September 2005
Environmental Technology
Verification Report

HORIBA INSTRUMENTS, INC.
APSA-360 AMBIENT
HYDROGEN SULFIDE ANALYZER
              Prepared by
               Battelle

              Baltelle
            7/jt> Business oj Innovalion
        Under a cooperative agreement with
          U.S. Environmental Protection Agency

-------
                                     September 2005
Environmental Technology Verification
                  Report
   ETV Advanced Monitoring Systems Center

     HORIBA INSTRUMENTS, INC.
         APSA-360 AMBIENT
    HYDROGEN SULFIDE ANALYZER
                     by
                 Ann Louise Sumner
                  Amy Dindal
                 Zachary Willenberg
                  Karen Riggs
                   Battelle
                 Columbus, Ohio

                 Richard Pfeiffer
                  Jerry Hatfield
            United States Department of Agriculture
              National Soil Tilth Laboratory
                  Ames, Iowa

                     and

                  Eric Winegar
              Applied Measurement Science
                 Fair Oaks, California

-------
                                        Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development, has financially supported and collaborated in the extramural program described
here. This document has been peer reviewed by the Agency. Mention of trade names or
commercial products does not constitute endorsement or recommendation by the EPA for use.
                                          11

-------
                                       Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
nation's air, water, and land 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, the EPA's Office of Research and Development provides data and science support that
can be used to solve environmental  problems and to build the scientific knowledge base needed
to manage our ecological resources  wisely, to understand how pollutants affect our health, and to
prevent or reduce environmental risks.

The Environmental Technology Verification (ETV) Program has been established by the EPA to
verify the performance characteristics of innovative environmental technology across all media
and to report this objective information to permitters, buyers, and users of the technology, thus
substantially accelerating the entrance of new environmental technologies into the marketplace.
Verification organizations oversee and report verification activities based on testing and quality
assurance protocols developed with input from major stakeholders and customer groups
associated with the technology area. ETV consists of six verification centers. Information about
each of these centers can be found on the Internet  at http://www.epa.gov/etv/.

Effective verifications of monitoring technologies  are needed to assess environmental quality
and to supply cost and performance data to select the most appropriate technology for that
assessment. Under a cooperative agreement, Battelle has received EPA funding to plan,
coordinate, and conduct  such verification tests for "Advanced Monitoring Systems for Air,
Water, and  Soil" and report the results to the community at large. Information concerning this
specific environmental technology area can be found on the Internet at
http ://www. epa.gov/etv/centers/centerl. html.
                                            in

-------
                                  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 would like to thank
Dr. Raul Dominguez, Jr., South Coast Air Quality Management District; Dr. D. Bruce Harris,
EPA; and Dr. William Ollison, American Petroleum Institute, for their technical review of the
test/quality assurance plan and for their careful review of this verification report. We also thank
Dr. Gary Norris, EPA, for his review of the test/quality assurance plan and Dr. Alan Vette, EPA,
for review of this report. We also would like to thank Mr. Kenwood Scoggin and
Dr. Steven Trabue, U.S. Department of Agriculture (USD A), for their assistance in performing
reference sample analysis, and Mr. Timothy Hart, USD A, for providing meteorological data and
programming the data logger used to collect the APSA-360 data. We thank the American
Petroleum Institute for providing funding to Applied Measurement Science for their participation
in this verification test and the USDA Agricultural Research Service for providing logistical
support.
                                           IV

-------
                                        Contents
                                                                                     Page


Notice	ii

Foreword	  iii

Acknowledgments	iv

List of Abbreviations	  viii

1 Background	 1

2 Technology Description	2

3 Test Design and Procedures	4
     3.1  Introduction	4
     3.2  Site Description	4
     3.3  Test Design	 5
          3.3.1   Accuracy, Bias, Precision, and Linearity  	 8
          3.3.2    Span and Zero Drift	 8
          3.3.3   Response Time  	 9
          3.3.4   Interference Effects 	 9
          3.3.5   Comparability  	 9
          3.3.6   Data Completeness 	  12
          3.3.7   Operational Factors 	  12

