting soda/lime glass.   This increase
 comes about because of the larger quantities of sulphate being added to
 the  furnace when the production rate doubled.
          NO  emissions can be also lowered when the furnace efficiency is
            X
 increased if the furnace temperature also drops.  A 10 percent decrease
 in fuel consumption will result in a 10 percent decrease in NO   emissions       .
                                                              X.
          Electric boosting is commonly used on fossil-fuel fired furnaces
 in the container glass industry, primarily to increase output.   About 40 percent
 of these glass-melting furnaces are equipped with boosters.  Boosting will
normally result in a reduction in emissions per unit of output.   '
          Electric melting furnaces essentially eliminate both particulate
and gaseous emissions from the glass-melting operation.  In 1975, less
 than 3 percent of container glass manufacturers used electric melting     .

-------
                                      73
          Adoption  of  Commercial-Control Apparai-i.a.   Participates  can be
 cleaned from the  glass furnace exhaust by scrubbers, fabric filters, or
 electrostatic precipitators (ESP).  Figure 14 shows  a breakdown of commercial-
 control apparatus presently in use.  Scrubbers can also be used to collect
 SO  emissions, while fabric filters and ESP's only remove particulates
       (51)
 Teller v    suggested spraying the stack gas with an  alkaline solution.  This
 would  cause the acidic gases (S0x, HF, or HCL) to react and form particulates
 that could then be  collected by the control device.

          Scrubbers .   One type of particulate scrubbing is a two-step process.
 Initially, particles in the exhaust gases are "contacted" or wet by a
 scrubbing fluid that draws the particles into agglomerates.   These agglomerates
 are then separated  from the gas stream by an inertial mist-elimination process.
          A low-pressure (<10-in. water) centrifugal scrubber used by one
                            (9)
 glass  company in  California     had two separate contacting sections within
 a single casing.  Separate 50.7 metric horsepower (50 horsepower)  circulating
 fans forced dirty gas  through each section containing two to three impingement
 elements similar  to fixed blades of a turbine.
          One reference     mentions a scrubber that uses a packed-bed
 preconditioning chamber.   Hot gases (538 C) containing volatilized sodium
 compounds enter the chamber, and the vapors condense out onto the  packing
 material.  This material is wet by a scrubbing solution and provides a large
 surface area for  condensation.  A standard Venturi-type scrubber completes
 the system.  This scrubber is presently installed on a 0.181 Gg/day (200 ton/day)
 container glass furnace and it reduced particulate loading from more than
 0.23 to less than 0.046 g/sdm3 (from more than 0.10  to less than 0.02 g/sdcf)
          One flint-glass manufacturer     installed a tower  scrubber  (2.9-meter
diameter) on a 44.8 meter2  (482  ft2)  melter.   Hot  effluent  from the  furnace  is
initially quenched and saturated with a caustic solution passing through  the
exhaust gas at 900 gal/min.   The gas  then passes into a 300 gal/min  variable
throat Venturi operating at  30 in.  of water.   This scrubber has been plagued by
malfunctions and breakdowns.  A  highly visible steam plume  is exhausted when
it is not working.

-------
                                            3.6%  Electrostatic Precipitator
                                            0.07% Centrifugal Collector
                                            2% Fabric Filters
                                            2% Scrubber
                                            7% Gravity Collector
-(No  collection equipment used)
              Amber
(No collection  equipment used)
           Soda/Lime
               FIGURE 14.  USE OF COMMERCIAL-COLLECTION EQUIPMENT FOR EMISSION,  .
                           CONTROL ON BOTH AMBER AND SODA/LIME GLASS FURNACESV  '

-------
                                      75
          Fabric Filters.  Fabric  filters,  also known as  "baghouses"(36),
collect particulates by filtering  exhaust gas  from  container glass-melting
furnaces through closely woven natural  or synthetic fabric filters that are
capable of trapping submicron particulates.  Unlike wet scrubbers, fabric
filters are unaffected by variations  in the gas flow rate.  Temperature control,
however, is critical for proper functioning and the type  of fabric filter
selected is dependent upon the temperature  of  the gases exhausted.  Fabric
filters are generally made of cotton  sateen, standard nylon, wool, dacron,
                                     / o ^_O^"S
orlon, NOMEX, teflon, and fiberglass        .   Maximum operating temperatures
for these fabrics are given in Table  16.

                    TABLE 16.  MAXIMUM  USE  TEMPERATURE FOR
                               VARIOUS  FABRIC-FILTER MATERIALS
Fabric
Cotton Sateen
Standard Nylon
Wool
Dacron
Orion
Nomex
Teflon
Fiber Glass
Maximum Temperature, C
99
93
107
135
135
204
232
288
Since stack gas from a glass container melting furnace is at 316 to 645 C
(600 to 1200 FK4  , the gas must be cooled to a temperature compatible
with the fabric filter bag.  This can be accomplished by using the following
methods, either alone or in combination with each other

          (1)   Air dilution
          (2)   Radiation-cooling columns
          (3)   Air/gas heat exchangers
          (4)   Water-spray chambers.

-------
                                     76
 Dilution of  off  gases with air is the simplest and most trouble-free method
 of reducing  temperature, but requires the largest baghouse because of the
 increased volume of gases.  Air-to-gas heat exchangers, and radiation and
 convection ductwork are subject to fouling from dust in the effluent.  A
 water-spray  increases humidity and requires careful temperature control to
 avoid  condensation, but does permit use of smaller baghouses.  Care must
 be taken, with all of these methods, to avoid cooling the gas to the
 temperature  where SO  and H.,0 would condense out and foul or react with the
 fabric filters.
          In addition to being selected for their thermal compatibility,
 fabricfilter ba,gs must also be corrosion and abrasion resistant.  Cotton,
                                                              ( 26}
 orlon,  and dacron can deteriorate from the S03 in the flue gas
          A  fabric-filter air-pollution-control system was installed in 1974
           ?       9                               (1T^
 on a 41.8 m   (450 ft ) melter producing amber glass v  '.  The 482 C  (900 F)
 effluent from the furnace was initially cooled to 177 C (350 F) .  A  fine
 powder  aluminate precoat was then introduced into the air stream at  18.1 kg/hr
 (40 Ib/hr) along with ambient air.  This further reduced the gas temperature
 to 121  C (250 F).  The baghouse contained 1200 m2 (12,915 ft2) of dacron-
 filter  cloth divided into six compartments, each containing 900 filter bags.
 During  normal operation, the air-to-cloth ratio was 1.55, but this would
 increase to  1.86 during the cleaning cycle.  The pressure drop ranged from
 3.5  to  4.5 in. of water across the bags.  An exhaust blower had to develop
 16  to 18 in. of water pressure to overcome the resistance of the checkers,
 heat exchanger,  baghouse, and about 46 meters of duct.  Initially, the heat
 exchanger required maintenance about 15 percent of  the time due to plugging
 with material condensing from the gas stream.  By blocking off about 40 percent
 of  the  tubes, the temperature increased slightly, but this permitted a
 normal maintenance schedule.  Discharge of particulates from the baghouse
 outlet was typically 1.1 kg/hr (2-3 Ib/hr).  Tests  using a Brinks  Impactor
 showed  these particulates to be <0.75 micron.

          ElectrostaticPrecipitator (ESP).  In an  electrostatic precipitator
 (ESP),  a voltage source creates a negatively charged area, usually by hanging
wires in  the gas flow path.  Grounded collecting plates composed  the sides  of
 the ESP.  A powerful electric field is created by the high potential difference

-------
                                     77
between these grounding plates and the discharging wires.  As the gas stream
passes through the field, the particles become electrically charged and are
drawn to the collecting plates.  Periodically, accumulated particles are
removed from these plates by vibration, rapping, or rinsing.  Thus, by applying
the collecting force only to the particles  to be collected, a much lower power
input is required (i.e., 200 watts per 0.5  m /s)^   .
          NAFCO Engineering, Ltd.  (a Japanese firm) has developed a new type
of ESP.  In contrast to the conventional units,  the NAFCO ESP uses thousands
of stainless steel needles affixed to the leading and trailing edges of
positively charged electrode plates.  Thirty five of these systems are now in
operation in Japan, with nine of them being used on soda/lime glass-melting
furnaces
          United McGill Corporation, who is the  licensed U.S. distributor
for the NAFCO ESP, has installed the unit on 10  soda-lime glass furnaces
       (25)
to date    .  All of these systems are designed  to have an outlet loading of
<0.046 g/std m3 (0.02 g/scfd).
                   2        2
          An 84.4 m  (908 ft ) melting furance, used for producing flint glass,
had an ESP installed in early 1974^   .  It consisted of dual chambers, where
the air flow could be directed to either chamber or divided between them.
Each chamber had three electrical fields connected in series.  Designed for
12.9-sec treatment time at 0.67 m/s  (2.2 fps) velocity through the treater,
one chamber was found to be as effective as two, the conclusion being that
the system was overdesigned.

          Technological Advances.  Collector systems previously discussed
are primarily useful for collecting particulates and for decreasing opacity
of gaseous emissions.  One company now offers dry and wet systems     to
control both particulate and gaseous emissions.  A nucleation scrubber is
used on their wet system to effect collection of submicron particulates and
acidic gases (HF and SO ).   A solid absorbent, on the other hand, is injected
into the gas stream to react with the noxious gases in their dry system.
The absorbant is then separated from the gas along with particulates in a
fabric filter.

-------
                                     78
          A patent  (U.S. 3,789,628) was issued for a scrubber where  an
 aqueous  solution of sodium silicate is sprayed into the gases as  they are
 exhausted in  the furnace stack.  Water from the solution evaporated  in the
 gas  stream and  the sodium silicate forms a small sticky sphere which the
 patent claims to react chemically with NO  and SO , and physically with
                                         X       X
 particulates.   These spheres can then be collected and recycled into the
           (43)
 glass batcb>    .
          The quantity of NO  from a glass-melting tank was studied  by Kitayama,
      fc.\                    ^
 et al    , in  order to evaluate methods for reducing fuel consumption under
 photochemical smog warnings.  A glass-melting furnace  (of unknown glass
 composition) with a 154.2 Gg/day (170 ton/day) capacity using preheated air
 at 1100  C, emits 850-1000 ppm of NO .  By varying the damper opening and
                                   2^
 reducing the  excess air by 10 percent, the NO  emissions were reduced to
                                             X
 480  ppm.  When  the excess air was reduced 20 percent, the NO  emissions were
                                                            X
 reduced  to 45 ppm.
                  (44)
          Takasaki     reports on a method for removing NO  from  flue gases by
                                                          X
 wet  oxidation and absorption.  This technique claims to eliminate more than
 90 percent of the NO  from the flue gas of a glass-melting furnace.  By
                    X                                                    3
 using activated carbon and chlorine acid soda, a pilot plant with 51 kg m /hr
 reduced  its NO  emissions by 95 percent.  This system consists of a  special
              X
 liquid-gas contact tower that utilizes a chlorine dioxide and chlorine oxidizing
 agent.  NO is converted into N0_, which is absorbed by the liquid and
 stabilized.  The existing gas contains no NO, <10 ppm N09, <5 ppm SO-, no
 chlorine oxide, chlorine, or hydrogen chloride, 13 percent C0_, 3.5  percent
                    2
 0?, and 0.29  mg/kg m  Of dusts.   Other details were not reported.
                   (45)                                                      1
          Kanenatsu     reports on scrubbers handling 377, 7,1, and  28.6 kg m /hr
m/hr of SO  in the flue gas.  By using a wet or dry desulfurization  method,
          X
 the sulfur oxides are absorbed by NaOH solutions and oxidized in  air, and
 the SO  recovered as tnirabilite.
                   f I £ \
          Kanematsu     suggested use of low sulfur fuels, high stacks,  and
 stack gas-desulfurization systems as methods for controlling SO   emissions.

-------
                                     79
Efficiency of Equipment

          Least effective and least expensive of the air-pollution control
devices is the wet scrubber     .  In addition to having numerous malfunctions
and breakdown, they have been found to exhibit particulate-collection
efficiencies as low as 66 percent^    to as high as 90 percent^33'* (if grain
loadings were low).  By fitting the column with impingement plates, efficiency
can range up to 95 percent with particles as small as 5 microns^3  .  A major
advantage of this system is its ability to remove acidic gases.
          Baghouses have a reputation for high efficiency and dependability.
Fabric filters are capable of >99 percent efficiencies and can collect
                                      (13 33)
particulates down to below 0.75 micron   '   .  Major disadvantages are that
exhaust gases must be pretreated to remove gaseous emissions and must be
cooled before they contact the  low-temperature fabrics.
          Electrostatic-precipitator performance is highly sensitive to
temperature and volume fluctuations.  Electrical characteristics of particulates,
which affect collection efficiency, vary with temperature, humidity, S02
content, and the type of particulate.  Conventional ESP's have been shown to
have efficiencies up to 95 percent and collect particulates down to submicron
size.  The NAPCO ESP, on the other hand, has a reported outlet loading of less
than 0.046 g/std m3 (0.02 grains/scf)    .  For an uncontrolled emission rate
of 1 kg of particulate/Gg (2 Ib of particulate/ton) of glass and an air flow
of 3119 std m3/Gg (100,000 scf/ton), the efficiency is reported to be
85 percent.  For an emission rate of 10 kg/Gg (20 Ib/ton), the efficiency
is reported to be greater than 98 percent.  This ESP was designed so
additional sections could be added so efficiencies greater than 99 percent
could be obtained(33»47).
          Wet or dry desulfurization methods, presently in use by a glass
company, in Japan, has shown efficiencies of better than 97 percent for
the wet and 80 to 90 percent for the dry for S0x removal.

-------
                                     80
                            Forming and Finishing

          As the glass leaves the forehearth of the furnace, the molten
 glass  is  cut into "gobs" by a pair of mechanical shears.  Chutes direct the
 gobs from the feeder into blank molds where it is formed by one or two
 methods.
          With the "blow-and-blow" technique, the gob is settled with compressed
 air and preformed into a parison with a counter blow.  The parison is inverted
 and transferred into the blow mold where it is blown into its final shape.
          Wide-mouth containers, on the other hand, are formed by a "press-
 and-blow" technique.  The gob is settled by pressing with a plunger and
 "puffed" with a counter blow.  The parison is inverted and transferred for
 final  blow forming.
          The surface of approximately 30 percent of glass containers are
 treated in an operation where hot bottles from the forming machine pass
 through a fume chamber containing vapors of tin or titanium tetrachloride.
 A surface layer of the metal oxide forms on the container.  Unreacted
 hydrated metal chlorides are exhausted into the atmosphere.  The containers
 are then annealed at 593 to 649 C (1100 to 1200 F) and uniformly cooled
 in gas-fired, continuous ovens called lehrs.
          A polymer coating may then be applied by spraying the containers
with an aqueous dispersion of coating material.  The heat from the containers
 evaporate the water and fuze the polymer into a uniform surface coating.
          Decorative coatings are applied to about 3 percent of the glass
 containers.   Vitrifiable glass enamels or organic resins are applied by
brush,  stencils,  banding machines, rubber stamps, offset processes,
electrostatic printing, and silk screen printing.  Metallic decorating materials
 (liquid bright metals,  such as gold, platinum, palladium, and silver, which
leave a mirror-like coating when fired on the glass) are also applied in
the same manner.   If a container is to be glazed, a water suspension of glass-
forming ingredients is  applied by spraying or dipping.  These decorative
coatings are then cured in annealing ovens at approximately 600 C.

