>
I
39
rsa
•ij
rs>
       EPA-R2 72-067

       October 1972
                      Environment*
Technology
e

to
CD
An Improved


Manual Method for


Emission Measurement
3_

M*
Vt
CD
tt
(A
CD


CD

                               UJ
                               CD
                                Office of **v..


                                U.S. t»Vf-;>"


                                     i ^ I-
                                      .id Monitoring


                                       utection Agency

-------
                                                 EPA-R2-72-067
               AN  IMPROVED
MANUAL  METHOD  FOR  NOX

   EMISSION  MEASUREMENT
                          By
                   L.A. Dee, H.H. Martens,
                 C.I. Merrill, J.T. Nakamura,
                     and J. Martone

             Air Force  Rocket  Propulsion Laboratory
                   Director of Laboratories
                  Air  Force Systems Command
                   United States Air Force
                    Edwards, California
                 Project Officer:  F. C. Jaye

              Division of Chemistry and Physics
            National Environmental Research Center
             Research Triangle Park, N. C.  27711
                         Prepared for
                OFFICE OF RESEARCH AND MONITORING
               U.S. ENVIRONMENTAL PROTECTION AGENCY
                    WASHINGTON, D. C. 20460


                         October 1972

-------
                            EPA REVIEW NOTICE





This report has been reviewed by the 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 comnercial products constitute



endorsement or recommendation for use.
                                   11

-------
                              FOREWORD
     Accurate determination of nitrogen oxide (NO ) emissions from
                                                 X


stationary combustion sources is necessary if the Environmental Pro-



tection Agency and the various Regional Air Quality Basin Agencies



are to set realistic NO  limits.  Furthermore the sampling and analy-



sis technique must be able to pass legal scrutiny by possessing



unique technical attributes, should a challenge occur.



     The basic problems hampering the currently used NO  methods are
                                                       X


an inherently nonrepresentative sampling technique and an inaccurate



analysis method at the projected NO  emission limits.  These disadvan-
                                   X


tages are aggravated by the fact that the presence of chloride ion



also causes erroneously low results in the NO  analysis method.
                                             X


Chloride is present in many combustion sources.



     This report describes the development of a unique manual sampling



and analysis method for NO  from stationary sources.  The recommended



system has none of the previously cited deficiencies and in addition



provides for time integrated sampling and a rapid accurate analysis of



the resultant aqueous solution.  The acquisition of NO  data could also
                                                      X


be performed in the field without return to a laboratory if necessary.



This work was performed during FY 72 at the Air Force Rocket Propulsion



Laboratory under Project EPAOOOCX at the request of and funded by the



Environmental Protection Agency.  The AFRPL has an interest, in that it



too needs accurate information on nitrogen oxide emissions at its



facilities  because such compounds are both rocket propellants and



engine exhaust products.
                                  111

-------
     Mr. F. C. Jaye of the Chemistry and Physics Div. was the EPA

Project Engineer.

     The laboratory assistance of Mr. M. F. Citro and Mr. B. Dixon

is gratefully acknowledged.
This Technical Report has been Reviewed and is Approved.
PAUL J. DAILY, ULt Colonel,
Chief, Technology Division
AIR FORCE ROCKET PROPULSION LABORATORY
                                 IV

-------
                             Abstract







     The current manual NO  sampling and analysis method was evaluated.



Improved time-Integrated sampling and rapid analysis methods were



developed.  In the new method, the sample gas is drawn through a heated



bed of uniquely active, crystalline PbO_ where NO  is quantitatively
                                       <-•         X


absorbed.  Nitrate ion is later extracted with water and the concentra-



tion subsequently determined by a NO,, Selective Ion Electrode.  A



simple selective precipitation eliminates electrode interferences derived



from Pb09 absorption of other combustion products such as HC1, SO ,
        i*                                                        X


HF, and CO.  Field tests were conducted at various stationary source



sites and the data is presented herein.

-------
                          TABLE OF CONTENTS









SECTION                                                        PAGE






  I.          INTRODUCTION 	  1




 II.          TECHNICAL DISCUSSION	  5




                1.0  Sampling Methods	  5




                2.0  Analysis Methods	 16




                3.0  Laboratory Evaluation	36




                4.0  Field Tests	 48




                5.0  Theoretical Considerations	 64




III.          SUMMARY AND CONCLUSIONS	 70




                APPENDIX 1	 75




                APPENDIX II	 78




                APPENDIX III	 89




                APPENDIX IV	 93
                                  vn

-------
                        LIST OF ILLUSTRATIONS
FIGURE


 1


 2




 3


 4


 5


 6


 7  -


 8


 9


10
                                                       PAGE


X-Ray Diffraction Spectra of Commercial Pb09	11


X-Ray Diffraction Spectrum of PbO? Precipitated from

Pb(Ac)4	7	13


PDS Method Variability	19


Improved PDS Method Regression Line	20


NO Absorption vs. Time (3% Neutral HO)	23


NO and NO  Absorption vs Time (3% Neutral HO)	25


NO % Recovery vs NO Concentration (3% Neutral HO)	26
SIE Calibration Curve (1000 ppm PO,  ),
                                                       .35
NO  Analysis Scheme (Field Tests)	51
  X.


Field Test Sampling Apparatus	53
                           LIST OF TABLES
TABLE
                                                       PAGE
  II


 III


  IV


   V


  VI




 VII




VIII
         Diesel Engine NO  Data (Grab Samples Analyzed by
                         x
PDS Method)	V	6


NO Recovery Using the PbO  Tube with *> 200ppm NO/N	 8


NO Recovery Using the PbO  Tube with 256ppm NO/He	 9


NO Recovery Using the Zllppm NO/N  Source	.10
Crystal Form Composition of MCB and Fisher PbO .
                                                       .12
Recovery From A 211ppm NO/N_ Source Using Electro-
lytically Produced PbO ,
                                                       .14
Comparison Of NO Recovery Using "As Received" and Ball-

Milled MCB Pb02	15


Summary of PDS Variability Study	18
                                  viii

-------
                       LIST OF TABLES, Cont'd
TABLE

   IX


    X

   XI

  XII

 XIII

  XIV

   XV

  XVI


 XVII

XVIII

  XIX

   XX

  XXI

 XXII

XXIII

 XXIV


  XXV

 XXVI

XXVII

XXVIII


 XXIX

  XXX
                                                       PAGE
Comparison Between KOH and NH.OH Neutralization
(Neutral 3% H) ............ 7
                                                        21
NO Recovery ............................................. 22

NO Absorption vs Time ................................... 22

NO Recovery vs Concentration (16 Hour Standing Time).... 24

Comparative NO Results ............. . . ....... . ........... 28

Comparison Between SIE and PDS With 211ppm NO/N  Source. 2y

Calibration Of The Orion Liquid Membrane Electrode ...... 31

pH Of Various Solutions Used During The PbO /SIE
Analysis ............... . ....... . . ....................... 33

Analysis Results From Sampling Rate Studies ............. 35

Analysis Results From NO Concentration/Recovery Studies. 38
NO Determination in the Presence of SO,.
                                                       .40
NO Determination in the Presence of CO	42

NO Determination in the Presence of HC1	43

NO Determination in the Presence of HF	45

NO Determination in the Presence of C0_	45

NO Recovery After Minimizing the Effects of Interfering
Combustion Products	47

Data Matrix	49

Ppm NO From 211ppm NO/N  Source	50

Raw Data, Hercules HNO  Plant Field Test No. 1	55

Raw Data, Hercules HNO  Plant Field Test No. 2
(PpmNO )	56
      X

Raw Data, Moapa Power Plant Field Test (Ppm NO )	58
                                              X

Raw Data, Gas Fired Boiler Field Test	59

-------
                      LIST OF TABLES, Cont'd
TABLE                                                           PAGE



  XXXI  Raw Data, Diesel Aux. Power Generator (Ppm NO )	,. 61
                                                     X


 XXXII  NO .Field Test Data Summary (Ppm NO ,X ± s)	, . 62
          X                                X


XXXIII  "Paired t" Test Results at 95% Level	 63



 XXXIV  Thermochemical Data	 66

-------
                           I - INTRODUCTION



     Approximately sixty percent (60%) of the total United States'




emissions of oxides of nitrogen, NO (NO + N00) are derived from fossil
                                   X        Z



fuel burning stationary sources.  The quantities involved were estimated



to be about 9.6 x 10  tons in 1968 according to Esso Research and Engineer-



ing Co. (1).  The levels of NO  in stack gases are reported to range from
                              X



20 ppm (v/v) for small gas fired boilers to 1400 ppm (v/v) for coal fired



power plants (2).  Regulatory agencies on the federal and state or region-



al levels are projecting the establishment of successively lower NO
                                                                   X


emission limits.  Such lower limits will require improvements in the



accuracy and sensitivity of current sampling and analysis methods.  This




need for an improved manual sampling and analysis method for NO  became
                                                               X


urgent as a result of Congressional mandate to set stationary source per-



formance standards.  The current NO  method (3) suffers from numerous dis-



advantages, the most serious of which is a loss of accuracy in the 100 ppm



to 200 ppm range.  This single characteristic can have a deleterious effect



on possible future legal actions which may be initiated by enforcement



agencies.  At the request of the Environmental Protection Agency, and based



on recommendations and conclusions contained in the Walden Research Corp.



Report (2), members of the Rocket Propulsion Laboratory Chemical and



Materials Branch investigated and concluded that it was feasible to



develop an improved sampling device for stationary emissions coupled with



a less tedious and more accurate NO  analysis.
                                   X


     In the current method (3) an evacuated 2 liter glass flask is



attached to a gas sampling apparatus and filled in 0.1 to 0.2 minute




to the source pressure.  A small amount of absorbing solution oxidizes

-------
the nitric oxide (NO) to the dioxide (NO-) which in turn reacts with



the oxidizing solution to form nitrate ion (NO,), a process which takes



up to 16 hours.  This type of sample, however, represents an instant-



aneous point in time and unless the emitting source is constant, the



calculated total hourly or daily NO  emission can be in serious error.
                                   X


Relatively few stationary combustion sources are sufficiently constant



for the acquisition of an accurate grab sample.



     The sampling time can be lengthened by using a liquid absorbing



solution and slowly bubbling a known amount of gas through it, or by use



of a solid absorber/reactor which forms solid products with the NO + NO-



contained in the sample gas stream.  This latter possibility appeared



quite feasible as a result of the work of Mishmash and Meloan (4) who



presented evidence for the quantitative reaction of lead dioxide (PbO.)



with NO and NO- at reaction temperatures from 40°C to 190°C to give



solid lead nitrate (Pb(NO-)-).  The Pb(NO )- is quite soluble in aqueous



solutions when compared to lead compounds formed with other possible



combustion gases, such as sulfur dioxide (SO-), hydrogen chloride (HC1),



hydrogen fluoride (HF), and carbon monoxide (CO).  Carbon dioxide (C0_),



also a combustion product, does not react with Pb02 according to Pregl (5).



     Other approaches to time-Integrated sampling were investigated,



using liquids in conjunction with ozone, electrolysis, or chemical oxi-



dizers including charge transfer agents (6), but none of ,these effectively



modified the insolubility of NO in an aqueous system.  In addition, the



lack of quantitative conversion of NO to NO- in the presence of other



reactive stack gases was responsible for abandoning liquid systems as



sampling devices.  It should be recalled that the NO  in combustion stack
                                                    X

-------
gases is generally 95  % NO (2) even with air in slight excess.



     A preliminary experiment with NO confirmed the quantitative



absorption with Pb02 as reported by Mishmash (4) and also showed that



N0_ was released when the PbO~ was washed with distilled water.  Varia-



bility in Pb02 reactivity to NO was encountered but was found to be due



to sources of PbO-, or more properly its physical state, as sold by



various vendors.  The literature (7-9) revealed much controversy over



"capacity" or "reactivity" with NO or NO- during some 60 years prior



to 1969.  It appears that of the four distinct forms of Pb02 only the



crystalline^- and  p -Pb02, or a mixture of the two, will yield quanti-



tative recovery with NO.



     The determination of NO- in aqueous solutions is well known (3, 10,



11) and efforts to improve the accuracy were carried forward concurrently



with the development of the Pb02 sampling device.  The difficulties con-



sisted primarily in minimizing or eliminating the anticipated interferences



to the phenoldisulfonic acid (PDS) and the NO. selective ion electrode (SIE)



methods.  For example, a coal burning power plant will emit HC1 and CO
which
      react with Pb02 to yield PbCl2 and PbCO».  The presence of Cl  or


  2-                                              -
CO _ in solution will cause inaccuracies in the NO,, response of the SIE.



Similarly Cl  will interfere with the PDS method (12) and cause erroneously



low results.  It was discovered that the SIE was even sensitive to changes



in pH (hydrogen ion concentration) and was utterly useless in. solutions



containing peroxide (H_0™).



     Nevertheless, the work described in the following sections resulted



in the development of a time-integrated manual sampling device coupled



with two accurate analysis methods for N0~.  The system was field tested

-------
and appears satisfactory for NO  emission measurement of stationary
                               X



sources above 20 ppm NO .  Its applicability to mobile sources and the
                       2t



measurement of ambient air NO  levels also appears possible.

-------
                      II  TECHNICAL DISCUSSION



1.0  Sampling Methods



     1.1  Flask Grab Sample Method



          The most advantageous feature of the grab sampling method is



that the only portable equipment required for its use is an evacuated



and calibrated vessel containing a measured volume of gas absorption



reagent.



          To collect a sample, the flask is connected to a sample port



and the flask closure is opened to admit the gas sample.  The operation



can take as little as ten seconds to accomplish.  Following collection,



the sample is returned to the laboratory for a final sample pressure



measurement and analysis of the absorbing reagent in the flask for the



specie of interest.  The NO  method described in the Federal Register (3)
                           X


is a modified version of the grab sampling technique.  The exceptions



are that the recommended sampling apparatus includes a vacuum pump for



evacuating the flask and a manometer for sampling pressure measurement.



     The major disadvantage of the grab sampling technique is that the



indicated emission level from any single sample can only represent the



value for that brief period of time required to take the sample.   Since



most emission levels fluctuate due to combustion variations or stack gas



turbulence, many grab samples must be taken and the results averaged for



accurate daily emission data.  Typical NO  emission data from a stationary
                                         X


diesel engine is shown on Table I.  The ten grab samples were randomly



spaced during a two-hour sampling period, and the analyses were performed



using the PDS method described in sections 2.2 and 3.3.

-------
     Table I, Diesel Engine NO  Data Grab Samples Analyzed by PDS Method.
                              X
Sample No.
1-3
1-5
2-1
2-2
3-1
NO (ppm)
X
353
318
454
353
398
Sample No.
3-5
4-3
4-5
5-1
5-2
NO (ppm)
X
382
364
339
411
346
           range • 318 to 454 ppm = 136 ppm NO
                                              X


     Even though this engine was operating at constant load, it is



apparent that a prohibitively large number of grab samples must be taken



to accurately show the twenty-four hour average NO  emission level of
                                                  X


this combustion source.  For other combustion sources where the fuel



feed rate is less carefully controlled, the emission level can be expected



to vary even more widely.



