4>EPA
           United States
           Environmental Protection
           Agency
           Office of Air Quality
           Planning and Standards
           Research Triangle Park NC 27711
EMB Report 79-ISC-7
August 1979
           Air
Industrial Surface
Coating Cans

Emission  Test Report
American Can
Forest Park, Georgia

-------
                                              INDUSTRIAL SURFACE COATING  CANS
                                                         EMISSION TEST REPORT
                                                                 AMERICAN CAN
                                                         FOREST PARK,  GEORGIA
Prepared For:
Mr. R.T. Harrison
Tech. Manager
ESED/EMB
Office of Air Quality Planning
and Standards
Environmental Protection Agency
Contract No. 68-02-2820
Work Assignment 13
          Prepared By:
         Samuel S. Cha
          Task Manager
      S. Dexter Peirce
Environmental Engineer
 TRC Contract 1170-E80

-------
                               TABLE OF CONTENTS
SECTION
  1.0
  2.0
INTRODUCTION
CONCLUSIONS
  3.0
  4.0
PROCESS DESCRIPTION
  5.0
      4.1

      4.2
      4.3
      4.4
      4.5
      4.6
      4.7
       5.1
       5.2
       5.3
METHODOLOGY 	
Coating-Solvent Mixture - Determination of Its
  Usage and Composition 	
Determination of Coating Applied  	
Determination of Moisture Losses  	
Sheet Counting  	
Oven Exhausts Flow Rate Measurement 	
Volatile Organic Compound (VOC) Measurements
Oxygen and Carbon Dioxide Measurements  . . .  .
RESULTS AND DISCUSSIONS 	
Coating and Solvent Used and Dried Coating Applied
Emission Measurements 	
Material Balance  	
 4

 7

 7
 8
 9
10
10
10
11

12
12
14
16
APPENDICES
    A

    B
VOC CONTENT OF COATING  .

PROCEDURE FOR MEASUREMENT OF TOTAL GASEOUS ORGANIC
  COMPOUND EPA REFERENCE METHOD 25
  "Determination of Total Gaseous Nonmethane Organic
  Emissions as Carbon" Manual Sampling and Analysis
  Procedure FR/Vol. 44, No. 195, October 5, 1979
     F

     G
PROCEDURE FOR DIRECT MEASUREMENT OF TOTAL GASEOUS
  ORGANIC COMPOUNDS USING A FLAME IONIZATION ANALYZER
  Published  in "Measurement of Volatile Organic Compounds"
  EPA-450/2-78-041 OAQPS No. 1.2-115, October, 1978

1.  Coating-Solvent Usage Date
2.  Sheet Counting Data

MASS OF COATING APPLIED ON SHEETS FIELD DATA
1.  By Weighing.  Corrected for metal sheet moisture
    weight losses.
2.  By Canco Resistance Gauge

FLOW RATE MEASUREMENT RESULTS

EMISSIONS MEASUREMENT RESULTS USING FLAME IONIZATION
  ANALYZER
                   EMISSION  MEASUREMENT  RESULTS USING EPA REFERENCE
                     METHOD  25  (TGNMO)
                   1.   Field Data  Sheets
                   2.   Analysis  Results
                                       -11-

-------
                                LIST OF FIGURES
FIGURE
  3-1
TABLE

  5-1




  5-2



  5-3



  5-4
                                                        PAGE

Sheet Base Coating Emissions  	      5




               LIST OF TABLES

                                                        PAGE

Coating - Solvent Mixture and Recovered Base
 Coating Operation - Three Piece Can Plant,
 October, 1979	     13
             i
Emission Measurement Results - TGNMO Base
 Coating Operation, 3 Piece Can Plant,
 October, 1979	     15

Comparison of TGNMO and FIA Mass Emission
 Results Oven Exhaust, Base Coating Operation
 3 Piece Can Plant, October, 1979	     17

VOC Material Balance Base Coating Operation
 3 Piece Can Plant, October, 1979	     18
                                      -111-

-------
1.0 INTRODUCTION




    As Work  Assignment 13 of  Contract 68-02-2820,  the  Office of Air  Quality




Planning  and  Standards,  Environmental  Protection  Agency  (OAQPS,   EPA)   has




assigned  TRC  Environmental Consultants,  Inc.,  to perform  a Volatile  Organic




Compound  (VOC)  material  balance  study  of  a  base  coating  operation  in  a




three-piece  can manufacturing  plant.   Results  from this  study  will  provide




EPA-OAQPS  with  additional  background  information  to  develop  New   Source




Performance Standards (NSPS)  for the can manufacturing industry.




