Ulalden
RESEAFOH CORPORATION
FINAL REPORT
IMPROVED CHEMICAL METHODS FOR SAMPLING AND
ANALYSIS OF GASEOUS POLLUTANTS FROM THE
COMBUSTION OF FOSSIL FUELS
VOLUME II
NITROGEN OXIDES
Contract No. CPA 22-69-95
JULY 1971
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Cincinnati, Ohio
toward a better environment...
-------
FINAL REPORT
IMPROVED CHEMICAL METHODS FOR SAMPLING AND
ANALYSIS OF GASEOUS POLLUTANTS FROM
THE COMBUSTION OF FOSSIL FUELS
VOLUME II
NITROGEN OXIDES
Prepared under
Contract No. CPA 22-69-95
by
J. N. Driscoll
and
A. W. Berger
Wai den Research Corporation
359 Allston Street
Cambridge, Massachusetts
June 1971
Prepared for
Environmental Protection Agency
Cincinnati, Ohio
-------
When U.S. Government drawings, specifications, or other data
are used for any purpose other than a definitely related
Government procurement operation, the Government thereby in-
curs no responsibility nor any obligation whatsoever, and
the fact that the Government may have formulated, furnished,
or in any way supplied the said drawings, specifications, or
other data, is not to be regarded by implication or other-
wise, or in any manner licensing the holder or any other per-
son or corporation, or conveying any rights or permission to
manufacture, use, or sell any patented invention that may in
any way be related thereto.
References to named commercial products in this report are not
to be considered in any sense as an endorsement of the product
by the Government.
WALDEN RESEARCH CORPORATION
-------
FOREWORD
The compilation of information contained in this publication was
performed pursuant to Contract No. CPA 22-69-95, Air Pollution Control
Office, Environmental Protection Agency.
The information was compiled by Wai den Research Corporation and
their subcontractor, Arthur D. Little, during the period June 12, 1969
to September 11, 1970.
Volume I of this report examines state-of-the-art manual methods
for sampling and analysis of sulfur oxides. The determination of sul-
fur dioxide (10-3000 ppm) and sulfur trioxide (5-300 ppm) is reviewed
for stationary fossil fuel combustion sources and for control equipment
which may be installed to reduce the emissions from such sources.
Volume II reviews methods for the determination of oxides of nitro-
gen in stationary fossil fuel combustion sources for the concentration
range 5-2000 ppm for nitric oxide (NO), nitrogen dioxide (N02), and
total oxides of nitrogen (NO + N02 or NOX).
Volume III of this report examines state-of-the-art manual methods
for sampling and analysis of carbon monoxide in stationary fossil fuel
combustion sources for the concentration range 10-1000 ppm.
The major pollutants of interest in this study are sulfur and nitro-
gen oxides. The report subsequently includes a literature search, re-
view of methodology and laboratory investigation on the most promising
methods. Carbon monoxide, the minor pollutant, was limited to a litera-
ture search and review of methodology.
This report has been reviewed and approved.
Robert L. Larkin
Project Officer
Process Measurements Section
Division of Control Systems
Office of Air Programs
Environmental Protection Agency
WALDEN RESEARCH CORPORATION
-------
TABLE OF CONTENTS
Section Title Page
1 INTRODUCTION 1
1.1 National NOX Emissions and Method Application
Priorities 2
1.2 Synopsis of Present State-of-the-Art in Determina-
tion of NOX in Combustion Effluents 4
2 THE COMBUSTION EFFLUENT ENVIRONMENT 5
2.1 Introduction 5
2.2 NOX Emission Concentrations 6
2.3 Control Technology 6
2.4 Interferences 11
3 CHEMISTRY OF NITROGEN OXIDES IN FLUE GASES 13
3.1 Gas Phase Equilibria and Kinetics 13
3.1.1 General 13
3.1.2 Equilibria and Kinetics Among the Oxides ... 18
3.1.3 Equilibria and Kinetics Among the Oxides
and Oxyacids 24
3.1.4 Kinetics of Formation of NO 27
3.1.5 Reaction with S02 36
3.1.6 Summary 36
3.2 Reactions of N0§ and N0j> in Solution 36
3.3 Chemistry of Absorption of Nitrogen Oxides in
Aqueous Solutions 39
3.3.1 Gas and Liquid Phase Oxidation Rates of
Nitric Oxide 39
3.3.2 Absorption Rates of the Oxides of Nitrogen
in Aqueous Solution 45
3.3.3 Experimental Studies on Oxidation Times for
NO 53
4 METHODS FOR DETERMINATION OF TOTAL OXIDES OF NITROGEN .. 54
4.1 Introduction 54
4.2 Sampling 55
4.2.1 General 55
4.2.2 Integrated Grab Sampling 58
4.3 Sampling and Analysis Methods Used for Total NOX
Determinations in Exhaust Gases 60
VIi WALDEN RESEARCH CORPORATION
-------
TABLE OF CONTENTS (continued)
Section Title Page
4.3.1 Phenol Disulfonic Acid (PDS) Method 60
4.3.2 Xylenol Isomer Methods 65
4.3.3 Iron (II) Sulfate (and Other Complexes) 68
4.3.4 Spectrophotometric Determination of N02 in
the Gas Phase 71
4.3.5 Modified Saltzmann Method 75
4.3.6 Neutralization Methods 77
4.3.7 Reactive Solid Sorbents 78
4.4 Analytical Methods Used for Determining Nitrate in
Aqueous Solutions 81
4.4.1 Nitrate Electrode 81
4.4.2 1-Aminopyrene 83
4.4.3 Chromatropic Acid 84
4.4.4 Strychnidine 86
4.4.5 Reduction of Nitrate to Nitrite with
Cadmium 87
4.4.6 Diphenylamine-diami nodiphenylmethane 88
4.4.7 Di ami nodi phenyl sulf one 89
4.4.8 Brucine 90
4.5 Analytical Methods Used for Measuring Nitrite Ion
in Aqueous Solutions 91
4.5.1 Griess-Saltzmann Diazotization Coupling
Reactions 91
4.5.2 Other Diazotization-Coupling Reagents for
Nitrite 94
4.5.3 Spectrophotometric Determination of Nitrite
in Aqueous Solutions 97
4.6 Nitrate Electrode Studies 98
4.6.1 Introduction 98
4.6.2 Electrode Assembly 99
4.6.3 Reference Electrodes 99
4.6.4 Electrode Response and Performance 99
4.6.5 Direct Measurements on Solutions Used for
Collection of Nitrogen Oxides 103
4.6.6 Effect of Sulfate on Nitrate Electrode
Response 104
4.6.7 Sulfate Removal 106
4.6.8 Alternate Approaches for Sulfate Removal ... 109
4.6.9 Test Results for the Measurement of NOX in
Flue Gas Samples with a Nitrate Ion Selec-
ti ve El ectrode Ill
VI ii WALDEN RESEARCH CORPORATION
-------
TABLE OF CONTENTS (continued)
Section
Title
Page
5 SAMPLING AND ANALYSIS METHODS FOR THE DETERMINATION OF
NOg AND/OR NO 113
5.1 Introduction 113
5.2 Sampling 113
5.3 Sample Collection and Analysis 114
5.3.1 Solid Sorbents 114
5.3.2 Aqueous Absorbents 115
5.3.3 Kinetic Approach for Simultaneous Measure-
ment of NO and N02 125
5.4 Conclusions 126
6 STATISTICS OF NOV DETERMINATION 128
A
6.1 Introduction 128
6.2 Power Plant Results 128
6.3 PDS Automotive or Diesel Exhaust Results 132
6.4 Comparison of the Precision of PDS to the Precision
of Other Methods for NOx Determination 132
6.5 Conclusions 138
7 RECOMMENDATIONS 139
7.1 NOX Determinations 139
7.2 NO/N02 Determinations 139
7.3 Simplified Methods 140
8 CONCLUSIONS 141
LITERATURE CITED 143
APPENDIX 1 LITERATURE SEARCH 155
APPENDIX 2 SAMPLING SYSTEM FOR S0¥ AND NOY 157
A A
APPENDIX 3 SIMULTANEOUS EQUILIBRIUM PROGRAM DESCRIPTION - DEQL .... 159
APPENDIX 4 LABORATORY STUDIES OF THE CHROMATROPIC ACID METHOD FOR
NITRATE 161
ix
WALDEN RESEARCH CORPORATION
-------
LIST OF ILLUSTRATIONS
Figure No. Caption Page
2-1 Oxides of Nitrogen Concentrations in Gases from
Various Gas-Fired, Oil-Fired, and Coal-Fired Steam
Generators 9
2-2 NOX Emission Concentration as a Function of Furnace
Type (Coal) 10
3-1 Equilibrium Constants of Formation of Nitrogen Oxides
(from the Elements) 14
3-2 Equilibrium Concentrations of Principal Oxides of
Nitrogen 15
3-3 Dissociation of N«0A at Constant Volume N90. v=^
2N02 17
3-4 Summary of Redox Potentials for Nitrogen Compounds ... 37
3-5 Equilibrium Vapor Pressures of NO? and NO over Nitric
Acid 49
4-1 Types of Grab Samplers (partial) 56
4-2 Additional Grab Samplers 57
4-3 Apparatus for Integrated Grab Samples - PHS NOX
Sampling Train 59
4-4 Effect of S02 Concentration on Apparent NOX Concentra-
tion by PDS and Saltzmann Methods 63
5-1 Dilution System for NO and N02 117
5-2 Effect of Moisture on the Absorption Time of N02 in
Caustic Solution 123
6-1 Precision of NOX Determination by PDS Method in Fossil
Fuel Fired Power Plants 134
A2-1 Probe Module 158
XT WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES
Table No. Title Page
1-1 Nationwide Emissions of NO , 1966 2
n '
1-2 Nationwide Emissions of NO , 1967 3
A
2-1 Size Categories for Stationary Combustion Sources 6
2-2 Emission Parameters for Fossil Fuel Combustion 7
2-3 Emissions of Oxides of Nitrogen from Industrial Boilers
and Heaters 8
2-4 Effect of Two-Stage Combustion on NOx Concentrations
from a Large Steam Generator at Normal Full Load 11
3-1 Computer Printout of Equilibrium Concentrations of
Nitrogen Compounds in a Flue Gas at 400°K 28
3-2 Comparison of Nitric Formation Rate Expressions 34
3-3 Comparison of Calculated Values of the Nitric Oxide
Formation Rate Expression of Bartok, et al. (444) and
Bortner and Go!den (599) 35
3-4 Computed Oxidation Times for Typical Flue Gas Samples . 42
3-5 Computed Oxidation Times for Typical Flue Gas Samples
wi th 02 Enrichment , 44
3-6 Comparison of Measured and Calculated NO Levels 53
A
4-1 Solubilities of Common Gases in Water at 20°C 55
4-2 Comparison of NOX Results Obtained by Simultaneous
Application of the ST, BM, CR and PDS Methods 74
4-3 Reported Stoichiometrics for Saltzmann Reagent 93
4-4 Diazotization-Coupling Reagents for Nitrite 96
4-5 Reproducibility of Nitrate Standards 100
4-6 Electrode Response in 0.1N Sulfuric Acid/0.03% Hydro-
gen Peroxide 103
4-7 Effect of Sulfate on Nitrate Electrode Measurement 105
4-8 Effect of Addition of BaHP04 on Electrode Response 107
4-9 Effect of BaCO^ on Measurements with the Nitrate
Electrode ....: 108
4-10 Comparison of Nitrate Electrode Results with PDS 110
4-11- Comparison of Nitrate Electrode and PDS Results for
N0¥ 110
A
4-12 Comparison of PDS and Nitrate Electrode Results for
NO in Combustion Effluents 112
A
WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES (continued)
Table No. Title Page
5-1 Comparison of Initial (5 min.) N02 Concentrations by
the PDS and Nitrate Electrode Analysis for "Dry" Air .. 119
5-2 Effects of Nitrite on the Nitrate Analysis by the PDS
Method 120
5-3 Absorption of N02 in Caustic 122
5-4 Comparison of Calculated and Measured N02 Concentra-
tions for NO + 02 Mixtures 124
6-1 Statistics of the PDS Method for NO in Coal-Fired
Power Plants 129
6-2 Statistics of the PDS Method for NO in Gas-Fired Power
Plants * 131
6-3 Statistics of the PDS Method for NO in Oil-Fired Power
Plants * 133
6-4 PDS Statistics for Automotive Exhausts 135
6-5 Comparison of the Precision of the PDS and Saltzmann
Methods for NO in Diesel Exhausts 136
A
Al-1 Primary Literature Sources Utilized in This Program ... 156
Al-2 Bibliographies 156
WALDEN RESEARCH CORPORATION
XIV
-------
ACKNOWLEDGMENTS
We are pleased to acknowledge the assistance of Robert L. Larkin,
the project leader for NAPCA. We thank Mr. Larkin and James Dorsey for
sharing their experience with the project staff. We are indebted to the
following people and organizations for making available to us results of
unpublished NO analyses:
^
D. J. Callaghan Bay Area APCD
B. Dimitriades Bureau of Mines
H. Lang Bureau of Mines
M. Stengel Combustion Engineering
H. Phillips Foster-Wheeler Corp.
P. Matthew Pacific Gas and Electric Co.
0. Kubik Riley Stoker Corp.
We thank Alfred Thompson of Riley Stoker Corp. for valuable consul-
tations. We also thank Mr. Cirvillo of Air Pollution Technical Informa-
tion Center for the literature search conducted at our request^
WALDEN RESEARCH CORPORATION
AT
-------
PROJECT STAFF
Major technical contributions to this report have been made by the
following:
Maiden Research Corporation
J. Becker
A. Berger
W. Dalzell
J. Driscoll
G. Margolis
Arthur D. Little, Inc.
J. Funkhouser
J. Valentine
XVI i WALDEN RESEARCH CORPORATION
-------
1. INTRODUCTION
This is Volume II of three volumes of the Final Report^on "Improved
Chemical Methods for Sampling and Analysis of Gaseous Pollutants from
the Combustion of Fossil Fuels (in Stationary Sources)." The three vol-
umes are divided by subject as follows:
I Sulfur Oxides
II Nitrogen Oxides
III Carbon Monoxide
This report was prepared under Contract No. CPA 22-69-95 covering the
period June 12, 1969 to September 11, 1970. The principal objectives of
this program were: (1) to review the state-of-the-art for manual
(wet chemical) methods for the determination of emissions of specified
gaseous pollutants from (indirect-fired) stationary fossil fuel sources,
(2) to select and develop improved procedures for high precision emis-
sions determinations, and (3) to select and develop simplified procedures
where high precision determinations were judged to be complex (i.e., re-
quired a highly trained research and development group).
The determination of the oxides of nitrogen, significant in fossil
fuel stationary combustion, is reviewed for the concentration range 5-
2000 ppm for nitric oxide (NO), nitrogen dioxide (NOp), and "total oxides
of nitrogen" (NO + N02 = NOX). We have shown, on the basis of equilib-
rium calculations, that N20 and HNO-, (g) may be present at significant
concentrations at low (i.e., near ambient) temperatures. For normal res-
idence times, the slow kinetics of N02 formation suggest that HN03 will
not be a significant product. Kinetic paths for the low temperature for-
• i
mation of N^O have not been described quantitatively but are believed to
be of minor importance. Where low temperature residence times become
long as a result of the application of control techniques, these species
may become of importance.*
We do not consider methods of analysis for N20 here. HN03 (g) would be
found as NOX in the usual phenol disulfonic acid (PDS) analysis. See
Sections 3, 4, and 7 for further discussions and recommendations.
WALDEN RESEARCH CORPORATION
-------
Since the practical utility of an analytical method is limited by the
fundamental problem of obtaining a representative sample, we have neces-
sarily discussed the principles of flue sampling and the limitations so
introduced to the overall accuracy and precision of a given determination.
General principles of flue gas sampling have been discussed in Volume I of
this report. However, problems specific to sampling for the nitrogen
oxides are included. The combustion effluent environment is treated in
similar fashion.
In the following subsections, we consider method development prior-
ities based upon national NO emissions and summarize the state-of-the-
/\
art in manual determination of NO in combustion effluents.
A
1.1 National NO Emissions and Method Application Priorities
A
A rough guide to method development priorities may be obtained
relative i
sions (Table 1-1).
from the relative contributions of various sources to national NO emis-
A
TABLE 1-1
NATIONWIDE EMISSIONS OF NO. 1966 (521)f
A
10 tons/year
Total from
By fuel
By sector*
fossil fuel combustion in stationary sources
Coal
Fuel oil
Natural gas
Power plants
Industrial combustion
Commercial and residential
6.5
4.0
0.9
1.6
3.5
2.1
0.9
Literature references are listed at the end of the report.
* 6
The "sector" figures given by NAPCA include 0.2 x 10 tons/year gener-
ated by combustion of wood. Arbitrarily subtracting these emissions
evenly from the industrial and commercial and residential sectors
yields the values given in Table 1-1.
WALDEN RESEARCH CORPORATION
-------
In contrast to sulfur emissions, NO emissions are primarily
A
equipment, rather than fuel, dependent (518). (However, both fuel heat
content and organic nitrogen content are among the factors determining
NO concentrations in the effluent.)
A
The categories above do not exclude fossil fuel effluents mixed
with process emissions. A detailed breakdown, which excludes (direct-
fired) process emissions, is available from work conducted at Walden (480)
and is summarized in Table 1-2. As may be seen from the boiler category,
92% of total non-process emissions are attributable to watertube boilers,
for which the smallest size range is about 25,000 pounds of steam per
hour (ca. 3 x 10 Btu/hr). The NO emission problem (as is also true for
A
SO ) is, therefore, centered upon these relatively large sources which re-
X
quite the highest priority for NO method development. Unlike SO , gas-
A A
fired equipment is, here, a significant fuel.
TABLE 1-2
NATIONWIDE EMISSIONS OF NOX, 1967
(excluding process emissions)*
106 tons
Total
By fuel
By sector
By boiler category
i
Coal
Residual fuel oil
Distillate fuel oil
Gas
Utilities
Industrial
Commercial
Residential
(excluding residential)
Watertube > 500,000 pph
Watertube <_ 500,000 pph
Firetube
Cast iron
5.03
3.47
.59
.23
.74
3.42
1.03
.34
.24
2.85
1.55
.24
.15
*
Also excludes stationary internal combustion engines
WALDEN RESEARCH CORPORATION
-------
It does not appear desirable to develop specially adapted high
precision methods applicable to the numerous but relatively unimportant
small sources of NO at this time.
A
1.2 Synopsis of Present State-of-the-Art in Determination of NO in
Combustion Effluents
Almost all reliable determinations of oxides of nitrogen in com-
bustion effluents have been made for total oxides (NO + ML) by the phenol
disulfonic acid method. There are no reliable manual wet chemical analy-
ses available for the N02 concentration in a combustion effluent so far as
we have been able to determine. This situation has arisen as a result of
the widely held, and correct belief that NO is by far the most important
species generated in combustion, and the absence of convenient and accu-
rate analytical methods for the individual oxides.
In the following section we summarize the effluent compositions
from fossil fuel-fired stationary combustion sources. The emphasis is
upon those factors which influence the selection of sampling and analyti-
cal methods for nitrogen oxides.
WALDEN RESEARCH CORPORATION
-------
2. THE COMBUSTION EFFLUENT ENVIRONMENT
2.1 Introduction
A prerequisite for the selection of manual chemical methods for
sampling and analysis of gaseous pollutants from the combustion of fossil
fuels is the description of the environments to be sampled. The combus-
tion products.from all stationary sources such as steam generators, proc-
ess heaters, water heaters, and air (space) heaters are to be considered,
except for those cases where process emissions are combined with the com-
bustion products, e.g., a lime kiln or blast furnace. The measurements
will normally be required at stack conditions; however, they may also in-
clude conditions at the inlet and outlet of control equipment.
The combustion effluent environment has been described in some
detail in Volume I of this report (Section 2 and Appendix 1). For con-
venience, we, here, briefly present the major conclusions from the pre-
vious discussion and include further detail relevant to NO determination.
^
i
For this evaluation, we have divided combustion equipment into
three arbitrary size groups as shown in Table 2-1 in several approxi-
mately equivalent ways. There are relatively few suppliers of equipment
in the large size range (four major sources) with 1968 annual sales of
the order of 75 units (556). On the other hand, there are several hun-
dred suppliers of small units with 1967 annual shipments of approximately
1,500,000 units (557). Useful life is perhaps 15 and 30 years, respec-
tively, for the small and large units. Combining the large number of in-
stalled units with the disparate designs, range of fuels used, load fac-
i i
tors, varying operating practices, etc., it is clear that it is totally
impractical to attempt to specify the environment for individual cases.
Rather, we have reviewed the range of environments reported in the liter-
ature for each of the three size categories shown in Table 2-1 for oil,
coal, and gas. A summary of emission parameters is given in Table 2-2
in consistent units (lb/10 Btu). From these data the various require-
ments for principal factors such as required sensitivity, possible inter-
ferences, transients and temperature regimes may be determined.
WALDEN RESEARCH CORPORATION
-------
TABLE 2-1
SIZE CATEGORIES FOR STATIONARY COMBUSTION SOURCES
Btu per hour
Boiler horsepower
Pounds of steam
Large
>. 5 x 108
>. 15,000
>. 500,000
Intermediate
3 x 105-5 x 108
10-15,000
350-500,000
Small
<3 x 105
<10
<330
per hour
Megawatts >_ 50 <50 MW Not Applicable
2.2 NO Emission Concentrations
A
The primary variable affecting the choice of analytical method-
ology is the concentration of the species sought. As may be seen from
Table 2-2, based upon average emission factors, NO concentrations will
/\
vary over an order of magnitude from small oil-fired equipment to large
coal-fired boilers. Greater variation is clearly to be expected over the
complete range of individual combustion equipment. The range of individ-
ual concentrations observed for industrial equipment is shown in Table
2-3 (93). The extreme range of NO concentrations are observed in large
A
equipment. A comparison of the effect of fuel type on NO emission con-
A
centration has been given by Barnhart and Diehl (464) (Figure 2-1).
Cuffe and Gerstle (431) have determined the effect of equipment
type and load on NO emissions for coal-fired equipment (Figure 2-2). The
A
very high emission rate of the cyclone furnace and the high sensitivity of
NO concentration to load condition for this unit are striking.
A
2.3 Control Technology
In contrast to SO control technology, NO control may best be
A /\
achieved at present, at least in part, by control of combustion condi-
tions. For gas- and oil-fired equipment the two-stage combustion
WALDEN RESEARCH CORPORATION
-------
TABLE 2-2
EMISSION PARAMETERS FOR FOSSIL FUEL COMBUSTION3 (pounds/106 BTU)
Velocity
ft/min
Moisture
Excess Air
Stack Temp.
OF (of
Breeching)
CO
so2
so3
NO
N02
Parti cu-
late
Hydrocarbons
Transients
Large
2000-
3000
15
275-
400
Neg.
4x1 O"4
0.35
0.014
Neg.
10%
GASb
~ Intermed.
18% @ 0% XS Air
1 R°/ 0 ?n# Y^ Ai
t-jfo \y c.U/0 Ao n 1
15-75
400-
750
4x1 O"4
4x1 O"4
0.2 .
0.016
Neg.
of Capacity Per
Small
75
750-
900
4x1 O"4
4x1 O"4
0.1
0.017
Neg.
Min.
Large
(Resid.
2000-
3000-
20
300-
400
3x1 O"4
1.65
0.03
0.68
0.07
0.022
10%
OILC
Intermed.
) (Resid. /Dist.
10% @ 0% XS Air
20-75
400-
750
0.01/0.01
1.65/0.31
0.01 6/3x1 O"3
0.47/0.51
0.15/0.10
0.01/0.01
of Capacity Per
Small
) (Dist.)
75
750-
900
.01
0.31
3x1 O"3
.08
0.06
.02
Min.
Large
2000-
3000
25
300-
400
0.02
3.5
0.8
0.46
(90% cont.
8x1 O"3
5%
COALd
Intermed.
o/o ^ IMUIII i Ma i )
25-75
400-
750
0.1-2
3.5
0.8-0.3
0.53
) (65% cont.)
0.04-0.4
of Capacity Per
Small
75-100
750-900
2
3.5
0.3
0.8
(uncont.)
0.4
Min.
5
TO
O
X
8
TO
TO
a. Details in Appendix 1, Final Report, Vol. I c. sdist. 0.39%, 142,000 BTU/Gal
b. 1100 BTU/SCF Sresid. 1.6%, 152,000 BTU/Gal
d. S = 2.5%, 8% Ash, 13,000 BTU/#
-------
TABLE 2-3
EMISSIONS OF OXIDES OF NITROGEN FROM INDUSTRIAL BOILERS AND HEATERS
00
I
O
m
O
X
O
O
TO
s
;o
O
Small oil heaters Less than 60
Natural gas
Fuel oil
Large refinery heaters 90 to 200
Natural gas
Fuel oil
Small boilers (less than 500 hp) Less than 20
Natural gas
Fuel oil
Large boilers (500 hp and larger) 20 to 90
Natural gas
Fuel oil
Power plant steam generators 200 to 2000
Natural gas
Fuel oil
20 to 100
25 to 137
5 to 92
15 to 387
45 to 149
214 to 282
75 to 320
275 to 600
47
59
33
122
91
258
205
420
0.06
0.33
0.25
0.52
0.14
0.49
0.28
0.62
0.36
0.78
-------
EL SEGUNOO
Figure 2-1. Oxides of nitrogen concentrations in gases from
various gas-fired, oil-fired, and coal-fired
steam generators.
WALDEN RESEARCH CORPORATION
-------
a.
a.
AVERAGE VALUES OF NOX MEASURED
1400 BEFORE AND AFTER COLLECTORS
- 1200
- 1000
- 800
- 600
- 400
- 200
Range
(avq. dev.)
Full
Load
Partial,
Load '
COAL RATES AT 52-66 TON/HR
EXCEPT FOR "SMALL" UNIT RATE
9-10 TON/HR
n
\
S
77\
A
4n
Vv
\x
!
$c
FURNACE
TYPE
o
h-
ce
UJ
o
UJ
o:
UJ,.,
ote^
< uj —
uji<: o
0:0 p
Q-H 5
to w^i
(C
UJ
z
oc
o
o
Z
O
oc
UJ
z
o
Figure 2-2. NOX Emission Concentration as a Function of Furnace Type
(Coal). (Data from Cuffe and Gerstle)
10
WALDEN RESEARCH CORPORATION
-------
technique of Barnhart and Diehl (464) has been utilized to achieve marked
reduction in NO concentration (Table 2-4).
A
TABLE 2-4
EFFECT OF TWO-STAGE COMBUSTION ON NOX CONCENTRATIONS
FROM A LARGE STEAM GENERATOR AT NORMAL FULL LOAD
(Barnhart and Diehl)
NO Concentration Two-Stage Combustion
A
All Air Through Air Through Mn
Fuel Burners Burners +
ppmbyvol. % of Theoretical*
Oil and gas
combined
Oil
525
580
95
90
385
305
27
47
Other techniques for reduction of NO concentrations, by modifica-
A
tion of combustion conditions, such as flue gas recirculation (518), are
under investigation. In general, these techniques do not affect NO con-
A
centrations or potential interferences very radically from the point of
view of analytical determinations.
SO control processes summarized in Volume I of this report are
A
not very efficient for NO control (see Section 3, following) but do af-
A
feet the flue gas environment, notably in reduction of particulate and SOg
content of the effluent. N09/N0 ratios may be significantly altered by
£ A
scrubbing or catalytic oxidation (Monsanto Catox) processes designed for
SOV control. As detailed in Section 3, N^O1 and HNO, (g) species may ap-
A t O '
pear where residence times are increased.
2.4 Interferences
Particulate matter, as discussed in Volume I, is an impediment
to analysis and must be removed by filtration during sampling. Since both
NO and N02 are relatively low boiling materials, adsorption should not be
an important factor, particularly for the principal component, NO.
•] -j WALDEN RESEARCH CORPORATION
-------
Catalytic (heterogeneous) oxidation is of possible importance only where
the individual oxides are sought, and is probably not significant given
the short residence times in both the duct and sampling probe.
The substance of greatest significance for potential interfer-
ence in NO determination is sulfur dioxide. S09, as is shown subse-
X c.
quently, is a direct interference for some analytical methods and a com-
plication in others. Thus, gas-fired equipment, and high efficiency SO
control equipment present interesting possibilities for application of
simplified methods. It should be noted that the PDS* method for NO is
not affected by S02 at levels ordinarily encountered.
Other gaseous species (CO,,, 02> CO, HC1, hydrocarbons, organic
acids, etc.) may produce interferences in the analytical determination of
NOV. Since these effects are specific, they are discussed individually
A
for each method considered.
Phenol disulfonic acid
12
WALDEN RESEARCH CORPORATION
-------
3. CHEMISTRY OF NITROGEN OXIDES IN FLUE GASES
3.1 Gas Phase Equilibria and Kinetics
3.1.1 General
There are seven oxides of nitrogen reported in the liter-
ature, namely:
N20 nitrous oxide
NO nitric oxide
NpO- dinitrogen trioxide
N02 nitrogen dioxide
N20. dinitrogen tetroxide
NpO,- dinitrogen pentoxide
NO, nitrogen trioxide
With the exception of N02> which is red-brown, the oxides are colorless
in the gas phase. The existence of the oxides N03 and N20~ has been
questioned but there is now sufficient evidence that they do exist in
the gas phase. The chemistry and equilibrium of the oxides of nitrogen
are reviewed here to establish ranges and limitations of chemical methods
of sampling and analysis of gaseous pollutants from combustion sources.