     1.2  Flask Extended Grab Sample Method



          To integrate or average NO  emission level variations the gas
                                    X


sampling time can be extended by installing a flow restrictor between the



flask and the sample point.  This idea was investigated using a labora-



tory NO  source (^ 200 ppm NO in N ).  The restrictor was constructed of
       X                          ^


glass and provided an extended fill time of approximately 12 minutes for



the 2 liter flask.  No detectable NO  analysis variation occurred which



could be attributed to the use of this extended grab sample technique.



Although the method showed promise when laboratory NO sources were sampled,



it can be readily seen that this method is doomed to failure for most



combustion sources.  Because the restrictor must be located between the



sample point and the 2-H flask and also must be of very small diameter

-------
 (i.e., ^ 0.002 inch) particulate matter or even water mist can plug
 the opening and thus preclude obtaining a representative sample.
     1.3  Time-Integrated Sample Method
          A more reliable means of obtaining a daily emission level for
 NO  is to collect a sample of the combustion gas at a constant rate for
  X
 some finite fraction of the emission period of interest.  Samples col-
 lected by this method, when analyzed, can be related directly to the
 average emission level.  The most important prerequisite for this sampling
 technique is that the collection medium used will quantitatively trap the
 species of interest. Thermodynamic studies (1) of combustion systems which
use air as the oxidizer have shown that the primary NO  species formed is
                                                      X
nitric oxide (NO).  NO is a relatively inert oxide of nitrogen with only
 limited solubility in liquids.  Collection of NO in gas streams using
aqueous scrubber systems has been an almost total failure.  Several at-
 tempts to absorb NO in aqueous media were made during this program.
          1.3.1  Gas Phase Pre-oxidation of NO.
                 Aqueous alkaline o-methoxyphenol has been reported to
quantitatively absorb NO  (6) and therefore if oxidation of NO to NO
 can be accomplished readily then absorption in this medium is possible.
Attempts to oxidize NO with 0« and U.V. light or with 0  resulted in
either no reaction at all or in the decomposition of the o-methoxyphenol.
In a third attempt to enhance the oxidative process, o-methoxyphenol was
prepared in alkaline 10% HO  solution.  Use of the absorption medium re-
sulted in NO recovery which was only 30% of theoretical.
          1.3.2  Electrolytic Oxidation of NO.
                 Anodic oxidation of NO to NO- appeared to be another

-------
means of time-integrated collection.  A device was constructed in such



a manner that the gas containing NO would directly contact a positively



charged porous screen immersed in an aqueous alkaline solution.  Later



the voltage was increased in order to generate oxygen at the anodic



absorption surface.  Neither system resulted in greater than 30% re-



covery of NO.



          1.3.3  Solid Absorbants



                 The Walden Research Corp. reports (2) describe several



potential solid sorbents for NO  and SO  such as Mn00, K0Cr00  and PbO
                               x       x            L   L  2. 7        2


which may be used in time-integrated samplers.  Of these solids, Pb09



appeared to be the most promising since it has a long history of success-



ful application as an N09 absorbent in the classical Pregl combustion



train.  Initial NO recovery tests using Pb09 were promising.  PbO , which
                                           £•                     £f


had been purchased from Fisher Chemical Co. approximately ten years ago,



was packed into glass 1/4 in. O.D. x 12 in. tubes and the ends were



plugged with glass wool.  The Fisher product was certified to be prepared



according to Pregl (5) which means that the PbO  was digested in concen-



trated HNO_, washed with water until NO  free, dried, sieved, and the
          J                            J


12/20 mesh particles retained.  Table II shows the NO recovery results



from PbO  tubes when an unverified 200 ppm NO source was sampled.



     Table II, NO Recovery Using the PbO  Tube with ^ 200 ppm NO/N



                      Test No.            ppm NO (found)



                         1                     178



                         2                     220



                         3                     145



                         4                     135



                                      X  =     170
                                   8

-------
The Table II analysis results show an unacceptable degree of scatter




and seem to contain a negative bias of ^ 15%.  Analysis was performed




for N0~ using the Orion Nitrate Selective Ion Electrode (SIE) in the




aqueous PbO. slurry.  Minor refinement of the method resulted in somewhat




better results shown in Table III when a verified 256 ppm NO/He source




(see section 2.1) was used.




     Table III, NO Recovery Using the PbO- Tube with 256 ppm NO/He




                  Test No.        ppm NO (found)
1
2
3
4
5
6
7
245
234
236
216
229
238
241
                                 X =   234




     The scatter and negative bias (^ 9%) are significantly less than




the earlier results.  The PbO» tube sampling method looked very promising




and a detailed investigation of the method followed.  A second batch of




PbO_ (prepared according to Pregl) was purchased from Matheson Coleman




and Bell (MCB) to conduct the more detailed study.  We observed that the




average particle size of this PbO~ was somewhat greater than that of the




Fisher's but both were used in the "as received" condition.  A large




quantity of 211 ppm NO/N,, was prepared to eliminate source variability from




the sum of the analytical errors.  Table IV shows the comparative results




obtained with the modified PDS method and flask (see section 2.2) and the

-------
MCB Pb02 using the 211 ppm NO source.
     Table IV, NO Recovery Using the 211 ppmNO/N  Source



                    Test No.         ppm NO (found)



                            PDS( flask)        MCB PbO



                      1        197               25



                      2        197               23



                      3        197               56



                      4        197               29



The extremely low recovery indicated with this PbO? was surprising since



both sources (MCB and Fisher) were specially prepared for use in the



Pregl combustion train.  Mishmash and Meloan (4) reported that Pb(NO.,)



was formed when NO reacts with PbO .  Therefore in order to determine



if the capacity of the coarser mesh MCB PbO  has been exceeded, x-ray



diffraction analyses were performed on the inlet and exit sides of a



used MCB PbO  tube.  Traces of Pb(NO_)  were detected at the exit end
            •^                       -J i,


of the tube thus indicating that the capacity had been exceeded.  In



addition, a used Fisher PbO  tube was subjected to the same analysis,



since slightly low recovery is also indicated.  No Pb(NO,.)  was detected
                                                        •J £


at the exit end of the tube.  However significant differences between the



PbO  spectra were observed.  Figure 1 represents the x-ray diffraction



spectra of Fisher and MCB PbO .  According to x-ray diffraction data



(15) PbO  can exist in three crystal forms .    Table V depicts the



crystal form composition of the two batches of PbO^ (assuming that no



amorphous materials are present and that the sensitivities of the three



forms are equal).
                                 10

-------
                         RELATIVE INTENSITY
  i  i  I  I  i  i  I T
0
                              11

-------
Table V, Crystal Form Composition of MCB & Fisher Pb02
              % x-Pb02         %
-------
                         RELATIVE   INTENSITY
CJ1
NO
oo
CO
OT
CO
NO
NO
00
NO
                                                                          CTQ
NO
NO
                                      13

-------
     Table VI, Recovery From a 211 ppm NO/N2 Source Using Electrolyti-




cally Produced PbO_









Sample No.                ppm NO Recovered




                  Non-acid washed       Acid washed




    1                   69                 217




    2                   80                 217




    3                   70                 194




    4                   69                 194




    5                   65                 212




    6                   70                 198




    7                   65                 198




    8                   64                 198
           Average      69                 205




% Recovered             33                  97




* Prepared according to Pregl.




    The low recovery illustrated by the non-acid washed PbO~ data




demonstrates the need to remove oxidizable material from the PbO  such




as elemental lead which can lead to formation of reduced nitrogen species,




i.e.,       2Pb + 2ND •* 2PbO + N .




    Since Cropper (8) demonstrated that surface area plays a role in the




NO- capacity of PbO~ the MCB product was ball-milled for three days and




retested with the 211 ppm NO/N- source.  Table VII shows those results.
                                     14

-------
      Table VII, Comparison of NO Recovery Using "As Received" and




Ball-Milled MCB PbCL








Sample No.              ppm NO Recovered

1
2
3
4

•ecover
"As Received"
25
23
56
27
X 33
ed 16
Ball-Milled
78
72
72
72
75
35
      Apparently the source (crystallinity) of the PbO  is a more




important factor than surface area.  Further refinement of the PbO?/SIE




method using electrolytically derived PbO~ is discussed in the following




sections.
                                  15

-------
2.0  Analysis Methods




     2.1  Laboratory Standard NO Sources.




          The necessity for obtaining or preparing verified sources of




NO for accomplishing Phase II of this program is recognized.




Therefore a lecture bottle of Matheson "Analytical Reagent Grade" NO




was analyzed on  a  CEC/DuPont 21-490 mass spectrometer.   The  only im-




purity found was less than 1% N2, and no N02 was detected. An NO source





in He was prepared using pressure measurements, which calculated as 256




ppm NO (v/v).  This dilute source was verified by gas chromatographic




(GC) analysis, after the GC was calibrated with the pure NO source using




an exponential dilution flask.  The average of seven determinations was




251 + 6ppm NO for the GC method.  The standard deviation  (6ppm NO) for the




GC analysis shows no significant difference between it and the calculated




256 ppm NO.  Both a 1% NO/N? and a 1% NO-/N,., source were prepared  (verified




by mass spectrometer) from which a 200 ppm source of each was obtained by




dilution.  The 211 ppm NO/N  source was prepared in a large volume "K"




bottle directly from the pure NO source by means of pressure measurements.




The use of pressure measurements was shown to be sufficiently accurate




for standard preparation.  The large volume assured the use of a single




source for the entire program.




     2.2 Phenoldisulfonic Acid (PDS) Method Improvement




         Initially, the published PDS method (3) was used for the  deter-




mination of N0~ derived from flask samples of the verified NO laboratory




source during the Phase II laboratory studies.  It was necessary to first




obtain a calibration curve using NaNO., standard solutions  (250 yg/ml)
                                   16

-------
added to 20-25 ml of the 0.1 N sulfuric acid/HO  solution.  The




evaporation was accomplished in beakers as well as evaporating dishes,




after carefully neutralizing the H«SO,.




     A large random variability was noticed in the color intensity of




the yellow nitro-PDS product, as well as undissolved silica turbidity




from time to time.  It was further noticed that when the acid H,jO? was





excluded from the NaNO« standards an acceptable calibration curve was




possible in the range from 1.0 to 5.0 pg NCL/ml or 100 to 500 yg N0_ per




100 ml vol. flask.  When the turbidity was present low values were re-




corded even after the yellow solution was filtered prior to measurement




in the 1 cm cell.  During the initial analysis studies we were unable to




obtain quantitative NO recovery when the 2 liter flask was filled with




the known laboratory NO source.  At this point it was decided to investi-




gate the causes of the difficulties in an attempt to improve the precision




and the accuracy of the PDS method.




          2.2.1  Evaporation Step




                 A literature survey (10, 12-14) revealed that much work




had been accomplished in investigating sources of errors in the PDS method




for the determination of NO  in potable water.  Chamot, et al, (12-14)




rarely mentioned encountering solids, or turbidity, following the nitration




step, because only little NaOH is necessary for neutralization of potable




water prior to the evaporation step.  Thus, the etching (dissolution) of




glass (silica) which produces turbidity was minimized.  Therefore, as an




initial step, it was decided to use 25 ml platinum (Ft) crucibles and




only a 4 to 6 ml aliquot of the acidic H^O  instead of evaporating the




entire 25 ml as mandated (3).
                                17

-------
     The variability study, using this procedure, consisted of ten




replicates each of four standard NO  concentrations equivalent to




1, 2, 3 and 4 yg NO /ml.  The standards were added to 3ml of a 0.1N




H2S04/3% H202 solution plus 1ml of IN NaOH for neutralization.  The




evaporated residue was treated with 2ml of PDS and allowed to stand for




2 to 3 minutes prior to dilution.  Some gas evolution was noted, but no




turbidity was observed.  Color intensity was measured at 405nm on a Gary




Model 14 Spectrophotometer.  The results are shown in Figures 3 and 4




and Table VIII:




     Table VIII, Summary of PDS Variability Study




Mg NO /ml (x)                     av. absorbance







      1




      2




      3




      4




s = standard deviation




y = average absorbance




A regression equation was derived from the above data and was found to




be y = 0.15x-0.003 where x = yg NO /ml.  The slope of the calibration




curve corresponds to about 0.15 absorbence per jug N07/ml if a 1 cm




cell is used.




          2.2.2  Neutral H 02




                 The variability, however, was still not satisfactory;




therefore, it was postulated that the gas evolution indicated earlier




was not only C09 escaping but also HNO  from the anhydrous PDS solution.




To test this hypothesis and to determine if a KOH instead of NH.OH final




                                 18

-------
                                     ABSORBANCE
                  ISJ
                                                   CJ1
                                                            ISJ
                                                            CO
                                                                                OP


                                                                                 CD

                                                                                 CO
 o
N5
                                         19

-------
                                 ABSORBANCE
                                                 p
                                                 CO
o
en
    rsj
OP
                                       20

-------
neutralization  (12) yields a more intense color, neutral 3% HO  was




substituted for the acidic peroxide, and both neutralization techniques




were used.  The results of this study using a concentration of 2.5ug N09/ml




are summarized in Table IX.




     Table IX, Comparison Between KOH and NH.OH Neutralization (neutral



o
-------
are given in Table X.

                         Table X, NO Recovery


                         ppm NO (v/v)               ppm NO (v/v)
     Sample No.          after 16 hrs               after 1 wk

         1                   236                       218

         2                   236                       218

         3                   225                       210

         4                   245                       236

         X                   235                       221

X = average of each set.

It is apparent that some NO, adsorption on glass occurs with standing.

This was again observed when six flasks were filled with the 256  ppm

NO/He standard and analyzed sequentially with time (hrs) in order to

determine the absorption time for the NO by the neutral 3% H^O^.   The

results are shown graphically in Figure 5 and Table XI.

                      Table XI, NO Absorption vs Time

                      Time                    ppm NO

                        1 hour                  10

                        3 hours                 31

                        7 hours                222

                       24 hours                240

                       30 hours                230

                      100 hours                230

It can be seen from Figure 5 and the above that the minimum of 16 hours

standing, as required by Method 7, is necessary before opening the flask

for the removal of a 5 ml aliquot.  Conversely, maximum absorption of
                                     22

-------
                             ppm NO
so
CO
                                                                 oo
                                                                 O
                                                                    cn
                                                                CO
                                                               10
                                   23

-------
NO  at the 200 ppm level was observed to take place in 3 hrs, see




Figure 6.  In a further experiment it was hoped to determine the




lowest NO level for which the 2-liter flask could be effectively




used.  At low NO levels the initial reaction of NO and 0  to give




NO  in the vapor space of the flask may proceed slowly enough so that




the 16 hour standing time is insufficient.  Therefore, four replicates




each of five concentrations of NO/N- over the range of 50 to 500 ppm NO




(v/v) were admitted to evacuated flasks.  The NO levels were determined




as before and the results are shown graphically in Figure 7 and summar-




ized in Table XII.  It can be seen that the recovery decreases rapidly




below 150 ppm NO if the flasks are allowed to stand only 16 hours.




     Table XII, NO Recovery vs Concentration (16 Hours Standing Time)




                               ppm NO




Calculated                    Recovered                   % Recovery




    58                            41                          71




   105                            84                          80




   199                           184                          92




   211                           197                          93




   540                           482                          88




Several conclusions can be drawn from the preceding recovery data:




     (a) the rate controlling step in conversion of gaseous NO to




aqueous NO  is the gas phase oxidation of NO to N0? and this oxidation




rate  is  inversely proportional  to  the NO  concentration.