    A  base  coating  operation  in  a  three-piece  can   manufacturing   plant,




operated by American Can Company at Forest  Park, GA,  was selected  by EPA-OAQPS




as the  object  of  the  material  balance evaluation.   The tests were  conducted




during the week of October 8,  1979  by  TRC  staff  along with  Mr. William  King of




Research  Triangle  Institute,  which  is also a contractor  of EPA.  and  an  EPA




staff member, Mr.  R.  T. Harrison.




    A  total of six  material  balance runs  were conducted.   For  each  run,




coating-solvent usage,  dry coating  applied on  the  can  sheets,  and volatile




organic  compound   emissions   from  oven  exhausts  were  quantified.   Volatile




organic  compound  material balances  for  the process  were  then derived.   This




report summarizes  TRC's approach,  test results and conclusions.

-------
2.0 CONCLUSIONS



    For  the  six  sets  of data  collected,  the weiqht  percentage of  the  total


coating-solvent used which  remained  on  the cured sheets as dry  coating ranged



from 42  to 55  weight percent  using  the  weighed sheet method and ranged from 53



to  58  weight  percent  using the electrical  gauging procedure.  Oven Volatile



Organic  Compound  (VOC) emissions  ranged  from 34 to  55 weight  percent of  the



total coating-solvent  used  when expressed as propane.  Because  these data  are



derived  from  the  TGNMO data  which  expresses VOC emissions as carbon, propane


was chosen as  a surrogate' compound  to estimate the molecular  weight  to carbon


ratio of the  VOC being  emitted.   Using propane as the surrogate  compound  and


the dry coating weights based on weighing  the  sheets  results  in an unaccounted



weight loss  ranging from 3 to  18  weight percent of  the  total coating-solvent



used.  Use  of the  dry coating weight  from the  electrical gauging  procedure


would  result   in  a  smaller unaccounted for weight  loss   (except  for  run  5) .



Based on the  coating  manufacturer's specifications  for  the  coating  used,  46
                            %

weight percent of the  coating was  solvent.   Assuming  this solvent  to be all of



the VOC  in the coating, then the percentage of the VOC  in  the  coating which is


accounted for  in  the  oven  VOC emissions  (including  the  cooling  zone)  ranges


from 74  to  79 weight percent  for  five  of the  six  runs  when the   oven  VOC



emissions are  reported with  propane as  the  surrogate compound.   (For run 1,


the oven emissions  appear to  be about 20  percent greater than  estimated  to be



available in the coating.)



    Finally,  additional  solvent was  used to provide  the proper  viscosity of


the coating during  application.  This additional solvent  was  not accounted  for


in the above calculations.  Data  from run 1 and 2 indicate that 6 to 7 weight


percent pure solvent was  added  to  the coating-solvent mixture.   The individual



solvent and coating weights were not recorded for runs 3 through 7.  For run 2,
                                      -2-

-------
this would  decrease  the percentage of  VOC emission accounted for  in  the oven




from 79 percent to 74 percent.   (For run 1,  the  oven  VOC emissions would still




be greater  than estimated  to  be  available  in the coating-solvent used by about



12 percent.)




    Run  7  was conducted with no coating  being  applied to the  sheets.   There




was a  substantial decrease   in  VOC emissions  from  the oven  exhaust  but  no




significant change  in the VOC  emission from the  cooling zone  exhaust  of the




oven.   It  is  assumed that most  of  these  emissions  represent residue  left  in




the oven from previous coating,  however, this assumption is not confirmed.
                                      -3-

-------
3.0 PROCESS DESCRIPTION




    The base  coating  operation tested  consists  of the application  and drying




of a white base coat onto metal sheets  prior  to  their  being printed and formed




into can  bodies.   The nominal size of  a sheet is  42  in.  x 35 in.  The coating




operation can  be  run  at a maximum  rate of  100 sheets per  minute.   Each sheet




is  first  fed  onto   a   conveyor   where  it   is  introduced  to  the  coating




application.  As shown in Figure 3-1,  the  application  roller applies the white




base coat to  the  top  of  the  sheet  and the cleaning roller cleans the bottom of




the sheet with backwash  solvent.  Both  the  application roller coating tray and




the cleaning  roller  backwash  solvent tray are  constantly  supplied  with,  and




recirculate, the coating  and  cleaning  solvent from two reservoir tanks.  These




two tanks are refilled  and  adjusted  for  proper viscosity  as  necessary.   The




sheets are counted by a counter before the coating application.




