&EPA
United States EPA-600/R-02-072
Agencymental Protection September 2002
Status of Semi-Continuous Monitoring
Instrumentation for Gases and Water-
Soluble Particle Components Needed for
Diagnostic Testing of Ambient Air Quality
Simulation Models
-------
EPA/600/R-02/072
September 2002
Status of Semi-continuous Monitoring
Instrumentation for Gases and Water-
Soluble Particle Components Needed for
Diagnostic Testing of Ambient Air
Quality Simulation Models
by
W.A. McClenny
Human Exposure and Atmospheric Sciences Division
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
and
K.G. Kronmiller, K.D. Oliver, H.H. Jacumin, Jr. and E.H. Daughtrey, Jr.
ManTech Environmental Technology, Inc.
Research Triangle Park, NC 27709
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and Development assembled the
information in this report with input from a number of prominent scientists and funded a research effort to
provide field data for this report under Contract 68-D-00-206 to ManTech Environmental Technology, Inc.
The report has been subjected to the Agency's peer and administrative review and has been approved for
publication as an EPA document. Mention of trade names or commercial products does not constitute endorse-
ment or recommendation for use.
-------
Abstract
This status report documents the effort by the National Exposure Research Laboratory (NERL), Office of
Research and Development (ORD) of the U.S. Environmental Protection Agency (U.S. EPA) to provide
methods of measurement, including calibration procedures, that meet the requirements for diagnostic testing
of models describing atmospheric photochemistry. These requirements include a measurement uncertainty of
±0.2 ppbv (±0.1 for N02) at 1.0 ppbv (or equivalent for particles) for compounds mentioned here as specified
by scientists in the Atmospheric Modeling Division, NERL. New developments for monitors of N02, HCHO,
H202, HN03, HONO, NH3, volatile organic compounds (VOCs) (including ^-aldehydes), particulate nitrate,
particulate sulfate, and particulate ammonium are discussed in the report. Descriptions of instrument design
and principles of operation for N02, HCHO, and water-soluble gases and particulate matter are presented.
Since field comparisons among different methods are often the accepted approach for identifying
monitoring problems and establishing consensus among scientists specializing in monitoring methods, recent
activities have emphasized field studies. Results from the 2002 Tampa, FL, Bay Regional Atmospheric
Chemistry Experiment (BRACE), the 2001 Philadelphia, PA, Northeast Ozone, Particle Study (NE-03PS)
study, and the 1999 Nashville, TN, Southern Oxidants Study (SOS) are shown. Of particular note are the
preliminary results of field monitoring of N02 in the BRACE study with systems using new commercial
photolytic converters for N02 to NO conversion followed by chemiluminescence detection. In the NE-03PS
field study, comparisons of HN03, HONO, S02, and nitrate and sulfate in particulate matter between the
Harvard School of Public Health (HSPH) EPA Denuder System (HEADS) (10-h average) and the average
response from the Texas Tech University (TTU) semi-continuous system (15-min updates) provide an
interesting (and so far unresolved) example of a multiplicative error in a study database. Current plans are for
a more definitive statistical comparison of methods using the existing field monitoring database in the NE-
03PS and BRACE studies and additional field comparisons of methods in the upcoming 2003 Seattle field
study.
-------
Foreword
The National Exposure Research Laboratory, Research Triangle Park, NC, performs research and development
to characterize, predict, and diagnose human and ecosystem exposure, giving priority to that research which
most significantly reduces the uncertainty in risk assessment and most improves the tools to assess and manage
risk or to characterize compliance with regulations. The Laboratory seeks opportunities for research
collaboration to integrate the work of ORD's scientific partners and provides leadership to address emerging
environmental issues and advance the science and technology essential for understanding human and ecosystem
exposures. One aspect of the Laboratory's mission is to support the iterative process of comparing model
predictions with experimental observations so that ultimately the reconciliation of differences is accomplished.
In the case of ozone production in the ambient atmosphere through photochemical processes, this support
includes the identification and/or development of monitoring instrumentation of sufficient quality to evaluate
the impact of control strategies.
EPA has established a strategic plan for its scientific program of research and development that includes
production of reports that respond to die perceived needs in support of public interest through compliance with
the Government Performance and Results Act (GPRA). The current report is GPRA Annual Performance
Measure (APM) 13; it contains a status report on the development of ambient air monitoring methods to meet
the needs for diagnostic testing of air quality simulation models. These needs were identified precisely in the
September 2000 EPA report titled "Recommended Methods for Ambient Air Monitoring of NO, N02, NOY,
and Individual NOz Species" (EPA/600/R-01/005) published by NERL. Chapter 4 of the report provides the
performance objectives for monitoring the reactive oxides of nitrogen, as the report title indicates, and also for
a number of other species, reaction rates, and process rates that are required to understand the prevalence and
extent of air pollution. If these performance objectives can be met, then the iterative process of refinement of
existing air quality simulation models can continue. The ultimate objective is to obtain an agreement between
predictions and measurements so that the efficacy of regulatory measures can be assured.
Gary J. Foley, Ph.D.
Director
National Exposure Research Laboratory
Research Triangle Park, NC 27711
iii
-------
Contents
Notice ii
Abstract ii
Foreword iii
Figures v
Tables vi
Acronyms and Abbreviations vii
Acknowledgments !... viii
Chapter 1 Conclusions 1
Chapter 2 Recommendations 3
Chapter 3 Introduction 4
Chapter 4 Status of Monitors for Trace Gases and Water-Soluble Particle Species 6
4.1 Status of Instrumentation for Photolytic Conversion of N02 to NO 6
4.1.1 The NOAA/TEI-Oriel Stand-Alone Photolytic Converter 6
4.1.2 Status of Laser-Based Photolytic Converter for N02 11
4.1.3 Comparison of N02 and NOY Monitors 11
4.2 HCHO and H2Oz Microchemical, Semi-continuous Monitors for Use in
Diagnostic Testing 13
4.2.1 The 1999 Nashville Summer Field Study 14
4.2.2 The 2001 NE-O3PS Philadephia Field Study 15
4.3 Instrumentation for H202 and MHP Monitoring 17
4.4 HNO3, HONO, S02, NH3, and Other Water-Soluble Gases—Semi-continuous
Monitors Based on Collection Using Diffusion Denuders and Analyses by IC 17
4.5 Semi-continuous Speciation Monitors for Water-Soluble Components of
Ambient Particulate Matter 19
4.6 Sample Integrity of VOCs Collected from the Ambient Air 21
4.6.1 Sampling Ambient Air Spiked with VOCs and 03—Comparison of
Adsorbent Tubes with Canisters and an AutoGC System 21
Chapter 5 References 26
Appendix A: Development of a Calibration Device for HONO 27
Appendix B: Measurement of Inorganic Gases and Particulate Matter in Philadelphia NE-03PS 2001 .. 35
iv
-------
Figures
4-1 Design of NOAA photolytic converter 7
4-2 Two-monitor analysis system for NO, N02 8
4-3 Factors related to photolytic converter performance in the Tampa BRACE study—
Gandy site 10
4-4 N0/N02 monitoring sequence for two days in the Tampa BRACE study at the
Gandy monitoring site .... 10
4-5 NO method comparisons for 9-10 May 2002 at the Gandy site: (a) NO vs time;
(b) scatter diagram of HCMS results vs EPA-PC results 12
4-6 N02 method comparisons for 9-10 May 2002 at the Gandy site: (a) N02 vs time;
(b) scatter diagram of HCMS results vs EPA-PC results 12
4-7 NOY ARA results and EPA-PC (NO + N02) results at the Gandy site: (a) Results vs
time; (b) scatter diagram of ARA results vs EPA-PC results 13
4-8 Schematic for monitor of HCHO ambient air concentrations 14
4-9 Example of results of HCHO monitor calibration 14
4-10 Picture of commercial prototype HCHO monitor 15
4-11 HCHO monitoring sequence from 1999 Nashville field study 15
4-12 Comparison of HCHOmonitoring results in same-day sequence from 1999 Nashville
field study— Polk Building and Cornelia Fort 16
4-13 Comparison of results for research prototype and commercial versions of a HCHO
monitor 16
4-14 Comparison of results of research prototype HCHO monitor and the DNPH procedure
for HCHO 17
4-15 Schematic of acid gas monitor 18
4-16 Example anion analysis by ion chromatography for acid gases 19
4-17 Schematic of water-soluble particle speciation monitor 20
4-18 Example of anion analysis by ion chromatography for particle component analysis 20
4-19 Octanal response by GC/MS as a function of ozone concentration—comparison of results
using canisters, sorbent tubes, and an autoGC 24
4-20 Performance of Mn02 ozone scrubber 25
v
-------
Tables
3-1 Monitoring Requirements for Diagnostic Testing of Air Simulation Models 4
4-1 Parameters for Photolytic Conversion of N02 to NO 9
4-2 Effect of Co-collecting Ozone on Integrity of Hydrocarbon Response (Selected Compounds) 22
4-3 Effect of Co-collecting Ozone with w-aldehydes on Graphitic Carbon Solid Sorbent 23
vi
-------
Acronyms and Abbreviations
APM
Annual Performance Measure
NE-O3PS
Northeast Ozone, Particle Study
ARA
Atmospheric Research and Analysis
NERL
National Exposure Research
Inc.
Laboratory
AutoGC
automated gas chromatograph
NMDS
Nation membrane diffusion scrubber
BRACE
Bay Regional Atmospheric Chemistry
NOAA
National Oceanic and Atmospheric
Experiment
Administration
BNL
Brookhaven National Laboratory
NOx
sum of NO plus N02
CF
Cornelia Fort
noy
sum of NOx and other reactive
CRADA
Cooperative Research and
nitrogen compounds
Development Agreement
NOz
NOY-NOx
DDL
1,4-diacetyldihydrolutidine
ORD
Office of Research and Development
DI
deionized
PAMS
Photochemical Assessment
DNPH
dinitrophenylhydrazine
Monitoring Stations
DVP
distributed volume pair
PPDD
parallel plate diffusion denuder
EPA
Environmental Protection Agency
ppbv
parts per billion by volume
EPCHC
Environmental Protection Commis-
ppm-L
parts per million per liter
sion of Hillsborough County
RH
relative humidity
FRM
Federal Reference Method
R&P
Rupprecht & Patashnick
GPRA
Government Performance and Results
SOS
Southern Oxidants Study
Act
STI
Sonoma Technology Inc.
HCHO
formaldehyde
TDL
Tunable diode laser
HEADS
Harvard EPA Denuder System
TEI
Thermo Environmental Instruments
HONO
nitrous acid
TTU
Texas Tech University
HSPH
Harvard School of Public Health
TVA
Tennessee Valley Authority
IC
ion chromatography
U.S. EPA
U.S. Environmental Protection
LED
light emitting diode
Agency
M
molar
USF
University of South Florida
MHP
methyl hydroperoxide
VOCs
volatile organic compounds
NAAQS
National Ambient Air Quality
Standards
vii
-------
Chapter 1
Conclusions
The following conclusions reflect the status of applied research
and development efforts performed under National Exposure
Research Laboratory (NERL) sponsorship to develop methods
for diagnostic testing of air quality simulation models.
• The current set of research prototype monitors for
diagnostic testing of atmospheric photochemical
models meets the requirements for time response of 1
hour or less as requested by the atmospheric modelers
for their list of primary and secondary chemical
species.
• The results from recent field studies in Tampa, FL, and
Philadelphia, PA, are currently being examined to
determine if the uncertainty requirements of ±0.2 parts
per billion volume (ppbv) (± 0.1 ppbv for N02) at 1
ppbv were attained.
• A critical supplement to the monitors themselves is a
set of calibration systems with certified or verifiable
standard calibration concentrations. Unless the
accuracy of the calibration systems can be established,
the accuracy of these measurements will remain in
doubt.
• Exchange of standards for participants in field study
comparisons is essential for method comparisons if a
common calibration system is not used.
• Field monitoring protocols should include provision for
full system calibrations with a frequency based on
experience operating the system. Calibration of the
sampling and analytical portions of the monitoring
system is critical to determine if the monitor is under
control and performing as expected. One example of a
calibration system for nitrous acid (HONO) is given in
Appendix A.
• A characteristic feature of gas monitors for the acid
gases HN03, HONO, and S02 is the sequence of
monitoring stages: (1) sample collection that involves
some interface with the air (a diffusion scrubber for
example) and (2) analysis of the target analyte (by ion
chromatography (IC) separation and electro-
conductivity detection). While part 1 is easy to calibrate
by introducing liquid solutions containing known
amounts of anions (N03* for instance), the calibration
of both 1 and 2 is not. It requires a stable source of the
target gas in a humidified, synthetic airstream for
venting past the inlet to the monitor. The total
instrument must be calibrated frequently during initial
operation and less frequently as experience dictates.
• Co-collected ozone destroys the sample integrity of
volatile organic compounds (VOCs) during sampling
and analysis to an extent depending on the compound,
the mode of sample introduction, and the amount of
ozone. An automated gas chromatograph (autoGC)
system with a Tenax GR concentrator, specially
prepared canisters, and sorbent tubes with a packing of
graphitic carbon followed by a carbon molecular sieve
were examined. Aromatic hydrocarbons beyond
toluene were noted to exhibit small but increasing
losses for canisters and sorbent tubes as ozone
concentration was increased to 300 ppbv. Under the
same conditions, nonanal, octanal, and other n-
aldehydes were dramatically reduced in canisters and
with tubes but were increased for the autoGC. Mn02
scrubbers proved effective in eliminating the artifacts
but had short useful lifetimes with respect to efficient
passage of the w-aldehydes.
• Several of these monitors have been engineered for
commercial manufacture and are now sold to the
public. These include the following:
1
-------
- Stand-alone N02 photolytic converters for use
with commercial NO, 03 chemiluminescence
monitors. Depending on instrument configura-
tion, real-time measurements can be obtained.