4 Quality Assurance/Quality Control  	  13
     4.1  Reference Method Quality Control Results 	  13
          4.1.1   Time-Integrated Reference Method Quality Control Results 	  13
          4.1.2   In Situ Reference Method Quality Control Results	  14
     4.2  Audits	  15
          4.2.1   Performance Evaluation Audits	  15
          4.2.2   Technical Systems Audits 	  15
          4.2.3   Audit of Data Quality  	  16
     4.3  Quality Assurance/Quality Control Reporting  	  16
     4.4  Data Review  	  16

5 Statistical Methods and Reported Parameters  	  18
     5.1  Accuracy	  18
     5.2  Bias	  18
     5.3  Precision	  18
     5.4  Linearity 	  19
     5.5  Span and Zero Drift	  19
     5.6  Response Time  	  19
     5.7  Interference Effects	20

-------
     5.8 Comparability	20
     5.9  Data Completeness 	21

6 Test Results	22
     6.1  Accuracy	24
     6.2  Bias	27
     6.3  Precision	27
     6.4  Linearity 	27
     6.5  Span and Zero Drift	29
     6.6  Response Time  	  30
     6.7  Interference Effects	  33
     6.8  Comparability 	  33
          6.8.1   Time-Integrated Comparability  	  34
          6.8.2   In Situ Comparability	  34
     6.9  Data Completeness 	  35
    6.10  Operational Factors	  35

7 Performance Summary 	  38

8 References  	40

Appendix A.  APSA-360 Checklist	A-l

Appendix B.  APSA-360 and Time-Integrated Reference Method Data  	 B-l

Appendix C.  APSA-360 and In Situ Comparability Data	 C-l


                                        Figures

Figure 2-1.  Horiba Instruments, Inc., APSA-360 Ambient Hydrogen Sulfide Analyzer	2

Figure 3-1.  Test Site	 5

Figure 3-2.  Swine Barn Interior	 5

Figure 3-3.  Test Site Lagoons	 6

Figure 3-4.  Teflon Manifold	 6

Figure 6-1.  Meteorological Conditions and APSA-360 Ambient H2S Measurements	23

Figure 6-2.  Hourly Averaged APSA-360 H2S Measurements Plotted as a
           Function of Wind Direction	24

Figure 6-3.  APSA-360 Accuracy Results	25

Figure 6-4.  APSA-360 Linearity Results   	28
                                          VI

-------
Figure 6-5. APSA-360 Baseline Response Results	29

Figure 6-6. Span and Zero Drift Control Chart	 32

Figure 6-7. Comparison of Time-Integrated Reference Measurements with
           Averages from the APSA-360  	 35

Figure 6-8. Comparison of Selected In Situ Reference Measurements with
           APSA-360 Averages and Measurement Data	 36

Figure 6-9. Scatter Plot of APSA-360 Results versus In Situ Reference Measurements	36


                                         Tables

Table 3-1.  Test Activities 	 7

Table 3-2.  H2S Concentrations and Order for Multipoint Challenges  	 8

Table 3-3.  Interferants and Approximate Concentrations for Interference Checks	 10

Table 4-1.  Reference Method Quality Control Requirements and
          Target Acceptance Criteria	 14

Table 4-2.  Summary of Data Recording Process	 17

Table 6-1.  Accuracy Results	26

Table 6-2.  Calculated Precision of the APSA-360  	27

Table 6-3.  Span and Zero Baseline Response	 30

Table 6-4.  Span and Zero Drift Check Results 	 31

Table 6-5.  Response Time Determinations	 32

Table 6-6.  Interference Effect Evaluation	 33
                                          vn

-------
                               List of Abbreviations
AFO              animal feeding operation
AMS             Advanced Monitoring Systems
ASTM            American Society for Testing and Materials
CCV             continuing calibration verification
CI                confidence interval
EPA              U.S. Environmental Protection Agency
ETV              Environmental Technology Verification
GC               gas chromatography
H2S               hydrogen sulfide
i.d.               internal diameter
Lpm              liter per minute
m                meter
mm               millimeter
NIST             National Institute of Standards and Technology
In                natural logarithm
PE               performance evaluation
PFPD             pulsed flame photometric detection
pg                picogram
ppb               part per billion
ppm              part per million
%D               percent difference
%R               percent recovery
QA               quality assurance
QC               quality control
QCS              quality control sample
QMP             quality management plan
RSD              relative standard deviation
sec               standard cubic centimeter
seem             standard cubic centimeter per minute
SD               standard deviation
SO2               sulfur dioxide
TSA              technical  systems audit
UHP              ultra-high purity
USD A            U.S. Department of Agriculture
                                         Vlll