The temperature and concentration ranges are far different from those of
nitric acid manufacturing (602) so that the oxides of importance may be
different in a flue gas environment.
The values of the equilibrium constants for the formation
of the nitrogen oxides from the elements over the temperature range 400-
2000°K are presented in Figure 3-1 (488). The three most stable oxides
2
are NO, N02> and N20 but all are in low concentrations (< 10 atm) at
equilibrium in flue gases. For a flue gas containing 0.8 atmosphere N2
and 0.03 atmosphere 02, the concentrations (in atmospheres) of these
three nitrogen oxides are shown in Figure 3-2. Since the total concen-
tration of nitrogen oxides is very small, the N2 and 02 levels are essen-
tially unchanged by reaction and each oxide species concentration was
13 WALDEN RESEARCH CORPORATION
-------
Fig. 3-1
Equilibrium Constants of Formation
of Nitrogen Oxides
(from the Elements) (488)
/2000°K\
\3500°R/
/1000°K\
\1800°R/
x 10
/ 600°K\
\1080°R/
+4, 0K'1
HO
720°R/
14
WALDEN RESEARCH CORPORATION
-------
V)
fc.
O)
10
c
o
to
i-
«
o
c
o
-------
calculated independently. At high temperatures, the only oxide of sig-
nificance is NO which is present in concentrations > 1000 ppm at tempera-
tures above 1600°K. Over the range of combustion temperatures, N02
reaches a few parts per million (^ 5 ppm at 2000°K).
As the combustion gases are cooled the equilibrium con-
centrations of the nitrogen oxides are decreased but the actual concen-
trations may remain high because of kinetic limitations. For instance,
the rates of decomposition of NO and N0? to the elements are slow so that
concentrations of nitrogen oxides in excess of equilibrium at sampling
temperature are expected. The rate of formation of NO may be signifi-
cantly increased by reaction of oxygen with bound nitrogen in the fuel.
Little quantitative work has been done (567,568) but nitrogen in fuel
(e.g., coal) is apparently of major importance in the rate of NO produc-
tion at lower temperatures (569).
The nitric oxide in the combustion products can react
with excess oxygen to form NOp as the gases are cooled and the N0« may
then dimerize to N204 at low temperatures (25-300°C) (Figure 3-3). How-
ever, in combustion gases the concentration of N,^ is very small at all
temperatures. Further, NpO may be present since it is the next most
stable oxide other than NO and NO,,. The dinitrogen trioxide is not pres-
ent at appreciable concentrations even at room temperature but forms
rapidly from NO and N02 and is the oxide species which is absorbed to
form nitrous acid. The higher oxides, NpCL and N03, are not present in
significant concentrations.
Besides the oxides of nitrogen, reaction products of ni-
trogen oxides and water vapor (HN03, HNOp) may be present. If liquid
water is present, these exist in both the liquid and gaseous phase.
Nitrous acid vapor is present at very low concentrations.
The oxides are discussed in the following sections in
more detail, including significant kinetic and equilibria calculations.
Equilibria including water are discussed in Section 3.1.3.
See Section 3.1.3 for equilibria including H?0.
WALDEN RESEARCH CORPORATION
-------
o
.oo
.36
.34
.32
.30
.28
.26
.24
.22
.20
.18
.16
.14
.12
.10
.08
.06
.04
.02
.00
C
i
1
FIGURE 3-5 C
DISSOCIATION OF N204 AT
*** *5^)
CONSTAN
VOLUME N204^ 2N02
LEGEND:
• * Equilibrium Calculations
X * Experimental IR Data
Total Nitrogen IV Oxides - 3.51 Mole
% NO 2
'
\
\
) 10 20 30 40
\
\
\
k
\
v
X
^*N<-;
i
- r
H
T
2319
\f
50 60 70 80 90 100 110 120 13
TEMPERATURE, °C.
17
WALDEN RESEARCH CORPORATION
-------
3.1.2 Equilibria and Kinetics Among the Oxides
3.1.2.1 Nitric Oxide, NO
Nitric oxide is the only oxide of nitrogen pres-
ent in significant concentrations at flue gas temperatures. Thermodynamic-
ally, the oxide is very unstable and should dissociate to the elements at
all temperatures encountered in a collection or sampling system. The
equilibrium for the formation of nitric oxide:
1/2 N2 + 1/2 02 = NO
has an equilibrium constant
PNO
Kp =
(po2)
'/2
which varies with temperature as follows (488):
Temperature °K LoglQ Kp
298 -15.171
500 - 8.783
1000 - 4.062
1500 - 2.487
2000 - 1.699
2500 - 1.227
3000 - 0.913
However, the rate of dissociation is extremely slow:
91.600
l= 13"
xl013e" RT (NO)
where T is in °K and the rate is expressed as g moles/cc-sec (444). At
temperatures below 1500°K this rate is negligible at the low concentra-
tions of nitric oxide normally present in stack gases. A more detailed
discussion of NO formation and decomposition is given in Section 3.1.4,
Kinetics of Formation of NO.
18 WALDEN RESEARCH CORPORATION
-------
Further, the reaction of NO with other gases (444)
such as \\2 and hydrocarbons is quite slow. However, NO can react with Op
to form N02 at slow but significant rates.
3.1.2.2 Nitrogen Dioxide, N02
Nitrogen dioxide, like NO, is an unstable oxide
and thermodynamically should decompose to the elements. However, the rate
is negligibly slow. At low temperatures, the nitric oxide present in
stack gases can react with oxygen to form N02 by the reaction:
NO + 1/2 02 = N02
The equilibrium constant for this reaction
K.
is given below (488):
Temperature °K Log,Q K
298 6.194
500 2.114
1000 -0.938
1500 -1.951
2000 -2.453
Thus, at temperatures below about 1000°K, N02 is more stable than NO and
the nitric oxide present in stack gases samples will form N02 provided
enough time is available. Near room temperature (up to 500°K) essentially
all of the NO will form N02 (at equilibrium).
The kinetics of the reaction of NO with 02 is
unusual in several ways. The reaction proceeds homogeneously in the gas
phase at a measurable but slow rate. It goes faster at lower temperatures
and proceeds by termolecular kinetics, with a rate proportional to the
square of the NO partial pressure times the partial pressure of 02.
19 WALDEN RESEARCH CORPORATION
-------
dlPN02)
~— = k
Expressing the partial pressures in atmospheres and time in seconds, the
value of k found by Bodenstein and reported by Chi 1 ton (602) is
Iog10 k = - 0.725
where T is in °K.
In order to gain some insight into the times re-
quired for oxidation, the above rate expression can be integrated. The
time in seconds is (602):
t =
where p refers to the initial partial pressures of the species in atmo-
spheres, x is the fraction of the initial NO converted to N02 and t is in
seconds. For a stack gas containing 1000 ppm NO and 3% Op, at 300°K, 10%
of the NO is converted to N02 in 144 seconds and 50% is converted in 1283
seconds, or about 21 minutes. Alternatively, one may say that only a small
amount (about 1-2 ppm) of the NO is converted to N02 in 10 seconds.
At 450°K, the rate constant, k, is 5, so that
the times involved would be four times longer than at 300°K.
Thus, the NO in a stack gas sample maintained at
or slightly above room temperature will be slowly converted to N02; if
several hours are allowed most of the NO is converted, but in seconds,
essentially none is converted.
3.1.2.3 Dinitrogen Tetroxide, N20,
The nitrogen dioxide present in a gaseous mix-
ture may dimerize to the dinitrogen tetroxide, N20,. The rate of the
reaction
20 WALDEN RESEARCH CORPORATION
-------
N02 = 1/2 N204
is very rapid in both directions so that N02 is in equilibrium with the
dimer. For the above reaction, the equilibrium constant is:
(PN o )V2
Kp = PN02
and has'the following values as a function of temperature (488):
Temperature °K Loglo \>
298 0.417
400 -0.851
500 -1.584
For a gas which contained initially 1000 ppm NO all of which was oxidized
to N02 only a small portion of the oxide would be in the dimer form. For
example, at 298°, 1C = 2.61 and the N20^ would be about 6.6 ppm and the
N02, 987 ppm. The fraction as N20^ is small but may be significant in
collection since it is the dimer which is rapidly absorbed in water to
form nitric acid, releasing nitric oxide (see Section 3.3).
3.1.2.4 Dinitrogen Trioxide, N20g
The dinitrogen trioxide, which is the anhydride
of nitrous acid, is formed by the reaction of NO and N02 in the gas phase.
The equilibrium constant for this reaction is:
,, . V3
P PNO PNOZ
and has the following dependence on temperature (488):
WALDEN RESEARCH CORPORATION
-------
Temperature °K
298 -0.278
400 -2.087
500, -3.139
where the oxides NO and N02 are at low concentration (<_ 1000 ppm), the
total amount of N203 is very small. For example, if N02 and NO are
present in equimolar quantities of 500 ppm, then p., n = 0.132 ppm.
N2U3
The dinitrogen trioxide is still of importance
since it is the species which is absorbed in water to form nitrous acid
(HN02) from an equimolar mixture of NO and N02- The nitrogen oxides, NO
and N02, react rapidly to form N203 which is then absorbed in the water.
Obviously, at low oxide concentrations, the rate of absorption will be
low because of the extremely low equilibrium partial pressure of N20~.
3.1.2.5 Nitrogen Trioxide and Dinitrogen Pentoxide, N03
and N20j-
The concentrations of N03 and N20r are extremely
low under all conditions of interest and need not be considered further.
3.1.2.6 Nitrous Oxide, N20
Nitrous oxide, N20 (laughing gas) is a non-toxic
oxide of nitrogen which supports combustion almost as well as oxygen.
The equilibrium constant for the reaction:
N02 = 1/2 N20 + 3/4 02
is
.3/4
2
K_ = *
P PN02
This equilibrium constant varies with temperature as follows (488):
22 WALDEN RESEARCH CORPORATION
-------
Temperature °K
298 -0.147
400. +0.215
500 +0.438
The equilibrium constant is approximately unity over the range of in-
terest in sampling and analysis. For a gas which originally contained
1000 ppm NO and 3% 02 at 298°K and was allowed to react to N02 and N20,
the equilibrium concentration of N02, 02 and N20 would be:
Compound Concentration (ppm)
NO < 1
N02 854
N20 73
02 29,790
At 400°K (starting with the same initial mixture of gases), the gaseous
species in equilibrium would be:
Compound Concentration (ppm)
NO < 1
N02 616
N20 192
02 29,790
Provided there is a kinetic route to N20 there will be a significant frac-
tion of the oxides in the form of N20. At these low temperatures it is
unlikely that a homogeneous gas phase route is possible. However, some
of the older preparations of N20 suggest a,possible route. Priestly (604)
formed N20 by passing NO over moist iron fillings. Also, nitrous oxide
can be found in the presence of sulfur dioxide by the reaction (604):
V
2NO + S02 + H20 = H2S04 + N20
Each of these mechanisms is possible in a sampling unit or in the stack
system of a combustion device. Recent work on either of these reactions
appears to be lacking.
23 WALDEN RESEARCH CORPORATION
-------
3.1.3 Equilibria and Kinetics Among the Oxides and Qxyacids
In addition to the oxides of nitrogen, there are two
oxyacids, nitrous acid, HNOp and nitric acid, HNO-. Both may be pres-
ent in the gas phase or in aqueous solution.
3.1.3.1 Nitrous Acid, HN02
Nitrous acid is formed in aqueous solutions when
N203 dissolves. Nitrous acid is unstable in solution and decomposes to
nitric acid and nitric oxide unless the solution is cool and alkaline.
3HN02 + HN03 + 2ND + H20
Nitrous acid can exist in the vapor state in equilibrium with the oxides
of nitrogen and water vapor. The equilibrium constant for the reaction:
NO + N02 + H20 = 2HN02
is
(PHNO?)
it = <-
P (Pun)
NO}(PN02)(PH20)
and varies with temperature as follows (488):
Temperature °K LoglQ Kp
cis(HNOo) trans (HNOJ
298 -1.212 -0.496
400 -2.789 -2.257
500 -3.718 -3.292
At higher temperatures, nitrous acid vapor decomposes to the oxides and
water vapor. Near room temperature a calculation of the amount of HN02
vapor present at equilibrium requires that the oxides NO, N02, N20., as
well as nitric acid vapor, HN03, be determined simultaneously.
24 WALDEN RESEARCH CORPORATION
-------
3.1.3.2 Nitric Acid, HN03
Nitric acid exists in the vapor phase and dis-
solves in water up to 68 wt %. Decomposition of nitric acid is very
slow and several days time is required to approach equilibrium at room
temperature. The equilibrium composition of HNOo vapor can be calculated
from the equation:
H20 + 2N02 + 1/2 02 = 2HN03
for which the equilibrium expression is:
("MNP
Kp =
Nitric acid vapor is stable at room temperature but decomposes at higher
temperatures as the values in the following table show (488):
Temperature °K Log,Q K
298 +3.838
400 -0.334
500 -2.784
The presence of other gases (NO, N02> N20. and HNOJ must be accounted
for in doing the simultaneous equilibrium calculations.
Nitric acid is normally formed in aqueous solu-
tion by passing a gas containing N02 through water. Each three moles of
nitrogen dioxide forms two moles of HNO, and a mole of NO, which must be
i J
reoxidized to N02 in the gas phase:
3N02 + H20 = 2HN03 + NO
Determination of the equilibrium of the oxyacids in the vapor require the
calculation of several simultaneous equilibria. Several specific examples
will be considered.
25
WALDEN RESEARCH CORPORATION
-------
For a flue gas mixture at one atmosphere and
298°K which originally contained NO, (L, H«0 and N2 at partial pressures
0.001, 0.03, 0.08 and 0.889 atmospheres, respectively, the amount of HN02
vapor can be estimated. Under these conditions most of the NO will be
converted (at equilibrium) to N02 with a small amount as N^O (see N^O
discussion). Consequently, the partial pressure of HN02 will be neg-
ligibly small because the partial pressures of the species on the left
side of the equation:
NO + N02 + H20 = 2HN02
are small (the amount of NO is extremely small). For example, using the
concentrations developed in the discussion of N20 and a value 0.1 for the
(average between cis and trans) equilibrium constant of the above reac-
tion forming HN02 and setting the partial pressure of NO at 10 (high
estimate) the partial pressure of HN02 is less than 1 ppm. Further
nitric acid vapor will consume some of the N02 so that this is a high es-
timate for HN02> Similarly, for other temperatures and similar concentra-
tion ranges the amount of HNO? in the vapor can be neglected in sampling.
However, the amount of HNO. will be far more im-
portant if equilibrium is assumed among the oxides and oxyacids since the
equilibrium constant for its formation from N02, hLO and 02 is very favor-
able at low temperatures. For instance, at 298°K the equilibrium composi-
tion of the nitrogen oxides and oxyacids at one atmosphere pressure from
a mixture originally containing 1000 ppm NO, 3% 02 and B% H20 is approxi-
mately:
Compound Concentration (ppm)
HN03 905
N02 94
All other fixed nitrogen species are negligible. As in all of the pre-
vious calculations, decomposition of any of the species to the elements
has been neglected.
26 WALDEN RESEARCH CORPORATION
-------
A more interesting calculation is one in which
the above initial mixture is allowed to reach equilibrium at 400°K and
1 atm. The HN03 is less stable so that all three species, N02> H20, and
HN03 are present in significant concentrations.
The computer program DEQL (see Appendix 3) was
utilized to conduct these equilibrium calculations. The results are shown
in detail in the computer printout, Table 3-1; the last four (data) lines
contain the species and concentrations. Major nitrogen-oxygen species
concentrations are summarized below.
Compound Concentration (ppm)
N02 589
HN03 (g) 47
N20 182
NO 0.8
Provided kinetic routes exist for convertine N02 to N20 and HN03, a sig-
nificant error could be introduced in stack analysis by the presence of
these species. Dei 1 in (449) has reported small amounts of N?0 in stack
gases.
3.1.4 Kinetics of Formation of NO
3.1.4.1 NO Kinetics
The concentrations of nitrogen oxides in stack
gases reported in the literature are below those calculated for equilib-
rium at flame temperature but above those in equilibrium at stack gas
temperature indicating that kinetics are important. The kinetics and
mechanism of NO formation are not simple but are well established and
are in close agreement with the model and results originally published
and discussed in detail by Zeldovish (448). The major results and con-
clusions of his work are discussed here.
(a) The rate of NO formation (except from ni-
trogen already fixed in the fuel) proceeds so much less rapidly than the
27 WALDEN RESEARCH CORPORATION
-------
TABLE 3-1
Computer Printout of Equilibrium Concentrations
of Nitrogen Compounds in a Flue Gas at 400°K
>fildef 2 dsk t400 data
K; T=0.26 17.25.46
>$ deql
EXECUTION BEGINS... .
GAS COMPOSITION FOR NO,H20,02,N2 1
T=400. KELVIN' TOTP= 1.000
FEED N0= 0.0010 02= 0.0300
H20= 0.0800
N2= 0.8890
REACTI
1
2
3
4
b
6
•7
8
GAS CO
81
ON LOG OF K E
3.6290
-0. 1660
0.4300
-0.8510
-^3.0760
-5.1330
-5.7160
-9.669.0-
MPOSITICN WITH
CYCLES
EQUILIBRIUM CONSTANT
0.426E 04
0.682E 00
0.269E 01
0. 141E 00
0.838E-03
0.735E-05
0. 192E-05
0.213E-09
ERROR IN K MOT
4 TEMPERATURE
MORE THAN, 0.00
400. KELVIN
ITERATIONS
I
1
2
3
4
5
6
~7
8
ABS VALUE (KCALC/K -1.0)
0.954E-06
0.939E-03
0.858E-05
0.987E-03
0.'793E-03
0.668E-03
0.519E-05
0.808E-03
1
SUM PARTIAL PR ESSUR£S= 1.00
SUM 0= 0.141
PNO PN02 P02
SUM N =
PH20
0.001
PHN03
SUM H= 0.160 '
PN20
0.8C1E-06 0.589E-03 0.293E-01 0.800E-01 0.472E-04 0.182E-03
PN204
PHN02
PN 205
PN203
PN03
PN2
0.687E-C8 0.336E-06 0.439E-12 0.385E-J1 0.217E-13 0.389E 00
IHCC02I STOP' 0
R; T=3.12 17.27.00
28
WALDEN RESEARCH CORPORATION
-------
combustion reactions that it takes place after the combustion process is
complete and is purely thermal in nature. Several corollaries follow:
the rate of NO formation does not depend on the chemical nature of the
fuel but on its heat of combustion; the rate depends on the concentration
of N2 and 02 remaining after combustion (excess 02); preheating the fuel
or air is equivalent to increasing the heat of combustion.
(b) The amount of NO formed is always less than
the equilibrium amount at peak flame temperature due to the finite rate
of the forward reaction and to the reverse reaction NO -»• 1/2 N« + 1/2 02
as the combustion products cool .
(c) The activation energies of NO formation
and dissociation are very large, 129 and 86 kcal, respectively.
(d) The forward reaction proceeds via a chain
mechanism. Oxygen is in equilibrium with oxygen atoms by the reaction:
02 + M j 02 + M
where M is any "third body." The oxygen atom then participates in the
exchange reactions.
k3
0 + N2 j NO + N
k4
k
5
k,
N + 02 j NO + 0
The reaction between oxygen atom and nitrogen molecule is rate limiting
and has an activation energy of 68 kcal. By assuming that 0 and N con-
centrations reach steady state, Zeldovich predicted a rate expression:
29 WALDEN RESEARCH CORPORATION
-------
, (0,,) ,
- k(0) 1 k(NO) iV5<°2> -
52
where K, is k,/k2> the equilibrium constant for oxygen dissociation. Fur-
ther, k^ and k& are of the same magnitude so that if ((L) » (NO), the
rate expression becomes:
where 1C, is the equilibrium constant for the overall reaction
N2 + 02 + 2ND
(e) The rate expression then becomes, numerically
.86.000 43
RT
RT ,„ WA ) _ (N0)2l
where t is in seconds, and concentration in gram moles per liter. This
expression was tested over a wide range of concentrations including very
low oxygen levels where the proposed chain mechanism might be questionable
and over the temperature range 2000-3000°K.
3.1.4.2 Discussion of Zeldovich Kinetics
Both the forward and reverse reactions have very
large activation energies, 129 and 86 kcal , respectively. Consequently,
the rates of both reactions become appreciable only at high temperatures,
i.e., at 2000°K
129,000
e" RT =e-32.6
For comparison, the same value of the exponential term would be obtained
for a reaction with an activation energy of 20 kcal at 323°K (50°C). At
30 WALDEN RESEARCH CORPORATION
-------
about 1300°K the reverse reaction "freezes," i.e., the reaction is so
slow that for residence times of a few seconds, no appreciable change in
NO concentration occurs. Since the reactions proceed rapidly only at
high temperatures, the temperature sensitivity of both reactions is rela-
tively small despite the large activation energies. The ratio of the re-
action rates at two temperatures, T, and T^, is approximately
EAT
X
where AT is T2-T1 and T is the average of TI and T2. Thus, at 2000°K for
an activation energy of 129 kcal, the reaction rate doubles with an in-
crease in temperature of 425°C. (Equivalently, the rate changes 17.7%
for a 10°C change in temperature.) The reverse reaction is less tempera-
ture sensitive and an increase of 64°C will double the rate at 2000°K.
The rate of formation of NO depends on the square
root of the oxygen concentration remaining after combustion (excess oxygen)
so that a doubling of the oxygen concentration changes the forward rate by
1.414. At peak flame temperatures (^ 2200°K), an increase in temperature
of about 26°C would be required to increase the rate an equivalent amount
(41.4%). At lower temperatures a smaller change in temperature is needed
for the same percentage change in rate.
Additional factors to be considered in NO kinet-
ics are the rate of cooling of the combustion gases and the maximum flame
temperature. At low flame temperatures (low heat of combustion, no pre-
heat, diluent added, etc.) and with rapid cooling, the amount of NO formed
will be low and far less than the equilibrium yield at flame temperature.
The reverse reaction can be neglected and rapid cooling of the combustion
gases will "freeze" the forward reaction. At high flame temperatures and
with slow cooling, the NO yield is again low. Large amounts of NO are
produced initially but the reverse reaction occurs as the gases cool and
the final yield of NO, determined by the temperature at which the reverse
reaction freezes, is very low. Obviously, in this case the NO level is
independent of maximum flame temperature. Between these two limits the
31 WALDEN RESEARCH CORPORATION
-------
nitrogen oxide yield goes through a maximum. Zeldovich showed this max-
imum to be 60-70% of the equilibrium yield at peak flame temperature in
his system,
An accurate calculation of NO produced in a par-
ticular combustion device depends on a knowledge of the temperature-time
and concentration- time (N2, excess 02) history of each parcel of the post-
combustion gases. With this information, the nitric oxide rate expression
can be integrated to find the yield of nitric oxide. Generally, one does
not have the desired spatial and temporal distribution of temperature and
concentration and approximate (e.g., time average, plug flow) profiles
are assumed.
The activation energies recommended by Zeldovich
were 129 ± 10 and 86 ± 10 kcal for formation and dissociation of NO, re-
spectively. His data showed values of 125 ± 10 and 82 ± 10 but he raised
the average values to account for non-uniform temperature distributions
and to simplify his later calculations. The difference between the acti-
vation energy for the forward and reverse reaction is ^ 43 kcal since the
equilibrium constant for the reaction
N + 0 + 2NO
is
PN P
N
43,260
- 91 i9 0" RT
21.12 e
More recent data (591-598) generally confirm the
mechanism and kinetics of Zeldovich. However, the activation energies re-
ported are higher than those of Zeldovich at high temperature (2000-3000°K)
but less than those of Zeldovich at low temperature (1 300-1 600°K). For
example, Glick, et al. (593) and Camac and Feinberg (591) report the acti-
vation energy of NO decomposition to be ^ 92 kcal between 2300 and 3000°K
whereas Kaufman and Kelso (594) and Wise and Freeh (596,597) report values
32 WALDEN RESEARCH CORPORATION
-------
of 63.8 and 82 kcal , respectively, at temperatures around 1500°K. Most
of these investigators report a ±5 kcal uncertainty associated with the
activation energies. However, many stationary combustion devices oper-
ate between 1600 and 2000°K necessitating an extrapolation of the kinetic
data with temperature. Although the data of the several investigators
are quite close (about a two-fold variation) within the low (< 1600°K)
and high (> 2300° K) temperature ranges, the activation energies reported
will greatly affect the extrapolation into the intermediate temperature
range of interest. Bortner and Golden (599) reviewed and correlated much
of the data available through 1960 on reaction rate constants for species
in high temperature air. They fitted curves through the data of several
investigators between room temperature and 7000°K for the individual re-
actions involved in nitric oxide formation and decomposition. Their
curves yield a forward rate expression for nitric oxide formation:
' 135,700
" RT
where t is in seconds and concentrations in gram moles per liter.
Several of the reported rate expressions for
nitric oxide formation are given in Table 3-2 with the temperature range
over which the data were collected. In each case the rate expression is
evaluated numerically at 2200°K to show the magnitude of variations that
are to be expected upon extrapolation of these kinetic expressions.
The expressions of Bartok, et al. (444) and
Bortner and Golden (599) (see Table 3-2) agree closely at 2200°K. These
rate expressions are based on recent data of several investigators on the'
individual steps in the chain mechanism. It is unlikely that data for
combustion equipment could be used to select one of these two expressions
For a recent summary of the rate constants of the elementary steps in
nitric oxide formation as well as those involving CO, C02» 0^ and H20,
see reference (600) and for a very complete list and discussion, ref-
erence (601).
33
OJ WALDEN RESEARCH CORPORATION
-------
TABLE 3-2
COMPARISON OF NITRIC FORMATION RATE EXPRESSIONS
CO
1
o
m
z
m
0
8
1
30
Investigator
Camac & Feinberg (591)
Click, et al. (593)
Bartok, et al . (603)
Zeldovich (448)
Bartok, et al. (444)
Bortner & Golden (599)
Freedman & Daiber (592)
Kaufman & Kelso (594)
Temperature Range °K
2300-3000 7 x
2000-3000 9 x
2000-3000 3.32
Recommended for sta- 3 x
tionary combustion
Best fit 300-7000°K 7 x
3000-4300 1.01
1300-1600 5.49
*
The rate constant is that value of k in the expression
five investigators (591,593,448,444,599), but is k in
last two citations (592,594).
Forward Rate Constant
,«14 / 134,300\
in ovn l_i . « .' T_. _ 1
IU CAJJ 1- py 1
1014 CXD / 135'000\
iu exp i- DT i
\ Kl /
x 1014 CXD / 129»000\
A IU CA^J 1 W-p 1
\ K' /
1014 CXD / 135'000\
1 U UAp 1 ~ PT J
10 i /o / ioc 7nn\
m 2 T1/2 rxn 1 135>700|
I U 1 UAp 1" n-r 1
1Q25 T-3/2 / 128,00
A 1 U 1 UAfJ 1™ DT
\ /
+ ^NOl=k(N2)(02)V2gr
Forward Rate Constant*
at 2200°K
29.4
31.05
46.1
10.35
10.80
^ i RI /in
1 O 1 fU
1193
moles/cc-sec for the first
)(02) gr moles/cc-sec for the
-------
over the other. The two expressions are compared numerically at three
temperatures (1000, 2200, and 3000°K) in Table 3-3, The expressions are
close except at temperatures where the rate is insignificantly small.
TABLE 3-3
COMPARISON OF CALCULATED VALUES OF THE NITRIC OXIDE
FORMATION RATE EXPRESSION OF BARTOK, ET AL. (444)
AND BORTNER AND GOLDEN ~^
Investigator 1500°K 2200°K 3000°K
Bartok, et al. (444) 6.32 x 10"6 10.35 5.88 x 103
Bortner and Daiber (599) 2 x 10"6 10.80 6.85 x 103
Values reported are for k in the rate expression:
= k(N)(0)1/2 gr moles/cc-sec
The values reported by Freedman and Daiber (592)
and Kaufman and Kelso (594) are expressed for a bimolecular forward reac-
tion. Obviously their values cannot be compared exactly to the other ex-
periments since oxygen concentration affects the comparison. For a bi-
molecular rate constant of 5000 (about midway between those in Table 3-2)
and for a chain mechanism rate constant of 10 [close to Bortner and Golden
(599)], the oxygen partial pressure would have to be 0.1385 atm at 2200°K
for the two rates to match exactly.
The rate of decomposition of NO to 02 and N2 is
(444):
91,600
_dMl=4.lxl013e~ RT (NO)2
WALDEN RESEARCH CORPORATION
-------
3.1.5 Reaction with SO,,
A further complication in the kinetics and equilibrium of
stack gases arises when S02 is present from the combustion of sulfur bear-
ing fuels. The S02 will react with N02 to form S03 and NO
NO, + SO, = SO. + NO
£ C. O
The kinetics of this reaction are rapid (452,605), but reduction of N02
in flue gas samples by this reaction is slow since the concentrations of
N02 and S02 are small. However, the rate of formation of NO by this re-
action may well be faster than its rate of reoxidation by 02.
i 3.1.6 Summary
Of the seven oxides and two oxyacids, only three of the
oxides (N20, NO, N02) and one of the oxyacids (HN03) need be considered
in flue gas analysis. At high temperatures (> 1000°K), only NO exists
in significant concentrations; at lower temperatures (25-400°K), all four
species may be important. Kinetic limitations prevent the dissociation
of these species to the elements. However, the three oxides and HN03 may
react slowly with each other and with 02 and H20. Generally, only NO is
expected to be present in significant amounts in stack samples at low
temperatures since reaction rates are slow, but if long times (minutes to
hours) are allowed, significant concentrations of all of the above species
may be generated.
