September 2009
Environmental Technology
Verification Report
Applikon MARGA Semi-Continuous Ambient Air
Monitoring System
Prepared by
Banene
The Business of Innovation
Under a cooperative agreement with
U.S. Environmental Protection Agency
-------
Environmental Technology Verification
Report
ETV Advanced Monitoring Systems Center
APPLIKON MARGA SEMI-CONTINUOUS
AMBIENT AIR MONITORING SYSTEM
by
Brad Goodwin
Dawn Deojay
Ken Cowen
Thomas Kelly
Zachary Willenberg
Amy Dindal
Battelle
Columbus, Ohio 43201
and
John L. McKernan
Michelle Henderson
U.S. EPA
26 W. Martin Luther King Dr., MS 208
Cincinnati, Ohio 45268
-------
Notice
The U.S. Environmental Protection Agency, through its Office of Research and Development,
funded and managed, or partially funded and collaborated in, the research described herein. It
has been subjected to the Agency's peer and administrative review and has been approved for
publication. Any opinions expressed in this report are those of the author (s) and do not
necessarily reflect the views of the Agency, therefore, no official endorsement should be
inferred. Any mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
This report was prepared by Battelle to summarize testing supported by the EPA Clean Air
Markets Division (CAMD). Neither Battelle nor any of its subcontractors nor the EPA CAMD;
nor any person acting on behalf of either
(a) Makes any warranty of representation, express or implied, with respect to the accuracy,
completeness, or usefulness of the information contained in this report, or that the use of
any information, apparatus, method, or process disclosed in this report may not infringe
privately-owned rights; or
(b) Assumes any liabilities with respect to the use of, or for damages resulting from the use of,
any information, apparatus, method, or process disclosed in this report.
Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement,
recommendation, or favoring; nor do the views and opinions of authors expressed herein
necessarily state or reflect those of the EPA.
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for
investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten human health and the environment. The focus of the Laboratory's
research program is on methods and their cost-effectiveness for prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments and groundwater; prevention and control
of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public
and private sector partners to foster technologies that reduce the cost of compliance and to
anticipate emerging problems. NRMRL's research provides solutions to environmental
problems by: developing and promoting technologies that protect and improve the
environ-ment; advancing scientific and engineering information to support regulatory and policy
decisions; and providing the technical support and information transfer to ensure implementation
of environmental regulations and strategies at the national, state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term research plan.
It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
Sally Gutierrez, Director
National Risk Management Research Laboratory
-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the
verification test, analyze the data, and prepare this report. We acknowledge the support of the
U.S. EPA Office of Research and Development for providing the Burdens Creek field site for
this verification. In particular we acknowledge the efforts of John Walker in support of the field
testing and collection of continuous ammonia reference measurements. We also acknowledge
Solomon Ricks and Nealson Watkins of EPA's Office of Air Quality Planning and Standards
(OAQPS) for providing sulfur dioxide reference measurements. Finally, the authors thank Rudy
Eden of the South Coast Air Quality Management District, Cliff Glowacki of Covenant
Associates, and Joann Rice of EPA/OAQPS for their review of this verification report.
in
-------
Contents
Notice i
Foreword ii
Acknowledgments iii
List of Figures v
List of Tables vi
List of Abbreviations vii
Chapter 1 Background 1
Chapter 2 Technology Description 2
Chapter 3 Test Design and Procedures 4
3.1 Introduction 4
3.2 Test Procedures 6
3.3 Field Site 8
3.4 Verification Schedule 9
Chapter 4 Quality Assurance/Quality Control 10
4.1 Deviations 10
4.2 Reference Methods 10
4.2.1 Denuder/Filter Pack Sampling 10
4.2.2 Denuder/Filter Pack Analysis 11
4.2.3 Gas Analyzers 12
4.3 Audits 13
4.3.1 Performance Evaluation Audit 13
4.3.2 Technical Systems Audit 13
4.3.3 Data Quality Audit 15
4.4 QA/QC Reporting 15
4.5 Data Review 15
Chapters Statistical Methods 17
5.1 Accuracy 17
5.1.1 Regression Analysis 17
5.1.2 MARPD Analysis 18
5.2 Precision 18
5.2.1 Comparison of Paired Results 18
5.2.2 Comparison to Pooled Reference Method Results 18
5.3 Data Completeness 18
5.4 Reliability 19
IV
-------
5.5 Operational Factors 19
Chapter 6 Test Results 20
6.1 Accuracy 24
6.1.1 Regression Analysis 25
6.1.2 MARPD Analysis 28
6.2 Precision 29
6.3 Data Completeness 30
6.4 Reliability 30
6.5 Operational Factors 31
6.5.1 Ease of Use 31
6.5.2 Maintenance 32
6.5.3 Consumables/Waste Generation 33
Chapter 7 Performance Summary 34
List of Figures
Figure 6-1. Time sequence plot of SO2 measurement results from duplicate MARGA
instruments 20
Figure 6-2. Time sequence plot of FINOs measurement results from duplicate MARGA
instruments 21
Figure 6-3. Time sequence plot of NHs measurement results from duplicate MARGA
instruments 21
9
Figure 6-4. Time sequence plot of SO4 " measurement results from duplicate MARGA
instruments 22
Figure 6-5. Time sequence plot of MV measurement results from duplicate MARGA
instruments 22
Figure 6-6. Time sequence plot of NH4+ measurement results from duplicate MARGA
instruments 23
Figure 6-7. Time sequence plot of SO2 measurement results from duplicate MARGA and
Continuous SO2 Reference Monitor 23
Figure 6-8. Time sequence plot of NFL? measurement results from duplicate MARGA and the
Mean of the Continuous NHs Reference Monitors 24
Figure 6-9. Scatter Plots Comparing Reference Data from Duplicate Trains for Target Analytes.
25
-------
List of Tables
Table 4-1. Summary of Field Blank Analyses 11
Table 4-2. Results of Duplicate Checks of Denuder/Filter Pack Reference Samples 12
Table 4-3. Calibration Standard Coefficients of Variation 12
Table 4-4 Summary of Denuder/Filter Pack Flow Rate Checks 14
Table 4-5 Summary of PE Audits of Analytical Methods 14
Table 6-1. Summary of Regression Analysis Results for the MARGA Systems
Relative to Reference Method Results 27
Table 6-2. Summary of Denuder/Filter Pack Reference Method and MARGA
Regression Analysis Results versus Target Performance Goals 28
Table 6-3. Summary of Calculated MARPD Results for Reference Data and
MARGA Systems 29
Table 6-4. Summary of Calculated MARPD Results for Duplicate MARGAs 29
Table 6-5. Comparison of MARPD of 12-Hour MARGA Average Measurementswith
95* Percentile of Pooled RPD Results of Duplicate Reference Measurements 30
Table 6-6. Summary of Data Completeness for MARGAs 31
Table 6-7. Summary of MARGA Reliability Assessments 31
Table 6-8. Summary of Maintenance Activities Performed on MARGAs During
Verification Testing 32
Table 7-1. Summary of Verification Test Results for the MARGA 34
VI
-------
List of Abbreviations
AC
AMS
ADS
ARPD
CAMD
CASTNET
CV
DL
DQO
EPA
ETV
FEM
1C
ICP-AES
IPC
MARGA
MARPD
NCSU
NIST
PE
ppb
QA
QC
QMP
RPD
RPD95
RTF
SJAC
ISA
WRD
automated colorimetry
Advanced Monitoring Systems Center
annular denuder system
absolute relative percent difference
Clean Air Markets Division (U.S. EPA)
Clean Air Status and Trends Network
coeffi ci ent of van ati on
detection limit
data quality objective
U.S. Environmental Protection Agency
Environmental Technology Verification
Federal Equivalent Method
ion chromatography
inductively coupled plasma atomic emission spectroscopy
Industrial PC
Monitor for Aerosols and Gases in Ambient Air
median absolute relative percent difference
North Carolina State University
National Institute of Science and Technology
performance evaluation
parts per billion
quality assurance
quality control
quality management plan
relative percent difference
95* percentile of pooled results
Research Triangle Park
steam jet aerosol collector
technical systems audit
wet rotating denuder
Vll
-------
Chapter 1
Background
The U.S. Environmental Protection Agency (EPA) supports the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative environmental
technologies through performance verification and dissemination of information. The goal of
the ETV Program is to further environmental protection by accelerating the acceptance and
use of improved and cost-effective technologies. ETV seeks to achieve this goal by providing
high-quality, peer-reviewed data on technology performance to those involved in the design,
distribution, financing, permitting, purchase, and use of environmental technologies.