Design of these monitors is still evolving with
the prospect of using new laser diode or light-
emitting diode (LED) light sources in lieu of
existing broadband light sources. Oriel Instru-
ments of Stratford, CT, markets a converter
with a Hg arc lamp. Use of this converter during
a month-long study in May 2002 by NERL in
the Tampa Bay Regional Atmospheric Chemis-
try Experiment (BRACE) was successful in
demonstrating the converter as a system com-
ponent for sub-ppbv measurements of N02.
Observations have so far been limited to moni-
toring results at the Gandy site in Tampa.
Limited comparisons with adjacent open-path
optical absorption monitors using differential
optical absorption spectroscopy and with col-
located high-performance "NOy" monitors are
consistent.
- Formaldehyde (HCHO) monitors based on
microchemistry and associated technology as
developed at Texas Tech University (TTU) by
Dr. P.K. Dasgupta and his associates. These
units provide 10-minute updates and can
operate unattended for 10 days before replen-
ishment of chemicals is required. Alpha-Omega
Power Technologies, Inc., of Albuquerque,
NM, markets this instrument. Use of both the
TTU research prototype and the Alpha-Omega
unit in the 2001 Philadelphia Northeast Ozone,
Particle Study (NE-03PS) indicates that agree-
ment of the two units in side-by-side compari-
sons is highly correlated but not precise enough
to meet requirements for diagnostic testing. An
offset of 0.73 ppbv is also indicated for the
commercial unit using a simple linear regression
analysis.
Equipment to monitor H202 and methyl hydro-
peroxide (MHP) that is functionally identical to
HCHO equipment. However, comparison with
other methods is still to be accomplished. A
commercial version is being offered by Alpha-
Omega Power Technologies, Inc.
• Other monitors are sufficiently well developed to be
reliably used as research prototypes and could be
manufactured commercially if sufficient incentives
were available. These include the following:
Acid gas monitors for ambient HN03, HONO,
and S02 as well as for ambient NH3, using
collection in water-based (wet wall) diffusion
denuders with ion chromatographic analysis.
The sample collection in water is followed by
accumulation of sample in a concentrator tube.
The collection period is short (e.g., 15 min) and
is followed by IC separation and detection of
individual compounds by electro-conductivity.
URG, Inc. of Carrboro, NC, has developed a
commercial prototype unit of this type.
- Particle speciation monitors for nitrate, sulfate,
and ammonium as well as other anions and
cations, using particle growth by the addition of
water vapor (or particle incorporation in water
mist), capture of the sample by impaction/
condensation, collection of sample over a short
time period (e.g., 15 min), and then separation
by IC and detection by electro-conductivity.
URG, Inc. has developed a commercial pro-
totype unit of this type.
Note: The particle speciation monitoring discussed in
this report addresses only the water soluble
particulate species which can be performed on
the same equipment and at the same time as
water soluble gases. This activity is being
coordinated with Dr. Paul Solomon, NERL,
Environmental Protection Agency (EPA), Las
Vegas, NV.
• Many of the instruments in or emerging from the
development phase require research personnel or
skilled technicians for optimum operation. New
operators typically require training and significant
hands-on experience to ensure successful operation.
2
-------
Chapter 2
Recommendations
Although techniques and instruments exist for meeting many of
the measurement requirements of diagnostic testing, additional
equipment testing and protocol development is needed in order
to ensure data quality. In particular, our recommendations
include the following:
• Complete data analyses from existing databases to
make comparisons more definitive.
• Perform interference testing of those photolytic
converters using broadband light sources to verify that
optical components are designed to eliminate inter-
ferences during photolysis of N02. Interference from
HONO is of particular concern with broadband sources
and is subject to the characteristics of the light source
and individual optical components (reflective and
absorptive) between the source and the photolysis cell.
• Continue efforts to provide an optimum light source for
photolytic converters, specifically diode laser sources
and light emitting diodes.
• Perform comprehensive testing of the new commercial,
semi-continuous HCHO monitors under different atmo-
spheric conditions and additional comparison with the
dinitrophenylhydrazine (DNPH) method of time-
averaged monitoring. The semi-continuous HCHO
monitors are potential replacements for the DNPH
method but must be proven user friendly for wide-
spread use.
• Continue comparison of acid gas and NH3 monitors
with time-integrative techniques to establish compar-
ability of results. Data collected in the 2001 Phila-
delphia NE-O3PS summer study with the Harvard EPA
Denuder System (HEADS), a system operated by the
Brookhaven National Laboratory (BNL), and a system
operated by personnel from TTU are being compared
(limited results are included in this report).
• Examine particle size effects using particle speciation
monitors. Speciation monitors are a main topic of
research and evaluation as part of the EPA Supersites
program where this effort is coordinated.
• Establish an in-house calibration facility for H202 and
field test the commercial H202 monitor being devel-
oped by Alpha-Omega Power Technologies, Inc.
• Develop an EPA in-house capability to assist users of
diagnostic testing instruments in developing a pro-
ficiency with these instruments and methods of calibra-
tion and maintenance.
• Establish an operational set of these instruments for
diagnostic testing to use in field tests.
• Test procedures to rejuvenate Mn02 ozone scrubbers
for use in conditioning air samples containing surface
reactive VOCs. Current Mn02 scrubbers are only good
for 0.5 parts per million per liter (ppm-L) of sample.
• Develop and demonstrate new ways to reliably measure
polar VOCs using, for example, two-dimensional
chromatography.
• Develop written protocols for the methods discussed in
this report, including specific verifiable calibration
procedures.
3
-------
Chapter 3
Introduction
This report is a status report on the effort by NERL to develop
and evaluate monitoring instrumentation for several gases and
for some water-soluble particulate species in the ambient air.
The species of interest are those specified by EPA atmospheric
modelers in a previous EPA report (EPA/600/R-01/005) dated
September 2000. Chapter 4 of that document provides the
rationale for monitoring requirements as well as for the required
response time and precision for each species. The species of
interest are shown in Table 3-1. For each, monitoring should be
continuous or semi-continuous with no greater than 1-hour
updates. Concentration measurement requirements for gases are
1 ppbv with no greater than ±20% uncertainty for gases and
particle equivalent concentrations except for the more stringent
requirement of ± 10% uncertainty for N02. With respect to Table
3-1, certain of the species (PANs, RONOz, ROz, total peroxides,
OH, N03 radical, speciatedRCHO [although the ^-aldehydes are
addressed herein], methacrolein, and methyl vinyl ketone) are
being addressed elsewhere. The particulate components nitrate,
sulfate, and ammonium were added to the report since they were
available with little additional effort when measuring the water-
soluble gases and are major particle components.
Table 3-1. Monitoring Requirements for Diagnostic Testing of Air
Simulation Models
First Priority 03, NO, N02, NOY, HzO, HCHO, HONO,
PAN(s), HN03i RON02 and total N03, H202
and total peroxides, H02, R02, CO, speciated
VOCs
Second Priority OH, NO, radical, speciated RON02, speciated
RCHO, methacrolein, methylvinyl ketone,
aerosol mass and size distribution
The rationale for this activity is that better (more accurate,
more precise, and semi-continuous) monitoring techniques are
necessary to develop accurate, predictive models based on
information about sources, sinks, meteorological conditions, and
atmospheric processes. Provided these techniques are available,
the predictions of the models can be verified during the evolu-
tion of atmospheric components from source to sink (emission
to reaction or deposition). Thus, at points in the aging of an air
mass the predictions can be checked by measurements. Agree-
ment between measurements and predictions implies an under-
standing of what is happening in the atmosphere and the
capability to base emissions controls on the predictions of air
simulation models. The term diagnostic testing is used in this
and related work to denote the procedure for testing the
agreement between theory (model predictions) and experiment
(reality) at various stages in the photochemical evolution of an
air mass.
The objective of this report is to provide a statement of
status as of May 2002 for achieving the modelers' requirements
for diagnostic testing and to document the notable advances that
have been realized since September 2000 through the efforts of
EPA and others in addressing monitoring needs. In particular,
the advances in instrumentation design have resulted in
compound-specific point monitors for several gas and particle
species. These achievements are of interest to state and local air
monitoring agencies as well as to atmospheric modelers. Of
particular note in this regard is the availability of new
instrumentation to monitor true N02. Since N02 concentrations
are generally reported as an upper limit by instruments used in
the national installed instrument base, the use of specific NOz
monitors would reduce the concentrations being reported. This
distinction may be crucial if the monitored concentrations are
above the national ambient air quality standards and would
otherwise trigger expensive control measures.
Other noteworthy advances include the refinement and
field evaluation of semi-continuous monitors for HCHO, HN03,
HONO, and water-soluble particle species, including nitrate and
sulfate. These achievements have occurred in large part due to
the efforts of Dr. P.K. Dasgupta and his associates at TTU and
have been supported, in part, by NERL. Much of this develop-
ment has occurred with techniques that cause the compound of
interest to be captured in water followed by some method for
detection in solution.
For water-soluble gases, a water-based diffusion scrubber
removes the gases while the particles move through the scrubber
under laminar flow conditions. Once in solution, the gas species
4
-------
can, in some cases, be detected by adding liquid reagents to the
solution, thereby creating a derivative compound that is more
easily detected, e.g., by fluorescence. Alternatively, water
containing the gases can be passed through an IC concentrator
column for a short period of time (typically 10 min), at the end
of which the concentrated ions are released onto an IC column
for separation and then to an electro-conductivity detector.
For collection of water-soluble particles, the potentially
interfering gases are first removed by a diffusion denuder (as
mentioned above) and the particles are grown to a larger size by
the attachment of water molecules. Then the enlarged particles
are impacted on some type of collection surface for subsequent
analysis. Once in a water solution, there are two primary ways to
detect the compounds: (1) by thermal decomposition to simple
gases and detection by gas analyzers or (2) by IC separation and
ion specific analysis. An instrument based on thermal decompo-
sition is currently commercially produced by Rupprecht &
Patashnick (R&P) Inc., with versions for both nitrate (decom-
posed to NO and N02 gases and detected by NO, 03 chemi-
luminescence monitors) and sulfate (decomposed to sulfur
dioxide and detected by ambient air S02 monitors). The
diffusion scrubber/lC-based technique has evolved over a multi-
year time frame by different groups of scientists (see, for
example, the recent review of methods for monitoring nitrate in
particles).1 Prototype research-grade instruments using this
technique have been compared in Atlanta, GA, during the
summer of 1999 in the Southern Oxidants Study (SOS), but the
results are still pending.2
Both the HCHO and H202 monitors are based on the
techniques of derivatized fluorescence in which the compound
of interest is mixed with a liquid reagent to create a derivative
compound, followed by the stimulation and detection of fluor-
escence from the derivative compound. The HCHO and H202
monitors have now been refined to provide highly reliable
research-grade instruments typically operated by a dedicated
scientist or experienced technician. The HCHO monitor is
marketed in a commercial version by Alpha-Omega Power
Technologies, Inc., and, based on recent testing in the NERL
laboratories, it can be operated reliably by a trained non-
specialist; unattended operation has been demonstrated for
several days. The design of the HCHO monitors has evolved
through several iterations, and prototype research-grade instru-
ments have been used in field tests beginning with the SOS
studies in Nashville and Atlanta in 1999, followed by the Texas
2000 field study in Houston, the Philadelphia NE-03PS field
study in 2001, and, most recently, the Tampa BRACE study.
EPA's NERL has been a major sponsor of this development,
beginning with the activities leading up to the Nashville study
and continuing to the present. Routine and widespread use of
this technology may or may not occur depending on the number
of clients for semi-continuous HCHO measurements. There is no
doubt, however, that these measurements are needed to
understand atmospheric chemistry and to verify atmospheric
models through diagnostic testing.
The evolution of a monitor for H202 and MHP has so far
paralleled that for HCHO and is now in a research prototype
stage with a commercial version being prepared, again by Alpha-
Omega Power Technologies, Inc. The design of the instrument
is almost identical to that for HCHO but with different
derivatizing agents and different instrument components associ-
ated with fluorescence detection.
The speciation of VOCs is a requirement as input to
atmospheric models. For this reason, a limited effort was
undertaken to address a recurring problem of n-aldehyde
detection. This class of compound is variously reported as a
sampling artifact or as a true component of ambient air. The
basis for the problem is that n-aldehydes can be both created and
destroyed by the reaction of Os with aldehydes after the
aldehydes are collected prior to analysis. The extent of some of
these problems has now been documented by EPA as part of its
in-house research project.
The information that follows includes specifics on the
development and evaluation topics mentioned above, including
Appendices A and B, which provide discussion of a HONO
calibration method and of preliminary field study results
obtained in the NE-03PS 2001 summer study in Philadelphia,
respectively. Where available, comparison of methods using
different principles of detection is shown and interpreted as an
indication of the degree of uncertainty in the current capabilities
of the measurement techniques and their approaches to
calibration and quality assurance.
5
-------
Chapter 4
Status of Gas and Water-Soluble Particulate Monitors
4.1 Status off Instrumentation for Photolytic
Conversion of N02 to NO
As noted in Table 3-1, specific measurement of N02 in ambient
air is a requirement for diagnostic testing of air quality
simulation models that predict the evolution of 03 and other
gases as a result of photochemical reactions. The overwhelming
majority of monitors used for ambient N02 are based on the use
of thermal converters that reduce N02 to NO and detect the NO
by its chemiluminescence with excess 03. Thermal converters
convert several other reactive oxides of nitrogen into NO and
make the monitor non-specific. Various research groups have
replaced the thermal converter with a photolytic converter to
obtain a more specific measurement of N02 and have used them
successfully for limited-duration field studies.3 Recently Thermo
Environmental Instruments (TEI)-Oriel has placed a stand-alone
photolytic converter on the commercial market. This chapter
discusses (1) the features of the Model 80800 TEI-Oriel
photolytic converter and its evaluation and use in the Tampa
BRACE field study and (2) a promising photolytic converter
based on a laser diode source.