    The coated sheets  travel across  a six  foot  flash-off  zone and  enter  a




wicket oven  where they  are  dried.   The wicket oven  consists of  five zones.




The first three  zones are  heated  and  used  to dry  the coated  metal sheets,




followed  by  a cooling zone  where  the coated  and  dried sheets are  allowed to




cool  before  exiting  the oven.  There  is  also  a  preheat  zone  underneath  the




other  four  zones  where  empty  wickets  are heated  and  returned  to  the front




(coater)   end  of the  coating line.   Upon exiting  the oven,  the  coated sheets




are  recounted  by  three different  counters,  then  stacked  and removed  for




further processing.   The air is  taken into  the  oven  through an  air intake




stack on  the  roof.  Air  is  preheated through  a  heat  exchange duct arrangement




within the  cooling zone.  Part of  the  air is sent through  the  oven from the



rear end  of  the oven toward  the front  end  (near the  coater)  of  the oven,  and




is  then  vented  to the  atmosphere  through  an oven  exhaust  stack.   The other
                                      -4-

-------
                                                               FXHAUST
                                                               EXHAUST
I
Ol
I
       SOLVENT
      ADDITION
   SOLVENT/COATING
         TANK
                                                                             EXHAUST
                               APPLICATION
                               	,
                              ROLLER
        SHEET (PLATE)
           FEEDER
   SOLVENT ADDITION
SOLVENT TANK
                                                                            I

                                                                 SHEET (PLATE)
                                                                     STACKER
                                       FIGURE 3-1:  SHEET BASE COATING EMISSIONS

-------
portion of  the  air is sent  through the cooling  zone  and then  to  the cooling


zone  exhaust  stack.  The  temperature of  air in each  zone is  controlled and

                                                                             o
monitored  on  a  continuous  basis,  and  is maintained  at  approximately   400 F


for the first three zones.
                                      -6-

-------
4.0 METHODOLOGY




    The  following  sections  present  the methodology  applied in  determining  a




material balance of VOC in the three-piece can base coating operation.




    During  the  program,   a  total  of  seven  runs  were  completed.   Each  run




consisted  of  approximately  1900   sheets  and   lasted  approximately  thirty




minutes. Runs  1  through 6 were  typical runs, during which  sheets  were coated



and dried. For the  purpose of gathering background information,  a  seventh run




was conducted, during  which  sheets were sent through the oven  without coating




applied.









4.1 Coating-Solvent Mixture - Determination of Its Usage and Composition




    Coating-solvent  usage  was  determined  by  weighing  the  coating-solvent




mixture and backwash solvent before and after each run.




    Prior to each run  (1 through 6),  the coating  applicator  was drained of all




coating-solvent  mixture  and backwash  solvent.   All coating material  was then



mixed with  make-up  solvent to meet  the viscosity  specification.   The initial




weights  of  the  coating-solvent  mixture  and  the backwash  solvent were  then




measured using a 0-1000 pound Toledo Scale.




    During   each   run,   no  make-up  solvent   was    added   to   either   the




coating-solvent  reservoir  or the backwash  solvent reservoir.   After  each run




was  completed,  the coating  applicator  was  drained  of all  coating-solvent




mixture and backwash solvent  and returned  to  the reservoir buckets.  The final




weights of  the coating-solvent  mixture and the  backwash  solvent were measured




and  recorded.   The difference between the initial and  final  weights  of the




coating-solvent  mixture  and  the  backwash  solvent yields  the  coating-solvent




usage for the run.
                                      -7-

-------
    The VOC  content  of the coating has been  estimated  from the manufacturer's




specifications (see Appendix A).  This has  resulted  in  a conservative estimate




of  VOC content  of  the coating  since  additional  solvent was  added to  the




coating to  adjust the viscosity of  the coating.   During  runs  1  and  2  the




additional  solvent added  accounted  for  6.5  and 5.5  weight  percent of  the




combined coating  solvent  mixture.   For subsequent  runs these  individual data




were not recorded.