ETV works in partnership with recognized testing organizations; with stakeholder groups
consisting of buyers, vendor organizations, and permitters; and with the full participation of
individual technology developers. The program evaluates the performance of innovative
technologies by developing test plans that are responsive to the needs of stakeholders,
conducting field or laboratory tests (as appropriate), collecting and analyzing data, and
preparing peer-reviewed reports. All evaluations are conducted in accordance with rigorous
quality assurance and quality control (QA/QC) protocols to ensure that data of known and
adequate quality are generated and that the results are defensible.
The EPA's National Risk Management Research Laboratory and its verification organization
partner, Battelle, operate the Advanced Monitoring Systems (AMS) Center under ETV. The
AMS Center recently evaluated the performance of Applikon BV's Monitor for Aerosols and
Gases in Ambient Air (MARGA) semi-continuous ambient air monitoring system at the
Burdens Creek ambient air quality monitoring site in Research Triangle Park, North
Carolina. Semi-continuous ambient air monitoring systems were identified as a priority
technology category for verification through the AMS Center stakeholder process.
-------
Chapter 2
Technology Description
The objective of the ETV AMS Center is to verify the performance characteristics of
environmental monitoring technologies for air, water, and soil. This report provides results
for the verification testing of Applikon BV's MARGA semi-continuous ambient air
monitoring system. The following is a description of the MARGA, based on information
provided by the vendor. The information provided below was not verified in this test.
The MARGA ADI 2080 is an on-line analyzer for semi-continuous measurement of gases
and soluble ions in aerosols. The MARGA utilizes a Wet Rotating Denuder (WRD) to collect
acid gases and ammonia by diffusion into an aqueous film. Particles pass through the WRD
and are collected in a Steam Jet Aerosol Collector (SJAC). Within the SJAC, a
supersaturated environment is created which grows particles by a process known as
deliquescence, allowing them subsequently to be collected by inertial separation. As cooling
takes place, steam condenses and washes the collected particles into an aqueous sample
stream. The aqueous solutions from the WRD and SJAC are subsequently analyzed by ion
chromatography (1C) for soluble anions and cations. Software integrated within the
MARGA calculates atmospheric concentrations based on air sample flow rate and the ion
concentrations in the collected solutions.
The MARGA ADI 2080 ambient air monitor components:
a A sampling box
a An analytical box
a Industrial PC (IPC) with Keyboard/Mouse and Screen
a ADI 2080 Ambient air monitor software
a Programmable Logic Control Input/Output modules, and software
a Applikon pump modules and stainless steel analyzer cabinet
a Polypropylene rack with steel inner body
a Uninterruptable power supply
a Air pump with mass flow controller
The analyzer consists of two boxes: the upper sampling box and the lower analytical box. Air
is drawn through the sampling system in the upper box where inorganic gases and aerosols
-------
are absorbed and collected into separate aqueous solutions. In the analytical box, the
inorganic compounds in the gases and aerosols are determined by 1C.
The analytical box also contains an IPC running instrument software that controls all
elements in the process with a fold-up liquid crystal display as well as a keyboard with
mouse. The MARGA software running on the IPC controls the instrument and provides a
user interface. In addition, the analyzer can be checked and controlled remotely via an
internet or modem connection. Figure 2-1 shows pictures of the sampling and analytical
boxes of the MARGA ADI 2080.
Sampling Box
Analytical Box
Screen/
Keyboard with
mouse
Figure2-1. MARGA ADI 2080 Sampling and Analytical Boxes
-------
Chapter 3
Test Design and Procedures
3.1 Introduction
EPA's Clean Air Status and Trends Network (CASTNET) is a regional long-term
environmental monitoring program, established in 1991 under the Clean Air Act
Amendments, which is administered and operated by EPA's Clean Air Markets Division
(CAMD). Presently there are a total of 86 operational CASTNET sites located in or near
rural areas and sensitive ecosystems collecting data on ambient levels of pollutants where
urban influences are minimal. As part of an interagency agreement, the National Park Service
sponsors 27 sites which are located in national parks and other Class-I areas designated as
deserving special protection from air pollution.
Throughout CASTNET, measurements are made to characterize the ambient concentrations
of the following species:
• Sulfur dioxide (SO2)
• Particulate sulfate (SO4"2)
• Particulate nitrate (NO"3)
• Nitric acid (HNO3)
• Particulate ammonium (NH4+)
• Particulate calcium (Ca2+)
• Particulate sodium (Na+)
• Particulate magnesium (Mg2+)
• Particulate potassium (K+)
• Particulate chloride (Cl~)
• Ozone (Os)
For all but ozone, ambient air sampling of particles and selected gases is performed by
drawing air at a controlled flow rate through an open face, three-stage filter pack that uses
four sequential filters (Teflon®, Nylon®, and dual Whatman® filters impregnated with
potassium carbonate). The filter packs are located at 10 meters above the ground surface and
accessed using a tilt-down aluminum tower. The filter packs are exchanged every week by a
-------
site operator and the exposed filter packs are shipped to a central analytical laboratory for
analysis. Although the filter pack is simple to use, reliable, inexpensive, and provides
sensitive measurements, it suffers from long sampling duration (7-day integrated average)
and is subject to bias and uncertainties in species of interest such as gaseous HNOs and
particle nitrate (NCV) due to reactivity and volatilization issues.1"3 In addition, due to the
time required for chemical analysis and reporting, preliminary concentration data from a
CASTNET site are typically not available until 4-6 months from the sample collection date.
Recent advancements in ambient air monitoring instrumentation now provide the capability
to observe operating status remotely and to allow real-time or near real-time (within 24
hours) access to monitoring data. The advantages of routine operation of such systems
include a much more timely data stream and improved air quality assessment capability.
Real-time, multi-pollutant monitoring in rural areas will help to better characterize the extent
of regional transport of pollutants (i.e., particulate matter and gaseous precursors), provide
improved regional dry deposition estimates, and help in both the development and validation
of air quality models.
This verification test was conducted according to procedures specified in the ETV Test/QA
Plan for Verification of Semi-Continuous Ambient Air Monitoring Systems.4 The purpose of
this verification test was to generate performance data on semi-continuous ambient air
monitoring technologies so organizations and users interested in installing and operating
these systems can make informed decisions about their potential benefit and, specifically, use
in CASTNET. The test was conducted over a period of approximately 30 days and involved
the continuous operation of duplicate semi-continuous monitoring technologies at an existing
ambient air monitoring station located near EPA laboratories in Research Triangle Park
(RTF), North Carolina. The accuracy of the monitoring technologies was determined
through comparisons to modified EPA reference methods for individual gaseous and
particulate species. Modifications to the reference methods primarily involved increasing the
sampling flow rate to reduce overall sampling times and help minimize measurement bias
and uncertainties, while still meeting the data quality objectives of this verification test. The
precision of the semi-continuous monitoring was determined from comparisons of paired
data from duplicate units, and through comparisons to pooled results of the reference
methods. Other performance parameters such as data completeness, maintenance
requirements, ease of use, and operational costs were assessed from observations by the
Battelle field testing staff. Target performance goals were established by EPA to demonstrate
if these monitoring systems are suitable for use in CASTNET. These specifications are referred
to as "target performance goals" in this report and were used to evaluate data generated by the
MARGA and reference methods. This test was not intended to simulate long-term (e.g., multi-
year) performance of semi-continuous monitoring technologies at a monitoring site. As such,
performance and maintenance issues associated with long-term use of the MARGA are not
addressed in this report.
The MARGA was verified by evaluating the following parameters:
• Accuracy as compared to reference measurements
• Precision between duplicate units
-------
• Data completeness
• Reliability
• Operational factors such as ease of use, maintenance and data output needs, power
and other consumables use, and operational costs
The MARGA was verified during a field test conducted from October 1 to October 31, 2008.
Testing was conducted at the Burdens Creek Air Monitoring Site in RTF, NC. The
monitoring systems were operated and maintained by the vendor throughout the field period.
Duplicate, integrated denuder/filter pack reference samples were collected over 12-hour
sampling intervals throughout the testing period, from 6:00 am to 6:00 pm and from 6:00 pm
to 6:00 am daily. The denuder/filter pack samples were collected and analyzed by North
Carolina State University (NCSU). Additionally, the MARGA units were collocated with
separate continuous gas analyzers for SO2 and ammonia (NHs), which were operated and
maintained by EPA staff throughout the testing period.