3.2 Reactions of NOl and NOp in Solution
The redox potentials for various nitrogen couples in acid and
basic solutions are given in Figure 3-4. Hydrolysis of the nitrogen
oxides NO, N02, and N203 in aqueous solutions leads to the production of
nitrite and nitrate ions.
The small positive standard free energy (AF°) for the reaction:
2HNO, + NO + NO, + H,0, AF° = 2.06 kcal
L Li.
35 WALDEN RESEARCH CORPORATION
-------
Figure 3-4
Summary of Redox Potentials For Nitrogen Compounds (479);.
Add Solution
+0.23
-1.275
0.05
-1.29
-1.41
1.87 -1.77 -1.59 -0.990 -1.07'—^-0.79
NH+-
-N2H+ NH3OH+ N2 N20 NO-
1-2.65 -0.711
-H2N2Or
-1.35
-HN02 N204 NO?
-0.496
-0.86
-0.94
Base Solution
1.16
-0.1
1.05
-0.15
-0.73
3.04 -0.94
NH4OH N2H4 NH,OH N, N2
I -1.52
-0.42
-0.76 0.46 I -0.88 0.86
-NO NO! N»0«
-0.18 I
0.76 || . 0.14
-0.01
37
WALDEN RESEARCH CORPORATION
-------
reveals that there is an appreciable concentration of NO and N02 in equi-
librium with the condensed phase (aqueous nitrous acid). A more important
reaction is the disproportionate of nitrous acid:
3HN02 + H+ + NO^ + 2ND + H20, AF° = -3.22 kcal
The rate of the above reaction is slow in cold dilute HN02 solutions but
becomes rapid if the solution is warmed.
The oxidation of nitrous acid to nitric acid (Figure 3-4) in
acid solution requires a powerful oxidizing agent such as hydrogen per-
oxide or permanganate, since the oxidation potential is very high.
H20 + HN02 + N0~ + 3H + 2e", E° = -0.94
Thermodynamically, nitrite may be oxidized in alkaline solution
by comparatively weak oxidizing agents. The half reaction for the oxida-
tion of nitrite is:
N02 + 20H" + NO" + H20 + 2e~, Eg = -.01*
The potential for oxidation by, for example, oxygen is, therefore, favor-
able (in alkaline solution).
NO" + 1/2 02 •* N0~, E° = 0.39
Although there are contradictory statements in the literature on
the ease of this reaction, slow kinetics prevent appreciable reaction (see
5.3.2 for discussion). The oxidation rate of nitrite by hydrogen peroxide
has been reported to be directly proportional to the hydrogen ion concen-
tration, and thus decreases rapidly as solution pH is increased (613,614).
*
Potential with respect to the hydrogen reference electrode in 1M OH ,
E° = 0.828.
38
WALDEN RESEARCH CORPORATION
-------
The reduction of nitrous acid by sulfur dioxide appears to be
another example of slow kinetics. This reaction goes in acidic but not
in basic media (479,613) although the thermodynamics of the reduction
(to, say, hyponitrous acid) are favorable in both media.
In summary, collection of oxides of nitrogen in basic media ap-
pears to be effective kinetically in inhibiting many common oxidation and
reduction reactions and thus preventing interferences. It is, however,
not clear that inhibition will necessarily occur in all cases, since
thermodynamically favored slow reactions are subject to both homogeneous
and heterogeneous catalytic acceleration.
3.3 Chemistry of Absorption of Nitrogen Oxides in Aqueous Solutions
Sampling procedures presently utilized in the determination of
NO may involve the following several major processes:
A
(1) oxidation of NO in the gas phase
(2) oxidation of NO in the liquid phase
(3) absorption of N0« and NO by aqueous solutions
In this section, we analyze the above reactions, insofar as pos-
sible, to permit a logical approach to the improvement of sampling pro-
cedures.
3.3.1 Gas and Liquid Phase Oxidation Rates of Nitric Oxide
N02 (or N20,) has a far greater solubility in aqueous
solution than NO (571,572). Upon absorption, N02 reacts with water to
produce nitric and nitrous acids. The latter species decomposes readily
in acidic solutions unless it is oxidized to nitric acid. On the other
hand, little has been reported about possible reactions following absorp-
tion of pure NO in aqueous solution. There is evidence, however, (571,
573) that NO is absorbed rapidly in the presence of N02> in effect as
N203, yielding nitrous acid. The data of Eagleton, Langer and Pigford
(574) suggest that the rate of absorption of N203 is as much as ten times
faster than that of N02 in 46% NaOH solution. This may be due to greater
39 WALDEN RESEARCH CORPORATION
-------
solubility of NgO.,, faster liquid phase reaction rates, or, possibly
both.
On the basis of the above brief description, it is con-
cluded that even where liquid phase oxidation is employed, gas phase
oxidation of NO must play a major role in the rate of overall absorption
since the most likely absorbing species is either ML and/or N^O^.
If the rates of absorption are rapid compared to the
oxidation reaction rates, it is possible to estimate the minimum times
necessary for the overall process of oxidation and absorption. This as-
sumption is a fairly good one (see Section 3.3.2) since commercial ab-
sorption units used in the production of HN03 are designed (575) on this
basis.
The gas phase oxidation of NO is a third order reaction
that has been widely discussed in the literature (see, e.g., 576). The
consumption rate of NO is given by:
dPMn
=
where PNQ = partial pressure of NO, atm
Pn = partial pressure of 09, atm
?
pi
k = reaction rate constant = 23.2 atm" sec" at 86°F (571)
3.3.1.1 Oxidation Time for N02 Absorption
If we consider only N02 (or N^O,) as the ab-
sorbing species, the overall oxidation can be represented by:
2NO + 02 + 2N02 (2)
Provided that the total number of moles in the
reaction system remains approximately constant (a good assumption for the
low NO concentrations generally encountered in an effluent sample),
40
WALDEN RESEARCH CORPORATION
-------
dP
NO
(3)
Integrating Equation (3) gives
so that,
[<2Initial -
Substitution of (5) into (1) therefore results in
dP
NO
(2(P02>In -
= |dt
(6)
Equation (6) is readily integrated to give:
In
<2P02>In
(PNo'ln (2P02)In"">NO)In'l'l>NO
[<2l>02>ln -
Considering a typical flue gas to contain 400 ppm
NO and 5% Op, Equation (7) may be directly applied to estimate oxidation
times necessary for the production of different final NO compositions. The
results of such computations are given in Table 3-4.
It is evident that if N02 (or N204) is the ab-
sorbing species, a minimum oxidation time of about 30 hours is required
if the amount of N02 absorbed is to be within 2-1/2% of the amount of NO
sampled.
41
WALDEN RESEARCH CORPORATION
-------
TABLE 3-4
COMPUTED OXIDATION TIMES FOR TYPICAL FLUE GAS SAMPLES
Initial (NO) = 400 ppm; 5$ 02
NO Final T1me
200 ' 0.63
40 6.25
10 28.5
1 285.0
3.3.1.2 Oxidation Time for N203 Absorption
If, on the other hand, it is assumed that NpO-
is the absorbing species, then the following equations apply:
2NO + 02 -»• 2N02, slow (8)
2N02 + 2NO •* 2N203, fast (9)
4NO + 02 -»• 2N203 (10)
Equation (8) is the rate controlling step, but the overall stoichiometry
is represented by Equation (10). Analysis similar to that previously
given results in Equation (11):
'in
PNO
-------
The oxidation times required to produce the final NO concentrations in
Table 3-4 are almost identical in this case to the previous values. This
results from the fact that for the conditions under consideration,
so that both Equations (7) and (11) reduce to:
'
P P *! ft 'Tn
HNO (Vln °2 In
These computed oxidation times are comparable to
the recommended "times of standing" for the phenol disulfonic acid method
(547,204,425). This correspondence adds further credence to the argument
that the oxidation step is rate controlling in the oxidation-absorption
process.
It is evident, therefore, that analysis times
could be considerably decreased if the oxidation times were reduced. This
concept has been described in the literature in connection with instrumental
methods of analysis for NO using ozone or dichromate paper for the oxida-
tion (6,10). The ozonization method, however, suffers from problems as-
sociated with the production of higher oxides of nitrogen, such as NpOg,
which may yield erroneous analytical results for gas phase spectrophoto-
metric determination of N02- The higher oxides of nitrogen (NpOc) should
not be a problem, however, if aqueous absorption is used to collect the
sample.
A simple and obvious method of reducing the oxi-
dation time is to increase the oxygen partial pressure (see Section 4.3
for further discussion). To illustrate this concept quantitatively, con-
sider addition of a flue gas of the composition previously assumed (NO =
400 ppm, 02 = 5%) to a sampling vessel containing 0.5 atmosphere of pure
oxygen, to a final pressure of one atmosphere. The partial pressures of
the components after mixing will be:
43 WALDEN RESEARCH CORPORATION
-------
Pn = 0.5 + 0.05 x 0.5 = 0.525 atm
U2
pwn = 0.5 x - 200 x 10"6 atm
NO 1Q6
Using these initial values, and substituting into Equation (12), the
oxidation times necessary to produce the various final NO concentrations
are given in Table 3-5.
TABLE 3-5
COMPUTED OXIDATION TIMES FOR TYPICAL FLUE GAS
SAMPLES WITH 02 ENRICHMENT
Initial (NO) = 400 ppm
ppm NO Final Time
40 0.45
10 2.3
1 23
The oxidation times have been reduced by an
order of magnitude compared to the previous results in Table 3-4. Thus,
this technique may provide a simple but effective means of reducing the
time required for analysis.
In summary, based on the premise that NOp or
NpO., are the only species absorbed into solution, it has been shown that
the times required to oxidize 97% of the NO in the sample are of the
order of 30 hours (for a flue gas containing 400 ppm NO and 5% Op). This
value compares well with the recommended "standing time" presently uti-
lized in the PDS method (Section 4.3.1). The oxidation is, therefore,
probably the rate controlling step in the oxidation-absorption process
and may be considerably speeded up by increasing the oxygen partial pres-
sure or by addition of other oxidants, such as ozone.
44
WALDEN RESEARCH CORPORATION
-------
3.3.2 Absorption Rates of the Oxides of Nitrogen in Aqueous
Solution
Wendel and Pigford (572), writing on the kinetics of ni-
trogen tetroxide absorption in water, state: "Despite the large quan-
tities of nitric acid made by the absorption of nitrogen oxides in water,
the kinetics and reactions involved are not yet fully explained ..." A
simple reason for the confusion that still exists in the analysis of the
absorption is the large number of possible reactions that can occur in
the liquid absorbent and the complex interaction between these reactions
and the transport processes.
The absorption of equilibrium mixtures of NOp, NpO. in
N2 (or air) into aqueous solutions has been extensively studied usually
in wetted wall columns. The results obtained in long wetted wall columns
by Bolshakoff (577) and Chambers and Sherwood (573) indicate that gas-
film resistance controls the absorption of nitrogen dioxide into aqueous
solutions of nitric acid and sodium hydroxide.
On the other hand, there is some evidence that the ab-
sorption rate may be controlled by a chemical reaction occurring in the
liquid phase and/or possibly the gas phase. Eagleton, Langer and Pigford
(574) reported that the slow rate of reaction between dissolved nitrogen
dioxide and liquid water limited the rate of nitrogen dioxide absorption
into strong caustic solutions in long wetted wall columns. Peters and
Hoiman (578), also working with long wetted wall columns, concluded that
both the gas and liquid phase reactions between NpO, and water controlled
the absorption. Peters, Ross and Klein (579) using a bubble cap plate
column |for absorption studies indicate that the gas phase reaction is the
controlling one. However, Denbigh and co-workers (580,581) using short
wetted wall columns, have reported that the rate of absorption of nitro-
gen dioxide into aqueous solutions was determined by the reaction of ni-
trogen tetroxide with water, was proportional to the N^O. concentration
in the gas phase and was not diffusion controlled.
45 WALDEN RESEARCH CORPORATION
-------
The existence of the gas phase homogeneous reaction be-
tween N02-N204 and water is an unsettled question which follows from the
controversial evidence discussed above. The evidence in favor of the gas
phase reaction results mainly from the absorption of N02-N204 in alkaline
solutions (573,574) in which nitric oxide and mist was observed in the off
gases. Since sodium nitrite does not decompose in alkaline solution, it
is postulated that N02 must have reacted with water vapor in the gas phase,
producing NO and acid mist. On the other hand, the absence of mist and NO
in many similar experiments (580,581) has been cited as opposing evidence.
The presence of nitric oxide in the N02-N204 absorbing gas
mixture complicates analysis of the absorption rates and mechanism.
Eagleton, Langer and Pigford (574) were the first to report that the pres-
ence of NO in the feed gas stream to a wetted wall tower, using 46% NaOH
solution as the absorbing medium, resulted in absorption rates as much as
ten times faster than those reported for N02 absorption in a similar ex-
periment. Recently, Peters and Koval (582) have suggested that nitric
oxide influences the rate of reaction of N0« with water even when NO is
not present in the inlet gas stream. Koval and Peters (583) attempted to
discern the effect of NO addition on N0« absorption in aqueous solution
in a wetted wall column, and concluded that a critical factor in analyzing
absorption data, particularly from a short residence time contactor, is
the nitrous acid concentration.
The principal reactions which control the absorption of
nitrogen oxides in aqueous solution are:
2N02(g) ^ N204(g) (13)
N0(g) + N02(g) ?=2 N203(g) (14)
2N02(g) or N204(g) ^ 2N02(aq) or N204(aq) (15)
N203(g) ^ N203(aq) (16)
2N02(aq) or N204(aq) + H20 ^ HN03 + HN02 (17)
46 WALDEN RESEARCH CORPORATION
-------
N203(aq) + H20 ^ 2HN02 (18)
3HN02 ^ HN03 + 2ND + H20 (19)
Reactions (13) and (14) occur in the gas phase, are extremely rapid and
are generally assumed to be in equilibrium; Reactions (15) and (16) are
physical equilibria; Reactions (17), (18), and (19) represent the liquid
phase reactions. In most studies, up to two of the latter three reactions
(17-19) are assumed to be in equilibrium; the remaining reaction(s) con-
trol the rate of absorption. The selection of reactions in equilibrium
depends on the ratio of the overall time for reaction to the half life of
the reaction. The difference in reported results on nitrogen oxide ab-
sorption may frequently be explained by the number of reactions (17-19)
assumed to be in equilibrium; i.e., the overall reaction (or absorption)
time has a major effect.
The overall chemical reaction may be represented by:
3N02(g) + H20 5* 2HN03 + N0(g) (20)
The equilibrium constant for Reaction (20) is:
NO
"20 = IT
PNO aHN03
N02 aH20
where a,,NQ and a,, Q represent the activities of HN03 and hLO in the
O L*
liquid. For convenience, this expression is usually separated into two
parts, the first a function of the equilibrium gas composition, and the
second dependent only upon the concentration of HN03 in the liquid (Equa
tion 21):
P
_
3 aH n
PN02 H2°
47 WALDEN RESEARCH CORPORATION
-------
The product, K, 1C,, can be computed from standard free
energy and vapor-pressure data as shown by Wenner (584) and Forsythe and
Giaque (585) permitting a plot of PNQ/PWQ versus acid concentration to
be obtained (Figure 3-5). The equilibrium constant for Reaction (13) has
been shown by Wenner (584) to be given by:
log1QK13 = -^p - 9.226 atm"1 (22)
The equilibrium concentration of Reaction (14) can be com-
puted from the data of Verhoek and Daniels (508) although the concentra-
tions of N203 are in most cases extremely small.
Figure 3-5 and Equation (22) may be employed to calculate
the equilibrium concentrations of N02 and NO above pure water, for a flue
gas containing initially 400 ppm NO which has been oxidized to 400 ppm N02
(total pressure 1 atm). Since from Equation (20), moles N02 absorbed
equals three times moles NO evolved, we have by material balance:
3 (moles NO evolved) = moles N02 absorbed
= initial moles N02
= (moles N02 reacted to NpO, + moles N02 remaining)
(23)
' "?P = 400 x 10 - 2P + P (24)
.. JKNQ «uu x iu |*KN204 KN02J (^}
" 3IVPN02) = 40° x 10 " [2K13^PN02^ + PN02J
where PNn = final N09 partial pressure, atm
IlUrt k
PNO = ^ina^ NO partial pressure, atm
PN 0 = final N2°4 partlal Press"'r'e» atm
48 WALDEN RESEARCH CORPORATION
-------
inz
.10'
\
i
^7-
\
^
,\
Vs
\
Wertne
II.
^
• - 1
Sx~—
\
\
\
\
%
A Epshtein, 32° F (587)
0 Denbigh and Prince, 77° F (580)
i Denbigh and Prince, 104° F (580)
D Abel, Schmidt and Stein, 77° F (586)
A Chambers and Sherwood, 77-95° (573)
1
-
\
VN. T!T
79 XX'
i-,7-!
1'p^yr
FV —
X V
V
\\
'^0
N
I04°F
\
\\
»
\
\
\
vjv
\ \
-Vx
86°F\
Wr\\,
\ Ny
\
\
N
\*
"i
u
\
\\
\— - :
V
"a
\
30 40 50 60
Acid concentration,per cent HNOjby weighf
Fig. 3-5 Equilibrium Vapor Pressures of NO,, and NO over Nitric Acid
49
WALDEN RESEARCH CORPORATION
-------
Extrapolating Figure 3-5 to 0% HN03 gives (at 77°F) Kj * 108 atm"2, and
from Equation (20), K- = 6.56 atm~^ . Substitution of these constants
4
into Equation (25) and solving by trial and error gives P..Q - 10 atm,
or a final concentration of N0« - 100 ppm. This simple example clearly
indicates the necessity for an oxidizing reaction to occur in the liquid
absorbing media in order to drive Reaction (20) continuously to the right.
Previous discussion of the kinetics of absorption indi-
cated that good correlation of experimental data could be obtained in
many situations if the following rate equation is used:
' k CN204 '26>
The basic assumptions in the theoretical development of
Equation (26) were:
(1) The rate of removal of NO^-N^O. is controlled by
Reaction (17).
(2) The chemical reactions occur under irreversible
consitions.
N204.
unchanging.
(3) Instantaneous equilibrium exists between N0~ and
(4) The nitrous acid concentration is negligible or
The fourth assumption has major implications for applica-
tion of the data reported for short contact time absorbers to batch ab-
sorption systems (of the type likely to be encountered in analytical sys-
tems) since the nitrous acid concentration cannot be considered unchanging
as the contact time increases. Indeed, Peters and Koval (582) in studying
the absorption of NO^-N^O. in an agitated reactor (i.e., a reactor of
fairly long gas residence time) observed that the effects of N«0, can no
£ O
longer be ignored because a small fraction of the total oxides was present
50
WALDEN RESEARCH CORPORATION
-------
as N203, and Reactions (14), (16) and (18) must be included in the analy-
sis. Clearly, where NO is added to the N02-N203 mixture, Reactions (14),
(16) and (18) must also be considered. Consequently, these reactions
should be considered in absorption processes for the analysis of nitrogen
oxides in flue gas which are N0-N02 mixtures. However, the equilibrium
position for Reaction (14) is such that the N203 concentrations are ex-
ceedingly small for the nitrogen oxide concentrations characteristic of
flue gases (obtained as the oxidation proceeds) so that including the ef-
fect of N203 in the analysis of absorption rates is probably not necessary.
As a first approximation, in studying the absorption of the nitrogen oxides
produced by the oxidation of flue gas nitric oxides, it will, therefore,
be assumed that the rate can be described by an equation similar to
Equation (26).
The data of Wendel and Pigford (572) for the kinetics of
the absorption of N20. in water are probably some of the best in the
literature and compare favorably to the results of Denbigh and Prince
(580). Their data can be approximated by the following equation:
NA = 35 x 10"6 PN 0 (27)
gm moles N20«
where N. = flux of absorbing species in *
cm sec
PM n = bulk partial pressure N90A, atm
N2U4 * *
Equation (27) can be used to estimate a typical time for
absorption of a gas consisting of 400 ppm N0? in an inert matrix (i.e.,
I
after the oxidation of NO has been completed). The gas is contained in a
volume, V liters, at 1 atm pressure and 25°C; the absorbing surface is A
p
cm . Considering the gas space as a system and performing a material
balance on N20, gives, therefore:
= - 35 x 10'6 A Pw n (28)
dVo,
dt
51
WALDEN RESEARCH CORPORATION
-------
where nw n
dpN 0
Equation (29) is readily integrated to give:
(PN70, Initial , .
In [p * i = 35 x 10"b RT §• t (30)
1 N204;Final v
Since it is assumed that N02 and N204 are in equilibrium given by:
(31)
Equation (31) can be substituted into Equation (30) to give:
N09 Initial fi .
2 In ip * x - = 35 x 10"° RT 5- t (32)
For a typical analytical absorption system we may take:
2
V = 1 liter and A = 100 cm . The time to absorb sufficient N09 to reduce
g ^
(PNO )Fl-nal to 10 x 10"° atm is given by:
t = - ^-^0 - = 86 sec
(35 x 10"°) x .082 x 298 x 100
It must be emphasized that it has been assumed in the
material balance, Equation (28), that the gas phase is well mixed, i.e.,
that no NpO. or N02 concentration gradients exist. In actual practice,
this will not be the case unless stirrers are placed in the gas space.
Furthermore, it is also assumed that the absorbing liquid does not begin
to saturate with nitrous and nitric acids, i.e., that the liquid is well
mixed and of such a volume that the acid concentrations are negligible.
Once again, this is probably not quite correct.
52 WALDEN RESEARCH CORPORATION
-------
However, it appears on the basis of the above sample cal-
culation that absorption rates are much faster than the oxidation rates,
confirming that, as assumed in the previous section, the overall absorp-
tion process is probably controlled by the rate of nitric oxide oxidation.
3.3.3 Experimental Studies on Oxidation Times for NO
In order to verify the hypothesis that the gas phase oxi-
dation of NO is the rate determining step in the absorption of nitrogen
oxides in aqueous solutions (Section 3.3.1), some preliminary laboratory
work was conducted. Tests were run in a gas-fired combustion effluent
doped to higher NO levels with NO. The results are given in Table 3-6
A
below. See also Section 5.3.2 for additional studies using caustic
absorption.
TABLE 3-6
COMPARISON OF MEASURED AND CALCULATED N0¥ LEVELS
A
Sample No.
1
2
3
4
Absorption
Time (hr)
0.5
>24
1
>24
ppm NOX
(nitrate electrode)
293
438
1028
1160
ppm NOX
(calculated)
287
—
1066
—
ppm NO
added
447
447
1174
1174
The calculated NO values were derived from Equation (12) using the mea-
A
sured oxygen content in the flue of 9%. The agreement between the cal-
culated and measured NO values for runs 1 and 3 is excellent. The de-
A
tails of the nitrate electrode method are given by Driscoll, et al. (619).
Thus, we see that the rate determining step in the absorption of the ni-
trogen oxides is presumably the gas phase oxidation of nitric oxide. This
indicates that substantial savings in analysis time can be made by using
the technique of oxygen enrichment described in Section 3.3.2 with the PDS
or other methods for NO analysis.
X
53
WALDEN RESEARCH CORPORATION
-------
4. METHODS FOR DETERMINATION OF TOTAL OXIDES OF NITROGEN
4.1 Introduction
Flue gases contain nitric oxide and a small but poorly defined
fraction of nitrogen dioxide. Manual methods for collection and analysis
of total oxides of nitrogen depend upon a number of reaction steps sche-
matically shown below:
NO \ N02 \ aq(jeous Saltzmann reagent
oxidation N ft ( absorption - analysis for N02
and/or _
*" analysis for NOl
H202 or KMn04 3
Since there are no "good" specific wet chemical methods for nitric oxide,
oxidation is a required step. The most commonly used oxidizing agent for
this reaction is molecular oxygen. Ozone has also been used for oxidation
prior to (continuous) gas phase spectrophotometric readout. The oxidation
methods in current use are described in detail in the following section on
sampling and analysis methods for total NO .
J\
After oxidation of nitric oxide, the (higher) nitrogen oxides may
be collected by absorption in aqueous solutions. If Saltzmann or other
Griess-type coupling reagent is used as the collection medium, the nitrite
determined may be used as a measure of the total nitrogen oxides concen-
tration. An oxidizing agent such as hydrogen peroxide or potassium per-
manganate may be utilized in the absorbing solution to oxidize nitrite to
nitrate for determination of total oxides of nitrogen as nitrate. The fol-
lowing sections, therefore, include discussion of measurement techniques
for both nitrate and nitrite.
Since we cannot separate sampling and analysis methods completely
(for realistic field methods), we have divided total NO methods into the
A
following classes:
(a) Those which have been used in effluents (power plant, auto
and diesel exhausts or nitric acid plant effluents); these methods include
sampling as well as analysis.
54 WALDEN RESEARCH CORPORATION
-------
(b) Methods which have been used mainly for measurement of
nitrate in aqueous solutions or nitrite (N02) in ambient air.
Methods for determination of the individual oxides, NO or NO^,
present completely different problems and are discussed subsequently.
4.2 Sampling
4.2.1 General
Sulfur oxides are ordinarily collected by methods which
concentrate the sample, e.g., absorption in a liquid. This procedure is
a feasible one as a result of the relatively high solubility of sulfur
oxides in aqueous media. The solubilities of a number of common gases
encountered in flue gas sampling are given in Table 4-1 (525).
TABLE 4-1
SOLUBILITIES OF COMMON GASES IN WATER AT 20°C
Gas Solubility, g/1 of Water
Nitrogen 0.010
Oxygen 0.044
Nitric oxide 0.063
Carbon dioxide 1.74
Hydrogen sulfide 3.98
Sulfur dioxide 117
, The solubility of nitric oxide in distilled water is com-
i
parable to that of the common permanent gases, N^ and 0«. The solubility
of S02 is, however, more than three orders of magnitude greater than that
of NO. Since high solubility is a prerequisite for quantitative collec-
tion by scrubber devices, the techniques used for nitrogen oxides have
been limited to grab samplers which function by mechanical retention fol-
lowed by absorption (Figures 4-1 and 4-2). Power plant sampling has ordi-
narily been conducted by collection in evacuated flasks, although plastic
bags have been widely used in automotive studies.
55
WALDEN RESEARCH CORPORATION
-------
BREAKING
SCRATCH
SCAltD WITH WAX-FIUCD
CARTRIDGE
290 lo 500 ct
CAPACITY
L I
1
, j
Evacuated Sample Container
I RUBBER CM
-— - • / ABSORBING \ '•"--" ~~ .
^ ••*•• LIOUIO • V':;":vy!
Evacuated Sample Container (or Col-
lection of Sample in an Absorbent.
-'&^~^^~"^ 350 lo 300 ee. CAPACITY
>..4- ,
^-T- '
I—T • 250 10 300 ee. CAPACITY »-'• f».
M :K
Gas Sample Tubes for Collection of
Sample by Displacement.
Figure 4-1. Types of Grab Samplers (partial).
56
WALDEN RESEARCH CORPORATION
-------
Stopper: J 14/20
outside joint, with
holes as shown.
I.D. of glass tube:
5 mm
Pyrex glass bottle and cap with 29/42
ground glass joint: openings in cap
and neck of flask to correspond
40 mm fr.----)! 40 mm
Standard wall
tubing 8 mm o.d.
5> 29/42 1 joint
8 mm hole
Figure 4-2. Additional Grab Samplers (507).
57
WALDEN RESEARCH CORPORATION
-------
: The selection of a probe for collection of total oxides
of nitrogen from a flue is not as difficult a problem as that for sulfur
oxides sampling. For the latter, losses due to condensation of H2SO. and
dissolution of S02 and S03 in condensate water must be prevented. Temper-
ature controlled probes are, therefore, employed for the quantitative col-
lection of SO . The very low solubility of the principal oxide of nitro-
A
gen, NO, considerably reduces the magnitude of this problem in sampling
for total NO . Simple stainless steel or pyrex tube is generally adequate
A
for the sampling probe. A complete NOX sampling train is illustrated in
Figure 4-3 (173).
4.2.2 Integrated Grab Sampling
A major difficulty in the interpretation of conventional
grab samples results from the short sampling time (seconds to minutes).
There is a significant advantage in collection of a sample integrated
over a full operational cycle characteristic of the equipment under test;
this is the direct relationship between the integrated sample and the
average emission, which is ordinarily the objective of the measurement.
Of course, a number of grab samples may be averaged to obtain the inte-
grated emission during a cycle, but this is clearly a more tedious ap-
proach.
An integrated grab sample may be obtained by inclusion of
a critical orifice in the sampling line as shown in Figure 4-3. Critical
orifices are available for flow rates as low as 25 ml/min (507) but the
small diameter of the orifice required for the lowest flow rates may lead
to erratic flow control as a result of plugging by particulates even when
a prefilter is installed. Brief and Drinker (260), for example, have reported
repeated plugging during stack sampling using an orifice sized for flow
of 25 ml/min. A 100 ml/min orifice may represent a reasonable compromise.
Such an orifice used with a 5-liter evacuated flask will provide 25 min
sampling time (since critical flow will continue only up to flask pres-
sures below 0.5 atmosphere). This technique would provide considerable
improvement over present grab sampling methods which usually have inte-
gration times as short as seconds.