Note: State agencies use N02 monitors with thermal
converters to check compliance with the National Ambient Air
Quality Standards (NAAQS). Although N02 is usually the
dominant reactive nitrogen oxide, especially during the periods
of high concentration of most interest and importance in
measuring human exposure, other nitrogen oxides are produced
in the air during the formation of photochemical smog (and
sometimes emitted directly from sources) and constitute
interferences in the measurement of NO?. These monitors
actually provide upper limits to N02. The annual average of this
upper limit does not exceed the NAAQS for N02 (53 ppbv
annual average) at any location in the U.S. Hence, the N02
concentrations being estimated with monitors incorporating
thermal converters are conservative estimates of the actual N02
concentration and provide the necessary information to deter-
mine compliance with the N02 NAAQS. Under this set of
circumstances, there has been no need to improve specificity for
N02 monitors for NAAQS monitoring.
4.1.1 The NOAA/TEt-Oriel Stand-Alone
Photolytic Converter
Recently scientists at the National Oceanic and Atmospheric
Administration (NOAA) Aeronomy Laboratory established a
Cooperative Research and Development Agreement (CRADA)
with TEI-Oriel to build and market the NOAA design for a
photolytic converter. Based on continuing cooperation between
EPA and these groups, arrangements were made to field test the
first units resulting from the CRADA. EPA was also asked to
provide N02 measurements at two monitoring sites for the
BRACE study in Tampa, FL, during May 2002 and decided to
assemble systems that incorporated the photolytic converters.
EPA arranged to place two photolytic converters in operation on
l May 2002. The objective of these measurements was to supply
point measurements of N02 for the BRACE research database.
Although the units had been tested in the laboratory by EPA,
their deployment in Tampa was their first field test. As such,
their performance was also considered a performance evaluation.
Since several other monitors for N02 and NOy were being used
in the study, interesting comparisons of data were also projected.
Selected results from these comparisons are shown here.
The NOAA photolytic converter design for the photolytic
source and cell is shown in Figure 4-1 as pictured in the journal
article describing the scientific approach.3 An Oriel 200-W Hg
arc lamp was also the light source used in the Tampa BRACE
study. As noted in Figure 4-1, rotation of a shutter shields the
photolytic converter cell from the light source so that only NO
is monitored. A number of instrument features were added to the
commercial unit to ensure unattended operation over extended
periods of time. For example, the Oriel photolytic converter
system (Model 80800) incorporates a system for circulating
water through the lamp housing and an air-cooled radiator for
removing heat from the instrument interior. Air enters the Model
80800 from two ports and is channeled through two independent
sampling lines, one containing the photolytic converter and the
other containing a reference cell that matches the volume of the
illuminated cell. The sample and reference lines lead to identical
chemiluminescence monitors (see Figure 4-2) where the NO, 03
6
-------
reaction occurs. The reference line
provides a measurement of NO, and
the sample line provides a measure-
ment of NO plus the NO resulting
from photolytic conversion of a per-
centage (40% is typical) of the N02.
Simultaneous measurement of the air
in the two lines allows a subtraction
of NO from the NO plus converted
N02. Dividing by the conversion
fraction gives the N02 occurring at
the same time. High variability of NO
arid N02 in the ambient air, as in a
source-impacted location, can be
accounted for in this case. To estab-
lish this approach in the BRACE
study, EPA set up two TEI Model
42C trace-level chemiluminescence
monitors and a TEI Model 146 cali-
brator. The monitors could be run in
real time (see qualification given in
next paragraph), and accurate real-
time values of NO and N02 could be
obtained as mentioned above.
The possibility of small inter-
ferences in the chemiluminescence
signal led to a decision to monitor the
low-level chemiluminescence that is
known to occur in the reaction cell
due to wall reactions or gas-phase
reactions with other species. This is a
feature available on the TEI Model
42C as the "pre-reactor" signal, i.e.,
ozone can be added to the ambient
airstream in a chamber just in front
of the chemiluminescence chamber
so that 99% of the NO, 03 reaction
can occur prior to entering the chemi-
luminescence chamber. The residual
signal is subtracted from the signal
obtained when bypassing the pre-
reactor signal to obtain the "true" NO
readings.
TEI provided new software to
give a sequence of readings (NO for
10 s and then pre-reactor signal for
10 s) on a cyclic basis so that the pre-
reactor signal could be automatically
subtracted each cycle and the result
displayed during the next cycle as the
analytical signal. Implementing this
solution meant that concentration
variability over the cycle periods of
20 seconds would occur unless the
Positioning stage
Macor insulator
a
Hg arc lamp (200 W shown)
Ellipsoidal reflector
Lamp enclosure
Pyrex window
BG-3 filter
WG-345 filter
Shutter
Sample In
Quartz cell
10 cm
Sample out
Cell stop
Figure 4-1. Design of NOAA photolytic converter.
7
-------
NO
NO Channel
To Inlet Lines
<
Hg Lamp
N02 Channel
NO + fN02
(f = fraction)
First Model 42C Chemluminescense Monitor
(Run in NO Mode Only)
Flow
Sensor
-o-
w
NC
NO
0
Dry
At In
Flow
Senaor
Pressure
Transducer
I Sana
r-^o
Reaction
Chamber
ix\ \
J
Qei
Ozonator
Prereector
Dry
Air Out
Chamber
LJ
Permeation
Dryer
FHsr
PMT
V
Fiber
i±
Pump
Second Model 42C Chemluminescense Monitor
(Run in NO Mode Only)
I
Flow
Sensor
-o-
-cW
NC
NO
~
Dry
Air h
Flow
Sensor
-o-
Pressure
Transducer
SOcs
Gel
Electronics
Reaction
Chamber
-y/\
|PMT
Prereaetor
Dry
Air Out
t-
Permeation
Dryer
"n
Balaton
RKer
Pump
Figure 4-2. Two-monitor analysis system for NO, N02
8
-------
cycle times were exactly matched. Exact matching was not
possible to implement given the time constraints for the BRACE
study, so some anomalous difference signals still exist in the
database. EPA is currently evaluating the extent of this problem
in the database obtained in the BRACE study. To obtain
undistorted NO and N02 readings using the two-monitor
approach, the instruments should read the two lines simul-
taneously with periodic measurement of the pre-reactor to
account for a changing pre-reactor signal.
In the Tampa study, the photolytic converters and chemi-
luminescence monitors were set up at two sites, a rural site east
of Tampa called the Sydney site and a Hillsborough County
monitoring site (Gandy site) located near Gandy Blvd. close to
the Gandy Bridge over Tampa Bay and situated on a U.S.
Marine installation. The different aspects of equipment
installation and calibration went well with one exception. The
source of N02 calibration gas contained significant concen-
trations of NO. Tests indicated that the source of con-
tamination was in the reference cylinder or in the cylinder
regulator. Repeated venting of the standard from the tank
lowered but did not eliminate the NO contamination. Fortu-
nately, the primary and only indispensable reason for using the
N02 standard was to establish the photolytic conversion
efficiency, which could be done by knowing the N02 actually
reaching the photolytic converter. Hence, the contaminant NO
was measured by blocking the light from the broadband source
and the resulting NO concentration was then subtracted from
the cylinder-derived N02 concentration. This N02 concen-
tration was used as the true concentration reaching the photo-
lytic converter. The N02 photolysis resulted in increased NO
signal. This increase was ratioed to the true N02 concentrations
to obtain a conversion efficiency. As a result of the problems
associated with the use of a compressed gas cylinder of N02,
the alternate approach of gas phase titration (excess 03 with
amounts of added NO) is now being considered.
Zero, span, and converter efficiency values were recorded
every day during the period from 1 May to 31 May. The
consistency of these values for the Gandy site can be inferred
from Table 4-1.
Zero values were very stable and always within 0.1 ppbv of
0.0. Matching of ambient values on the two monitors was
obtained by closing the shutter on the light source and com-
paring the NO values recorded on the two monitors. These
values were typically within 0.1 ppbv at sub-ppbv concentrations
as seen in Figure 4-3. Instrument-specific span values and the
conversion efficiency of the photolytic converter are also shown.
The range of values for the photolytic converter efficiency was
36.3 to 42.1 with a slight downward trend as a function of time.
If the daily span values changed significantly for either of the
two chemiluminescence monitors, the span value was reset to
give the expected value. The span values were expected to be
stable to within 5% based on the manufacturer's specifications,
Table 4-1. Parameters for Photolytic Conversion of NO, to NO
May
Span Factor,
Monitor #1
Span Factor,
Monitor #2
Converter
Efficiency
30-31
1.04
1.13
35.5
29-30
1.00
0.83
36.6
28-29
0.97
0.93
38.0
27-28
1.01
1.04
38.3
26-27
1.00
0.98
36.2
25-26
0.99
0.99
35.7
24-25
0.99
1.01
34.9
23-24
0.99
0.88
35.7
22-23
1.02
0.90
36.8
21-22
1.05
1.02
36.2
20-21
0.99
1.02
36.4
19-20
1.00
1.00
36.0
18-19
0.98
1.03
36.7
17-18
0.97
1.17
36.4
16-17
0.95
0.94
36.3
15-16
0.98
0.98
36.6
14-15
1.01
1.03
36.7
13-14
1.00
1.00
36.6
12-13
1.04
0.98
37.2
11-12
1.03
1.02
37.2
10-11
1.00
1.01
37.2
09-10
1.01
1.10
36.9
08-09
0.98
0.88
36.6
07-08
1.00
1.07
37.6
06-07
0.99
0.94
38.9
05-06
1.01
1.02
37.5
04-05
0.99
1.04
37.0
03-04
1.00
0.97
38.5
02-03
1.02
0.97
41.3
01-02
1.00*
1.00*
42.1
Average
1.00
1.00
36.44
One SD
0.02
0.07
0.54
SD/Avg.
0.02
0.07
0.15
'Consistent nominal values
although this value was exclusive of the variations expected
from the TEI Model 146 calibrator.
NO and N02 concentration values from the two monitors
were obtained throughout the study except during calibration and
preparation of the monitoring systems and on the last three days
of the study at the Gandy site (photolytic lamp outage). The data
were post-processed to obtain the N02 values averaged over
1 minute and will be placed in the BRACE database. Examples
of the 5-minute averaged data are given in Figure 4-4 for NO
and N02 for 9-10 May and 23-24 May 2002 at the Gandy site.
Interpretation of these data and data from the Sydney monitoring
site are taking place as part of the Tampa BRACE study. In
general, the Gandy site N02 concentrations were strongly
influenced by wind direction and speed with occurrences of local
sources of automotive traffic and both local and remote point
9
-------
45
25
20
Day 1 to 30, Gandy Site, May 2002
20
15
10
5
0
-5
-5
0
5
10
15
20
Monitor #1, ppbv
1.2
1
£0.7
o.e
0.5
~ay 1-so. candy site. May 2002
1.2
0.6
0.5
Day 1-is, candy Site, May 2002
Figure 4-3. Factors related to photoiytic converter performance in the Tampa BRACE study—Gandy site.
Gandy Sit* • Ambient N02 (ppbv)
M»y 9 - May 10. 2002 (Tim* EST)
-
: J
IL 1 ill
| 1 i S 1 ! 8 1 1 1 s s 5 1 1 ! 1 I II
Gandy 9 Ho » Ambtant NO (ppbv)
May 9 - May 10, 2002 (Tim* EST)
1.. jjMnUsid^1
S 1 | 1 | 1 1 3 1 § I % 2 1 1 ! 1 1 | 5
Oandy Sit® - Ambient NQ2 (ppbv)
May 23 - May 24, 2002 (Time EST)
1 ! I S 3 S I I ! § I 1 i 1 1
Gandy 9Ha • Ambient NO (ppbv)
May 23 - May 24. 2002 (T»m* EST)
8
1
1 S 8 1 S S I 2 i J I I I 1
Figure 4-4. NO/NO-, monitoring sequence for two days in the Tampa BRACE study at the Gandy monitoring site.
10
-------
sources. For example, on 9-10 May the figure shows slow
variations in N02 concentrations preceding a large increase and
then a period of abrupt and frequent variations that coincided
with a change in wind direction so as to bring local traffic
emissions to the Gandy site. Comparisons of monitors using the
thermal and photolytic converters in the BRA CE study are given
in section 4.1.3.
Based on the field tests, the TEI Model 80800 photolytic
converters performed without interruption during the first 28
days of the study, providing photolytic conversion of N02 to NO
with a converter efficiency ranging from 36.3% to 42.1%. The
photolytic conversion of N02 to NO as implemented by EPA
using commercially available converters in the Tampa BRACE
study approaches the performance criteria set by the atmospheric
modelers in EPA publication EPA/600/R-01/005. Comparisons
over the entire time period and at the Sydney site are in progress.
4.1.2 Status of Laser-Based Photolytic
Converter for N02
NERL purchased the first stand-alone, laser-based photolytic
converter for use in N02 detection from Sonoma Technology
Inc. (STI). This unit uses new technology to provide an efficient
light source for photolysis of N02. The light source is a laser
diode (GaN), which has output power of approximately 30 mW
over a 3-nm wavelength interval between 390 and 400 nm
delivered as a collimated, coherent beam of light of narrow
diameter. The photolytic cell consists of four 0.6-cm-i.d. tubes
mounted in parallel and connected optically by the use of turning
prisms. Air is pulled through the tubes in sequence as light from
the laser diode is directed along each of the tubes in a long but
folded path configuration. A mirror is placed at the end of the
tube path to reflect the beam back through the cell. The
converter efficiency of this system is approximately 30% or
1%/mW for a sample airflow rate of 500 cc/min. However, in
typical chemiluminescenee monitors most often used with the
photolytic cells, the sample flow rate is usually 1-1.5 L/min. The
high flow rate minimizes the residence time in sample inlet
systems. Since the converter efficiency is inversely proportional
to flow rate, it would be reduced to 10-15% for the SIT unit if
the higher flow rate were used. Increases in the output power of
the diode laser is only a matter of time and will proportionally
increase the converter efficiency.