-------
                                     81
          Forming Emissions.  Molds  on  forming  machines,  gob shears, and
delivery chutes are lubricated with  solutions ranging  from  grease and oils to
graphite and silicone-based emulsions.   During  the  past decade, silicone
emulsions and water-soluble oils have replaced  grease  and oil  lubricants on
                                    (9)
gob shears and gob delivery systemsv '.   Grease and oils  are still used on
molds and causes white smoke emissions  during flash vaporization of the swab.
Although the smoke dissipates in a few  seconds, hydrocarbon vapors are released.
These emissions are released inside  the plant but probably  drawn outside
through the large ventilators above  the melting furnace and are minor.

Forming and Finishing Control Technology

          Efforts to control the hydrocarbons emissions have centered on
finding lubricants capable of withstanding  high temperature (900 C) and not
volatilize.  Use of silicone emulsions  are  water-soluble  oils  (90 to 150 parts
of water to 1 part oil or silicone)  can eliminate these emissions.  Unfortunately,
                                                       (9)
they have not performed well as mold-release compounds    .  Emissions from
the forming machinery are dispersed  within  the  plant and  exhausted by the
room ventilating systems.  No companies were identified which  used any control
device for these emissions.

Surface Treatment

          Emissions.  Emissions from the surface treatment  of  glass containers
with tin or titanium tetrachloride include  metal oxides,  hydrated metal
chlorides, and HC1 that are released into the atmosphere.   The emission rate
is estimated to be 0.02 g/kg for metal  oxides,  0.03 g/kg  for hydrated metal
chlorides, and 0.02 g/kg for HC1.
          Particulates exhausted are generally  composed of  submicron size
metal chloride and oxide.  A calculated particulate emissions  rate is
0-05 g/kg.

-------
                                     82
 Surface Treatment Control Technology

          A U.S. Patent 3,789,109 describes^    an apparatus to be used
 for cleaning solid, liquid, and gaseous pollutants from a hot-end surface
 treatment station of a container glass manufacturing plant.  In this apparatus,
 the air discharging from the hood is heated until the metal chlorides in
 the air disassociate to metallic oxides and hydrogen chloride gas.  Exhaust
 gases are then sprayed with fresh water to cool the stream.  Water reacts
 with the hydrogen chloride to form hydrochloric acid.  Exhaust air passes
 into a scrubber where the pollutants are removed ,  and then conveyed to a gas
 scrubber where metal oxides are removed.
 Decorating

          Emissions.  Hydrocarbon emissions from organic solvents and binders
 used in coatings on containers are released when decorative coatings are
 cured in annealing lehrs.  A calculated emission rate for these hydrocarbons
 is 4.37 g/kg.  Only 3 percent of the containers are decorated, giving a total
                           (2)
 annual emission of 1.44 Gg.

 Decorating-Control Technology
          Process modifications are difficult to accomplish without harming
                          (49)
the quality of the coating    .  In addition, they do not completely  eliminate
hydrocarbon emissions.  Several such changes involve the substitution of
solvents and a reduction of solvent concentration in the coating.
          Hydrocarbon emissions can be controlled by incineration, absorption
(activated charcoal or silica gel), or condensation     .

-------
                                      83

                                   SECTION VI

                       FUTURE GLASS-CONTAINER PRODUCTION

          The  future  production levels for glass containers is highly
 unpredictable  because of three major factors:  (1)  shortages in natural gas
 and substitute energy sources, (2)  potential large-volume penetration  of  the
 beverage-container market by plastics plus continuing penetration  by cans, and
 (3) possible legislation restricting or outlawing nonreturnable containers.
          The  shortage of natural gas and the allocation of petroleum
 products places  a constraint on the container-glass industry.   In  recent
 years, most of the industry has used petroleum products primarily  as a
 reserve or standby fuel and, therefore, does not have a base period of any
 significant usage.  At the same time, the industry  has incurred reductions in the
 use of natural gas, its primary fuel.  Since oil is the only normal replacement
 fuel, allocations based on historical demand would  indeed constrain the
 production on  container glass in the United States.
          The  primary substitute fuel oil used by the glass industry is a
 distillate, such as No.  2 fuel oil.   However, both  distillates  and residuals
 are used by the industry in the melting operation with properly designed fuel-
 handling and burner systems.
          Oil  cannot  normally be substituted for natural gas in other  nonmelting
 operations, such as in the refiner,  forehearth,  and annealing  lehrs, for
 reasons of glass quality.   In such  areas, propane is the only  substitute.  If
 a limited amount of natural gas is  available in the glass plant, it is
 usually reserved for  use in the nonmelting areas, assuming oil  is  available for
 substitution in the melting area.
          The extent  to  which plastic containers will penetrate the beverage
 market is still highly  speculative.   But, extensive research to resolve
 limiting factors such  as cost,  gas permeability,  and creep has  led to  test-
market introduction of plastic containers by the two largest soft-drink
 companies Coca-Cola Company and Pepsico,  Inc.   The  initial plastic soft-
drink containers have concentrated on the 32-oz  capacity or larger sizes of
which they are  most cost  competitive.   Also,  a large beer manufacturer has
announced intent to package in metal  and  plastic in the future  rather  than

-------
                                     84
 glass.  If energy costs continue to escalate, the competitive position of
 plastic containers may be improved both in manufacturing and transportation
 (distribution) costs.
          Metal cans have been claiming an increasing percentage of the
 smaller capacity soft-drink containers and may be expected to increase these
 inroads barring any legislative effects.
          Federal legislative action relative to returnable versus nonreturnable
 beverage  containers has been under consideration for some time.  If such
 legislation does become law, it could have a major impact on both the total
 numbers of glass containers manufactured and the competitive position of
 glass, plastics, and metal.  The details of such legislation would determine
 how it would effect glass-container production.
          The future of glass container production will also depend upon the
 success of the industry's research on developing very light-weight beverage
 containers.  If gross reductions in glass weight are achieved through
 improved  forming techniques, strengthening processes, plastic coatings, etc.,
 then the  competitive position with metal and plastic containers could lead to
 improved  growth potential.
          Historically, the shipments by the industry have grown an average
 of 6 percent annually since 1967, as can be seen in Table  17.

               TABLE 17.  GLASS CONTAINER PRODUCTION STATISTICS^
Year
1974
1973
1972
1967
Tg Production, 10 tons
11.00
11.32
10.77
8.39
12.13
12.48
11.87
9.25
106 Bottles
39,800
39,790
38,550
33,271
This growth is attributable primarily to the increased  popularity of the
nonreturnable bottle.  In recent years, the growth has  been less.  It is likely
that 1980 production will be 20 percent higher  than  for 1974.   Total National
emission will also increase by this amount without changes in control technology.

-------
                                     85

                                 SECTION VII


                               UNUSUAL RESULTS

          As mentioned in the previous section, the use of natural gas is
declining as a primary energy source by glass container manufacturers.  Its
usage has dropped from approximately 83 percent in 1971     to about 70 percent
       C2)
in 1975   .   Oil usage is increasing and this will have a direct effect on
SO  and possibly particulate air emissions.  Tending to counter this effect
  X
is improved furnace efficiency achieved by process modifications that act to
reduce stack emissions.  It is not possible at this time to predict the
quantitative impact of these changes on future air emissions.

-------
                                      86
                                  REFERENCES


  (1)   Current Industrial Reports:  Glass Containers Summary for 1974. Series
       M-32G (74)-13 (May 1975).

  (2)   Personal communication with Glass Containers Manufacturer Institute
       (GCMI).

  (3)   "Directory Issue", The Glass Industry, .56 (10), 1975.

  (4)   "Glass Factory Directory Issue", American Glass Review, .95 (8A), Febru-
       ary 28, 1975.

  (5)   Private communication with anonymous manufacturer.

  (6)   Kitayama, Hiroshi, Hideo Hayashi, Sataro Iwasaki, Tadashi Fujimura,
       Tomohiko Mujano, Hideaki Murayama, Tomihiro Myuhata, "Effect of Com-
       bustion Conditions on Nitrogen Oxides Formation of Furnaces", presented
       at Japan,Soc. Air Pollution, 14th annual meeting, Fukushima, Japan,
       (Jov. 1973).

  (7)   Hutchins, J. R. Ill, and Harrington, R. V., "Glass" from the Encyclopedia of
       of Chemical Technology, 2nd Edition, 10, John Wiley & Sons., Inc.  (1966)
       533-604.

  (8)   Bauer, W. C., Tooley, F. V., and Manring, W. H., "Batch Materials  Hand-
       ling and Preparation", The Handbook of Glass Manufacture. 1, 57-94
       (1974).

  (9)   Danielson, J. A., Air Pollution Engineering Manual, 2nd Edition, EPA
       Publication No. AP-40 (May 1970).

(10)   Anon, "A Screening Study to Develop Background Information to Determine
       the Significance of Glass Manufacturing", prepared by Research Triangle
       Park Institute for EPA, Contract No. 68-02-0607-Task 3  (December 1972).

(11)   Private communication with E. Stable, Owens-Illinois, Inc., Toledo, Ohio.

(12)   Chemical and Process Tech. Encyclopedia, Ed. Considine, 551-561  (1968)


(13)   Simon, Herbert and Wiliamson, E., "Control of Fine Particulates  from
       Continuous Melting Regenerative Glass Furnaces", presented at  the  68th
       Annual Meeting of the APCA, Boston, Massachusetts  (June  15-20,  1975).

(14)   Anon, National Emission Data System, Environmental Protection Agency
       Research Triangle Park, North Carolina (1974).

(15)   Schorr,  J. R. and Anderson, G. A., Final Report on Industrial  Energy
       Study of the Glass Industry to FEA and DoC", Battelle  Columbus  Labora-
       tories,  Contract No. 14-01-0001-1667, pages 80-142  (1974).

-------
                                       87
(16)   Lillis, E. J., and Young, D. "EPA Looks at  'Fugitive Emissions'", J. Air
       Pollution Control Assoc., 2J5 (10), 1015-18  (1975).
(17)
Air Pollution, Vol.  1,  Edited  by A.  C.  Stern,  2nd  Edition, Academic  Press,
N. Y. (1968),  "Nonviable  Particles  in the  Air1,  (M.  Corn). 49-52.
(18)   Ryder,  R.  J. and McMackin, J. J., "some Factors Affecting Stack Emissions
       from a  Glass Container Furnace", The Glass Industry, 50, 307-11, 346-350
       .(June 1969).

(19)   Arrandale, R. S., "Air Pollution Control in Glass Melting", Symposium
       Sur La  Fusion du Verre, Brussels (October 1968), 619-644.

(20)   Anon, State-by-State Listing of Source Types that Exceed the Third Decision
       Criteria,  Special Project Report, Monsanto Research Corp., Contract 68-02-
       1874, 1-3  (1975).

(21)   Reed, R.  J., "Combustion Pollution in the Glass Industry", The Glass Industry,
       54 (4), 24-26, 36 (1973).

(22)   Arrandale, R. S., "Pollution Control in Fuel Fired Tanks", The Glass Industry,
       55 (12),  12ff (August & November 1974).

(23)   Davis,  R.  E., Manring, W. H., and Bauer, W. C., "Carryover Studies in Glass
       Furnaces", presented at the 34th Annual Conference on Glass Problems, 109-
       126,  U. of 111.  (November 1973).

(24)   Stockham,  John D., "The Composition of Glass Furnace Emissions", Journal
       of the  Air Pollution Control Assoc., 21 (11), 713-715 (1971).

(25)   Custer, W. W., "Electrostatic Cleaning of Emissions from Lead, Borosilicate,
       and Soda/Lime Glass Furnaces", presented at the 35th Annual Conference on
       Glass Proglems,  Ohio State University (Nov. 14-15, 1974).

(26)   Frantz, C. J., Miser, D. L., Troy, H. N., and Stabbe, E. D., collected papers
       from the 32nd Annual Conference on Glass Problems, Dept. of Ceramic Engineer-
       ing,  University  of Illinois, 25-38 (1971).

(27)   Turner, D. B., Workbook of Atmospheric Dispersion Estimates. EPA Publication
       No. AP-26  (1970),  Figure 3-9.

(28)   Climatic Atlas of the United States. U.S. Dept. of Commerce (1968).

(29)   Recommended  Guide  for the Prediction of the Dispersion of Airborne Effluents.
       Edited  by  M.  Smith,  ASME (1968).

(30)   Reznik, R. B., Source Assessment; Flat Glass Manufacturing Plants. EPA En-
       vironmental  Protection Technology Series, Monsanto Research Corporation,
       Dayton  (October  1975).

-------
                                       88
 (31)   Mills, H. N., and Jasinski, J., "Evaluating Batch Changes", The Glass Ind.,
        51 (5), 223-227 (1970).

 (32)   Tooley, F. V., "Raw Materials", Handbook of Glass Manufacture. Vol. 1,
        Books for Industry, New York (1974),  Chap. 2.

.(33)   Rymarz, Ted M., and Lipstein,  David H.,  "Removing Particulates from Gases",
        Chemical Engineering Deskbook. 82  (21),  The McGraw-Hill Publishing Company,
        New York 113-129 OctOber 1975).

 (34)   Swift, P., "Dust Control Related to  the Bulk Delivery of Particulate
        Materials", The Chemical Engineer,  143-150 (March 1975).

 (35)   Edmundson, J. N.,  Rietz, L., Weise, R.  L., and Fraas, J., Collected papers
        from the 32nd Annual Conference on  Glass Problems, Dept. of Ceramic Engineer-
        ing,  University of Illinois, 39-54  (1971).

 (36)   Arrandale, R. S.,  "Furnaces, Furnace  Design, and Related Topics", Handbook
        of Glass Manufacture.  Vol.  1,  Books for  Industry, New York (1974), Section
        5, 249-387.

 (37)   Hibscher,  William,  Stertz,  R., The  U.S.  Glass Industry's Challenge in These
        Energy Critical Times", presented at  the 35th Annual Conference on Glass
        Problems,  The Ohio State University,  85-101 (November 1974).

 (38)   Bartz,  D.  V., KVB  Engineering, Inc.,  Control of Oxides of Nitrogen from
        Stationary Sources  in  the South Coast Air Basin of California, California
        State  Air  Resources Board (1974).

 (39)    Anon,  "Symposium on Pollution, Stratford-Upon-Avon, 30 May-1 June 1973",
        Glass  Technology, J. (6),  140-144.

 (40)    Teller, A.  J.,  "Control of  Emissions  from Glass Furnaces", Ceramic Bulletin
        51, 637-640 (1972).

(41)    Keller, G.,  "Scrubber  System Lightens Load of Glass Furnace Emissions",
        Chemical Processing, 38.  9  (Jan. 1975).

(42)   Teller, A.  J.,  "Control of  Emissions  from Glass or Ceramic Manufacture",
       presented to ACS, St.  Louis, Missouri, September 1971.