-------
                                      Chapter 1
                                      Background
The U.S. Environmental Protection Agency (EPA) supports the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative environmental tech-
nologies 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 tech-
nologies 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
(QA) protocols to ensure that data of known and adequate quality are generated and that the
results are defensible.

The EPA's National Exposure Research Laboratory and its verification organization partner,
Battelle, operate the Advanced Monitoring Systems (AMS) Center under ETV. The AMS
Center, in collaboration with the U.S. Department of Agriculture's (USDA's) National Soil Tilth
Laboratory and Applied Measurement Science, recently evaluated the performance of the Horiba
Instruments, Inc., APSA-360 ambient hydrogen sulfide (H2S) analyzer in quantifying H2S in
ambient air at a swine  finishing farm.

-------
                                      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 verification report provides
results for the verification testing of the APSA-360. Following is a description of the APSA-360,
based on information provided by the vendor. The information provided below was not verified in
this test.

The APSA-360 continuously measures the concentration of H2S (including other sulfide
compounds) in ambient air using an H2S converter and sulfur dioxide (SO2) ultraviolet
luminescence as the measurement principle.  A hydrocarbon reduction membrane eliminates
hydrocarbon interference in the sample gas. The APSA-360 can be configured to measure SO2
and/or H2S by switching measuring lines into and out of the built-in H2S converter at regular
intervals or with the measuring line fixed to SO2 or H2S. The APSA-360 verified in this test was
configured to  measure only H2S. The basic system can be operated by controls on the front panel
when it is connected to a calibration gas, but it can also be upgraded for remote monitoring by
adding a computer, a controller, and a recorder using the APSA-360 AP-Remote software for
Microsoft Windows. The APSA-360 can be calibrated automatically or manually and has a lower
detection limit of 4 parts per billion (ppb).

                                         Data logged by the AP-Remote software can be
                                         exported into Microsoft Excel. The APSA-360 has
                                         internal storage for up to several weeks of data
                                         depending on the sample rate.  The data are
                                         accessible by the front panel or the AP-Remote
                                         software. Data may also be recorded by an external
                                         data logger that is connected to the analog and
                                         digital outputs of the APSA-360. As configured for
                                         the verification test, the 4-20 mA instantaneous
                                         H2S reading was output to an analog input channel
                                         of a Campbell Scientific Model CR43 data logger
                                         that was  made available by the USDA. The data
                                         logger program sampled the APSA-360 signal
                                         every ten seconds and recorded one-minute
Figure 2-1. Horiba Instruments, Inc.,
APSA-360 Ambient Hydrogen Sulfide
Analyzer
averages calculated from six instantaneous readings. The external data logger was used for the
verification test because the AP-Remote software at the time was operational only for the Horiba
ambient carbon monoxide, nitrogen oxides, SO2 (only Model APSA-360 CE), ozone, and total
hydrocarbon analyzers. In the future, AP-Remote will be  available for the APSA-360 H2S
Analyzer.

-------
The APSA-360 weighs 25 kilograms (55 pounds); it is 221 millimeters (mm, 8.7 inches) high,
430 mm (17 inches) wide, and 550 mm (22.7 inches) deep (excluding front and rear extrusions).
The list price of the APSA-360 H2S-only analyzer is approximately $18,000. The APSA-360 that
alternately measures SO2 and H2S is list priced at approximately $24,000.

-------
                                      Chapter 3
                             Test Design and Procedures
3.1 Introduction
H2S is formed at animal feeding operations (AFOs) during the bacterial decomposition of sulfur-
containing organic compounds present in manure. Also known as a component of sewer gas, H2S
has the characteristic odor of rotten eggs and, at high levels [greater than 500 parts per million
(ppm)], can cause death from even brief exposure. As a result, H2S analyzers were identified as a
priority technology category through the AMS Center stakeholder process.