     (b) for samples which contain low concentrations  (<150ppm) of NO,




a compromise is necessary between sufficient NO oxidation time and NO.





losses in the glass flask,




                                    24

-------
                                               ppm
ISJ
CD
£
O5
CD
                                  CO
                                  O
—*     IS5    IS3
oo     CD    ro
O     CD    CD
N5
CJ1
TO
OO
    IS3
    CD
   CO
   CD
                                                   25

-------
                %  RECOVERED
                             oo
CJ1
                    CD   0
^   S^ m
                 CD

                         O
                    oa    m
                 O
                ro
                                                          CD
                                                      CO
                                                     to
                          26

-------
     (c) a larger flask will likely not reduce the NO losses at low




levels because of the diffusion controlled gas phase and absorption




reactions.




          2.2.4  PbO  Nitrate Analysis by PDS




                 At about the time that improvements in the accuracy




and precision of the PDS method were accomplished and incorporated




into a "Suggested Amendment" to Method 7 (3), the PbO  sampling




"variability" problems had also been solved (see Section 1.3).   Thus




it seemed desirable from a statistical point of view to be able to




determine the NO  content of the centrifuged aqueous solution obtained




from the PbO  tubes by both SIE and PDS.




     The ten PbO_ tube samples obtained during the Hercules Field




Test No. 1 (see Section 4) presented the first opportunity for ob-




taining such comparative data.  The Table XIII results, shown below,




indicate that the methods agree within 10% most of the time.
                                    27

-------
                Table XIII,  Comparative NO  Results
                                          X


                           ppm NO  (v/v)
    Sample #                    SIE                  PDS



      1-1                       450                  570



      1-2                       640                  620



      1-3                       660                  600



      2-1                       560                  550



      2-2                       620                  580



      2-3                       590                  620



      2-4                       890                  910



      2-5                       640                  690



      2-6                       530                  580



      2-7                       540                  580



The data also revealed several problem areas which may account for the



observed random differences.  A pH dependence (see section 2.3.3)  was



observed to be responsible for poor precision in the PbO_/SIE results.

                                            3-
This was corrected by adding a phosphate (PO,) buffer to  the extract,



thus stabilizing the pH at about 11.  Five replicates of  the 211 ppm



NO/N- source using PKL tubes were analyzed for NO by both SIE and PDS.



The results shown in Table XIV indicate an improved precision.
                                   28

-------
Table XIV, Comparison Between SIE and PDS with 211 ppm NO/N2 Source



      Test #                          ppm NO (v/v)



                                SIE                  PDS



        1                       216                  215



        2                       204                  201



        3                       206                  208



        4                       204                  191



        5                       208



The results averaged 208 and 204 ppm NO for the SIE and the PDS methods.



At this point it must be emphasized that the acceptable accuracy and



precision of the two analysis methods are "tailored" for stationary



sources from which other reactive gases are absent.  These would be HC1,



HF, CO and SO-, all of which are assumed to react at 180 C with the



highly active PbO_ (see section 5.1).  A more comprehensive development



program was undertaken which will be described in Section 3.0.



      2.3  The N0~ Selective Ion Electrode (SIE)



           2.3.1  Description


                                                          -4      -6
                  The usual determination of trace NO- (10   to 10  M)



in aqueous solutions has been accomplished for many years by colorimetry,



of which the PDS method is only one example.  Other techniques such as



polarography have been used more recently.  Since the advent of selective



ion electrodes (SIE) an acceptably accurate result can be obtained in



minimum time and by any operator.  As with other methods, some knowledge



of other ions present is necessary in order to minimize their inter-



ference.  This aspect and the operation of such an electrode system is



reviewed by Durst (18).  The SIE data contained herein was obtained with
                                  29

-------
an Orion liquid membrane nitrate electrode, a Model 801 "lonalyzer",




and a Model 605 electrode switching unit.   A Ag/AgCl reference




electrode was used to develop the AV in mv.  With the Model 605




electrode switch it is possible to obtain the N0~, Cl , and F  con-




centrations, if desired, as well as the pH of a sample with time savings




and convenience.  While the NO., SIE has a limited lifetime, the restora-




tion is relatively simple, and consists of replacing the membrane and




the internal filling solutions.




            2.3.2  Evaluation




                   The electrode system was calibrated seven times




in eight days with standard solutions of NaNO,.  This should show what




variability might be expected when the 100 ppm NO. response is held




constant.  This calibration is necessary in order to define that range




of concentrations where the least error occurs.  Table XV shows the




results and it appears that the range from 50 ppm NO, to 500 ppm N0_




is most useful.
                                  30

-------
Table XV, Calibration of the Orion Liquid Membrane Nitrate Electrode
N0~ CONG Electrode Response (mv.)
(ppm) 15 Nov 16 Nov 16 Nov 17 Nov 17 Nov 19 Nov 22 Nov
10
20
50
*
100
500
1000
10,000
197.3
180.0
151.0
130.1
85.1
62.7
7.5
198.0
180.5
152.5
130.1
84.0
63.0
5.0
202.3
177.1
149.7
130.1
85.9
63.1
3.0
197.6
180.8
146.5
129.9
82.8
66.2
5.5
197.6
178.6
156.8
130.0
84.5
62.9
4.6
191.2
177.3
153.2
130.0
87.2
65.5
7.0
192.5
176.5
142.6
130.1
82.6
60.5
-4.5
*100 ppm N0~ electrode response adjusted to 130.0 ± 0.1 mv prior to




each run.




     A linear regression analysis was performed on the electrode response




data in order to determine the calibration curve and also to obtain an




indication of the precision of calibration.  Since the electrode response




varies logarithmically with respect to the NO- concentration, the log of




the NO, concentration was used for the regression analysis.




      Results of Least Squares Analysis of N0_ Electrode Data





      a.  Best fit line is:




          y =-1.536    + 4.101




          where:




          y = log1Q (N03 cone, (ppm) )




          and




          x = electrode response (mV/100)
                                  31

-------
      b.  Standard error in y.,



          *S- - 0.0515 - 1.1 ppm N0~



      c.  Standard error of intercept,



          (x = 0) - 0.0153 = 1.0 ppm N0~



All laboratory SIE data through the first (preliminary)  Hercules nitric



acid plant field test were obtained with the calibration techniques



described.  Some variation due to ambient temperature fluctuation was



noted, since the electrode is subject to the Nernst equation which contains



a temperature term:


                tr _L 2.3RT   .
          E   = K + _. i, —  log a
           mv                 e
      where a = activity



          K = constant



         2L = ionic charge


          RT                                         A
      and — TT = 59.16 mv for singly charged ions at 25 C.



          2.3.3  Sources of Error



                 As was mentioned in Section 2.2 the first Hercules



results  (Section 4.0) indicated some unexplained drifting in the electrode



response.  Some drift was observed even when Pb(NO»)2 standards were used



for calibration (they more closely resemble the aqueous composition of



the samples) .  The somewhat erratic results when compared with the PDS



analysis of the same samples were thought to be due to this drifting



response.  In order to obtain better correlation between the two methods,



another N0« calibration curve was prepared using conditions which more



nearly duplicate the actual sample.  PbO~ was added to each Pb (N0,)»



standard solution (10 ml Pb (N0_)2 solution/3 g PbO~) .  These mixtures
                                 32

-------
were treated in exactly the same manner as an actual sample and a


calibration curve was prepared.  These solutions gave rise to a large


signal drift similar to that encountered initially with the Hercules


samples.  A series of analyses of the 211 ppm NO/N« source using the PbO~


sampling technique and this SIE calibration technique gave results which


averaged 165 ppm NO.  Grab samples and PDS analyses of the same source


gave an average analysis of 205 ppm.


      These results indicated that further refinement of the PbO_/SIE


system was necessary.  A brief study showed that the SIE is pH sensitive


to the extent of 5 mv per pH unit in the range pH 3-6.  A study of the


pH of several solutions encountered during the analyses gave the following


results:


Table XVI, pH of Various Solutions Used During the Pb02/SIE Analysis


                    Solution                   2JH


          distilled H-0 + NaNO                 6.5


          distilled H20 + Pb(N03)2              5


          distilled H_0 used to rinse PKL     3.2


      It can be seen from Table XVI that the NaNO- solution is slightly


greater than pH 6.  The Pb(NO )- solution is less than pH 6 and within


the pH dependent region.   The distilled H~0 used to rinse PbO? was even


more acidic.  The literature revealed that H~0 and PbO~ will react to a


small extent as shown:


                    Pb02 + H20 -> H+ + HPbO~


      It became apparent that some means to control the pH of the


solution prior to the SIE measurement would be required.  A 1000 ppm

  3_
PO,  buffer solution added to the Pb(NO_) standard solutions and to the


                                 33

-------
PbO? for N0_ extraction was sufficient to give a pH of 11 and also


solved the drift problem that was initially encountered with the

                          2-
Hercules samples.  The HPO.  ion formed,  reportedly constitutes a minimal


interference to NCL determinations (18).   Figure 8 is a typical calibration


curve using Pb(NO_)_ standards which were prepared in 1000 ppm PO,  buffer.

                  3_
The addition of PO,  apparently has no deleterious effect on the SIE


calibration.


      2.4  Final Analysis Methods Development


           The PDS and SIE methods developed to this point have not been


tested with simulated stack gas samples containing reactive or inter-


fering gases such as HC1, HF, SO- and CO.  Most of the later method changes


were responses to deviations in results observed during the laboratory


studies with HC1, HF, CO, and S0« in the presence of the NO stream.  Details


of the changes are described in Section 3.3 and Figure 9.
                                  34

-------
                                          to
                                               T3
                                               •o

                                                3
    CJl
3   co
    ro
                                                                      I
    C=) •	
                                                                                        r%  
-------
3.0  Laboratory Evaluation of Candidate Methods



    3.1  NO recovery vs sampling rate.  The 211 ppm NO/N_ source was



used throughout this series of tests and in all cases the Pb(NO_)9 was
                                                               J ^


extracted at ambient temperature with 10.0 ml of 1000 ppm PO ~ buffer.



Table XVII shows the analysis results using both the N0~ Selective



Ion Electrode (SIE) and the Phenol Disulfonic Acid (PDS) methods.



      TABLE XVII, Analysis Results From Sampling Rate Studies



    Test No.       Flow Rate (ml/min)  Volume Sampled (1)   Ppm NO Found



                                                            	      PDS



       1                  18



       2                  41



       3                  68                  3.94          199      214



       4                 135                  3.58          204      196



       5                 253                  3.80          216      204



       6                 256                  3.84          223      202



       7                 513                  4.10          212      190



       8                 513                  4.10          203      196



                                                                     200



                                                                     8.3



                                                                     -11



All samples were collected at 190 C except test 3 which was collected at



room temperature (^ 25 C).  This result indicates that NO is quantitatively



collected and oxidized to N0_ in the absence of gaseous 0^ even at room



temperature.  Test 1 represents an NO capacity test for the 4 mm I.D. tube

17.8
2.85
3.94
3.58
3.80
3.84
4.10
4.10



SIE
191
212
199
204
216
223
212
203
X = 208
s = 10.2
X - 211 - -3
                                  36

-------
 (2-4 g Pb02) that was described in Section 2.3.  The fact that the


                                                                     2+
 cannot be quickly extracted from the PbO. when large quantities of Pb



 salts are present unless the PbO./extract slurry is heated for a short



 period of time may explain the slightly low result.  This result does



 demonstrate that the capacity of the PbO_ tube is much greater than the



 required amount of NO, dictated by either analysis method.



    3.2  NO recovery vs NO concentration.  A 1.0% NO/N2 mixture was used



 to prepare various NO/N» mixtures by dynamic flow dilution.  The analysis



 results are presented in Table XVIII.  Tests designated with "*" indicate



 that the sample was absorbed at ambient temperature instead of 190 C.



 Table XVIII illustrates that the PbO~ sampling method can be expected



 to generally yield results that are accurate to within + 5% of the actual



 NO value with either analysis method (SIE or PDS).  The ambient tempera-



 ture reactivity of electrolytically derived PbO« with NO greatly enhances



 its versatility as a sampling device.  Almost any heating device can



 now be used because the tube needs only to be maintained at some temp-



 erature above the dew point of the sample gas.



    3.3  NO recovery in the presence of other combustion products.



 Small flow rates of the pure combustion products (i.e., S0_, HC1, ...,



 etc.) were diluted with the 211 ppm NO/N» source in a dynamic flow



dilution system.  This gas sample preparation method allows independent



variation of the combustion product level without significantly affect-



 ing the NO concentration.  Each combustion product was evaluated inde-



 pendently with the 211 ppm NO/N» source.
                                 37

-------
TABLE XVIII, Analysis Results  From NO Concentration/Recovery  Studies
Test No.     Ppm NO (theo.)   Volume Sampled  (liters)    Ppm  NO (Found)
                                                       SIE        PDS
    1          64                 10.49                 51         55
    2          64                 10.54                 60         65
    3*         64                 10.58                 61         65
    4         101                  8.92                102        101
    5         101                  7.68                105        116
    6*        101                 10.08                104         96
    7         211                  4.37                203        228
    8         211                  4.31                210        220
    9*        211                  4.78                228        247
   10         420                  3.12                426        421
   11         420                  3.07                446        427
   12*        420                  3.21                428        409
   13         664                  2.42                633        656
   14         664                  2.32                624        673
   15         664                  2.37                609        624
   16         900                  1.44                880        897
   17         900                  1.66                863        891
   18*        900                  1.49                836        855
   Linear regression analyses yielded the following data:
   a.  SIE:  x = 10.3 + 0.945y; std error 19.1 ppm, std error of
                                intercept 7.3 ppm, std error of slope 0.015
   b.  PDS:  x = 10.2 + 0.969y; std error 16.6 ppm, std error of
                                intercept 6.4 ppm, std error of slope 0.013
             where x = theo and y = found
                                     38

-------
          3.3.1  S0?.  Data contained in the Walden Research Corporation

Report, Part 1, "Sulfur Oxides", demonstrate that SO,, quantitatively

reacts with Pb00.  Therefore, since PbSO  is an insoluble salt it seemed
               2                        4
that SO- would not interfere with the NO reaction or analysis except

through competition for oxidative sites on the PbO-.  Table XIX shows

that this assumption was not valid and also demonstrates that NO can be

determined in the presence of a ten-fold excess of SO..  The PbO_ was
                 3_
extracted with PO,  buffer, centrifuged, and the supernatant liquid

analyzed for NO- in tests 1-3 of Table XIX.  The data show an abnormal

amount of scatter (PDS vs SIE) and indicate that something may be inter-

fering with the SIE analysis.  Since Pb«(PO,)- is much less soluble than
                                3-
PbSO, it is probable that the PO,  contained in the aqueous extraction
                           2-
liquid is displacing the SO,  in the PbSO, as follows:

           3PbSO, + 2 P0,~ ->• Pb (P0,)9 + 3SO ~

To test this hypothesis, cold water was used to extract the Pb(NO_)-

in tests 4a - 6a and the aqueous extract was decanted from the Pb07

prior to adding sufficient buffer to make a rinal concentration of

1000 ppm PO, .  The low results indicate that the Pb(NO_)- either was

not extracted quantitatively or had not been formed at the expected level,
      3_
The PO,  buffered extracts were returned to the appropriate PbO- samples,

mixed, and allowed to stand overnight.  These SIE results (tests 4b - 6b)
                                                      2-
are much too high, thus indicating a high degree of SO,  interference.