In the test reported here, the MARGA performance was verified for measurement of SO2,
HNOs, and NHs in the gas phase; and MV, SO42", and NH4+ in the particle phase. In
addition, data completeness was evaluated for Cl", Ca +, and Na+ in the particle phase.
3.2 Test Procedures
During testing, duplicate semi-continuous ambient air monitoring systems were installed
inside an environmentally controlled shelter at the Burdens Creek Air Monitoring Site. The
monitoring systems were operated and maintained by the vendor, and intended to operate
continuously over the 30-day testing period. Maintenance performed on the monitoring
systems was conducted by the vendor, documented by Battelle, and is reported in Section 6.4
of this report. Data from the monitoring systems was retrieved by the vendor and provided to
Battelle within 24 hours of collection.
Annular Denuder Systems (ADS) based on Compendium Method IO-4.25 were used as the
reference comparison method and consisted of a sodium carbonate (IS^COs) coated denuder
and phosphorus acid (H3PO3) coated denuder in series for the collection of acid and base
gases, respectively, followed by a Teflon filter for the collection of parti culate matter, a
Nylon filter for the collection of volatilized particulate nitrate, and a HaPOs coated denuder
"chaser" for the collection of volatilized particulate ammonium. The denuder/filter pack
samplers were installed on the roof of the trailer housing the monitoring systems being tested
and collected ambient air samples at a flow rate of 10 liters per minute (L/min). Figure 3-1
shows the sampling shelter with denuders and filter packs set up on the roof.
-------
Figure 3-1. Sampling Shelter with Denuder/Filter Packs Deployed
Each reference sampler manifold contained eight measurement channels, which were used
for seven denuder/filter pack sample trains and one blank train. Thus, when fully loaded the
reference samplers could carry out three and one-half days of routine sampling (i.e., seven
successive 12-hour samples). Consequently, changeout of collected reference samples and
reloading of the samplers was conducted twice each week. The denuder/filter pack samples
were retrieved and returned to the analytical laboratory for disassembly, extraction, and
analysis. Figure 3-2 shows the retrieval process for the denuder/filter packs. After
disassembly in the laboratory, the filters and denuders were extracted using deionized water
and analyzed for target analytes. The denuder extracts were analyzed for SO2 (as SO42~),
nitrous acid (HONO) (as nitrite (NOf), HNO3 (as NO3'), NH3 (as NH4+), and HC1 (as Cl').
The Teflon filter extracts were analyzed for SO42", NO3", NH4+, Cl', Ca2+, Mg2+, Na+, and K+.
The Nylon filter extracts were analyzed for NO3", and the backup denuder chaser extracts
were analyzed for NH4+. Analysis for each of the target analytes was performed by 1C based
on the procedures described in EPA Method 300.0.6 Additional analysis for NH4+ was
performed by automated colorimetry (AC) based on the procedures described in EPA
Method 350.1.7
-------
I
Figure3-2. Denuder/Filter Pack Retrieval
Specific sample handling procedures were implemented to minimize handling of the
denuder/filter pack components and limit the number of transfers of the denuder/filter packs.
When not in use, the denuders and assembled filter packs were sealed or capped, to prevent
contamination. Clean lint-free gloves were used when handling the denuder/filter pack
components. Clean forceps were used when handling filters. The denuders and filter packs
were assembled in NCSU's analytical laboratory facilities and transferred by NCSU staff to
the Burden's Creek Air Monitoring Site for sampling. Special care was taken to avoid
breathing on components of the denuder/filter pack reference samples, to minimize ammonia
contamination.
3.3 Field Site
The Burdens Creek Air Monitoring Site is near the EPA offices in RTF and is maintained by
EPA staff. The site consists of an open area within surrounding forested land, and is subject
to restricted access at all times. A variety of routine measurements are performed at this site
and it is periodically used for special studies. The MARGAs evaluated during this
verification were housed in an environmentally controlled shelter along with the continuous
NHa instrument. The denuder/filter pack samplers were located on a platform on the roof of
the trailer. Pumps for the denuder/filter packs were located in pump boxes adjacent to the
-------
trailer. Continuous SC>2 Federal Equivalent Method (FEM) measurements were collected in
a separate trailer approximately 30 yards from the trailer housing the MARGAs.
3.4 Verification Schedule
The MARGA verification field effort took place from October 1 through October 31, 2008.
Duplicate MARGA units had been installed by the vendor and were operating at the Burdens
Creek site for several weeks prior to the start of the verification test. The vendor performed
routine maintenance on both units to prepare them for the start of the verification test. The
continuous SO2 and NH3 analyzers used for reference measurements were also installed and
operating at the site before the start of the MARGA evaluation.
Denuder/filter pack reference measurements began on October 1 at 6:00 am and ran in 12-
hour integrated samples through October 31 at 6:00 pm. However, each week two of the
denuder/filter pack samples were collected on a time period shorter than 12 hours. The
sample that should have started on Tuesdays at 6:00 am was started three hours later to allow
for changeout of the denuder/filter packs. This sample was started on the hour after the
changeout was completed to allow for comparison with hourly MARGA data. Similarly, the
sample that should have finished sampling on Fridays at 6:00 pm was stopped at 3:00 pm to
allow for changeout of the denuder/filter packs. All start and stop times were recorded by
NCSU staff as part of the deployment/collection process.
-------
Chapter 4
Quality Assurance/Quality Control
QA/QC procedures and all verification testing were performed in accordance with test/QA
plan for this verification test4 and the quality management plan (QMP) for the AMS Center8
except where noted below. QA/QC procedures and results are described below.
4.1 Deviations
There was one documented deviation from the test/QA plan during this verification test. The
deviation involved the use of 1C rather than inductively coupled plasma atomic emission
spectroscopy (ICP-AES) for the analysis of the metal cations (Na+ and Ca2+) from the
collected reference method samples. 1C was chosen as the preferred method since 1C
allowed for reanalysis of samples if needed. These cation data were used only for
determination of data completeness, and the change in analytical methods did not negatively
impact data quality
4.2 Reference Methods
The following sections describe the QA/QC procedures employed in the collection and
analysis of reference samples.
4.2.1 Denuder/Filter Pack Sampling
This verification test included a comparison of MARGA results to those of the duplicate
denuder/filter pack reference measurements. Quality control activities for the filter pack
sampling included flow rate checks performed on the sampling trains and the collection of
field blank samples. Prior to each sampling event, each sampling train was checked for leaks
to ensure proper operation.
On each of the duplicate denuder/filter pack reference sample manifolds, one of the eight
channels was reserved for the collection of field blank samples. The field blanks were
collected by installing the sampling media (i.e., denuder and filters) in the sampling train but
without drawing any air through the train. The field blank samples remained installed until
the denuder/filter pack reference samples collected at the same time were retrieved. The
10
-------
field blank media were then recovered along with the other samples, and were handled and
analyzed like normal samples. Additionally, travel blank samples were collected for the
Na2CC>3 and HsPOs coated denuders. Table 4-1 presents a summary of the denuder/filter
blank analyses, including the detection limit for each species and the number of blanks with
results above the detection limit. The results of the field blank analyses were subsequently
subtracted from the corresponding denuder/filter pack reference samples that were collected
at the same time as the field blank samples. When the result of the field blank analysis was
below the detection limit, the detection limit value was subtracted from the corresponding
denuder/filter pack reference sample result.
Table 4-1. Summary of Field
Medium
Teflon filter
Nylon filter
Na2CO3 denuder
H3PO3 denuder
H3PO3 chaser
Analyte
NH4+
NO3-
so42-
NH4+
NO3-
NO3- -
so42- -
NH4+ -
NH4+
Blank Type
Field blank
Field blank
Field blank
Field blank
Field blank
Field blank
Travel blank
Field blank
Travel blank
Field blank
Travel blank
Field Blank
Det. Limit
(UK)
0.252
O.I2
0.25
0.052
0.152
0.22
0.22
1.0
1.0
O.I2
O.I2
O.I2
Blank Analyses
# of Blank
Samples
26
26
26
26
26
26
30
26
30
26
30
26
# above
D.L.
4
12
2
20
7
o
J
0
25
26
26
16
18
Average (u.g)
(St. Dev.)1
1.41 (0.55)
1.25 (1.22)
1.1(1.1)
0.34 (0.34)
0.29(0.12)
1.07 (0.71)
~
2.19(0.61)
2.03 (0.52)
0.57 (0.42)
0.21 (0.03)
0.28(0.12)
- Average and standard deviation of the results above the detection limit.