4.2 Determination of Dry Coating Applied




    Selected  sheets  were  weighed  before  and  after  the  coating  and  drying




operation to determine the mass of dry coating applied to each sheet.




    For each run  (1   through  6) ,  thirty sheets  were  numbered  with  a grease




pencil  and  an initial mass  was determined  for each sheet.   The  sheets were




then  returned  to the  sheet  stack  and were coated and  dried as  a  normal run.




For runs  1  and  2, the preweighed  thirty sheets  were placed within  the  first




sheets  in the  stack  to be coated.  In runs 3  through 6,  the preweighed sheets




were placed  on  line  approximately  half way through  the stack  (or  run) .   This




latter  procedure was  recommended  by  plant personnel who  indicated  that this




procedure  should  result   in  a  more  accurate determination of  average  dry




coating  weight,   since minor  adjustments  in  coating  thickness  are  normally




required at  the  start  of  a  run.   After  the coating and drying operation,  a




final  mass  was  determined for  each  preweighed sheet.  The  difference between




the initial and final  weights yielded the mass  (weight)  of coating  applied per




sheet.




    All sheet mass measurements were made  using  an Ohaus  triple beam balance




having  a capacity of  2610 grams.   Prior  to testing, the balance  was checked




against a  set of  standard weights.  The  balance pan  was  also fitted with  a




wire basket to facilitate the weighing of the  sheets.




                                      -8-

-------
    The wire basket portion  of  the total tare weight, changed  between initial




and final mass  determinations.   An adjustment was made  to  the  average coating




applied, determined by  weighing, according  to  the change  in tare weight,  by




addina  the   difference  between  the  final  and  initial  tare weights  to  the




calculated coating  applied.   For runs 1,  2 and  3  the  long  time (  2  months)




between the  initial and  final mass determinations caused a  difference  in tare




weight of 1.95 gram which  was added to the  average coating  applied  determined




for each of the three runs.




    A  method  that  measures  coating  weight  as  a   function   of  electrical




resistance,  developed by Canco, was also used.  For each of  the  thirty weighed




sheets, electrical  resistances  at  five different positions on each  sheet were




measured and  expressed  as mg  per   4  sq.  in.  These  five values  were  further




averaged and  converted  according to  the  size of sheets  to derive the  coating




weight per sheet.









4.3 Determination of Moisture Losses




    To  determine   metal   sheet  weight   losses,  if  any,   due   to  moisture




evaporation  during the  drying  operations,  a  blank  run  (run   7)  was  made.




During  the  blank  run,  all  sheets  were  sent through the  oven  without  first




being coated.




    Before run 7,  ten sheets were  numbered with  a grease pencil and an initial




mass  was  determined   for each sheet.    These  preweighed  sheets  were  then




introduced to  the  drying  oven approximately midway through  a run 7.   Once the




preweighed sheets  exited  the  drying  oven  and  were allowed  to  cool,  a  final




mass  was  determined  for  each  sheet.   The  difference  between  the  intial and




final masses of each sheet yields the moisture loss per sheet.
                                      -9-

-------
4.4 Sheet Counting




    As briefly described in Section  3.0,  several  sheet  counters are located at




the front end and the back end of the base coating operation line.




    Readings of  sheet  counting  at both ends were  made  and recorded.   Since 30




testing sheets (for coating weight determination)  were  removed before the back




end sheet  counters,  the back end counter readings were always  30  sheets less




than the front end counter readings.









4.5 Oven Exhausts Flow Rate Measurement




    Flow  rate  measurements were  conducted  at  both the  oven exhaust  (at the




front end of the oven)  and the cooling zone exhaust during each of the runs.




    A  standard  type  pitot  and  an  inclined  manometer  were   used  and  EPA




reference methods 1 and  2  were  followed  to  determine  flow rates.  Temperature




of  air  flow was also measured  by  using  a  chromel-alumel thermocouple  and  a




potentiometer.




    On the  first day  of the test program, all  exhaust  flow  rates were checked




and the oven  air flow was found  out of  balance,  i.e., the air  flow  rate from




the cooling zone exhaust  was  much  lower than its operating specifications.