In further tests (7 - 9) the sample and aqueous extract volumes were

decreased in view of the possibility that the PbO- capacity may have

been exceeded.
                                   39

-------
          TABLE XIX, NO Determination in the Presence of SO,
Test No.
Treatment   Ppm NO/Ppm S02 (theo)  Volume Sampled
1
2
3
4a
5a
6a
4b

5b
6b
7

8
9
10
11
12
13
14
15
16

17
18
j—
it
it
Cold H20
it
it
P03~
4
it
"
Cold H00
2
"
it
Hot H20
it
ii
it
ii
it
Cold H00
2
ii
it
211/1900
it
"
211/2200
it
ii
it

11
ti
it

it
"
211/2200
it
"
211/1080
it
"
n

it
it
.me Sampled
liters)
3.06
2.98
2.99
2.88
3.13
2.88
	
	
	
1.94
1.94
2.02
1.94
1.92
1.93
1.94
1.92
1.93
2.12
2.41
2.06
Ppm NO
SIE
223
245
229
191
187
182
277
266
275
192
184
183
227
223
227
211
211
222
204
193
195
(Found)
PDS
197
195
228
170
169
166
	
	
	
186
174
195
187
193
191
211
217
217
222
216
216
No corresponding increase in the indicated NO concentration is apparent.




Therefore, if the NO is being quantitatively trapped by PbO. it is then




not being extracted completely by the water during the brief contact




time (ca. 20 minutes).  To increase the Pb(NO,)- extraction rate, the




Pb02/water slurries were heated in boiling water with occasional shaking




for 30 minutes in tests 10-15.  These samples were then cooled in an




ice bath and centrifuged prior to decanting the extracts.  After adding
                                   40

-------
                            3-               -
the appropriate amount of PO,   buffer, the NO- concentrations were


determined.  Acceptable results were obtained (ca.  + 5% error probably

         2_
due to SO,   interference) thus demonstrating that the presence of SO-


does not preclude accurate NO determination.  To determine if the same


extraction difficulty exists at the 5:1 excess SO-  level,  cool (25 C)


water was again used to extract the N0_ and the SIE measurement was made

                                         3_
after adding the appropriate amount of PO,   (tests  16-18).  The buffered


extracts were then returned to the PbO~ and the slurries were allowed to


stand for 20 minutes after shaking.  The corresponding PDS analysis re-


sults indicate that heating the aqueous extract is  necessary even at the


5:1 excess SO- level, probably due to co-crystallization of PbSO,  with
Pb(N03)2.
                 3.3.2  CO.  Carbon monoxide was combined with the
211 ppm NO/N- source and the PbO« sampling method was used in conjunction


with both the SIE and PDS analysis techniques.   The analysis results


are shown in Table XX.
                                   41

-------
2.67
2.65
2.66
2.33
1.94
1.97
136
140
109
210
218
222
218
218
221
222
224
220
           TABLE XX, NO Determination in the Presence of CO


Test No.   Treatment   Ppm NO/Ppm CO    Volume Sampled   Ppm NO (Found)

                            (theo)         (liters)

                                                         SIE       PDS


   1       Cold H20      211/3000


   2          M             it


   Q          II             II



   4       Hot H20          "


   C          II             II



   6          "             "
In tests 1-3, cold water was used to extract the PbO? and the extract


decanted.  The SIE measurements were made after the appropriate amount

     3_
of PO,  buffer was added.  The corresponding PDS results were obtained


after the buffered extracts were allowed to stand for 16 hours in contact


with the PbO-.  The low SIE results indicate that CO also reacts with


PbO™ to form a slightly soluble co-crystal with Pb(NO ), (i.e., PbCO_»

                                                                     3_
Pb(NO_)2.     Since Pb_(PO.)  is much less soluble than PbCO , the PO,

               2-
releases the C0_  and NO, by displacement precipitation as evidenced by


the corresponding PDS results.   In further tests (4-6), the aqueous


slurry was heated as in the S0» study for about 30 minutes followed by


cooling the mixture in an ice bath prior to separating the extract.  SIE

                                                                 3_
and PDS measurements were made after the appropriate amount of PO,  buffer


was added to the separated extracts, and quantitative NO recovery is


again indicated.  Thus far, SO,, and CO interference can be eliminated


simply by heating the aqueous PbO- slurry.
                                   42

-------
          3.3.3  HC1.  Gaseous hydrogen chloride was combined with the
211 ppm NO/N_ source in the dynamic flow dilution apparatus and Table XXI
shows the SIE and PDS analysis results of the PbO? samples.
          TABLE XXI, NO Determination in the Presence of HC1
   Test No.  Treatment   Ppm NO/Ppm
                         HCl(theo)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
             Hot H20     211/2400
             Electrolysis   "
             Hot H20/PbF2   "
                         211/1500
                            it
s Sampled
.iters)
1.97
1.97
1.95
1.94
1.95
1.93
1.95
1.93
1.93
1.95
1.93
1.97
2.45
2.47
2.54
Ppm NO
SIE
258
274
274
256
266
248
207
172
188
192
209
207
218
216
211
(Found)
PDS
	
	
215
200
172
182
162
162
182
212
199
190
196
187
SIE results obtained from tests 1-6 indicate that PbCl- is not sufficiently
insoluble and therefore interferes with both the SIE and PDS analyses when
the earlier hot water extraction technique is used.  The PDS results
corresponding to tests 4-6 were obtained after Ag-SO,  was added to the
extract.   These results are also somewhat erratic,  possibly due to absorp-
tion of N0« by the AgCl and Ag_ PO,  which was formed.   An attempt to
electrolytically strip the Cl  from the extract with a silver billet
electrode was made in tests 7-9.  Low results from both the SIE and PDS
                                   43

-------
analyses indicate that NCL was either absorbed by the AgCl,  electrolyti-



cally reduced at the cathode, or evaporated as HNCL from the acidic



solution (necessary for electrolysis).  The complex salt, PbCIF,  has been


                                                           2+    —      -
used for many years for the gravimetric determination of Pb   ,  Cl , or F .



The solubility of PbCIF is somewhat lower than PbF2 and F~ interferes


                                     _2
with the SIE to the same degree as SO, .  Therefore, the addition of



excess PbF2 to the Pb02/water slurry prior to heating the mixture should



result in the precipitation of PbCIF as follows:



                      PbCl2 + PbF  + 2 PbCIF



The results of tests 10 - 15 indicate that the assumption is valid and



acceptable NO recovery is indicated.  Thus, chloride interference is



eliminated through selective precipitation of PbCIF,



          3.3.4  HF.  Gaseous hydrogen fluoride was combined with the



211 ppm NO/N7 source in the dynamic flow dilution apparatus.  The HF con-



tent of the gas mixture was calculated based on the assumption that only



the monomer (HF) was present.  However, it is well known that hydrogen



fluoride can exist in various molecular multiples up to H.F, in the



gaseous state.  Therefore, the HF level indicated in Table XXII may be



conservative.
                                   44

-------
2.44
2.46
4.54
2.46
201
203
197
199
232
234
224
244
         TABLE XXII, NO Determination in the Presence of HF



Test No.     Treatment   Ppm NO/Ppm HF    Volume Sampled   Ppm NO  (Found)

                             (theo)           (liters)



   1         Hot H.O     211/1700


   2


   O            It           It


   4


* Calibration standards did not contain F  which may account for the 4-10%


bias.


The hot water extraction method was used in tests 1-4 and acceptable


recovery of NO is indicated for the SIE analysis method.


          3.3.5  C0_.  Carbon dioxide was combined with the 211 ppm NO/N_


source in the dynamic flow dilution apparatus and the SIE/PDS analysis


results from PbO- samples are reported in Table XXIII.


        TABLE XXIII, NO Determination in the Presence of CO-


   Test No.     Ppm NO/Ppm CO-    Volume Sampled (1)   Ppm NO (Found)
      1


      2


      3


Instead of extracting the Pb(NO_)2 with hot water as before, the 1000 ppm

  3_
PO,  buffer was added directly to the PbO  and the slurry was shaken for


only two minutes prior to centrifuging and decanting the extract.   If C0?


reacts with PbO. a large error would be evident.   The acceptable results


verify that CO- did not react with the PbO-.
(theo)
211/14,000
it
ii

2.53
2.94
2.79
SIE
219
215
214
PDS
222
218
201
                                   45

-------
     3.4  Analysis of results.  In the foregoing sections it has been



demonstrated that the PbO? sample tube will quantitatively collect NO



at flow rates of 20 to 500 cc/tnin, at NO levels from 50 to 900 ppm, and



at reactor temperatures from 25 C to 190 C.  In addition, it was also



demonstrated that accurate NO analysis results can be obtained even in



the presence of tenfold excesses of SO-, HC1, CO, HF, and C0_ by using



a single sample preparation method and either of two analysis methods.



Table XXIV describes the overall scatter obtained from those results



of section 3.3 where the sample preparation techniques demonstrated that



quantitative recovery occurred and interfering anions were eliminated



through selective precipitation.  In short, all PbO- tube samples known


                          2-     2-     -
to be contaminated with SO,   , CO-  , Cl  and F  can be extracted with



hot water containing excess PbF2, cooled to 0 C, centrifuged, and the


                                   3-
extract decanted prior to adding PO,  buffer.  This sample preparation



technique minimizes the effects of interfering combustion products and



allows accurate NO analyses.
                                   46

-------
TABLE XXIV, NO Recovery After minimizing the Effects of Interfering
Combustion Products
Test No.        Interfering Combustion          Ppm NO Found  (211 ppm theo)
                      PraduCt                          SIE       PDS
   10                 so2
   11
   12                  "
   13
   14
   15
    4                 CO
    5
    6
   10                 HC1
   11
   12                  "
   13
   14
   15
    1                 HF/SiF4
    2
    3
    4
    2
    3
227
223
227
211
211
222
210
218
222
192
209
207
218
216
211
201
203
197
199
219
215
214
X = 212
s - 9.6
X - 211 - +1
187
193
191
211
217
217
222
224
220
182
212
199
190
196
187
232
234
224
244
222
218
201
210
17.5
-1
                                  47

-------
4.0  Field Tests of Manual Methods




     4.1  Field test experimental design.  The field test program was




designed in such a manner that the analysis error portion of the test




could be isolated from the sampling error portion.  This was accomplished




by analyzing the tube and flask samples by both analytical methods (PDS




and SIE).   In addition, to compare sampling methods, two flask (grab)




samples were taken concurrently with each tube (time-integrated) sample.




The time period required to take each tube sample was divided into five




segments and the two flask samples were randomly spaced within the five




segment periods.  Economics and space limitations dictated the number of




flask samples to a maximum of thirty per field test.  Therefore, each




field test was divided into three equal sampling periods (3 days) during




which five tube samples and ten concurrent flask samples were taken each




period.  The data matrix used is shown in Table XXV.
                                   48

-------

Day
1




2




3




Pb02 Tube
Sample SIE PDS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Sub
Sample
Order

3,5
1,2
1,5
3,5
1,2
2,3
2,5
2,4
3,4
3,5
3,5
1,2
1,3
2,4
2,3
TABLE XXV, Data Matrix




                        2 Liter Flask




                      I                II




                 SIE      PDS     SIE      PDS
          49

-------
     4.2  Field Test Analysis Scheme



          The effects of other gases on the hot PbO~ sampling tube



(reactor) have been reported in section 3.3.  Because some of these


          -   -    -2
anions (Cl , F , SO, ) will interfere with the SIE and/or PDS analysis



methods, a special sample preparation procedure was developed for both



the PbO- tube and flask grab samples.  This procedure (Figure 9)  was



tested with aqueous Pb(N03>2 and NaNO, standards for the SIE and  PDS



analysis methods respectively.  In addition, triplicate tube and  flask



samples were taken using the 211 ppm NO/N2 source, and each sample was



analyzed by the SIE and PDS methods.  This was necessary since the



statistical data matrix (Table XXV) requires that the NO  of each sample
                                                        A


be determined by both analysis methods.  Table XXVI shows the averages



of three NO samples,  (SPbO, tubes and 3 flasks) each analyzed by SIE



and PDS.



            TABLE XXVI, Ppm NO from 211 ppm NO/N2 Source



        Sampling Methods                     Analysis Methods


                                             SIE          PDS



        Flask                                257 ± 10     206 ± 2



        Pb02                                 204 ±3      221 + ^




An analysis bias definitely exists for the flask-SIE and possibly for the



tube-PDS combinations.  However, the laboratory NO  source does not contain
                                                  X


the other anticipated stack gases for which the sample preparation pro-



cedure was developed.  The Flask-SIE bias is likely introduced by the



excessive amount of F  which  is released when the Na~PO, buffer is added.



     4.3  Field Test Data.



          Sampling and analysis methods were identical for all of the




                                   50

-------
 TJ


 CO
co  22
3  "n
                 ro

                 3


                 o
                 £   -•
                 m   JQ

                 —I   >

                 §5   C/>
                 m   2
                                  en  <=
                                  -^O  _^
                                  o~-  ui
           CO

           m
                                        CD   OO

                                       it   i.

                                        •o
                                            IS3
                                                       ro
                          o

                          oo
                                        mS
                                        o =
                                                    o

                                                   O
        ro

        3
                                                   8
                                                    o
                                                   O
                                                           o
                                                           m
CD
m
                                                  CD
                                                  m
                                                         CO  Z

                                                         CO  CO
                                                            ro

                                                            BO
                                                                          •o

                                                                          §
                                                                         ro
                                                                          oo
                                                                          m

                                                                          co
                                                                                         O

                                                                                        X
      s   I
oo
ro
CJI

3
                o

                oo
             CO
                           ro
                               orq
                              ro
                                                  OFQ

                                                  -o
                                                  a-
                                                  o
                                                  ro
                                              en

                                              a
                                              zo

                                              -o
                                              CO
                                                                                      DTP



                                                                                      CD


                                                                                      CO
                                O
                                     ro



                                     o
                                     C?
                             00   OO
                                ro
                                en
                                                             ro
                     yo


                     co
                     CO

                     3
                                         co   -n
                                         35   m
                                              yo

                                              ro
                                              o

                                              3
                                           ro

                                           o
                                           ro

                                           a
                                           m
                                           o
                                           O

                                           2
                                           TJ


                                           CO
                                           m
                                           co
                                                                    e
                                                    00
                                                    o
                                                         CO
                                                                  5    3
                                                                  ?»    U>
                                                                                  CO
                                                                            CO
                                                                            2

                                                                            en
                                                51

-------
field tests except Hercules Test No.  1.  This test was conducted



primarily to establish the reliability and sampling equipment require-



ments for a remote operation.  The analysis methods used for  this test



are described in section 2.0.  The sampling apparatus used for the re-



mainder of the field tests is schematically represented by Figure 10.



This apparatus was designed to allow flask and tube samples to be taken



simultaneously from a common gas stream.   The PbO» tubes (D)  were, main-



tained at 180 C. during the sampling process and the sample flow  through



each tube was metered to ^ 100 cc/min by the valve at (G).  The volume



of gas sampled with each tube or flask (C) was determined by the  following



relationship:






     P  - P
V  «  f    o  V,

 ^     BP                       where:



                                      V  = Volume of Sample

                                       g

                                      P, - P  = Pressure change from A, or A»



                                      V,  = Volume of ballast or flask
                                       D


                                      BP = Barometric pressure



The size of the ballast (5 liters) was chosen so that an adequate sample



could be collected for analysis without allowing the final sample pressure



(Pf) to become high enough to condense the moisture in the combustion gas.