2 - Detection limit for nitrogen compounds is the mass of nitrogen only, not the mass of the compound.
4.2.2 Denuder/Filter Pack Analysis
The analysis of the denuder/filter pack samples was conducted by 1C based on EPA Method
300.06 and by AC based on EPA Method 350.1.7 Analysis of these samples was subject to
the data quality criteria of the respective methods, which included the analysis of duplicate
samples, blanks, and calibration check standards with every batch of samples analyzed by the
different analytical methods. For each duplicate analysis the absolute relative percent
difference (ARPD) between the measured results was calculated. Table 4-2 summarizes the
results of the analysis of the duplicate samples for the collected denuder/filter pack reference
samples. Only those samples for which both duplicate results are above twice the detection
limit are included in this summary. The duplicate analysis exceeded the acceptance criterion
of 20% ARPD established in the test/QA plan4 a total of seven times. The ARPD exceeded
20% in one instance for the analysis of SC>4 " from the Teflon filters, in three instances for the
analysis of NH4+ from nylon filters, in one instance for the analysis of NCV from nylon
filters, and in two instances for the analysis of NH4+ from the HaPOs denuders. In most
cases, the exceedances were a result of limitations in the reporting precision of the analytical
equipment because the ambient concentrations of the target analytes were very low. The
causes for the other exceedances were not apparent.
11
-------
Calibration curves were run after every set of 25 samples using aqueous solution calibration
standards and method blanks. Table 4-3 presents the coefficient of variation (CV), defined as
the standard deviation of the instrumental results divided by the mean of the results, for the
individual standards used in the calibrations. Shown for each analyte is the concentration of
the standard in ng/L, accompanied by the C V of all such standards (as a percent, in
parentheses). The CV results are all <4.0% for SO42", <5.7% for NO3", and <1.1% for NH4+.
Table 4-2. Results of Duplicate Checks of Denuder/Filter Pack Reference Samples
Medium
Teflon filter
Nylon filter
Na2CO3 denuder
H3PO3 denuder/chaser
Analyte
NH4+
NO3
SO42
NH4+
NO3
NO3
SO42
NH4+
#of
Samples
17
17
17
16
16
22
22
35
# above
2 x D.L.
9
10
8
11
14
10
14
30
Average
ARPD
5.1%
2.9%
7.3%
9.7%
2.3%
1.7%
2.1%
4.4%
Max.
ARPD
18.9%
8.7%
40.8%
28.6%
22.2%
11.8%
7.5%
23.3%
Table 4-3. Calibration Standard Coefficients of Variation
Analyte
Standard
Concentration
fna/T ^
\}*&L')
(CV)
Blank
1
2
3
4
SO42
Oa
400(3.6%)
800 (4.0%)
1600 (2.5%)
4000(1.8%)
N03
Oa
90 (5.7%)
160 (3.3%)
360 (2.0%)
900(1.8%)
NH4+
Oa
310(0.6%)
620(1.1%)
1240 (0.7%)
3100(0.8%)
a CV results are not reported for blank samples since the mean result is set to zero.
4.2.3 Gas Analyzers
The continuous gas analyzers used for this verification test were already in operation at the
Burdens Creek site and were included in routine QC activities at the site. Quality control
activities associated with the SC>2 continuous gas analyzer included multipoint calibrations of
the analyzer, routine zero/span checks, and biweekly precision checks. A multipoint
calibration of the NH3 continuous gas analyzer was performed before and after the
verification test. The continuous NHs data were also corrected for water vapor interference
based on an observed linear relationship between the Pranalytica baseline and atmospheric
dewpoint. No additional QC activities were implemented specifically for this verification
test although documentation of the QC activities performed during testing was provided to
Battelle by EPA.
12
-------
4.3 Audits
Three types of audits were performed during the verification test: a performance evaluation
(PE) audit of the denuder/filter pack reference method sampling and analysis, a technical
systems audit (TSA) of the verification test performance, and a data quality audit. Audit
procedures are described further below.
4.3.1 Performance Evaluation Audit
PE audits of the denuder/filter pack reference method sampling procedures were performed
by measuring the sample flow rate through the denuder/filter pack inlet during sampling.
The flow rate was measured using a National Institute of Science and Technology (NIST)-
traceable flow transfer standard. During the testing period, a total of 42 flow rate checks
were performed. The results of those checks are summarized in Table 4-4.
Table 4-4 shows that in 30 of the 42 flow checks, the sampler flow rate was within the target
±5% tolerance of the nominal flow rate. In the 12 cases where the measured flow rate was
outside that tolerance, the sampling trains were inspected for any apparent problems. In two
cases, an obstruction that completely blocked the air pathway was found and removed. In the
other cases the flow rate was only slightly outside the target tolerance with a range from
+5.1% to -11.1%. No apparent cause for the discrepancy in flow rates was found and in
those cases the measured (rather than nominal) flow rate was used in calculating the ambient
concentrations.
Additionally, a PE audit of the analytical methods was performed by supplying the analytical
laboratory with samples prepared from independent NIST-traceable standard solutions. The
samples were analyzed and the results are summarized in Table 4-5. The target acceptance
criteria for the PE audit results were 5% for the 1C results and 10% for the AC results. In all
but one case (NH4+ by AC at 1,000 ng/L), the results of the PE audit met the target
acceptance criteria.
4.3.2 Technical Systems Audit
The Battelle Quality Manager performed a TSA of the testing procedures during the first
week of the verification test. The purpose of this audit was to ensure that the verification test
was being performed in accordance with the AMS Center QMP,8 the test/QA plan for this
verification test,4 published reference methods,5"7 and any SOPs used by the analytical
laboratory. In this audit, the Battelle Quality Manager reviewed the reference methods used,
compared the actual test procedures being performed to those specified or referenced the
test/QA plan, and reviewed data acquisition and handling procedures. The TSA was
performed at both the verification test site and the analytical laboratories at NCSU where the
13
-------
Table 4-4 Summary of Denuder/Filter Pack Flow Rate Checks
Date
10/1/2008
10/1/2008
10/1/2008
10/1/2008
10/2/2008
10/2/2008
10/2/2008
10/2/2008
10/7/2008
10/7/2008
10/7/2008
10/7/2008
10/10/2008
10/10/2008
10/14/2008
10/14/2008
10/14/2008
10/14/2008
10/17/2008
10/17/2008
10/17/2008
10/17/2008
10/21/2008
10/21/2008
10/24/2008
10/24/2008
10/28/2008
10/28/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
10/31/2008
Train
Train 1-3
Train 1-4
Train 2-3
Train 2-4
Train 1-5
Train 1-6
Train 2-5
Train 2-6
Train 1-1
Train 1-2
Train 2-1
Train 2-2
Train 1-7
Train 2-7
Train 1-3
Train 1-4
Train 2-3
Train 2-4
Train 1-5
Train 1-6
Train 2-5
Train 2-6
Train 1-7
Train 2-7
Train 1-1
Train 2-1
Train 1-2
Train 2-2
Train 1-1
Train 1-2
Train 1-3
Train 1-4
Train 1-5
Train 1-6
Train 1-7
Train 2-1
Train 2-2
Train 2-3
Train 2-4
Train 2-5
Train 2-6
Train 2-7
Measured
Flow (L/min)
9.33
0.00
9.82
10.00
9.70
0.00
10.01
9.95
9.52
9.51
9.88
9.69
9.90
10.47
10.14
10.30
10.42
10.36
9.66
9.52
10.00
9.51
10.51
9.76
9.97
10.31
9.62
9.51
9.22
9.33
8.89
9.35
9.18
9.37
9.20
9.99
9.43
10.09
10.04
9.98
10.02
9.98
%Difference
from Nominal
-6.7%
-100%
-1.8%
0.0%
-3.0%
-100%
0.1%
-0.5%
-4.8%
-4.9%
-1.2%
-3.1%
-1.0%
4.7%
1.4%
3.0%
4.2%
3.6%
-3.4%
-4.8%
-0.04%
-4.9%
5.1%
-2.5%
-0.3%
3.1%
-3.8%
-5.0%
-7.8%
-6.7%
-11.1%
-6.5%
-8.2%
-6.3%
-8.1%
-0.1%
-5.7%
0.9%
0.4%
-0.3%
0.2%
-0.3%
Comment
No cause identified3
Flow obstruction
Flow obstruction
No cause identified3
No cause identified3
No cause identified3
No cause identified3
No cause identified3
No cause identified3
No cause identified3
No cause identified3
No cause identified3
a: Sampler operating normally, no cause
tolerance. Measured flow rate used to
was found for flow measurement outside of ±5% target
calculate ambient concentrations.