Damper and  fan adjustments were  then made by  plant  personnel  to  balance the




oven  flow  according   to  design  specifications,   prior  to   starting  the  VOC




emission tests.









4.6 Volatile Organic Compound (VOC)  Measurements




    To measure the VOC  emissions at the oven exhaust  and  the cooling zone




exhaust  of the  base  coat  drying oven,  two different approaches were  used.




They were:
                                      -10-

-------
         "Determination of Total Gaseous Non-Methane Organic  Compounds  as Car-
         bon -  Manual Method",  known as  the TGNMO  method,   or  proposed  EPA
         reference method  25  (see  Appendix B) .  Samples  were anisokinetically
         drawn from the stack through a chilled condensate trap by means  of an
         evacuated gas  collection  tank.   Total  gaseous non-methane organics
         (TGNMO)  were determined by  combining the analytical  results obtained
         from independent  analyses  of the condensate  trap and evacuated tank
         fractions.  After sampling was completed, the organic contents of  the
         condensate trap  were  oxidized to carbon  dioxide which was  quantita-
         tively collected in an evacuated  vessel;  a  portion of the carbon  di-
         oxide was reduced  to  methane and measured  by a flame ionization  de-
         tector (FID).  A portion of  the  sample  collected in  the gas  sampling
         tank was  injected  into a gas chromatographic  (GC)  column to  achieve
         separation of the  non-methane organics  from  carbon monoxide, carbon
         dioxide and methane; the non-methane organics were oxidized  to carbon
         dioxide,   reduced  to methane  and  measured  by  FID.   This analytical
         work was  conducted by  TRC's subcontractor -  Pollution Control Sci-
         ence,  Inc.,  of Miamisburg,  Ohio.

         "Direct Measurement of  Total Gaseous Organic Compounds Using  a  Flame
         Ionization Analyzer",   (see  Appendix  C) .   This method is known as  FIA
         or FID.   A  sample was drawn  from the source through a  heated sample
         line and  glass  fiber  filter to  a   flame  ionization analyzer   (FIA).
         Ions formed in the combustion of  a  specific hydrocarbon  compound in a
         H2-O2  flame   established  a  current  that  is  approximately  propora-
         tional to the mass  flow rate of  the  hydrocarbon to  the flame.   This
         current was measured at two polarized electrodes, and was  read out on
         a potentiometric recorder and compared with a calibration curve  based
         on  propane   (C%g) .    The   results   are  reported  as  equivalents  of
         propane (C3Hg)  or methane (CH,})  or Carbon (C) .
4.7 Oxygen and Carbon Dioxide Measurements

    An integrated bag sample was  taken  at the  oven exhaust during Run 7.  Car-

bon dioxide and  oxygen  concentrations of  the  sample were determined  using  an

Orsat technique  known as EPA Method  3.   Data  were used for  the  indication  of

performance of the oven heat supplier burner.
                                     -11-

-------
5.0 RESULTS AND DISCUSSIONS




5.1 Coating and Solvent Used and Dried Coating Applied




    Coating and  solvent usage was determined  by weighing  the  coating-solvent




mixture  and backwash  solvent  before  and  after  each  run.   Results of  these




weighings are listed in Appendix D.




    Thirty  (30)  numbered  sheets  in  each  approximately  1900-sheet  stack  run




were weighed  before and  after  the coating-drying  operation to  determine  the




mass (weight) of  coating  applied to each sheet.  Results  of these weights are




listed  in  Appendix  E.   Results  of  the  weight  increase  on  each  sheet  were




averaged for each  run,  corrected  for  both tare weight and moisture loss during




drying.  Each  of  the  weighed  30 sheets per  run  were  also  measured by  an




electrical  resistance  gauge  to  determine  the  mass  of  coating  applied  per




sheet.  These measurements were made at  five  locations on  each  sheet.  These




results  are also  listed  in Appendix E.  Table  5-1 summarizes the results in




Appendix D  & E.   As shown in  Table  5-1 and  Appendix E,  the  coating-solvent




usage  for   each  1900-sheet run  ranged  from  87.5  Ibs  (39.7  kg)  to  135.0  Ibs




(61.3 kg).  After  the  two  test  runs  on the first testing  day  (Oct. 10, 1979),




the coating line  supervisor at  the plant  modified the application procedure to




increase the coating-solvent  usage because the amount  of  coating-solvent  used




in the first two runs was  lower than the standard specification.