For sources containing > 300 ppm NO , 1 liter of gas is sufficient for



either the SIE or PDS method.  This sampling technique was chosen because



it is independent of flow rate changes through the tube and moisture con-



densation can be eliminated thus precluding a separate moisture determination



for accurate emission level calculation.
                                     52

-------
 o   o   oo

 -o   ro    '_
 O-        CJI
 o   r   r-
N5
 00
      oo
00

r—



OO




o
Kg

2  f
»  3

m  >
£?  oo

S  8
«r>  i—
m  c=
 oo
 m
              no

              E
              -o
ffp
c

CD
        m
        m
        oo
                                                  53

-------
          4.3.1  Hercules HN03 Plant.  This plant uses the ammonia



oxidation process for synthesis of N20, and HNO .  The exhausted gas



is saturated with H»0 at •». 110 F and contains both NO and NO- in about



equal concentrations.  Although the plant operates on a continuous basis,



the NO  emission level varies rapidly over a wide range as shown by the
      X


raw data from the two field tests (Tables XXVII and XXVIII).

-------
               TABLE XXVII, Raw Data Hercules HN03 Plant
                           FIELD TEST NO. 1
                            1 February 72
TIME OF DAY


   1300
   1315
   1335
   1350
   1410
   1500
   1520
   1545

   1005
   1020
   1040
   1100
   1117
   1215
   1315
   1340
   1410
   1440
   1510
   1540
                  NO
PPM/Vol
                  PDS
                  600
                  560
                  650
                  630
                  570
                     (1)
   (2)
   V '
       (PDS)
                          450
                          640
                          660
       (570)
       (620)
       (600)
2 February 72
                  500
                  64.0
                  720

                  550
                          560
                          620
                          590
                          890
                          640
                          530
                          540
       (550)
       (580)
       (620)
       (910) *
       (690)
       (580)
       (580)
                                        X =  602
                                        s =   66
                          581
                           68
       (599)
         41
(1)  Flask/PDS
(2)  Pb02 tube/SIE (PDS)
*    Not included in X and s because plant changed operating parameters
     during sample period.
                                   55

-------
             TABLE XXVIII,  Raw Data,  Hercules HN03  Plant
                      FIELD TEST NO.  2 (PPM NO )
                          Tube      Sub             2 Liter  Flask
Day Sample
1 1
2
3
4
5
2 6
7
8
9
10
3 11
12
13
14
15
X
s
£.
SIE
568
565
557
491
440
35 3
365
370
472
596
624
575
579
542
551
510
90
PDS
705
605
535
506
435
335
346
370
514
640
755
660
650
587
648
553
132
Sample
Order
3,5
1,2
1,5
3,5
1,2
2,3
2,5
2,4
3,4
3,5
3,5
1,2
1,3
2,4
2,3


" I
SIE
319
398
296
468
413
402
345
362
486
695
773
713
694
625
628


PDS
412
517
415 *
532
439
446
385
414
551
610
874
723
741
647
584


II
SIE
323
314
526
386
433
400
368
368
509
619
664
761
648
630
672
544
142
PDS
378
410
621
437
440
443
428
413
567
612
644
776
711
563
822
576
141
* Samples above the line leaked air and are not included in statistics

Paired "t" results (95% level):
      tube/SIE t tube/PDS 1 Flask/PDS
                                   56

-------
An extended grab sample technique was used with the 2 liter flask samples



during Test No. 1 and all tube samples were analyzed by both the PDS and



SIE methods.  The use of the extended grab sample technique with this



variable source may account for the significantly better agreement between



the tube and flask samples when compared to the tube and flask data from



Test No. 2.



          4.3.2  Moapa Power Plant.  The Nevada Power Company coal-fired



power plant at Moapa, Nevada, uses a low sulfur type of coal.  The SO



level of the flue gas is reportedly about 300 ppm.  The source variability



was much less than that of the Hercules HNO~ plant as shown in Table XXIX.



The significantly higher values for the flask samples are due to an unde-



termined amount of moisture condensation in the flasks.



        4.3.3  Gas Fired Boiler.  The third source sampled was a natural



gas fired boiler which is rated in excess of 150 h.p.  Samples were taken



from a probe located directly above the heat exchanger tubes.  The gas



temperature at this point is normally about 400 F.  However, since the



field test was conducted in the summer, the boiler was not operating at



full capacity as evidenced by the low NO  level indicated by Table XXX.
                                        X


Only the PbO? tube data is shown because the 2 liter flasks contained



too little sample for analysis.  The gas volume passed through each tube



was approximately 4 liters in order to obtain a large enough sample for



analysis.  The ability to vary the sample size commensurate with the NO



emission level represents a marked improvement, among others, over the



simple grab sample technique.
                                   57

-------
               TABLE XXIX, Raw Data, Moapa Power Plant




                         Field Test  (ppm NO )
Day Sample
1 1
2
3
4
5
2 6
7
8
9
10
3 11
12
13
14
15
X
s
Pb00
SIE
619
620
622
621
615
598
589
569
550
561
530
538
539
537
527
576
38
Tube
PDS
688
582
615
686
661
566
586
616
570
535
524
474
467
492
500
571
73
Sub
Sample
Order
3,5
1,2
1,5
3,5
1,2
2,3
2,5
2,4
3,4
3,5
3,5
1,2
1,3
2,4
2,3

SIE
676
674
665
683
646
692
693
650
652
675
640
593
549
567
564

2 Liter
I
PDS
691
734
679
639
817
682
575
535
573
606
628
526
519
576
530

Flask
II
SIE
654
651
674
666
664
691
663
632
643
635
619
594
592
554
596
638
43
PDS
650
711
625
764
730
635
688
587
617
580
518
610
619
498
499
621
82
Paired "t" results (95% level):
             tube/ SIE = tube/PDS ^ Flask/PDS
                                    58

-------
            TABLE XXX, Raw Data Gas Fired Boiler Field Test
            Day       Test             Ppm NO  (tube)
                                             X

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15


SIE
17
16
15
18
15
17
15
13
3
15
19
19
15
17
16
X - 16
s = 1.7
PDS
18
17
16
17
17
23
23
20
7
19
28
26
18
24
18
20
3.8
     This field test clearly illustrates that the PbO~ tube time-integrated



sampling method is much more versatile than the grab sampling technique



which is strictly limited to higher NO  levels.
                                      X


          4.3.4  Diesel Auxiliary Power Generator.  The final stationary



source tested was the exhaust of a large six-cylinder diesel powered



electric generator.  This test was unique with respect to potential inter-



ferences.  The engine was operated under load but with an extremely fuel-



rich mixture ratio as evidenced by a noticeable amount of unburned fuel





                                   59

-------
which condensed on the inside walls of the 2 liter flasks.  The NO  level




appeared to be very constant and apparently moisture condensation in the




flasks did not significantly affect the results.  This data is shown in




Table XXXI.  It should be noted that the unburned hydrocarbon mist did




not interfere with the PbO» collection of NO.
                                    60

-------
    TABLE XXXI, Raw Data From Diesel Aux. Power Generator (ppm


Pb00
Tube
Sub
2 Liter Flask
Sample I
Order
Day
1




2




3





Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

SIE
411
354
384
391
390
330
337
366
363
372
406
377
361
362
375
X = 372
s - 23
PDS
365
277
291
289
301
324
298
299
299
302
343
502
379
372
364
335
58
3,5
1,2
1,5
3,5
1,2
2,3
2,5
2,4
3,4
3,5
3,5
1,2
1,3
2,4
2,3

SIE
444
443
501
446
479
383
361
386
406
398
460
494
466
446
453

PDS
353
454
398
364
411
385
350
350
332
339
424
403
362
355
373

II
SIE
422
487
438
459
451
383
383
416
433
420
461
489
439
457
449
438
36
PDS
318
353
382
339
346
298
349
363
341
349
367
431
416
372
369
368
21
Paired "t" results (95% level):
        tube/SIE = flask/PDS ^ tube/PDS
                                   61

-------
     4.4  The Field Test Data is summarized in Table XXXII.





       TABLE XXXII, NO  Field Test Data Summary (ppm NO .X ± s)
                      *»•                                X

          Hercules          Moapa         Diesel        Boiler       Std.

       (1)        (2)       (Coal)         (Oil)         (Gas)       (211 ppm)



Tube


SIE    581±68     510±90    576±38        372±23        16±1.7       204±3



PDS    599±41     5531132   571±73        335158        2014         221±23



Flask



SIE      —       5441142   638143        438136         —          257110



PDS    602166     5761141   62H82        368121         —          20612



This data summary shows the mean and standard deviation of each sampling/



analysis method combination for all of the field tests.  Repeated also is



the data from Table XXVII for comparison.  At this point, it should be



recognized that each field test represents a unique situation and only the



data for a single field test can be compared with respect to analysis and



sampling bias.  The combustion sources showed the lowest NO  level
                                                           X


variability as evidenced by the lower standard deviations.  The high



standard deviations for the Hercules No. 2 Flask data clearly show the



need for a continuous sampling method for highly variable sources.  Use



of the extended grab sampling technique (Hercules No. 1) greatly improved



the agreement between the means of the sampling methods.  The effect of



condensable moisture on the analysis results is apparent with the Moapa



Flask data.  Moisture condensation in the flasks leads to gas samples



which are larger than those which can be calculated from flask volumes.



Apparently there was less moisture in the diesel exhaust because the



Flask/PDS and Tube/SIE data agree quite well.  The unburned diesel oil
                                   62

-------
condensed on the walls of the flasks but since it was likely an aerosol


in the first place, it did not contribute to the sample size measure-



ment error.  The excellent agreement between this Flask/PDS and Tube/SIE



data also indicates that unburned hydrocarbons do not seriously interfere



with the PbO. tube sampling method.  A series of "paired t" tests was



performed using the hypothesis that the means of the various Sample



Method/Analysis Method combinations for each field test were equal at



the 95% confidence level.  This data is shown in Table XXXIII.



             TABLE XXXIII, "Paired t" Results at 95% Level



Sample/Analysis Methods   Hercules(No. 2)   Moapa    Diesel    Boiler



Flask/PDS vs_ Tube/SIE           -             -         +


Tube/PDS vs_ Tube/SIE            -             +         -         -



Flask/PDS v£ Tube/PDS           -             -



- Significantly different at 95% level



+ Not significantly different at 95% level






The Flask/SIE data is not compared since an analysis bias is definitely



shown to exist (Table XXVI).  No explanation can be offered for the



Hercules data except that the inequalities are due to the high standard



deviations shown in Table XXXII.  The Moapa data illustrates the signifi-



cance of the sampling method when condensable moisture is present.  The



data from the Diesel engine shows that when condensable moisture is absent



and the source is constant, the tube/SIE method is equivalent to the



flask/PDS method.  The inequality demonstrated with the boiler data can



only be explained by the loss in analysis accuracy at extremely low NO
                                                                      X


levels.
                                   63

-------
     In summary, the tube/SIE method is numerically equivalent to Method



7 (3) for the Diesel generator test.  The lack of equivalency for the




Hercules nitric acid plant and the Moapa power generator was shown to be




due to uncorrected errors in the flask sample.  If the errors were deter-




mined and the results corrected, the numerical equivalency would probably




hold for those sources also.  Thus it is obvious that the tube/SIE method




can be recommended as a replacement for and an improvement over the current




Method 7.




5.0  Theoretical Considerations




     5.1  Discussion of reactions between proposed flue gas substances




and crystalline lead dioxide.




     In Table XXXIV are given proposed reactions between a number of




probable flue gas substances with lead dioxide along with enthalpy and




free energy values (20).  Since the chemistry of most of the reactions




have never been observed in detail, some of the proposed interactions may




not occur in practice.  However, the free energies and possibly the en-




thalpies of the reactions of Table XXXIV may be used as guides for pre-




dicting whether or not absorption of the probable flue gas materials will




occur.




     Data have been accumulated that show nitric oxide, nitrogen dioxide,




carbon monoxide, and sulfur dioxide to be absorbed by electrolytic lead




dioxide while carbon dioxide is unabsorbed.  The free energies of reaction




between lead dioxide and nitrogen dioxide, carbon monoxide, or sulfur




dioxide all exceed minus forty-two kilocalories per mole.  A most interesting




anomaly occurs when the free energies of reaction of nitric oxide and carbon
                                    64

-------
dioxide with lead dioxide are compared.   While both free energies are



negative, the more negative value lies with absorption of carbon




dioxide.  Carbon dioxide absorption does not take place under the usual




experimental conditions (temperatures ranging between room temperature




and 185 C).   In general  chemical experience, reactions that have large




free energies usually occur under relatively mild conditions, but reactions




that have small free energies may be unlikely to occur under the same




conditions.   Thus, quantitative absorption of nitric oxide by lead dioxide




would be predicted to be unlikely on the basis of the small free energy




involved and the observation that carbon dioxide with a more negative




free energy of reaction is unaffected under identical experimental con-




ditions.  From this and other data that  have been obtained, Reaction 1




of Table XXXIV originally proposed by Mishmash and Meloan (4) may be an




incorrect description of the interaction between nitric oxide and electro-




lytic PbO».    This subject will be discussed in more detail later.
                                   65

-------
                    TABLE XXXIV, Thermochemical Data
Probable Reactions between Flue
     Gas Species and PbO_
AH         AF
Kcai/mole  Kca^/mole

1. NO + Pb02 -> N02 + PbO
2. 2N02 + Pb02 + Pb(N03)2
3 . CO + Pb02 -> PbCO,
4. C02 + Pb02 ->• PbC03 + % 02
O O A
6 SO H~ PbO *^" Pb^O "H * Q ^
7. 2HF + Pb02 -*• PbF2 + H20(g) + J-
8. 2HC1 + Pb02 -»• PbCl2 + H20(g) -
9. 2HC1(1) + Pb02 ->• PbCl2 + H20 -
10. Cl2(g) + Pb02 -»• PbCl2 + ^ 02
11. 2HBr + Pb00 + PbBr. + H00 + %
2 Li.
12. 2HN03 + Pb02 -^ Pb(N03)2 + H20i
13. 2HN00(1) + PbO. -»• Pb(NO.)0 + 1
3 2 j /
14. H_SO, + PbOo •*• PbSO, + H_0 + ?
15. H_SO. (1) + PbO- -»• PbSO. + H_0
24 Z 42
16. %S±F, + Pb00 -> PbF_ + %Si00 +
Absorption?
yes
yes
yes
no
yes
unknown
2 02 unknown
1- *S 02 yes
t- ^S 0- unknown
unknown .
09 unknown
^
(g) + % 0_ unknown
120 + ^ 0» no
5 0« unknown
+ h Oj no
*S 00 unknown
. Uv /*
- 1.53
-58.16
-76.3
-16.7
-84.0
-60.36
-23.81
-35.33
- 9.93
-21.65
-41 . 91

-35.94
- 9.01
-18.20
+ 4.15
-12.05
r uvj-
- 2.81
-42.
-66.1
- 4.7
-71.38
-54.43
-16.17
-33.70
- 8.61
-24.62
-40.42

-36.4
	
-32.83
	
- 7.15
                                    66

-------
     Since hydrogen chloride is absorbed by electrolytic lead dioxide,




nitric and sulfuric acid vapors should also be absorbed because of the




similarity of the free energies of reaction.  However, aqueous nitric and




sulfuric acids do not react with lead dioxide at temperatures up to the




boiling point of water, and the thermodynamic values of the enthalpy and




free energy for these reactions are much lower than for the gaseous acid




vapors.  Lead dioxide should also be unaffected by aqueous hydrochloric




acid since the enthalpy and free energy values for such an interaction is




very similar to those with nitric acid.