The disadvantages of the current system include (1) the low
converter efficiency at typical sampling rates as mentioned, and
(2) the lack of an alignment protocol for the laser beam through
the tube channels. The advantages are several: (1) heat genera-
tion is minimal so no cooling system will be required as with
broadband systems, (2) laser wavelength can be chosen to match
the optimum wavelength for photolysis, (3) component size is
reduced significantly, and (4) interferences caused by the use of
a wide bandwidth of radiation as with broadband sources will be
minimized subject to choice of optimum wavelength.
4.1.3 Comparison of N02 and NOY Monitors
Two N02 monitors and two monitors giving an upper limit to
N02 values were used at or near the Gandy site. One of the N02
monitors was the EPA system described above, and the other
was an OPSIS open-path monitoring system operated by
University of South Florida personnel. One output from the
OPSIS system was the N02 concentration averaged over a path
length of approximately 200 m (source at one end and receiver
at the other) and located approximately 100 m from the Gandy
monitoring station with end points roughly equidistant from the
Gandy site. The path length was parallel to Gandy Blvd. on the
opposite side from the station. Based on these locations, mea-
surements ofN02 were expected to agree at times and disagree
at others because of the wind direction relative to the roadway.
However, for remote sources the concentrations were expected
to be approximately equal.
The Environmental Protection Commission of Hills-
borough County (EPCHC) operated a "N02" monitor at the
Gandy monitoring station using the Federal Reference Method
(FRM) approach. The FRM approach uses a thermal converter
and a single chemiluminescenee monitor. The instrument opera-
tion is cyclic, providingNO and NOz readings every 30 seconds.
The two instruments sampled from a common manifold that ran
vertically downward through the center of the monitoring
station. The remaining monitor was located at an adjacent site
and was operated by employees of Atmospheric Research and
Analysis (ARA) Inc. In this monitor, two thermal converters
were placed externally to the station and routed to the
chemiluminescenee monitor through two separate lines, each
with a thermal converter. One thermal converter had an open line
to the ambient air, while the other had a specific in-line HN03
scrubber but was otherwise the same. The difference between
the two monitors provided the ambient HN03 concentration. For
the N02 monitor comparison, the NOY reading was available as
an upper limit.
Comparison of the point monitors and the path monitor
allows an estimate of the comparability of these systems for
monitoring NO and N02. For this purpose, superimposed re-
sponse versus time plots and scatter plots of the data are shown
in Figures 4-5 and 4-6 for NO and N02 for 9-10 May using
5-minute averaged data. Figure 4-5 (left) shows the EPCHC and
EPA point monitors and the USF path monitor for NO. Samples
for the EPA and EPCHC monitors are taken from the same
manifold and show similar results as noted for the scatter
diagram in Figure 4-5 (right). Figure 4-6 shows the three
monitors again for N02 comparisons. In this case, the scatter
diagram indicates excellent correlation but a difference in
calibration and a zero offset. Figure 4-7 shows a comparison of
the EPA sum of NO and N02 compared to the ARA NOY
results. The NOY results are consistently above the NO plus N02
results for values higher than 20 ppbv but are scattered above
and below for lower values. Three of the monitor types (all
11
-------
Gandy Site - 5 rnin NO Comparison (ppbv)
May 9 - May 10,2002 (Time EST)
20
15
-EPA NO
SCNO
USF NO
I ! 1111i ! ! I i i ! 2 S i | § 111
Gandy Site - 5 rrm EPCHC NO vs EPA NO (ppbv)
May 9 - May 10,2002 (Tame EST)
y = 0.9965X - 0.3876
R2 = 0.976
18
Figure 4-6. NO method comparisons for &-10 May 2002 at the Gandy site, left: NO vs time (EPC denotes EPCHC); right: scatter diagram of EPCHC
results vs EPA results (abscissa).
Gandy Site - 5 min N02 Comparison (ppbv)
May 9 - May 10,2002 (time EST)
60
50
40
30
20
10
0
EPA N02
EPCN02
USF N02
?.! u i n 11 ! n i i 1 i! ! 11
«» O «r- W |A ^ r- » O »- *3 <">"*?
i" N N N ril
Gandy Site - 5 miri EPCHC N02 vs EPA N02 (ppbv)
toy 9 - May 10,2002
50
40
y = 0.8373X - 0.7046
R2 = 0.9929
10 20 30 40 50
Figure 4-6. N02 method comparisons for &-10 May 2002 at the Gandy site, left N02 vs time (EPC denotes EPCHC), right scatter diagram of EPCHC
results vs EPA results (abscissa).
12
-------
Figure 4-7. N0Y ARA results and EPA (NO + NO,) results at the Gandy site, left: results vs time; right scatter diagram of ARA results vs EPA results
(abscissa).
Gandy Site - ARA NOy vs EPA (N0+N02) (ppbv)
May 9 - May 10, 2002 (Time EST)
60
50
40
30
20
10
0
-EPA "NOX*
ARA NOY
1
A -
1
W a
Y 1
— « fti
Gandy Site - 5 rrin ARA NOY vs ffA (NO+N02)
May 9-May 10,2002
y=0.9469x +1.3388
R? = 0.9101
20 30 40 50 60
except an EPCHC monitor) were present at the Sydney site.
However, there were two OPSIS paths at this site, one
perpendicular to die other and with a common receiver location
at the station. In the absence of a strong emission source nearby,
the open-path monitor was expected to show close agreement
with the EPA unit. These data are currently being prepared for
a more definitive comparison.
4.2 HCHO and H302 Microchemical,
Semi-continuous Monitors for
Use in Diagnostic Testing
The techniques used by Dasgupta and associates for HCHO
monitoring are well documented in the open literature1 and in the
operational manuals provided with prototype commercial
instruments made by Dasgupta through his company and with
the commercial instrument made by Alpha-Omega Power Tech-
nologies, Inc. However, the Hj02 monitor is a fairly new
application of microchcmistiy. The basic design of these instru-
ments and the results of recent comparisons are presented below.
These instruments are similar in design and both are based on
collection of HCHO and H202 in a flowing airsiream, mixing
with a reagent chemical that reacts with the dissolved gas to
create a derivati ve compound, exposure of the compound to light
to cause fluorescence, and detection of the fluorescence using a
photomultiplier tube.
The HCHO instrument consists of a sampling unit and an
analytical unit With reference to Figure 4-8, the sampling unit
consists of a diffusion scrubber (DS) that contains a single tube
of Nation through which water is drawn by the action of a
peristaltic pump (LP). The unit is set up for constant temperature
operation and is referred to as a thermostated Nation membrane
diffusion scrubber (NMDS). Sample air moves countercurrent
to the water flow and passes around the Nafion tube so that any
HCHO in the sample has a chance to collide with and be
retained by water held near the outer tube surface. The retained
HCHO permeates through the Nafion and into the water stream.
As water is removed from the water reservoir, air passes into the
voided area through a Carulite HCHO removal trap (T), Thus,
the water passes continuously through the Nafion tube and joins
the solutions of ammoni um acetate (B) and of 2,4-pentanedione
(PT), which is the reagent chemical. The mixture reacts as it
moves through the heated (70 °C) reaction coil (L) over a
3-minute time interval. A derivative compound, 1,4-diacetyl-
dihydrolutidine (DDL) is formed and moves to the detector,
consisting of a pencil-size liquid core waveguide (unique to
ITU) and a photomultiplier tube. Excitation of DDL occurs
using three blue LEDs as the source, and subsequent fluor-
escence of DDL is monitored by channeling fluorescence
through a fiber-optic element to the photomultiplier tube. After
passing through the excitation chamber, the liquid mixture
moves into a waste bottle. To shorten the reaction time of this
13
-------
Sample in
I
AP
\ /
Vent
Cartridge
\ /
Zero gas
TC
DBP
FM
PT
Timer
118
28
28
DS
LP
Figure 4-8. Schematic for monitor of HCHO ambient air concentrations.
system and improve the stability of response, the instrument is
operated in a cyclic manner, first with zero air and then with
sample air presented at the entrance to the Nafion tube. To
generate the zero air from the sample air after it has passed
through the Nafion tube, additional ambient air is added by a
controlled leak (N) in the exhaust line and the resulting sum of
flows is passed back through the system after HCHO is removed
(cartridge) and vented past the sample inlet. The availability of
excess zero air at the inlet is controlled (timer) so that a cycle of
operation consisting of 7 minutes for establishing a zero and
3 minutes for establishing a sample response is typical. The
resulting response to a step calibration of a HCHO instrument is
shown in Figure 4-9. The element I in Figure 4-8 is the location
of a valve that can be used to inject a liquid solution for
calibration of the detector or to analyze a liquid sample. The
element DBP is the debubble port by which methanol is injected
into the detector to dislodge air bubbles that can be formed in the
detector. The commercial unit being manufactured by Alpha-
Omega Power Technologies is shown in Figure 4-10.
4.2.1 The 1999 Nashville Summer
Field Study
The system used in Nashville was the first commercial prototype
for the HCHO unit. Deployment and operation of this unit
occurred as an EPA initiative both to provide an operational test
c
O)
<73
1.00
0.80
0.60 -
0.40
0.20
0.00 -
6.0
4.5
3.0
1.5
u
7.5
I |
I l! [|
il
II
11,1 ppbv HCHO
5.7ppbv
ill!
L
4000
16000
8000 12000
Time/S
Figure 4-9. Example of results of HCHO monitor calibration.
2000
14
-------
Figure 4-10. Picture of commercial prototype HCHO monitor
for the instrument and to provide data to the study. EPA was
assisted by BNL in the daily operation of the unit during a
portion of the study and by the Tennessee Valley Authority
(TVA) who established and provided space at the Polk Building
in downtown Nashville. The station was approximately 100 m
above the ground: sampling occurred from a 2.5-cm-diameter
Teflon tube through which a high flow rate of ambient air was
pulled from a location approximately 12 m outside die monitor
ing station. Calibration was achieved by zero air dilution of a
calibration gas from a compressed gas cylinder containing
5 ppmv of HCHO inN2. Results of the HCHO monitoring over
a multi-day period is shown in Figure 4-11. A second HCHO
monitor based on absorption spectroscopy using a tunable diode
laser was located at the Cornelia Fort airport site approximately
5 miles to the northeast of die Polk Building. Exchange and
measurement of a common HCHO cylinder standard was per-
formed and showed a difference in response of 12%. Under
certain atmospheric conditions the concentrations at the two sites
were very similar, as shown in Figure 4-12.
4.2.2 The 2001 NE~03PS Philadephia
Field Study
Semi-continuous monitoring of HCHO was performed by
TTU as a subcontractor to the EPA in-house contractor,
providing a high-quality database. As part of this study, die
HCHO measurements from the TTU research prototype and the
commercial Alpha-Omega monitor were compared as shown in
Figure 4-13. The local Philadelphia monitoring group (arranged
by Ted Erdman, Region HI) also provided measurements of
HCHO using die DNPH time-integrated sampling method.
Results of these comparisons are shown in Figure 4-14.
Additional comparisons between the two methods are planned
as part of studies in Seattle, WA, during the summer of 2003.
•
1
ir.
y
'~
£
a
i
i
/
fi.
&
> '<¦
. J !
f
X
r
L
Spi
x
k
- A
/
K
J
'
V
V
J
~
V
V
1
r ™
179 179J3 179.3 179.79 180 1B0.29 1SO.I 180.7S 1B1 111-23 181.8 181.79 182 182JB 1*2.5 182.75 183 183.28 183.S 183.7S 184
Julian Day, 28 Juris - 2 July 1999
14
12
m 10
8
6
4
2
0
18U 184.75 188 18&S 1*5-5 1*5.7*
18* 186.39 1M-5 188.79 187 187JS
Julian Day, 3-7 July 1999
r.S 187.79 188 188J9 188.9 198.78
184
Figure 4-11. HCHO monitoring sequence from 1339 Nashville field study.
15
-------
10
8
6
~ TDLatCF
Meth. at Polk
2§P
184
184.25 184.5 184.75
Julian Day 184, July 3,1999
185
Figure 4-12. Comparison of HCHO monitoring results in same-day sequence from 1999 Nashville field
study—Polk Building and Cornelia Fort (CF). (TDL - tunable diode laser; Meth, - methanatyzer)
R-square = 0.846 # pa = 446$
V = 0.755 + 1.08X
TTU Research Prototype Response, ppbv
Figure 4-13. Comparison of results for research prototype and commercial versions of a HCHO
monitor.
16
-------
25
> 20
0)
(0
15
R 10
8
cr
Me th analyzer
DNPH
1 2 3 4 5 6 7
Comparison Period
Methanalyzer
State DNPH
Difference
1/2 Sum
% Difference
4.14
4.19
-0.05
4.165
-1,2004802
5.24
3.48
1.76
4.36
40.366972
4.18
3.23
0.95
3.705
25.641026
3.31
3.11
0.20
3.21
6.2305296
4.08
3.S7
0.51
3.825
13.333333
19.73
19.21
0.52
19.47
2,6707756
19.72
16.93
2.79
18.325
15.225102
Average % Difference:
14.609608
Figure 4-14. Comparison of results of research prototype HCHO monitor arid the DNPH
procedure for HCHO.
4.3 Instrumentation for H202 and
MHP Monitoring
As mentioned earlier, the equipment and instrument design
required for measurement of H202 and MHP are obtained by
simple modification of the HCHO instrument. Different reagent
compounds and different components for the light source,
filtering, and detection are required, but the other parts of the
instrument are essentially identical. At this point, no one has
performed comparison studies for Hj02 and MHP, although
these gases have been measured by Dasgupta and his group in
at least two studies. Although comparisons were anticipated for
2002, none are available at this point. H202 and MHP diurnal
variations are very similar to those for HCHO. A commercial
prototype system is being manufactured by Alpha-Omega Power
Technologies, Inc., and is expected to be very similar in function
and appearance to the HCHO instrument shown in Figure 4-10.