58 WALDEN RESEARCH CORPORATION
-------
PROBE
12 5-
12 5
\
-.
/
TO VACUUM
^ORIFICE ASSEMBLY
(
1
Xio » •'THREE-WAV
IB M STOPCOCK
*" MERCURY _/
MANOMETER
•^ 24/40
^^ X24.ITER FLASK
. . „„.*. .
i
1YGON
SLEEVE
DETAIL A
Figure 4-3. Apparatus for integrated grab samples- PHS NO
sampling train (173)
59
WALDEN RESEARCH,CORPORATION
-------
It is of interest to note that Brief and Drinker employed
a partially opened stopcock as a simple critical orifice (260). They ob-
served that sonic velocity is maintained up to flask pressure of 0.4 atm.
4.3 Sampling and Analysis Methods Used for Total NO Determinations
in Exhaust Gases
4.3.1 Phenol Disulfonic Acid (PDS) Method (1.9,76.77.134.278.
388,404,428,433,425,512,536,547)
4.3.1.1 Description
An aqueous sample containing nitrate (from ab-
sorption of nitrogen oxides in 0.1N H^O./HpOp) is made slightly alka-
line and evaporated to dryness. The solids are dissolved in PDS reagent
(30% PDS in concentrated HpSO.) to give nitrophenol-disulfonic acid.
When made sufficiently alkaline (generally with NhLOH), the nitro com-
pound has a yellow color (absorption maximum = 410 nm); the optical den-
sity is proportional to the amount of nitrate in the original sample.
Some standard procedures (134,404) recommend
diluting the PDS sample mixture to 70% H^SO. and heating for a few min-
utes before the final dilution and addition of base; however, none of the
cited references give justification for this step.
The final concentration of NH^OH (in excess of
that to neutralize the H2SOJ in the solution varies from 0.3 molar (278)
to 0.75 molar (76,77,428) to 3 molar (404), the higher concentrations hav-
ing a higher pH; use of KOH in place of NH.OH gives a higher pH, and 5 to
10% more intense color (428,433).
4.3.1.2 Application
This is the most widely used method for measuring
NOX (absorbed into dilute H2S04/H202) in flue gases (1,134,404,547,76,137,
425,512,347,204,536). It has also been used for automotive (7) and diesel
exhausts (9). The method is also widely used for nitrate in agricultural
material, soil and water.
WALDEN RESEARCH CORPORATION
-------
4.3.1.3 Sampling
Nitrogen oxides in power plants are collected by
taking grab samples in 1- or 2-liter evacuated flasks. The most common
absorbent is acidic (0.06N H2S04) peroxide (7,76,137,404,425,512,347,204,
536). The peroxide serves a two-fold purpose: oxidation of NOZ to NOl
and SOI to SOT. For flue gas samples which contain 3 to 5% oxygen, over-
night contact time is generally employed (7,137,76,404). Gill (77) and
Beaty, et al. (209) use a 2 to 3 hour contact time. However, their work
is with ambient air samples and/or tunnel atmospheres where the oxygen
content is generally 3 to 4 times higher than in combustion sources. The
low oxygen concentration in fossil fuel combustion would account for the
longer absorption time required since the gas phase oxidation of NO is
the limiting step.
In the early work on the PDS method for air
analysis, alkaline peroxide absorbing solutions were used. These were
discontinued when Francis and Parsons (453) found low results. They at-
tributed this to incomplete oxidation of nitrite in caustic solution as
a result of low reaction rate. Since nitrite gives nearly the same re-
sponse to PDS as nitrate, the actual reason is not known.
An alternative to the acidic peroxide absorbent
is acid permanganate which has been used by API (204). This method offers
no apparent advantage over peroxide. On the basis of limited data (204),
the precision is somewhat poorer than for peroxide absorbent.
4.3.1.4 Applicable Concentration Range
I The absorbance-concentration curve generally
covers the range of 0-100 ppm N02 (134,404). The method is generally em-
ployed for measuring from a few ppm to a few thousand ppm NO (76,77,134,
/\
404), the upper limit being extended by appropriate aliquoting or dilution.
The upper range can also be extended by measuring the absorbance at 500 nm.
61
WALDEN RESEARCH CORPORATION
-------
f' 4.3.1.5 Precision and Accuracy
See Section 6 for precision and accuracy data
on the PDS method.
4.3.1.6 Interferences
Halides cause variable losses, particularly at
low NOX levels. Br" equivalent to 2.5 ppm HBr added to NOZ equivalent to
40 ppm N02 resulted in a 5% loss in N0« (7). For a solution of NOZ equi-
valent to 400 ppm N02, addition of Cl~ equivalent to 16 ppm HC1 caused 2
to 8% loss, and Cl" equivalent to 48 ppm HC1 resulted in 13 to 17% loss
(433). The halogen (approximately 25 ppm) expected from gasoline in
automotive exhaust gases will cause less than 2% loss in apparent NO
(assumed 1200 ppm NO ) by the PDS method (7).
/\
Nitrite gives an interference equivalent to 80%
of the nitrate response with the PDS method (see Section 5.3.2). Excess
halides can be removed by adding slightly less than the stoichiometric
amount of Ag^O* and filtering off the AgCl before evaporating the sample
(278,433).
The effect of S09 on NO determinations by the
w /\
PDS method has been checked by NAPCA (548). They find no decrease in ap-
parent NOV concentrations even when the S09 is present in a ten-fold ex-
X t
cess (2000 ppm). (See Figure 4-4.)
With the strong H2$04, unburned hydrocarbons may
give colored products which will interfere with the photometry (323). This
problem should not be serious, however, in large power plants which are
operated with close control of combustion conditions.
4.3.1.7 Comments
This method is the one most widely used in the
United States for measuring nitrate in aqueous solution. The method was
developed by Chamot, et al. (433) around 1910, and adapted to give a work-
ing method for nitrogen in air (including sampling) by Beatty, Berger and
Schrenk in 1943 (428).
62 WALDEN RESEARCH CORPORATION
-------
i.
Q.
O
a
o
S5
1 I I I I ' i M i I i : | i i ! I i il I
600 1000 1500
ppm SO (added)
2000
0 PDS
A Saltzaann
Figure 4-4 Effect of SO. concentration on apparent NO concentration
by PDS and Saltzmann methods (548)
63
WALDEN RESEARCH CORPORATION
-------
Beatty, et al. (428) used 0.1N H2S04 plus 0.1%
H202 as the absorbing solution. More recent versions of this method have
employed H202 concentrations of from 0.06% to 0.4% (134,404,76). The
peroxide serves a two-fold purpose, both to oxidize any nitrate and to
convert sulfite, an interference, to sulfate.
After absorption, the solution is made alkaline
with potassium or sodium hydroxide and slowly evaporated to dryness to
avoid spattering losses. This evaporation step is time-consuming and gen-
erally requires from a few hours to overnight depending on the amount of
solution and the equipment. Obviously, minimizing the solution volume
would decrease the evaporation time.
The PDS reagent is composed of 25 g of phenol in
150 ml of concentrated H2$04 + 75 ml 15% fuming H2$04. Two ml of this
reagent is sufficient to wet and dissolve the solids from the absorbent
solution since this amount of reagent has been shown to react with 1.2 mg
N02 (equivalent to 600 ppm in a liter) (77).
Virtually, all procedures allow a 2-minute> nitra-
tion reaction time. This was originally established by Chamot (433) and
checked by Gill (77) on samples containing up to 1.2 mg N02> In fact,
Gill found 96% recovery for 1.2 mg NOp.
Some standard procedures (134,404) call for ad-
dition of 1 ml of FLO and four drops of concentrated H2$04 and warming
for 3 minutes after the nitration is completed. No reason is cited for
this operation (although it is probably included to help dissolve all the
solids), and it was not deemed necessary by the developers of the method.
The yellow color produced on making the solution
alkaline is due to the ionization of the phenolic proton from the 6-nitro-
phenol-2,4-disulfonate (433). Since this species is a weak acid (pK of o-
nitrophenol = 7.2, pK of phenol-p-sulfonate = 9.0), the final solution
must either be sufficiently alkaline so that complete ionization takes
place, or be very reproducibly buffered. Thus, the nitrate calibration
solutions must be run exactly as the samples, so that the same volumes of
reagents are used.
64 WALDEN RESEARCH CORPORATION
-------
\
NH.OH is recommended as the base in most proce-
dures and is useful for avoiding precipitation of silver oxide when silver
salts are used to remove halogen. KOH does give some 5 to 8% more color
intensity (428,433) presumably because of the higher pH of the final
solution.
The sampling technique described by Dimitriades
(7) provides more precise results than the conventional PDS grab sampling
techniques (see Section 6.3). He collects a dried, proportional sample of
the exhaust in a 50-liter plastic bag. The replicate grab samples are then
taken from the plastic bag. This procedure minimizes sampling errors and
smooths source fluctuations. If the residence times in the bag are too
long, the accuracy may be of some question because of absorption of NO,
on the walls.
'2
4.3.1.8 Conclusions
The advantages include wide concentration range,
minimum number of sample handling steps, and no interference from the S02
in flue gases. The disadvantages are the long elapsed time and a possible
interference from halides. It may be possible, however, to shorten the
long elapsed time by addition of oxygen or ozone in the flask after the
oxides of nitrogen are collected (Section 3). The sampling technique
used by Dimitriades may provide increased precision for the PDS method.
The accuracy of this sampling method should be checked, however. The PDS
method would be more useful if: (1) elapsed time could be shortened (see
footnote), (2) if halide interferences could be more easily removed.
4;3.2 Xylenol Isomer Methods
2,4-xylenol (189,323,379); 3,4-xylenol (279);2,6-xylenol
(379).
Coulehan and Lang (620) have recently modified the^PDS method to eliminate
the time-consuming step of evaporation to dryness. This had been done by
minimization of the volume of the absorbing solution and running the nitra-
tion in 50% H2S04. They obtain good agreement between conventional PDS and
their modified version of PDS.
55 WALDEN RESEARCH CORPORATION
-------
4.3.2.1 Description
Aqueous nitrate is mixed with the xylenol and
H2S04 to give an acid concentration of from 50% (2,6-xylenol) to 80% (3,4-
xylenol), and allowed to react (with or without heating) for approximately
30 minutes. The sample is then diluted with water and steam-distilled into
dilute NaOH. The yellow-orange color produced is measured with a spectro-
photometer in the 420-440 nm region.
The steam-distillation step can be replaced by
extracting the nitroxylenol into toluene and back-extraction into dilute
NaOH for measurement (323).
The 2,6-isomer is measured directly in the nitra-
tion mixture by UV spectrophotometry (320-330 nm region).
4.3.2.2 Areas of Application
Nitrous fumes in air (189); nitrogen oxides in
diesel exhaust (379), nitrate in water (397).
4.3.2.3 Sampling
In the 2,4-xylenol method (379), the oxides of
nitrogen are collected in an evacuated sampling bottle (250 ml) contain-
ing 10 ml of 0.05N NaOH solution. To collect the sample, the evacuated
flask is inverted and the sample of flue gas is bubbled through the ,
sodium hydroxide solution. After the NO is oxidized and absorbed (3 min-
utes),^" potassium permanganate is added to oxidize nitrite to nitrate.
The sample is then analyzed for nitrate by the 2,4-xylenol method.
4.3.2.4 Applicable Concentration Range
2,4-xylenol: 50 to 1000 ppm N02 in 1 liter of
*
BSI states that this method should be suitable for flue, gas. It is not,
however, apparent that such tests have actually been conducted.
This short oxidation time is not consistent with kinetic data available.
(See Section 3.)
66 WALDEN RESEARCH CORPORATION
-------
diesel exhaust (379); also from 4 to 1000 ppm by DuBuboscq (visual)
colorimetry (189).
3,4-xylenol: up to 1.1 mg N02 (equivalent to
550 ppm N02 per liter) (279).
2,6-xylenol: up to 65 yg N02 (equivalent to 30
ppm N0« per liter) (397). Upper limits can be extended by aliquoting
sample.
4.3.2.5 Precision and Accuracy
2,4-xylenol: relative precision is about 5%, no
data (379).
3,4-xylenol: coefficient of variation (C.V.) =
1 to 2% from data summary (279).
I.
2,6-xylenol: C.V. = 0.5 to 2%, data summary (397),
No accuracy data found.
4.3.2.6 Interferences
Chlorides - For 2,4-xylenol, 0.4 mg HC1 (equi-
valent to 270 ppm HC1 in 1 liter flue gas) caused an apparent decrease of
5 ppm N02 at the 34 ppm level (323), while 500 ppm HC1 was reported to
give a negative error of less than 5% in the measurement of from 20 to 400
ppm N0« in flue gas (379). With 2,6-xylenol, chloride causes a small,
positive error; the equivalent of 600 ppm HC1 causes only a 3% error at
the 40 ppm N02 level.
Nitrites - Interference by formation of colored
products with xylenols (279) and formation of nitrosoxylenols (397).
Oxidants - Give colored products with 2,4-xylenol
which are co-distilled with nitroxylenol (279,323). Traces of KMn04 ap-
parently do not interfere (189,379).
67
WALDEN RESEARCH CORPORATION
-------
Organics - Up to 20 mg of sucrose (added as a
model compound) gave losses equivalent to only 1 ppm N02 (323). Organics
which are found in ground water result in negative errors with 2,6-xylenol
(397).
4.3.2.7 Comments
(a) Steam distillation is a tedious step. The
double extraction into toluene and back into aqueous NaOH (323) may be
useful, but this is still a problem due to emulsion formation.
(b) Losses of nitro compound from volatiliza-
tion can occur if the temperature rises more than 40°C during addition
of acid or subsequent nitration period.
(c) Oxidants, except for traces of KMnO., must
be destroyed.
(d) Sensitivity and acidity requirements are
such that the sample (or an aliquot) can be run directly without the
evaporation step.
4.3.2.8 Conclusions
The lower acidity required for nitration is one
advantage over PDS since it is not necessary to evaporate the sample.
Xylenols are quite sensitive to oxidants, and require a separation step
before measurement. Although the total elapsed time is shorter for
xylenols, there are more sample handling steps than with PDS. In view
of this, the xylenol procedures offer no outstanding advantage over the
PDS method.
4.3.3 Iron (II) Sulfate (and Other Complexes)
4.3.3.1 Description of Method
The nitrogen oxides in the gas are absorbed in
75% H2S04, then iron (II) sulfate in 75% H2S04 is added. Nitric oxide
68 WALDEN RESEARCH CORPORATION
-------
reacts with ferrous sulfate to produce ferrous nitrosyl sulfate. This
complex absorbs strongly in the visible region and after 5 minutes is
measured spectrophotometrically at 520 nm. The color is stable for 2
hours. The reactions with the oxides of nitrogen are:
FeS04 + NO -»• FeS04 • NO
3FeS04 + N02 + H2$04 - FeS04 • NO + Fe2(S04)3 + H20
4.3.3.2 Areas of Application
NO and N02 mixtures in air, and in real and
synthetic mixtures of gases from detonation of explosives (5), automotive
exhausts (455).
4.3.3.3 Sampling
Samples collected in bubblers at a rate of 0.5
1/min for 5 minutes (455); evacuated flasks containing 75% H2SO, (5).
4.3.3.4 Applicable Concentration Range
50 to 10,000 ppm N02 with an 800 ml gas sample;
sensitivity could be increased by increasing the sample size (.5).
4.3.3.5 Precision and Accuracy
Coefficient of variation, 1 to 2% in analysis of
gas mixtures containing from 125 to 10,000 ppm of NO. Using NaN03 solu-
tions for calibration, recoveries of NO and N02 added in separate runs
were 97|% and 90%, respectively (5).
4.3.3.6 Interferences
Nitrate gives identical response; 400 ppm H2S
interferes but 6000 ppm S02 did not (5). One possible interference is
due to the oxidation of ferrous sulfate by molecular oxygen followed by
reaction of the corresponding ferric sulfate with NO. The ferric sulfate
does not absorb and, therefore, would produce a negative interference.
69 WALDEN RESEARCH CORPORATION
-------
4.3.3.7 Comments v •'
Other inorganic salts which absorb nitric oxide
are (487): :,
(a) ferric chloride to produce ferric nitrosyl
hexachloride 2(FeCl3)NO
(b) ferrous selenate to produce ferrous
nitrosyl selenate
(c) potassium nitrosyl disulfonate (454)
Some preliminary work at NAPCA (548) indicates that Fremy's salt (nitrosyl
disulfonate) might prove useful for collecting nitric oxide. This is based
on the following reaction:
(KS03)2NO + NO -> (KS03)2NO-NO
H20
(KS03)2NOH + HN02
In basic solution, the salt is highly colored (absorption maximum 540 nm).
When NO is bubbled through the solution, bleaching occurs. Collection was
evaluated by scrubbing NO (in N2) with bubblers containing alkaline solu-
tions of Fremy's salt. There were several problems associated with the
technique:
(1) What is exact stoichiometry?
(2) Absorbance of bleached solution changed
with time indicating a slow reaction of NO in the scrubbing solution.
The latter problem could be solved by taking a grab or integrated grab
sample. The former problem would involve additional laboratory work to
clear up. To use this method in sources which contain oxidizable com-
ponents (e.g., S02 in large amounts), a prescrubber such as peroxide so-
lution would have to be used.
70 WALDEN RESEARCH CORPORATION
-------
4.3.3.8 Conclusions
These methods may lead to poorer accuracy than
the nitration methods (PDS, etc.), but they may, indeed, prove useful as
a fast readout at "high" NO concentrations. The method should be con-
A
sidered for a "simplified" method for NO in power plants.
A
4.3.4 Spectrophotometric Determination of N00 in the Gas Phase
—IJT u.-- - - — -._._.._. - i _.— -jr -
4.3.4.1 Description
Automotive exhaust samples containing NO are
oxidized to N02 by solid (dichromate paper) or gaseous oxidizing agents
such as oxygen or ozone. The NOp formed is then determined spectrophoto-
metrically (at ^ 400 nm) by optical density measurements in the gas phase
(6,7,10,458).
4.3.4.2 Areas of Application
Automotive exhausts.
4.3.4.3 Sampling Equipment
The oxides of nitrogen are collected by grab
sampling techniques using evacuated flasks (10) or syringes (458).
Methods which employ ozone or dichromate paper (for rapid oxidation of NO)
achieved a continuous readout of the absorbance of NOp and, therefore, the
NO concentration.
A
4.3.4.4 Applicable Concentration Range
10-2000 ppm and up (6,10).
4.3.4.5 Precision and Accuracy
Depends on oxidation step.
4.3.4.6 Interferences
Hydrocarbons and water vapor. S02 would be the
71 WALDEN RESEARCH CORPORATION
-------
major problem with gas phase spectrophotometry of N02 from power plant ef-
fluents since S02 is present at high concentrations and reacts with N02
in the gas phase (452) to produce S(L and NO.
4.3.4.7 Comments
(a) Solid oxidants. Solid KoCr207 supported on
glass fiber paper (10) was used to oxidize nitric oxide from an auto ex-
haust and, therefore, determine N02 continuously. Ripley, et al. (10)
found that 200 ppm of NO could be oxidized quantitatively but at 1000 ppm
the efficiency was only 85%. The dichromate paper lost efficiency at a
relative humidity higher than 75% or when low concentrations of aromatic
hydrocarbons such as toluene were present (91). Dichromate has been used
successfully by Wilson and Koplinski (91) for ambient NO concentrations
/\
as a replacement for the irreproducible permanganate bubbler.
(b) Ozone. Ozone has been used as an oxidizing
agent for NO by Singh, et al. (6) and DuPont Instruments (457). DuPont
found that, although the oxidation of NO was complete within minutes, not
all the NO was converted to N0«. A substantial portion (30%) was con-
verted to higher oxides (NoOg) which do not absorb in the same spectral
region as NO,,. Singh, et al. (6) found that the ozone method was in
agreement (5%) with PDS analysis for NOX only for an 0.,/NO ratio of 0.5.
At higher ratios, low results are obtained, presumably as a result of the
oxidation to N205. At ratios below 0.5, the oxidation was incomplete and
results were again low. This dependence of NO concentration on the ozone
/\
concentration leaves much to be desired if gas phase spectrophotometry is
used for N02 readout. If the higher oxides of nitrogen were collected in
aqueous solutions for NOI analysis, this would not be important in ozone
oxidation.
(c) Oxygen. Dried auto exhaust gas is passed
into a 50-liter tedlar bag (7). A sample is transferred into an evacuated
stainless steel tank until a final pressure of one atmosphere is reached.
Then an equal volume of oxygen is added to this mixture to oxidize the
nitric oxide to NO,,. A modification of this procedure by the Bureau of
72 WALDEN RESEARCH CORPORATION
-------
Mines involves prior removal of hydrocarbons by absorption on a heated
Hopcalite column. After 45 to 60 minutes, the absorbance of NCL is meas-
ured with a spectrophotometer (7).
The Chevron research method (458) involves
sampling the dried exhaust gas with a 200 cc syringe. The syringe is
charged with a 50 cc sample of the exhaust and 100 cc of oxygen. It is
then injected into a 113 cc spectrophotometer cell and the absorbance
measurement is made when the N02 concentration ceases to rise. Although
this method is not as rapid as ozone oxidation, less than one hour is
usually required. Although the oxygen content of the automotive exhaust
gases is low (5%), the presence of hydrocarbons speed up this oxidation
process (456). Dimitriades has found that the NO values where hydrocar-
A
bons (^ 400 ppm) are removed, are about ten percent higher. The high
hydrocarbon content is generally not a problem, however, in gas- or oil-
fired power plants but this may be a problem to be considered for the
coal-fired plants. The hydrocarbon content for an intermediate-size
plant may approach 400 ppm but is generally below this value (see Final
Report, Volume I).
A comparison of the above three spectrophoto-
metric methods with the PDS method is given in Table 4-2 [from
Dimitriades (7)].
4.3.4.8 Conclusions
Ozone and dichromate paper appear to be interest-
ing oxidants since they permit immediate readout of the NO concentrations.
J\
The ozone technique is limited in gas phase spectrophotometry because of
the formation of higher oxides which do not absorb in the same spectral re-
gion. It may, however, be used for continuous NO measurement if the N09
A £
and higher oxides are collected in an aqueous solution. This technique
(continuous determination of NO ) is not in the mission of the present
J\
program. However, addition of ozone to grab samples to decrease the oxi-
dation time is a possible approach. The dichromate paper appears to suffer
from incomplete oxidation at higher NO concentrations and is also subject
/\
to difficulties in preparation.
73 WALDEN RESEARCH CORPORATION
-------
Legend: ST = Static Tank
BM = Bureau of Mines
CR = Chevron Research
= Phenol disulfonic acid
Fuel
Indolene
Isooctenc
Jlravy with aromatics
Heavy with olefins
Engine Operal ion
Crui.se, 40 jnph
Califor
i
t
i
lia cycle
Modified California cycle"
1 1
it
California cvcle
tf
ti
Modified California cycle"
ti
it
n
n
California cycle
1 1
C(
Modified California cycle1"
California cycle
K
((
Modified California cycle"
ST
297
27!)
331
315
22J<
2-10
9!)
250 _
121
270
632
032
678
220
230
520
513
487
113
17
217
462
MS
105
UK)
422
NO,, ppm
BM Cl!
311
305
340
329
241
276
111
278
127
283
645
647
702
237
247
545
521
500
119
185
US
—
—
—
417
369
150
244
224
47!)
156
198
205
433
—
—
: —
. — .
—
—
—
—
—
—
. — .
—
—
225
—
—
500
496
—
165
—
—
—
—
—
375
225
202
46(>
—
186
ISC.
410
PUS
—
—
—
—
—
—
—
—
—
274
647
657
600
220
230
—
496
401
100
172
113
427
527
633
440
348
152
227
20S
465
147
185
181
409
Table 4-2
Comparison of NOX Results Obtained by Simultaneous
Application of the ST, BM, CR and PDS Methods (7)
74
WALDEN RESEARCH CORPORATION
-------
Some of the automotive exhaust sampling methods
offer advantages over the conventional grab sampling methods. Collection
in a 50 liter plastic bag allows integration over a 25 or 30 minute cycle
and thereby increases the accuracy of the emission determination and de-
creases sampling errors due to source variation. An interesting feature
of these automotive exhaust results is that the overnight absorption time
of the, conventional PDS method may be reduced by increasing the oxygen con-
tent of the collected sample.
4.3.5 Modified Saltzmann Method
4.3.5.1 Description
This method involves oxidation of NO to N02 in
the gas phase by the excess oxygen in the sample. The N02 formed is ab-
sorbed in the Saltzmann reagent. (The chemistry of the Saltzmann method
is discussed in detail in Section 4.5.1.) Davis and O'Neil (9) collect
exhaust gas directly in an evacuated flask containing Saltzmann reagent
as the absorbing solution. The diesel exhaust contains sufficient oxygen
C\Q% or more) to oxidize all the NO to N02 within an hour. The PHS method
involves grab sampling with a 100 cc syringe. The syringe is charged with
80 cc of Saltzmann reagent and 20 cc of gas is collected. The syringe is
used as the reaction vessel. After 24 hours, the absorbance is measured
at 550 nm and the NO concentration is computed.
rt
4.3.5.2 Areas of Application
(a) Diesel exhausts (9), power plants (590).
(b) Nitric acid plants (173).
4.3.5.3 Sampling Equipment
(a) 250 cc evacuated flask containing 25 cc of
Saltzmann reagent.
(b) 250 ml gas sampling tube or syringe.
75 WALDEN RESEARCH CORPORATION
-------
4.3.5.4 Applicable Concentration Range
Wide if grab sampling techniques are used to in-
crease the range.
4.3.5.5 Precision and Accuracy
Analysis of a series of 20 diesel exhaust samples
with the modified Saltzmann method gives a coefficient of variation of 3.5%
at 846 ppm NO for duplicate samples compared to 1.6% (838 ppm NO ) for
A X
the phenol disulfonic acid method. Accuracy when compared to mass spec-
trometry is the same as PDS, ±5%.
4.3.5.6 Interferences
A major problem in this method is the interfer-
ence from S02. A series of tests with synthetic mixtures of NO and S02
in air was conducted at NAPCA with the PDS and the Saltzmann methods (548).
The NO and S02 concentrations are typical of those found in power plants.
The PDS method shows no decrease in NO concentration as the S02 concen-
tration is increased. The apparent NO concentration in the Saltzmann
A
method, however, decreases markedly with concentration (Figure 4-4). This
may be due to the reaction:
N02 + H2S03 -»• H2S04 + NO
which occurs at the gas liquid interface (451). H2SO~ will be oxidized
by the peroxide in the PDS method. Thus, S02 would have to be removed
quantitatively for the Saltzmann method to product accurate results in
coal- and oil-fired power plants.
4.3.5.7 Comments and Conclusions
This method may be useful for a rapid determina-
tion of NOX In yas-fired power plants where S02 is not present. One should
be able to obtain a rapid estimation by this method provided sufficient
oxygen is present and the NO concentration is high. The stoichiometry
A
remains a major problem.
76 WALDEN RESEARCH CORPORATION
-------
4.3.6 Neutralization Methods
4.3.6.1 Description
The PHS hydrogen peroxide method (173) as well
as the modified Galliard (DuPont) method (173) involve collection of gas
samples in evacuated flasks containing dilute aqueous peroxide solution
as the absorbent. The nitrogen oxides are oxidized to nitric acid, then
determined by titration with standard sodium hydroxide solution using
methyl red (PHS) or methyl purple (DuPont) indicators. In the DuPont
method, a trace of cupric sulfate is added to increase the rate of ab-
sorption of nitric oxide.
4.3.6.2 Areas of Application
Nitric acid plants.
4.3.6.3 Sampling Equipment
(a) Evacuated flasks containing acidic peroxide
as the absorbent (PHS).
(b) 300-500 ml gas sampling bulb; sample col-
lected by displacement (DuPont).
4.3.6.4 Applicable Concentration Range
Has been used for high NO concentrations (> 500
^
ppm) only.
4.3.6.5 Precision and Accuracy
Given as ±0.12% at 20% NO based on 30 determina-
tions in a weekly period by the same analyst. DuPont (173).
4.3.6.6 Interferences
Since it is a non-specific method (determines
total acidity), any acidic (COp, S0p)» or basic compound in the flue gas
will interfere.
77 WALDEN RESEARCH CORPORATION
-------
4.3.6.7 Comments
In the PHS method, the sample flask is thor-
oughly shaken every ten minutes for the first half hour. After another
half hour,* the sample is titrated with 0.01N NaOH. The DuPont procedure
required the addition of peroxide and freshly prepared cupric sulfate
solution to the gas sampling flask. The contact time is one hour.
Another non-specific method"*" is that developed
by Johnson (450) for determining the oxides of nitrogen. This method in-
volves reacting the oxides of nitrogen with a "mixed acid" of HpSO. and
to form HNOSO,. After the sample is completely reacted, excess
standardized KMnO^ is added to oxidize the nitrous acid (HNOp) to nitric
acid according to the following reactions:
HNOS04 + H20 ^ H2S04 + HN02
2KMn04 + 5HN02 + 3H2$04 ?* 5HN03 + 2MnS04 + K^ + 3H20
This method has been used for high NO concen-
A
trations and is a non-specific method. Interferences such as S02 would
be a major problem in fossil fuel-fired power plants (oil and coal).
4.3.6.8 Conclusions
The hydrogen peroxide method (PHS) may prove use-
ful for a fast analytical readout in a gas -fired power plant where S02 is
not present and the NO concentration is high.