This verification test was conducted according to procedures specified in the Test/QA Plan for
Verification of Ambient Hydrogen Sulfide Analyzers at a Swine Finishing Faring with the
exception of three deviations that are addressed later in this report. The testing was conducted at
a large swine finishing farm near Ames, Iowa. This verification test was conducted for six weeks
between April 25 and June 3, 2005. As discussed in Section 3.3, the APSA-360 was not installed
at the test site until May 16, 2005. Testing was conducted on the APSA-360 between May 17
and June 3, 2005, during which time the APSA-360 continuously measured H2S concentrations
in ambient air or synthetic air samples of known concentration ("standards"). The performance
of the APSA-360 was evaluated in terms of

•  Accuracy
•  Bias
•  Precision
•  Linearity
•  Span and zero drift
•  Response  time
•  Interference effects
•  Comparability
•  Data completeness
•  Operational factors.
3.2 Site Description

The layout of the swine finishing farm is shown in Figure 3-1. The farm had 10 animal barns,
arranged in two parallel rows of five, with each barn housing up to 2,000 swine. Figure 3-2
shows the interior of a swine barn; natural ventilation was regulated by raising or lowering
curtains, shown in the foreground. The urine and feces from the swine leave the barns through
wood slats in the floor and are flushed through underground piping into a nutrient lagoon

-------
located on the southern end of the farm;
supernatant liquid from the primary lagoon
is pumped into a secondary storage lagoon
and used to fertilize nearby fields. The
primary H2S source was expected to be the
lagoons. The perimeter of the farm is lined
with trees, and agricultural fields surround
the perimeter.  A temperature-regulated
instrument trailer was placed on-site during
the test to house the monitoring equipment
and to provide a sheltered work space.
Figure 3-3 shows the test site as
photographed  from the south of the
lagoons, showing the instrument trailer and
swine barns in the background. The APSA-
360 was installed inside the instrument
trailer, and a Teflon inlet line (sampling
ambient air through the east window) was
connected to a Teflon manifold and was
used to sample ambient air. The Teflon
inlet line was protected from rain by an
inverted funnel. The Teflon manifold used
for supplying ambient air and gas standards
to the APSA-360 is shown in Figure 3-4.
Sample tubing lengths were minimized
both for ambient air sampling and for
delivery of gas standards.
   s
   s
   $
   §
   9
   5
 N$
                                      Entrance
                         Trailer
   9
   9
   9
   9
   9
Primary
Nutrient
Lagoon
Secondary
 Storage
  Lagoon
Figure 3-1. Test Site
                                                               3.3 Test Design

                                                               Table 3-1  shows the
                                                               activities involved in
                                                               preparing for and
                                                               conducting the verification
                                                               test.

                                                               The APSA-360 evaluated
                                                               during this verification test
                                                               was manufactured in
                                                               Germany  in July 2004 and
                                                               shipped to Horiba in
                                                               California on April 15,
                                                               2005. Delays were
                                                               encountered in shipping the
                                                               APSA-360 to the United
                                                               States and clearing
                                                               customs. To accommodate
Figure 3-2. Swine Barn Interior

-------
Figure 3-3. Test Site Lagoons


these delays, the verification test start date (originally April 18, 2005)
was delayed by one week, and the APSA-360 was first installed at the
test site on April 22, 2005, by a vendor representative with assistance
from Battelle and USDA staff. Trial operations revealed that the
APSA-360 was not performing up to the vendor representative's
expectations.  After Horiba, Battelle, and USDA staff examined the
APSA-360 for possible malfunctions for several days, it was decided on
April 25, 2005, that the APSA-360 should be shipped to the California
Horiba office for inspection and repair, if needed. The APSA-360 was
repaired by Horiba staff and returned to USDA on May 16, 2005.
USDA staff installed the APSA-360 at the field site the same day and
began testing the APSA-360 on Tuesday, May 17, 2005. A Campbell
Scientific Model CR43 data logger, provided by USDA, was used to
collect the APSA-360  data from the analog output. Data were output by
the APSA-360 using the 4- to 20-milliamp range; a resistor was used to
produce a voltage signal, which was collected by the data logger.