4.4 HN03i HONO, SQ2s NH3i and Other
Water-Soluble Gases—Semi-continuous
Monitors Based on Collection Using
Diffusion Denuders and Analyses by IC
With respect to meeting the requirements for diagnostic testing
of atmospheric models, the set of point monitors currently
available as research prototypes for HN03, HONO»,SOz< and
NHj from Dr. P.K. Dasgupta of TTU and from other university
laboratories around the country are sufficient. With proper
calibration and attention to factors indicating interferences in
measurement, research scientists and property trained operators
can produce the data needed for diagnostic testing of these
gases. However, built-in quality control measures are necessary
to avoid measurement biases due to changes in collection
efficiency and in sample accumulation prior to analysis. The
development of specific operational protocols that document the
17
-------
steps necessary to ensure the uniform quality of data using these
semi-continuous monitors is still needed.
The different monitoring approaches for HN03 andHONO
are covered in a recent EPA report,1 to which the reader is
referred. NERL has continued over several years to sponsor and
promote the techniques involving collection of the gases on a
wet-wall denuder in which water or water with additives flows
over a surface through which the ambient air sample is counter-
streaming. The water-soluble gases diffuse to the wall and go
into solution. At the end of the diffusion denuder, the water is
collected and directed to a short "concentrator" column, which
allows accumulation of the target anions/cations over an
operating cycle. After collection, the collected anions/cations are
released by a counter-streaming liquid mixture that displaces the
anions/cations and carries them through an interface to the IC
column and electro-conductivity detector. Two "concentrator"
columns are used so that one can be collecting while the other is
being analyzed. A schematic of a typical experimental arrange-
ment is shown in Figure 4-15, and the result of a typical analysis
is shown in Figure 4-16. As noted in Figure 4-15, a 0.5 mM
solution of H202 is pushed to the top of the parallel plate
diffusion denuder (PPDD) and wets the inner glazed surface of
the parallel plates as the water moves under gravity to be
collected at the bottom of the PPDD. This occurs continuously
as sample air enters at the bottom of the PPDD and is pulled
between the plates to the top of the PPDD by a pump (P) and
through the mass flow controller (MFC). Soluble gases diffuse
to the walls, go into solution, and then move with the water
through the peristaltic pump (PP), through one of the anion
concentrators (TAC-LP1), to a waste line (W). The second
anion concentrator is part of the stream of water moving
through the isocratic pump (IS25), through the eluent generator
(EG40), through the guard and separator columns, through the
self regenerating suppress (SRS), and finally to the conduc-
tivity detector (CD25). The anion concentrator columns switch
after 15 minutes typically.
NERL has recently sponsored activities including the
documentation of a method for calibrating HONO monitors (see
Appendix A) and additional field monitoring of gases in the
2001 Philadelphia NE-03PS field study. Data from the EPA-
sponsored group (TTU) using semi-continuous monitoring for
S02, HNOj, and HONO as described above and the time
integrative method of HSPH have been compared by EPA.
However, results are still preliminary subject to review by the
two groups. Sample comparisons, however, are shown in Appen-
dix B with the permission of the two groups.
H20
H202
0.5mM
IS25
EG40
MFC
SRS
CD25
AS11-HC
PPDD
TAC-LP1 AG11-HC
PP
LC30
Figure 4-15. Schematic of acid gas monitor.
18
-------
20.00
15.00
CL
^ 10.00 -
CM
o
o
CL
CL
CO
n
N
5.00
T3
o
CM
10X
0.00
¦5,00
4.00
6.00
8.00
10.00
2.00
Time, min
Figure 4-16. Example anion analysis by ion chromatography for acid gases.
Also in the NE-03PS study, ammonia was measured for the
first time as a supplement to the analysis procedure described
above for anions. The basic technique begins with the procedure
for monitoring anions originating as gases as noted above. After
the anions are detected, cations can be extracted from the
concentrator cartridge and analyzed for ammonium using a
permeable membrane and electrochemical detection (see Figure
4-17). Because of the uniqueness of the ammonium as a cation
in the ambient air, this process has minimal interferences from
other anions. The alternative is analysis of the liquid sample on
a separate cation IC column. Commercial interest in providing
a semi-continuous monitor for acid and basic gases is evident,
and one company, URG, Inc. has a prototype unit that uses
components similar to those shown in Figure 4-15.
4.5 Semi-continuous Speciation Monitors
for Water-Soluble Components of
Ambient Particulate Matter
A review of techniques for nitrate monitoring is available in a
recent journal article by EPA.5 As mentioned in its introduction,
the monitoring approach using a diffusion scrubber to remove
gases prior to particle capture and subsequent separation by IC
is one of the two most prevalent approaches and is used by a
number of laboratories. For example, this class of monitor has
gone through several iterations by Dasgupta and his associates
at TTU, including the use of steam to grow sampled particles for
later impaction,6 the use of a dual filter approach for alternating
sampling and sample extraction,7 and finally the use of a mist
chamber to capture particles.8 Although the method of particle
capture can be different, the use of column concentrators for the
target compounds followed by IC analysis has been a common
component. A schematic of a particle speciation monitor of the
latter type is shown in Figure 4-17, and a typical analysis is
shown in Figure 4-18. In the schematic, the IC unit in the lower
right hand corner is identical to the gas monitor. Air is pulled
through the PPDD as before, although in this case the soluble
gases are simply removed from the PPDD. Particles move into
the particle collector (PC) and mix with a water mist generated
at the center of the chamber. Anions are dissolved in the mist,
which falls to the bottom of the chamber and is pumped away by
a peristaltic pump, through a cation concentrator column (CC),
and then into the anion concentrator column. The anions are
analyzed as in the gas monitor. The cations are analyzed by an
independent system that uses a porous membrane device (PMD)
to allow ammonium to move through the membrane to a con-
ductivity detector (CD25).
19
-------
H202
0.5mM
EG40
PPDD
NaOH
TAC-LPl
AS11-HC
AGIl-HC
H20
Figure 4-17. Schematic of water-soluble particle speciation monitor.
3.00
2.00 -
E
.o
CO
-------
The 1999 SOS summer study in Atlanta is expected to
yield a definitive database for the comparison of instruments of
this type and of the thermal desorption type based on instrumen-
tation available in 1999. The dual-filter approach was used by
TTU during the Atlanta SOS study. With the emphasis on semi-
continuous monitoring of particle components, this work was
continued to compare techniques using different principles of
operation. Dasgupta and his associates were funded to develop
a database for nitrate, sulfate, and ammonium in the 2001
Philadelphia NE-03PS study using the so-called mist chamber
technique. Preliminaiy results of comparisons between the
HSPH time-integrative method and the Dasgupta method for
sulfate are shown in Appendix B.
4.6 Sample Integrity of VOCs Collected from
the Ambient Air
The composition and prevalence of VOCs in the ambient air are
important factors in the atmospheric chemistiy leading to
photochemical product formation. The accuracy and precision
with which they are known is critical to correctly model the
processes leading to ozone, nitric acid, and other photochemical
products. Artifact formation during the sampling and analysis
procedure inherent to determination of the VOCs adds un-
certainty and inaccuracy to model calculations and must be
considered in any monitoring program that tests atmospheric
models. In so much as temporaiy storage of VOCs on solid
sorbents and in canister sampling systems is typically used
during the determination of concentrations, these procedures
must be carefully examined. In the present case, artifact
formation due to the interaction of ozone with VOCs was
investigated as a follow-up to previously published work
showing artifact formation on solid sorbents.9 The artifacts were
determined in a limited comparison of sampling based on three
approaches to sampling: solid sorbents, evaculated canisters, and
the selective VOC concentrators used with autoGCs. The results
discussed in this report are taken from the text of a manuscript
accepted for publication ill the Journal of Environmental
Monitoring (JEM) titled "Ambient VOC Monitoring Using Solid
Adsorbents—Recent U.S. EPA Developments."
4.6.1 Sampling Ambient Air Spiked with
VOCs and 03—Comparison of
Adsorbent Tubes with Canisters
and an AutoGC System
Samples were taken from an ambient air manifold that was
designed to allow spiking of ambient air with ozone and selected
VOCs. Duplicate canister samples were taken and two sorbent
tubes were taken in the manner referred to as a distributed
volume pair (DVP), i.e., simultaneous sampling for a common
period but at a different sampling rate. Samples were also
analyzed on an autoGC equipped with two traps that operated
alternately either to capture sample air or to release concentrated
VOCs by thermal desorption. The VOCs released were entrained
for downstream separation by gas chromatography and then
detection by an ion trap mass spectrometer. Canisters were
analyzed on the autoGC system, and sorbent tubes were
analyzed on a separate but essentially identical GC/ion trap
system. Details of the experimental arrangement are available in
the JEM article mentioned above. Each comparison run resulted
in nine measured values of each target VOC concentration, one
from the autoGC system, four from the canisters (both canisters
were run on both autoGC traps), and four from tubes (two 1 L,
4 L DVP sets). The target compounds consisted of compounds
in the two primaiy target compound lists that are widely used,
i.e., the list of mostly chlorinated and aromatic VOCs in the EPA
Compendium Method TO-14A and the list of hydrocarbons in
the list for the Photochemical Assessment Monitoring Stations
(PAMS). A list of polarVOCs and the w-aldehydes were added
to complete the target compound list. Most of the non-polar
compounds were not appreciably affected by co-collected ozone.
The results for the hydrocarbons on the TO-14A list of
compounds are shown in Table 4-2. Results have been analyzed
for the chlorinated compounds but are not presented here; a
similar table for the PAMS list of hydrocarbons is still in
preparation. Co-collection of ozone did result in reduction of
some of the hydrocarbons shown in Table 4-2, most notably
styrene. The emphasis here is, however, on the occurrence of
significant artifacts associated with the analysis of w-aldehydes.
In the case of co-collection of n-aldehydes and ozone in
the same sample, a significant positive artifact resulted for the
autoGC, and a significant negative artifact resulted for the
canister sampling train and for the sorbent tubes. This
information is shown in Table 4-3 for all the w-aldehydes tested.
The artifacts for n-aldehydes with lower n values show lower
artifacts than octanal due to ozone (except for the very high
autoGC readings for butanal), and nonanal showed comparable
artifact formation. As noted in recently published research,9
ozone destroys octanal and other of the target w-aldehydes when
these VOCs are adsorbed on graphitic carbon solid sorbents. The
destruction occurs to an extent depending on the ozone concen-
tration. The current results verify the earlier observations and
also indicate a significant positive artifact for sampling w-alde-
hydes and ozone onto a Tenax GR sorbent.
Figure 4-19 shows the behavior of octanal as a function of
ozone concentration for each of the three methods. The ozone
values are associated with the closest multiple of 50 ppbv. Each
point is the mean of multiple runs as noted in the additional
information in Figure 4-19, and the number of data points
represented in the figure was reduced from the total number
available (those outside of a 1 a value were discarded) to form
a reduced data set. The inclusion of the total data set increased
the a values as noted in the figure.
21
-------
Table 4-2. Effect of Co-collecting Ozone on Integrity of Hydrocarbon Response (Selected Compounds)
1
2
3
4
5
6
7
8
9
Compound Names I
50 ppb 0, AutoGC Trap 1
Average (ppbv)
2.95
2.99
2.93
5.54
1.92
3.03
1.25
2.80
2.92
1
Benzene
n = 7
Standard deviation
0.37
0.24
0.39
0.77
0.31
0.43
0.31
0.52
0.41
50 ppb O, AutoGC Trap 2
Average (ppbv)
3.09
2.81
2.90
5.77
1.98
2.95
1.28
2.79
3.14
2
Toluene
n = 5
Standard deviation
0.36
0.37
0.34
0.79
0.17
0.35
0.33
0.43
0.51
3
Ethylbenzene
50 ppb O] Cans Trap 1
Average (ppbv)
2.95
2.98
2.93
5.60
1.79
2.97
1.21
2.72
2.86
4
m,p-Xylene
n = 24
Standard deviation
0.35
0.34
0.33
0.76
0.25
0.41
0.29
0.45
0.46
5
Styrene
50 ppb 03 Cans Trap 2
Average (ppbv)
3.04
2.82
2.81
5.59
1.73
2.78
1.19
2.56
2.79
6
o-Xylene
n = 24
Standard deviation
0.37
0.33
0.36
0.77
0.18
0.37
0.28
0.40
0.47
7
4-Ethyltoluene
50 ppb O, Tubes 1 L
Average (ppbv)
3.03
3.05
2.71
5.48
1.72
2.83
1.20
2.57
2.69
8
1,3,5-Trimethylbenzene
n =24
Standard deviation
0.37
0.33
0.37
0.72
0.18
0.37
0.29
0.37
0.38
9
1,2,4-Trimethylbenzene
50 ppb 0] Tubes 4 L/4
Average (ppbv)
2.90
3.04
2.77
5.75
1.82
2.87
1.29
2.69
2.88
n = 24
Standard deviation
0.30
0.30
0.34
0.72
0.13
0.34
0.30
0.36
0.38
150 ppb 0, AutoGC Trap 1
Average (ppbv)
3.18
3.35
3.13
6.24
2.20
3.27
1.49
2.97
3.35
n = 4
Standard deviation
0.16
0.24
0.09
0.16
0.09
0.15
0.08
0.11
0.16
150 ppb O, AutoGC Trap 2
Average (ppbv)
3.10
3.03
3.04
6.10
2.01
3.03
1.51
3.04
3.39
n = 6
Standard deviation
0.10
0.23
0.07
0.16
0.04
0.08
0.07
0.10
0.14
150 ppb O] Cans Trap 1
Average (ppbv)
3.09
3.11
3.02
6.05
1.60
3.10
1.37
2.64
3.03
n = 20
Standard deviation
0.16
0.20
0.17
0.24
0.45
0.19
0.15
0.26
0.34
150 ppb 03 Cans Trap 2
Average (ppbv)
3.13
3.01
2.95
5.87
1.46
2.90
1.34
2.67
2.96
n = 20
Standard deviation
0.14
0.16
0.13
0.27
0.43
0.14
0.13
0.30
0.32
150 ppb O, Tubes 1 L
Average (ppbv)
3.20
3.26
2.81
5.74
1.31
2.92
1.19
2.57
2.69
O
CM
H
c
Standard deviation
0.30
0.15
0.18
0.37
0.22
0.17
0.12
0.20
0.20
150 ppb O, Tubes 4 L/4
Average (ppbv)
3.04
3.17
2.67
5.76
0.84
2.82
1.20
2.60
2.78
n = 20
Standard deviation
0.15
0.11
0.13
0.20
0.19
0.12
0.10
0.18
0.18
300 ppb O, AutoGC Trap 1
Average (ppbv)
3.23
3.11
3.07
6.01
2.05
3.21
1.34
2.95
3.03
n = 5
Standard deviation
0.18
0.26
0.08
0.08
0.13
0.16
0.09
0.14
0.16
300 ppb O, AutoGC Trap 2
Average (ppbv)
3.23
2.98
2.98
5.95
1.99
3.00
1.38
2.76
3.08
n =4
Standard deviation
0.12
0.14
0.06
0.15
0.11
0.11
0.15
0.20
0.21
300 ppb O, Cans Trap 1
Average (ppbv)
3.06
3.01
2.93
5.57
1.12
3.04
1.11
2.39
2.43
n = 18
Standard deviation
0.17
0.32
0.18
0.37
0.31
0.21
0.16
0.36
0.37
300 ppb O, Cans Trap 2
Average (ppbv)
3.09
2.90
2.82
5.54
1.07
2.80
1.09
2.17
2.39
n = 18
Standard deviation
0.15
0.22
0.21
0.45
0.31
0.21
0.17
0.35
0.41
300 ppb O, Tubes 1 L
Average (ppbv)
3.13
3.05
2.33
4.98
0.67
2.55
0.85
2.03
2.07
n = 18
Standard deviation
0.18
0.28
0.25
0.50
0.22
0.24
0.12
0.26
0.23
300 ppb Oj Tubes 4 L/4 .