A
4.3.7 Reactive Solid Sorbents
Solid sorbents for collection of oxides of nitrogen offer
several advantages over conventional grab sampling techniques now in use:
Note the short oxidation time in this procedure is due to the very high
NO (several thousand ppm) concentrations found in nitric acid plants.
Not a neutralization method.
78 WALDEN RESEARCH CORPORATION
-------
(1) Integration over long sampling times (30 to 60 min-
utes) provides more reliable and accurate measurement of the source
strength. The resulting large sample permits replicate analyses to be
performed for improved precision.
(2) The number of individual samples required may be re-
duced compared to grab sampling techniques; L.A. APCD requires taking NO
A
samples in quadruplicate for PDS analyses.
(3) There is the possibility that a multi-pollutant col-
lector for both NO and SO may be developed.
A A
Peters and Straschil (570) have tested reactive solid sor-
bents for the simultaneous determination of NO and N09 or NO in mixed
* e. A.
nitric acid plant off gases. (Only determination of NO is discussed
A
here; determination of NO and N02 is treated in Section 5.) The sorbents
developed were AgMnO. and NaClOp supported on aluminum hydrate. These
supported oxidants have a high surface area and are reported to react
rapidly and quantitatively with both NO and N02. The resulting nitrate
is determined by Devardas method (reduction to NH-).
A brief discussion of potential solid sorbents for collec-
tion of SOp was given previously (Volume I). Much thermodynamic data and
some kinetic results are available for the sulfur oxides from studies on
the application of solid sorbents for S02 control processes. However,
little data are available for the oxides of nitrogen. The reactive sor-
bents which appear promising for total NO determination are the metal
A
oxide oxidants which were the most promising materials for S02 collection:
Pb02, Mn02 and Ce02<
The reactions of Pb02 with oxides of nitrogen have been
studied recently (391). Many previous workers have considered the reac-
tion of nitrogen oxides with Pb02 to be represented by the following
equations:
This method has not been used in combustion flue gases which contain
other active reducing agents.
79 WALDEN RESEARCH CORPORATION
-------
2NO + Pb02 = Pb(N02) • PbO + 0.5 02
and
2N02 + 2Pb02 = Pb(N03) • PbO + 0.5 02
Mishmash and McLoan (391) reinvestigated these reactions by gas chromato-
graphic and polarographic techniques. They found that chemical reaction
rather than physical absorption does indeed occur. Evolution of oxygen
was not observed; the product was divalent lead, as determined by polar-
ography and spot tests. Mishmash and McLoan (391) rewrite the reactions
as follows:
Nitrous oxide (N90)
Nitric oxide (NO)
Pb02 + N20 = N.R.
Pb02 + NO = PbO + N02
2N02 + Pb02 = Pb(N03)2
Nitrogen dioxide (NOp)
Pb02 + 2N02 = Pb(N03)2
Since they showed that oxides of nitrogen are quanti-
tatively collected by a Pb02 column, determination requires selection of
a suitable method for analysis of NOl in the Pb02 matrix. One possibility
is thermal decomposition (if it can be shown to be quantitative) as
f ol 1 ows :
2Pb(N03)2 -> 2PbO + 4N0
Trie N02 + 02 product resulting may be readily determined
by a variety of techniques. Direct analysis by dissolution or leaching
of the PbOp matrix is, of course, also feasible.
80 WALDEN RESEARCH CORPORATION
-------
A possible advantage (and complication) of this technique
is that SOp also reacts with the lead dioxide and would be retained as
lead sulfate. This is the principle of the sulfation candle used fre-
quently in ambient air SO^ measurements. Thus, from one test, one might
expect to obtain total NO., and SO., which are the major gaseous pollutants
-• — ~ ~~ ~ T' "•-"• •"-"-'••"• j{ X
from fossil fuel combustion. Although these solid sorbents clearly re-
quire much developmental work, there are many possible advantages in-
cluding ease of use and potential high precision multiple pollutant (NO
«
and SO ) determinations.
A
4.4 Analytical Methods Used for Determining Nitrate in Aqueous
Solutions
4.4.1 Nitrate Electrode (317,390,417,435,439,440,441,607,612)
4.4.1.1 Description of Method
The nitrate-selective specification electrode (a
liquid ion-exchange membrane electrode) and a reference electrode (calomel)
are inserted into the aqueous nitrate sample. The potential difference be-
tween the electrodes is directly related to the nitrate concentration in
the sample via a calibration curve prepared using aqueous nitrate standards,
4.4.1.2 Areas of Application
Nitrate in water (390,607) in plant materials
(417,441), soil extracts (435,440), and microbiological media (439).
4.4.1.3 Applicable Concentration Range
From approximately 2 to 2200 ppm N02 assuming a
1-liter flue gas sample and 10 ml of absorbent solution. Higher concen-
trations can be obtained by dilution with more absorbent.
4.4.1.4 Accuracy and Precision
Coefficient of variation (C.V.) approximately 3%
when using standards which bracket the unknown (317,417): C.V. = 2-3%
81 WALDEN RESEARCH CORPORATION
-------
(435) and 3-6% (440) from analyses of a large number of soil extracts,
and 1% when adding electrolyte to get high ionic strength (441). Use of
a solid-state fluoride electrode and standard addition reported to have
C.V. = 0.5-0.7 for measurement of nitrate solutions (612).
Electrode results agreed with the phenol disul-
fonic acid method within 10 to 15% in analysis of plant extracts (417),
water (607), and with 2,4-xylenol method in analysis of soil extracts
(435).
4.4.1.5 Interferences
Iodide, bromide, and some organic acids inter-
fere (390,417) but these can be removed by ion-exchange (417). Since the
electrode measures activity rather than concentration, changes in the
ionic strength of the sample solution will result in variations in appar-
ent concentration. This effect can be minimized by maintaining the ionic
strength of the solution at a constant level with 1% sodium sulfate. This
results in a slight loss in sensitivity (441).
The response to nitrite is only about 0.06 that
of nitrate (390).
4.4.1.6 Comments
(a) The procedure is very rapid.-' Measurement of
nitrate generally takes 1 to 2 minutes after equilibrium of the nitrate
solution with the electrode.
(b) Interferences such as bromide and iodide can
be removed via ion-exchange (417), if necessary.
(c) With a battery-operated, expanded scale pH
meter, field analysis of nitrate would be simple provided that the
absorption ah'd oxidation time of NO can be reduced comparably.
(d) Effects of flue gas constituents such as SOp
have not been checked.
82 WALDEN RESEARCH CORPORATION
-------
(e) Electrode response was not changed and sen-
sitivity decreased slightly when measurements were made in 0.07M sodium
sulfate solution (441); this would indicate that the absorber solution
now used with the PDS method could not be applicable.
(f) In the range of 2 to 9, pH appears to have
little effect on the response (317).
4.4.1.7 Conclusion
This method should be investigated as an alter-
native to the nitration of organic reagents methods now in use (see
Section 4.6).
4.4.2 1-Aminopyrene (7)
4.4.2.1 Description of Method
One volume of aqueous nitrate sample is added to
three volumes of 0.01% 1-aminopyrene in concentrated H^SO.. The mixture
is heated at 100°C for 15 minutes, cooled, and the absorbance is measured
with a spectrophotometer at 456 nm. The color is stable for at least one
hour.
4.4.2.2 Areas of Application
Aqueous extracts of airborne particulate matter.
4.4.2.3 Applicable Concentration Range
Up to 2.3 lag N02 (1.2 ppm N02 in 1 liter) so that
aliquoting is necessary, the absorbance concentration curve is non-linear
below ca. 1 yg NOp.
4.4.2.4 Precision and Accuracy
Coefficient of variation = 4% and 6% for repli-
cate measurement of 6.2 and 3.1 yg nitrate, respectively. No accuracy
data.
83 WALDEN RESEARCH CORPORATION
-------
4.4.2.5 Interferences
Nitrite gives nearly the same response as nitrate;
iron (III) gives positive interference, but less than l/20th that of
nitrate.
4.4.2.6 Comments
(a) The reagent is more sensitive but less pre-
cise than most other colorimetric methods.
(b) It is likely to be oxidant-sensitive as
shown by iron (III) reaction.
(c) Chemistry of reaction is unclear; lack of
adherence to Beer's law may mean mixed product.
(d) This procedure is reasonably rapid compared
to PDS and xylenols.
4.4.2.7 Conclusion
Since this high sensitivity is not needed for
total NO measurements in flue gases, the relatively poor precision and
/\
other problems make this method unattractive as an analysis method for
our purposes.
4.4.3 Chromatropic Acid (276,327)
4.4.3.1 Description of Method
Three ml of aqueous nitrate is mixed with 7 ml
of 0.01% chrcmatropic acid (in concentrated H2S04) and the resulting color
is measured spectrophotometrically at 415 nm. Addition of 1% HC1 to the
HpSO. in the reagent gave a higher absorptivity. An intense peak was re-
ported at 357 nm (327) but this was not present in other investigations
(276). The color is stable at least four days.
84 WALDEN RESEARCH CORPORATION
-------
4.4.3.2 Areas of Application
None tested.
4.4.3.3 Applicable Concentration Range
Five to 55 ppm per liter gas measured at 415 nm
in the absence of chloride; the 357 nm peak gives non-linear absorbance
versus concentration curves in the 0 to 5 ppm N0« range.
4.4.3.4 Precision and Accuracy
No data.
4.4.3.5 Interferences
Nitrite gives a positive interference. Other
interferences include: iron (III), bromide, chloride (if the chloride-
containing reagent is not used), formaldehyde, bromate, chlorate, and
iodate.
4.4.3.6 Comments
(a) The usefulness of the 357 nm peak is
greatly impaired by the non-linear calibration curves. This peak may,
in fact, be an artifact (276).
(b) The reagent appears to be oxidant-sensitive
since iron (III) and halates interfere.
(c) Chloride-containing reagent would eliminate
chloride interference found in most nitration reagents.
4.4.3.7 Conclusions
Chromatropic acid reagent might be useful if
formaldehyde and oxidants do not interfere. Initial laboratory work,
however, indicated that oxidants did interfere with this method. Another
problem was the very broad-;absorption spectrum of the compound. Thus,
this method was not investigated further (see Appendix 4).
85 WALDEN RESEARCH CORPORATION
-------
4.4.4 Strychnidine (331)
4.4.4.1 Description of Method
An aqueous sample is mixed with an equal volume
of Strychnidine reagent (0.03% in concentrated H2SO.). The red
color produced is allowed to develop for 20 minutes (±1 minute) and meas-
ured at ca. 540 nm with a spectrophotometer. The color is not stable.
4.4.4.2 Areas of Application
Trace nitrate in water.
4.4.4.3 Applicable Concentration Range
1.6 to 16 yg nitrate in a 5 ml sample.
4.4.4.4 Precision and Accuracy
No data.
4.4.4.5 Interferences
Nitrite.
4.4.4.6 Comments
The absorbance-concentration curve does not obey
Beer's law, and the increase of absorbance with time necessitates a timed
reaction.
4.4.4.7 Conclusion
The method is too tedious and problematic. It
is not useful as other organic reagents.
86 WALDEN RESEARCH CORPORATION
-------
:r4.4;5 Reduction of Nitrate to Nitrite with Cadmium (332)
4.4.5.1 Description of Method
An aqueous sample is buffered at pH 8 and passed
through a column of amalgamated cadmium to reduce nitrate to nitrite. The
nitrite is then determined by addition of sulfanilamide and 1-naphthyl-
ehtylenediamine and spectrophotometric measurement at 540 nm.
4.4.5.2 Areas of Application
Trace nitrate in sea water.
4.4.5.3 Applicable Concentration Range
0.06 to 0.6 mg nitrate per liter (equivalent to
from 8 to 85 ppm ML per liter using 25 ml absorbent).
4.4.5.4 Precision and Accuracy
Coefficient of variation = 2% for replicate
analyses of distilled water containing up to 0.6 mg nitrate per liter.
No data are given on the accuracy.
4.4.5.5 Interferences
Nitrite.
4.4.5.6 Comments
(a) The reduction is a slow procedure (10 to
15 minutes through column), and the cadmium amalgum is tedious to prepare.
(b) This multi-step procedure is prone to errors.
4.4.5.7 Conclusions
Not enough advantages over organic nitration
reagents to warrant consideration.
87 WALDEN RESEARCH CORPORATION
-------
4.4.6 Diphenylamine-diami nodiphenylmethane (267)
4.4.6.1 Description of Method
An aqueous nitrate sample is mixed with 0.05%
diphenylamine plus 2% diaminodiphenylsulfone in 60$ HpSO.-lO^ H-PO, and
ca. 5 mg sodium chloride to give a blue color which is measured spectro-
photometrically at 600 nm.
4.4.6.2 Areas of Application
Trace nitrates in H^SO,.
4.4.6.3 Applicable Concentration Range
0.25 to 3 yg nitrate per ml sample.
4.4.6.4 Precision and Accuracy
Coefficient of variation estimated to be 2 to
4%. No accuracy data.
4.4.6.5 Interferences
Peroxide is a positive interference but other
oxidants are not.
4.4.6.6 Comments
(a) The complex is very sensitive to tempera-
ture (extinction coefficient varies about 3% per degree C).
(b) Nitrite does not react with this reagent.
4.4.6.7 Conclusions
This method has no advantages over other organic
reagents; the temperature effect is very detrimental.
88 WALDEN RESEARCH CORPORATION
-------
4.4.7 Pi aminodiphenylsulfone (267)
4.4.7.1 Description of Method
An aqueous nitrate sample (1 ml) is added to Q.2%
4,4-diaminodiphenylsulfone in concentrated HgSO, (3 ml) and heated at 100°C
for two minutes. The solution is diluted to 25 ml with H20 and the color
is measured spectrophotometrically at 410 nm.
4.4.7.2 Area of Application
Aqueous nitrate samples.
4.4.7.3 Application Concentration Range
Twenty to 200 yg NOp per ml of absorbent (equi-
valent to 100 to 1,000 ppm N02 per liter using 10 ml of absorbent).
4.4.7.4 Precision and Accuracy
Coefficient of variation = 1 to 1.5% based on
replicate analysis of aqueous samples containing from 30 to 300 yg nitrate.
No accuracy data.
4.4.7.5 Interferences
Nitrite, bromide, halates give positive inter-
ferences: chloride negative interference; lower alcohols and esters
cause negative interference.
4.4.7.6 Comments
(a) Reagent sensitive to oxidants and halides.
(b) Relatively less sensitive to nitrate than
other reagents.
89 WALDEN RESEARCH CORPORATION
-------
4.4.7.7 Conclusions
Reagent offers no special advantages over other
organic reagents, and is apparently more subject to interferences.
4.4.8 Brucine (329)
4.4.8.1 Description of Method
An aqueous nitrate sample is mixed with aqueous
brucine and concentrated H^SO, to give a final mixture containing 0.04%
brucine in 50% H^SO,. The mixture is stored in the dark, cooled to 30°C,
and measured spectrophotometrically at 410 nm.
4.4.8.2 Areas of Application
Nitrates in soil extracts and water.
4.4.8.3 Applicable Concentration Range
Up to 50Y N02 (25 ppm N02 per liter sample).
4.4.8.4 Precision and Accuracy
No data.
4.4.8.5 Interferences
Nitrite is a positive interference but chlorides
do not interfere.
4.4.8.6 Comments
The color is sensitive to reaction temperature,
concentration of the acid, and exposure to light.
4.4.8.7 Conclusions
Method has no advantages over other organic
reagents, and is considerably more tedious.
90 WALDEN RESEARCH CORPORATION
-------
4.5 Analytical Methods Used for Measuring Nitrite Ion in Aqueous
Solutions
4.5.1 Griess-Saltzmann Diazotization Coupling Reactions
4.5.1.1 Description
This method as generally used involves absorp-
tion of the N(L or N203 directly in the coupling reagent. Saltzmann's
reagent has been the most widely used method for atmospheric monitoring
(1,9,77,85,97,100,105,133,144,151,187,191,375,387,420,436). The
Saltzmann reagent consists of 0.5% sulfanilic acid and 0.002N-(l-naphthyl)
ethyenediamine in aqueous 14% acetic acid (133). Other coupling agents
used include 1-naphthylamine (modified Griess Ilsovay) (176), phenol
(used with sulfanil amide) (241), and N-(l-naphthyl) ethlenediamine
dihydrochloride in aqueous HgPO. (142,251) coupled with sulfanilamide.
After a short time (ranging from 15 to 60 min-
utes), the absorbance of the dye (formed by the diazotization reaction
initiated by HNOp) is measured at 540-560 nm with a spectrophotometer.
4.5.1.2 Areas of Application
Used mainly for atmospheric monitoring but mod-
ified methods have been used for diesel exhausts (9) (Section 4.3.5),
oxides of nitrogen in flue gas (241), and nitric acid plants (Section
4.3.5).
4.5.1.3 Sampling
(a) Bubblers containing coupling reagent
(b) Bubblers containing 0.1N NaOH (241,172).
(c) KMnO^ bubblers plus (3) bubblers contain-
ing Saltzmann reagent in series (436).
(d) (Grab sampling technique described earlier
in Section 4.3.5.)
91 WALDEN RESEARCH CORPORATION
-------
4.5.1.4 Applicable Concentration Range
Primarily for low N02 levels (with bubblers),
however, the concentration range has been extended with grab sampling
techniques for the higher NO concentrations (9,77).
J\
4.5.1.5 Precision and Accuracy
Collection in three bubblers in series - coeffi-
cient of variation (C.V.) is 1.3-2.6% for gases containing 15 and 31 ppm
N02 (436). Collection in 50 ml syringe - C.V. = 2-3% for 1260 ppm N02 in
air (77).
Accuracy is very much in doubt because of the un-
certain stoichiometry of hydrolysis and reactions of N02 and NO with the
coupling reagent. Amounts of N02 required to give absorbance equal to
known concentrations of NaN02 vary from 0.60 to 1.0 for up to 2 ppm N02
in air (77,109,133,151,191) using bubbler samples, and from 0.77 to 0.57
for 500 to 1400 ppm N02 (77,97) using grab samples. A recent grab sample
study (611) has also shown the change in factor from 0.85 to 0.65 when
NOp concentration increased from 500 to 2800 ppm.
Various investigators have compared the color
intensity from mixtures of N02 and air (or N2) with the color intensity
from known amounts of NaN02 (in solution). Representative results il-
lustrating the lack of agreement found using Saltzmann's reagent are
summarized in Table 4-3.
4.5.1.6 Interferences
S02 results in loss of color, for example: 15 ppm
S02 causes an 8% loss in 100 ppm N02 (436); 90 ppm S02 caused 4% loss in
2-3 ppm N02 (133), but the error could be minimized by measuring the solu-
tion within 45 minutes; 900 ppm S02 caused 11% loss in 2-3 ppm N02 (133).
Addition of 1% acetone cuts loss due to S02 by a factor of 4 (133).
92 WALDEN RESEARCH CORPORATION
-------
10
TABLE 4-3
REPORTED STOICHIOMETRICS FOR SALTZMANN REAGENT
Sampling Method
Bubbler
Bubbler (change in
reagent)
Bubbler
Bubbler
Flask (syringe)
Bubbler
Flask
N02 Cone. , ppm
< 1
< 1
1-2
'X/ 1
5
1200
< 1
500
2800
NOo Source/ Carrier D
i Gas R
N02 cylinder/air
N02 cylinder/air
N204/air
N02 in air via dilution
in bottles
N02 from PbN03 in carboy
N02 from PbN03 in carboy
N02 in air stream
N02/air in carboy
N02/air in carboy
Mole NaN09
-4.,' .» . £•
atl°* Mole N02
0.72
0.72
0.95
0.60
0.69
0.57
~ 1.0
0.85
0.65
Reference No.
133
85
- 191
109
77
71
151
611
611
-------
NO does not react with the coupling reagent but
it will if oxidized by Og. Light exposure causes 20-25$ loss in color
after 24 hours (436). Storage in dark at 0°C reduces losses.
4.5.1.7 Comments
SOp interferences in flue gases is a major prob-
lem (Figure 4-4) with this method.
Many changes in the Saltzmann method have been
made, i.e., use of isopropanol (100) instead of acetic acid, etc. (with-
out enough extensive testing).
The stoichiometry problem has not been resolved
to date.
4.5.1.8 Conclusions
This method has not been checked sufficiently for
effect of real flue gas components. However, the SOo interference at levels
found in flue gases has been shown to be a major problem of this technique.
The S02 interference may be controlled by absorbing the gases directly in
caustic rather than Saltzmann's reagent (see Section 3.3). Extent of oxi-
dation of NO during sampling and reaction not known. Calibration (stoichi-
ometry) factor for use of aqueous nitrite standard is uncertain; individual
operators should check their results against known gas mixtures.
Overall, the method should not be used for flue
gases which contain sulfur dioxide.
4.5.2 Other Diazotization-Coupling Reagents for Nitrite
4.5.2.1 Description
These reagents are generally composed of a diazo-
tizable amine and a coupling agent (also generally an amine) which may be
added simultaneously or sequentially to the aqueous nitrite sample. The
resulting colored azo compounds are measured spectrophotometrically.
94 WALDEN RESEARCH CORPORATION
-------
Sawacki, et al. have published two comprehen-
sive reviews and listings of diazotization reagents for nitrite (326,
362). Table 4-4 lists the reagents noted by them (326) and gives the
colors which are stable for at least one hour and capable of detecting
6 micrograms of nitrite (equivalent to 1 liter of gas containing 3 ppm
(v/v) of N02). Saltzmann's reagent is included for comparison.
Shinn used sulfanilamide and N-(l-naphthyl)-
ethylenediamine for the colorimetric measurement of trace aqueous nitrite
in the range of 0.05 to 1 microgram per ml sample (177).
Another reagent proposed for nitrite is a mix-
ture of p-diamino-diphenylsulfone and dipheny1 amine in ethanol contain-
ing nitric acid which gives a yellow color (277). The Beer's law range
is 0.3 to 8 micrograms nitrite per ml of sample.
The use of (4-aminophenyl)-trimethylammonium
salts as the diazotizable amine and N,N-dimethyl-l-naphthylamine as the
coupling agent has been proposed as a reagent for aqueous nitrite with
a Beer's law range of 3 to 45 micrograms per ml in a sample (255).
4.5.2.2 Application
Most of these reagents have not been tested to
any great extent. The p-nitroam'line/azulene reagent has been applied
to measurement to trace nitrite in water (268). This reagent also gave
a similar color when N02 was bubbled through it, but no quantitative
measurements were made.
4.5.2.3 Applicable Concentration Ranges
As noted under 4.5.2.1, the reagents are the-
oretically capable of measuring 3 ppm N0~ per liter sample.
4.5.2.4 Precision and Accuracy
No data on flue gas analysis.
95 WALDEN RESEARCH CORPORATION
-------
I
5
m
z
TO
m
C/J
•so
o
8
TO
3
TABLE 4-4
DIAZOTIZATION-COUPLING REAGENTS FOR NITRITE
Amine
Coupler
Maximum
(nm)
Beer's Law Range
(yg N02 in
sample taken)
to
CT)
p-phenylazoaniline
4-nitro-l-naphthylamine
p-phenylazoaniline
p-phenylazoaniline
p-nitroaniline
anthranilic acid
p-nitroaniline
sulfanilic acid
(Saltzmann's reagent)
1-naphthylamine 643
1-anilinonaphthalene 695
benzaldehyde + 2-benzothiazolyhydrazone 590
none added 595
azulene 515
1-anilinonaphthalene 571
benzaldehyde + 2-benzothiazolylhydrazone 595
N-(1-naphthyl)-ethylene-di ami ne 550
0.7-20
0.6-15
0.8-20
0.8-16
0.4-9
0.4-9
^ 1-20
2.5-60
-------
4.5.2.5 Interferences
Oxidants such as chlorine are general interfer-
ences.
4.5.2.6 Comments
Any one of these reagents might be used for
measurement of nitrite, once it is trapped, but none would appear to
hold any advantage over the Saltzmann reagent with regard to stoichi-
ometry.
4.5.2.7 Conclusions
No reason to prefer any of these over the
Saltzmann reagent.
4.5.3 Spectrophotometric Determination of Nitrite in Aqueous
Solutions (69.621)
4.5.3.1 Description
The nitrite ion has been measured directly in
aqueous solutions by spectrophotometry at 355 nm (69) after absorption
of NCv, in 10% potassium hydroxide solution. The simultaneous measurement
of both nitrate and nitrite in aqueous solution at varying pH's has been
reported (621).
4.5.3.2 Areas of Application
High concentrations of N(L in N« (69).
4.5.3.3 Applicable Concentration Range
Used for measurement of a few hundred up to
several thousand ppm N02 (69).
4.5.3.4 Precision and Accuracy
No useful data.
97 WALDEN RESEARCH CORPORATION
-------
4.5.3.5 Interferences
This is actually a non-specific method since any
substance which absorbs in the near UV spectral region is a potential in-
terference (aromatic compounds, ketones, etc.).
4.5.3.6 Comments
Collection in strongly alkaline solution ap-
parently gives a constant stoichiometry ratio of 0.5 mole nitrite per
mole NOg (69), which would be an improvement over the direct collection
situation where the stoichiometry is uncertain (see Section 3.3).
The SOo, C0«» and other acidic species in the
flue gases will be rapidly absorbed by the alkaline solution. These may
result in changes in alkalinity, and, hence, in stoichiometry.
4.5.3.7 Conclusions
Probable interferences in flue gases as well as
the limited sensitivity make this method unattractive for present purposes.
4.6 Nitrate Electrode Studies
4.6.1 Introduction
In view of the tedious nature of the phenol disulfonic
acid (PDS) technique for NO (discussed earlier), several alternate ap-
J\
preaches were considered for the determination of nitrate ion in aqueous
solutions. The chromatropic acid method seemed to be an obvious improve-
ment, however, preliminary laboratory work (Appendix 4) was not encour-
aging. Therefore, considerable effort was spent on a feasibility study
of the nitrate electrode. This included ease of maintenance, durability,
reliability, accuracy, sensitivity, and the relative interference pro-
duced by other species. Since sulfate proved to be a problem in a pre-
vious study (548), we evaluated a variety of methods for the removal of
sulfate ion from the solution. The results are discussed in the follow-
ing sections.
98 WALDEN RESEARCH CORPORATION
-------
4.6.2 Electrode Assembly
Work has been conducted with an Orion Research nitrate
electrode (Model No. 92-07). No substantial problems were encountered in
assembling and filling the electrode. However, it is necessary to center
the membrane disk in order to get an effective seal on the exchanger
reservoir.
4.6.3 Reference Electrodes
In the initial work we used a saturated calomel electrode
(SCE) as the reference electrode. In very dilute nitrate solutions (10 M),
a sleeve-type electrode repeatedly produced discernible drift toward more
negative potentials (increasing nitrate activity). This drift was prob-
ably due to leakage of the concentrated KC1 electrolyte, since chloride
ion does give a response approximately 0.06 times that of nitrate. A
fiber-tip SCE, however, was found to perform satisfactorily.
4.6.4 Electrode Response and Performance
The sensitivity of the nitrate electrode and the repeat-
ability of the nitrate electrode-SCE pair were evaluated by measuring the
potential developed in dilute (10 to 10 M) solutions of nitric acid in
distilled water prepared fresh daily by diluting l.OM stock solution. The
measurements were made with either a Beckman Expandomatic pH meter or an
Orion Model 701 digital millivolt meter.
Immediately after assembly, the electrode often had a
sensitivity of 59 to 60 mV per tenfold change in nitrate concentration
over the range of 10"1 to 10"4M and slightly less (ca. 45-50 mV) in the
-4 -5
10 to 10 M range. After a few measurements had been made, the sen-
sitivity decreased to 55 to 56 mV per tenfold concentration change and
remained at this level for at least two weeks. This sensitivity of 55
to 56 mV is in the range previously reported for the electrode.
Table 4-5 shows the variation in response of the elec-
trode vs concentration found in measurements of the nitric acid
gg WALDEN RESEARCH CORPORATION
-------
TABLE 4-5
REPRODUCIBILITY OF NITRATE STANDARDS
o
o
I
JO
m
m
I
8
JO
3
1
O
Readings
Date
4/9
4/10
4/13
4/14
4/16
4/24
4/30
5/1
a.
Run No.
A
B
A
B
C
A
A
B
A
B
C
A
B
C
A
A
B
a(mV)b
Calculated
(mV)
Calibration Slope3
"Error" in Measuring 10"
Calibration Solution0
Nitrate Concentration
10"2M 10"3M 10"4M A B C Av. From
Slo
38.
37
34.
23.
22
17.
9.
14
46
46.
46.
46.
45.
45
43.
43.
43
3.
from:
5
5
5
5
5
5
5
5
5
5
5
4
m\
92.
90.
87
76.
74.
67.
63.
67
102
101.
101.
103
100
100
98
99.
98
3.
'lO~4M
5
5
5
5
5
5
5
5
5
4
1 "
144.
140
138.
129
127
115.
122.
123.
157.
157.
157
157
157.
157
155
155.
156.
3.
