DETERMINATION OF TOTAL GASEOUS



   NONMETHANE ORGANIC EMISSIONS AS CARBON
     SUMMARY OF COMMENTS AND RESPONSES
        Emission Measurement Branch



Emission Standards and Engineering Division
    U.S. Environmental Protection Agency



Research Triangle Park, North Carolina 27711








                August 1987

-------
                             TA3LE OF CONTENTS



                                                                  Page



Chapter 1.  Introduction 	    2





Chapter 2.  Summary of Changes Since Proposal	    3





Chapter 3.  Summary of Comments and Responses	    4





     Table 1.  List of Acronyms Used in Summary of Comments



               and Responses	    1



     Table 2.  List of Commenters.	   16

-------
                      Taole 1.  LIST OF ACRONYMS UStJ
EPA    - Environmental Protection Agency



VOC    - Volatile Organic compounds



NSPS   - New Source Performance Standards



QA     - Quality Assurance



VOC    - Volatile Organic Compounds

-------
                                  >.apter 1
                                 INTRODUCTION

      On  November 7,  1986,  the U.S.  Environmental  Protection Agency (EPA)
 published in  the Federal  Register (51 FR 40448)  the revised Method 25,
 "Determination  of Total  Gaseous  Nonmethane  Organic Emissions as  Car-Don."
 This  method was  proposed  under the  authority  of  Sections  101,  111, li<:,
 116,  and  301  of  the  Clean  Air Act,  as amended.
      Public comments were  solicited  at  the  time  of proposal.   To  provide
 interested persons the opportunity  for  oral presentations of data,  v^ews,
 or arguments  concerning the proposed  revisions of  the test  method,  a
 public hearing was scheduled  for  December 22, 1986, at the  Research
 Triangle  Park, North Carolina, but no person desired to make an oral
 presentation.   The public comment period was from  November  7,  1986, to
 January 21, 1987.
     Three comment letters were received concerning issues  relative to the
proposed test  methods.   A detailed d.iscussion of  these comments and
responses  is  summarized in this document.  The comments  have been carefully
considered,  and  appropriate changes  have been  made  in  the  proposed
revision.

-------

-------
                                 Chapter 2



                     SUMMARY OF CHANGES SINCE PROPOSAL






     The following changes have been made in the revised method since




proposal:



     1.  Section 4.1.1 has been added to include a contamination cne<;:<




of the sampling equipment.



     2.  Section 4.1.2 has been revised to allow the use of a vacuum




gauge in place of a mercury manometer for measuring system pressure in




the field.



     3.  Section 4.1.4 has been revised to reduce the leak check period




from 10 to 5 minutes.

-------
                                  Chapter 3



                      SUMMARY  OF  COMMENTS AND  RESPONSES



 Commenter IV-D-1



      1.1   Comment:   Although  it  was  not  the  intent  of  the  writers  of the



 method  that  it  be  applied  universally  to any  source emitting  volatile



 organics, past  practice  has  shown  that State  and  local  -egulatory  agencies,



 and  even  the  EPA,  will adopt  the method  without  regard for its  limitea




 applicability.  Thus,  it  behooves  the  EPA to  write  reference  methods t'lal



 have applicability  beyond  a  narrow scope of  specific sources.   An  arbitrary



 method  with diminished value  should  be avoided.




      For  the  method to be  applicable to  a wide  range of  sources  not



 covered by new  source  performance  standards  (NSPS),  or studied  during tne



 development of  these  changes,  the  method should  be  written in the  form  of



'general guidelines.   These would take  the form  of a  stringent quality



 assurance (QA)  program to  insure acceptable performance  for the  sources  to



 be sampled.   The changes  in the  method needed to make  it applicable  to



 specific  sources would then be sanctioned, rather than have poor or



 meaningless data delivered with  a  rigid  procedure.   Thus,  much  data  al-eady



 collected using variations of  the  present  Method 25  wi11 maintain  izs



 validity.