14
-------
Table 4-5 Summary of PE Audits of Analytical Methods
Analyte
NH4+
NH4+
NH4+
NO3
NO3
SO42
NH4+
Analytical Standard
Method Concentration ((Ag/L)
AC
AC
AC
AC
1C
1C
1C
1000
100
100
1000
1000
1000
100
Measured Percent
Concentration (ng/L) Difference
945
95
103
997
1007
1050
107
-5.5
-5.0
3.0
-0.3
0.7
5.0
7.0
reference method analyses were performed. During the ISA, the Battelle Quality Manager
observed the reference method sampling and sample recovery; inspected documentation of
reference sample chain of custody; and reviewed laboratory record books. He also checked
data acquisition procedures, and conferred with the vendor, EPA, and NCSU testing staff.
As noted in Section 4.1, one deviation from the test/QA plan was identified as a result of the
TSA. The deviation involved the use of 1C rather than ICP-AES for the analysis of the metal
cations (Na+ and Ca2+) from the collected reference method samples. 1C was chosen as the
preferred method since a significantly smaller volume of sample was required and thus 1C
allowed for reanalysis of samples if needed. These cation data were used only for
determination of data completeness, and the change in analytical methods did not negatively
impact data quality since the detection limits for the two methods are approximately equal.
4.3.3 Data Quality Audit
At least 10% of the data acquired during the verification test were audited. Battelle's Quality
Manager, or designee, traced the data from the acquisition, through reduction and statistical
analysis, to final reporting, to ensure the integrity of the reported results. All calculations
performed on the data undergoing the audit were checked.
4.4 QA/QC Reporting
Each audit was documented in accordance with Sections 3.3.4 and 3.3.5 of the QMP for the
ETV AMS Center. The results of the TSA were submitted to the EPA.
4.5 Data Review
All data received from the Vendor from the two MARGA units, from the EPA for the SC»2
and NHs analyzers, and from NCSU for the denuder/filter pack reference measurements
underwent 100% validation by Battelle technical staff before being used for any statistical
calculations. Data were assessed technically and results that appeared anomalous, based on
comparisons to other comparable data, were flagged and removed from additional statistical
calculations. When possible, the cause of the anomalous data was investigated through
15
-------
analysis of other testing records (e.g., logbook entries) to assess if unusual events occurring
during testing may have potentially affected the data in question.
Records generated in the verification test received a one-to-one review before these records
were used to calculate, evaluate, or report verification results. Data were reviewed by a
Battelle technical staff member involved in the verification test. The person performing the
review added his/her initials and the date to a hard copy of the record being reviewed.
16
-------
Chapter 5
Statistical Methods
The statistical methods used to evaluate the quantitative performance factors listed in Section
3.1 are presented in this chapter. Qualitative observations were also used to evaluate
verification test data.
5.1 Accuracy
The accuracy of the MARGA monitoring systems was evaluated in two ways for each of the
target analytes (SO2, HNO3, NH3, SO42", NO3~, and NH4+).
5.1.1 Regression Analysis
Firstly, the accuracy was determined from a linear least squares regression analysis of the
measured concentrations of the target analytes determined from the MARGA systems and the
corresponding reference methods. For comparison to the denuder/filter pack reference
samples, average concentrations from each of the two MARGA systems were determined
separately for each of the 12-hour sampling periods during the testing period, by averaging
the 1-hour results over the corresponding sampling periods. For each of the duplicate
MARGAs, these averages were plotted separately against the mean of the corresponding
duplicate reference method measurements. The slope and intercept of these plots were
determined from a linear regression analysis and are reported independently for each of the
duplicate monitoring systems, and for each target analyte. For the continuous gas
measurements (862 and NH3), 1-hour average concentration readings from each monitoring
system were plotted against the corresponding 1-hour average reference measurements,
excluding data below twice the instrument detection limit. Again, the slope and intercept of
these plots were determined from a linear regression analysis and are reported independently
for each of the duplicate monitoring systems.
17
-------
5.1.2 MARPD Analysis
The accuracy of the MARGA systems in terms of median ARPD (MARPD) was calculated
as the median value of the ARPD results determined using Equation 1:
ARPD =
C,-C(ref\
•100
(1)
where Ct and C(ref)i are the average target analyte concentration measured by a MARGA
system and the mean of the duplicate reference method concentrations, respectively, for the
i' reference sampling period.
The accuracy of the MARGA systems was determined for all periods for which
concentrations determined by both the reference method samplers were greater than twice the
detection limit.
5.2 Precision
5.2.1 Comparison of Paired Results
For this assessment of precision, the MARPD between the paired measurements from the
duplicate MARGA monitoring systems was calculated as the median value of the ARPD
values determined using Equation 2:
ARPD =
where C(l)t and C(2)t are the target analyte concentration measured by the first and second of
the two duplicate monitoring systems. For this calculation, periods when measurement data
for either of the MARGA systems was below twice the instrumental detection limit was
excluded from the analysis.
Precision was assessed independently for each target analyte.
5.2.2 Comparison to Pooled Reference Method Results
Precision was also assessed through comparisons of the MARPD to the 95* percentile of the
pooled relative percent difference of the duplicate reference method measurements. For this
calculation, periods when measurement data for either of the reference method samples were
below twice the instrumental detection limit were excluded from the analysis.
Precision was assessed independently for each target analyte.
5.3 Data Completeness
Data completeness was assessed in two ways, based on the overall data return achieved by
each MARGA system during the testing period. For each of the duplicate MARGA systems,
this calculation used the total hours of apparently valid data reported by the monitoring
18
-------
systems and available within 24 hours, divided by the total hours of data in the entire field
period. Also, the number of hours of valid monitoring system data was assessed relative to
the number of hours in each reference method sampling period. The performance goals for
both of these measures of data completeness were > 80%. The causes of any substantial
incompleteness of data return were established from operator observations or vendor records,
and noted in the discussion of data completeness results.
5.4 Reliability
Instrument reliability was assessed in two ways. Firstly, reliability was assessed in terms of
the percentage of time that the monitoring systems operated in measurement mode over the
duration of the test period. This assessment is reported independently for the two duplicate
MARGA systems. Additionally, reliability was assessed in terms of the ability of the
instruments to perform a controlled shut-down in the case of a power failure, followed by an
automated return to measurement mode within four hours after power has been restored. For
this assessment, the testing staff imposed a temporary power outage at the test site and
monitored the performance of one of the monitoring systems during and after the power
outage.
5.5 Operational Factors
Operational factors such as maintenance needs, data output, consumables used, ease of use,
repair requirements, etc., were evaluated based on observations recorded by Battelle and
facility staff, and explained by the vendor as needed. Battelle staff were at the monitoring
site whenever the vendor was present and recorded all activities performed on the monitoring
systems. A laboratory record book was maintained at the test site, and was used to enter
daily observations on these factors. Examples of information recorded in the record book
include the daily status of diagnostic indicators for the monitoring systems; use or
replacement of any consumables; the effort or cost associated with maintenance or repair;
vendor effort (e.g., time on site) for repair or maintenance; the duration and causes of any
down time or data acquisition failure; and Battelle testing staff observations about ease of use
of the monitoring systems.
19
-------
Chapter 6
Test Results
Figures 6-1 through 6-6 show time sequence plots of the duplicate MARGA data recorded on
an hourly basis during the verification testing period for 862, HNOs, NHs, SC>42~, N(V, and
NH4+, respectively. For comparison the mean denuder/filter pack reference method results
for the respective sampling periods are also presented in these figures, where the vertical
error bars on the reference measurements represent the range of the duplicate reference
measurements, and the horizontal extent of the reference result indicates the time frame of
sampling. Where no error bars are shown, only one reference measurement was above
detection limit. The quality of the reference method results as a basis for assessing MARGA
performance is discussed in subsequent sections of this chapter.
20
MARGA1
MARGA 2
Mean Reference
SO,
10/1/08
10/8/08
10/15/08
Date
10/22/08
10/29/08
Figure 6-1. Time sequence plot of S02 measurement results from duplicate MARGA and mean
denuder/filter pack reference method measurements.