(43)   Mahoney, W. P.,  "Method  for Controlling  Furnace Emissions", U.S. Pat.
       3,789,628  (1974).

(44)   Takasaki, Shoichi,  "Flue  Gas Denitration by West Oxidation and Absorption",
       Heat Management  Pollution Control,  26 (1), 57-62 (Jan. 1974).

(45)   Kanematsu, Jado, "Air  Pollution Control  in Glass Industry", Seramikk
        (Ceramics), .9  (1), 49-55  (Jan.  1974).

-------
                                      89


(46)   Kanematsu,  Jado,  "Countermeasures for Preventing Air Pollution Caused by
      Glass  Industry",  Seramikksu (Ceramics), j)  (1), 15-21 (1974).

(47)   Wright,  R.  W.,  "Application of Electrostatic Precipitators for the Control
      of  Container Glass Emissions", IEEE Trans, on Industry Applications. 1A-11
       (No. 4),  447-456  (July 1975).                                        	

(48)   Lyon,  R.  S. and Lyon, R. L., "Method for Cleaning a Gas", U.S. Pat. 3,789,104
       (1971).

(49)   Troy,  H.  N. and Kalter, P. A.,"Pollution Control and Glass Decurating", The
      Glass  Ind., 52  (3), 102-105 (March 1971).

(50)    LeMaire,  W. H., "Why Bill  Coors Wants A Plastic Bottle",  Packaging Engineer-
       ing.  January, 1976.

 (51)  Teller, A.  J.,  "Control of Glass Furnace Emissions", Glass Industry,  5_7 (2),
      15-19, 22 (February,  1976).

 (52)  Roos,  P.  W.,  "Lehr Priority:   Design Concepts to Save Energy", The Glass
      Industry, 56, 18-22 (April, 1975).

 (53)  Hangebrauck,  R. P., Von Lehmden,  D. J., and Meeker, J«  E., "Emissions of
      Polynuclear Hydrocarbons and Other Pollutants from Heat-Generation and
      Incineration Processes", Journal of the Air Pollution Control Association,
      14, 267-278 (July,  1964).

-------
         APPENDIX A
GEOGRAPHICAL LISTING OF THE
122 CONTAINER GLASS PLANTS

-------
TABLE A-l.  GEOGRAPHICAL LISTING  OF  THE  122  CONTAINER GLASS PIANTS
                                                                   County
                                                                  Population
                                                                   Density,
                                                                  persons/km^
State
Alabama
Arkansas
California















Colorado
Connecticut
Florida



Plant
Brockway Glass Co., Inc.
Arkansas Glass Containers Corp.
Anchor Hocking Corp.
Anchor Hocking Corp.
Ball Corp.
Brockway Glass Co., Inc.
Brockway Glass Co., Inc.
Gallo Glass Co.
Glass Containers Corp.
Glass Containers Corp.
Glass Containers Corp.
Kerr Glass Mfg. Corp.
Latchford Glass Co.
Madera Glass Co.
Owens -Illinois
Owens-Illinois
Owens-Illinois
Thatcher Glass Mfg. Co.
Columbine Glass Co.
Glass Containers Corp.
Anchor Hocking Corp.
Industrial Glass Co., Inc.
Owens-Illinois
Thatcher Glass Mfg. Co.
City
Montgomery
Jonesboro
Los Angeles
San Leonadro
El Monte
Oakland
Pomona
Modesto
Antioch
Hayward
Vernon
Santa Ana
Los Angeles
Madera
Los Angeles
Oakland
Tracy
Saugus
Wheat Ridge
Day vi lie
Jacksonville
Bradenton
Lake land
Tampa
County
Montgomery
Craighead
Los Angeles
Alameda
Los Angeles
Alameda
Los Angeles
Stanislaus
Contra Costa
Alameda
Los Angeles
Orange
Los Angeles
Madera
Los Angeles
Alameda
San Joaquin
Los Angeles
Arapahoe
Windham
Duval
Manatee
Polk
Hillsborough
AQCR
2
20
24
30
24
30
24
31
30
30
24
24
24
31
24
30
31
24
36
41
49
52
52
52
(persons /mi2)
80 (206)
28 (73)
662 (1714)
558 (1445)
662 (1714)
558 (1445)
662 (1714)
50 (127)
290 (752)
558 (1445)
662 (1714)
696 (1802)
662 (1714)
7 (19)
662 (1714)
558 (1445)
78 (201)
662 (1714)
76 (196)
61 (159)
259 (670.3)
49 (127)
46 (119)
180 (467)
Furnaces
3
3
3
1
2
2
1
4
1
1
2
1
4
2



3
2
4
3
2

2

-------
TABLE A-l.   (Continued)
State
Georgia


Illinois











Indiana








Louisiana.


Plant
Glass Containers Corp.
Midland Glass Co.
Owens - 11 linois
Anchor Hocking Corp.
Ball Corp.
Hillsboro Glass Co.
Kerr Glass Mfg. Co.
Metro Containers
Obear-Nester Glass Co.
Obear-Nester Glass Co.
Owens-Illinois
Owens-Illinois
Owens -11 linois
Thatcher Glass Mfg. Co.
Universal Glass Products Co.
Anchor Hocking Corp.
Brockway Glass Co., Inc.
Foster-Forbes Glass Co.
Glass Containers Corp,
Glass Containers Corp.
Kerr Glass Mfg. Co.
Midland Glass Co.
Owens-Illinois
Thatcher Glass Mfg. Co.
Laurens Glass Co.
Owens -Illinois
Underwood Glass Co.
City
Atlanta
Warner Robin
At lanta
Gurnee
Mundelein
Hillsboro
Plainfield
Dolton
East St. Louis
Lincoln
Alton
Chicago Heights
Streator
Streator
Joliet
Winchester
Lapel
Marion
Indianapolis
Gas City
Dunkirk
Terre Haute
Gas City
Lawrenceburg
Rust on
New Orleans
New Orleans
County
Fulton
Houston
Fulton
Lake
Lake
Montgomery
Will
Cook
St. Clair
Logan
Madison
Cook
LaSalle
LaSalle
Will
Randolph
Madison
Grant
Marion
Grant
Jay
Vigo
Grant
Dearborn
Lincoln
Orleans
Orleans
AQCR
56
54
56
67
67
75
67
67
70
75
70
67
71
37
67
76
76
76
80
76
76
84
76
79
22
106
106
County
Population
Density
persons /knr
(persons/mi^)
433 (1122)
63 (164)
433 (1122)
317 (821)
317 (821)
16 (42.3)
275 (712)
2197 (5689)
161 (417)
20 (52.9)
130 (337)
2197 (5689)
37 (96)
37 (96)
275 (712)
24 (63)
117 (304)
77 (199)
758 (1963)
77 (199)
23 (60.3)
105 (272)
77 (199)
37 (94.9)
27 (69.2)
1102 (2854)
1102 (2854)
Furnaces
2
1

4
1
1
2
3
3
2
8

9
2
1
3
2
5
2

3
3
3
2
4
2
5
                                                                         I
                                                                         l-o

-------
TABLE A-l.   (Continued)
State
Maryland


Massachusetts

Michigan
Minnesota

Mississippi


New Jersey











Plant
Carr-Lowrey Glass Co.
Columbia Glass Co.
Maryland Glass Corp.
Foster-Forbes Glass Co.
Owens -Illinois
Owens-Illinois
Brockway Glass Co.
Midland Glass Co.
Chattanooga Glass Co.
Glass Containers Corp.
Underwood Glass Co.
Anchor Hocking Corp.
Brockway Glass Co.
Gayner Glass Works
Kerr Glass Mfg. Corp.
Leone Industries
Metro Containers
Metro Containers
Midland Glass Co.
Owens-Illinois
Owens-Illinois
Thatcher Glass Mfg. Co.
Wheaton Industries
City
Baltimore
Baltimore
Baltimore
Milford
Mansfield
Charlotte
Roseroount
Shakopee
Gulf port
Jackson
Mineral Wells
Salem
Freehold
Salem
Millville
Bridgeton
Jersey City
Carteret
Cliffwood
Bridgeton
N. Bergen
Wharton
Millville
County
Baltimore
Baltimore
Baltimore
Worchester
Bristol
Eaton
Dakota
Scott
Harrison
Hinds
Union
Salem
Monmouth
Salem
Cumberland
Cumberland
Hudson
Middlesex
Monmouth
Cumberland
Hudson
Morris
Cumberland
AQCR
115
115
115
120
120
125
131
131
5
5
135
45
43
45
150
150
43
43
43
150
43
43
150
County
Population
Density,
persons/km^
(persons/mi^)
976 (2527)
976 (2527)
976 (2527)
162 (419)
307 (796.3)
46 (120)
93 (241)
35 (91)
87 (224)
92 (239)
17 (44.2)
7 (17)
369 (955)
7 (17)
92 (238)
92 (238)
4907 (12703)
714 (1849)
369 (955)
92 (238)
4907 (12703)
314 (813)
92 (238)
Furnaces
3

4
1
2
3
2
2
4
1
8
3
2
3
5
1
2
2
4
8
4
2
23
                                                                         >

-------
                                              TABLE A-l.   (Continued)
State
Plant
City
                                                                    County
                                                         AQCR
  County
 Population
  Density,
 persons/km2
(persons/nd.2)
Furnaces
New York



North
Carolina

Ohio

Oklahoma





Oregon
Pennsylvania







Glenshaw Glass Co.
Leone Industries
Owens-Illinois
Thatcher Glass Mfg. Co.
Ball Corporation
Laurens Glass Co.
Owens ~I1 linois
Brockway Glass Co., Inc.
Chattanooga Glass Co.
Ball Corp.
Bartlett-Collins Co.
Brockway Glass Co.
Brockway Glass Co.
Kerr Glass Mfg. Corp.
Liberty Glass Co.
Owens -Illinois
Anchor Hocking Corp.
Brockway Glass Co.
Brockway Glass Co.
Brockway Glass Co.
Brockway Glass Co.
Diamond Glass Co.
Foster-Forbes Glass Co.
Glass Containers Corp.
Orangeburg
Rochester
Brockport
Elmira
Asheville
Henderson
Winston-Salem
Zanesville
Mt . Vernon
Okmulgee
Sapulpa
Ada
Muskogee
Sand Springs
Sapulpa
Portland
Connellsville
Brockway
Cr ens haw
Washington
Washington
Royersford
Oil City
Marienville
Rock land
Monroe
Monroe
Chemung
Buncombe
Vance
Forsyth
Muskingum
Know
Okmulgee
Creek
Pontotoc
Muskogee
Tulsa
Creek
Multnomah
Fayette
Jefferson

Washington
Washington
Montgomery
Venango
Forest
43
160
160
164
171
166
136
183
175
186
186
188
186
186
186
193
197
178
178
197
197
45
178
178
502 (1300)
404 (1047)
404 (1047
94 (243)
81 (210)
51 (131)
189 (490)
45 (116)
31 (79)
19 (49)
19 (48)
15 (38)
27 (70)
14 (37)
19 (48)
500 (1295)
75 (193)
25 (66)
25 (66)
94 (244)
94 (244)
490 (1268)
35 (90)
4 (10)
2
3
3
3
1
1
3
3

1
2
1
2
1
4
4
5
2
2
4
3

2
1

-------
                                                TABLE A-l.  (Continued)
  State
           Plant
    City
                                                                     County
AQCR
  County
 Population
  Density,
 persons/km?
(persons/nd.2)
Furnaces





Rhode Island
South
Carolina
Tennessee
Texas




Washington
West
Virginia




Glass Containers Corp.
Glass Containers Corp.
Glenshaw Glass Co.
Metro Containers
Owens-Illinois
Pennsylvania Glass Products Co.
Pierce Glass Co.
Star City Glass Co.
Laurens Glass Co.

Chattanooga Glass Co»
Anchor Hocking Corp.
Chattanooga Glass Co0
Glass Containers Corp.
Kerr Glass Mfg. Corp.
Owens-Illinois
Northwestern Glass
Erockway
Chattanooga Glass Co.
Kerr Glass Mfg. Corp.
Owens-Illinois
Owens-Illinois
Universal Glass Products Co.
Parker
KnoK
Glenshaw
Washington
Clarion
Pittsburgh
Port Allegheny
Coventry
Laurens

Chattanooga
Houston
Corsicana
Palestine
Waxahachie
Waco
Seattle
Clarksburg
Keyser
Huntington
Fairmont
Huntington
Parkersburg
Butler
Clarion
Allegheny
Washington
Clarion
Allegheny
McKean
Kent
Laurens

Hamilton
Harris
Navarro
Anderson
Ellis
McLennan
King
Harrison
Mineral
Cabell
Marion
Cabell
Wood
197
178
197
197
178
197
178
120
203

55
216
215
22
215
212
229
235
113
103
235
103
179
61 (157)
24 (63)
842 (2180)
94 (244)
24 (63)
842 (2180)
20 (51)
315 (817)
26 (68)

170 (441)
386 (1000)
11 (28)
10 (26)
19 (48)
54 (139)
207 (536)
66 (171)
27 (69)
144 (373)
75 (194)
144 (373)
88 (229)
1
2
4
1
2

3
2
9

5
4
3
2
1
4
6
6
1
1
5
5
1
Wisconsin
Foster-Forbes Glass Co.
Burlington
                                                                   Kenosha
 239
  166 (429)

-------
  APPENDIX B
EMISSIONS DATA

-------
                                APPENDIX B


                              EMISSIONS DATA

                  Raw Materials Preparation and Handling

           Five typical points for participate emissions were considered for
the raw materials preparation and handling operations:  (1) unloading and
conveying, (2) crushing of cullet (scrap glass),  (3) filling and emptying of
storage bins, (4) weighing and mixing of batch, and (5) feeding of batch to
glass melting furnace (batch charging).  Source test data are summarized in
Table B-l.
           Emissions from the raw materials preparation and handling operation
consist entirely of particulates from raw batch materials.  In practice, only
fugitive dust emissions should be considered, since particulate emissions
remaining within the plant may constitute an OSHA health and safety considera-
tion distinct from plant emissions.  As discussed in the text, only particles
below 100 micron are considered as contributing to figitive dust emissions.
Actual measurements of plant emissions from these operations were not avail-
able; however, personal observation indicates that there are no visible
                                                                      (14)
emissions from the batch house.  Measurements were available from NEDS
of particulate emissions within a few plants.  These were used to determine
particulate emissions on a worst-case basis.
           The average emission factors for the various raw material prepara-
tion and handling operations were taken to be the following, calculated on
a worst-case basis.

                                               rig/kg.
                   1.  Handling               22 + 100%
                   2.  Crushing                1 + 100%
                   3.  Storage                 1 + 100%
                   4.  Mixing                  5 + 100%
                   5.  Charging                <0.1
                                              29 + 100%

-------
                                            B-2
                   TABLE B-l.  SUMMARY OF SOURCE TEST DATA  FOR MATERIALS
                               PREPARATION AND HANDLING^
Particulate Emissions
Mg/yr
3.63

0.91

(Tons/yr)
(4.

(1.

0)

0)

trace (a)
0.91

4.54

trace <">
0.91
(1.

(5.