Battelle and USDA staff worked with the vendor representative to
establish procedures for operating the APSA-360 during this
verification test. The vendor representative trained Battelle and USDA
staff to check several instrument parameters to verify the operation of
the APSA-360 and identify signs of malfunction. A checklist, provided
by the vendor representative and included as Appendix A, was
completed daily (Monday through Friday) by Battelle or USDA staff. In
general, Battelle or USDA staff verified that the APSA-360 power was
on, checked for alarms, and downloaded the APSA-360 data from the
data logger (recorded as one-minute averages of instantaneous data
logged every  10 seconds) on a daily basis. In the event  of an instrument
malfunction, Battelle and/or USDA staff could contact  the vendor
representative and conduct minor troubleshooting procedures as
necessary, but were not expected to make any major repairs. All the
testing activities, which are described in the following sections, were
conducted by Battelle and/or USDA staff.
Figure 3-4. Teflon
Manifold

-------
Table 3-1. Test Activities
Week of
April 11
April 18
April 25
(Testing Week 1)
May 2
(Testing Week 2)
May 9
(Testing Week 3)
May 16
(Testing Week 4)
May 23
(Testing Week 5)
May 30
(Testing Week 6)
Activities
• Testing preparations by USDA and Battelle staff
• Initial installation of APSA-360
Training of USDA and Battelle staff by vendor representative
Conduct trial operations
• Troubleshoot APSA-360
• Ship APSA-360 to Horiba (Irvine, California) for inspection and repair
Ongoing testing activities, excluding APSA-360
Three time-integrated reference samples collected and analyzed
• Ongoing testing activities, excluding APSA-360
Three time-integrated reference samples collected and analyzed
• Install in situ reference method instrumentation at test site
• Ongoing testing activities, excluding APSA-360
Five time-integrated reference samples collected and analyzed
• Install APSA-360
Zero air/H2S standard challenge for analyzer response (baseline) and
analyzer response time
• Two zero/span checks
Four time-integrated reference samples collected and analyzed
• Multipoint H2S standard challenges for accuracy, bias, precision, linearity
• Routine operation
Two zero/span checks
• Multipoint H2S standard challenges for accuracy, bias, precision, linearity
Gas standard challenges for interference check
• Troubleshoot in situ reference method instrumentation
• Begin in situ reference measurements
Routine operation
Three zero/span checks
• Continue in situ reference method measurements
• Demobilize in situ reference method instrumentation
• Remove APSA-360 from test site
Individual data files from the data logger (comma-delimited text), containing the day of year,
time, and voltage readings, were opened in Microsoft Excel. The APSA-360 voltage readings
were converted to H2S concentrations using a calibration produced from a linear regression of
APSA-360 panel readings (in ppb) versus the voltage signal (y = 1.68x - 2.05). This calibration
was applied to all APSA-360 data collected during this verification test. The resulting H2S
concentration data were analyzed using the procedures outlined in Chapter 5 of this report. The
final data file containing the full APSA-360 data set from this verification test was less than
700 kilobytes.

Gas standard dilutions were supplied to the APSA-360 during testing activities for 20 minutes
using a programmable dilution system (Environics Series 4040, with silanized internal
components) that supplied each mixture to the Teflon manifold at flow rates at least 1  liter per
minute (Lpm) in excess of the APSA-360 sampling flow rate (approximately 0.8 Lpm). The data
logger program sampled the APSA-360 signal every 10 seconds and recorded one-minute

-------
averages. The average APSA-360 response to each gas standard was calculated from the last
5 minutes of data from each delivery period (5 data points). The last five minutes were used
because the APSA-360 response appeared to be stable during that period (i.e., a general increase
or decrease in the response was not apparent). These average APSA-360 response values were
used in the calculations described in Chapter 5 of this report.

The APSA-360 H2S readings when sampling ambient air were compared to concurrent
measurements by two H2S reference methods. For comparison with the time-integrated reference
method  (described in Section 3.3.5.1), the APSA-360 H2S readings were averaged to the same
time period over which the reference method samples were collected (approximately 7.5 hours).
For comparison with the in situ reference method (described in Section 3.3.5.2), APSA-360
readings were averaged over 15-minute periods, centered on the in situ reference method sample
times. The performance results of the APSA-360 during this verification test are presented in
Chapter 6 of this report and summarized in Chapter 7.

