>
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
------- |