    Table 5-1 also shows that the  values  of coating mass applied  on each sheet




derived  through  two different  measurement methods  are  different  from  each




other. The  value  derived  by using the  electrical  resistance gauge was higher




than that derived  by using the  weighing technique,  except for run  5.  Since we




are not  familiar  with the calibration  of the electrical  gauge  technique,  the




values derived by using our weighing technique are used in further discussions.
                                      -12-

-------
                                   TABLE 5-1

                  COATING-SOLVENT MIXTURE USED AND RECOVERED
                            BASE COATING OPERATION
                     THREE PIECE CAN PLANT, OCTOBER, 1979
Coating Weight Applied
on Sheet (Ib) % Recovered
Coating-Solvent (D Derived by Derived by
Run # Usage (Ib) Weighing Gauging Weighing Gauging
A B C B/A
1(2), (3) 87.5 36.8 50.4 42
2(2), (3) 97.0 47.0 56.4 48
3 129.5 62.6 70.5 48
4 135.0 74.2 77.6 55
5 126.5 69.0 67.4 55
6 ' 110.8 54.8 61.0 49
C/A
58
58
54
57
53
55
(1)


(2)


(3)
Assumes that  the  weight gained by the  backwash  solvent during the run
is coating-solvent recovered from the applicator.

Process modifications were  made  by the plant after  run 2 and prior to
run 3 to increase the coating usage.

During  runs  1  and  2,  the  30 test  sheets used  to determine  the dry
coating weight  were  introduced  near  the  beginning of  the  run;  for
subsequent runs the test sheets  were  introduced  later  in  the  run to
obtain a better estimate of  average  coating weight applied since minor
process modifications  are  sometimes  required  at  the  start of  a run.
                                     -13-

-------
5.2 Emission Measurements




    Flow  rate  measurements  at  both the  oven  exhaust  and  the  cooling   zone




exhaust are listed  in Appendix  F.   Measurement results indicated that the  flow




rate  at the oven  exhaust averaged  approximately 4,500  SCFM,  or  127 M /MIN,




and that  the  cooling zone exhaust  averaged approximately 15,000  SCFM,  or 425




M3/MIN.




    Volatile organic compounds  (VOC)  measurement results using  both  the Flame




lonization  Hydrocarbon  Analyzer   (FIA)  and  the  Total  Gaseous  Non-methane




Organic  (TGNMO or  Method 25)  are  listed  in  Appendices G  &  H.    Table  5-2




summarizes  the flow rates and  VOC  measurement  results  derived by  using the




TGNMO method.   As  indicated  in the  table,  the oven  exhaust  emission averaged




25.48 Ibs  (11.55  kg)  as carbon  among the  6  runs  and ranged between 21.95 Ib




(9.95  kg)   and 28.06  Ib  (12.73  kg)  as  carbon.   The  cooling  zone  exhaust




emission averaged  9.80  Ib (4.44  kg)  as  carbon  with  a  range between  4.85 Ib




(2.20  kg)  and  16.39  Ib (7.43  kg)  as carbon  among  the  six  runs.   The  total




emission, therefore, ranged from 28.83 Ib  (13.07  kg)  to  38.46 Ib (17.44 kg) as




carbon with an average of 35.28 Ib  (16.00 kg) per run.




    Since coating-solvent  usage data (see  5.1)  are  based  on  the  weight of




actual material but not  the weight  of the carbon contained in  the  material,  a




conversion  is  made  to  convert the emission  mass  of  carbon  to the  emission of




propane   (it   is   realized   that   even   the   propane   basis   may   not   be




representative).  After  the  conversion,  the  total  emission  ranges from 35.24




Ib (15.97 kg)  to 47.01 Ib (21.32 kg) for the six runs.
                                     -14-

-------
                                                                   TABLE  5-2

                                                      EMISSION MEASUREMENT RESULTS - TGNMO
                                                   BASE COATING OPERATION,  3 PIECE  CAN PLANT
                                                                 OCTOBER,  1979
Run
1
1
2
3
Ul
1 4
5
6
7*3
AVG
Run
Time
Min.
30
28
27
34
37
39
33
( Run 1 to