     From this brief inspection of the thermodynamic values it appears




that those materials having free energies of reaction with lead dioxide




greater than -30 kilocalories per mole are readily absorbed.   Those mater-




ials whose free energies of reaction with lead dioxide are less than -10




kilocalories per mole do not appear to interact under normal experimental




conditions.




     As yet, it is unknown whether or not materials with free energies




of reaction with lead dioxide with intermediate values(hydrogen fluoride)




will be absorbed or not.  The glaring anomaly with this simple reasoning is




the fact that nitric oxide is quantitively absorbed.  However, it is




interesting, where experimental data is available, that acid vapors will




be absorbed while the aqueous solutions are inactive to lead dioxide.




     The outstanding feature in the application of lead dioxide to flue




gas analysis is its surprising ability to quantitatively absorb nitric oxide




at low concentrations.  However, amorphous lead dioxide and other commer-




cial forms of lead dioxide will not quantitatively absorb nitric oxide.
                                    67

-------
Inertness of lead dioxide to nitric oxide is indicated by the free energy



of reaction as was discussed above.  The night-and-day aspect of  the re-



activity of amorphous lead dioxide and electrolytic a and/or  B lead dioxide



with nitric oxide demonstrates that an unusual chemical reaction  is ob-



tained with the electrolytic materials.  X-ray diffraction scans  of the



powder obtained before and after saturation of the "reactive" lead dioxide



with nitric oxide indicates that no new crystalline material  is formed.



That is, the product formed by reaction between nitric oxide  and  "reactive"



lead dioxide is amorphous.  X-ray scans after saturation of the lead dioxide



with nitrogen dioxide definitely show the presence of crystalline lead


                                   —1          —1
nitrate.  Infrared spectra (5000 cm   to 625 cm  ) before and after satura-



tion of the "reactive" lead dioxide with nitric oxide exhibits the appearance



of a very broad band at 1340 cm   which is not the same shape as  the princi-



pal nitrate absorptions in lead nitrate near the same wavelength, and



other small bands are missing.  Nitrogen dioxide saturated electrolytic



lead dioxide, however, does give an infrared spectrum identical with lead



nitrate.  Treatment of the nitric oxide saturated lead dioxide with water



yields nitrate ion.  Additionally, the capacity of the "reactive" lead



dioxide is surprisingly high, up to 75% or more of that predicted by



equations 1 and 2 of Table XXXIV.  Thus, no coating or passivation of the



lead dioxide surface occurs to an appreciable degree.



     The extraordinary reactivity of carefully prepared electrolytic



lead dioxj.de with nitric oxide must arise from a chemical reaction different



from that proposed by Mishmash and Meloan (4).  The data reported above



shows that lead nitrate is not a primary product of reaction between nitric



oxide and lead dioxide although nitrate is formed through later hydrolysis.





                                    68

-------
The nitric oxide reaction is with « or 3-crystalline lead dioxide and not

amorphous lead dioxide.  One possible explanation of the nitric oxide

reaction would be that nitric oxide inserts between two or, preferably,

three oxygen atoms of the crystalline lead dioxide forming two or three

oxygen nitrogen bonds simultaneously in a bridged structure.  An inter-

mediate material with a structure such as,          0 - Pb - 0
                                                 /            \
                                            0=N-0-Pb-0-N-0

                                                  ^ 0 - Pb - 0

could explain the observations that were made.  This model compound fits the

data for the following reasons:  (a)  such an insertion mechanism as

described above readily explains why amorphous lead dioxide, lacking the

necessary geometric array of oxygen atoms, will not interact with nitric

oxide, while the proper crystalline forms of lead dioxide are highly reactive;

(b)  if three nitrogen to oxygen bonds are formed simultaneously by an in-

sertion, the energetics are roughly similar to direct conversion of nitric

oxide to nitrate;  (c)  the unusual infrared spectrum could be explained by

such an intermediate;  (d)  an intermediate such as shown above could be

expected to form nitrate ion during hydrolysis;  (e)  such an intermediate

could be expected to separate out from the lead dioxide crystal leaving a

fresh active surface which explains the high capacity;  (f)  such an inter-

mediate results in the same overall stoichiometry as indicated by the

combination of equations 1 and 2 of Table XXXIV; e.g., 3Pb02 + 2ND -»- Pb(NO )2


+ 2PbO.  Postulation of a rather unusual bridged structure explains the data

obtained, but final verification awaits the isolation and identification

of the crystalline lead dioxide-nitric oxide intermediate.
                                   69

-------
                      Ill - SUMMARY AND CONCLUSIONS



     The grab sampling technique/PDS analysis method for NO  described
                                                           X


in the Federal Register has been evaluated and several critical deficiencies



were discovered:  (a) use of the grab sampling technique requires that many



samples must be taken from a stationary source to accurately define the



24 hour emission levelj (b)    tiam acidic H-0- absorbing reagent in the



sample flask necessitates the use of NaOH for neutralization and intro-


                   2-
duces sufficient CO-  during the analysis to give erratic and erroneously



low results; (c) during the evaporation step, silica from the glassware is



dissolved and later produces particulate matter which, upon filtration,



gives low results; (d) no provision for removal of interfering substances



such as Cl  is mentioned.



     A modified PDS analysis scheme is presented in Appendix I which



minimizes some of the foregoing deficiencies through the use of a neutral



H202 absorbing reagent and platinum crucibles for sample evaporation.  Re-



maining, however, are the sampling and interference problems.



     In order to overcome the problems that are inherent with grab sampling



techniques, a search was conducted for a means of sampling NO  on a rela-



tively continuous basis.  The principal oxide of nitrogen emitted from



stationary sources is NO which unfortunately is a relatively stable and



insoluble species.  A number of liquid absorbing solutions and systems



were evaluated, but all proved to be ineffective as continuous sampling



devices for NO.  The solid sorbent, PbO?, was found to quantitatively ab-



sorb NO and NO- (by converting both to Pb(NO-)2>.  During the initial



studies of this solid sorbent it was determined that the crystal form



of the PbO_ was critical to its reactivity toward NO.  The results of x-ray




                                    70

-------
crystallographic analyses Identified the electrolytically derived PbC>2



(o(and 3 crystal forms) as the only NO reactive source of this solid



sorbent.  A time-integrated sampling scheme for NO  was successfully
                                                  X


developed and evaluated which uses Pb09 as the solid sorbent.  In addition,



the Orion NO- Selective Ion Electrode (SIE) was used successfully as the



analytical device for measurement of NO  in the aqueous PbO  extract.  A



complete sampling and analysis scheme for NO  which describes a Pb09 time-
                                            X                      £»


integrated sampler and a SIE analysis method is detailed in Appendix II.



     Although PbO» absorbs other combustion products such as S0_, CO,



and HC1 which can interfere with the SIE, the analysis scheme presented



describes a simple selective precipitation technique which virtually



eliminates the possibility of other anion interferences.



     Similar techniques for elimination of anion interferences were



developed for the grab sample/PDS method (Appendix III) and field tests



were conducted to evaluate the performance of the two sampling and analysis



schemes.  The field tests consisted of NO  emission measurements of a
                                         x


nitric acid plant, a coal fired power plant, a gas fired boiler, and a



diesel engine exhaust.  During the field tests, grab samples and PbO~ time-



integrated samples were collected simultaneously and both SIE and PDS



analyses were performed on the collected samples.  In all cases the pre-



cision of the results reflected the advantages of the time-integrated PbO?



sample/SIE combination.  An additional feature of this sample technique



was demonstrated by the gas fired boiler tests where the NO  emission level



was insufficient for the collection of an adequate flask sample whereas



by simply lengthening the sampling time, adequate NO  was collected in the
                                                    X


Pb02 sampler for analysis by both the SIE and PDS methods.
                                   71

-------
     While all objectives of the program were met there are some important



additional benefits and features which cannot be overlooked:  (a)  the over-



all cost of an NO  analysis in time and manhours has been appreciably re-
                 X


duced by the PbO_ tube/SIE sampling/analysis system.  For example, five



tube samples of a stationary source effluent can be obtained in about three



hours.  The analysis of the samples requires another hour of laboratory time



so that final calculations could be finished within a half an hour more.



The 16 (or more) hour waiting period requirement of the flask sample is



eliminated as is the more lengthy and tedious PDS analysis; (b) the PbO»



tube sampling device lends itself to interlaboratory comparisons because



an issuing laboratory can expose a number of tubes to a standard NO source,



mail them to recipients, and have^/ig N0_ results for calculations within a



very short time; (c) the accuracy, precision,and sensitivity of the PbO_



tube sampling device have been reported in Section 3.0 of the Technical



Discussion.  The unique character and specificity of 0( and 6 PbO~ for the



quantitative reaction with NO from 25°C to 190°C is of significance.  An



added feature is the relatively large capacity of this material for NO



and NO- absorption.  These factors in addition to insensitivity to unburned



hydrocarbons argue strongly for its consideration also as an NO  sampling
                                                               X


device for mobile and ambient air sources (see Appendix IV); (d)  while HCl,



HF, S09,and CO are apparently absorbed and initially caused some analysis



problems during the program, the "reactivity" of
-------
                                 BIBLIOGRAPHY
 1.  W. Bartok, A.  R.  Crawford, and A.  Skopp,  "Control of  NOx Emissions
     from Stationary Sources",  Chem. Eng.  Prog.,  67,  64(1971).

 2.  "Improved Chemical Methods for Sampling and  Analysis  of  Gaseous
     Pollutants from the Combustion of  Fossil  Fuels",  Walden  Research
     Corp., Environmental Protection Agency, APTD 1291,  February 1970,
     Research Triangle Park,  North Carolina.

 3.  Federal Register, Method #7,  Vol.  36, #247,  pp 24891-24893, Dec 23,  1971.

 4.  H. E.  Mishmash and C. E. Meloan, "The Reactions  of  Nitroc Oxide, Nitrous
     Oxide and Nitrogen Dioxide with Lead  Dioxide," Microchem. J.,  14,
     181(1969).

 5.  F. Pregl, "Quantitative Organic Microanalysis,"  trans, by E. Fyleman,
     2nd ed., 1924, Blakiston's &  Sons, Phila.,  Pa.

 6.  T. Nash, "Absorption of Nitrogen Dioxide by  Aqueous Solutions," J.
     Chem.  Soc. (A), 3032(1970).

 7.  M. Dennstedt and  F. Hassler,  "Ueber das Bleisuperoxydals Absorptionsmittel
     bei der Elementaranalyse", Z. Anal. Chem.,  42, 417(1903).

 8.  F. R.  Cropper, "Microanalytical Determination of  Carbon  and Hydrogen,"
     Mikrochim Acta, p. 25(1954).

 9.  W. R.  Kirner,  "Mechanisms of  Absorption of  Oxides of  Nitrogen by Lead
     Peroxide in Microcombustions," Ind. Eng.  Chem., Anal. Ed.,  10,  342(1938).

10.  American Public Health Association, "Standard Methods for the Examination
     of Water and Wastewater," p.  234,  13th Ed.,  Washington,  DC, 1971.

11.  D. Langmuir and R. L. Jacobson, "Specific Ion Electrode  Determination  of
     Nitrate in some Freshwaters Sewage Effluents," Envir. Sci.  and Tech.,
     4,835(1970).

12.  E. M.  Chamot,  D.  S. Pratt and H. W. Redfield, "A  Study of the Phenoldi-
     sulphonic Acid Method for the Determination  of Nitrates  in Water,"
     J.A.C.S., 33,  366(1911).

13.  E. M.  Chamot and  D. S. Pratt, ibid, 31, 922(1909).

14.  E. M.  Chamot and  D. S. Pratt, ibid, 32, 630(1910).

15.  Inorganic Index,  "The Powder  Diffraction  File, 1968,  "ASTM Publication
     No. PD1S - 18i, Phila.,  Pa. 1968.
                                      73

-------
16.  R. A. Baker, "Conditions for the Formation ofO$ or 3 -Lead  Dioxide
     during the Anodic Oxidation of Lead," J.  Electrochem.  Soc., 109,
     337(1962).

17.  H. Bode, "Chemische Vorgaenge auf Electroden von Galvanischen
     Stromquellen," Angew. Chem., 73, 553(1961).

18.  R. A. Durst, Editor, "Ion Selective Electrodes," NBS,  Special Publica-
     tion No. 314, pp 57-87, Nov 1969.

19.  B. A. Coulehan and H. W. Lang, "Rapid Determination of Nitrogen Oxides
     with the Use of Phenoldisulfonic Acid," Envir.  Sci. and Tech., 5,
     163(1971).

20.  JANAF Thermochemical Tables, NBS Circular No. 500, Washington, DC  and
     Zhumal Prikladnoi Khlmii, 40(11), 2583-2586(1967).
                                   74

-------
                                                                       1-1
                               Appendix I

                     Recommended Changes to Method 7

        (Fed. Register, Vol. 36, No 247, Part II,  dated 23 Dec 71)


1.  Principle and Applicability

    1.1  Principle.  A grab sample is collected in an evacuated flask con-

taining a neutral 3% hydrogen peroxide absorbing solution and the nitrogen

oxides, except for nitrous oxide, are measured colorimetrically using the

phenol-disulfonic acid (PDS) procedure.

    1.2 to 2.2.3 inclusive:  no changes

    2.3  Analysis

         2.3.1  Steam bath or resistance heated hot plate.

         2.3.2  Platinum or gold-plated nickel crucibles (15-25 ml size),

one for each sample, standard and reagent blank.

         2.3.3  Volumetric pipettes, 1 ml, 2 ml, 5 ml, 10 ml.

         2.3.4  Transfer pipette, 10 ml with 0.1 ml graduations.

         2.3.5  Volumetric flasks, 100 ml, one for each sample, standard

and reagent blank, 1000 ml for std. solution.

         2.3.6  Spectrophotometer, narrow band, to measure absorbance at

410-420 nm.
         2.3.7  Graduated cylinder, 100 ml, 1 ml graduations.

         2.3.8  Analytical balance, lOOg capacity, 0.1 mg sensitivity.

         2.3.9  Ice bath - 1 liter plastic beaker and chipped ice + H~0.

3.  Reagents

    3.1  Sampling

         3.1.1  Absorbing solution - Dilute 100 ml of 30% hydrogen peroxide

to 1 liter with distilled water.  Mix well and store in a clean bottle away
                                    75

-------
                                                                       1-2
from heat and light.   Prepare a fresh solution weekly.




     3.2 to 3.3.3 inclusive:   no change




          3.3.4  Standard solution - Dissolve exactly 2.1980g  of  dried




potassium nitrate (KNO.,) in distilled water and dilute to 1 liter in a




1000 ml volumetric flask (1 ml - 1000jug NO-).  For the working standard,




dilute 10 ml to 100 ml with distilled water in a 100 ml volumetric flask




(1 ml - 100jug N02)•   Store both solutions in screw-capped plastic




containers.