3.3.1 Accuracy, Bias, Precision, and Linearity

During Week 4 and Week 5 of the verification test, the APSA-360 was challenged with a
certified compressed H2S gas standard (5.12 ppm, H2S  Scott Specialty Gases) diluted in zero air
to achieve measurements over a range of concentrations from approximately 0 to 300 ppb. Three
non-consecutive measurements were recorded at each of five nominal concentration levels. Each
concentration was supplied to the APSA-360 for 20 minutes. Table 3-2 shows the nominal H2S
concentrations supplied to the APSA-360 and the order in which they were supplied. As
Table 3-2 indicates, the H2S concentrations were supplied to the APSA-360 in increasing order,
then in random order, and finally in decreasing order. After the last measurement was recorded,
the APSA-360 was returned to sampling ambient air.

Table 3-2. H2S Concentrations and Order for Multipoint Challenges
Concentration
Measurement
Number
Oppb
1
7
15
30 ppb
2
10
14
90 ppb
3
6
13
150 ppb
4
9
12
300 ppb
5
8
11
The APSA-360 response to the series of H2S gas standards was used to evaluate accuracy, bias,
precision, and linearity. The statistical procedures used are presented in Section 5. Accuracy was
calculated at each concentration and for each replicate relative to the nominal H2S concentration.
Bias was calculated for each series of multipoint H2S challenges. The APSA-360 precision was
demonstrated by the reproducibility of the average APSA-360 response at each nominal H2S
concentration. Linearity was assessed by establishing a multipoint calibration curve from the
APSA-360 response.

3.3.2  Span and Zero Drift

The baseline response of the APSA-360 to zero air and a 30-ppb dilution of a compressed H2S
gas standard was determined on the first day it was tested (Week 4 of the verification test). The
APSA-360 was challenged alternately with the diluted H2S gas standard and  zero air, for a total

-------
of five replicates of both the gas standard and zero air. Each gas was supplied sequentially for
20 minutes and the average response calculated for each replicate using data from the last
5 minutes of each delivery period. The overall average and standard deviation of the APSA-360
response to zero air and to the 30-ppb H2S standard were calculated from the average response
for the five replicates. Control charts showing the [±2 standard deviation (SD)] warning and
(±3 SD) action limits were constructed for the span and zero response for use in evaluating drift.

At least twice each week, zero air and a 30-ppb H2S standard were supplied to the APSA-360 for
20 minutes for a total of seven zero/span checks. The gas standard dilution system was not
flushed with the H2S gas standard before performing two of the span checks. Thus,  the results of
five span drift checks were used to evaluate span drift. Each response was compared to the
baseline response to determine whether drift occurred in the APSA-360 sensitivity to zero air or
the 30-ppb H2S standard.

3.3.3  Response Time

The data collected during the zero/span baseline response checks were used to determine the
APSA-360  response time. The 95% rise time was calculated for changes from zero  air to the
30-ppb H2S standard, and the 95% fall time was calculated for changes from the 30-ppb standard
to zero air. A minimum of three individual measurements was used to determine the average rise
and fall times.

3.3.4  Interference Effects

The APSA-360 was challenged with a series of gases (Table 3-3) that may be present at an AFO
and could interfere with the APSA-360 response to H2S. Each interferant was supplied at either
100 or 500  ppb, as listed in Table 3-3, in the presence and absence of 100 ppb of H2S. A 100-ppb
H2S standard was supplied to the APSA-360 for 20 minutes, and the responses were recorded.
The APSA-360 was then supplied with zero air for five minutes. The first interferant was diluted
with zero air and delivered to the APSA-360 for 20 minutes. After the responses were recorded,
the APSA-360 was supplied with zero air for five minutes. A mixture of the first interferant
(SO2) at 100 ppb with 100-ppb H2S  in zero air was supplied to the APSA-360 for 20 minutes.
The APSA-360 responses were recorded, and zero air was  supplied to the APSA-360 for
approximately five minutes.  This process was repeated for each interferant at the concentrations
listed  in Table 3-3.