20
-------
-------
10
so
2-
—MARGA1
—MARGA2
— Mean Reference
10/1/08
10/8/08
10/15/08
Date
10/22/08
10/29/08
Figure 6-4. Time sequence plot of S042" measurement results from duplicate M ARG A and
mean denuder/filter pack reference method measurements.
D)
o
s
-I— •
-------
NH4+
— MARGA1
— MARGA 2
— Mean Reference
10/1/08
10/8/08
10/15/08
Date
10/22/08
10/29/08
Figure 6-6. Time sequence plot of NH4+ measurement results from duplicate MARGA and
mean denuder filter pack reference method measurements.
In addition to the denuder/filter pack reference samples, separate continuous gas analyzers
were used to monitor 862 and NHs Figures 6-7 and 6-8 show time sequence plots of the
measurements from these gas analyzers with the corresponding MARGA measurements.
10/01/08
10/08/08
10/15/08
Date
10/22/08
10/29/08
Figure 6-7. Time sequence plot of S02 measurement results from duplicate MARGA and
Continuous S02 Reference Monitor.
23
-------
10/01/08
10/08/08
10/15/08
Date
10/22/08
10/29/08
Figure 6-8. Time sequence plot of NH3 measurement results from duplicate M ARG A and the
Mean of the Continuous NH3 Reference Monitors.
As discussed in Chapter 4, a number of QA/QC activities were performed to assure the
quality of the reference method data. Figure 6-9 shows scatter plots of the duplicate
denuder/filter pack reference trains (Train 2 vs. Train 1) for each of the target analytes. The
reference method data failed to meet some of the target performance goals established by
EPA in consideration of the monitoring needs for CASTNET and are not always of sufficient
quality to allow a definitive assessment of the MARGA in terms of these target performance
goals. Nonetheless, the results of the verification tests of the MARGA semi-continuous
ambient air monitoring system are presented below for each of the performance parameters.
Where appropriate, the performance of the reference method relative to the target
performance goals is shown to indicate the relative utility of the reference method results as
an appropriate standard for testing against the target performance goals.
6.1 Accuracy
The accuracy of the MARGA systems was determined in two ways. Firstly, accuracy was
determined from a linear least squares regression analysis of the measured concentrations of
the target analytes determined from each of the two MARGA monitoring unit and the
corresponding reference methods as described in Section 5.1.1. Also, accuracy was
determined from the MARPD of the differences between the MARGA data and the mean of
the reference method data for all sampling periods in which the measured reference
concentrations were greater than twice the detection limit, as described in Section 5.1.2. The
results of these analyses are presented below.
24
-------
Soluble Gases
Soluble Aerosol Ions
6 -
.E 4 -
8 2H
y= 1.04X + 0.11
r2 = 0.94
MARPD = 24%
246
SO2 - Train 1 (ng/m3)
6 -
I 4
o
co
y= 1.03X + 0.10
r2 = 0.81
MARPD= 15%
246
SO42' - Train 1 (ng/m3)
E
ra 2 -
1 -
y = 0.76x + 0.09
r2 = 0.64
MARPD = 29%
1 2
HNO3 - Train 1 (ng/m3)
2 -
1 -
y = 0.72x + 0.24
^ = 0.21
MARPD = 50%
1
NO3" - Train 1 (ng/m3)
.E 2 -
1 -
y= 1.25X + 0.11
r2 = 0.46
MARPD = 23%
1 2 3
NH3 - Train 1 (ng/m3)
D) 2
y = 0.57x + 0.30
R2 = 0.34
MARPD = 37%
1 2
NH4+ - Train 1 (ng/m3)
Figure6-9. Scatter Plots Comparing Reference Data from DuplicateTrains for Target
Analytes.
6.1.1 Regression Analysis
Figure 6-10 shows regression plots of the results from the duplicate MARGA instruments
versus the denuder/filter pack reference results for each of the target analytes. Figure 6-11
25
-------
so,
15
10 -
15
i=
CD
O
c
O
O
5 -
MARGA1 =1.09x + 0.60
r2 = 0.89
MARGA2 = 1.09x + 0.57
r2 = 0.90
0 5 10
Mean Reference Method Concentration
15
12
f
15
i=
(D
O
C
O
O
SO,
MARGA 1 = 0.92x + 0.68
r2 = 0.91
MARGA 2 = 0.87x +0.63
r2 = 0.89
048
Mean Reference Method Concentration
12
HNO,
4 -
3 -
o
1
4
15 3
-fc
CD
O
g 2
O
1 -
MARGA 1 =0.82x+0.08
r2 = 0.67
MARGA 2 = 0.85x+0.21
r2 = 0.68
01234
Mean Reference Method Concentration
Figure6-10. Regression Plots of MARGA data versus mean reference method data for S02, HN03,
NH3, S042~, N03~, and NH4+.(*Oneoutlier excluded from theSO2 regression for MARGA 1.
This was the first data point recorded by the MARGA after a 15 hour power outage. With
this point included slope is 1.00, intercept is 1.0, and r2 is 0.31 for MARGA 1.)
26
-------
shows scatter plots of the hourly MARGA readings versus hourly average readings from the
reference continuous gas monitors.
25
20 -
|15H
o
MARGA 1 =0.79x + 0.36
^ = 0.88
MARGA2=0.78x + 0.34
r2 = 0.86 ,
1.5 -
MARGA 1 =0.56x + 0.09
r2 = 0.18
MARGA°2 = 0.47x + 0.13
r2 = 0.08/'
Reference SO2 Conc.(|^g/m
Reference NH3 Conc.(|^g/m )
Figure 6-11. Regression Plots of MARGA data versus reference data from the continuous gas
monitors for S02 and NH3.
Table 6-1 presents a summary of the linear regression analysis of these data for each target
analyte. Asterisks in this table indicate that the reference method measurements did not meet
the target performance goals and should not be used as a basis for evaluating the performance
of the MARGA.
Table 6-1. Summary of Regression Analysis Results for the MARGA Systems Relative to
Reference Method Results.
Target Analyte
(reference method)
SO2 (12-hour)a
SO2(l-hour)c
HNO3
NH3 (12-hour)a
NH3(l-hour)c
SO42
NO3
NH4+
MARGA 1
Slope
1.00(1.09b)
0.79
1.53*
0.14
0.56
0.92
0.48*
0.82*
Intercept
M,g/m3
1.00(0.60b)
0.36
-0.07*
0.20
0.09
0.68
0.19*
0.08*
r2
0.31(0.89b)
0.88
0.60*
0.21
0.18
0.91
0.24*
0.67*
MARGA 2
Slope
1.09
0.78
1.51*
0.21
0.47
0.87
0.40*
0.85*
Intercept
|xg/m3
0.57
0.34
-0.01*
0.10
0.13
0.63
0.25*
0.21*
r2
0.90
0.86
0.59*
0.26
0.08
0.89
0.19*
0.68*
a-Based on comparisons to 12-hour denuder/filter pack reference measurements.
b-Values for slope, intercept, and r2 with removal of one outlier that occurred immediately after a power failure.
c-Based on comparisons to 1-hour continuous gas analyzer measurements.
* Duplicate denuder/filter pack reference method results do not meet target performance goals.
27
-------
The target performance goals for semi-continuous ambient air monitoring systems call for a
slope of the regression analysis between 0.80 and 1.20, and an intercept between 0 ± 10 parts
per billion (ppb) for each analyte. Table 6-2 summarizes the performance of the duplicate
MARGAs relative to these goals for each target analyte (note that the intercept values in
Figure 6-10 and Table 6-1 were converted from ng/m3 to ppb for comparison to the target
performance goals). This table also considers the regression results from the duplicate
reference method results as shown in Figure 6-9. A check mark in Table 6-2 means that the
indicated method met the target performance goal for the indicated target analyte. Table 6-2
shows that whereas all target goals for regression intercept were met by both the reference
methods and the MARGAs, the goals for regression slope were met by both reference and
MARGA results only for 862 and SC>42~. In addition, the denuder/filter pack results met the
slope goal for NHa, and the MARGAs met the slope goal for NH4+. When the reference
method results do not meet the performance goals, it is not appropriate to use them as a basis
of comparison for the MARGA.
Table 6-2. Summary of Denuder/Filter Pack Reference Method and MARGA Regression
Analysis Results versus Target Performance Goals
Target Analyte
SO2
HNO3
NH3
SO42"
NO3-
NH4+
Slope
Reference MARGA 1 MARGA 2
V S S
000
•/ o o
,/,/,/
000
o •/ •/
Intercept
Reference MARGA 1 MARGA 2
V S S
s s s
s s s
s s s
s s s
s s s
NA - Not applicable. Only a single continuous analyzer was used.