(1.
0)

0)


0)
Production
Gg/yr
7

87

121
231

231

12
276
.53

.8

.5
.2

.2

.8
.6
(Tons/yr)
(830,

( 96,

(134,
(255,

(255,

( 14,
(305,
000)

800)

000)
000)

000)

100)
000)
Rate
mg/kg
(Ib/ton)
5
(0.
10
(0.
0
4
(0.
19
(0.
0
3
.0
010)
.5
021)
.0
.0
008)
.5
039)
.0
.5
Control
fabric

fabric

fabric
fabric

fabric

fabric
fabric
Equipment
filters

filters

filters
filters

filters

filters
filters
Operation
hand and storage

batch house

crushing
mixing

delivery

storage
mixing
(0.007)



3.63
trace
trace
trace
(4.
(b)
(b)
(b)
0)
134
276
257
165
.2
.6
.6
.1
(148,
(305,
(284,
(182,
000)
000)
000)
000)
0
0
0
22
.0
.0
.0
.0
fabric
fabric
filters
filters
fabric filters
fabric
filters
storage
storage
mixing
conveying
(0.044)
(a)   Source NEDS
                (14)
(b)   Trace < 1.0.

-------
                                    B-3
Total annual emissions were based on 15.8 Tg of raw materials being processed
to melt 12.7 Tg of glass.  This assumes that 85 percent of glass melted
produces a saleable container.
           Stack heights for these and other plant operations are listed in
Appendix C.  They range from 5 m (16 feet) to 44 m (144 feet).
           The accuracy was only obtainable for batch mixing where the sample
mean was 4.5 mg/kg and the sample standard deviation was 2.0  mg/kg.  The
95 percent confidence level was - 3.187 mg/kg.  The accuracy of engineering
estimates was assumed to be - 100 percent.

                               Glass Melting

Nitrogen Oxides
           Source test measurements of NO  emissions from NEDS     are
                                         X
listed in Table B-2.  Emission factors vary from 0.58 to 6.29 g/kg (1.60 to
12.40 Ib/ton),  which clearly reflect the wide range of operating conditions
found in glass  melting furnaces.  The average emission factor of 3.07 g/kg
(6.14 Ib/ton) assumes 85 percent of the glass melted is flint (clear or
green) which has an average emission factor of 3.40 g/kg and 15 percent is
amber, which has an average emission factor of 1.22 g/kg.  The average
emission factors for these two glass types was determined by adding the
average emission factors together and dividing by the number of values.
Alternatively,  the average found by dividing the total emissions by total
production was  3.20 g/kg (6.40 Ib/ton).  The difference is not significant
because the standard deviation is - 3.2 g/kg, and the 95 percent confidence
level is - 1.469 g/kg.
           Standard deviations (a) were determined by the following method.
\  -   .
|_m   i
                                    m
                                    E

-------
                                  B-4
       TABLE B-2.
NO  EMISSIONS FROM GLASS CONTAINER FURNACES
  x
            Production
                 Emissions
Emission Factor

Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Amber
22
23
24
Average*
Gg/yr

46.1
58.8
70.1
38.7
56.2
40.0
47.6
31.8
11.8
34.7
47.8
36.5
41.3
98.9
34.7
71.0
161.5
66.2
53.5
41.1
49.9

14.3
28.6
37.8
50.1
(tons/yr)

(50.800)
(64,800)
(77,300)
(42,700)
(62,000)
(44,100)
(52,500)
(35,100)
(13,000)
(38,300)
(52,700)
(40,300)
(45,500)
(109,000)
(38,300)
(78,300)
(178,000)
(73,000)
(59,000)
(45,300)
(55,000)

(15,800)
(31,500)
(41,700)
(55,200)
Mg/yr

5.08
50.8
349.3
38.1
151.5
51.7
299.4
167.8
62.6
184.2
253.1
226.8
117.9
381.0
49.9
78.0
733.9
198.7
198.7
198.7
29.9

17.2
35.4
46.2
153.8
(tons/yr)

(56)
(56)
(385)
(42)
(167)
(57)
(330)
(185)
(69)
(203)
(279)
(250)
(130)
(420)
(55)
(86)
(809)
(219)
(219)
(219)
(33)

(19)
(39)
(51)
(170)
§/kg

1.10
0.86
4.98
0.98
2.69
1.29
6.29
5.27
5.31
5.30
5.29
6.20
2.86
3.85
1.44
1.10
4.55
3.00
3.71
4.83
0.58

1.21
1.24
1.22
3.07
(Ib/ton)

(2.20)
(1.72)
(9.96)
(1.96)
(5.38)
(2.58)
(12.58)
(10.54)
(10.62)
(10.60)
(10.58)
(12.40)
(5.72)
(7.72)
(2.88)
(2.20)
(9.10)
(6.00)
(7.42)
(9.66)
(1.16)

(2.41)
(2.48)
(2.45)
(6.14)
Assumes 85% of glass is Flint and 15% Amber.

sum of averages for these two glass types.
                           Average  is  a weighted

-------
                                     B-5

 where:  m  =  number  of samples
        N.  =  sample  value
        jj,  =  sample  mean.
 The confidence  interval (CI)  was determined by
                                        ka
                                         1/2
                                        m
 where: k - "Student's t" variable  for m-1  degrees  of  freedom.
 Sulfur Oxides

                                                              (14)
          Source  test measurements of SO  emissions  from NEDS     are listed
                                         X
 in Table B-3.  Emission  factors  vary from 0.21 to 8.35  g/kg  (0.41 to 16.7 Ib/ton).
 The average emission factor  of 1.70 g/kg (3.4 Ib/ton) is based upon 85 percent
 flint glass having an average emission factor of  1.84 g/kg and 15 percent amber
 having an average emission factor  of 0.93 g/kg.   The standard deviation is
 2.2 g/kg and the  95 percent  confidence level is - 0.66  g/kg.
          SO  emissions  are  believed to come entirely from natural gas fired
            X
 glass melting furnaces.  Increased use of oil would  increase both the rate and
 amoung of emissions.

 Particulates

          Source test measurements of  particulate emissions obtained from
NEDS^  '  are listed in Table B-4.   Emission  factors  vary from 0.13 to 1.95 g/kg
 (0.27 to  3.90 Ib/ton).   The average emission factor  of  0.68 g/kg (1.35 Ib/ton)
is based  on 85 percent flint glass having an emission factor of 0.71 g/kg and
15 percent amber glass having an average  emission factor of 0.48 g/kg.  The
standard  deviation is 1.0 g/kg and the 95 percent confidence interval is - 0.25.

-------
                           B-6
TABLE B-3.  SO  EMISSIONS FROM GLASS CONTAINER FURNACES
              x
     Production
Emissions
Emission Factor

Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39 •
40
41
42
43
Gg/yr

46.1
49.9
15.9
56.6
40.0
73.9
47.6
55.3
31.8
58.8
11.8
38.7
34.8
45.6
47.8
80.6
34.8
57.4
71.0
54.5
41.0
57.9
30.0
64.3
81.7
103.9
81.5
44.5
81.5
57.2
161.5
50.8
66.2
60.3
53.5
72.6
41.1
33.6
51.7
14.3
14.3
34.5
37.2
(tons/yr)

(50,800)
(55,000)
(17,500)
(62,400)
(44,100)
(81,500)
(52,500)
(60,900)
(35,100)
(64,800)
(13,000)
(42,700)
(38,300)
(50,300)
(52,700)
(88,800)
(38,300)
(63,300
(78,300)
(60,100)
(45,200)
(63,800)
(33,100)
(70,900)
(90,000)
(115,000)
(90,000)
(49,000)
(90.000)
(63,000)
(178,000)
(56,000)
(73,000)
(66,500)
(59,000)
(80,000)
(45,300)
(37,000)
(57,000)
(15,800)
(15,800)
(38,000)
(41,000)
Mg/yr

27.2
36.3
8.2
79.8
94.4
104.3
122.5
435.5
16.3
27.2
6.4
49.9
17.2
381.0
24.5
184.2
10.9
208.7
52.6
381.0
10.9
208.7
9.1
381.0
65.3
208.7
39.0
69.0
70.0
11.8
110.7
72,6
79.8
49.9
59.9
108.9
59.9
64.4
96.2
39.0
39.0
145.2
38.1
(tons/yr)

(30)
(40)
(09)
(88)
(104)
(115)
(135)
(480)
(18)
(30)
(07)
(55)
(19)
(420)
(27)
(203)
(12)
(230)
(58)
(420)
(12)
(230)
(10)
(420)
(72)
(230)
(43)
(76)
(77)
(13)
(122)
(80)
(88)
(55)
(66)
(120)
(66)
(71)
(106)
(43)
(43)
(160)
(42)
g/kg

0.59
0.67
0.51
1.41
2.36
1.41
2.57
7.88
0.51
0.46
0.54
1.28
0.50
8.35
0.51
2.59
0.31
3.63
0.74
6.99
0.27
3.61
0.30
5.92
0.80
2.00
0.48
1.55
0.86
0.21
0.69
1.43
1.21
0.83
1.12
2.14
1.47
1.92
1.86
2.72
2.72
4.21
1.02
(Ib/ton)

(1.18)
(1.34)
(1.02)
(2.82)
(4.72)
(2.82)
(5.14)
(15.76)
(1.02)
(0.98)
(1.08)
(2.56)
(1.00)
(16.70)
(1.02)
(5.18)
(0.62)
(7.26)
(1.48)
(13.98)
(0.54)
(7.22)
(0.60)
(11.84)
(1.60)
(4.00)
(0.96)
(3.10)
(1.72)
(0.42)
(1.38)
(2.86)
(2.42)
(1.66)
(2.24)
(4.28)
(2.94)
(3.84)
(3.72)
(5.44)
(5.44)
(8.42)
(2.04)

-------
                                    B-7
                         TABLE B-3.   (Continued)
Production

44
45
46
Amber
47
48
49
Average*
Gg/yr
73.2
98.9
35.3

14.3
28.6
37.8
50.1
(tons/yr)
(80,700)
(109,000)
(38,900)

(15,800)
(31,500)
(41,700)
(55,200)
Emissions
Mg/yr
53.5
45.4
10.0

4.5
58.9
15.4
85.2
(tons/yr)
(59)
(50)
(11)

(05)
(65)
(17)
(94)
Emission Factor
g/kg
0.73
0.46
0.28

0.32
2.06
0.41
1.70
(Ib/ton)
(1.46)
(0.92)
(0.56)

(0.63)
(4.13)
(0.82)
(3.40)
*  Assumes  85%  of  glass  is Flint and 15% Amber.   Average is  a weighted
   sum of averages for these two glass types.

-------
                               B-8
TABLE B-4.  PARTICULATE EMISSIONS FROM GLASS CONTAINER FURNACES
         Production
Emissions
Emission Factor

Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Gg/yr

46.1
81.7
15.9
161.5
40.0
66.2
47.6
53.5
31.8
41.1
11.8
49.9
34.7
60.3
47.8
56.6
752.8
73.9
34.8
55.3
71.0
58.8
65.4
38.7
68.0
45.6
65.3
80.6
41.0
57.4
30.0
54.5
66.0
57.9
81.7
64.3
81.7
104.3
44.5
57.2
50.8
60.3
" (tons/yr)

(50,800)
(90,000)
(17,500)
(178,000)
(44,100)
(73,000)
(52,500)
(59,000)
(35,100)
(45,300)
(13,000)
(55,000)
(38,300)
(66,500)
(52,700)
(62,400)
(830,000)
(81,500)
(38,300)
(60,900)
(78,300)
(64,800)
(72,100)
(42,700)
(74,900)
(50,300)
(72,000)
(88,800)
(45,200)
(63,300)
(33,100)
(60,100)
(72,800)
(63,800)
(90,000)
(70,900)
(90,000)
(115,000)
(49,000)
(63,000)
(56,000)
(66,500)
Mg/yr

35.4
39.0
5.4
161.5
23.6
43.6
26.3
39.9
23.6
36.3
9.1
11.8
26.3
17.2
26.3
20.9
677.7
28.1
33.6
31.8
61.7
35.4
64.4
27.2
43.6
16.3
48.1
39.9
35.4
16.3
22.7
16.3
36.3
39.9
69.9
16.3
49.9
39.9
11.8
21.8
27.2
17.3
(tons/yr)

(39)
(43)
(6)
(178)
(26)
(48)
(29)
(44)
(26)
(40)
(10)
(13)
(29)
(19)
(40)
(23)
(747)
(31)
(37)
(35)
(68)
(39)
(71)
(30)
(48)
(18)
(53)
(44)
(39)
(18)
(25)
(18)
(40)
(44)
(77)
(18)
(55)
(44)
(13)
(24)
(30)
(19)
g/kg

0.77
0.48
0.34
1.00
0.59
0.66
0.55
0.75
0.74
0.88
0.77
0.24
0.76
0.29
0.76
0.37
0.90
0.38
0.97
0.57
0.87
0.60
0.98
0.70
0.64
0.36
0.74
0.50
0.86
0.28
0.76
0.30
0.55
0.69
0.86
0.25
0.61
0.38
0.36
0.38
0.54
0.29
(Ib/ton)

(1.54)
(0.96)
(0.68)
(2.00)
(1.18)
(1.32)
(1.10)
(1.50)
(1.48)
(1.76)
(1.54)
(0.48)
(1.52)
(0.58)
(1.52)
(0.74)
(1.80)
(0.76)
(1.94)
(1.14)
(1.74)
(1.20)
(1.96)
(1.40)
(1.28)
(0.72)
(1.48)
(1.00)
(1.72)
(0.56)
(1.52)
(0.60)
(1.10)
(1.38)
(1.72)
(0.50)
(1.22)
(0.76)
(0.72)
(0.76)
(1.08)
(0.58)

-------
           B-9
TABLE B-4.  (Continued)
Production

43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
Amber
67
68
69
70
71
72
73
74
75
76
/ w
77
78
79
80
81
82
83
Gg/yr
72.6
33.6
51.7
14.3
14.3
34.5
74.4
66.2
69.9
76.2
10.9
37.2
16.4
8.9
19.3
64.8
55.9
57.3
165.1
73.2
98.9
35.3
30.9
35.7

14.3
28.6
37.8
3.9
3.9
3.9
3.9
3.9
2.8
2.8
*•• 0 \x
2.8
1.7
1.1
0.98
0.98
0.98
0.98
(tons/yr)
(80,000)
(37,000)
(57,000)
(15,800)
(15,800)
(38,000)
(82,000)
(73,000)
(77,000)
(84,000)
(120,000)
(41,000)
(18,100)
(9,800)
(21,300)
(71,400)
(61,600)
(63,200)
(182,000)
(80,700)
(109,000)
(38,900)
(34,100)
(39,300)

(15,800)
(31,500)
(41,700)
(4,345)
(4,345)
(4,345)
(4,315)
(4,345)
(3,121)
\ y *
(3,121)
\ y *
(3,121)
(1,836)
(1,224)
(1,083)
(1,083)
(1,083)
(1,083)
Emissions
Mg/yr
57.2
8.2
25.4
13.6
13.6
26.3
47.2
45.4
89.8
107.1
119.8
34.5
3.6
7.3
10.9
111.6
104.3
111.6
99.8
33.6
25,4
21.8
14.5
20,0