Average (ppbv)
3.03
2.84
2.09
4.77
0.45
2.34
0.80
1.94
2.04
n = 18
Standard deviation
0.09
0.19
0.22
0.38
0.17
0.19
0.12
0.24
0.23
-------
Table 4-3. Complete Data from the Methods Comparison Study for n-Aldehydes in Spiked Ambient Air for AutoGC, Canisters, and Sorbents
Butanal
Hexanal
Heptanal
Octanal
Nonanal
Aldehydes
Sum
Compound #
32
62
73
88
99
(ppbv)
50 ppb 03 AutoGC TRAP 1
n = 7
Average (ppbv)
Standard Deviation
2.58
0.58
1.86
0.12
1.60
0.15
1.77
0.11
1.42
0.24
9.23
50 ppb 03 AutoGC TRAP 2
n = 5
Average (ppbv)
Standard Deviation
2.14
0.54
1.24
0.25
1.64
0.30
2.37
0.39
1.53
0.32
8.92
50 ppb 0, Cans TRAP 1
n = 24
Average (ppbv)
Standard Deviation
2.37
1.54
1.48
0.36
0.92
0.37
0.62
0.36
0.35
0.24
5.74
50 ppb 0, Cans TRAP 2
n = 24
Average (ppbv)
Standard Deviation
1.63
1.36
1.11
0.39
0.92
0.26
0.76
0.44
0.35
0.27
4.77
50 ppb 03 Tubes 1 L
n = 24
Average (ppbv)
Standard Deviation
1.60
0.19
1.53
0.25
1.07
0.27
0.85
0.38
0.60
0.47
5.66
50 ppb 0, Tubes 4 L/4
n = 24
Average (ppbv)
Standard Deviation
1.40
0.18
1.23
0.19
0.79
0.14
0.55
0.15
0.26
0.08
4.24
150 ppb 03 AutoGC TRAP 1
n = 4
Average (ppbv)
Standard Deviation
10.03
2.07
2.11
0.14
1.81
0.14
2.11
0.26
2.45
0.37
18.50
150 ppb O, AutoGC TRAP 2
n = 6
Average (ppbv)
Standard Deviation
5.45
0.73
1.47
0.26
1.85
0.30
2.95
0.56
2.50
0.49
14.23
150 ppb 03 Cans TRAP 1
n = 20
Average (ppbv)
Standard Deviation
2.71
0.64
1.19
0.53
0.71
0.52
0.59
0.67
0.86
1.56
6.06
150 ppb 03 Cans TRAP 2
n = 20
Average (ppbv)
Standard Deviation
1.84
0.60
0.88
0.53
0.71
0.49
0.60
0.57
0.62
1.11
4.65
150 ppb O, T ubes 1 L
n = 20
Average (ppbv)
Standard Deviation
1.99
0.16
0.95
0.33
0.50
0.21
0.36
0.13
0.37
0.11
4.17
150 ppb 0, Tubes 4 L/4
n = 20
Average (ppbv)
Standard Deviation
2.10
0.36
0.51
0.21
0.27
0.11
0.20
0.08
0.15
0.07
3.23
300 ppb O, AutoGC TRAP 1
n = 5
Average (ppbv)
Standard Deviation
12.72
1.15
2.09
0.22
1.75
0.13
2.17
0.25
2.68
0.23
21.39
300 ppb Oa AutoGC TRAP 2
n = 4
Average (ppbv)
Standard Deviation
8.67
1.50
1.24
0.12
1.90
0.56
2.92
0.55
3.43
0.48
18.15
300 ppb 03 Cans TRAP 1
n = 18
Average (ppbv)
Standard Deviation
2.05
0.47
0.65
0.32
0.28
0.19
0.23
0.11
0.34
0.19
3.55
300 ppb Oa Cans TRAP 2
n = 18
Average (ppbv)
Standard Deviation
1.38
0.32
0.40
0.16
0.26
0.11
0.24
0.11
0.23
0.12
2.52
300 ppb 03 Tubes 1 L
n = 18
Average (ppbv)
Standard Deviation
2.40
0.37
0.41
0.12
0.21
0.05
0.18
0.05
0.24
0.08
3.43
300 ppb Oa Tubes 4 L/4
Average (ppbv)
2.27
0.25
0.12
0.10
0.10
2.83
to
II
c
Standard Deviation
0.36
0.05
0.02
0.03
0.04
23
-------
Octanal
>
JQ
a
a
a>
to
c
o
a
to
zc
AutoGC
2.5
1.5
0.5
SD = .30
i^7
n = 9 ./
SD = .36
SD = .22 /
3
II
\ SD= .16
n = 35
SD = .20
n = 35
n = 29
SD = .20
SD = .06
n = 30 ak _____
—WM n = 29
SD = .07
m SD = .04
50
100
150 200
Ozone(ppbv)
250
300
350
Figure 4-19. Octanal response by GC/MS as function of ozone concentration—comparison of results using canisters, sorbent
tubes, and an autoGC. The ordinate intercept value equals the concentration intentionally added to the manifold.
The loss of octanal and other w-aldehydes in canisters
shows almost the same dependence on ozone concentration as
for the solid sorbent. Even though the samples taken directly
from the manifold and those stored temporarily in the canisters
were both analyzed on the same analytical system, there were
two differences in the monitoring procedure: (1) for samples
taken in the canister, ozone was destroyed in the canister and did
not reach the concentrator on the analytical system; and (2) the
possibility existed for reactions between the ^-aldehydes and
ozone during the sampling and storage procedures associated
with the canisters. Apparently the positive artifact observed
when monitoring with the autoGC is due to ozone reaction with
the autoGC concentrator and the procedure of sampling and
storage in canisters. However, eliminating ozone, and hence
eliminating any subsequent reaction between ozone and the
autoGC concentrator, results in a negative artifact that increased
with the ozone concentration. The use of a canister sampling
train containing in-line elements, including an in-line air pump
and needle valve so that the canisters can be slightly pressurized
for venting past the inlet to the autoGC system, has complicated
investigation of the effects that are associated only with storage
in the canister. Additional experiments with a simplified system
24
-------
are needed. Hence the negative artifacts observed with the
canisters have been associated with the canister sampling train,
not just the canister.
The use of a Mn02 scrubber, under certain conditions, was
successful in reducing the ozone in the sample stream to near
zero and in reducing the artifacts otherwise observed in all the
methods. Figure 4-20 shows the results of three sequential
comparisons of the methods for octanal in which ozone
concentration was 150 ppbv. Each comparison involves the
sampling at 5 L/min over a period of 57 min through a stack of
eight Mn02-coated, 1 -inch-diameter screens placed between the
main and secondary sampling manifolds. This scrubber is similar
to the one recommended by Calogirou, et al10 for sampling
ambient air so that ozone is destroyed and the terpenes can be
detected. In the first comparison, the Mn02 scrubber was in
place and the three methods provided nearly the same results
(canister values were slightly low) at the expected ppbv level. In
the second comparison the scrubber was removed and the results
show the type of behavior expected based on the Table 4-3
results, i.e., positive artifacts for the autoGC and negative arti-
facts for the canisters and sorbent tubes. In a third comparison,
the same scrubber was used again, but the responses for all three
methods were reduced. Subsequent use of the same scrubber
gave similar results to the third comparison; however, the use of
a new scrubber resulted in a repeat of the same sequence of
results. Similar results were obtained for the other aldehydes.
These experiments imply that the ^-aldehydes are being lost on
the filter by changes that occur as the result of sampling the
spiked ambient air mixture.
In summary, ozone reacts with ^-aldehydes adsorbed on
Tenax GRio create significant positive artifacts, whereas signif-
icant negative artifacts are created on graphitic carbon increasing
with the value of the aldehyde and the ozone level. Signifi-
cant negative artifacts are also created in canister-based sam-
pling, although the mechanism is not known. The use of a
previously unused Mn02 scrubber eliminates artifact formation
in all methods, but the lifetime of the scrubber, as configured
and applied for this work, is limited with respect to passing the
w-aldehydes. Repeated use of the same scrubber resulted in low
readings for all methods. Use of a new scrubber resulted in the
same behavior.
The results of data analysis for the co-collection of
H-aldehydes and ozone and the preliminary analysis of similar
data for the co-collection of terpenoid compounds and ozone
(not included here) indicates that significant positive and
negative artifacts occur and could limit the usefulness of such
data unless ozone is selectively eliminated from the sample
stream. Further investigation is warranted to determine a means
to recondition Mn02 ozone scrubbers after initial use so that
they do not retain n-aldehydes and other compounds of interest.
C
o
<0
w
c
-------
Chapter 5
References
1. Recommended Methods for Ambient Air Monitoring of
NO, N02, NOy, and Individual NOz Species. W.A.
McClenny, ed. EPA Report EPA/600/R-01/005, September
2000.
2. Weber, R., Orsini, D., Duan, Y., Baumann, K., Kiang,
C.S., Chameides, W., Lee, N., Brechtel, F., Klotz, P.,
Jongejan, P., ten Brink, H., Slanina, J., Boring, C.B.,
Genfa, Z., Dasgupta, P., Hering, S., Stolzenburg, M.,
Dutcher, D.D., Edgerton, E., Hartsell, B., Solomon, P.,
Tanner, R. Intercomparison of Near Real-Time Monitors
of PM2 5 Nitrate and Sulfate at the EPA Atlanta Supersite,
accepted for publication in J. Geophys.
Res.—Atmospheres.
3. Ryerson, T.B., Williams, E.J., Fehsenfeld, F.C. 2000. An
Efficient Protolysis System for Fast-Response N02
Measurements. J. Geophys. Res. 105:26447-26461.
4. Fan, Q., Dasgupta, P.K. 1994. Continuous Automated
Determination of Atmospheric Formaldehyde at the Parts
Per Trillion Level. Anal. Chem. 66:551-556.
5. W. A. McClenny, W.A., Williams, E.J., Cohen, R.C., and
Stutz, J. 2002. Preparing to Measure the Effects of the
NOx SIP Call - Methods for Ambient Air Monitoring of
NO, N02, NOy, and Individual NOz Species, J.Air &
Waste Manage.Assoc. 52, 542-562.
6. Simon, P.K., and Dasgupta, P.K. 1995. Continuous
Automated Measurement of the Soluble Fraction of
Atmospheric Particulate Matter, Anal. Chem. 67, 71-78.
7. Boring, C.B., Al-Horr, R., Genfa, A., Dasgupta, P.K.,
Martin, M.W., and Smith, W.F. 2002. Field Measurement
of Acid Gases and Soluble Anions in Atmospheric
Particulate Matter Using a Parallel Plate Wet Denuder and
an Alternating Filter-Based Automated Analytical System,
Anal. Chem. 74, 1256-1268.
8. Dasgupta, P.K., and Al-Horr, R., Texas Tech University,
Lubbock, TX, Personal Communication.
9. McClenny, W.A., Colon, W., Oliver, K.D. 2001. Ozone
Reaction with ^-Aldehydes (w=4-10), Benzaldehyde,
Ethanol, Isopropanol, and n-Propanol Adsorbed on a
Graphitic Carbon Solid Adsorbent, J. ChromatogrA., 929,
89-100.
10. Calogirou, A., Larsen, B.R., Brussol, C., Duane, M.,
Dotzias, D. 1996. Decomposition of Terpenes by Ozone
During Sampling on Tenax, Anal. Chem. 68, 1499-1506.
26
-------
Appendix A
Development of a Calibration Device for HONO
by
Keith G. Kronmiller and Dennis D. Williams
ManTech Environmental Technology, Inc.