      Response:  We  agree that  it was not  our  intent  that the metnod  be



 applied universally to any source  emitting volatile  organic compounds



 (VOC).  we also agree that it would  be beneficial if the method could be



 written to have a broad applicability  to  all types of  sources, but we do



 not  agree that this should be done at  the expense of the method's  accuracy



 and  precision at those sources for which  it is the compliance test method.
                                     4

-------
Our experience v/ith performance guidelines for "lec.iod ^3 ;j cia: t :e>  /;.'-



become so cumbersome that the method would not be useable for routine



testing.  However, this does not preclude any individual or organization



from performing the necessary testing to demonstrate that their procedures



or equipment are adequate for the requirements of Method 25.



Commenter IV-0-1



     1.2  Comment:  The intent of the filter system is to remove



condensible organic and other particulate material from the sample gas.



The filter temperature was chosen to correspond to the filter temperature



of EPA Method 5, which was selected to prevent collection of sulfjric acid



mists on the filter and has nothing to do with photochemical reactivity of



volatile organic materials.  This arbitrary filter temperature should oe



replaced with one consistent with removal of condensible organics or



particulate that does not react photochemically to form ozone.



     Some organic compounds that condense in the atmosphere and, therefore,



do not react pho-tochemically, pass through a filter at 250°F and are



measured by the method.  This biases the measurement to the high side.



These materials should be excluded from the sample.  We suggest  that the



filter temperature be reduced to 190°F.  Even at this temperature,  some



nonphotochemical ly reactive condensible organic materials will pass th-i



filter and be measured, but the potential problem of the sample  gas



temperature falling below the dew point will  be avoided.



     There is no need for the Method 25 filter temperature  to correspond



to that of Method 5.  There is no relationship between organic emissions



and particulate emissions.  Each test method reports its results in units



that are not comparable.   If the agency wants to measure the condensible
                                     5

-------
  organic .naterial in the emissions,  It should designate a inet.icc 'idS'r-e:



  do so.  There are satisfactory methods in use for this purpose that coula



.  be appropriated.  We suggest the Oregon Department of Environmental



  Quality Method  7.



       Response:  We do not agree that all photochemically reactive  organic



  compounds must  be gases at ambient temperatures.  Even those compounds that



  are normally  liquids at ambient temperature may have a significant  vapor



  pressure and thus exist partly in the gas phase as evidenced by the



  evaporation of  organic liquids stored in open containers.   If the  filtration



  temperature were reduced from 250°F to 190°F, we believe that conpounds



  that should be  counted as VOC would be artificially excluded from  the



  sample.



  Commenter IV-D-1



       1.3  Comment:  The filtering system could be greatly simplified and



  the oven design improved.  The design of these components should not be



  rigidly specified such that possible improvements are precluded.   For



  example, the solvents used to clean the probe and filter system between



  samples should  be thoroughly removed before assembly of the equipment to



  preclude contamination of the sample.  The long purge time could then be



  eliminated.  As another example, it would be desirable to reduce the volur-e



  of the filtering system so that the error in sample volume would then be



  less than 1 percent.  The filter system volume could be reduced further by



  using  1/8-inch  tubing throughout except for thermocouple wells.  A 3/8-inch



  filter holder can be constructed of Swagelok parts.  This would eliminate



  the need for a  quarter of the equipment that would be hauled to the



  sampling'site.  Lastly, we see no need for an elbow in the probe.

-------
               :  There is no requirement  in tne method to clean the prooe
 and filter system between samples so this was not a factor in determining
 the length of purge time.  The purge time of 10 minutes was chosen to
 insure that the sampling system was thoroughly purged with sample gas and
 at equilibrium and does  not seem excessive.   Since there is a sample p.,.-^
 before sampling begins,  there  is no error in  sample volume and the voi.^e
 of the filtering system  is  not critical.
      The elbow at  the end of the probe  allows  the opening  of  the  probe to
 be pointed away from the direction  of  sample  gas  flow  to reduce the  amoun:
 of particulate matter entering the  sampling system..
 Commenter IV-D-1
      1.4  Comment:   The  leak check  can  be  reduced  from  lu minutes  to
 5  minutes.  For a  probe  system with  0.1 liters volume, a sample size  of
 3  liters,  and  a leak  rate of 1  percent, the pressure would  change  19  mm in
 5  minutes.  Reducing  the probe  system volume would reduce the time needed
 to  conduct the leak check.  It  should be noted that a leak   can be detected
 in  the same manner by pressurizing the system, with the added advantage
 that in the event of a leak, its location  can  be easily found by wetting
 the component connections and  looking for  bubbles.
     ResP°nse:   We a9ree  that  the leak-check  procedure can  be reduced from
 10 minutes to 5 minutes.   The  method has been  revised to reflect  this.
     While a leak  check could  be conducted under  positive pressure, it is
more appropriate to conduct  the leak  check  under  vacuum which  is how  the
sample train  would  be operated  while  collecting the sample.