S indicates that target performance goal was met.
o indicates that target performance goal was not met.
6.1.2 MARPD Analysis
Table 6-3 presents a summary of the calculated MARPD results and whether the results meet
the target accuracy goals for each target analyte for each of the duplicate MARGA systems.
The MARPD of the paired denuder/filter pack reference data has been included to show the
precision of the reference data. In all cases except for SC>42~, the duplicate reference method
measurements failed to meet the precision DQO of < 20%. In most cases this was likely
because of the very low and narrow range of concentrations of the target analytes during the
testing period. Because the reference method failed to meet the testing DQO, comparisons of
the MARGA data to the reference data should not be considered as conclusive evidence of
MARGA performance. Nonetheless, in the one instance where the reference data did meet
the DQO, both MARGA units met the target accuracy goal (for SO42")- Both MARGA units
also met the target goals in one other case where the reference method failed to meet the
testing DQO (for NH4+).
28
-------
Table 6-3. Summary of Calculated MARPD Results for Reference Data and MARGA Systems
Target
Analyte
SO2
HNO3
NH3
SO42
NO3
NH4+
Reference Method*
MARPD
(DQO < 20%)
21%
22%
21%
15%
56%
32%
MARGA 1
MARPD
59%
42%
41%
35%
66%
24%
Target Goal
o
o
o
S
o
S
MARGA 2
MARPD
53%
43%
64%
35%
69%
29%
Target Goal
o
o
o
s
o
s
S indicates that target performance goal was met.
o indicates that target performance goal was not met.
* Reference method data (other than SO42~) did not meet the data quality objectives (DQOs) set forth for this
evaluation (MARPD < 20%). Comparisons of reference method data to MARGA data are presented, but it
should be noted that the reference data did not meet the DQOs.
6.2 Precision
Precision was assessed in two ways as described in Section 5.2. Firstly, the MARPD of
paired measurements from the duplicate MARGA units was calculated for each of the target
analytes, when both measurements exceeded twice the detection limit for the respective
analyte. Table 6-4 presents a summary of the resulting MARPD results for the duplicate
MARGAs. This table also presents a summary of the number of data points for each analyte
where both MARGA results exceeded twice the detection limit as well as the number of
points below twice the detection limit for each MARGA. For all analytes, the duplicate
MARGAs met the target precision goal of MARPD < 25%, with the MARPD values ranging
from 5% for SO2 to 20% for NH4+.
Table 6-4. Summary of Calculated MARPD Results for Duplicate MARGAs.
Target
Analyte
SO2
HNO3
NH3
SO42
NO3
NH4+
MARPD
(Goal < 25%)
5%
12%
18%
9%
8%
20%
Number of Data Points
with Both Monitors above
2xDL
552
341
239
651
368
575
Number of Data Points below 2 x
DL
MARGA 1
59
314
237
1
279
13
MARGA 2
113
296
413
0
253
112
Additionally, the MARPD of the duplicate MARGA results was calculated for each 12-hour
th
reference sampling period and compared to the 95 percentile of the pooled RPD results
29
-------
(RPDgs) of the duplicate denuder/filter pack reference measurements (Table 6-5). The target
performance goal for this measure of precision is for the MARPD of the MARGAs to be less
than the corresponding 95th percentile of the reference data. The MARGAs met this
performance goal for all target analytes, exhibiting better duplicate precision than did the
reference measurements.
Table 6-5. Comparison of MARPD of 12-Hour MARGA Average Measurements with 95th
Percentile of Pooled RPD Results of Duplicate Reference Measurements
Target Analyte
SO2
HNO3
NH3
SO42
NO3
NH4+
RPD95
76%
51%
121%
57%
168%
137%
MARPD (%)
5%
10%
26%
10%
6%
23%
Met Target Goal
S
•/
•/
•/
•/
s
6.3 Data Completeness
The data completeness for the duplicate MARGA systems was calculated in two ways as
described in Section 5.3. Data completeness was calculated both as the average percentage
of valid data collected per day and as the average percentage of valid data collected during
each reference period when detectable levels were observed in both reference method
samples. Data validity, as reported by the MARGA systems, was used to evaluate data
completeness; no external validation procedure was used. Completeness was calculated
independently for each MARGA and for each target analyte. Included in this calculation
were the MARGA analytes Na+, Ca2+, and Cl", which were not included in evaluations for
other performance parameters. Table 6-6 summarizes the results of the data completeness
calculations. All of the completeness results in Table 6-7 exceed the target value of >80%,
except for the valid data per reference period result for Cl" for MARGA 1.
6.4 Reliability
Instrument reliability was assessed in three ways. First, reliability was assessed in terms of
the percentage of time that the monitoring systems operated in measurement mode over the
duration of the test period. The target goal for this metric is >90%. Second, reliability was
assessed in terms of the ability of the instruments to perform a controlled shut-down in the
case of a power failure, followed by an automated return to measurement mode within four
hours after power was restored. For this assessment, the testing staff imposed a temporary
power outage at the test site and monitored the performance of one of the monitoring systems
during and after the power outage. Additionally, the average number of site visits per week
30
-------
Table 6-6. Summary of Data Completeness for MARGAs
Average % of Valid Data per Day Average % of Valid Data per
_ . , (e.g., per 24 hours) Reference Sampling Period (e.g., per
TargetAnalyte ^ 12hours)
SO2
HNO3
NH3
SO42
NO3
NH4+
Na+
Ca2+
cr
MARGA 1
91%
91%
90%
90%
91%
90%
90%
90%
91%
MARGA 2
90%
90%
90%
90%
90%
90%
90%
90%
90%
MARGA 1
94%
85%
93%
94%
95%
93%
94%
NA
40%
MARGA 2
98%
98%
97%
98%
99%
98%
96%
NA
90%
NA - not available, as Ca was never detected in both reference method samples for a given sampling period.
that were required to keep the MARGA units operating was recorded; the target goal is <2
visits per week. These assessments are reported independently for the duplicate monitoring
systems in Table 6-7, which shows that the MARGA units met all reliability goals. Note that
MARGA startup after a power outage can occur in less than one hour, provided the outage
occurs early enough in the instrument's hourly sampling/analysis cycle. However, even if
the outage occurs very late in the current hourly cycle the startup will occur by the end of the
subsequent hour. As a result, the time to startup shown in Table 6-7 (< 2 hours) covers all
outage conditions.
Table 6-7. Summary of MARGA Reliability Assessments
% of Time in Time to Start-up after Average Site Visits
Operating Mode Power Interruption (per week)
MARGA 1 95% NA 1.6
MARGA 2 96% < 2 hours 1.6
NA: Not applicable, power interruption test only conducted for MARGA 2.
6.5 Operational Factors
Table 6-8 presents a summary of the maintenance activities performed on the MARGA
systems during the verification test.
6.5.1 Ease of Use
The MARGA systems were installed by a single representative of Applikon BV prior to
testing. No documentation of the time required for the installation was available, however,
31
-------
the representative of Applikon B V indicated that the installation was completed over the
course of 3-4 days. The MARGA systems operated primarily unattended during the
verification test, and all maintenance activities performed on the MARGA systems were
performed by a representative of Applikon B V.
Table 6-8. Summary of Maintenance Activities Performed on MARGAs During Verification
Testing
Date
10/03/08
10/06/08
10/07/08
10/08/08
10/14/08
10/15/08
10/21/08
10/22/08
10/23/08
10/24/08
10/28/08
10/29/08
10/30/08
Duration
2.5 hours
1.5 hours
10 minutes
5 minutes
10 minutes
5 minutes
10 minutes
5 minutes
10 minutes
20 minutes
5 minutes
10 minutes
10 minutes
Activity
Replacement of wetted rotating denuder on
MARGA1, alignment screw adjustment, tubing
length on new denuder shortened.
Check syringe pump operation on MARGA 1.
Add solution to anion eluent, cation eluent, and
absorbing solution reservoirs. Empty waste
container.
Filter exchange on MARGA 1 and MARGA2.
Add solution to anion eluent, cation eluent, and
absorbing solution reservoirs. Empty waste
container.
Filter exchange on MARGA 1 and MARGA2.
Add solution to anion eluent, cation eluent, and
absorbing solution reservoirs. Empty waste
container.
Filter exchange on MARGA 1 and MARGA2
Instruments put back in measurement mode after a
power outage at the site.