5 53 51.5 52.3
5
52 53.3 52.5 52.6
5 49 (49)
5 56.5 54.8 55.7
5
5 55.8 55.5 55.3 55.5
5 55.3 56 56 55.8
55.8 (55.8)
j? 56 56.8 56.4
1
91
89
87
76
74
66
65
101
102
102
102
101
101
99
100
99
1 c* Measured Deviation
92
90
87
76
.5 74
.5 67
.0 63
.5 102
101
101
.5 103
.5 100
100
.5 98
.5
.5
.5
.5
.5
.5
.5
.5
99.5
.5 98
Average "error"
Standard Deviation
+1.5
+1.5
0
+0.5
0
+1
-1.5
+0.5
-0.5
-0.5
+0..5
-1.5
-1
-1.5
-0.5
-1.5
= -0.2 mV
= ±1.0 mV
-------
TABLE 4-5 (continued)
I
i
30
m
CO
s
TO
S
JO
b. Calculated from:
(, - X)2
s =*f n
E d.f.
where: X. = individual readings
1( = average mV reading for that day
d.f. = number of degrees of freedom for that day (number of measurements -1)
_2
c. Assumes that 10 M calibration point was accurate. Calculated by adding the average calibration
slope (mV per 10-fold concentration change) to the 10~2 M readings and rounding to the nearest
0.5 mV.
-------
calibration solutions over a two-week period. In performing this study,
-2 -3
measurements were made on aqueous solutions containing 10 , 10 , and
-4
10 M nitric acid concentrations. When the response became erratic, we
renewed the electrode. It can be seen that the electrode performs very
reproducibly.
With a slope of 56 mV change in potential per tenfold
change in nitrate concentration, a difference of 1 mV corresponds to an
error of 4.3% in concentration. Since, as shown in Table 4-5, the ab-
solute potential can vary over a few mV, it will be necessary to check
calibration standards frequently in order to minimize measurement un-
certainty.
The standard deviation of the variation in potential
readings for any one day is nearly the same (3.4-3.1 mV) for all three
concentration ranges. This represents more than 10% error, clearly
showing that frequent standardization during the day will be needed in
order to achieve acceptable precision (i.e., within a few percent rela-
tive error). The benefit of frequent "bracketing" calibration is shown
by a calculation of the "error" (uncertainty) in the measurement of the
_3
10 M calibration solution. For this calculation, we have assumed that
_p
the 10 M solution was accurate, and calculated the mV reading expected
_3
for 10 M nitrate. The average error found (-0.2 mV) and the standard
deviation (±1.0 mV) for the "error" in any one measurement, are much
lower than those of the raw data. As noted previously, for a slope of
56 mV per tenfold concentration, a 1 mV error corresponds to a relative
error of 4.3% in concentration.
In connection with methods for determination of N02 by
caustic absorption (Section 5.3), the response of the nitrate electrode
was determined at pH 11.5. The system was calibrated by preparation of
standards containing known concentrations of KNO, dissolved in .005N NaOH.
The response observed in NaOH is very close to that in pure water to con-
centrations down to 10" M nitrate ion. At lower concentrations of nitrate
ion in the caustic solution, the response (millivolts/decade) decreases
102 WALDEN RESEARCH CORPORATION
-------
rapidly so that the effective lower limit for the analysis at pH - 11.5
is ~ 5 x 10~4M nitrate ion (EMF vs .005N NaOH * 20 mV). In neutral (pH *
7.5) and acid (pH = 4.5) solutions the EMF response remains linear vs log
(NOl) down to 10 M, dropping rapidly at lower concentrations so that the
effective sensitivity limit (EMF vs nitrate-free solution - 20 mV) is
10" M NOl. In dilute basic solution, electrode response in the range 10
to 10"3M
.
NOj is Nernstean (58 mV/decade).
It is necessary to stir the solutions in order to obtain
steady potentials. If the solutions are not stirred, the potentials tend
to drift in a negative direction by 10 to 20 mV.
4.6.5 Direct Measurements on Solutions Used for Collection of
Nitrogen Oxides
The data of Table 4-6 show the response of the nitrate
electrode to various concentrations of nitric acid in 0.1N sulfuric acid
containing 0.03% hydrogen peroxide (the absorbing solution used in the
phenol disulfbnic acid (PDS) method for nitrogen oxides).
TABLE 4-6
ELECTRODE RESPONSE IN 0.1N SULFURIC
ACID/0.03% HYDROGEN PEROXIDE
HN03
Concentration (M)
Electrode Response (mV vs SCE)
0.1.N H2S04/0.03% H202 H20
ID'1
ri
_ *j
10 i
10'3
10'4
r
10'5
Reading "
- 28
+ 35
+ 91
+123
+123
Slope Factor
63
56
32
0
Reading "
- 18.5
+ 38.5
+ 92.5
+144.5
+191
Slope Factor"
57
54
52
46.5
103
WALDEN RESEARCH CORlPORATlON
-------
The response of the electrode to the same nitric acid con-
centrations in aqueous 0.03% hydrogen peroxide was identical to that for
water alone, so that the sharp decrei
apparently due to the sulfuric acid.
_3
water alone, so that the sharp decrease in response below 10 M nitrate is
This decrease can probably be attributed either to re-
sponse of the electrode to sulfate and bisulfate (the electrode response
to sulfate is approximately 0.006 times that for nitrate). It is ap-
parent from these data that direct measurement of the PDS scrubber solu-
tion is not very useful for low nitrogen concentrations, since the lower
limit of 10 nitrate necessary for accurate measurement corresponds to
approximately 220 ppm of nitrogen oxides for a 1-liter gas sample ab-
sorbed into 10 ml of solution.
4.6.6 Effect of Sulfate on Nitrate Electrode Response
In order to gain a clearer idea of the extent to which
the presence of sulfate influences the response of the nitrate electrode,
the potential of this electrode was studied in a series of aqueous mix-
tures of HpSO* and HN03, designed to simulate the absorbing solution from
a range of flue gas compositions. The results, shown in Table 4-7, indi-
cate that in the range of mole ratios of 50:1 and 100:1 sulfate to nitrate,
the sulfate begins to cause substantial positive error in the measured
nitrate concentration. This interference is presumably due to a positive
response of the electrode to sulfate rather than to an ionic-strength ef-
fect, since increasing ionic strength (more sulfate) should cause a de-
crease in apparent nitrate activity. The 50:1 sulfate:nitrate limit cor-
responds to a flue gas containing approximately 3500 ppm sulfur oxides
and 70 ppm nitrogen oxides, so that sulfate interference may be a prob-
lem only in the extreme case of high sulfur fuel and very low values of
nitrogen oxides.
This problem of a serious positive interference by sul-
fate at molar ratio of greater than 50:1 over nitrate has also been noted
by Orion Research. In view of the interference from sulfate, we tried a
number of approaches for the removal of sulfate. These are described in
the following section.
1 04 WALDEN RESEARCH CORPORATION
-------
TABLE 4-7
EFFECT OF SULFATE ON NITRATE ELECTRODE MEASUREMENT
Added
o
en
|
TO
m
1
CORPORATIC
H2S04 (M)
2.5 x 10"4
1.25 x 10"3
2.5 x 10"4
2.5 x 10"4
2.25 x 10"3
5.0 x 10"3
5.0 x 10"3
1.0 x 10"2
1.0 x 10"2
2.0 x 10"2
5.0 x 10~3
1.0 x 10"2
a. ppm (y&
b. Sulfate
HN03 (M)
1.0 x 10"3
1.0 x 10"3
1.0 x 10"4
1.0 x 10"3
1.0 x 10"4
1.0 x 10"4
5.0 x 10"5
1.0 x 10"4
5.0 x 10"5
5.0 x 10"5
b l.OxlO"4
b 5.0 x 10'5
/£) based on a 1
added as «2S04
Found
HN03 (M)
1.0 x 10"3
9.9 x 10"4
1.15 x 10"4
9.7 x 10"4
9.4 x 10"5
8.5 x 10"5
1.03 x 10"4
1.6 x TO"4,
1.52 x 10'4
1.4 x 10~4 -
1.15 x 10'4
1.7 x 10"4
1.03 x 10"4
7.7 x 10"5
liter sample absorbed
instead of H2S04.
S and N oxides
added - expressed
as ppm in 1 liter
sample3
% Difference
in HN03
0
- 1
+ 15
- 3
- 6
- 15
+ 100
+ 60
+ 52
+ 180
+ 130
+ 240
+ 3
+ 55
into 25 ml of
Mole Ratio
S0=/N03
0.25
1.25
2.5
2.5
22.5
50
100
100
200
400
50
200
solution.
S°x
140
700
140
140
1260
2800
2800
5600
5600
11200
2800
5600
NOX
560
560
56
560
56
56
28
56
28
28
56
28
-------
4.6.7 Sulfate Removal
Sulfate may be removed by precipitation of an insoluble
sulfate salt, for example, with barium. Since the electrode also responds
to many anions, we decided to use a system where the end product was an
anion which gave a minimum response, e.g., monohydrogen phosphate. An
excess of solid BaHPO. (prepared from barium hydroxide and phosphoric
acid) was added to aqueous solutions of nitric acid of various concentra-
tions, and to two concentrations of sulfuric acid containing nitric acid,
and then the potential of the nitrate electrode was measured for each so-
lution. The data are shown in Table 4-8.
The data in Table 4-8 indicate that removal of sulfate by
direct addition of barium phosphate, or any more soluble barium salt, does
not allow accurate measurement of nitrate with the electrode. It was
thought that barium carbonate might work as a substitute because the lib-
erated CCL could be removed by boiling of a dilute acid (0.001M) solution.
Aqueous solutions containing various concentrations of HpSO. and HNO« were
prepared to cover a range of concentrations of sulfate and nitrate, and
the apparent nitrate content was measured with the electrode. These solu-
tions were treated with powdered BaCO- (0.6 g per 25 ml of solution),
boiled for four minutes, cooled, diluted back to volume, and measured
again with the nitrate electrode. The results are given in Table 4-9.
The data in Table 4-9 indicate that addition of
and subsequent boiling, did not result in a simple and immediate removal
of sulfate. The increase in the measured nitrate concentration in two
tests may be an indication that the boiling period was too short to re-
move all the bicarbonate, although Kolthoff and Sandell reported that
five minutes of boiling is sufficient to remove 5 x 10" moles of COp
from 50 ml of solution. In any case, this does not appear to be an at-
tractive approach for removal of sulfate.
To test the possible use of anion exchange for selective
removal of sulfate from the absorbent solutions, we used a strong base
resin, to hold up sulfate and other multivalent ions and let nitrate pass
106 WALDEN RESEARCH CORPORATION
-------
TABLE 4-8
EFFECT OF ADDITION OF BaHP04 ON ELECTRODE RESPONSE
Concentration
of HN03 (M)
10'1
ID'2
ID'3
5 x 10"4
10-"
5 x 10"5
H20
-8.5
+46
+102.5
V
+157.5
+171.5
H20 +
BaHP04
-6.5
+50.5
+103.5
V
+157.5
+171.5
0.012 M
H2S04a + BaHP04
V
V
111.5
+128
+173.5
V
0.0036 M
H2S04a + BaHP04
V
V
+110
+129.5
+168
V
Average Apparent
Nitrate Concentration
(M)
--
-_
7 x 10'4
3.3 x 10'4
5.3 x 10"4
—
% Error
_ _ •
—
-30%
-35%
-45%
—
a. HpS04 concentrations correspond to those expected for absorption of 1 liter samples containing
approximately 2700 and 2400 ppm of S02 absorbed into 10 ml and 30 ml (0.012 M and 0.0036 M,
respectively) of absorbing solution.
b. Calculated from the average mV reading for the two H2$04 solutions.
V = Not measured
o
S
3
-------
I
o
m
z
TO
o
8
TO
s
TO
TABLE 4-9
EFFECT OF BaC03 ON MEASUREMENTS WITH THE NITRATE ELECTRODE
o
00
Added
H2S04
2.5 x
1.25 x
2.5 x
2.5 x
5.0 x
1.0 x
1.0 x
(M)
ID"4
ID'3
ID'4
io-3
io-3
io-4
ID'2
HN03
1.0 x
1.0 x
1.0 x
1.0 x
1.0 x
1.0 x
5.0 x
(M)
io-3
ID'3
io-4
ID'3
io-4
io-4
io-5
Mole Ratio
SO:/NO~
*t o
0.25
1.25
2.5
2.5
50
100
200
HNO
1.0
9.9
1.15
9.7
8.5
1.5
1.15
Before
3
x 10"3
x 10"4
x 10"4
x 10"4
x 10"5
x 10"4
x 10"4
HN03 Found
BaC03
% Error
0
-1
+15
-3
-15
+50
+130
HNO
9.9
9.7
1.3
9.3
1.7
5.0
8.8
3
x
x
X
X
X
X
X
After
(M)
ID'4
io-4
io-4
io-4
io-4
ID'5
io-5
BaC03
% Error
-1
-3
+30
-6
+70
-50
+76
-------
2
through for measurement. A 9 cm x 0.8 cm column of Amber!ite IRA-410
(20/50 mesh) was washed with IN HC1, H20, HN NaOH and H20 until the ef-
fluent was neutral and free from chloride.
To test the ion-exchange take-up of sulfate, we applied
10 ml of 0.025M ICSO. (equivalent to 1 liter of gas containing 5000 ppm
S02) and then washed with 20 ml of H20. After 4 ml of KgSO^j solution
had passed through, the column effluent became strongly alkaline, in-
dicating displacement of hydroxide from the resin, but no sulfate was
detected in the effluent. At the end of the 20 ml H20 wash, the column
effluent had become neutral.
A solution (1 ml) of 0.01M KNOj (equivalent to 1 liter of
gas containing 225 ppm NO) was applied to the column, and washed with 10
ml of H20. After 3 ml of wash had been applied, the column effluent be-
came alkaline (pH greater than 10), indicating that hydroxide was being
displaced by nitrate. Since the hydroxide form of the resin takes up
nitrate as well as sulfate, this anion exchange approach does not appear
to be feasible.
4.6.8 Alternate Approaches for Sulfate Removal
An alternate approach to the nitrate electrode was needed
since the original PDS absorbing solution gave increased sensitivity at
the high sulfate levels (0.1N) and methods for the removal of sulfate by
precipitation with barium salts or ion exchange proved unsuccessful. The
effect of sulfate on the nitrate electrode potentials (Table 4-7) showed
clearly that fairly high levels of S02 (sulfate) could be tolerated with-
out major interference with the nitrate electrode potential. Therefore,
we investigated the use of the nitrate electrode with a new absorbing so-
lution, e.g., neutral 0.1% hydrogen peroxide. Samples were taken in a
gas-fired home heating unit using the neutral peroxide absorbing solution.
In order to check this technique, PDS samples were taken at the same time.
The results are shown in Table 4-10 below.
109 WALDEN RESEARCH CORPORATI&N
-------
TABLE 4-10
COMPARISON OF NITRATE ELECTRODE
RESULTS WITH PDS
ppm NOX
(nitrate electrode)*
25
27
ppm NOX 0/
(PDS) '"
37
44
&1 ectrode
Recovery . «\4<-
67
62
100
Neutral (pH 7.0) peroxide absorbing solution
These low results are presumably due to the incomplete
oxidation of nitrite in neutral solution. Therefore, we acidified the
solution with sulfuric acid and repeated the measurement. For a pH ad-
justed to the range 2 to 3, we obtained good agreement between the
nitrate electrode and PDS measurements (Table 4-11)
TABLE 4-11
COMPARISON OF NITRATE ELECTRODE AND
PDS RESULTS FOR NO.
ppm NOX ^ ppm NOX
(nitrate electrode) (PDS)
41 37
42 44
*pH 2.5
In view of the good agreement between nitrate electrode
and PDS, we ran a number of tests on the nitrate electrode. These are
described in the following section.
110 WALDEN RESEARCH CORPORATION
-------
4.6.9 Test Results for the Measurement of NQy in Flue Gas Samples
with a Nitrate Ion Selective Electrode
In order to determine the feasibility of the nitrate elec-
trode for measuring nitrogen oxides (as nitrate following oxidation and
absorption steps), we have compared this technique to the phenol disulfonic
acid (PDS) method for several combustion sources. The samples were col-
lected in 2-liter evacuated flasks. The absorbing solution for the nitrate
electrode samples was 25 cc of Q.OQ3N H2$0.-0.3% H202. The PDS absorbing
solution was 25 cc of 0.1N H2S04-0.06% H202. The flue gas samples were
collected in the evacuated flasks and allowed to stand overnight. The
nitrate electrode samples could be measured immediately after absorption
was complete. The PDS samples had to be evaporated, reacted, etc., and
required 5 to 6 hours before the analysis was complete. The results of
these tests are given in Table 4rl2. The agreement between the two methods
is excellent, i.e., within ±4% of the mean of the two methods.* When one
considers that this difference represents errors in sampling as well as
analysis, these results are encouraging.
Note that neither method yields results which are consistently low.
Ill WALDEN RESEARCH CORPORATION
-------
TABLE 4-12
COMPARISON OF PDS AND NITRATE ELECTRODE RESULTS
FOR N0 IN COMBUSTION EFFLUENTS
Test No , ppm N0* , p?m N9* Source
lest N0- (nitrate electrode) (PDS) bource
1 42 37 Gas-fired home heating
9 /n HA uni't run at low excess
* ^ ^ air
3 229 217 Tangentially-fired (oil)
, ,cQ 1Q1 boiler at 500 MW power
4 lbb Iai plant with 1310 ppm S02
5 137 155 in effluent
6 132 129 Oil-fired unit (home
7 llfi 1n, heating) with 150 ppm
1 M0 lw
. S0 in the effluent
8 52 53
112
WALDEN RESEARCH CORPORATION
-------
5. SAMPLING AND ANALYSIS METHODS FOR THE DETERMINATION OF NOo AND/OR NO
••—^~ £
5.1 Introduction
Although a number of ML measurements in fossil fuel-fired ef-
fluents have been reported (512,590), it is not clear that the measure-
ments made were, in fact, NCL measurements. In one case (512), the N02
determination was by the PDS method, apparently based upon rapid absorp-
tion. N02 (?) concentrations were reported as 6 to 8% of the NO
concentrations (589). .The contact time is quite important as indicated,
e.g., Equation 12 (Section 3.3.1.1). The time is very crucial. Unless
very short absorption times were used, < 2 min., the error due to oxida-
tion of NO would be significant. Unfortunately, we cannot determine the
procedures utilized.
Another method, used by Lee, et al. (590) determines N02 (?) by
the modified Saltzmann method (9). The major problems in application of
this method, however, are major interference by S02 (see Figure 4-4) and
the uncertain stoichiometry. A further problem in these measurements is
presented by the contact time. Presumably, the contact time is long since
their results reveal that N02 (?) concentrations apparently increase with
the initial oxygen concentration, suggesting the possibility of NO oxida-
ti on.
Similar critical evaluations could be made where N0/N02 measure-
ments have been attempted by others. Insofar as we have been able to de-
termine, no reliable NCL measurements have been made in power plants
effluents.
5.2 Sampling
For collection of total NO samples, very simple probes have been
A
used (Section 4.2) since NO is the major nitrogen oxide present and its
solubility in water is very small. This procedure is not valid where N02
is sought, since it is both a minor component and is, in addition, fairly
soluble in water. Therefore, for collection of a sample for NO or N02
determination, as opposed to a total NO sample, the probe should be
A
113 WALDEN RESEARCH CORPORATl6N
-------
heated to stack temperatures to prevent changes in the gas composition by
loss of N02 due to absorption in the condensate.
5.3 Sample Collection and Analysis
Several options exist for collection and analysis of NO and ML.
These include use of solid sorbents, aqueous absorbents or kinetic analy-
sis. The alternatives are discussed in the following sections.
5.3.1 Solid Sorbents
Peters and Strachil (570) have developed and tested two
classes of solid sorbents: supported oxidants (discussed in Section 4.3.7
for total NO determination) and supported alkali carbonates. Sodium car-
A
bonate supported on an aluminum hydrate substrate reacts rapidly and quan-
titatively with N02 and (NO + N02 = N203) but does not react with excess
NO (570). The supported carbonate must be "poisoned" to prevent catalytic
oxidation of "excess" NO by air.
The principle of the analysis is based upon the formation
of equimolar amounts of N02 and NOZ upon absorption of N02 in alkali, com-
pared to the formation of pure N02 upon absorption of N20o. Peters and
Strachil (570) determine N02 by KMn04 titration and total NOg (after KMn04
oxidation) by Devardas method. A third determination, for excess non-
absorbed NO, is also required to complete the analysis and is obtained
from the secondary absorbent (NaC102 - see Section 4.3.7). The authors
suggest (570) that a significant advantage to the use of solid sorbents
in the primary absorption step is the rapid reaction rate which permits
precise determination of the time of sampling for a reacting (i.e.,
oxidizing) mixture containing NO + 0^.
Another approach is the use of inert solid absorbents
(such as silica gel) as the N02 collector. Silica get appears to be a
good candidate for collection of N02 and separation of NO. Gill (77)
*
These analytical methods are not recommended for flue gas analysis, how-
ever; see previous discussion of analytical methods.
114 WALDEN RESEARCH CORPORATION
-------
using 4 g of silica get for collecting N02 in the presence of NO, reported
efficiencies of 40% in the range 15 to 60 ppm ML. At NO concentrations
as high as 200 ppm, no more than 2 ppm is absorbed. Gill's measurements
were apparently made at room temperature. An obvious way to increase the
absorption efficiency of silica gel for N02 is to reduce the absorbent
temperature.
Due to the limited time and effort, it was decided that
the aqueous absorbents would be more fruitful (e.g., require less develop-
ment) for NOp. Thie approach is discussed in the following section.
5.3.2 Aqueous Absorbents
5.3.2.1 Introduction
The primary and secondary absorption approach to
NCL analysis may also be conducted by use of more conventional aqueous ab-
sorbents. Peters and Strachil (570) recommend 0.1N NaOH for the primary
adsorption and report complete separation of N02 and NO provided that both
gas and solution phases are well stirred and adequate reaction time is em-
ployed. A grab sample may be taken in an evacuated flask, e.g., Figure
4-1 (bottom), containing 0.1N NaOH as the absorbent. After the sample is
shaken for 5 to 10 minutes, the liquid containing the absorption products
arising from N«03 is drawn off. The gas sample is retained for an analy-
sis of "excess" NO. The nitrate component of the liquid sample may be de-
termined from an aliquot by a nitrate method. The nitrite component may
be determined in the caustic solution by the Saltzmann method since sul-
fite will not reduce nitrite in alkaline solutions or by permanganate
titration as recommended by Peters and Strachil (570). The "excess" NO
in the grab sampling flask could be determined spectrophotometrically as
N02 after oxidation in the gas phase (since all the S02 will be absorbed
in the caustic solution) or by addition of an aqueous peroxide absorbent
and subsequent measurement of total nitrate by the PDS or nitrate elec-
trode methods. Thus, the original NO and N02 concentrations may be de-
termined from the results of the two-stage absorption.
WALDEN RESEARCH CORPORATION
-------
5.3.2.2 Experimental
The apparatus used for the two-stage absorption
studies is that of reference (96) without the limiting orifice. Twenty-
five ml of 0.005N NaOH was pipetted into a clean (calibrated) 2-liter
flask. The flask was evacuated and the initial pressure was recorded.
Samples of NO were prepared using the dilution system shown in Figure
5-1 a. The N02 was prepared initially using a lecture bottle of N02 as
the source and the dilution system in Figure 5-la. This system required
a heated flowmeter to eliminate condensation of N02- The NO and N02 flow-
meters were calibrated using a soap film flowmeter. The NO calibration
remained constant over long periods of time, however, the N02 calibration
changed and yielded irreproducible N02 concentrations. Therefore, we
eliminated the tank N02 and used an N02 permeation tube (Figure 5-lb).
The evacuated flask was filled from the sampling line over a two to three
minute period and the final pressure was recorded after periodic shaking
during the specified absorption time. The contents were poured out and
the flask was rinsed with an additional 25 ml of .005N NaOH which was
added to the initial solution.
Analysis for nitrate was performed using the
nitrate electrode and phenol disulfonic acid (PDS) method (described in
the following section).
The concentration of N02 in ppm was calculated
in the volume originally sampled at the sampling temperature (24.5°C)
corrected to standard pressure by the following equation:
m v T Pn V f
N02 (ppm) = Vf T° $
where m = NOl ion concentration, moles/1
V = NOZ sample volume, ml
T = sampling temperature = 24.5°C
PQ = 760 mm Hg
11 6 WALDEN RESEARCH CORPORATION
-------
FLOW
METER
(SS TEE) EXPANSION
VITON 0-PUNG COIL
SEALS
\
EXHAUST
5 LITER FLASK
a. NO DILUTION' SYSTEM
N02
PERMEATIO
TUBE
CONSTANT
'TEMPERATURE
BATH
TO 5 LITER MIXING FLASK
b. N02 DILUTION SYSTEM
Figure 5-1. Dilution System for NO and N02-
117
WALDEN RESEARCH CORPORATION
-------
V = molar volume =22.4 I/mole
f = stoichiometric factor = 0.5 for nitrate electrode, 0.9 for
PDS
Vf = sampling flask volume =2.01 liters
TQ = 273.2°C
AP = final-initial pressure (of sampling vessel), mm Hg
Introducing the constants and fixed terms, this
equation reduces to the formula:
N02 ppm = 9.0 x 108 x -j^ (for the nitrate electrode)
N02 ppm = 4.05 x 108 x ^ (for PDS)
where m is in moles NOZ/l and AP is in mm Hg.
5.3.2.3 Stoichiometry and Analytical Methodology
Initial experiments on N0? absorption were run
using the nitrate electrode for analysis of the nitrate formed by adsorp-
tion of NOp. The low values for the ratio of measured to theoretical NOp
(^ 1:5) (Table 5-1) prompted us to investigate further. For absorption
of NOp in caustic, the stoichiometric factor has been reported as 0.5 cor-
responding to the reaction:
2N02 + 20H" + NO" + N02 + H20
The above stoichiometric factor which we measured as ^ 0.2 (Table 5-1)
assuming no oxidation of nitrite in alkaline solutions. However, Moeller
(487) and Latimer (479) state that comparatively weak oxidants convert
Absorption of 1.0 moles of N02 in alkaline solution yields 0.5 moles
nitrite and 0.5 moles of nitrate. The nitrate electrode does not re-
spond to nitrite, hence a stoichiometric factor of 0.5. The PDS method
does respond to nitrite (80% of nitrate) as well as nitrate. Thus, we
have a factor of 0.5 for nitrate and 80% of 0.5 for nitrite or a
stoichiometric factor of 0.9 for PDS.
118
WALDEN RESEARCH CORPORATION
-------
TABLE 5-1
COMPARISON OF INITIAL (5 MIN.) N02 CONCENTRATIONS BY THE
PDS AND NITRATE ELECTRODE ANALYSIS FOR "DRY" AIR
Run No. NOg Measured (ppm) N02
Nitrate Electrode Analysis
3 11 0.18
4 11 0.17
5 17 0.24
6 14 0.20
11 10 0.16
12 10 0.16
18 12 0.21
19 11 0.20
21 12 0.23
0.19 - Av.
PDS Analysis
25 14 0.30
28 8 0.41
31 19 0.36
0.35 - Av.
nitrite to nitrate in alkaline medium, the former citing 0« as an example.
Mellor, on the other hand, states that oxidation does not occur in neutral
or alkaline solution, as implied by the Peters and Strachil technique (570),
A brief experimental investigation revealed that
p
np_ oxidation occurs even with 100% (tank) oxygen for 10 M NOZ in .005N
NaOH with reaction times as long as 16 hours. The analyses were obtained
by absorption spectroscopy and are estimated as accurate at the ±1% level
(since the measurements are relative not absolute).
Due to the limited sensitivity of the nitrate
electrode techniques, we decided to.investigate the PDS method as a read-
out for nitrate. Initial results using 5-minute absorption times with
119
WALDEN RESEARCH CORPORATION
-------
"dry" air revealed that the "apparent" stoichiometric factor (0.35) in
Table 5-1 was nearly twice the value obtained by the nitrate electrode.
Nitrite is a reported interference (547) with the PDS method but the mag-
nitude was not given. Accordingly, we measured the effect of nitrate and
nitrite mixtures by PDS analysis. The results are given in Table 5-2.
These data reveal that nitrite gives 80% of the response of nitrate. The
stoichiometric factors which were obtained by PDS and nitrate electrode
confirm that PDS responds nearly the same to nitrate and nitrite but the
nitrate electrode responds only to nitrate. Thus, the actual stoichio-
metric factor for N02 absorption in caustic becomes the ratio of nitrate
electrode to PDS or 0.19/0.39* = 0.49 in good agreement with previous
results (69).
TABLE 5-2
EFFECTS OF NITRITE ON THE NITRATE
ANALYSIS BY THE PDS METHOD
Sample No.
1
2
3
4
Absorbance at 400 nm
0.080
0.130
0.173
0.300
Remarks
25 ppm NO^
25 ppm NOZ + 25
50 ppm NOZ
50 ppm NO^ + 50
ppm NOZ
ppm N02
5.3.2.4 Results and Discussion
The initial analytical results for N02 absorp-
tion were performed by nitrate electrode (as discussed in the previous
section). The results for N02 absorption, however, have been run by the
PDS technique. The low recovery of nitrate in the analysis required a
*
Corrected for 80% nitrite response.
120 WALDEN RESEARCH CORPORATION
-------
study of the absorption of NOg as a function of time. Runs were made at
5, 30, and 60 minutes, respectively. The results are given in Table 5-3.
These data showed that the absorption time for N02 in caustic was long
(% 1 hour) since only the values at one hour contact time were close to
the N02 concentration added to the system. The higher (close to
stoichiometric) N02 concentrations obtained in runs 31-33 compared to 25-
27 appeared to be due to the higher relative humidity of the former runs.