-------
Commencer IV-D-i



     1.5  Comment:   The proposed sampling procedure ignores the need for



isokinetic sampling if participate is present.   If the stack temperature :s



less than the filter temperature, condensed organic material, if present,



will be disproportionately sampled.   If the stack temperature is g^eate"



than the filter temperature, a nonisokinetic sampling system can be 'jsea.



     A lower filter temperature would reduce the need for isokinecio



sampling.  The filter temperature would be lower than the stack teuro erasure



in more situations.  Also, errors induced by nonisokinetic sampling when



the filter temperature is greater than the stack temperature would be



reduced because the difference in temperature would be smaller and the



amount of particulate volatilized in the sampling train would be less.



     Response:  We believe that isokinetic sampling is not necessary



because particles formed by condensation are typically submicron in



size and will behave as a gas.  Thus, even if the sample is not collected



isokinetically, it will not contain  a disproportionate amount of



condensed organic matter.



Commenter IV-D-1



     1.6  Comment:   Mercury manometers are not  suitable for field use.



A vacuum gauge is suitable to estimate vacuums  in the field.  The initial



and final tank pressures can be measured before leaving and upon return  to



the laboratory.  We routinely hold evacuated sample tanks for several  weeks



without experiencing even small leaks.



     Response:  We agree that a vacuum gauge of sufficient accuracy would be



suitable to measure vacuum in the field.  The method has been revised to



reflect this.

-------
Commenter  IV-0-1; Commenter IV-D-2



     1.7   Comment:   A post-test leak check should not be eliminated,  a



'post-test  leak check is necessary to assure sample integrity.  If the



promulgated method incorporates the elaborate filter system, a post-test



leak check will be even more of a necessity considering the number of



connections suggested in the proposal.



     Response: At the beginning of the run, the system vacuum is  at its



highest and the potential for leaks is greatest.  This is the most



critical time for leak checking the sampling train.   If no leaks  are



detected at this point, it is very unlikely that any wou-d develop during



sampling.  At the end of the sampling period, the vacuum remaining in the



cylinder will be near zero and the potential  for leaks in the system will



be very low.  Also if a post-test leak check were added, it would lengthen



the time required to complete a test and would be very unlikely to add any



additional information.



     1.8   Comment:   A trap burnout temperature of 2(JO°C is specified by



the proposed method.  This temperature may suffice for the more volatile



solvents.  However, in some sources, specifically some for which  there



are no NSPS, organic material  captured in the cryogenic trap may  pyrolyze



rather than vaporize at these temperatures and will  not be recovered.



     To insure complete removal of organic material, the traps should be



heated to  600°C.  At this temperature, carbon residuals left by pyrolysis



of heavy organics are burned to carbon dioxide and removed from the trap.



The EPA chose the 200°C trap heating temperature out of fear from C02 would



diffuse through the stainless steel trap at higher temperatures and
                                     9 .

-------
 contain: nata tne sa.np'e.   Ae  ha/a  -leva"  had  a  p-oi'e;:,  f.-j.r.C-';  ::•-.;  -..



 into the traps at 60U°C trap  heating  temperatures.  Our  experience was



 that trap blanks of  less  than  5 ppm were  readily  achieved  at  the 600°C



 temperature, even when the trap was bathed  in  C02 from an  acetylene torch



 EPA's own studies did not indicate a  problem  with C02 diffusion  ;ito  :,ie



 trap when heated to  60U°C.