Remove air from denuder syringe pumps on
MARGA1.
Add absorbing solution to MARGA1.
Reboot both computers to restart data writing to
files.
Add solution to anion eluent, cation eluent,
suppressor regenerant, and absorbing solution
reservoirs. Empty waste container. Filter exchange
on MARGA1 and MARGA2.
Down Time
None
3 hours
None
None
None
None
None
None
15 hours
1 1 hours
None
None
None
6.5.2 Maintenance
Routine maintenance consisted of preparation and changeout of cation and anion 1C eluents,
absorbing solution, internal standard, and suppressor regenerant. Cation eluent, anion eluent,
and absorbing solution were refilled on a weekly basis. The suppressor regenerant and
internal standard were prepared once at the beginning of the study and lasted for the entire 30
day duration. Additionally, particle filters downstream of the aerosol collection device were
changed on a weekly basis. The Applikon BV representative chose to change these particle
filters on a different day than the solution changeouts, but those activities would generally be
performed in one site visit. Other maintenance activities included replacement of the wetted
rotating denuder, syringe pump maintenance, reboot of instrument software, and maintenance
related to an unplanned power failure during the study.
32
-------
6.5.3 Consumables/Waste Generation
During the verification test, the MARGA systems required the use of several standard
consumable materials. The consumables that were used included absorbing solution
(deionized water), internal lithium bromide standard, suppressor regenerant (sulfuric acid),
cation eluent (sodium carbonate and sodium bicarbonate solution), anion eluent (nitric acid)
and particle filters. The eluents were consumed at a rate of approximately 5 L per week for
each instrument. Absorbing solution was consumed at a rate of approximately 15 L per week
for each instrument. Internal standard and suppressor regenerant were consumed at
approximately 1 L per week for each instrument. Approximately 20 L of waste were
generated by each instrument in one week. Waste was emptied once per week. All wastes
are considered non-hazardous and do not require any special treatment for disposal.
33
-------
Chapter 7
Performance Summary
Table 7-1 presents a summary of the results of the verification of the MARGA systems
during this verification test. It is important to note that the reference method data for this
verification test failed to meet some of the target performance goals specified by EPA in
consideration of the monitoring needs for CASTNET. Specifically, the reference data for
HNOs, NCV, and NH4+ failed to meet the target performance goals for accuracy based on a
regression analysis. Also, the duplicate reference method data failed to meet the data quality
objectives of this verification test for MARPD for all target analytes except SC>42~. In those
cases, the data are annotated in Table 7-1 and should not be considered to be of sufficient
quality to allow a definitive assessment of the MARGA. Bolded entries indicate that the
target performance goal was met.
Table 7-1. Summary of Verification Test Results for the MARGA
Parameter
Evaluated
Accuracy
Accuracy
Method of
Evaluation
Regression analysis
comparison to
reference samples
Calculation of
MARPD between
MARGA results
and reference
method results
Results
Analyte
SO2
HNO3
NH3
SO42
NO3
NH/
MARGA 1
Slope
1.00 (1.09a)
1.53*
0.14
0.92
0.48*
0.82*
Analyte
SO2
HNO3 **
NH3 **
SO42
NO3
NH4+**
Intercept
(ug/m3)
1.00 (0.60a)
-0.07*
0.20
0.68
0.19*
0.08*
MARGA 2
Slope
1.09
1.51*
0.21
0.87
0.40*
0.85*
Intercept
(u.g/m3)
0.57
-0.01*
0.10
0.63
0.25*
0.21*
MARPD
MARGA 1 MARGA 2
59% 53%
42% 43%
41% 64%
35% 35%
66% 69%
24% 29%
34
-------
Table 7-1. (Continued)
Parameter
Evaluated
Precision
Precision
Data completeness
Reliability
Reliability
Reliability
Method of
Evaluation
Comparison of results
from duplicate
monitoring systems
Comparison of
MARPD of 12-hour
average MARGA data
and 95th percentile of
pooled RPD results
from reference
measurements
(RPD95)
Ratio of number of
samples successfully
collected to number of
potential samples that
could have been
collected
Percentage of time in
operating mode
Time to start-up after
power interruption
Number of site visits
1
per week
Results
Analyte
SO2
HNO3
NH3
SO
2_
4
NO3
NH4+
Analyte
SO2
HNO3
NH3
SO42
NO3
NH4+
Target
Analyte
SO2
HNO3
NH3
SO42
NO3
NH4+
Na+
Ca2+
cr
Ref. Method
RPD95
76%
51%
121%
57%
168%
137%
Average % of Valid
Data per Day (e.g., per
24 hours)
MARGA
1
91%
91%
90%
90%
91%
90%
90%
90%
91%
MARGA
2
90%
90%
90%
90%
90%
90%
90%
90%
90%
1-Hour MARPD
5%
12%
18%
9%
8%
20%
MARGA 12-Hour
MARPD
5%
10%
26%
10%
6%
23%
Average % of Valid
Data per Reference
Sampling Period (e.g.,
per 12 hours)
MARGA
1
94%
85%
93%
94%
95%
93%
94%
NA
40%
MARGA
2
98%
98%
97%
98%
99%
98%
96%
NA
90%
MARGA 1 : 95% MARGA 2: 96%
MARGA 1 : Not tested MARGA 2: < 2 hours
MARGA 1
: 1.6 MARGA 2: 1.6
35
-------
Table 7-1. (Continued)
Ease of use
Operator observations
Routine operations of the instrument were generally
easy with the only regularly scheduled tasks being
solution preparation and changing, and filter
replacement
Installation performed by Applikon over 3-4 days (not
independently observed as part of this verification)
Maintenance
Operator observations
Routine maintenance consists of preparing and
changing/refilling solutions and replacement of
particle filters
Non-routine maintenance observed included wet
rotating denuder replacement, syringe pump
maintenance, and PC reboot to restart data acquisition
Consumable s/waste
generated
Operator observations
• Cation and anion eluents, and absorbing solution
refilled weekly
• Supressor regenerant and internal standard refilled
monthly
• Internal filters replaced weekly
• Waste emptied weekly
* Duplicate denuder/filter pack reference method results do not meet target performance goals for the
MARGA and should not be used for performance evaluation.
** Reference method data (other than SO42~) did not meet the data quality objectives (DQOs) set forth for this
evaluation (MARPD < 20%). Comparisons of reference method data to MARGA data are presented, but it
should be noted that the reference data did not meet the DQOs.
a Values for slope and intercept with removal of one outlier that occurred immediately after a power failure.
36
-------
Chapter 8
References
1. Allegrini, I, De Santis, F., Di Palo, V., Febo, A., C.,Possanzini, M., and Liberti, A. 1987.
"Annular Denuder Method for Sampling Reactive Gases and Aerosols in the
Atmosphere." Sci. Total Environment. Vol. 67, 1-16.
2. Harrison, R.M., Kitto, A.M.N. 1990. "Field Intercomparison of Filter Pack and Denuder
Sampling Methods for Reactive Gaseous and Particulate Pollutants", Atmos.
Environment 24A(10):2633-2640.
3. Sickles, I.E., Hodson, L.L., McClenny, W.A., Paur, R.J., Ellestad, T.G 1990. "Field
Comparison of Methods for the Measurement of Gaseous and Particulate Contributors to
Acidic Dry Deposition." Atmos. Environ. 24A(01): 155-165.
4. Test/QA Plan for Verification of Semi-continuous Ambient Air Monitoring Systems,
Battelle, Columbus, Ohio, September 26, 2008.
5. EPA, Compendium Method IO-4.2 Determination of Reactive Acidic and Basic Gases
and Strong Acidity of Atmospheric Particles (< 2.5 |j,m). Center for Environmental
Research Information, Office of Research and Development, Cincinnati, OH. June,
1999.
6. EPA, Method 300.0 Determination of Inorganic Anions by Chromatography.
Environmental Monitoring Systems Laboratory, Office of Research and Development,
Cincinnati, OH. August, 1993.
7. EPA, Method 350.1 Determination of Ammonia Nitrogen by Semi-Automated
Colorimetry. Environmental Monitoring Systems Laboratory, Office of Research and
Development, Cincinnati, OH. August, 1993.
37
-------
Battelle, Quality Management Plan for the ETV Advanced Monitoring Systems Center,
Version 6.0, U.S. EPA Environmental Technology Verification Program, prepared by
Battelle, Columbus, Ohio, November 2005.
38
------- |