6.4
10.0
8.2
2.7
2.6
2.9
2.6
2.6
1.2
1.3
1.2
1.3
0.91
0.64
0.54
0.64
0.73
(tons/yr)
(63)
(09)
(28)
(15)
(15)
(29)
(52)
(50)
(99)
(118)
(132)
(38)
(04)
(08)
(12)
(123)
(115)
(123)
(110)
(37)
(28)
(24)
(16)
(22)

(7)
(ID
(9)
(3)
(2.9)
(3.2)
(2.9)
(2.9)
(1.3)
(1.4)
(1.3)
(1.4)
(1.0)
(0.7)
(0.6)
(0.7)
(0.8)
Emission Factor
g/kg
1.13
0.24
0.49
0.95
0.95
0.76
0.63
0.68
1.29
1.40
1.10
0.93
0.22
0.81
0.56
1.72
1.87
1.95
1.61
0.46
0.26
0.62
0.47
0.56

0.44
0.35
0.22
0.69
0.67
0.74
0.67
0.67
0.43
0.46
0.43
0.76
0.83
0.65
0.55
0.65
0.74
(Ib/ton)
(2.26)
(0.48)
(0.98)
(1.90)
(1.90)
(1.52)
(1.26)
(1.36)
(2.58)
(2.80)
(2.20)
(1.86)
(0.44)
(1.62)
(1.12)
(3.44)
(3.74)
(3.90)
(3.22)
(0.92)
(0.52)
(1.24)
(0.94)
(1.12)

(0.88)
(0.70)
(0.44)
(1.38)
(1.34)
(1.48)
(1.34)
(1.34)
(0.86)
(0.92)
(0.86)
(1.42)
(1.66)
(1.30)
(1.10)
(1.30)
(1.48)

-------
                                 B-10
                      TABLE B-4.  (Continued)
Production

84
85
86
87
88
89
Average*
Gg/yr
47.9
43.3
43.3
29o3
44.7
44.6
58.1
(tons/yr)
(52,836)
(47,712)
(47,712)
(32,327)
(49,316)
(49,140)
(64,000)
Emissions
Mg/yr
2.1
6.6
12.0
3.9
12.4
8.2
39.3
(tons/yr)
(2.3)
(7.3)
(13.2)
(4.3)
(13.7)
(9.0)
(43.3)
Emission Factor
g/kg
0.04
0.15
0.28
0.13
0.28
0.18
0.68
(Ib/ton)
(0.08)
(0.30)
(0.56)
(0.27)
(0.56)
(0.36)
(1.35)
Assumes 85% of glass is Flint and 15% Amber.  Average is a weighted
sum of averages for these two glass types.

-------
                                     B-ll
Carbon Monoxide

          Source test measurements  of  carbon monoxide  emissions  are limited
because this is not a major  glass furnace emission.  It  can form because  the
industry uses a long diffusion  flame to  effect  uniform heat release.  Properly
controlled, the emissions  are negligible.  The  emissions are listed in
Table B-5.  Emission factors vary from 0,05 to  0.13  g/kg (0.09 to  0.25
Ib/ton).  The average emission  factor  of 0.7 g/kg. (0.13  Ib/ton)  is based  on
85 percent flint glass having an emission factor  of  0.06 g/kg and  15 percent
amber glass having an emission  factor  of 0.11 g/kg.  The standard  deviation
was 0.2 g/kg and the 95 percent confidence level  was 0.10 g/kg.

Hydrocarbons

          Source test measurements  of  hydrocarbon emissions are  also limited.
Formation occurs for the same reasons  as cited  for carbon monoxide.  These
emissions are listed in Table B-6.   Emission factors vary from 0.01 to
0.53 g/kg (0.02 to 1.06 Ib/ton).  The  average emission factor of 0.08 g/kg
(0.150 Ib/ton) is based on 85 percent  flint glass having an average emission
factor of 0.08 g/kg and 15 percent  amber glass  having  an average emission
factor of 0.05 g/kg.  The  standard  deviation is 0.5  g/kg,  and the  95 percent
confidence interval is - 0.178  g/kg.

Selenium

          No source test measurements  are available  for  selenium emissions
                                                       (2)
from flint glass furances.   Using data supplied by GCMIV   and obtained from
the technical literature,  a  worst-case engineering calculation was made.
Selenium is used as a decolorizer to neutralize the  tint from transition  metal
oxide contaminants such as iron.  Approximately 0.36 Gg  (395  tons) of selenium
are consumed annually in the U.S.,  of  which about one-sixth is used by glass
container industry (0.06 Gg) . ^  If it  is  assumed that  40 percent of the
selenium used volatilizes  and is emitted  from the melting furnace, then the
emission rate would be 0.002 g/kg.  The  accuracy  of  this calculation was  taken

-------
                                 B-12
        TABLE B-5.  CO EMISSIONS FROM GLASS CONTAINER FURNACES
           Production
Emissions
Emission Factor

Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Amber
20
21
22
Average*
Gg/yr

65.3
65.3
53.5
41.1
49.9
60.3
56.6
73.9
55.3
22.9
16.4
18.6
102.5
74.4
66.2
69.9
76.2
30.9
35.7

14.3
28.6
37.8
50.3
(tons/yr)

(72,000)
(72,000)
(59,000)
(45,300)
(55,000)
(66,500)
(62,400)
(81,500)
(60,900)
(25,200)
(18,100)
(20,500)
(113,000)
(82,000)
(73,000)
(77,000)
(84,000)
(34,100)
(39,300)

(15,800)
(31,500)
(41,700)
(55,500)
Mg/yr

3.6
3.6
2.7
2.7
4.5
5.4
3.6
3.6
3.6
1.8
0.91
0.91
6.4
3.6
3.6
3.6
3.6
2.7
3.6

1.8
2.7
3.6
3.1
(tons/yr)

(4)
(4)
(3)
(3)
(5)
(6)
(4)
(4)
(4)
(2)
(1)
(1)
(7)
(4)
(4)
(4)
(4)
(3)
(4)

(2)
(3)
(4)
(3)
g/kg

0.06
0.06
0.05
0.07
0.09
0.09
0.06
0.05
0.07
0.08
0.06
0.05
0.06
0.05
0.05
0.05
0.05
0.09
0.10

0.12
0.09
0.10
0.07
(Ib/ton)

(0.11)
(0.11)
(0.10)
(0.13)
(0.18)
(0.18)
(0.12)
(0.10)
(0.13)
(0.16)
(0.11)
(0.10)
(0.12)
(0.10)
(0.10)
(0.10)
(0.10)
(0.18)
(0.20)

(0.25)
(0.19)
(0.19)
(0.13)
Assumes 85% of glass is Flint and 15% Amber.
sum of averages for these two glass types.
           Average is a weighted

-------
                          B-13
TABLE B-6.  HC EMISSIONS FROM GLASS CONTAINER FURNACES
Production

Flint
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Amber
34
35
36
Average*
Gg/yr

32.7
34.8
71.0
65.4
67.9
65.3
95.2
105.2
65.3
65.3
53.5
41.1
49.9
60.3
56.6
73.9
55.2
22.9
102.5
72.6
33.6
51.7
14.3
14.3
34.5
74.4
66.2
69.9
76.2
64.4
56.3
30.9
35.7

14.3
28.6
37.8
53.4
(tons/yr)

(36,000)
(38,300)
(78,300)
(72,100)
(74,900)
(72,000)
(105,000)
(116,000)
(72,000)
(72,000)
(59,000)
(45,300)
(55,000)
(66,500)
(62,400)
(81,500)
(60,900)
(25,200)
(113,000)
(80,000)
(37,000)
(57,000)
(15,800)
(15,800)
(38,000)
(82,000)
(73,000)
(77,000)
(84,000)
(71,000)
(62,000)
(34,100)
(39,300)

(15,800)
(31,500)
(41,700)
(58,900)
Emissions
Mg/yr

17.2
5.4
7.3
0.8
7.3
10.0
12.7
12.7
0.9
0.9
0.9
0.9
2.7
2.7
0.9
1.8
1.8
0.9
0.9
6.4
3.6
3.6
1.8
1.8
2.7
1.8
1.8
1.8
1.8
9.1
7.3
0.9
1.8

0.9
0.9
1.8
4.0
(tons/yr)

(19)
(6)
(8)
U)
(8)
(11)
(14)
(14)
(1)
(1)
(1)
(1)
(3)
(3)
(1)
(2)
(2)
(1)
(1)
(7)
(4)
(4)
(2)
(2)
(3)
(2)
(2)
(2)
(2)
(10)
(8)
(1)
(2)

(1)
(1)
(2)
(5)
Emission Factor
g/kg

0.53
0.16
0.10
0.01
0.11
0.16
0.13
0.12
0.01
0.01
0.02
0.02
0.05
0.05
0.02
0.02
0.03
0.04
0.01
0.09
0.11
0.07
0.12
0.12
0.08
0.02
0.03
0.03
0.02
0.14
0.13
0.03
0.05

0.06
0.03
0.05
0.08
(Ib/ton)

(1.06)
(0.31)
(0.20)
(0.03)
(0.21)
(0.31)
(0.27)
(0.24)
(0.03)
(0.03)
(0.03)
(0.04)
(0.10)
(0.09)
(0.03)
(0.05)
(0.07)
(0.08)
(0.02)
(0.18)
(0.22)
(0.14)
(0.25)
(0.25)
(0.16)
(0.05)
(0.05)
(0.05)
(0.05)
(0.28)
(0.26)
(0.06)
(0.10)

(0.13)
(0.06)
(0.10)
(0.15)
*  Assumes  85% of glass  is  Flint and 15% Amber.
   of  averages for these two  glass types.
                                      Average is a weighted sum

-------
                                   B-14

as + 100 percent.  This is determined by dividing the total selenium
(0.06 Gg) by the total glass melted (12.7 Tg) and multiplying by 0.4  (the
amount of selenium emitted) alternately, the selenium represents no more than
                                                           (32)
0.001 percent of the glass or 0.008 g/kg of glass produced.      If 75 percent of
the production contains selenium, then the emission rate would also be 0.002 g/kg.
This is determined by dividing O.Oi g/kg (amount of selenium) by 1.25 (batch to
produce 1 unit of glass) and then multiplying by 0.40 (percent selenium emitted)
and 0.75 (percent glass using selenium).

                          Forming and Finishing

          Emissions measurements from the forming and finishing operations
were not available; hence, engineering calculations, considering worst-case
situations, were used to determine the severity of emissions from this area.

Forming

          During forming, an emulsion containing oil or silicone and water
is sprayed onto the molds.  Approximately 1.4 g (1/20 ounce) of liquid is
sprayed per container produced.  If one assumes that only oil is used as a
lubricant (silicone is also popular) and that the mixture of oil to water is
1:125, a normal situation.  Then under these conditions, the emission rate
would be 0.035 g/kg, considering the production in 1974 of approximately 40
                               (2)
billion containers and 12.7 Tg.     These were considered minor.  The calculation
is shown below.  Accuracy is taken to be +_ 100%,
     Oil/container   -   1.4 g * 125 » 0.0112 g
                                  12
     Container weight  - 12.7 x 10   g * 40 x 1<
     Emission rate   -   0.0112 * 0.318 = 0,035 g/kg

Surface Treatment

          Approximately 30 percent of containers produced receive a surface
treatment to improve resistance to scratching and to facilitate handling.
                             12            9
Container weight  - 12.7 x 10   g * 40 x 10  containers = 0.318 kg/container

-------
                                    B-15
 Hot containers are  subjected to  a tin or titanium chloride vapor.   Emissions
 consist of metal oxide and hydrated metal chloride particulates and HC1.
          Data received  from glass container manufacturers showed  the  average
 consumption of material  in surface treatment operations was 0.12 g/kg  (0.24 Ib/ton),
 of which 60 percent or 0.07  g/kg (0.14 Ib/ton)  escapes to  the  atmosphere.  This  is
 composed of 22 percent metal oxide, 44 percent  hydrated metal  chloride, and 34
 percent HC1.  Emission rates were then calculated for tin  compounds to be 0,02
 g/kg  (0.03 Ib/ton)  of tin  chloride, 0.03 g/kg (0.06 Ib/ton)  of hydrated tin
 chloride, and 0.02  g/kg  (0.05 Ib/ton) of HC1.  These are based  on the tin
 compounds comprising approximately 40 percent of  the total weight  input and
 HC1 approximately 20 percent of  the total weight  input.    Accuracy of  the data
 was taken as + 100  percent.
Annealing

          No emission data were available  for  gas-fired annealing lehrs;
therefore, emission factors were estimated from  other data on gas combustion.
A modern recirculating air type lehr consumes  11 to 17 m /hr (400 cfh to
600 cfh) when annealing 91 Gg  (100 tons) of glass per day.  Lehrs of older
design can consume 34 to 57 m3/hr  (1200 cfh to 2000 cfh) ^    .   On a worst case
           3                            3
basis (57 M /hr) would require 0.0062  m /kg of glass produced.  For a plant
                                                             3
producing 319 Mg/day (352 ton/day) this would  amount to 91 m /hr.  With a
                                                                  3
heating value of natural gas  (1000 Btu/cf  or 37.3 million joules/m ) this
amounts to 0.93 million joules per second  or about 0.23 million joules per
kg of glass (200,000 Btu/ton).
                                           (53)
          Using tests on gas-fired burners    ,  emission data was determined
as shown in Table B-7.  Converting these on a  basis of 0.24 million joules/kg
of glass gave the emission factors for annealing shown in Table B-8.
Decorating

          Glass containers are sometimes decorated with  vitrifiable glass
enamels or organic materials.  Emissions are derived  from organic  solvents
and binders used in the coatings.  Data supplied by glass manufacturers

-------
                      B-16
TABLE B-7.  EMISSIONS FROM GAS-FIRED BURNERS
                                            (53)
EmissionSj
mg/fcg
Partic- Hydro-
Test NO SO ulates CO caroons
X X
15 60.2 - 9.0
17 150 0 2.58
18 38.7 0 3.0
19 25.8 - 11.2
Average and 69 0 5.2
95% confi- -113% -196%
dence limits
Standard -47 0 -5.2
deviation
5.6 1.29
8.6
11.2 9.5
12.9 6.9
9.5 6.0
-55% -144%

-2.75 -3A


TABLE B-8. EMISSION FACTORS FOR ANNEALING LEHRS

Emission
Emission mg/kg
NO +
x 16 - 113%
SO 0
X
Participates 1.2 - 196%
CO 2.2 - 55%
Hydrocarbons 1.4 - 144%
factor
Clb/ton)
(0.032)
(0)
(0.0024)
(0.0044)
(0.0028)

-------
                                   B-17
would indicate that only 3 percent of containers are coated.  Point source
emission data were not available.  Materials balances taken from NEDS
were used as the basis for estimating the emission rate for hydrocarbons from
decorating lehrs.  Table B-9 lists the emission data.  The hydrocarbon emission
rate was 4.37 g/kg (8.73 Ib/ton).  The total hydrocarbons emitted annually from
all sources is estimated to be 1.44 Gg (1,590 tons).  This was determined in the
following manner.
                                                             12
               1974 container shipments     =     11.005 x 10   g
                                                             Q
               Decorated container (3%)     =      330.1 x 10  g
                                                                      9
               Total Emissions = Shipments x Emission Rate = 1.44 x 10  g