Flow
SCFM
4750
8
4510*
4500
4460
3840
4292
4443
Run 6)
Oven Exhaust
Cone . * 6
Rate (C )
M /M MG/L
134.52 2.378
2.583
127.72 2.783
127.44 3.570
3.294
Emission

Mass Emission
(Carbon)
GM Ib
9597
10424 ,
AVG 10010*
9953
12284
11334
AVG 11809*
126.31 2.824*^ 12128
1.446*
108.75 3.163
121.55 2.425
2.932
125.83 0.267

12727
11496
13899
AVG 12698*
1109
11554
21.16
22.98 j
22.07*
21.95
27.09
24.99
26.04*
26.74
28.06
25.34
30.65
27.99*
2.45
25.48
Cooling
Flow Rate
SCFM M /M
17700*7 501.61
17700 501.61
13550 383.74
13680 387.42
16480 466.71
14960 423.67
14000 396.48

Zone Exhaust
Cone.* 6
(C )
MG/L
0.494
0.222
0.445
0.371
0.333*5
0.133
0.280
0.333
Emission
Mass Emission
(Carbon)
GM Ib
7434 16.39
3118 6.88
4611 10.17
4887 10.77
4421*5 9.75*5
2198 4.85
3663 8.08
4445 9.80
Total
As
KG
17.44
13.07
16.42
17.02
17.15
14.90
4.77
16.00
Emission
Carbon
Ib
38.46
28.83
36.21
37.51
37.81
32.84
7.22
35.28
Total Emission
As C H *4
KG 3 8 Ib
21.32 47.01
15.97 35.24
20.07 44.26
20.80 45.85
20.96 46.21
18.21 40.14
5.83 8.82
19.56 43.12
•^Duplicate samples.
^Duplicate samples, second sample was deleted due to inconsistence with other  data.
      run, metal sheets only,  no coating applied.   Emission represents residues  remained  in  the  oven.
* Total emission as C, H. = Total emission as Carbon x  —
                     3  8                               36

*5No data, average of other runs were used.
*6TGNMO results, (see Appendix H)
*'fiun 2 flow rate use.
*8Run 2C flow rate used.

-------
    Results derived by  using  FIA method (at oven exhaust  only)  are summarized


in Appendix  G.   When evaluating  FIA results,  it was  found that  the analyzer



still indicated that a  certain  amount of emission was present even though the


coated  sheets  had  left  the  oven.   It was  believed  that both   the  gaseous


organic residue in  the  oven and  the  condensed organic compound in the  sample

                                     o
line  (sample line  temperature;  250 F)  were  the  major  contributors  of  the


observed  emission  "tail".  Mass  emission  calculations  (both  as  C,Hg and  as


carbon)  were  made  to quantify  both  with  the  "tail"  and  without  the  "tail".


Results are  listed in  the  summary  table  of Appendix  G.   These  calculations


show that  the  above mentioned  emission  "tail"  would account  for  15 -  30%  of


the total emission (except Run  #5,  in  which  the  emission tail  contributed  47%


of total emission).


    While comparing TGNMO and FIA  results  at  the  oven  exhaust,  TGNMO  result


averaged  23.1%  higher  than FIA  results without  the emission  "tail".   When



emission tails are  included,  FIA results are higher  in  4  runs and lower  in  2


runs (Runs II and  #3).  In  the  6 runs, TGNMO results generally  agree with FIA


results with the emission tail;  the  differences  are within ± 35%  of  the  TGNMO


results. Table  5-3 shows  the  comparison.    (For  material  balance, only  TGNMO


results were used.)






5.3 Material Balance


    Table  5-4  summarizes  the  VOC  material  balance  of  a  base  coating


operation.  For the six sets  of data  collected,  the weight percentage of the



total coating-solvent used  which remained  on the cured  sheets  as  dry  coating


ranged from 42 to  55 weight percent  using  the  weighed sheet method and  ranged


from 53  to 58 weight  percent  using  the electrical  gauging  procedure.   Oven


Volatile Organic Compound (VOC)  emissions ranged from  34 to 54  weight percent
                                     -16-