     Resp_on_se:  Because the  sample is filtered  at 121JC, all  mate^'a "is



 reaching the condensate trap  must be  a  gas  at  that  temperature.   We



 believe that a material which can exist as  a  gas  at 121°C  will be



 vaporized easily when heated to 200°C.  At  the  same time,  a tempest j~o



 of 200°C should be low enough to avoid  decomposing  the sample during  the



 vaporization process.  Thus, the 20U°C  temperature  was chosen because  it



 provided a temperature that was high  enough to  vaporize the sample  while



minimizing any problems due to decomposition  from temperatures that are



too high.  Diffusion of C02 through stainless steel at high temperatures



was not a factor in choosing a heating temperature.



Commenter IV-D-1




     1.9  Comment_:   Materials other than  chromia  on alumina are  superior



catalysts.   An oxygen donating cacu'ys-; s«ic:i as manyanese  ;i^ije,  ,  j«.;



be allowed  so that  when more organics  are present than oxygen is  available,



the catalyst will  provide the additional  oxygen -»c'j :-?<-; *  ; ^n^rt 3.: '  c



the organic carbon  to carbon dioxide.




     The  EPA contractor developing this method evaluated manganese dioxide



as an oxidation catalyst  but found it  unacceptable.  However,  the

-------
contractor evaluated tne catalyst ac an i napprcpri ate cpe"i:~i"iG i-'-Cc~;~, .~;



(4o4°C).  The chromia on alumina also failed to perform adequately at tn:s



temperature.  When operated at the proper temperature (6bO°C), manganese



dioxide is a superior catalyst to chromium on alumina because of its



oxygen donating properties.



     The manganese dioxide catalyst does degrade at temperatures above



700°C.  Thus, proper temperature control is necessary.



     Chromia on alumina tends to spread the peaks.  Sharper peaks from toe



chromatography can be obtained with manganese dioxide.  The EPA contractor



reported a peak asymmetry factor of 1.38 for the chromia on alumina



catalysts.  The National Council of the Paper Industry for Air and Stream



Improvements, Inc. found a peak asymmetry  of 1.24 for the manganese dioxide



catalyst.  A factor of 1.0 indicates asymmetrical peak,  and is preferred.



The sharper peaks allow for more accurate  determination  of the



concentrations.



     Use of an oxygen donating catalyst alleviates the need for adding



oxygen or air to the carrier gas prior to  the oxidation  catalyst,  thus



simplifying the analysis system.  It avoids an additional  hazard in ^he



work place.  There is SJ.""f icient oxygen ',rt c.ie sample i; - Hj -eye'ic*- :,x-



the catalyst upon each injection.  We have used a manganese dioxide



catalyst for several years without "egeneration  nth r!~ )th,?r ^har Vr1-.:.



in the sample.  We believe that ;ne manganese Jiox:de Cacalyst -:noun



be specified, or at least be allowed, provided that the  QA portion of the



method is satisfied.



     Response:  Although there is good evidence that manganese dioxide



based catalysts have limited ability to oxidize organic  compounds  in an
                                     11

-------
overloading.   There is also evidence that, in the absence of sufficient



oxygen, some  of the organic material is converted to carbon which deposits



on the surface of the catalyst and is subsequently converted to carbon



dioxide when  sufficient oxygen is present.  It appears that this process



does not always go to completion and may, over a period time, leaa to loss



of oxidation  efficiency.  Even though manganese dioxide has the ability to



donate oxygen for oxidation in an oxygen poor environment, it would still



be necessary  to add auxiliary oxygen to ensure good performance with



samples having high organic levels.  Thus, it's ability to donate oxygen



does not give it any special advantage over other catalysts.



     Although most people who use manganese dioxide catalysts operate them



at 850°C, the manufacturers of two of the most widely used manganese



dioxide-based catalysts recommend that they be operated at temperatures no



higher than 454°C.  This is the temperature at which they were operated in



our tests, as well as 850°C.



     Even though manganese dioxide may have shown a slightly lower



asymmetry factor than that of chromia on alumina, the difference is not



significant.   As a result of our evaluation,  we concluded tnat cnro'.rii or,



alumina represented the best choice of a catalyst considering all the



factors that  affect a catalyst's oerformance.