-------
                            B-18
TABLE B-9.  HC EMISSIONS FROM GLASS CONTAINER DECORATING

            OPERATIONS

Production

1
2
3
4
5
Average
Gg/yr
21.2
20.2
15.7
15.7
20.2
18.6
(tons/yr)
(23,400)
(22,300)
(17,300)
(17,300)
(22,300)
(20,500)
Emissions
Mg/yr
80
80
80
80
80
80
(tons/yr)
(88)
(88)
(88)
(88)
(88)
(88)
Emission Factor
g/kg
3.76
3.95
5.09
5.09
3.95
4.37
(Ib/ton)
(7.52)
(7.89)
(10.17)
(10.17)
(7.89)
(8.73)
                                       (141
* Material Balance Data Taken from NEDSV  J

-------
          APPENDIX C
   STACK HEIGHTS FROM THE
VARIOUS PHASES OF GLASSMAKING

-------
TABLE C-l.  TYPICAL STACK HEIGHTS OF BATCH HANDLING, TREATMENT
            AND DECORATING OPERATIONS FOR SODA/LIME CONTAINER GLASS
Batch Handling
<40 m >40 m
No. of Stacks Height, m No. of Stacks Height, m
(3) 5 44
22
24
(4) 36






Total 9 Average 23 Total 1 Average 44
Median 24 Median 44


Treatment
No. of Stacks



(4)
(3)


(3)
(3)

Total 19

Height, m
8
12
13
14
15
17
20
23
25
38
Average 19
Median 15

Decoration
No. of Stacks Height, m
(3) 12
(2) 13








Total 5 Average 12
Median 12
                                                                                            o

-------
                      C-2
TABLE C-2.  TYPICAL STACK HEIGHTS OF FLINT.AND
            AMBER CONTAINER GLASS FURNACES11 '

<40
No. of Stacks






(9)
(2)

(7)
(5)
(15)
(2)
(3)
(3)

(7)
(4)

(6)
(25)


Total 99

Flint

Height, m
5
6
10
12
13
14
17
18
19
20
21
23
24
25
26
27
30
33
34
36
38
39
40
Average 27
Median 25
Glass
>40
No. of Stacks
(2)
(8)
(3)
(2)
(7)
(2)
(3)
















Total 27


m Amber Glass
Height, m No. of Stacks Height, m
41 17
43 20
44 23
45
46
47
49
















Average 45 Total 3 Average 2
Median 45 Median 2

-------
     APPENDIX D
STATE LISTING OF TOTAL
 EMISSIONS AS OF 1972

-------
                          D-l
TABLE D-l.  STATE LISTING OF TOTAL EMISSIONS AS OF 1972
State
1 ALA&AfA
2 ALASKA
2 ARIZONA
4 ARKANSAS
5 CALIFORNIA
6 COLORADO
7 CONNECTICUT
8 DELAWARE
9 FLORIDA
10 GEORGIA
11- HAKAII
12 IDAHO
13 ILLINOIS
14 INDIANA
15 IOWA
16 KANSAS
17 KENTUCKY
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
2002000.0
1.53000
16340000.0
12.50000
3265000.0
2.49000
1619000.0
1.24000
5675000.0
4. 33000
3156000.0
2.1*1000
365600.0
0.27900
130200.0
0.09930
2430000.0
1.86000
2331000.0
1.78000
251200.0
0.19200
2430000.0
1.65000
3584000.0
2.74000
2202000,0
I.fe8000
2579000.0
1.97000
3358000.0
2.56000
1654000, D
1.42000
SO2
1226000.0
1.91000
222600.0
0. J4700
200200.0
0.31100
205400.0
0.31900
2557000.0
3.98000
473300.0
0.73600
1227000.0
1.91000
420700.0
0.65500
1755000.0
2.73000
1635000,0
2.54QOO
232000.0
0.36100
59140.0
0.09200
3714000.0
5.78000
3036000.0
4.72000
397400.0
0.61800
225000.0
0.35000
1627000.0
2.53000
NO
X
26160U.O
2.27000
31990.0
0.27700
75100.0
O.feblUO
77310.0
0.67UOO
796800.0
6.91000
116600.0
1.01000
152200.0
1.32000
45720.0
0.396UO
410300.0
3.56000
294200.0
2.55000
40790.0
0.35400
33220.0
0,28800
665100.0
5.77000
414400.0
3.59000
137700.0
1.19000
109900,0
0.95300
30200U.O
2.62000
Hydro-
carbons
342100.0
1.29000
iw&oo.o
0.53200
171100.0
0.64700
281700.0
1.07000
1914000.0
7.24000
294400.0
1.11000
259400.0
0,98100
77510.0
0.29300
536200.0
2,03000
526700.0
1,99000
62720.0
0.23700
163600.0
0.61900
1343000.0
5.08000
675100.0
2.55000
4Q0800.0
1.52000
742800.0
2.81000
274600. 0
1.0400B
CO
372600.0
2.04000
472200.0
2.50000
178300.0
0.97600
225800.0
1.24000
1987000.0
10.90000
105800.0
0.57900
92690.0
0.50700
24580.0
0.13500
3502000.0
19.20000
705400.0
3.86000
84750.0
0.46400
516300.0
2.84000
412500.0
2.26000
182100.0
0.99700
90720.0
0.49700
174600.0
0.95600
2193UO.O
1.20000

-------
          D-2
TABLE D-l. (Continued)
State
16 LOUISIANA .
19 MAINE.
20 MARYLAND
Zl MASSACHUSETTS
Z2 MICHIGAN
23 MINNESOTA
24 MISSISSIPPI
25 MISSOURI
26 MONTANA
27 NEBRASKA
28 NEVADA
29 NEW HAMPSHIRE
30 NEW JERSEY
31 NEW MEXICO
v
42 NEW YORK
S3 N CAROLINA
31 N DAKOTA
35 OHIO
36 OKLAHOMA
37 OREGON
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
1651000.0
1.26000
1036000.0
0.79200
657300.0
0.50200
802700.0
0.61300
2801*000.0
2.14000
3056000.0
2.33000
m90000.0
1.11000
2839000,0
2.17000
4975000.0
3.60000
3049000.0
2.33000
3155000.0
2.41000
326500.0
0.24900
615800.0
0.62300
354SOOO.O
2.71000
2704000.0
2.0600U
2203000.0
1.66000
2654000.0
2.16000
3054000.0
2.33000
2276000.0
1.74000
2885000.0
2.20000
S02
585800.0
0.^1100
770700.0
1.20000
1352000.0
2.10000
3640000.0
5.97000
3513000.0
5,46000
846600.0
1.32000
280300.0
0.43600
1259000.0
1.96000
177000.0
0.27500
137100,0
0.21300
263100.0
0.40900
325600.0
0.50700
2922000.0
4.55000
441400.0
0.68700
5137000.0
7.99000
2298000.0
3.56000
326700.0
0'. 51100
4062UOO.O
6.32000
163400.0
0.25400
372500.0
0.57900
NO
X
219000.0
1.90000
54270.0
0.47000
2151UU.O
1.86000
322300.0 '
2.79.000
548000.0
4.75000
165000.0
1.60000
67010.0
0.75400
287500.0
2.49000
34650.0
0.30UOO
50940.0
0.44200
56500.0
0.50700
3606U.O
0.31300
323400.0
2.80000
109800.0
0.95200
721400.0
6.25000
336400.0
2.93000
61110.0
0.53000
735800.0
6.61000
130000.0
1.13000
62710.0
0.54400
Hydro-
carbons
1741000.0
6.58000
71970.0
0.27200
302300.0
1.14000
463100.0
1.75000
734000.0
2.78000
386000.0
1.47000
350200.0
1.32000
S88400.0
2.22000
174200.0
0.65600
255600.0
0.96600
36140.0
0.13700
44430.0
0.16800
786600.0
2.97000
310200.0
1.17000
1353000.0
5.11000
465100.0 .
1.76000
73930iO
0.28000
1244000.0
4.70000 '
674700.0
2.55000
204800.0
0.774UO
CO
139900.0
4.60000
61430.0
0.33600
163400,0
0.89400
190400.0
1.04000
299400.0
1.64000
150700.0
0.62500
228200.0
1.29000
268500.0
1.47000
230500.0
1.26000
59590.0
0.32600
28700.0 ^
0.15700
30200. 0>
0.16500
281400.0
1.54000
49460.0
0.27100
551600.0
3.02000
371500.0
2.03000
22340.0
0.12200
"482700.0
2.64000
200800.0
1.10000
304900.0
1.67000

-------
         D-3
TABLE D-l.  (Continued)
State
3s PENNSYLVANIA
39 RHOOi ISLAND
*0 S CAROLINA
HI S DAKOTA
H2 TENNLSSEE
43 TEXAS
44 UTAH
45 VERMONT
16 VIRGINIA
47 WASHINGTON
48 nl VIRGINIA
49 WISCONSIN
SO WYOMING
US TOTALS
Mass of emissions, metric tons/yr (upper entry)
Percent of U.S. totals (lower entry)
Partic-
ulate
3132000.0
2.39000
113200. U
0.06640
1209000.0
0.923UO
2861000.0
2.18000
1769000.0
1.37000
9302000. U
7.10000
2461000.0
1.68000
292100.0
0,22300
1607000.0
1.23000
2204000.0
1.66000
1261000.0
0.96200
2180000.0
1.66000
2851000.0
2.18000
131000000.0
SO2
5603000.0
0.72000
519900.0
0.80900
1076UOO.O
1.67000
69420.0
0.10800
1307000.0
2.03000
1817000.0
2.63000
285400.0
0.44400
112600.0
0.17500
1388000.0
2.16000
626400.0
0.97500
1455000.0
2.2600C
1216UOO.O
1.69000
513000.0
0.79600
64300000.0
NO
X
782200.0
*.7«OUO
3F.760.0
3.33600
146300.0
1.27000
16560.0
0.16100
261<100.0
2.29000
635500.0
6.03000
46410.0
0.42UOO
13710,0
0.11900
197600.0
1.71000
126300.0
> 1.09000
306500.0
2.66000
231300.0
2.00000
70570,0
0.61200
1150000U.O
Hydro-
carbons
1331000.0
S.0300U
93730.0
0.354UO
260500.0
0.985UO
91110.0
0.34400
340900.0
1.29000
4139000.0
lb. 60000
112800.0
0.42600
2S460.0
0.0963U
41S200.0
1.57000
361800.0
1.37000
172800.0
0.65300
362600.0
1.37000
275200.0
1.04000
26400000.0
CO
527000.0
2.68000
29390.0
3.16100
483900.0
2.65000
23480.0
0.12900
200300,0
1.10000
1501000.0
8.22000
46840.0
0.25600
14190.0
0.07770
235100.0
1.29000
425500.0
2.33000
435100.0
2.36000
161300.0
0.68300
20870.0
0.11400
16300000.0

-------
      APPENDIX E
.   rnwVKRSION FACTORS

-------
              E-l
TABLE E-l.  CONVERSION FACTORS
To Convert From
Btu
degree Fahrenheit (F)
foot (ft)
foot3 (ft3)
inch (in.)
mile2 (mi)
pound (mass, Ib)
ton (short)


Prefix Symbol

tera T
giga G
mega M
kilo k
milli m
micro M-
To
joule (J)
degree Celsium (C)
meter (m)
3 3
meter (m )
meter (m)
2 2
meter (m )
kilogram (kg)
gigagram (Gg)
PREFIXES
Multiplication
Factor
10
io12
IO9
io6
io3
io-3
io"6
Multiply By
1.055 x IO3
t°c • (t°p - 32)71.8
3.048 x IO"1
2.832 x IO"2
2.540 x IO"2
2.590 x IO6
4.536 x 10'1
9.072 x 10'4


Multiply By
,«12
1 Tg = 1 x 10 g
1 Gg = 1 x IO9 g
1 Mg = 1 x 10 g
3
1 km = 1 x 10 g
-3
1 mm = 1 x 10 m
1 pm = 1 x 10 m

-------
   APPENDIX F
GLOSSARY OF TERMS

-------
                                     F-l
                                 APPENDIX F
                             GLOSSARY OF TERMS


ANNEALING - Controlled heating and cooling of glass to remove objectionable
stresses.

BATCH - Mixed glass raw materials.

BATCH HOUSE - Structure where raw materials are sorted, weighed, and mixed.

BOOSTING - Supplemental electrical heating in the glass furnace.

CHECKERS, CHECKERWORK - A network of refractory ducts on both sides of a
glass furnace, used as heat exchangers.

GULLET - Scrap glass that is to be recycled.

FINING - Process of removing gas bubbles from molten glass.

LEHR - A long oven for annealing glass continuously.

MELT - The molten glass in the glass furnace.

REFINING - Process of conditioning the molten glass to remove gas bubbles
and undissolved grains of sand.

TANK - That part of the glass melting furnace which holds the molten glass,
made of refractory material.

REGENERATORS - Chambers of refractory checkerwork on both sides of the
melting furnace.  Hot exhaust gases from the furnace pass through one
regenerator and heat it while combustion air passes through the other
regenerator and is heated.  At intervals of 15-20 minutes, the flow is
reversed.

-------
    APPENDIX G
LETTERS OF COMMENT

-------
                                  APPENDIX G
                              LETTERS OF COMMENT

         This  appendix contains letters from the Glass Packaging Institute
(GPI)  (formerly GCMI),  EPA, and Battelle that address certain points
raised by GPI during a  review of the Preliminary Source Assessment Document.
Some of the comments and suggestions for revision were utilized.  The
letters from EPA and Battelle are in response to questions raised by GPI,
and were included at the request of GPI and EPA.

-------
                                 G-2
              UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                    INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
                              RESEARCH TRIANGLE PARK
                              NORTH CAROLINA 2771 1
 May 28, 1976
 Dr. John Turk
 Glass Container Manufacturers  Institute
 1800 K Street,  N.W.
 Washington, D.C.  20060
 Dear John:
      I appreciate the  opportunity  for Dr.  Schorr and myself to
 get together with the  Air Quality  Task Force of GCMI to discuss
 your comments on the PSAD for  the-glass container industry.  At
 that meeting, several  questions were left  unresolved, and I
 indicated that  I would attempt to  provide  answers for or obtain
 further clarifications as to EPA position  regarding certain policy
 issues. These  issues  or questions are outlined below along with
 my response.

 1) Question:  What is  the national primary ambient air quality
 standard for  nitrogen  oxides?  What reference method is used for
 ambient sampling for nitrogen  oxide and what species is measured?
 What  is the national primary ambient air quality standard for
 sulfur  oxides?

 Answer:  The  national  primary  ambient air  quality standard for
 nitrogen oxides  is nitrogen dioxide ,(N09).  The standard for
 NC^ is  100 micrograms/cubic meter-annual arithmetic mean.  The
 proposed reference method 'for  measuring N0? in ambient conditions
 is chemiluminesce.  This method, however,  has not been promulgated,
but is  expected  to be  so in about 2 months.  This method indirectly
measures N02 by  first measuring NO in the  sample and then reducing
the N02 to NO in a reducing chamber and remeasuring the total NO
concentration in the sample.   The difference in the two measurements
represents the NO- concentration.