We proceeded to investigate the influence of moisture of the absorption
of N02> The striking conclusion which may be drawn from runs 36-44 is
that the absorption time for N0« in aqueous caustic becomes very rapid
in moisture saturated air compared to unsaturated air of relatively high
humidity. The reason for this behavior is not known, but it strongly sug-
gests that the hydrolytic reactions discussed in Section 3 have more than
academic significance. Note that the slopes rather than the absolute
values in Figure 5-2 are significant. There is some scatter in the N02
supply level. Curve (f), for example, appears to reflect a 75 ppm N02
supply. Note curve (a) was run at 20 ppm N02 feed concentration. The
data indicate that the absorption in a moist effluent will be very rapid.
In Section 3.3, we have shown that the rate of
oxidation can be predicted from Equation 12:
'
in ^ I
In J
We have run additional studies on NO using caustic absorption. The re-
sults are given in Table 5-4. Measurements were made at different con-
tact times, 02, and NO levels. The measured N02 concentrations are in
good agreement with the calculated values again verifying the above
equation.
If we assume a flue gas with an NO concentration
of 400 ppm, and an oxygen content of 4%, for 30-second contact time the
N02 concentration would be 12 ppm. Thus, a significant error is intro-
duced unless the N02 level is high *> 1/10 the NO level. Thus, the Peters
and Strachil technique (570) of two-stage absorption appears to work since
1 21 WALDEN RESEARCH CORPORATION
-------
TABLE 5-3
ABSORPTION OF N02 IN CAUSTIC
Run No.a
25
26
27
28
29
30
31
32
33
36
37
38
39
40
41
42
43
44
a. Two runs
b. Nitrate
c. Nominal
was ^0
ppm N02 Absorption Time
Measured^'0 (min)
15
25
29
8
9
13
19
34
47
81
72
72
53
53
53
49
53
53
discarded
analyses by modified
N02 concentration 50
ppm.
5
30
60
5
30
60
5
30
60
5
30
60
5
30
60
5
30
60
PDS
to
(RH
(RH
(RH
Remarks
•x-60%) room air
^71%) room air
^2%) room air
Room air (saturated
water vapor at 23°C)
Room air (saturated
water vapor at 23°C)
Room air (saturated
water vapor at 23°C)
method
60 ppm except
runs 28-30 where
with
with
with
N02
122
WALDEN RESEARCH CORPORATION
-------
ca 50ppm in dirdt ca 23C (except curve a",
Cc3 20 ppm N0)
80
o
i 40
CL
20
0
A
A ~~~_J
A~~~~
/ ^
•^ __
A
— — A
• ^ ""
— — 9" ~~"
t
A
— — ^
^ .1
. ^ "*^"
_ —
>
>- e
; d
x c
i
,- b
>- d
RELATIVE
HUMIDITY %
IOO
100
100
80
60
70
20 40 60
ABSORPTION TIME (MIN)
• ROOM AIR
A MOIST (SATURATED) AIR
Figure 5-2, Effect of Moisture on the Absorption Time of N0? in
Caustic Solution.
123
WALDEN RESEARCH CORPORATION
-------
TABLE 5-4
COMPARISON OF CALCULATED AND MEASURED NO?
CONCENTRATIONS FOR NO + 02 MIXTURES
Run
25
36
35
45
37
43
49
47
a.
b.
N Initial NO
no' Cone, (ppm)
217
217
217
533
533
533
1009
1009
Nitrate analysis by
Calculated from:
(see Section 3.3)
02(X)
6.5
20.0
20.0
6.5
20.0
20.0
6.5
20.0
nitrate
^N?
Contact
Time (min.)
15
30
60
60
60
60
70
90
electrode
1
In
N02 Measured9
(ppm)
118
140
148
282
448
468
752
932
k
-------
the N(L is absorbed rapidly and only a small amount of NO is oxidized.
However, if the N02 level is considerably less than 1/10 the NO concen-
tration, the fraction of NO oxidized to N0« becomes a significant portion
of the NOp measured.
5.3.2.5 Discussion and Recommendations
Additional work on actual flue gas samples is
needed to confirm and evaluate the results obtained. The influence of
moisture on the absorption of N0« should be more thoroughly investigated.
Although the chemistry of the system is complex,
a two-stage caustic absorption method appears feasible for combustion ef-
fluents after further study. The N02 concentration should be at least
1/10 the NO concentration in order to achieve accurate results.
5.3.3 Kinetic Approach for Simultaneous Measurement of NO and
NP_2
The gas phase air oxidation of NO and N02 is represented
by the equation: 2NO + 02 = 2N02, for which the following expression re-
lating the rate of disappearance of NO and the concentrations of 0« and
NO can be written:
^f1 • k[02][NO]2 (1)
where [ ] denote concentrations of the various species and t = time. If
02 is present in large excess over the NO and N02, [02] can be regarded
as a constant which leads to the integrated rate equation:
where [NO] is concentration at zero time. This equation can be rewritten
in terms of NO and NOp concentrations at:
[NO + N02](1-X) 2 [NO + N02](1-XQ)
125 WALDEN RESEARCH CORPORATION
-------
s
where X = mole fraction of NCL. and further simplified to:
] w> + N + (4)
From Equation (4), it can be seen that a plot of the re-
ciprocal of [1- (fractional conversion of NOp at time "t")] versus time
will give a straight line whose slope is proportional to the total [NO
+ ML] and whose intercept gives the fraction of N02 present in the
sample at the start.
It appears technically feasible to apply the kinetic ap-
proach to the measurement of NO and N02 in flue gases. N02 would be ab-
sorbed in a suitable collecting medium and the increase of nitrate or
nitrite concentration would be monitored at known time intervals after
sampling. There are many possible absorbing media including water, buf-
fered aqueous systems, oxidative solutions (such as the presently used
H^/HpSO^ absorbers for PDS) and non-aqueous solvents. The kinetic ap-
proach was quite recently applied to the measurement of NO and N0? in
cigarette smoke, where the N02 produced was measured by absorption Into
Saltzmann's reagent (438).
It is clear, however, that this technique would require
fairly elaborate analytical facilities at the sampling site.
5.4 Conclusions
The use of solid sorbents or aqueous absorbents in the two-stage
sequence discussed is a fairly simple approach to the determination of N02
and NO in a flue gas. The kinetic method, on the other hand, is relatively
complex and not readily adaptable for field measurement. Since both the
kinetic method and aqueous absorbents are best approached by grab sampling,
the sensitivity for N02 is critical. If we estimate the N02 concentration
as 2% of NO , N09 concentrations in power plants effluents may be expected
*
to be in the range 5 to 20 ppm. These concentrations are at the lower
These values are not even as low as those suggested by the thermodynamic
and kinetic calculations of Section 3.
126 WALDEN RESEARCH CORPORATION
-------
limit of sensitivity for both PDS or nitrate electrode analysis of grab
samples.
These concentrations could, on the other hand, be readily deter-
mined utilizing solid sorbents which permit longer integration times and
larger sample volumes to be taken.
127 WALDEN RESEARCH CORPORATION
-------
6. STATISTICS OF NOX DETERMINATION
6.1 Introduction
There are major differences between the collection and analyti-
cal methods used for the oxides of sulfur and those used for the oxides
of nitrogen. First, sulfur oxides are collected by techniques which con-
centrate the sample; nitrogen oxides are collected by grab sampling tech-
niques. Second, many different techniques have been employed for collec-
tion and analysis of sulfur oxides in power plants (Volume I); virtually
all the published and unpublished power plant analyses for NO have been
A
performed by the PDS method.
Since PDS or similar techniques (grab sampling) are almost uni-
versal for NO , we have examined the statistics for this method in detail.
A
6.2 Power Plant Results
Grab sampling methods such as PDS require replicate samples
(duplicate, triplicate or quadruplicate) and averaging to obtain the
total NO concentration for a single firing condition with reasonable ac-
A
curacy. We have calculated the precision of the PDS method in field
sampling by computing the pooled standard deviation for a series of runs
at a given power plant. For example, in Table 6-1, Test Series #1 con-
tains nine runs in duplicate at the same power plant under different
firing conditions. The results for a series of coal-fired units are sum-
marized in Table 6-1. As expected, the precision for these analyses is
considerably poorer than that reported for the determination of inorganic
nitrate (analysis alone) (447). In this instance, we contend not only
with the errors due to sampling and source variations, but also with the
problem of the reactivity of the nitrogen oxides.
Inspection of the data in Table 6-1 reveals a correlation between
the value of the mean NO concentration and the coefficient of variation,
A
in the expected sense; i.e., the precision tneds to improve at higher con-
centrations. For comparison, we may calculate the pooled coefficient of
variation for these results from the pooled variance:
128 WALDEN RESEARCH CORPORATION
-------
o
8
TO
s
TO
TABLE 6-1
STATISTICS OF THE PDS METHOD FOR N0y IN COAL-FIRED POWER PLANTS
A
ro
vo
Identification courcp Pairs of Duplicate Pairs Mean NOX
Number source Samples Used Discarded1" Cone, ppm
1
2
3
4
5
6
*cv - std-
548
548
442
548
548
548
dev. x 100
mean
Values were discarded
9
14
8
10
4
15
only where the departure
0
1
1
0
1
3
from the mean
236
387
418
424
511
964
exceeded
Range Pooled Standard cu* «
(ppm) Deviation ppm
150-400
300-460
220-520
300-560
400-630
700-1400
§p, including
32.8
21.0
17.0
19.0
41.6
35.0
the rejected
13.9
5.6
4.1
4.7
8.1
3.6
values.
-------
2 Z (n-1) S*
(E n.) - i
where S. is the sample standard deviation and n. is the statistical sample
size.
For the weighted mean NO concentration, 527 ppm, the pooled coefficient
A
of variation is calculated to be 5.3%.
An additional datum has been provided by D. Barnhart (425) who
summarized PDS precision in coal-fired power plant sampling as ±25 ppm
at 250 ppm NO .
X
Since these data on the precision of the PDS method in coal-fired
power plants shows some degradation at NO concentrations less than 400 ppm,
A
we have investigated the possible importance of fuel type by searching for
additional NO field data from various sources.
/\
Measurements of NO concentrations at a number of gas-fired plants
A
were obtained from Combustion Engineering (442), Foster Wheeler (560),
Riley Stoker (561), and Bary Area APCD (562). Another series of measure-
ments at a gas-fired power plant was obtained from NAPCA (548). The re-
sults are summarized in Table 6-2. The precision of these results is in
the same range as those of the coal-fired power plants; the trend toward
decreased precision at low NO concentrations is apparent in both data
A
sets.
There are several differences, however, between the data in
Tables 6-1 and 6-2. The measurements at the gas-fired power plants were
obtained from five different laboratories but the coal-fired power plant
data are essentially from one laboratory (NAPCA). Thus, the gas-fired
data may be expected to have a greater variance due to differences in
techniques between laboratories. The important thing to note is that the
trend in the data for an individual laboratory is consistent (i.e., pre-
cision decreases with concentration).
130 WALDEN RESEARCH CORPORATION
-------
TABLE 6-2
STATISTICS OF THE PDS METHOD FOR NO IN GAS-FIRED POWER PLANTS
m
v>
8
TO
3
TO
o
Identification
Number
Data
D =
T =
Q-
1
2
3
4
5
6
7
8
9
10
not incl
duplicate
Qr.nv.r-0 Number of Number
source Samples Used Discarded
562
561*
442
548
442
562
560*
561*
560*
442
uded in Figure
samples
7 D
12 D
7 D
12 D
6 T
4 0
6 0 1
4 Q
8 T 1
8 T
6-1
Mean NOX
Cone, ppm
99
114
152
286
294
331
433
443
545
589
Range Pooled Standard cv *
(ppm) Deviation (ppm)
10-150
70-150
90-220
260-315
230-250
210-450
300-600
320-500
310-680
320-720
12.9
9.0
13.8
12.1
14.9
25.0
38.1
29.8
38.1
27.8
13.1
7.9
9.1
4.2
5.1
7.6
8.8
6.8
7.0
4.8
triplicate samples
quadruplicate samples
-------
Some additional results for oil-fired power plants (563,561) are
given in Table 6-3.
To determine the variation of the precision with NO concentra-
A
tion, we have combined the data of Tables 6-1 to 6-3 and conducted loga-
rithmic regression analysis of the pooled data (weighted for sample size)
to obtain best estimates of the coefficient of variation as a function of
NO concentration (Figure 6-1). The pooled estimate of the coefficient of
A
variation is 3% at 1000 ppm and 10% at 100 ppm NO .
A
6.3 PDS Automotive or Diesel Exhaust Results
Data on the analysis of NO in auto exhaust (26 tests) were ob-
A
tained from B. Dimitriades of the Bureau of Mines (564). Grab samples
were taken in quadruplicate from a mylar bag in which the exhaust was col-
lected (7). This method should eliminate source variation from calculated
precision values. We have divided the data into three concentration ranges
and have calculated the precision statistics given in Table 6-4. This tech-
nique yields precision values comparable to the precision of the analytical
method alone, i.e., sampling errors are essentially eliminated. Additional
data were obtained from H. Lang at the Bureau of Mines (20 sets) for NO in
A
diesel exhausts by the PDS method (565,566). These samples were taken "on
stream" and, therefore, include sampling errors. The coefficient of varia-
tion for these samples is, as expected, somewhat higher than Dimitriades'
results (2.5 compared to 1.8%).
The precision for 21 individual diesel exhaust samples obtained
under standard test conditions by David and O'Neil (9) over an extended
period of time is C.V. = 6.9% at mean NO = 636 ppm. The influence of
A
source variation on the precision of NO sampling is readily seen by com-
A
parison to the precision obtained by Lang for duplicate samples.
6.4 Comparison of the Precision of PDS to the Precision of Other
Methods for NO Determination
A
The PDS method has been used almost exclusively in power plants.
The British Standards Institution recommends the 2,4-xylenol method which
132 WALDEN RESEARCH CORPORATION
-------
TABLE 6-3
STATISTICS OF THE PDS METHOD FOR NOV IN OIL-FIRED POWER PLANTS
y\
CO
CO
Identification crtllv.-,Q Number of
Number source Samples Used
*
Data
D =
Q =
1
2
3
4
5
not included
561*
561*
561*
563
563
in Figure
4
10
4
18
11
6-1
D
D
Q
D
D
Number Mean NOX
Discarded Cone, ppm
167
221
320
2 445
6 672
Range Pooled Standard cv «
(ppm) Deviation (ppm)
110-240
160-265
230-390
340-520
560-690
22.
8.
51.
30.
34.
7
5
5
2
2
13.
3.
16.
6.
5.
6
9
0
8
1
duplicate samples
quadruplicate
samples
I
CO
o
8
TO
s
TO
-------
Figure 6-1
Precision of NOV Determination by PDS Method in Fossil Fuel Fired Power Plants
.X _
(100%) 2.0-1-
CO
-pa
I
o
m
JO
m
•33
O
o
o
•33
-o
O
33
d)
o
o
o
en
o
(10%)
l.O
0
log y = 2.0 - 0.51 log X
coal
oil
gas
O coal estimate
by B&W
10% (100 ppm)
O
(1000 ppm)
1.0
(10)
2.0
(100)
3.0
(1000)
log N0x (ppm)
, ppm)
-------
TABLE 6-4
PDS STATISTICS FOR AUTOMOTIVE EXHAUSTS
No. of Samples
13*
8*
5*
10f
10f
Mean NOx
Cone, (ppm)
169
449
633
667
1003
Range
37-237
347-526
633-690
445-855
890-1200
Standard Deviation
(ppm)
4.8
4.6
11.4
16.8
13.0
CV %
2.9
1.03
1.8
2.5
1.3
Run in quadruplicate (564)
fRun in duplicate (566)
also utilizes evacuated flask sampling and colorimetric readout. It is
not surprising, therefore, that the precision of the determination, re-
ported to. be ±5% (379), is similar to that found for the PDS method.
In a series of tests on diesel exhaust gases, the PDS and modi-
fied Saltzmann methods were run in duplicate and compared to mass spectro-
metric results (565). The results of the individual duplicate wet chemical
samples were obtained from H. Lang (566). The precision of these results
is given in Table 6-5, series #1 and 2. The PDS method gives slightly
better precision than the modified Saltzmann method. Results from com-
parison of PDS and modified Saltzmann methods, also for diesel exhausts,
taken from (9) are given in Table 6-5, series #3-10.
In order to compare these two sets of data, we calculate the
pooled coefficient of variation for all the data in Table 6-5. The re-
sults are given below:
35 WALDEN RESEARCH CORPORATION
-------
o
m
TO
O
I
8
y>
3
TABLE 6-5
COMPARISON OF THE PRECISION OF THE PDS AND SALTZMANN
METHODS FOR NO IN DIESEL EXHAUSTS
01
Series No.
1
2
3
4
5
6
7
8
9
10
Source
566
566
9
9
9
9
9
9
9
9
Number of
Tests
20f
21
21
5
5
4
4
2
2
Mean NOx
Cone, ppm
835
846
636
678
937
915
1096
1118
1811
1817
Pooled Standard
Deviation ppm
15.0
29.1
44.0
27.0
29.0
20.0
18.0
15.0
42.0
18.0
CV %
1.8
3.5
6.9
2.5
3.1
2.2
1.7
1.3
2.3
1.0
Method
PDS
Saltzmann
PDS
Saltzmann
PDS
Saltzmann
PDS
Saltzmann
PDS
Saltzmann
t
Duplicate samples
O
-------
Mean NOX Cone. Pooled Coefficient
Method (ppm) of Variation (%)
PDS 825 3.4
Modified Saltzmann 842 3.2
Thus, we find that the precision of the Saltzmann is essentially the same
as for the PDS method. As above, both methods which again use similar
techniques (evacuated flasks) and colorimetric readout yield the same pre-
cision for diesel exhaust.
Lang, et al. (565) compared both the PDS and the modified Saltz-
mann methods to nitric oxide determination by the mass spectrometer. If
we assume that the mass spectrometer results are accurate, we can compute
the accuracy of these wet chemical methods using the simple linear regres-
sion method.
Taking the mass spectrometer as the dependent variable and the
PDS as the independent variable, the regression equation for the 20 set
population is:
Yms = -7.4 + 1.0 XpDS
mean (S~) = 840 ppm
(SE) = 45 ppm
(CV) = 5.4%
(CD) = 0.961"
For the Saltzmann method as the independent variable and the mass spec-
trometer as the dependent variable, the regression equation for the 20
This should be a valid assumption since the major oxide component is
nitric oxide, and the sample is introduced directly into the inlet sys-
tem where the pressure is immediately reduced, minimizing the possibil-
ity of oxidation.
CD = coefficient of determination - the square of the correlation coef-
ficient which may be interpreted so the fraction of the total sample
variance accounted for by the regression equation.
1 37 WALDEN RESEARCH CORPORATION
-------
set population is:
Yms = -43.4 + 1.0 X$A
mean (S) = 840 ppm
(SE) = 48 ppm
(CV) = 5.7*
(CD) = 0.96
The unit slope, small intercept, and high coefficient of determination for
both wet chemical methods indicates that interferences in NO determinatior
A
in diesel exhaust are minimal at these concentrations.
6.5 Conclusions
The coefficient of variation for the NO measurements in power
A
plants by the PDS method is dependent on the concentration (CV = 10% at
100; 3% at 1000 ppm NOJ.
A
No apparent differences have been found for variations in fuels
but we intend to examine this effect in more detail by obtaining data on
the three major fuels run by the same laboratory.
Accuracy of the order of 5 to 6% is found for diesel exhausts,
however, the accuracy may be poorer for power plants where more severe
sampling problems are encountered.
The precision of the PDS method for NO (in diesel exhausts) is
A
about the same as that for the similar Saltzmann and 2,4-xylenol methods.
1 38 WALDEN RESEARCH CORPORATION
-------
7. RECOMMENDATIONS
7.1 NO Determinations
A " ~"~"~* '~
The phenol disulfonic acid method has been found to be a reli-
able and the most generally applicable method for total oxides of nitro-
gen in combustion effluents. The rate controlling step for the sampling
procedure is the gas phase oxidation of NO to N0«. The colorimetric
analysis is also slow as a result of the long (2 to 6 hour) evaporation
time required.
The PDS procedure may be substantially improved by (successful)
development of:
(1) Addition of oxygen or ozone to the sampling vessel
to reduce absorption time.
(2) Use of rotary (or other) fast evaporation equipment
and smaller aliquots for the present colorimetric
determination of nitrate.
(3) Direct determination of nitrate in the absorber solu-
tion with the specific ion electrode.
(4) Nitration in 50% sulfuric acid (instead of evapora-
tion to dryness). See Section 4.3.1.
An alternative approach of great promise is the use of reactive
solid sorbents. If successfully developed, solid sorbents may yield con-
siderably higher precision and accuracy than presently attainable, greater
east of use, and possibly permit simultaneous SO and NO determination.
A A
The nitrate electrode yields a rapid analysis for nitrate and gave good
agreement with PDS for the limited samples taken. This technique should
be validated as a replacement for the time-consuming PDS technique.
7.2 NO/NOo Determinations
The most promising approach to NO/N02 analysis is development of
new sampling techniques to be used with existing nitrite and nitrate methods.
The sampling methods include solid sorbents and/or aqueous absorbents used
in two-stage sequence. The two-stage caustic absorption technique appears
1 39 WALDEN RESEARCH CORPORATION
-------
to be feasible for N0«/N0 ratios of 1:10. Kinetic analysis may be a
£+ A
feasible alternative if simple analytical determinations are developed.
7.3 Simplified Methods
The modified Saltzmann method or hydrogen peroxide collection
followed by NaOH titration are simplified methods which may be used for
total NOX determinations in a gas-fired power plant, i.e., in an effluent
free of SOp. Ferrous sulfate and other complexes may be considered as
candidate simplified methods for total NO determinations. Solid sor-
/\
bents because of their ease of use may also be considered as simplified
methods. The nitrate electrode promises to be a simplified method with
nearly the same precision as the PDS method.
1 40 WALDEN RESEARCH CORPORATION
-------
8. CONCLUSIONS
Thermochemical and kinetic data reveal that N.,0,,, N00C, and NO, do not
C. "t C. 0 O
exist in significant concentrations in flue gases. At equilibrium, sig-
nificant concentrations of NpO and HNO^g) may exist in a flue gas at low
temperatures. NO and NOp are the major oxides emitted in combustion proc-
esses. Analysis of the (homogeneous) kinetics shows that for typical flue
gas conditions and 600 ppm NO, the N02 concentration is only 1 ppm (as-
suming no SOp in the flue gas) for typical residence times. There appears
to be no experimental basis for the high N09/N0 ratios (10%) commonly re-
£ A
ported for flue gases. No reliable measurements of NOp in power plant ef-
fluents have been located.
All sampling procedures for NO determination in stack gases require
/\
grab sampling techniques because of the low solubility of NO in aqueous
solutions. Integrated grab sampling should provide improved results in
emissions determinations.
A large number of analytical methods for both nitrate and nitrite have
been evaluated. Many of these methods have drawbacks such as limited range,
interferences, many procedural steps, etc. The methods which determine
nitrite present a problem because SOp, which is present in high concentra-
tions in oil- and coal-fired plants, will reduce nitrite in acid solutions.
The PDS method (for nitrate) is the only one which has been widely used for
the determination of NO in power plants. The major drawback to this tech-
A
nique is the long time required for analysis. Two steps in this procedure
are time consuming:
(1) oxidation of NO to N02
(2) evaporation of absorbing solution
The time required to carry out the oxidation may be reduced by an order of
magnitude or more by Op or 0~ enrichment. The evaporation time may be re-
duced by use of a rotary evaporator in place of a steam bath or by nitra-
tion in 50% H2S04 (620).
WALDEN RESEARCH CORPORATION
-------
The nitrate electrode has been tested for NO determination. Prelim-
/\
inary results at oil-fired power plants show good agreement (within ±5%)
with PDS. This method has the advantage of rapid readout of the nitrate
concentrations.
Solid sorbents are promising candidates for development of methods
for total NO as well as N0/N02» which are both convenient and of high
precision. Two-stage absorption in aqueous NaOH appears to be a feasible
method for the determination of N0« in combustion effluents provided that
the N09/NOV ratio is about 1:10.
£ A
The coefficient of variation for NO determination in power plants by
A
the PDS method is dependent on the concentration (CV = 10% at 100; 3% at
1000 ppm NO ). No apparent differences in precision have been found as a
X
function of fuel, although we are seeking analytical data for the three •
major fuels conducted by the same laboratory.
Accuracy of the order of 5 to 6% is found for diesel exhausts; how-
ever, the accuracy may be poorer for power plants where more severe
sampling problems are encountered.
The precision of the PDS method for NO (in diesel exhausts) is about
"
the same as that for the similar Saltzmann and 2,4-xylenol methods.
142 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED*
1. J. R. Taylor, W. 0. Holland, R. D. MacPhee and K. H. Schoenemann,
Editors, "Laboratory Methods," Air Pollution Control District, County
of Los Angeles (1958).
5. 6. Norwitz, "A Colorimetric Method for the Determination of Oxides of
Nitrogen," Analyst 91_, 553-8 (September 1966).
6. T. Singh, R. F. Sawyer, E. S. Starkman and L. S. Caretto, "Rapid Con-
tinuous Determination of Nitric Oxide Concentration in Exhaust Gases,"
JAPCA ]i (2), 102-5 (1968).
7. B. Dimitriades, "Determination of Nitrogen Oxides in Auto Exhaust,"
JAPCA 17. (4), 238-43 (1967).
9. R. F. Davis and W. E. O'Neill, "Determination of Oxides of Nitrogen in
Diesel Exhaust Gas by a Modified Saltzmann Method," U.S. Dept. of the
Interior, Bureau of Mines Report of Investigations 6790, 5 pp. (1966).
10. D. L. Ripley, J. M. Clingenpeel.and R. W. Hurn, "Continuous Determina-
tion of Nitrogen Oxides in Air and Exhaust Gases," International J. of
Air and Water Pollution 8, 455-63 (1964).
69. A. P. Altshuller and A. F. Wartburg, "Ultraviolet Determination of
Nitrogen Dioxide as Nitrite Ion," Analytical Chemistry 32_ (2), 174-7
(I960).
76. J. F. Smith, J. A. Hultz and A. A. Orning, "Sampling and Analysis of
Flue Gas for Oxides of Sulfur and Nitrogen," U.S. Dept. of the In-
terior, Bureau of Mines Report of Investigations 7108, 21 pp. (April
1968).
77. W. E. Gill, "Determination of N02 and NO in Air," AIHA J. 21_, 87-96
(February 1960).
85. B. E. Saltzman, "Modified Nitrogen Dioxide Reagent for Recording Air
Analyzers," Analytical Chemistry 32_ (1), 135-6 (January 1960).
89. N. A. Lyshkow, "A Rapid and Sensitive Colorimetric Reagent for Nitrogen
Dioxide in Air," JAPCA 1_5 (10), 481-4 (1965).
91. D. Wilson and S. L. Kopczynski, "Laboratory Experiences in Analysis of
Nitric Oxide with ''Dichromate" Paper," JAPCA 18. (3), 160-1 (1968).
93. J. L. Mills, K. D. Luedtke, et al., "Emissions of Oxides of Nitrogen
from Stationary Sources in Los Angeles County," Los Angeles County,
Air Pollution Control District Report No. 4, 34 pp. (July 1961).
These references were selected from a larger bibliography and are not
necessarily continuous. i
] 43 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
96. S. Hochheiser, "Methods of Measuring and Monitoring Atmospheric Sul-
fur Dioxide," U.S. Dept. of Health, Education, and Welfare, Public
Health Service Publication No. 999-AP-6 (August 1964).
97. M. Buck and H. Stratmann, "The Joint and Separate Determination of
Nitrogen Monoxide and Nitrogen Dioxide in the Atmosphere," Staub
(English Translation) 27 (6), 11-5 (1967).
100. S. Yanagisawa, N. Yamate, S. Mitsuzawa and M. Mori, "Continuous De-
termination of Nitric Oxide and Nitrogen Dioxide in the Atmosphere,"
Bulletin of the Chemical Society of Japan 39_ (10), 2173-8 (1966).
105. J. T. Shaw, "The Measurement of Nitrogen Dioxide in the Air,"
Atmospheric Environment 1, 81-5 (1967;.
109. M. Fugas, "Determination of Nitrogen Dioxide in Air," Arhib Za
Higijenu Radi i Toksikologiju 1_3 (3), 207-29 (1962).
133. B. E. Saltzman, "Colorimetric Microdetermination of Nitrogen Dioxide
in the Atmosphere," Analytical Chemistry 26_, 1949-55 (1954).
134. H. Devorkin, R. L. Chass, A. P. Fudurich, C. V. Kanter, "Air Pollution
Source Testing Manual," Los Angeles County, Air Pollution Control Dis-
trict, 179 pp. (1965).
137. H. Stratmann, "Microanalytical Methods for Determining Sulfur Dioxide
in the Atmosphere," Mikrochimica Acta, pp. 668-78 (1954).
142. M. B. Jacobs and S. Hochheiser, "Continuous Sampling and Ultramicro-
determination of Nitrogen Dioxide," Analytical Chemistry 3J) (3),
426-8 (1958).
144. B. E. Saltzman and A. F. Wartburg, "Precision Flow Dilution System for
Standard Low Concentrations of Nitrogen Dioxide," Analytical Chemistry
37 (10), 1261-4 (1965).
151. M. D. Thomas, J. A. Macleod, R. C. Robbins, R. C. Goettelman and R. W.
Eldridge, "Automatic Apparatus for Determination of Nitric Oxide and
Nitrogen Dioxide in the Atmosphere," Analytical Chemistry 28 (12),
1810-5 (1956).