A.1 Introduction
This report details the design and evaluation of a nitrous acid
(HONO) generator capable of producing concentrations such as
would be found in the air at ground level (0 to 10 parts per
billion by volume [ppbv]). This generator will be used to
develop methods for generating known HONO mixing ratios for
calibrating and evaluating a new nitrous acid continuous
analyzer that the U.S. Environmental Protection Agency (EPA)
acquired during fiscal year 2001.
After searching the literature for techniques and methods
to produce HONO, one paper in particular was found to clearly
describe and show potential for our intended purpose. This
paper, by A. Febo et al., was found in the journal Environmental
Science and Technology (1995, 29:2390-2395). We contacted
the scientists who performed the original work and they provided
assistance, clarifications, and technical support, which made
reproducing the generator an easier task.
The generation system was built and assembled from
various components available in any laboratory. All components
were assembled inside an aluminum box fabricated by EPA's
machine shop personnel. Much of the evaluation was performed
with assistance from the ManTech-staffed ion chromatography
(IC) laboratory.
A.2 Experimental Design
The HONO generator consists of a reverse permeation device
(see below) for releasing hydrogen chloride (HC1) into an
airstream that then passes to an in-line container of sodium
nitrite (NaN02). Nitrous acid is produced when HC1 combines
with NaN02 according to the following reaction:
HC1 + NaN02 - NaCl + HONO (1)
We chose to incorporate the reverse permeation device described
by Febo and colleagues because it eliminates several problems
inherent in other gas collection methods as described below.
HC1 vapor can be obtained by various means. The easiest
method is to collect the evaporated gas over a container of
hydrochloric acid. However, because hydrochloric acid has a
vapor pressure greater than 600 mm Hg at room temperature, the
concentrations produced in this manner are too large for an
ambient-level generator.
A suitable alternative is to release HC1 by permeation
through Teflon tubing. A typical permeation tube consists of a
length of Teflon tubing with plugged ends containing an aqueous
solution of the substance. Partial pressure differences between
inside and outside provide a continuous flow of the substance
through the Teflon walls. The permeation rate through Teflon-
walled tubing is determined by the density of the Teflon, the
wall thickness, the length of wall in contact with the substance,
and the temperature at which it is held. Over time the mass loss
of solute inside the closed-tube permeation system results in total
evaporation and therefore ends the tube's useful life.
The paper published by Febo and colleagues describes a
reverse permeation system that can control all of the afore-
mentioned factors that determine permeation rate. In reverse
permeation, Teflon tubing is placed in a vessel containing an
aqueous solution of HC1. The HC1 permeates into the tubing as
a constant flow of air (or Nj) is passing through the tube. This
type of system has two distinct advantages: First, the quantity of
hydrochloric acid contained inside the cell is so large there are
no end-of-life worries. Second, because the HC1 liquid is on the
outside of the Teflon tubing, it can be easily changed for more
or less dilute concentrations, which directly affects the overall
HC1 output. Thus, the capabilities of this system are ideally
suited for our purposes, and we incorporated this type of cell
into our HONO generator.
27
-------
The HONO generator was developed in two parts. First, the
HC1 reverse permeation device was assembled and tested alone.
Figure A-1 shows the reverse permeation cell, which consists of
a 500-mL Pyrex glass bottle with two 1/16-in. stainless steel
Swagelok bulkhead fittings mounted through the plastic cap.
FEP Teflon tubing is attached to the inside of each bulkhead
fitting and coiled inside the glass cell (1/16-in. o.d., 1/32-in. i.d.,
1/64-in. wall, and 210-cm length). The glass bottle is wrapped
with electrical tape to control the operational temperature.
HC1 IN
HONO OUT
Nitrogen Gas IN
HC1 in N2 OUT
Heater
9MHC1
Heater
NaN02
Magnetic Stirrer
Stir Bar
Figure A-2. Schematic drawing of the NaN02 cell.
Figure A-1. Schematic drawing of reverse permeation cell.
We tested the cell to determine its permeation character-
istics using various flow rates and temperatures. Data from these
tests are presented in the Experimental Results section.
After the reverse permeation cell was found to generate
sufficient quantities of HC1 vapor, the NaN02 cell was con-
nected and the combined apparatus was tested. Figure A-2
shows the NaN02 cell and illustrates how the reaction between
the HC1 vapor and granular sodium nitrite is produced by intro-
ducing HC1 gas through small-diameter Teflon tubing with its
outlet located in the sodium nitrite (contained inside a gas-tight
vial). In addition, the NaNQ2 granules are continuously stirred
with a small magnetic stirring bar. This helps to maintain a
constant surface area without which the reaction rate becomes
unstable. Finally, the reaction vial is surrounded by electrical
heating tape for temperature control.
Tests were performed using the HC1 cell along with the
NaN02 cell to determine the HONO output characteristics. The
Experimental Results section contains data from these tests.
The system components were then assembled inside an
aluminum case fabricated by the in-house machine shop (see
Figure A-3). A hinged front door provides access into the
interior where the various components are securely mounted.
Two digital temperature controllers mounted on the front door
are connected to the two heated components (HC1 reverse
permeation cell and the NaN02 cell).
Refinements were made during development. The Febo
paper described enhancement of HONO production by using
humidified carrier gas (UHP N2). Our tests verified that this
method did enhance HONO production, so we included a beaker
of deionized water (DIH20) with a bypass control valve in the
design. The N2 carrier gas is split after a flow-controlling
rotameter such that a portion of the gas is directed over the DI
H20. The bypass control valve provides a means to adjust the
overall relative humidity (RH) from 0% to 90% (maximum).
Figure A-4 shows the completed HONO generation device.
All components are commercially available except the ManTech-
designed RH sensor. (Note: For information on the RH device,
contact the author.) Tubing and connectors were either Teflon or
stainless steel.
28
-------
HCI perm temp control Humidity disomy NaNQa cell temp control
Humidity control
M? intet
tfeflow rotometer
Figure A-3. HONO generation system front panel components.
Humidity Sensor
N2In
«l
HCI Cell
Rotameter | \
DIH20
HONO Out
I—~
NaN02Cell
A
Mag. Stirrer |
Figure A-4. Schematic diagram of HONO generator.
Figure A-5 shows the calibration and testing apparatus for
determining the HONO output concentration over various
conditions. A strip chart recorder was attached to the chemi-
luminescence monitor to record the generator's performance
over time. The reference chemiluminescence analyser was a TEI
Model 42S (Thermo Environmental Instruments, Inc., Franklin,
MA), which was calibrated using a TEI Model 146 calibration/
dilution system with NIST-traceable NO span gas and a certified
N02 permeation tube. All flows were measured using a certified
DryCal DC-2 flow meter (BIOS International, Butler, NJ).
Dilution air was supplied from an in-house clean air system to
make up flow volumes needed by the analyzer and annular
denuders. Although the HONO generator was located over 5 m
from the analyzers, tests verified that there was no appreciable
loss of HONO.
DIONEX IC
Analysis
N02 and CI
Anions
Annular
Denuder
Strip Chart
Recorder
TEI 42S
Analyzer
System
TEI 146
Calibrator
N02 Perm
MO Span
Gas
Figure A-5. HONO system test and calibration apparatus.
A.3 Experimental Results
As mentioned earlier, the system was developed in two stages.
The HCI reverse permeation cell was tested first in order to
determine the efficiency of the selected tubing, HCI con-
centration, and temperature. These tests were performed inside
the wet chemical laboratory operated by ManTech personnel, DI
H20 (20 mL) was used inside a standard impinger to collect the
HCI vapor. This was then followed by IC analysis using the
Dionex Model DX500 for chloride and nitrite ions. We
performed a total of 44 analytical tests, which are tabulated in
Table A-l. Thirteen of these were made using two impingers
connected in series to verify 100% collection efficiency within
the first stage. Chloride anion was never found within the second
impinger's DI H20, therefore verifying complete recovery inside
the first stage. The average chloride ion concentration converted
to permeation rate was 120 ng min'1 with a standard deviation of
±34 ng min*'. The resolution of the Dionex system (0.01 |ig
mL4) is equivalent to a permeation resolution of 13 ng min
These tests confirmed the reverse permeation system generated
HCI with title desired rate for producing HONO at ambient levels
as shown by the following calculations:
29
-------
Assume: HONO concentration range of 0 to 25 ppbv
HONO flow rate set constant = 500 cm3 min"1
mol wt of HONO = 47.013 g
Molar constant (Km) for HONO = 24.46 L/47.013 g
= 0.520 L g"1 (STP)
K,. = 0.555 Lg"1 (at 45 °C)
Dilution airflow (variable) minimum =
2000 cm3 min"1
Calculate: (1) Desired HONO concentration from HONO
generator before dilution
Cone (HONO output) = (HONO cone)
(FIoWhono + Flowdilution))/Flow
HONO
= (0.025)(500 + 2000)/500
= 0.125 ppmv
(2) Convert ppbv to permeation rate equivalent
Permeation rate (ng min"1) = (Cone x
FIoWhonoVKtii
= (0.125 x 500)/0.555
= 113 ng min'1
Conclusion: Assuming HC1 and NaN02 reaction efficiency
of 100%, then the measured HC1 rate found by
the IC analysis can be expected to generate the
desired HONO concentration.
Measured avg. HC1 and calculated HC1 concen-
tration are approximately equal.
Next, the NaN02 cell was connected and heated to 45 °C,
and the magnetic stirrer was used to agitate the NaN02.
Additional tests were performed inside the IC laboratory to
verify the HONO generator was operational. To test for HONO,
an annular denuder coated with sodium carbonate (1% Na2C03
and 1 % glycerol in methanol) was attached to the HONO outlet
port and allowed to collect sample. At the end of sampling, the
denuder contents were extracted and subsequently analyzed by
using the Dionex Model DX500 IC. The results of these tests are
shown in Table A-2.
The testing showed that there was some variation in the
HONO output (as seen by the nitrite ion analysis numbers). This
variation was considered and evaluated. Samples 25 and 26 were
higher than the others. This was due to the carrier gas (N2) run-
ning out overnight, which caused the buildup of HC1 inside the
permeation cell tubing. When the carrier gas was reconnected,
this high concentration of HC1 was combined with the NaN02
during subsequent sampling. Other than these two samples, the
output rate ofHONO was fairly consistent at about 125 ng min"1.
After the IC tests were completed, the system was installed
inside laboratory S-213 in the ERC Annex building and
connected to the air dilution system and sampling manifold. A
strip chart recorder was attached to the NOx instrument's N02
channel for observing such characteristics as long-term stability,
recovery time after temperature change, and output HONO
concentration range versus HC1 cell temperature.
Long-term stability. The HONO generator was set up and
operated for more than 80 h while we observed the NOx
instrument's N02 channel. An electrical timer controlled a
Teflon solenoid valve that switched the HONO pollutant gas into
the glass mixing manifold every 0.5 h. The data showed very
little variation during this time:
HC1 cell temp = 40.0 °C
NaN02 cell temp = 45 °C
N2 carrier flow = 290 cm3 min"1
Zero air dilution flow = 8020 cm3 min"1
Start test (9/23 08:40)
NO 0.12 ppb
N02 8.22 ppb
NOx 8.35 ppb
End test (9/26 14:00)
NO 0.50 ppb
N02 8.34 ppb
NOx 8.84 ppb
Calculated difference = (8.34-8.22)/8.22 = 1.2%
Recovery time after temperature change. As the HC1 cell
temperature can be used to adjust the range ofHONO concen-
trations, it is important to understand how long the system takes
to stabilize after changing the temperature. This recovery time
can be readily measured by recording the HONO concentration
(as measured by the NOx instrument's N02 channel) on the strip
chart recorder.
Example:
(!) TNaN02= 45 °C, FN2 = 290 mL min"1, RH = 47%, FlowZA =
2090 mL min"1
HONO stable 9/20 07:30 THC, = 40 °C
NO 0.73 ppb
N02 30.27 ppb
NOx 31.04 ppb
Change temperature of HC1 cell to 35 °C @ 07:35
HONO stable 9/20 12:30 THC1 = 35 °C
NO 0.25 ppb
N0219.13 ppb
NOx 19.38 ppb
Time to reach new stable HONO output = 5 h (approx.)
(2) TNaN02 = 45 °C, FN2 = 290 mL min"1, RH = 47%, FlowZA =
2090 mL min"'
HONO stable 9/20 13:00 THCI = 35 °C
NO 0.25 ppb
N0219.13 ppb
NOx 19.38 ppb
Change temperature of HC1 cell to 30 °C @ 13:05
HONO stable 9/20 18:00 THC1 = 30 °C
NO 0.59 ppb
N02 11.33 ppb
NOx 11.99 ppb
Time to reach new stable HONO output = 5 h (approx.)
30
-------
Q.