-------
                                  G-3
 the national primary ambient air quality standard for sulfur oxides
 is sulfur oxides (S0x> measured as sulfur dioxide (SO ).  The
 standard for S(>x measured as S02 is 365 micrograms per cubic meter-
 maximum 24 hours concentration.

 Conforming with these ambient -air quality for these pollutants, the
 PSAD will referenced these standards as N00 and SO .
                                           2       x

 2) Question:   If the ambient air quality standard for nitrogen
 oxide is for N02,  how is EPA justified in comparing an NO  emission
                                                          X
 which is predominately NO species emitted from a glass melting
 furnace to N0_ ambient air quality standard.

 Answer:   I have discussed this issue of comparing a  predominately
 NO emission species  against a NO,, standard  with our  physics  and chemistry
 experts  at RTF.  In  making this comparison,  this assumes  that  the
 NO-  emission  factor  is equal to the NO  emission factor.   They feel
 (and I concur)  that  this assumption is a fairly good  one  for the
 following  reason.  Once the plume has  been  diluted sufficiently with
 air  (dispersion calculations show that the  plume is diluted
 approximately  1000 to  1 at the  point where  it  touches  ground),
 the  photochemical conversion of NO  to  NO- is quite rapid
 particularly in  a urban environment.   Studies  has shown that
 the NO half life due to conversion  to  NO- is approximately one
 hour.  In  a rural environment,  it is approximately an  order  of
 magnitude  higher.  Based  on this evidence,  our assumption of using
                                    t
 an NO  emission  factor  to represent NO-  emission factor appears
     x                       c          2.
 to be good.  Unless  the task force  has ambient air sampling  data
 to the contrary, we  feel  justified  in  comparing  the N0x emission
 factor to the NO- standard.

Secondly,  if we were to  use NO  as  the reference instead  of  NO,,
for our source severity calculation, the risk  factor  (TLV X  8_ X _JL_ )
                                                             24   100

-------
                                  G-4
 would be the same as the ambient air quality standard for
                                3
 N0_.  The TLV for NO is 30 mg/m  and when adjusted using our safety
                                                             3
 factor of  1   (_8 x  1  ),  the risk factor becomes 100 pg/m  for the
 source severity calculation which is the same as ambient air quality
 standard for NO-.   Therefore,  in either case,  the source severity
 calculation would remain the same.
 3) Question:   The safety  factor  of   1   appears  to  be  conservative
 in the risk factor based  on a  comparison between TLV's  and  ambien
 air quality standards  for criteria pollutants.
Answer:   I have  reviewed  the data comparing our risk  factor  for
 source severity  based on  TLV's and then on an Ambient Air Quality
 Standard  for  criteria pollutants.  It would appear  that we have
 one to one correspondence between the two basis for determining
 exposure  which would cause an adverse health effect if for SO
 and particulates, a safety factor of 10 is used; for  N0_, a  safety
 factor of 33  is  used; and for CO, a safety factor of  1000 is used.
 I have discussed this issue with several people at RTF.  Our
conclusion is to leave the safety factor at 100.  The reasoning
is  as  follows: Expert judgement recommends a safety factor of 100
to  account  for exposure to a general population instead of a healtier
working population.  Also, there is no reason to believe that all other
pollutants will have the  same safety factor as evidenced by  CO being
off by a  factor of 1000.  Additionally, the health effects are better
known and  studied for the criteria pollutants for which ambient ao.r
quality standards have been"set, than for those pollutants for which
no standards have been set to date.  For these reasons, we will
continue to use the safety factor of 100 in our source severity
calculations for non criteria pollutants.

-------
                                    G-5
If my responses to these questions or issues are not satisfactory to

the Task Force, I will be glad to discuss these issues with them in

person at their convenience.  However, I do hope they shed some
light as to our current thinking on these issues.  Again,  let me
thank you and the Task Force for your help in reviewing this document.


Sincerely,
E. J. Wooldridge
Chemical Processes Branch

cc:
Dr. Dale A. Denny
Dr. J. R. Schorr

-------
                                     G-6
  Glass Container Manufacturers Institute, Inc.
                                                      1800 K Street NW, Washington, D C 2i
                                                                 (202)872-1280

                                                          TWX: 7108229337 GLASS VVSH
 July 2,  1976
 Mr.  Ed Wooldridge
 Reasearch Triangle Park
 Environmental Protection Agency
 Research Triangle Park, NC  27711

 Dear Mr. Wooldridge:

 Enclosed herein are limited comments from the GCMI Air  Quality Task Group
 relative to your June 1976 Draft Source Assessment Document.   In view of the
 time limitation imposed by your current schedule for  finalizing this report,
 we have had to hastily distribute copies of the draft,  and  have been unable
 to give it enough attention for a comprehensive review.   A  full  word by word
 analysis of this report has not been attempted.  However, several  questions
 have been raised where treatment was questionable, and  the  apparent impact
 of a glass manufacturing operation was seriously in error.   I  would like to
 point out that this draft still contains several errors,  significant internal
 contradictions, incomplete editing, and also-shows a  lack of attention to
 errors pointed out in earlier drafts for which there  were agreed  changes to
 be included .

 The  incorporation of the attached suggestions will still  not mean that we
 are  in accord with all of the remaining content of the  report  for the reasons
 given above.

 The  Air Task Group is very much concerned at the rush to  issue this report  in
 view of its current state, and the lack of time for comment allotted to the
 GCMI Task Group, which has attempted the fullest cooperation possible in the
 preparation of this document.  Ue wish to call to your  attention to earlier
 communication between Battelle, EPA, and GCMI concerning  the time available
 for  proper preparation of this source assessment.  At our first meeting,
 September 30, 1975, in Washington, you clearly stated that  there were no time
 constraints in publishing on this report.  It was with  this understanding  that
 there would be adequate time for constructive comment and corrective action
that the Committee agreed to cooperate in the source  assessment study. We
 regret that we consider further time necessary to complete  this document so  that
appropriate data is included and objective analysis is  performed, as this

-------
                                    G-7


 process  is  time  consuming  and expensive  to our membership,  as well as to
 Battelle and  EPA.  We  sincerely  request  that the  following  conments be
 considered  objectively and that  appropriate revisions  to  the source draft
 be made  and reviewed before any  consideration of  a  printing of  the final
 document is considered.

 Nitrogen Oxides  from Glass Melting Furnaces

 In earlier  meetings, we have questioned  several aspects of  the  treatment of
 nitrogen oxides  from glass melting in determining ambient impact of the plant.
 First we questioned  representation of total nitrogen  oxides as  the regulated
 pollutant.   It was clarified that nitrogen dioxide  is  the regulated pollutant.
 Then we  were  informed  that in converting a threshold  limit  value for nitric
 oxide to a  simulated ambient standard the value for nitrogen oxide would come
 out equivalent for nitrogen dioxide.   (This relation  depending  on the use of
 1/100 "safety factor").  We had  earlier  commented on  the  questionable use of
 the safety  factor, and it  is appropriate to further point out that were this
 a recognized  approach  for  setting a standard there  would  have been no justifi-
 cation not  to regulate total nitrogen oxides, rather  than consider nitrogen
 dioxide  the pollutant.

 We also  questioned treatment of  the furnace emissions  as  nitrogen dioxide, since
 in nitrogen oxide emissions for  combustion processes,  glass melting not
 exclusive,  it is well  established that approximately  95%  of the total oxides
 are emitted as nitric  oxide (NO),   We seriously questioned  this treatment in
 determination of a maximum ground level  concentration  only  a few minutes
 residence time from  the point of discharge.  The  EPA  response indicated the
 treatment is  based on  "rapid" conversion of nitric  oxide  to nitrogen dioxide in
 the atmosphere,  with the half life on one hour cited.  Our  own  reference
 searches have not turned up a good documentation of  atmospheric  residence,
 rather referring to  a  range of anywhere  from a few  hours  to a few days.  However,
 we consider our  question still valid, for reasons which will be shown later.

 Referring to  Table 12  of the June draft, we note  the  "representative" plant
 contains  a  single glass  melting  furnace  discharging through a single stack.'
 We note  the glass tonnage  of this  furnace has been  increased from 80,000 tons
 per year in the  preceeding draft,  to  115,000 tons per year  in the current
 draft.   Using a  typical  production  schedule of 350 operating days per year,
 115,000  tons  per year  equates  to  329  tons  of glass per day, a very high
 tonnage  for a "representative" furnace.   Since the nitrogen oxide emission
 rate  is  directly correlated  to production  tonnage, this illustration would be
 more  representative of near  the maximum  level  of  nitrogen oxide emissions,
 not a  "representative" level.

 We note in the same section  of Table  12  gas  flow  rate and temperature are
 shown.  These are unchanged  from the  earlier draft.   Considering a conservative
 5.5 million  BTU's per ton  to melt glass, and a typical 8% stack oxygen level,
we can only  account for the  production of  about 250 tons of glass per day at
this exhaust flow.   This,  together with  a  nitrogen oxide emission rate based
on the higher tonnage,  produces an imbalance  which shows nitrogen oxide
emission  concentrations (which will be directly reflected in downwind level
concentrations) about 32% too high.   (Supporting calculations are attached to
this letter).

-------
                                    G-8


 From Table 16, we find the maximum downwind concentration is located a
 distance of 880 meters (2,887 feet) from the stack.  From Table 12, we pick
 the  average wind speed used in the dispersion calculation of 6 meters per
 second  (19.8 feet per second).  Relating these two, we can determine the
 residence time of the nitrogen oxide from the point of stack discharge to the
 point of maximum downwind concentration is 2.43 minutes.

 Using the one hour half life for the oxidation of nitric oxide, as provided
 by Mr.  Wooldridge, we can calculate that in the 2.43 minutes atmospheric
 residence time of the nitrogen oxides only 3% of the emitted nitric oxide
 will  have time to convert to nitrogen dioxide.  Assuming that we start with
 the  well established 95% nitric oxide:  52 nitrogen dioxide ratio, at the end of
 2.43 minutes in the atmosphere we calculate that a total of 7.85% of the nitrogen
 oxide is nitrogen dioxide at the point of maximum downwind concentration.

 If we concede (ignoring the 32% high value for nitrogen oxides) 11.25 grams
 per  second emmitted, only 1.35 grams per second of this can legitimately be
 treated as nitrogen dioxide for the purpose of determining ambient dispersed
 concentration at the maximum downwind concentration location.  If this correction
 is not  made, we calculate that the nitrogen oxide value represented is 833% of
 the  correct value.

 In referring to Table 14, we first note the regulated pollutant is again
 represented as total oxides of nitrogen.  The pollutant, of course, should be
 nitrogen dioxide.  We next note the represented quality standard is the annual
 arithmetic mean of 100 micrograms per cubic meter, which, for the purpose of
 this  exercise is assumed here as a 24 hour standard.  We do not believe this
 is a  legitimate assumption.  In fact, we believe it introduces further error
 in the  analysis.  We also believe other authors have related an annual mean
 to a  24 hour standard by a ratio of about .3.   Indeed,  in reviewing the ratios
 between annual  and 24 hour for particulates (primary and secondary) and for
 sulfur  dioxide, pollutants where both annual  and 24 hour standards exist, we
 find  the Federal EPA related them as ratios of .29, .40, and .22, respectively.
 We submit that it would be consistant with EPA practice for a ratio of .3 to
 be used in converting from a one year to a 24 hour standard.  With reference to
 Table 14, an appropriate 24 hour standard for nitrogen  dioxide would, then, be
 333 micrograms  per cubic meter.   Calculating in the other direction,  converting
 a 3 minute maximum to an annual  mean, one would insert  the multiple of (.36) x
 (.3), or .11,  as the scaling factor.  We take note that Turner, who provided
 the basis for the scaling factor from the three minute  sample to a 24 hour
 sample, had questioned using his own data for scaling for sampling times of
 greater than two hours.   This points out that even experts in the field of
 ambient dispersion,  consider these statistical treatments uncertain, and
 certainly,  in  our view,  adds question to the practice of weighting every
 calculation to  the most severe disadvantage of the source.

To summarize the corrections we feel are legitimate, we have prepared a new
Table 14 for your study.

 If we use the  new figures that we calculate for maximum downwind concentration
of nitrogen dioxide, we would expect to see an increase in background nitrogen
dioxide concentration of only 4.55 micrograms per cubic meter, which we doubt

-------
                                    G-9


 •is measurable with current analytical accuracy.  To further illustrate the
 low ambient nitrogen dioxide impact within the vicinity of the plant, we
 are attaching a report summarizing ambient sampling of a members operating
 factory.  This was done under the indirect supervision of the San Joaquin
 County Air Pollution Control District of California and EPA Region 9, and has
 been given to Region 9 in its entirety.  The ambient samplers were located
 where results from EPA's UMAMAP computer program indicated a maximum ground
 level concentrations would be found.  Refer to Table 7, Figures 1 and Figures
 6-9, for the nitrogen dioxide analyses.  As is apparent in the data, the
 plant contribution to the nitrogen dioxide background is not evident.  It
 should be specifically noted, that the measured total  nitrogen dioxide
 background is not as high as the predicted contribution from the furnace
 alone as represented in the Battelle report.

 Hydrocarbons from Forming and Finishing - (Principally Decorating)

 Table 2 shows an emission factor of hydrocarbons from forming and finishing
 as .36 grams per kilogram of glass produced (.72 pounds per ton).  Based on
 an annual decorated tonnage of 364,000 tons (3% of total annual production)
 the industry emits from the forming - finishing operation a calculated 131
 tons per year of hydrocarbons.  While we believe the factor of .36 grns/Kg
 is high, our real question of the treatment of hydrocarbons relates to the
 assumptions used for the dispersion calculations.

 From Table 15, Source 2, we see the single source used in the calculation
 emits 79 tons of hydrocarbons per year, or an astonishing 60.3% of the total
 annual emission for the industry.

 Calculating another way with the Table 2 and Table 15  data, we can calculate
 this sample source has to decorate 219,000 tons of glass a year, or 190* of
 the glass produced in this sample plant (reference Table 12).   Either of these
 calculations illustrates that the emission assumptions used for dispersion
 calculations are absurdly high.

 Even assuming the hydrocarbon factor from Table 2 to be correct, and further
 assuming the rest of the calculation is right, choosing the realistic
 decorated tonnage we have given above and recalculating, would show a greatly
 reduced severity factor.

 Summary Comment

 A strong motivation of the glass industry in cooperating in the development
 of this source assessment study has been to see that a fair and impartial
 treatment is given the glass container industry and its impact on the surrounding.
 area and effected population.  We are certainly not attempting to obscure or
 cloud facts or potential problem areas.

While we would appreciate your incorporating the above comments into the
 final  report, we still would not be in a position to indicate to concurrence
with the balance of it since we have not had ample time to review it in detail.