3.3.5  Comparability

The comparability of the APSA-360 response to ambient air was evaluated by comparing its
response to two H2S reference methods (time-integrated and in situ), which were carried out by
USDA and Applied Measurement Science. The two reference methods were based  on American
Society for  Testing and Materials (ASTM) Method D5504-01,(2) with the following  substitution:
pulsed flame photometric detection (PFPD) was used instead of sulfur chemiluminescence
detection. Reference H2S measurements in ambient air were conducted using gas chroma-
tography (GC) with PFPD using two sample collection techniques. Although the analytical
approach of the two methods was the same, they differed in sample collection and handling. The

-------
Table 3-3. Interferants and Approximate Concentrations for Interference Checks

                         T  ,  ,    ,                  Approximate
                         Interferant               „      ,   .    ,  ,,
                                                 Concentration (ppb)
                Sulfur dioxide                             100
                Carbonyl sulfide                            100
                Carbon disulfide                            100
                Methyl mercaptan                          100
                Dimethyl sulfide                            100
                Hydrocarbon blend                      ,„„ ,.  . 1X
                ,  .  .     cr.^ +  f-.f.  „    x             500 (total)
                (mixture of Cl to C6 alkanes)
                Ammonia                                 500
two reference methods were not conducted simultaneously; therefore the results of the two
methods could not be compared. As discussed in Section 4.1, not all of the QC requirements of
the time-integrated and in situ reference methods were satisfied and, consequently, the quality of
the reference method data was not confirmed. Therefore, in addition to the linear regression
analysis described in the test/QA plan,(1) the reference method data were compared to the APSA-
360 data in a more qualitative manner. The APSA-360 data were compared to the reference
method data to determine whether the measured H2S concentrations were statistically
significantly different at the 95% confidence level and the linear regression analysis was
repeated including only those data that were not significantly different.

3.3.5.1 Time-Integrated Comparability

Time-integrated reference measurements were conducted by collecting ambient air samples over
relatively long periods (up to eight hours) in evacuated 1.4-liter Silonite canisters (Entech
Instruments, Inc.) and were taken to the USDA laboratory for analysis. Ambient air was drawn
into the evacuated canisters from the same Teflon manifold to which the APSA-360 was
connected. The canisters were fitted with a silanized Entech flow controller and pressure gauge
to restrict the air flow to  approximately one to three standard cubic centimeters per minute
(seem), allowing the canisters to fill slowly over approximately eight hours. A performance
evaluation (PE) audit of the canister sampling flow rate revealed that the flow rate varied
between 1.01 and 2.52 seem over 7.5 hours. The variability in the canister sampling flow rate
could result in uneven weighting of the time-integrated air  sample collected in the canister,
potentially resulting in biased results.  Samples were collected during the following time periods:
April 29 to 30, May 4 to 5, May 11 to 13, and May 18 to 21. Up to three samples were collected
over eight-hour intervals on each sampling day according to the following  approximate
schedule: 10:00 p.m. to 6:00 a.m., 6:00 a.m. to 2:00 p.m, and 2:00 p.m. to 10:00 p.m.

The Silonite canisters were cleaned before sample collection using an  Entech 3120a Canister
Cleaning System by heating under vacuum at 120°C, filling with humidified nitrogen, and
evacuating to a pressure  of 50 millitorr. This process was repeated for 50 cycles. Canisters were
then transported to the test site for sampling and returned to the laboratory for analysis. Canisters
were sampled using an Entech 7500 Series Robotic Autosampler, which was connected to an
Entech 7100A Preconcentrator and an Agilent 6890 GC with an OI Analytical PFPD. Canisters