Commenter IV-D-1



     1.1U  Comment:  Our experience with nickel reduction catalysts is



that they fail rapidly.  Any sample containing sulfur compounds poison the



catalyst.  We have used a rhodium catalyst on a variety of sources with no



degradation.   We have not replaced our reduction catalyst in 7 years of
                                     12

-------
 concinua!  service,   .ve  S'jyges:  t.ia.1:  t.ie rnojiuc, catilys:
 at  least  be  allowed  provided  that  the QA portion  of  the nethod is
 satisfied.
      Response:   We would  not  expect  to collect  sulfur containing compounds
 from  sources  where Method 2b  is  specified  as  the  compliance  test -nethod.
 We  did  not evaluate  rhodium as a reduction  catalyst  and could  not spec:''/
 it  as the  reduction  catalyst  for Method  25.   However,  the  rhodium catal /st
 could be approved as  an alternative  if sufficient  data  were  presented  to
 support its use.
 Commenter  IV-D-2
     2.1  Comment:   The proposed method  does  not address the prooleiTi of
 residual contamination in the sampling train  which users of Method  25  have
 confirmed to  De a serious problem.   A  laboratory confirmation  of  the
 presence of residual  contamination  from  a previous test  or laboratory
 contamination needs to be specified.   The proposal should incorporate  a
 pretest conditioning and QA check of the sampling equipment which will
 contribute to the overall  reliability of Method 25.  Such a procedure  could
 consist of baking the probe and trap assembly while purging with  a suitable
 organic-free  gas and performing an  analysis on tne trap and Sainp'.e tank.
     Response:  We agree that  such  a procedure should be included in
Method 25.   We have revised tne nethod to incorporate a c :nt.Tni Tat; on
check.
Commenter IV-D-2
     2.2  Comment:   The  proposed  particulate filtration system is
unnecessartly complex since particulate concentrations from applicable

-------
incorporate this elaborate filtering system is estimated (by EPA in 19db)



to cost approximately $1,700.  Although we agree that a filtration system



is needed, we believe that an in-stack filter (as mentioned in the proposal



as an option when electrical  systems are prohibited) would be adequate.   nw



using an in-stack filter, the probe, from the back half of the filter



holder to the condensate trap, could then be baked along with the trap to



determine the condensible organics as in the present system.  The method



could alternatively allow the more elaborate filtration system for problem



sources.



     Response:  We believe that an in-stack filter system is appropriate



only at those sources where a heated out-of-stack filter would be



prohibited for safety reasons.  Because many organic compounds can change



physical state between ambient temperature and 250°F, the temperature at



which the sample is filtered  determines whether some organic compounds



are counted as VOC or not.  Under these circumstances, we believe that tne



only equitable way to measure VOC emissions is to filter all samples at,  the



same filtration temperature regardless of the source's operating



temperature.



Commenter IV-D-2



     2.3  Comment:  The tolerances ostao1 : shed by E0/n ^jr 0* •; - c a 3 r to '-,-



very stringent for such a complex analytical system.  Although we have had



limited experience with the system, we are concerned wirn the achievaoi1ity



of the proposed accuracy standards.  We oel ieve the present standards are



adequate and should be maintained.

-------
     Re_sp_pnse:  Our laboratory program to Devise trie net.'; a ;-rons : "itc :



that the tolerances proposed in the revised method should be achiavaole



by any well operated system.  Relaxing the tolerances would be li,
-------
                                       OF
                           Docket Number A-36-U5
Docket Item Number
      IV-D-1
      IV-D-2
      IV-0-3
Commenter/Af f 11 iati oji

John £. Pinkerton, Manager
Ai r Quality Program
National Council of the Paper Lidii
 for Air and Stream Improvement,  {
260 Madison Avenue
New York, New York 10016

Ben A Brodovicz, Chief
Division of Technical  Service and
 Monitoring
Department of Environmental Resour
Commonwealth of Pennsylvania
Post Office Box 2063
Harrisburg, Pennsylvania 17120

U.V. Henderson,  Associate Director
Environmental  Affairs  Research and
 Environmental Affairs 'epartment
Post Office Box 509
Beacon, New York 125U8

-------