CO
O
CN1
X
Q
d>
c
o
D
(0
d>
q:
<
o
XI
(0
o
00 CO
00 CO
00 CO
00 O)
^ T-
h-
(O CO N
CO
O)
co co r-
COOOOCDOOOCDOOO
N M ^ r O (D
CO
^ CM
00 CD
o o o
CO
(N O if) O
CO
O "
o csi o
CO
O
OOOOOCOOCOOCOOCOCOCOO
CMCO^-CDOCMOCMOCMCOCOCOCM
00^00^1^00000000 CO
O^O^CMCsiCsi^T-' t-'^t-^CMCMCM^
oi ^
T- CO
H
" >
D)
-------
Table A-2. Results from IC Analysis of Na?C03/Glycerine-Coated Denuder Tests
CI ion
CI ion
CI equiv
N02 ion
N02 ion
NOz equiv
ID
(pg/mL)
(ug/10 mL)
(ng/min)
(ug/mL)
(Mg/10 mL)
(ng/min)
Na2C03 B
0.02
1 HONO
0.09
0.7
46.6
0.19
1.9
63.33
2 HONO
0.08
0.6
40
0.24
2.4
80
3 HONO
0.04
0.2
13.3
0.23
2.3
76.7
Na2C03 B
0.1
4 HONO
0.08
0.8
53.3
0.67
5.7
335.2
5 HONO
0.05
0.5
33.3
0.43
3.3
220
6 HONO
0.07
0.7
46.6
0.33
2.3
153.3
Na2C03 B
0.03
7 HONO
0.07
0.4
26.6
0.41
4.1
273.3
8 HONO
0.02
0.23
2.3
153.3
9 HONO
0.01
0.2
2
133.3
Na2C03 B
0.03
10A HONO
0
0
0.3
3
200
10B HONO
0
0
0
11AHONO
0.02
0
0.31
3.1
206.6
11B HONO
0
0'
0
12AHONO
0
0
0.27
2.7
180
12B HONO
0.02
0
0
13 HONO
0.02
0
0.24
2.4
160
14 HONO
0.03
0
1.45
14.5
241.6
15 HONO
0.05
0.2
3.33
1.11
11.1
185
16 HONO
0
0
0.85
8.5
141.6
Na2C03 B
0.03
0
17 HONO
0.04
0.1
3.33
0.59
5.9
196.6
18 HONO
0.02
0
0.45
4.5
150
19 HONO
0.02
0
0.4
4
133.3
20 HONO
0.02
0
0.55
5.5
183.3
21 HONO
0.03
0
0.42
4.2
140
22 HONO
0
0
0.47
4.7
156.6
23 HONO
0.02
0
0.37
3.7
123.3
Na2C03 B
0.08
24 HONO
0
0.26
1.8
120
25 HONO
0.02
1.06
9.8
653.3
26 HONO
0
0.58
5
333.3
27 HONO
0.01
0.16
0.8
53.3
28 HONO
0
<
0.15
0.7
46.7
29 HONO
0.03
0.14
0.6
40
30 HONO
0.02
0.16
0.8
53.3
32
-------
Output HONO concentration range versus HCl cell
temperature. Table A-3 shows data obtained by the TEI 42S
oxides of nitrogen monitor with changing HCl cell temperatures
and dilution flow rates.
Output Concentration versus Relative Humidity. Table A-4
shows data collected to characterize how humidified carrier gas
(N2) affects the HONO output concentration (as measured by the
chemiluminescence monitor). Humidifying the carrier gas (to
40% RH) resulted in an average increase of 75% over the same
temperatures and flows shown in Table A-4.
Table A-3. Range of HONO Concentrations versus HCl Cell Temperature
HCl Temp
°C
ZA Flow
"mL/min
NO
ppb
no2
PPb
NOx
PPb
30
2090
0.59
11.33
11.99
30
8020
0.36
3.22
3.57
30
17750
1.05
1.80
2.84
35
2090
1.00
19.20
20.20
35
8020
0.55
4.98
5.53
35
17750
0.29
2.61
2.91
40
2090
0.73
30.27
31.04
40
8020
0.12
8.22
8.35
40
17750
0.41
3.96
4.37
45
2090
0.35
49.68
50.00*
45
8020
0.20
13.63
13.82
45
17750
0.49
6.29
6.77
*DI H20 ran dry.
Note: The NaN02 cell temperature was constant at 45
°C, and the N2
flow was set to 290 mL min"1
(nominally).
Table A-4. HONO Output Concentration versus Carrier Gas Humidity (average 75% increase for relative humidity change from 0%
to 40% RH)
HCl Temp ZA Flow RH=40% RH=0% % HONO Increase
X mL/min HONO (ppb) HONO (ppb) with 40% RH N,
30 2090 11.33 6.34 78.71
30 8020 3.22 1.82 76.92
30 17750 1.8 0.82 119.51
35 2090 19.01 10.87 74.89
35 8020 4.98 3.14 58.60
35 17750 2.61 1.57 66.24
40 2090 30.27 16.23 86.51
40 8020 8.22 4.75 73.05
40 17750 3.96 2.13 85.92
45 2090 49.68 28.61 73.65
45 8020 13.63 7.94 71.66
45 17750 6.29 3.77 66.84
Avg = 77.71
STDev= 15.35
33
-------
A.4 Conclusions and Recommendations
The HONO generator constructed and tested as described
in this paper performs well. Overall the system is stable and can
be used for its intended purpose with confidence. The HONO
output has very few impurities such as nitric acid (HN03) or
N02. The system can be used to produce HONO concentrations
from very close to zero through SO ppb by making simple
adjustments to the HC1 cell temperature and/or to die dilution
air, which is combined with the N2 carrier flow from the unit
Figure A-6 shows the system performance over HC1 cell
temperature at various dilution flow rates.
solution should produce only half as much HC1 vapor at similar
cell temperatures. Another method would be to remove the HC1
cell, empty the contents, and then shorten die 1/16-in. Teflon
tubing loop inside die cell. A shorter permeation length would
also result in lower HC1 concentrations
Finally, it is recommended that the HONO generator be
used with care and only by operators familiar with laboratory
safety practices. The HC1 contained inside die unit is inside a
glass bottle subject to breakage if the device is dropped or turned
over accidently. Use utmost care when transporting the device,
preferably removing the HC1 cell altogether before moving it to
another site. Use gloves and face mask when near the unit and
always operate it in a hood
2.09 t_/mIn
*.02 L/mln
17.78 Li win
Figure A-6. HONO generator output vs. HCI temperature at
various dilution airflow rates. Minimum concentration is 1.80 ppb
and the maximum concentration is 49.68 ppb.
While tests were being performed, certain observations in
the use of the system were noted. One is that the DI H20
humidifier has a limited volume and is therefore subject to
becoming depleted at some time. In our experience, die water
should be replenished every 2 weeks of continuous use. Another
observation is that if the HCI cell's temperature is changed to
obtain a different range of concentrations, at least 5 h are
required before the output becomes stable. As seen by the data
in Table A-3 and Figure A-6, the lowest concentration we
produced was about 2 ppb using a HCI cell tempe rature of 30 °C
and a dilution flow rate of greater than 17 L min"1. If it is
necessary to produce lower concentrations, two means exist The
simplest would be to remove the HCI cell, empty the 9 molar
(M) HCI, and refill it with a more dilute concentration. A 4.5M
34
-------
Appendix B
Measurement of Inorganic Gases and Particulate Matter in
Philadelphia NE-03PS 2001
Because comparisons involving this data are still being
processed,, only a preliminary look at the data comparisons will
be considered in this appendix. Final conclusions must still be
developed. However, it is important to note the status of this
comparison in that it involves two significantly different
methods for sampling target compounds, one a semi-continuous
method and the other a time-integrated one.
The Philadelphia Northeast Ozone, Particulate Study (NE-
03PS) summer field study of 2001 provided an opportunity to
compare measurement techniques for the inorganic gases and
particulate matter. The Harvard School of Public Health (HSPH)
and the Texas Tech University (TTU) (with NERL, EPA
sponsorship) monitoredHONO, HN03, S02, fine particle nitrate
and fine particle sulfate at adjacent locations. HSPH used
day/night collection periods of 10 hours (1000 - 2000 hrs and
2200 - 0800 hrs) for accumulation of samples on the Harvard
EPA denuder system (HEADS) sampling system1 followed by
shipment of the sampler to the HSPH laboratory for analysis.
The TTU group used semi-continuous monitoring (every 15
minutes) of gases (one system) and of particles (second system)
at the site. Both groups made their data available to NERL for
comparison. The sampling occurred from 1 July until 30 July
and included 60 day/night sequences. Measurements from the
TTU instruments were averaged over each of the HSPH
sampling periods and then compared to the results forwarded by
the HSPH. The comparisons are summarized in Table B-l as
summary values for a number of interesting parameters describ-
ing the data and the comparison.
The linear regression parameters for the data set indicates
that the TTU values can be best represented by a regression
slope of 0.41, 0.40, and 0.40 for HONO, HN03 and nitrate,
respectively and that the regression slopes for S02 and sulfate
are 0.56 and 0.54, respectively.
R-square values are 0.73, 0.95, 0.94, 0.92, and 0.89 for
HONO, HN03, S02, nitrate, and sulfate, respectively. HONO
has a ordinate intercept that at 0.37 is more than half of the
average value, a fact that may explain the relatively low R-
square value.
Based on this preliminary look at the data comparison, the
NERL is corresponding with HSPH and TTU to resolve what
appears to be common factor errors, one for the nitrogen-
containing compounds and a different one for the sulfur-
containing compounds. For example, the low value of the
common slope for the nitrogen-containing compounds is likely
to be an error in measurement of or calculations for the
calibration standard used for the analytical system(s), or possibly
a flow measurement error for one or both of the groups.
The error occurs throughout the study since the difference
in slope applies over portions of the data set as well as to the
overall data set. Since a number of calibrations and flow checks
were made during the tests, it is difficult to see how the error
persisted if the error were due to calibration or flow rate
problems. Note that the day/night ratios of the nitrate and sulfate
are almost the same for nitrate and sulfate. This fact implies the
differences are due to multiplicative factors.
At this point in evaluating the comparison, these facts are
being discussed with the two laboratories in an attempt to find
the missing factors. Assuming that these factors will be evident
in the future, the resulting comparisons would look like the
Figure B-l. Comparisons of the TTU values with a real-time
instrument operated at the Philadelphia by Brookhaven National
Laboratory (BNL) for HN03 may help to resolve the current
comparison differences. However, this data has not been
processed yet.
1. Koutrakis, P., Wolfson, J.M., Slater, J.L., Brauer, M.,
Spengler, J.D., Stevens, R.K., Stone, C.L. 1988. Evaluation
of an annular denuder/filter pack system to collect acidic
aerosols and gases, Environ. Sci. Technol. 22,1463-1468.
Note: One of the multiplication errors referred to above has
been identified (during publication of this report) as an error in
the measurement of a sample loop used for calibration.
35
-------
Table B-1 Prfeiirointuy Comparisons between HSPH
and TTU in NEOaPS *01 Philadelphia Fiaid Study
NEOjPS '01
Philadelphia
Data Summary
HONO
ppbv
HNO,
ppbv
SO,
ppbv
Nitrate
nmoies/m5
Sulfate
nrnoles.'rrr5
HSPH
TTU
HSPH
TTU
HSPH
TTU
HSPH TTU
HSPH TTU
Ave Cone.
0.66
0.64
1.28
0.51
3.18
1.55
9.84 3.65
50.8 34.0
Ratio HSPHTTU
1.03
2.51
2.05
2.7
1.49
Linear Reg.
TTU = 0.41'HSPH
+ 0.37
TTU - 0.40'HSPH
+ 0.00
TTU = 0.56*HSPH
-0.18
TTU = 0.40*HSPH
-0.27
TTU = 0.54*HSPH
+ 6.76
R-square
0.73
0.95
0.94
0.92
0.89
Ave Ratio, Std HSPH/TTU
1.69,1.05
2.67, 0.6S
2.25, 0.54
3.04,1.07
1.38.0.45
Night/Day Ratio
4.S3
1.73
0.25
0.33
0.79
0.92
2.37 2.57
0.68 0.69
Comparison of Day/Night Averages for HN03
7
6
tvi 4
3
2
Both Average 1.28
R-square = 0.953 # pts = 60
y = 0.00426 + 0.993x
1
0
0
3
4
5
6
7
HSPH, HN03, ppbv
36
-------
TECHNICAL REPORT DATA
1. Report No.
3. Recipient's Accession No.
4. Title and Subtitle
Status of Semi-continuous Monitoring Instrumentation for Gases
and Water-Soluble Particle Components Needed for Diagnostic
Testing of Ambient Air Quality Simulation Models
5. Report Date
6. Performing Organization Code
7. Author(s) W.A. McClenny (U.S. EPA, RTP, NC); K.Kronmiller,
K. Oliver, H.H. Jacumin, Jr.. E.H. Daughtrey, Jr. (ManTech
Environmental Technologies, Inc., RTP, NC)
8. Performing Organization
Report No.
9.Performing Organization Name and Address
National Exposure Research Laboratory
RTP, NC 27711
10. Program Element No.
11. Contract/Grant No. 68-D00-
206
12.Sponsoring Agency Name and Address
National Exposure Research Laboratory
RTP, NC 27711
13. Type of Report and Period
Covered, Annual Performance
Measure, 2002
14.Sponsoring Agency Code
15. Supplementary Notes
This report documents the status of development for monitoring methods that are required to meet
criteria established by NERL modelers. This supports the NERL's research efforts regarding APG16,
APM#13.
16. Abstract: This status report documents the effort by the National Exposure Research Laboratory (NERL), Office of
Research and Development (ORD) of the U.S. Environmental Protection Agency (U.S. EPA) to provide methods of
measurement, including calibration procedures, that meet the requirements for diagnostic testing of models describing
atmospheric photochemistry. These requirements include a measurement uncertainty of ±0.2 ppbv (±0.1 for N02) at 1.0
ppbv (or equivalent for particles) for compounds mentioned here as specified by scientists in the AMD, NERL. New
developments for monitors of N02, HCHO, H202, HN03, HONO, NH3, VOCs (including ^-aldehydes), particulate nitrate,
particulate sulfate, and particulate ammonium are discussed in the report. Descriptions of instrument design and
measurement principle for N02, HCHO, and water soluble gases and particlulate matter are presented. Results from the
2002 Tampa,PA BRACE study, the 2001 Philadelphia, PA NE-03PS study , and the 1999 Nashville, TN SOS field study
are also shown. Of particular note are the preliminary results of field monitoring of N02 in the BRACE field study with
systems using new commercial photolytic converters for N02 to NO conversion followed by chemiluminescence
detection.
17. KEY WORDS AND DOCUMENT ANALYSIS
A. Descriptors
Methods development, Ambient air
measurements, Diagnostic testing of models
B. Identifiers / Open Ended
Terms
C. COSATI
-------
TECHNICAL REPORT DATA
18. Distribution Statement
19. Security Class (This
Report)
21. No. of Pages
20. Security Class (This
Page)
22. Price
Form Available: Network Ncighborhood\Knight\Groups\HEASD\Forms\Technical-Report-Data-2220-l
------- |