          3.3.5 to 4.1.1 inclusive:  no changes




     4.2  Sample Recovery




          4.2.1  Connect the flask to a mercury-filled U-tube  manometer,




open the valve from the flask to the monometer and record the  flask pres-




sure, temperature and barometric pressure.  Allow the flask to stand for




a minimum of 16 hours and shake the contents for 2 minutes. If the contents




of the flask are to be shipped, transfer the 25 ml to a dry plastic con-




tainer (bottle).  Add 5 drops IN NaOH prior to shipping.




     4.3  Analysis




          4.3.1  Place 1 drop of IN NaOH into each of the platinum or gold-



plated nickel crucibles.  Pipet 5 ml of the peroxide absorbing solution




into one crucible for a reagent blank.  Pipet a 1/5 (one fifth) aliquot




of the sample solution into another crucible (5 ml for a non-shipped sample




and 10 ml from a shipped sample after adjusting its volume to  50 ml).




Allow the peroxide decomposition to proceed at room temperature for 5-10




minutes.  Place the crucibles on a 100 C hot plate or steam bath and




evaporate solutions to dryness.  Remove the crucibles and allow to cool,




and add 2 ml of the PDS reagent.  Allow 5 minutes for dissolution of solids






                                  76

-------
                                                                          1-3
(tilt and rotate the crucible to wet the walls with PDS).   Transfer the




PDS solution from the crucible to a 100 ml volumetric flask, rinsing 3 or




4 times with ^ 10 ml portions of distilled water.  Use a plastic funnel.




Place the 100 ml volumetric flask into an ice bath and slowly add 8 ml




concentrated NH,OH(28% NH ) with swirling.  Make up to the mark with dis-




tilled water and mix thoroughly.  Set the spectrophotometer to "0" absorbance




(100%T) at 410-420 nm with a 1 cm cell containing the blank which was




treated the same as the sample.  Record the absorbances of the yellow




sample solutions in a 1 cm cell.  Determine the/tig NO-/100 ml for the




aliquot from a previously or concurrently prepared calibration curve.




5. to 5.1 inclusive:  no changes




    5.2  Spectrophotometer Calibration - Add 0.0 ml, 1.0 ml, 2.0 ml, 3.0 ml and




4.0 ml of the nitrate (lOO^ug NO- = 1 ml) working standard solution to each




of five (5) crucibles.   Add 1 drop of 1 N NaOH plus 5 ml of the peroxide




absorbing solution.  Evaporate the solutions on a hot plate or a steam bath




and continue the procedure of section 4.3.  Plot the absorbances of the




standards on linear graph paper and draw a smooth line through the origin.




The slope should be 0.15 ± 0.007 absorbance per 100^/ig NO  /100 ml.  Calcu-




late the total^g NO- per sample as follows:




            M = a x F




            where M = Total lag NO- (sec 6.2)




                  a =yig NO-/100 ml from calibration curve




                  F = aliquot factor (i.e., 25/5, 25/10, etc.)
                                   77

-------
                                                                   II-l
                              Appendix II




   Determination of Nitrogen Oxide Emissions from Stationary Sources




1.  Principle and Applicability




    1.1  Principle.  A gas sample is slowly drawn through a heated  glass




tube containing an active form of lead dioxide thereby converting the




nitrogen oxides, except nitrous oxide to lead nitrate.  The lead  nitrate




is extracted with water and the nitrate ion concentration is measured with




the Nitrate Selective Ion Electrode.




    1.2  Applicability 	




2.  Apparatus




    2.1  Sampling.  See Figure 1.




         2.1.1  Probe - Borosilicate glass, heated.  Heating is unnecessary




if the dewpoint of the sample gas is equal to or below ambient temperature.




         2.1.2  Air Pump - Equipped with inlet filter and water separator.




Capable of pumping 20-50 standard liters/min.




         2.1.3  NOx absorption tube - Borosilicate glass, h in. O.D. (6 mm)




and packed with 2-4g of 30/50 mesh Pb02.  The granular PK>2 is held in place




with borosilicate glass wool plugs.




         2.1.4  Compression Fitting - For \ in. O.D.  (6 mm) tubing, either




"0"-ring or polytetrafluoroethylene ferrule type, vacuum tight, 2 required




for glass to metal seals.




         2.1.5  Tube furnace or heater - capable of maintaining 100-180 ± 10 C.




         2.1.6  Flow Control Valves - Vacuum tight - A shut off valve




followed by a fine metering valve.  System should be  capable of flow con-




trol from 10-200 cc/min at one atmosphere pressure differential.
                                   78

-------
                                                                                              II-2
             en   -P*
CT)
     £5
     o
ro
cr>
oo
oo
-
        m
        CO  Ni   —*   o  CD  OO
        £  £   5   g  2  s
        g  o   p   ^  3  «
        c  ^   *•       rn  P^
        ^  ^   CO   ~O  ^  m
        ^  S   —4   30  ^  _
        ^  ^   3  «"
        i  ?   m  <»
        •°  ^   >  S?
                          o
                en  c?
                     m
                m  3O
                3D  -3-
                C3  orq
                %  S
                2  z
                —<  o
                •<  S
                                                             CO
                                             79

-------
                                                                     II-3






         2.1.7  Pressure Gauge - Absolute pressure gauge calibrated in




1 mm increments from 0-760 mm or a similarly equipped mercury "U-tube"




manometer.




         2.1.8  Ballast Volume - Metal, thin wall, approximately 5 liter




volume for NOx sources of 50-1000 ppm.  A larger volume should be substi-




tuted if the source contains  \50 ppm NOx.




         2.1.9  Vacuum Valve - Vacuum tight, bellows seal.




         2.1.10  Vacuum Pump - Two stage, 50 liter/minute capacity, ulti-




mate vacuum capability of 0.1 mm Hg or better.




         2.1.11  Thermometer - 0-100°C range, 1°C graduations.




    2.2  Sample Recovery




         2.2.1  Centrifuge Tube - Borosilicate glass, 12 ml capacity and




equipped with a screw cap, one for each absorption tube.




         2.2.2  Microspatula - sized to fit inside the absorption tube.




         2.2.3  Stirring rod - sized to fit inside the absorption tube but




somewhat longer.




         2.2.4  Syringe - 10 ml capacity, 0.2 ml graduations.




         2.2.5  Hot Plate - 36 sq in. heating surface, capable of boiling




water.




         2.2.6  Beaker - Borosilicate glass, 2 required, for boiling water




and for ice bath, 600 ml capacities.




         2.2.7  Clinical Centrifuge - Head cavities sized to fit 12 ml




centrifuge tubes, 2500 rpm capability.




         2.2.8  Test tube rack - to fit 12 ml centrifuge tubes.




         2.2.9  Plastic Bottle - Polyethylene, narrow mouth, screw capped,




100 ml.
                                   80

-------
                                                                        II-4





    2.3  Analysis



         2.3.1  Syringe - 250^/11 (0.250 ml) capacity, lOyul (0.01 ml)



graduations.



         2.3.2  Beaker - Plastic, disposable, 25 ml capacity, one for each



sample and calibration standard.



         2.3.3  Volumetric pipettes - 2, 3, 5, 10, 50 and 100 ml capacities.



         2.3.4  Volumetric flasks - 2-100 ml, 6-1000 ml.



         2.3.5  Plastic Bottles - Polyethylene, narrow mouth,  screw-capped,



2-100 ml, 6-1000 ml.



         2.3.6  Nitrate Selective Ion Electrode, and kit for regeneration



of electrode.



         2.3.7  Ag/AgCl - Reference electrode, and AgCl/KCl filling solution.



         2.3.8  Analytical balance - To measure to 0.1 mg.



         2.3.9  Electrometer Amplifier - With millivolt readout, readable



to 1 mv.



3.  Reagents



    3.1  Sampling



         3.1.1  Lead Dioxide - 30/50 mesh, electrolytically derived ( ' and


                            (2)
purified according to Pregl



    3.2  Sample Recovery



         3.2.1  Lead Fluoride - Anhydrous, purified, nitrate free.



         3.2.2  Phosphate Buffer - Dissolve 20g of ACS reagent grade



Na-PO, .12H 0 in 80g (ml) of distilled water.  This is nearly a saturated



solution therefore gentle heating of the mixture may be necessary.  When



dissolved, transfer the solution to a 100 ml plastic bottle.  Minimize



exposure of this solution to the atmosphere to prevent C0» absorption.
                                   81

-------
                                                                        II-5



    3.3  Analysis

                           3_
         3.3.1  2000 ppm PO,   solution - Dissolve 0.800 g  of ACS  reagent


grade Na_PO,  .12H20 in distilled water contained  in a 100  ml volumetric


flask, and dilute to the mark.  Mix and transfer  the contents  of  the flask


to a 100 ml plastic bottle.

                           3_
         3.3.2  1000 ppm PO,   solution - Dissolve 4.002 g  of ACS  reagent


grade Na,PO,  .12H 0 in distilled water contained  in a 1000 ml  volumetric


flask and dilute to the mark.  Mix and transfer the contents of the flask


to a 1000 ml plastic bottle.   At least 6000 ml will be required.


         3.3.3  10,000 ppm N0~ Standard - Dissolve 2.6708g of  dried ACS


reagent grade Pb(NO-)2 in distilled water contained in a 100 ml volumetric


flask and dilute to the mark.  Mix and transfer the contents of the flask


to a 100 ml plastic bottle.


         3.3.4  NO- Selective Ion Electrode calibration solutions.


                3.3.4.1  1000 ppm N0~ Standard - Pipette 10.0  ml  each of the

                                           3-
10,000 ppm NO- standard and the 2000 ppm PO,  buffer solution  into  a 1000 ml


flask and dilute to the mark with 1000 ppm PO,  buffer solution.  Mix and


transfer the contents of the flask to a 1000 ml plastic bottle.


                3.3.4.2  500 ppm N0~ Standard - Pipette 5.00 ml each of  the


10,000 ppm N0~ standard and the 2000 ppm P0^~ buffer solution  into  a 1000 ml

                                                        3_
volumetric flask and dilute to the mark with 1000 ppm PO,   buffer solution.


Mix and transfer the contents of the flask to a 1000 ml plastic bottle.


                3.3.4.3  300 ppm N0~ Standard - Pipette 3.00 ml  each of  the

             -                             3-
10,000 ppm NO- standard and the 2000 ppm PO,  buffer solution  into  a 1000 ml

                                                            3_
volumetric flask and dilute to the mark with the 1000 ppm PO^   buffer


solution.  Mix and transfer the contents of the flask to a plastic  bottle.
                                   82

-------
                                                                        II-6
                3.3.4.4  200 ppm W^ Standard - Pipette 2.00 ml each of

                 _                             3_
the 10,000 ppm NO, standard and the 2000 ppm PO,   buffer solution into a


1000 ml volumetric flask and dilute to the mark with the 1000 ppm PO ~


buffer solution.  Mix and transfer the contents of the flask to a 1000 ml


plastic bottle.


                3.3.4.5  100 ppm N0~ Standard - Pipette 100.0 ml of  the


1000 ppm NO, standard into a 1000 ml volumetric flask and dilute to  the  mark


with the 1000 ppm PO,  buffer solution.  Mix and transfer the contents of


the flask to a 1000 ml plastic bottle.


                3.3.4.6  50 ppm N0~ Standard - Pipette 50.0 ml of the 1000


ppm N0_ standard into a 1000 ml volumetric flask and dilute to the mark  with

               3_
the 1000 ppm PO,  buffer solution.  Mix and transfer the contents of the


flask to a 1000 ml plastic bottle.


4.  Procedure


    4.1  Sampling


         4.1.1  Assemble the probe and air pump as shown in Figure 1 but


with the sample inlet side arm capped.  Insert the probe into the sample


port and turn on the air pump so that sample gas is drawn through the system


at 20-50 liters/min.  Heat the probe to approximately the sample gas tempera-


ture if the dewpoint of the gas is above ambient temperature.  Allow the


system to purge for several minutes before continuing.  Insert a sample


absorption tube into the heater (adjusted to 100 - 180 ± 10 C) and connect


the tube to the flow control valves as shown.  With the toggle valve closed,


open the vacuum valve and evacuate the ballast volume and pressure gauge


(or manometer).  Close the vacuum valve and record the initial pressure  (Po).
                                   83

-------
                                                                        II-7
Po should be constant for at least 20 min.  to insure that the system is



free from leaks.  If no leak is detected connect the sample absorption




tube to the sample probe side arm.  Open the toggle valve and adjust the




metering valve to the desired sampling rate (determined by the rate of




pressure increase in the ballast volume).  After the desired sample size




has been accumulated, close the toggle valve and remove the sample absorp-




tion tube.  Record the final pressures (P.) in the ballast volume and ambient




temperature.  In addition, use the gauge (or manometer) to determine the




barometric pressure.  The volume of the ballast vessel and the sample volume




should be correlated with the moisture and NOx content of the sample gas




such that adequate NOx for analysis can be collected without exceeding a




partial pressure of water vapor in the ballast volume whereby condensation




will occur.  Evacuate the ballast volume and repeat the sequence for sub-




sequent samples.  The sample absorption tubes should always be protected




from atmospheric contamination because the PbO^ will absorb NOx even at



ambient temperature.




    4.2  Sample Recovery




         4.2.1  Remove one glass wool plug with a microspatula from the




sample absorption tube and transfer it and the PbO  to a centrifuge tube.




Remove the remaining plug by pushing it through with a stirring rod in the




same direction so that the PbO,, adhering to the inside wall of the tube is




also transferred to the centrifuge tube.  With a microspatula add approxi-




mately 0.1 g of PbF? to the contents of the centrifuge tube.  Use a 10 ml




syringe to transfer 8.0 ml of distilled water to the mixture and cap the




centrifuge tube securely.  Thoroughly mix the contents of the centrifuge
                                   84

-------
                                                                         II-8
tube and place it in boiling water.  Allow the slurry to remain at ^ 100°C


for 15-30 minutes with occasional mixing.  After the heating period, cool


the slurry to 0 C in an ice bath.  Allow the tube to cool for 15-30 minutes


and then centrifuge the mixture at ^2500 rpm for 5-10 minutes.  Remove the


tube from the centrifuge -and place it in a test tube rack.  Allow the con-


tents of the centrifuge tube to return to ambient temperature and decant


the supernatant liquid into a disposable beaker.  Add ISOyil (0.15 ml) of

     3-
5% PO,  buffer to the contents of the beaker and swirl the mixture.


    4.3  Analysis


         4.3.1  Immerse the Nitrate Selective Ion Electrode and the Ag/AgCl


reference electrode in the buffered solution and rotate the beaker several


times to remove any gas bubbles which may be adhering to the electrodes.


Allow the millivolt meter to stabilize (^30 sec.) and record the millivolt


reading.  Compare the reading with the calibration curve prepared in Section


5.2 to determine the nitrate ion concentration of the sample solution and


use the formulas in Sections 6.2 and 6.3 to calculate the NOx concentration


of the gas sample.


5.  Calibration


    5.1  Sampling Apparatus (Figure 1)


         5.1.1  With the toggle valve closed open the vacuum valve and


evacuate the apparatus.  Following evacuation, close the vacuum valve and
                                                             i

disconnect the vacuum line from the vacuum valve.  Record the system


Pressure (Po) and connect a 1 to 10 liter vessel containing dry air at one


atmosphere pressure to the vacuum valve.   The water volume of the vessel


should be previously determined.  Open the vacuum valve and record the


equilibrium pressure (Pf).  Also record the barometric pressure.   Use the


formula in Section 6.1 to determine the system volume.