172. I. Popa and T. Stan, "p-Hydroxyazobenzenesulfonamide as a Standard for
the Determination of Nitrogen Oxides and Nitrates," Farmacia (Bucharest)
1_0, 745-50 (1962).
173. "Atmospheric Emissions from Nitric Acid Manufacturing Processes,"
U.S. Dept. of Health, Education, and Welfare, Public Health Service
Publication No. 999-AP-27 (1966).
144 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
176. F. A. Patty and G. M. Petty, "Nitrite Field Method for the Determina-
tion of Oxides of Nitrogen," Journal of Industrial Hygiene/Toxicology
25, 361-5 (1943).
177. M. B. Shinn, "Colorimetric Method for Determination of Nitrite,"
Industrial Engineering Chemistry, Analytical Edition 13, 33-5
(15 January 1941).
187. J. P. Lodge, Jr., "Determination of Nitrogen Oxides," International
Journal of Air and Water Pollution 7^, 84-5 (1963).
189. H. Yagoda and F. H. Goldman, "Analysis of Atmospheric Contaminants
Containing Nitrate Groupings," Journal of Industrial Hygiene/Toxic-
ology £5 (10), 440-4 (December 1943).
191. F. Nietruch and K. E. Prescher, "Photometric Determination of Small
Amounts of Nitrogen Dioxide," Zeitschrift fur Analytische Chemie 234
(2), 118-9 (1968).
197. J. L. Mills, K. D. Luedtke, P. Woolrich and L. Perry, "Emissions of
Oxides of Nitrogen from Stationary Sources in Los Angeles County,"
Report No, 2: Oxides of Nitrogen Emitted by Small Sources, Los
Angeles County, Air Pollution Control District, 73 pp. (September
1960).
204. Anon., "Manual on Disposal of Refinery Wastes," Vol. V: Sampling and
Analysis of Waste Gases and Particulate Matter, American Petroleum
Institute, Division of Refining, 1271 Avenue of the Americas, New
York (1957).
205. E. R. Hendrickson, "Air Pollution Sampling and Analysis with Special
Reference to Sulphate Pulping Operations," TAPPI 42 Supp. (5), 173A-
176A (1959).
209. R. L. Beatty, L. B. Berger and H. H. Schrenk, "Phenol Disulfonic Acid
Method for Determining Oxides of Nitrogen," U.S. Dept. of the Interior,
Bureau of Mines Information Circular 7728, pp. 41-7 (January 1956).
241. H. A. Liebhafsky and E. H. Winslow, "Spectrophotometric Determination
of Nitrite," Industrial and Engineering Chemistry, Analytical Edition
Jl, 189-90 (15 April 1939).
251. 0. Tada, "Determination of Nitrogen Oxides in the Air," J. of Science
and Labour 60, 7-26 (1962).
255. J. L. Lambert and F. Zitomer, "Differential Colorimetric Determination
of Nitrite and Nitrate Ions," Analytical Chemistry 32_, 1684-6 (1960).
145 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
260. R. S. Brief and P. A. Drinker, "Collection of Integrated Samples of
Gaseous Effluents," AMA Archives of Industrial Health 17 (6), 654-8
(1958). ~~
267. E. Szekely, "Spectrophotometric Determination of Nitrate with p-
Di ami nodi phenylsulphone and Di phenylami ne-p-Di ami nodi phenylsulphone,"
Talanta 14. (8), 941-5 (1967).
268. E. E. Garcia, "Determination of Nitrite Ion Using the Reaction with
p-Nitroaniline and Azulene," Analytical Chemistry 39_ (13), 1605-7
(November 1967).
276. J. W. Robinson and C. J. Hsu, "Spectroscopic Studies of the Chromo-
tropic acid-Nitrate Reaction," Analytica Chimica Acta 44 (1), 51-8
(1969). ~~~
277. E. Szekely, "Colorimetric Determination of Nitrites with p-
Diami nodiphenylsulphone-Diphenylamine as Reagent," Talanta 15 (8),
795-801 (1968). ~~
278. M. J. Taras, "Phenol Disulfonic Acid Method of Determining Nitrate
in Water," Analytical Chemistry 22. (8), 1020-2 (1950).
279. A. C. Holler and R. V. Huch, "Colorimetric Determination of Nitrates
and Nitric Acid Esters," Analytical Chemistry 2J[ (11) 1385-9 (1949).
317. S. S. Potterton and W. D. Shults, "An Evaluation of the Performance
of the Nitrate-Selective Electrode," Analytical Letters 1 (2), 11-22
(1967).
323. H. Barnes, "Modified 2,4-Xylenol Method for Nitrate Estimations,"
Analyst 75, 388-91 (1950).
326. E. Sawicki, T. W. Stanley, J. Pfaff and A. D'Amico, "Comparison of
Fifty-Two Spectrophotometric Methods for the Determination of Nitrite,"
Talanta 1_0, 641-55 (1963).
327. P. W. West and G. L. Lyles, "A New Method for the Determination of
Nitrates," Analytica Chimica Acta 23 (3), 227-32 (September 1960).
329. F. L. Fisher, E. R. Ibert and H. F. Beckman, "Inorganic Nitrate,
Nitrite, or Nitrate-Nitrite," Analytical Chemistry 30_ 02), 1972-4
(December 1958).
331. A. D. Westland and R. R. Langford, "Determination of Nitrate in Fresh
Water," Analytical Chemistry 28 (12), 1996-8 (December 1956).
WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
332. A. B. Iseyeva and A. N. Bogoyavlenskiy, "The Determination of Nitrate
in Seawater by Reduction to Nitrite with Amalgamated Cadmium,"
Okeanologiya 8 (3), 433-8 (1968).
347. P. F. Corbett, "The $03 Content of the Combustion Gases from an Oil-
Fired Watertube Boiler," J. of the Institute Fuel 26, 92-106 (August
1953).
362. E. Sawicki, J. Pfaff and T. W. Stanley, "Review of Methods of Analysis
for Nitrite and Nitrite Precursors Particularly for Spectrophotometric
Techniques," Revista de la Universidad Industrial de Santander 5, 337-
46 (1963).
375. T. Nash, "Chemical Status of Nitrogen Dioxide at Low Aerial Concentra-
tion," Annales of Occupational Hygiene Y\_ (3), 235-9 (July 1968).
379. Anon., "Methods for the Sampling and Analysis of Flue Gases," Part 4:
Miscellaneous Analyses, British Standards Institution B.S. 1756:
Part 4: 1965, 68 pp.
387. Anon., "Nitrogen Dioxide in Air (Saltzmann Method)," Bay Area Air Pol-
lution Control District, San Francisco, Method N-l (15 November 1961).
388. Anon., "Sampling and Analytical Methods for Total Nitrogen Oxides,"
Bay Area Air Pollution Control District, San Francisco, Method N-3
(14 November 1961).
390. Anon., "Guide to Specific Ion Electrodes and Instrumentation," Orion
Research, Inc., Cat. 961, 23 pp. (1969).
391. H. E. Mishmash and C. E. Meloan, "The Reactions of Nitric Oxide,
Nitrous Oxide, and Nitrogen Dioxide with Lead Dioxide," Microchemical
Journal 14., 181-9 (1969).
397. A. M. Hartley and R. I. Asai, "Spectrophotometric Determination of
Nitrate with 2,6-Xylenol Reagent," Analytical Chemistry 35_ (9), 1207-
13 (August 1963).
404. Anon., "Oxides of Nitrogen in Gaseous Combustion Products (Phenol
Disulfonic Acid Procedure)," American Society for Testing and Mate-
rials, ASTM Designation: D1608-60, 81-5, 6 pp. (1960).
417. J. L. Paul and R. M. Carlson, "Nitrate Determination in Plant Extracts
by the Nitrate Electrode," J. of Agriculture and Food Chemistry 1_6 (5),
766-8 (Sept.-Oct. 1968).
147 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
420. E. D. Robinton, Editor, "Intersociety Committee Methods for Ambient
Air Sampling and Analysis," Health Laboratory Science £ (2), 129 pp.
(Special Issue) (April 1969).
425. D. Earnhardt, Babcock & Wilcox, Alliance, Ohio, Private Communication
(1969).
428. R. L. Beatty, L. B. Berger and H. H. Schrenk, "Determination of the
Oxides of Nitrogen by the Phenol Disulfonic Acid Method," Bureau of
Mines, Report of Investigations RI 3687 (February 1943).
431. S. T. Cuffe and R. W. Gerstle, "Emissions from Coal Fired Power
Plants," PHS Pub. #999-AP-35 (1967).
433. E. M. Chamot, D. S. Pratt and H. W. Redfield, "A Study on the Phenol
Disulfonic Acid Method for the Determination of Nitrates in Water --
The Chief Sources of Error," J. of the American Chemical Society 33,
366-81 (1911).
435. A. 0ien and A. R. Selmer-Olsen, "Nitrate Determination in Soil Ex-
tracts with the Nitrate Electrode," Analyst 94, 888-94 (1969).
436. R. H. Groth and D. S. Calabro, "Evaluation of Saltzmann and Phenol
Disulfonic Acid Methods for Determining NOX in Engine Exhaust Gases,"
JAPCA 19, 884-7 (1969).
438. C. H. Sloan and J. E. Kiefer, "Determination of NO and N02 in Cigarette
Smoke from Kinetic Data," Tobacco 169 (26), 85-7 (1969).
439. S. E. Manahan, "Nitrate Ion-Selective Electrode in Microbial Media,"
Applied Microbiology 1J3 (3), 479-81 (September 1969).
440. M. K. Mahendrappa, "Determination of Nitrate Nitrogen in Soil Extracts
Using a Specific Ion Activity Electrode, Soil Science 108 (2), 132-6
(1969).
441. P. Voogt, "Nitrate Determination in Spinach with a Nitrate-Selective
Electrode, Deutsche Lebensmittel-Rundschau 65_ (7), 196-8 (1969).
442. M. Stengel, Combustion Engineering, Windsor, Connecticut, Private
Communication (1969).
443. Bodenstein & Linder, Z. Phy. Chem. 100. 82 (1922).
444. W. Bartok, et al., "Systems Study of Nitrogen Oxide Control Methods
for Stationary Sources," Esso Research & Engr. Co., Washington, D.C.
(May 1969).
148 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
445. Jellinek, Z. Anorg. Chem. 49, 229 (1906).
446. Bodenstein, Z. P-ys. Chem. TOO, 68 (1922).
447. Breiner, Pfeiffer and Malet, J. Chem. Phys. 21_, 25 (1924).
448. J. Zeldovich, Acta Physiochemica USSR 21_ (4), 577 (1946).
449. B. Deilin, et al., "The Recovery of Nitric and Sulphuric Acids from
Flue Gases," presented at the Symposium on Air Pollution Control
Technology 65th National Meeting, American Inst. of Chem. Engineers,
Dleveland, Ohio (May 1969).
450. H. Johnson, Anal. Chem. 24, 572 (1952).
451. E. R. Riegal, Industrial Chemistry, Reinhold Publishing Co., N.Y.,
p. 12 (1949).
452. D. Yost and M. Russel, Systematic Inorganic Chemistry, Prentice Hall
Inc., N.Y. (1944).
453. A. G. Francis and A. T. Parsons, "The Determination of Oxides of
Nitrogen in Small Concentrations in the Products of Combustion of
Coal Gas and in Air," Analyst 50, 262 (1925).
454. Emelius and Anderson, Modern Aspects of Inorganic Chemistry.
455. E. L. Kothny and P. K. Mueller, "A Method for Oxides of Nitrogen above
100 ppm," 7th Conference on Methods in Air Pollution Studies, Calif.
State Dept. Public Health (Jan. 1965).
456. W. A. Glasson and C. S. Tuesday, "The Atmospheric Thermal Oxidation
of Nitric Oxide in the Presence of Dienes," General Motors Res. Pub.
GMR-454 (1965).
457. R. Saltzmann, Dupont Instruments Div., Newark, Delaware, Private
Communication (1969).
458. S. W. Nicksic and J. Harkins, "Spectrophotometric Determination of
Nitric Oxide in Auto Exhaust," Anal. Chem. 34, 985 (1962).
464. D. H. Earnhardt and E. K. Diehl, "Control of Nitrogen Oxides in Boiler
Flue Gases by Two-Stage Combustion," JAPCA 1£, 397 (1960).
479. W. Latimer, Oxidation Reduction Potentials, 2nd Ed., Prentice Hall,
N.Y. (1952).
149 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
480. J. R. Ehrenfeld, et al.. "Systematic Study of Air Pollution from
Fossil Fuel-Fired Combustion Equipment," Final Report, Contract No.
CPA 22-69-85, Maiden Research Corp. (1970).
487. T. Hoeller, Inorganic Chemistry, John Wiley and Son, Inc., N.Y.
(1953).
488. D. R. Stall, JANAF Thermochemical Tables, Dow Chemical Co. (1965-68).
507. M. B. Jacobs, The Analytical Toxicology of Industrial Inorganic
Poisons, Interscience (1967).
508. C. V. Kanter, R. G. Lunche and A. P. Fuderich, "Techniques of Testing
for Air Contaminants from Combustion Sources," JAPCA 6_, 191 (1967).
512. Wasser, et al., "Effects of Air Fuel Stoichiometry on Air Pollution
from Oil-Fired Furnaces," JAPCA 18, 332 (1968).
518. R. T. Walsh, "Combustion Equipment," in Air Pollution Engineering
Manual, J. A. Daniel son (Ed.), NAPCA, PHS Pub. No. 999-AP-40,
Cincinnati, Ohio (1967).
521. "Reference Book of Nationwide Emissions, June 1969 Estimates," Bureau
of Criteria and Standards, Division of Air Quality and Emission Data,
NAPCA, Durham, N.C.
525. B. D. Bloomfield, Air Pollution. Vol. II, A. Stern, Ed., p. 487 (1968).
536. 6. C. Jefferis and J. D. Sensenbaugh, "Effects of Operating Variables
on the Stack Emissions from a Modern Power Station Boiler," ASME Paper
59-A-308 (1959).
547. ASTM, "Oxides of Nitrogen in Gaseous Combustion Products, D1608-60
(1970).
548. R. Larkin, NAPCA, Cinncinati, Ohio, Private Communication (1969).
556. ABMA Reports for 1968.
557. , Bureau of Census, Dept. of Commerce, Form M34N (1967).
560. H. Phillips, Foster Wheeler Corp., Livingston, New Jersey, Private
Communication (1970).
561. J. P. Kubik, Riley Stoker Corp., Worcester, Mass., Private Communica-
tion (1970).
562. D. J. Callaghan, Bay Area APCD, San Francisco, Calif., Private Com-
munication (1970).
1 50 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
563. P. Matthew, Pacific Gas and Electric Co., California, Private Com-
munication (1970).
564. B. Dimitriades, Bureau of Mines, Bartlettsville, Oklahoma, Private
Communication (1970).
565. H. Lang, et al., "Continuous Monitoring of NO, CO " Bureau of
Mines Publication, R.I. 7241 (1969).
566. H. Lang, Bureau of Mines, Brewston, Pa., Private Communication (1969)
567. J. T. Shaw and A. C. Thomas, "Oxides of Nitrogen in Relation to the
Combustion of Coal," paper presented at the 7th International Con-
ference on Coal Science, Prague (June 10-14, 1968).
568. A. J. Elshout and H. Van Duren, Electro-techniek 46_, 251 (1968).
569. A. A. Jonke, "Reduction of Atmospheric Pollution by Fluidized Bed
Combustion," Argonne National Laboratory, Monthly Progress Report
No. 8 (March 1969).
570. K. Peters and H. Strachil, "Neue Methoden zur Analyse Nitrose-
haltiger Case," Angewandte Chem. 68, 291 (1956).
571. T. K. Sherwood and R. L. Pigford, Absorption and Extraction,
p. 368-383, McGraw Hill Book Co., Inc., New York (1952).
572. M. M. Wendel and R. L. Pigford, "Kinetics of Nitrogen Tetroxide Ab-
sorption in Water," AIChEJ 4, 249 (1958).
573. F. S. Chambers and T. K. Sherwood, "Absorption of Nitrogen Dioxide
by Aqueous Solutions," I. & E.C. 29, 1415 (1937).
574. L. C. Eagleson, R. M. Langer and T. H. Pigford, S.M. Thesis, M.I.T.
(1948).
575. G. B. Taylor, T. H. Chilton and S. L. Handforth, I. & E.C. 23_, 860
(1931).
576. M. Bodenstein, Helv. Chem. Acta 18., 743 (1935).
577. P. E. Bolshakoff, S.M. Thesis, M.I.T. (1934).
578. M. S. Peters and J. L. Holman, I. & E.C. 47, 2536 (1955).
579. M. S. Peters, C. P. Ross and J. E. Klein, AIChEJ 1, 105 (1955).
580. K. G. Denbigh and A. J. Pornie, J. Chem. Soc. 59_, 316 (1937).
1 51 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
581. P. G. Candle and K. 6. Denbigh, Trans. Faraday Soc., 49_, 39 (1953).
582. M. S. Peters and E. J. Koval, I. & E.C. 51_, 577 (1959).
583. E. J. Koval and M. S. Peters, I. & E. C. 52_, 1011 (1960).
584. R. R. Wenner, Thermochemical Calculations, McGraw-Hill Book Co., Inc.,
New York (-1941T
585. W. R. Forsythe and W. F. Giaque, J. American Chem. Soc. 64, 48 (1942).
586. E. Abel, H. Schmidt and M. Z. Stein, Electrochem. 3£, 692 (1930),
587. D. A. Epshtein, J. Gen. Chem. (USSR) 9, 792 (1939).
588. F. H. Verhoeh and F. Daniels, J. American Chem. Soc. 53_, 1250 (1931).
i
589. J. H. Wasser, NAPCA, Cinncinati, Ohio, Private Communication (1J969).
590. G. K. Lee, F. D. Friedrich and E. R. Mitchell, "Control of 503 in Heat-
ing Boilers by an Additive," J. Inst. Fuel 337, 69 (Feb. 1969).!
591. N. Camac and R. N. Feinberg, Eleventh Symposium (International) on
Combustion, 137, The Combustion Institute (1967).
592. E. Freedman and J. W. Daiber, J. Chem. Phys. 34_, No. 4, 1271 (T961).
593. H. S. Glock, J. J. Klein and W. Squire, J. Chem. Phys. 27, No. 4
(1957).
594. F. Kaufman and 0. R. Kelso, J. Chem. Phys. 23_, No. 9, 1702 (1955).
595. K. Vetter, Z. Electrochem 53, 369, 376 (1949).
596. H. Wise and M. F. Freeh, J. Chem. Phys. 20, No. 1, 22 (1952).
597. H. Wise and M. F. Freeh, J. Chem. Phys. 20, No. 11, 1724 (1952).
598. E. L. Yuan, J. I. Slaughter, W. E. Koerner and F. Daniels, J. Phys.
Chem. 63, 952 (959).
599. M. H. Bortner, et al., Report AD 600629, Army Rocket and Guided Mis-
sile Agency Contract No. DA-36-034-ORD-3187(RD), General Electric
(February 1961).
600. H. K. Newhall, Twelfth Symposium (International) on Combustion, 603,
The Combustion Institute (1969).
1 52 WALDEN RESEARCH CORPORATION
-------
LITERATURE CITED (continued)
601. D. L. Baulch, D. D. Drysdale, D. G. Dome and A. C. Lloyd, "Critical
Evaluation of Rate Data for Homogeneous, Gas Phase Reactions of In-
terest in High-Temperature Systems," School of Chemistry, The
University, Leeds, supported by Science Research Council, England
(December 1969).
602. T. H. Chilton, "Strong Water, Nitric Acid: Sources, Methods of Manu-
facture and Uses," The MIT Press, Cambridge, Mass. (1968).
604. J. R. Partington, A Textbook of Inorganic Chemistry. 5th Ed., MacMillan,
London, 579-580 (19397:
605. E. Saunders and E. Rissman, "Study of the Feasibility of Oxidizing
the S02 in Power Plant Flue Gases to Sulfuric Acid," First Interim
Progress Report of Tyco Laboratories, Inc., on Contract No. PH-86-
68-75 (July 1968).
607. H. H. Haaland, Editor, "Methods for Determination of Velocity, Volume,
Dust and Mist Contents of Gases," Bulletin WP-50, 7th Ed., Western
Precipitation Division, Joy Manufacturing Co., Los Angeles (1968).
609. "Determining Dust Concentration in a Gas Stream," PTC 27-1957, The
Amer. Soc. of Mech. Engrs., N.Y. (1957).
611. K. Fukui, "Alkaline Filter Paper Method for Collection of Sulfur and
Nitrogen Oxides," 2nd Int. Conf. on Air Pollution, London (1966).
612. W. K. Lewis and A. H. Radash, Industrial Stoichiometry. McGraw-Hill,
N.Y. (1926).
613. Prof. Paul Giever, University of Michigan, Private Communication (1969).
614. ASTM D-22 Subcommittee VI, Tentative Standard Method for Sampling
Stacks (1970).
619. J. N. Driscoll, et al., "Determination of Oxides of Nitrogen in Com-
bustion Effluents with a Nitrate Ion Selective Electrode," APCA Paper
#71-149 presented at the 64th Annual APCA Meeting (June 1971).
620. B. A. Coulehan and H. W. Lang, "Rapid Determination-of Nitrogen Oxides
with Use of Phenol Disulfonic Acid," Environmental Science and Tech-
nology 5_, 163 (1970).
621. J. H. Wetters and K. L. Uglum, "Direct Spectrophotometric Simultaneous
Determination of Nitrite and Nitrate in the Ultraviolet," Anal. Chem.
42, 335 (1970).
1 53 WALDEN RESEARCH CORPORATION
-------
APPENDIX 1
LITERATURE SEARCH
A comprehensive search of the literature was conducted to learn of
all methodology which might be applicable to this program. This includes
methods currently in use, those used in the past which seemed appropriate,
and information from other areas which seemed relevant.
In order to perform this study, we made a thorough search of the lit-
erature both past and present. Thus, we covered journals, abstracts, in-
dices, and bibliographies shown in Tables Al-1 and Al-2. By this approach,
we feel we covered all journals in the western world and many others which
might have any pertinence to this problem. To supplement this, we re-
quested and obtained a literature search by APTIC on sampling and analyti-
cal methods for the pollutants of interest.
As articles of interest were uncovered by the literature research
staff, the papers were studied by analytical chemists and significant
information abstracted for further evaluation. As the study progressed,
group discussions were held with chemists with relevant experience in
order to both evaluate existing methods and seek fresh alternatives.
155
WALDEN RESEARCH CORPORATION
-------
TABLE Al-1
PRIMARY LITERATURE SOURCES UTILIZED IN THIS PROGRAM
Air Pollution Control Abstracts 1956-1969
Air Pollution Titles Thru 1969
American Petroleum Institute Proceedings 1962-1969
Analytical Abstracts 1954-1969
Applied Science and Technology Index 1939-1969
Chemical Abstracts 1907-1968
Chemical Titles Thru 1969
Engineering Index 1939-1969
Fuel Abstracts and Current Titles 1960-1969
SAE Progress in Technology Series 1961-1967
TABLE Al-2
BIBLIOGRAPHIES
Air Pollution Publications: A Selected Bibliography
(DHEW) 1963-1966
DDC Bibliography on Air and Water Pollution (1968)
AD 679210
Research into Sampling, Analysis and Monitoring of
Gaseous Pollutant Emissions from Stacks
Literature Searches I and II (Engineering Found)
1963,1965
Sulfur Oxides and Other Sulfur Compounds
A. G. Cooper, DHEW 1965
1 56 WALDEN RESEARCH CORPORATION
-------
APPENDIX 2
SAMPLING SYSTEM FOR SOV AND NOV
A X
Equipment for sampling SO (and NO ) gases from combustion flue gases
X X
has been designed. The objectives of the design were first to provide high
precision and secondly, to make the equipment convenient to transport, in-
stall, and operate at field locations. The system is comprised of the fol-
lowing three modules:
1. dual probe module
2. SOV and N0v col
X . X
3. control module
2. SO and NO collection module
X . X
The above modules constitute a complete system for sampling for SO.,.
S09, and NO in the flue gas.
£~ r\
Probe Module - A schematic of the probe module is shown in Figure
A2-1. The features of this module are:
1. Dual probes, a 1/4-inch diameter stainless steel probe for NO
X
mechanically coupled to 11 mm pyrex probe for SO .
X
2. Pyrex probe electrically heated to prevent condensation of SO.,
in the sampling line.
3. Stack adapter assembly that allows various probe insertion
depths.
gas,
4. Filter in the probe to remove particulates from the SOV sample
X
5. Quick attachment tees for NO grab samplers.
X
6. Pyrex socket joint for connection to collector module.
157 WALDEN RESEARCH CORPORATION
-------
Uer
tn
Co
. Asbestos
Insulation
Heating
Wire
\ s
"
/\ /y A /\ A A
/ \ / \ / \ / \ / \/~\)
f \f XT \j i \j XT vr
\
11 mm Dia,
Pyrex
1 inch
Stainless
Tube
Stack Adapter
1/4" Diameter
Stainless Steel Probe
for NCL
Quick Tee
r-
Pyrex Socket Ooir,
5
30
m
Figure A2-1. Probe Module.
o
o
o
•yo
s
•
o
-------
APPENDIX 3
SIMULTANEOUS EQUILIBRIUM PROGRAM
DESCRIPTION - DEQL
The program DEQL was designed to carry out the equilibrium calcula-
tion for a series of chemical reactions occurring simultaneously. The
calculation follows an iterative procedure, examining each reaction sepa-
rately and, in each case, bringing the concentrations of the reactants to
equilibrium. When the change in reactant concentrations on successive
iteration steps becomes less than some preset margin of precision, the
iteration is stopped and the final reactant concentrations are printed
out. Experience with this program has shown that convergence is quite
rapid. In one case with eight reactions and with a margin of precision
of .1%, only four iteration cycles (each cycle calculates equilibrium for
all eight reactions) were required for convergence on a solution.
In its present form, the program allows up to eight reactions with
six reactants to be included. As input it requires the reaction tempera-
tures, the reaction equilibrium constants, and the initial concentrations
of the reactants. The program output gives the equilibrium concentrations
of the reactants.
DEQL has been written in FORTRAN IV and has been used at Wai den on a
time shared IBM 360/67 computer. Running times for the program have been
about four seconds, CPU time (including program loading and execution) for
a calculation involving eight reactions and about four iteration cycles.
WALDEN RESEARCH CORPORATION
-------
APPENDIX 4
LABORATORY STUDIES OF THE CHROMATROPIC
ACID METHOD FOR NITRATE
We have investigated the use of chromatropic acid (2,7-dihydroxy-
naphthalene-4,5-disulfonic acid; abbreviated "CA") (10) for analysis of
nitrogen oxides because of its apparent high selectivity for nitrate and
ease of application. CA is made up as an 0.01% solution in concentrated
H2S04, and 7 ml of this solution are mixed with 3 ml of aqueous sample.
The reaction mixture is allowed to cool to room temperature, and the
yellow color resulting from nitration of the CA is measured spectrophoto-
metrically. The CA reagent in 70% H2S04 lias absorPtion maxima at 234 nm,
310-316 nm (broad), 332 nm and 346-348 nm. Reaction of the CA reagent
with aqueous samples containing 5-15 ppm nitrate, resulted in decreases
in all the reagent peaks (above) and the appearance of two new ones at
357 nm (appearing as a shoulder on the 347 nm reagent peak) and atJ420 nm.
When high concentrations (300-1000 ppm) of nitrate were used, the absorb-
ance in the 420 nm region did not obey Beer's law but decreased with in-
creasing amounts of nitrate.
The CA reagent also reacts with formaldehyde (CA is a well-known
colorimetric reagent for formaldehyde) to give a purple color, which ab-
sorbs slightly in the 420 nm region, A^Q for a 30 ppm CH20 sample is
equivalent to that from a 3 ppm of nitrate. Thus, the formaldehyde in-
terference is not very important. Since the CA gives a yellow-brown
color even with traces of oxidants, excess oxidant must be removed be-
fore resetting the sample with the CA reagent. We used a 30 ppm solution
of formaldehyde in water as the standard, although this represented a
larger amount than would be expected. We tested as oxidants 1% H^Op,
0.1 M KMn04 and 0.3 M K^Og. Up to 2 ml of H202 were allowed to react
with 10 ml of 30 ppm CH20 for up to one hour with no observable effect
on the CH20-CA color. The product of the reduction of KMn04 with Na2S03
was Mn02, which oxidized the CA reagent in the 70% H2SO, color develop-
ment solution. The KS^ (aPProxlmately °-3 M close to saturation) did
WALDEN RESEARCH CORPORATION
-------
remove the formaldehyde after standing overnight, but not after only six
hours of standing. Our experiments indicate that removal of CH20 at these
low levels will be very difficult to accomplish.
Since CA is extremely sensitive to oxidants (such as H202) normally
present in the acidic absorption solution used for sampling nitrogen oxides,
it would have to be removed by addition of a reagent such as Na2S03 or
boiling to decompose the peroxide. Another problem with CA is that the re-
action with nitrate produces a very broad absorption band in the 400-450 nm
region which may suffer from interferences due to unburned hydrocarbons.
Because of the lack of a unique absorption spectrum, the possible
interferences from traces of unburned fuel and QUO and sensitivity to
oxidants, chromatropic acid offers no advantages over the phenol disul-
fonic acid reagent. Investigation of this reagent for an improved method
was, therefore, abandoned.
162 WALDEN RESEARCH CORPORATION
------- |