-------
                    TABLE 5-3

COMPARISON OF TGNMO AND FIA MASS EMISSION RESULTS
      OVEN EXHAUST, BASE COATING OPERATION
        3 PIECE CAN PLANT, OCTOBER, 1979
FIA Results
TGNMO W W/0 % Ratio
Results Emission Tail FIA/TGNMO
Run as Carbon as Carbon W W/0
# Ib Ib Ib Emission Tail
1 20.07 20.67 17.01 93.7% 77.1%
2 21.95 26.12 19.24 119.0% 87.7%
3 26.04 22.53 18.59 86.5% 71.4%
4 26.74 29.12 21.80 108.9% 81.5%
5 28.06 37.29 21.36 132.9% 76.1%
6 27.99 28.07 20.33 100.3% 72.6%
AVG (1-6) 25.48 27.30 19.72 107.1% 77.4%
% Difference
FIA VS TGNMO
W W/0
Emission Tail
-6.3% -22.9%
19.0% -12.3%
-13.5% -28.6%
8.9% -18.5%
32.9% -23.9%
0.3% -27.4%
+7.1% -22.6%
*Percent difference (FIA vs TGNMO) = FIA ~ TGNMO x 100%
                        TGNMO
                     -17-

-------
                                   TABLE  5-4

                              VOC MATERIAL BALANCE
                             BASE COATING  OPERATION
                        3 PIECE CAN PLANT, OCTOBER, 1979
            Coa t ing-SoIven t
                Usage
 Total Oven
VOC Emission
Coating
Applied
Run (as Actual Material)
# lb
1 87.5
(100%)
2 97.0
(100%)
3 129.5
(100%)
4 135.0
(100%)
5 126.5
(100%)
6 110.8
(100%)
(As C3H8)
lb
47.0
(54%)
35.2
(36%)
44.3
(34%)
45.9
(34%)
46.2
(37%)
39.8
(36%)
(as Actual Material)
lb
36.8
(42%)
47.0
(48%)
62.6
(48%)
1 74.2
(55%)
69.0
(55%)
54.8
(49%)
Difference*
lb
3.7
(4%)
14.8
(15%)
22.6
(18%)
14.9
(11%)
11.3
(9%)
16.2
(15%)
*Difference = Coating-Solvent Usage - Oven  Emissions-  Coating  Applied
                                     -18-

-------
of  the  total coating-solvent  used when  expressed  as propane.   Because these




data are derived  from  the  TGNMO data which expresses  VOC  emissions as carbon,




propane was chosen as a surrogate  compound  to  estimate the molecular weight to



carbon  ratio of  the  VOC  being  emitted.   using  propane  as  the  surrogate




compound and  the  dry coating weights based on weighing the sheets  results in




an unaccounted  weight  loss ranging  from  4  to  18 weight percent of  the total




coating-solvent  used.   Use  of  the dry  coating  weight  from the .electrical




gauging procedure would result  in a smaller  unaccounted weight  loss (except




for  run  5).    Based  on  the  coating   manufacturer's  specifications  for  the




coating used,  46 weight percent of the  coating was  solvent.  Assuming  this




solvent to be all of the VOC in  the coating, then the  percentage of the  VOC in




the  coating  which  is  accounted for  in  the  oven  VOC emissions  as  propane




(including the  cooling  zone)  ranges from 74 to  79  weight  percent  for  five of




the six runs.   (For  run 1,  the oven emissions  appear to be  about  20 percent




greater than estimated to be available  in the coating.)




    Finally,  additional  solvent was used to  provide  the  proper  viscosity of




the coating during application.  This additional solvent was not  accounted for




in the above  calculations.  Data from run 1 and 2  indicate  that 6  to 7 weight




percent pure  solvent was added  to  the coating-solvent mixture.  The individual




solvent and  coating weights  were  not  recorded for  runs  3  through 7.   For




run 2, this would decrease the  percentage of VOC emission  accounted for  in the




oven from 79  percent to  74  percent.   (For run  1, the  oven VOC emissions would




still be greater  than estimated  to  be available  in  the coating-solvent used by




about 12 percent.)




    Run 7  was conducted with  no coating  being  applied  to  the sheets.   There




was  a  substantial decrease  in  VOC emissions   from  the oven  exhaust but  no
                                     -19-

-------
r
         significant  change in  the  VOC emission from  the  cooling zone  exhaust  of the



         oven.   It is assumed  that  most of these  emissions  represent residue  left in




         the oven  from previous coating, however, this assumption is not confirmed.
                                               -20-

-------