                                   85

-------
                                                                    II-9
    5.2  Nitrate Selective Ion Electrode (SIE)




         5.2.1  Rinse the SIE and reference electrode with distilled water




and dry with a clean tissue.  In turn,  starting with the most dilute NO.,




standard, transfer approximately 10 ml  to a clean plastic disposable beaker




and immerse the electrodes in the liquid.  Rotate the beaker several times




to remove any gas bubbles which may adhere to the electrodes and allow ap-




proximately 30 seconds for the millivolt meter to stabilize.  Record the




millivolt reading and the corresponding N0_ concentration.  Remove the




beaker, wipe the electrodes with a clean tissue, and proceed to the next




most concentrated standard without rinsing the electrodes.  Plot log N0~




concentration vs millivolt reading (use of semilog graph paper is most con-




venient) .  After the millivolt reading  for the most concentrated N0_ standard




is recorded, place the dried electrodes in one of the NO- standards which




approximates the NO  concentration of the sample.  This serves both to check




the stability of the electrode and also decrease the equilibration time of




the electrode with the sample solution.  Do not readjust the calibration




of the meter until the reading has been checked several times with fresh




solutions of the more dilute N0_ standard.  Once calibrated, the electrodes




should only be carefully wiped with a clean tissue between each standard




and each sample.  In the event that the millivolt reading cannot be adjusted




to the normal value for a given standard or that the reading drifts con-




tinuously, the electrode should be reconditioned in accordance with the




manufacturer's instructions.
                                  86

-------
                                                                       11-10
6.  Calculations

    6.1  Sampling apparatus volume.
                           \  - v
                          /    ^
                                      where:
                                            Va
                                            Vs
                                            Pbp
                                            P
     equation 1


  sampling app.  vol  (liters)
  water  vol.  of  std.  (liters)
 = barometric pressure (mmHg)
  initial  pressure (mmHg)
  final  pressure (mmHg)
    6.2  Gas Sample Volume.
             Vg
                                                   equation 2
                                      where:
                                            Vg » gas sample volume (liters)
                                            Va = sampling app.  vol (liters)
                                            Pbp = barometric pressute at
                                                  sample site (mmHg)
                                            P.. = final apparatus pressure (mmHg)
                                            P  = initial apparatus pressure  (mmHg)
    6.3  NOx concentration.
                    APR
                    VgM
                                      where:
                                            A =

                                            F =

                                            R =



                                            Vg
                                                  equation 3
 ppm N0_  in sample solution
 c  n  i    7.0 + 0.15    Q  ,,  T
 8.0 ml x 	=—=	  = 8.17 ml
 22.4  x
        760
7.0
 x
(°K)  at
        Pbp   ~   273*
 sample site,  PbP = (mmHg)
 barometric pressure at  sample  site
= sample volume  (liters)
                                            M = formula wt.  N0~ (62.01)
                                   87

-------
                                                                        11-11
7.  Bibliography

    1.  Final Report,  EPAOOOCX,  etc.

    2.  Pregl. F., "Quantitative Organic Microanalysis" trans,  by E.  Fyleman,
        2nd Ed., 1924, P. Blakiston's Son & Co.,  1012 Walnut Street,  Phila,  Pa.
                                   88

-------
                                                                      III-l
                              Appendix III






    The following procedure is designed to determine^ug nitrate ion in



the presence of chloride ion when Method 7 flask samples have been obtained.



It is important that the final sample pressure measurement be made at the



end of paragraph 4.1.1., and prior to allowing the sample to stand for a



minimum of 16 hrs.  Assuming that HC1 was a stack gas constituent and the



25 ml absorbing solution is neutral 3% H~09, the modified analysis begins



with paragraph 4.3, while apparatus and reagent additions are listed in



paragraphs 2.3 and 3.3.  The reaction with phenoldisulfonic acid takes



place in an aqueous medium which helps prevent loss of NO  by volatilization
                                                         X


as NOC1.



    2.3  Analysis



         2.3.1  Steam bath or resistance heated hot plate.



         2.3.2  Beakers, 50 ml size, one for each sample, standard and



reagent blank.



         2.3.3  Volumetric pipettes, 1, 2, 5, 10, 20 ml.



         2.3.4  Transfer pipette, 10 ml with 0.1 ml graduations.



         2.3.5  Plastic bottles, about 75 ml size, one for each sample, std.,



and reagent blank, with screw cap.



         2.3.6  Volumetric flasks, 1000 ml for Std. solutions, 25 ml and 100 ml



one for each sample, std ._, and reagent blank.



         2.3.7  Spectrophotometer, narrow band, to measure absorbance at



405-410 nm.



         2.3.8  Graduated cylinder, 100 ml with 1 ml graduations.



         2.3.9  Analytical balance, lOOg cap., 0.1 mg sensitivity.



         2.3.10  Ice bath, 1 liter, plastic or metal container.




                                   89

-------
                                                                     III-2
         2.3.11  Centrifuge, clinical,  capable of accepting 12-15 ml




conical tubes.




         2.3.12  Centrifuge tubes,  glass,  12-15 ml,  one for each sample,




Stdv and reagent blank, with teflon seals  in the screw cap.




    3.3  Analysis




         3.3.1 to 3.3.3 inclusive - no  changes.




         3.3.4  Standard Solution - Dissolve exactly 2.1980g of dried ACS




Reagent grade potassium nitrate (KNO,)  in  distilled  water and dilute to




the mark in a 1000 ml vol flask (1 ml = 1000 ug N0_).   The working standard




is prepared by diluting 10 ml to 100 ml with distilled water in a 100 ml




volumetric flask (1 ml = 100 ug NO-).




         3.3.5 to 3.3.6 inclusive - no  changes.




         3.3.7  Lead Dioxide, PbO?, nitrate-free, electrolytically derived




and prepared according to Pregl.




         3.3.8  Lead Fluoride, PbF_, nitrate-free, purified grade.




4. to 4.1 inclusive - no change, except  to  measure flask pressure at end




of paragraph 4.1.1.




    4.2  Sample recovery.




         4.2.1  Let the flask stand for a  minimum of 16 hours and then




shake contents for 2 minutes.  Remove flask valve stopper and carefully




transfer contents to a 75 ml plastic bottle into which 5 drops of IN NaOH




have been added.  Use a plastic funnel.  For a blank add 25 ml neutral H^O^




to another bottle with 5 drops of IN NaOH.  Cap the bottle.  The sample is




now ready for shipment or analysis.




    4.3  Analysis




         4.3.1  Shake the plastic bottle to remove condensed moisture from




the sides.  Carefully loosen the screw cap to allow gas pressure to escape^




                                   90

-------
                                                                      III-3
and remove screw cap.  Add approximately O.lg PbC^ powder to the solution




and replace cap loosely and allow the H20~ decomposition to proceed to




completion.  Transfer 20 ml to a 25 ml volumetric flask with a 20 ml




volumetric pipette, dilute to the mark with distilled water and mix.




Transfer a 10 ml aliquot with a 10 ml volumetric pipette to a 12-15-ni/




glass centrifuge tube to which has been added approximately O.l^PbF , cap




and shake.  Cool the centrifuge tube to 0 C in an ice bath (^ 10 mins.),




remove and place into centrifuge.  Centrifuge at 2000-3000 rpm for 10




minutes.  Return the tube to the ice bath and remove a 5 ml aliquot into




a 50 ml beaker.  Add 1 drop IN NaOH, mix and place the beaker on a steam




bath or hot plate (^ 100 C) and evaporate to 1 or 2 ml.  Remove the beaker




and allow to cool.  Add 2 ml of the phenoldisulfonic acid reagent.  Mix




gently and place on the steam bath or hot plate for a least 10 minutes.




Remove and allow to cool.  Transfer the solution to a 100 ml volumetric




flask, rinsing 3-4 times with ^ 10 ml portions of distilled water.  Use a




plastic funnel.  Place the 100 ml volumetric flask into an ice bath, swirl,




and slowly add 8 ml concentrated NH.OH (28% NH-) using a transfer pipet.




Dilute the solution to the mark with distilled water and mix thoroughly.




Set the spectrophotometer to "0" absorbance (100% T) at 405 to 415 nm with




a 1 cm cell containing a reagent blank which was prepared in the same




manner as the sample.  Record the absorbance of the sample solution (yellow)




using the same 1 cm cell.  Determine the^ug N02/100 ml for the aliquot




from a previously or concurrently prepared calibration curve.




5. to 5.1 inclusive - no change.




    5.2  Spectrophotometer.  Add 0.0 ml, 1.0 ml, 2.0 ml, 3.0 ml and 4.0 ml




of the nitrate working standard solution (100 Xig N02 = 1 ml) to each of five




plastic bottles containing 25 ml, 24 ml, 23 ml, 22 ml and 21 ml of neutral





                                   91

-------
                                                                       III-4
3% H-O-.  Add 5 drops of IN NaOH to each and mix.   Continue the procedure




as outlined in paragraph 4.3 by adding approximately O.lg PbO,, to each.




Plot the absorbances of the standards on linear graph paper and draw a




smooth curve through the origin.  The slope should be 0.14 to 0.145




absorbance per 100 ug NO-/100 ml.  Calculate the total jug NO- per sample




as follows:




                           m = 4a




                  where:   m * total^g N02 (see para 6.2)




                           a =jug NO-/100 ml from calibration curve




                           4 = aliquot factor (i.e., 20/5).
                                   92

-------
                                                                         IV-1
                               Appendix IV



      Application of PbO- to Mobile Source and Ambient Air Sampling
I.  Mobile Sources



    1.  NO  concentration range.  The stationary diesel engine which was
          X.


tested  (see Sec. 4.4) produced approximately 370 ppm NO .  In addition, a
                                                       X


392 cu. in. 1970 automobile engine which used premium grade gasoline was



tested  (not reported in the  text) at fast idle and no load.  The average



emission level for this source was 50 ppm.  In view of the above, it



seems reasonable to assume that mobile source NO  emission levels should
                                                x


range between 10-1000 ppm.



    2.  Required Sampling Parameters.  The greatest difficulty in sampling



for NO  will occur at the lowest emission levels (i.e., 10 ppm).  At low
      X


emission levels large gas samples are required to collect sufficient NO



for analysis with the S.I.E.  By rearrangement of equation 6.3 (Appendix II),



the necessary gas sample volume for an adequate N0» concentration in the



final solution can be determined.  If 25 ppm N0« is the lowest acceptable



level for accurate analysis with the SIE, then using equation 6.3 the sample



volume necessary for a 10 ppm NO  source is:
                                x




                      = AFR = 25x8x22.4

                    8   CM    10x62




                   Vg = 7.2 liters



If it is assumed that CO is the major competitor with NO  for active sites



on the PbO  and that only 75% of the theoretical PbO  capacity is avail-



able (see Sec.  5.1),  then by using the CO/PbO» equation from Sec. 5.1 and



the required sample volume (Vg)  for 10 ppm NO  it is possible to calculate
                                   93

-------
                                                                        IV-2
the highest CO level that can be tolerated when sampling a mobile source



as follows:






                         UP            f\

              CO ppm  - r~-  0-75 x 10°
                        M Vg



              CO ppm  -  39,0001



              where:  W - 4g Pb02/r,ube



                      M - 239g Pb02/mole



                      R - 22.4 I/mole @ STP



                      Vg - 7.2 liters



Therefore, the PbO_ tube (4g Pb02) capacity will be exceeded if  the source



containing 10 ppm NO  also contains 39,000 ppm CO (3.9%) when 7.2 liters of



gas are sampled.  A recent National Air Pollution Control Administration



(NAPCA) publication (1) reports that the average CO emission level of an



internal combustion engine is between 10,000 and 20,000 ppm (1-2%).



In view of this data, the PbO_ sampling tube can be used for mobile sources



with at least a 100% safety factor even under the worst possible conditions.








II.  Ambient Air



     1.  Mean NO  Concentration.  A much larger gas sample will be required
                X


for measurement of NO  levels in ambient air since the mean level is much
                     x


lower than the normal emission levels of both stationary and mobile sources.



However, if a nominal NO  level in ambient air is chosen, then the minimum
                        X


sample size for SIE measurement can be calculated as before.  At this point



it must be recalled that Pb02 reacts with both NO and NO  to form Pb(NO )2>



If 25 ppm N0~ is again chosen as the minimum acceptable level in the final
                                   94

-------
                                                                        IV-3
solution for SIE measurement,  then the only remaining information necessary
is a nominal ambient air NO  level.  Data summarized in a recent NAPCA
                           x
report (2) shows that 0.05 ppm is a reasonable low nominal NO  level for
ambient air in U.S.  cities.  Use of the above and equation 6.3  defines
the minimum gas sample size for accurate NO  measurement.
                      AFR  _ 25x8x22.4
                  g  = CM   ~ 0.05x62
                 Vg  = 1445 liters
                                   95

-------
                                                                         IV-4
    2.  Sampler Configuration.   The sample volume is not as unwieldy as



it may seem if the collection time is expanded to a 24  hour period  (i.e.,


1445
-, ,n—   = 1 liter/min).   With a modification in the sampler  configuration,



such as that depicted in Figure 1, sample collection should be relatively



simple.  On the basis of  other  NAPCA data (3)  the major competitor  with  NO



for active sites on the PbO- is again CO.  A calculation for the maximum



CO level which can be tolerated in a gas sample of 1445 liters is  shown  below:



       CO ppm = ^—  0.75 x 10         where:  W = 4g  PbO
                                                M = 239g PbO_/mole

       CO ppm = 195

                                                R = 22.4 I/mole @ STP



                                                Vg = 1445 liters
This CO level is far above the nominal daily average (10-50 ppm)  which was



reported in the NAPCA publication.   Therefore,  it also appears that PbO-



can be used for ambient air sampling as well as for mobile and stationary



sources.  An additional feature of  the Pb09 sampler is that SO ,  CO, HC1,
                                          ^                   X


and possibly other pollutants  can  be determined on the same sample when



the applicable analysis techniques  are perfected.
                                   96

-------
                                                                                        IV-5
                                                                           O
                                                                           O
                                                                   O
                                                                                  trp

                                                                                  CD
       O



       O
       O
o
r—

<

r—
       -o

       1
       -o
=O


O
o
m
                     O
                    1X5
                            3

                            3 ~
                            a o
                         m
                         oo
                               CO
                               3
                               3
 00

 -o
 C3-
 o
ro

 o
 o

 -4
 *»

 m
-a  o
P~  CD
m  ...
    m
    5O

    -n
    o
    TO
                                         97

-------
                                                                        IV-6
                              Bibliography








1.  "Control Techniques for Carbon Monoxide,  Nitrogen Oxide,  and  Hydro-




carbon Emissions from Mobile Sources", NAPCA Publication No.  AP-66,  1970,




U. S. Government Printing Office, Washington, D.C.,  20402








2.  "Air Quality Criteria for Photochemical Oxidants", NAPCA  Publication




No. AP-63, 1970, U. S. Government Printing Office, Washington,  D.C., 20402








3.  "Air Quality Criteria for Carbon Monoxide",  NAPCA Publication No.  AP-62,




1970, U. S. Government Printing Office, Washington,  D.C., 20402
                                    98

-------