-------
                                   G-10
As noted previously, our comments  simply  point out some of the  more  obvious
errors.  We note that this  report  still contains  many  errors  and  significant
internal contradiction.
                                     Sincerely,
                                    John  G. Turk
                                    Vice  President  - Technical
JGT/bb

Attachments )
     Dr.  Richard  Schorr

-------
                                   G-ll
                                                         eaneiie
                                                         Columbus Laboratories
                                                         ">(!"> kini> Avt-nui.-
                                                         fi.lumbus, Ohio 4H201
 A     *.  on   im£
 August  20,  1976
 Dr. John G. Turk
 Vice President - Technical
 Glass Containers Manufacturers  Institute
 1800 K  Street, N.W.
 Washington, B.C.   20006

 Dear Dr. Turk:

 In reply to your letter  of  July 2,  1976,  regarding  changes in the draft
 final of the  source  assessment  document  (SAD)  on  glass containers, we
 have reviewed these  comments  along  with representatives  from EPA and would
 like to offer the  following consensus  opinions.

 Your letter refers to  several serious  errors,  significant internal contra-
 dictions, and several  other faults, but we  can only deal with the ones
 which were actually  called  out;  and this  has been done.
                               Nitrogen  Oxides


We do recognize your position  regarding nitrogen oxide emissions but fail
to find any reasonable basis on which to  change the report.  Nitric oxide
is not the only oxide of nitrogen leaving the plant site, and little evidence
is available to document its relative proportion.  However, the report does
point out the NO  is the predominant  specie.  The report from Owens-Illinois
was most informative but does  not really  seem to conflict with the general
conclusions of the SAD.  It is well  demonstrated that NOX emissions can vary
widely depending  upon the  operating  conditions of  the melting furnace.
Mr. Wooldridge's  letter of May 28, 1976,  we  believe clearly indicated that
the calculations  for source severity were not changed by assuming NO to be the
pollutant instead of N02,  because of the  safety factor used for non-criteria
pollutants.  Again, it is  important  to  remember that the purpose of these
studies is not to set standards but  rather to define the needs for control
technology development in  the  glass  industry.

Regarding other specific comments in your letter,  Table 12 represents the
emissions from the entire  plant's melting furnaces and not a single furnace.
The title will be changed  to more clearly reflect  this.  The treatment of the
data in Table 14  has been  checked and while  we agree that there are a variety
of opinions regarding the  calculation of  maximum pollutant concentrations and
source severity,  the treatment which we used is generally consistent with that
done in source assessment  documents  for other industries.

-------
 Dr. John Turk                        G~12               August 20, 1976
                                 Hydrocarbons


 The hydrocarbon emissions in Table 15 came from NEDS.  We are currently
 reevaluating why this number was so high.  We believe the number to be in
 error and that we perpetuated the error.   Dropping this number would lower
 the emission rate in Table 15 to 2.88 x 10~2 gm/sec (1 ton/year) and change
 the values for XMAX (3 minutes)  and Xj^ (3 hours) to 3.11 and 1.58 ym/m3,
 respectively, and the severity factor to 9.9 x 10~3.
                                   Summary


 I would like to note that very little actual emission measurement data was
 supplied to Battelle by GCMI for use in the conduct of this study.  This
 is contrary to what is implied by your letter of July 2, 1976.

 We are most concerned that the report be consistent with other studies of
 like nature and that it accurately reflect the data available to us.  Your
 letter contains many statements inferring just the opposite.

 Should you have any additional questions regarding the report, you may call
 me on Extension 3624.   After making final changes, the report will be
 duplicated and copies sent to you, as per Ed Wooldridge's requests.
 Very truly yours

-------
                              G-13
      Gloss
      PockooinQ                  180°K street NW' Washin9ton> D-c- 20005
                                               (202)872-1280
                                         TWX: 7108229337 GLASS WSH

                                    October  18, 1976
 Dr.  Dale Denny
 Environmental  Protection Agency
 Research Triangle Park,  North Carolina 27711

 Dear Dr. Denny:

      We appreciate the opportunity extended the  Glass Packaging
 Institute (formerly Glass Container Manufacturers  Institute) Air
 Quality Task Group to attach comment to your  final report "Source
 Assessment of the Glass  Container Manufacturing  Industry",
 Contract Number  68-02-1323,  Task 37.

      As  we understand from our phone conversation on or around
 September 17,  1976 with  you,  your procedure does not allow us time
 to see  and comment on the final  draft document before it is made
 public.   Our comment, therefore,  is not a  critique of the wording
 and  data of the  final draft,  as  obviously  we  have no confirmation
:of what  is to  be in that draft.

      Rather than submit  further  detailed comment, as we have done
 in. reviewing the two earlier drafts of the Source Assessment
 report,  we wish  to relate to you some general observations we have
 about the Source Assessment  project,  the format of this project, and
.the  handling of  data. Specifically,  we wish  to comment on four
 general  points to which  we attach considerable significance.


 1.    General Report Format and Project Perspective

      The major concern we have with the format of this study, as it
was  presented in the earlier  report drafts, is that the summary
presentation does  not allow  us,  or others, to develop a perspective
about the  significance,  or severity,  of the emissions.  Although we
understand  it is  not the responsibility of the contractor to draw
conclusions, we  do  believe the summary format could include the
criteria  to  be used  by EPA in  evaluating the sources, and in defining
the  need  for further programs.

-------
                                G-14
 Dr.  Dale  Denny
 October 18,  1976
 Page -2-

 2.    Significance of Emission Data

      The  Glass Container Manufacturers, through the Glass Packaging
 Institute, have attempted to provide assistance in collecting
 emission  data for this study.  Information has been provided through
 response  to  requests from the contractor for emission data, or by
 way  of comment concerning the draft reports.  Referring to the draft
 of June,  1976, the industry is satisfied that a fair representation
 of the process emissions is reported.  We question comments by the
 authors,  inferring inadequate data still exists.

      Comparing the emission summary of Table 2 with comments in the
 text,  we  note some contradictions.  The emission summaries of Table 2
 are  reported in a manner indicative that statistically significant
 data has  been collected.  However, for some minor process sources,
 the  authors  have indicated relatively few individual measurements
 were available to them.  We question performing a large number of
 expensive measurements on sources which initial data and process
 analysis  indicate to have negligible emissions and insignificant
 ambient impact.  We believe for those processes which have a detectable
 ambient concentration, there is represented in the report an adequate
 body of data from which to draw conclusions.
 3.   Calculation of Ambient Concentration  in Source Severity
     Determination	   __ .	:  .   	

     The format provided the contractor for determining source
 severity is the calculation of maximum downwind ambient concentration
 (by common equations after Turner) and direct comparison with
 ambient air quality standards or converted threshold limit values.
 Where the emitted specie is chemically the same as the "regulated"
 compound, the resultant source severity comparison can be reasonably
 represented as a worst case example.  However, where the emitted
 compound is not a regulated compound, but is a precursor (that is,
 by atmospheric conversion or reaction may become the regulated	
 compound), the question of reaction time and  reaction rate cannot
 be ignored in the projection of the maximum ambient concentration
 of the regulated pollutant.

     In the glass container source assessment, the treatment of
 nitric oxide (NO), where it is assumed to react immediately to form
 nitrogen dioxide (N02) is questionable.  The reaction of nitric
 oxide to form nitrogen dioxide is not instantaneous, but is more
on the order of hours to days, depending on atmospheric conditions.
The point of maximum downlevel concentration of the effluent from
 a glass furnace is, timewise, only two to three minutes.  Only a
 small  portion of the nitric oxide can be converted in this time
 interval,  (3 to 5%).  The simplification then, in assuming
 instantaneous reaction in this case, predicts the source impact
approximately ten times higher than it should.  (Additional comments
about the mathematics of ambient disbursion calculations, and the

-------
                                G-15

Dr. Dale Denny
October 18, 1976
Page -3-

conversion units used in the Glass Industry Source Assessment are
attached to this letter.)

     The treatment of nitric oxide emitted is of concern to us in
the glass industry.  However, we wish to point out a generalized
treatment of this form will be a problem to any industry, so evaluated,
which emits a material considered precursor to any pollutant when
instantaneous reaction to form that pollutant is assumed rather than
a reaction time representative of the known reaction rate and
chemistry.
4.   Treatment of Control Technology

     A stated purpose of this project was to determine where
pollution control technology or process modification is needed,
Another purpose of this report was to determine where control  or
process technology is currently unavailable or unsuitable.   It is
our opinion that any useful discussion of emission reduction
technology must consider technical feasibility and economic
practicality considerations.

     The treatment in the two draft reports has considered  neither
technical nor economic factors but, instead, has been simply a
listing of claims and promotions from various sources, with no
screening of fact.  We'believe such a treatment has no value in
enlightening one as to what technology is available, adequate or
economical, or what development needs exist.

                                   Sincerely,
                                   John 6. Turk
                                   Vice President •*. Technical

OGT:drh

cc:  E. Wooldridge - EPA
     R. Schorr ~ Batten e

-------
                                    G-16
APPENDIX

Ambient Source Severity Determination of Pollutant Precursor

The preponderant nitrogen oxide formed in a combustion process,  such  as
in a container glass melting furnace, is nitric oxide (NO).   Measurements
on operating furnaces and literature references indicate no  more than 5%
of the total oxides of nitrogen is nitrogen dioxide (N02),  the regulated pollutanl
                                                                          *
Nitric oxide is considered a precursor to nitrogen dioxide  in atmospheric
reaction.  References to the rate of atmospheric conversion  of nitric oxide
to nitrogen dioxide are not specific.  Rather,  they refer to a range  of  a
few hours to a few days, depending on atmospheric conditions.  EPA project
personnel, responding to our original comments  about atmospheric reaction,
indicated the conversion to be "rapid," and cited a half-life for the
reaction of one hour.  A half-life of one hour  is used in calculations
presented in this correspondence.

Furnace ambient dispersion calculations by the  contractor (June, 1976, draft
report) located the point of maximum downwind concentration  of emissions
a distance of 880 meters (2,887 ft) from the furnace stack  (Table 16).      ' ""
The average ground wind speed used in the dispersion calculation was
6 meters (19.8 ft) per second (Table 12).  Using these values, the calculated
residence time from the point of stack egress to the point  of maximum down-
wind concentration would be 2.43 minutes.

Using the one hour half-life for the oxidation  of nitric oxide, only 3%  of
the emitted nitric oxide will be converted to nitrogen dioxide in the 2.43
minute atmospheric residence time.  Assuming an emission of 95% nitric oxide:
5% nitrogen dioxide, only 7.85% of the total nitrogen oxides would exist as
nitrogen dioxide at the point of maximum downwind concentration.

-------
                                    G-17
The ambient air quality standard for nitrogen dioxide is 100 micrograms
per cubic meter, annual arithmetic mean.  The mathematical ambient dispersion
calculation used represents a 3 minute maximum sample.  The question,  then,
is conversion from a 3 minute maximum to an annual mean.  A common scaling
factor for converting from a 3 minute to a 24 hour maximum is the multiple,
0.36.

The comparison of a 24 hour maximum and an annual arithmetic mean is a
point of disagreement.  The format of the June, 1976, draft was  to assume
                                                                         «
the 24 hour and annual values to be the same.  We disagree.  Most authors
relate an annual arithmetic mean to a 24 hour standard by a ratio of
about 0.3.  In reviewing the ratios between annual and 24 hour standards
for particulates (primary and secondary) and for sulfur dioxide  pollutants,
we find the Federal EPA related them as ratios of 0.29, 0.40 and 0.22,
respectively.  This is in good agreement with a general ratio of 0..3 for
the conversion.  In converting', then, from the 3 minute maximum  to an
annual mean, the multiple of (0.36) x (0.3), or 0.11, should be  used as
the scaling factor.

A further point of disagreement is the ~ ratio used in the formula to
calculate maximum ground-level concentration (Page 55, June, 1976 draft).
For D (neutral) atmospheric stability, the stability class used, the authors
cite a value of 1.6 (and appear to use the reciprocal, 0.63, in  the calcu-
lation).  The actual ratio shown in the cited Turner reference is 0.46.

The following restatement of the NOX portion of Table 14 illustrates the
effect of correcting the scaling factor, correcting the ^Jy ratio in the
dispersion formula and considering only the portion of NOX calculated  to
exist as the pollutant, N02, at the point of maximum ground-level concen-
tration.

-------
 . REPORT NO.
 EPA-600/2-76-269
                          ,p.     TECHNICAL REPORT DATA
                          (f lease read Instructions on the reverse before completing)
12.
                            3. RECIPIENT'S ACCESSION NO.
  TITLE AND SUBTITLE
 SOURCE ASSESSMENT:  GLASS CONTAINER
  MANUFACTURING PLANTS
                            5. REPORT DATE
                             October 1976
                            6. PERFORMING ORGANIZATION CODE
 J.R. Schorr, Diane T.  Hooie, Philip R. Sticksel, and
  Clifford Brockwav	
                                                       8. PERFORMING ORGANIZATION REPORT NO.
                           4D ADDRESS
 Battelle-Columbus Laboratories
 505 King Avenue
 Columbus, Ohio  43201
                            10. PROGRAM ELEMENT NO.

                            1AB015; RQAP 21AFA-013
                            11. CONTRACT/GRANT NOV

                            68-02-1323, Task 37
 12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC  27711
                            13. TYPE OF REPORT AND PERIOD COVERED
                            Task Final; 9/75-9/76	
                            14. SPONSORING AGENCY CODE
                             EPA-ORD
 is. SUPPLEMENTARY NOTESTask officer for this report is E. J. Wooldridge, Mail Drop 62
 919/549-8411 Ext 2547.  Earlier related reports are: EPA-650/2-75-019a and the
 EPA-600/2-76-032 series.		
 is. ABSTRACT
               repOrj- summarizes results of a study to gather and analyze background
 information and technical data related to air emissions from glass container manufac-
 turing operations.  It covers emissions from three plant areas: raw materials pre-
 paration and handling, glass melting, and forming and finishing operations.  Melting
 furnace emissions account for over 95 percent of the total plant emissions.  The major|
 pollutants are NOx,  SOx, and submicron particulates consisting mainly (over 90 per-
 cent) of mineral sulfates.  NOx has the largest emission factor (3 g/kg) with annual
 emissions of 36. 5 x 10 to the 9th power g.  Compared with national emissions from
 stationary sources,  NOx emissions from glass melting furnaces contribute 0.34 per-
 cent of the total.  Source severity factors determined by this study were 0. 38 for NOx,|
 0. 56 for SOx, and 0. 035 for particulates , with others being less than 0. 01.  Source
 severity is a measure of the potential environmental effect of air emissions and is the
 ratio of the maximum average ground level concentration to the primary ambient air
 quality standard for criteria pollutants.
 7.
                 DESCRIPTORS
Air Pollution        Dust
Assessments        Minerals
  lass Industry       Sulfates
  ontainers
Nitrogen Oxides
Sulfur Oxides
 s. DISTRIBUTION'STATEMENT

 Unlimited
EPA Form 2220-1 (9-73)
   KEY WORDS AND DOCUMENT ANALYSIS
                b IDENTIFIERS/OPEN ENDED TERMS
                __	—	
                 Air Pollution Control
                 Source  Assessment
                 Stationary Sources
                 Particulate
                 Mineral Sulfates
                                                                   c. COSATI Field/Group
                19. SECURITY CLASS (This Report)
                Unclassified
                20 SECURITY CLASS (Thispage)
                Unclassified
11G
08G
13B
14B
11B
13D
07B
                                         21. NO. OF PAGES

-------