                                            10

-------
were heated to 100°C during sample transfer, and all transfer lines were maintained at 100°C.
Helium carrier gas was used at a flow of 16 sccrn. A sample of known volume [10 to
400 standard cubic centimeters (sec), depending upon the expected concentration] was with-
drawn from the canister and trapped on glass beads at -20°C (the bead trap was subsequently
desorbed at 10°C), and collected on a Tenax® trap at -80°C to reduce water in the system. The
Tenax trap was heated to 180°C, and the desorbed components were cryofocused at - 150°C
before a final heating and transfer to the GC column. The column was a GS-Gaspro, 60 meter
(m) x 0.32  mm inner diameter (i.d.) capillary column (J & W Scientific). The column was held
at 35°C for 0.5 minutes, ramped to 230°C at 12°C per minute, and held at 230°C for the
remainder of the approximately 20-minute run. The test/QA plan(1) stated that samples would be
analyzed within 24 hours of collection. It was not always possible to analyze the canisters within
the 24-hour time frame; in some  cases, samples could not be analyzed until 4 days after
collection because of instrument  availability. The longer holding times may have resulted in H2S
loss in the canisters, and consequently to artificially low H2S reference measurement results.
Sample degradation in the canisters was not verified since a holding time study was not
performed on ambient air samples. The test/QA plan(1) stated that the acceptability of the 24-hour
holding time would be verified on an ambient air sample. A deviation report was filed to address
the holding time issues. A multipoint calibration curve from approximately 150 to 2,300
picograms (pg) for H2S was constructed daily (before reference analyses were conducted) by
injecting several volumes of a diluted H2S compressed gas standard (5.12 ppm H2S, Scott
Specialty Gases) onto the GC-PFPD. Instrument blanks (i.e., zero-volume injections) were
included in each analytical run. Based on the instrument blank results, the quantitation limit
(average blank result plus 10 times the standard deviation of the blank) for a 10-scc injection
was 2.2 ppb.

3.3.5.2 In Situ Comparability

In situ reference measurements were conducted by Applied Measurement Science. The
instrumentation for the in situ method was installed in the instrument trailer at the test site. Air
samples were drawn from a Teflon tube whose inlet was collocated with the Teflon manifold
sampling inlet at a flow rate of approximately 5 Lpm to reduce the residence time of ambient air
in the inlet. Volatile compounds in the samples were cryotrapped, thermally desorbed, and
injected directly onto a Varian 3800 GC with PFPD. The duration of sample collection was
adjusted so that the mass of H2S  was maintained, to the extent possible, within the range of the
PFPD system,  nominally from 30 pg to 3,000 pg per sample. Sample collection times varied
between 6 seconds and 8 minutes. The column was a GS-Gaspro 30-m x 0.32-mm i.d. capillary
column (J & W Scientific). Helium carrier gas was used at a flow of 2 seem. The column was
held at - 10°C for 2 minutes, ramped to 200°C at 40°C per minute, and held at 200°C for the
remainder of the approximately 20-minute run. Multilevel calibrations were performed using the
same  certified H2S gas standard (5.12 ppm H2S, Scott Specialty Gases) and programmable
dilution system used for performing testing activities. In situ reference measurements were
conducted as frequently as possible (usually every 16 minutes) over a four-day period  at the end
of the verification test. Due to technical problems with the reference method air sampling valve
system, measurements could not be conducted over the ten days specified in the test/QA plan.(1)
Of the ambient air measurements conducted by the in situ reference method, 41 of the 53
reference measurements could be used for comparison to the APSA-360 results. The other 12
measurements were presented as upper (result below quantitation limit) or lower (saturated H2S
peak) limits.

                                           11

-------
3.3.6 Data Completeness

Data completeness was assessed based on the overall data return achieved by the APSA-360.

3.3.7 Operational Factors

Operational factors such as maintenance needs, data output, consumables used, ease of use, and
repair requirements were evaluated based on the observations of Battelle and USDA staff.
                                           12

-------
                                      Chapter 4
                         Quality Assurance/Quality Control
QA/quality control (QC) procedures were performed in accordance with the quality management
plan (QMP) for the AMS Center(3) and the test/QA plan for this verification test(1) with the
exception of three deviations, which have been addressed in this report. First, the time-integrated
reference method canister flow rate was lower than expected. This deviation did not impact the
quality of this verification test. The second deviation from the test/QA plan(1) involved the
reference method QC requirements, which were not fully satisfied. Third, the pre-analytical
holding time for ten of the 15 time-integrated reference samples was longer than  24 hours. As
discussed in Section 4.1 and Section 3.3.5.1, the second and third deviations, respectively, did
impact the comparisons that were performed with the reference method data.
4.1 Reference Method Quality Control Results

Table 4-1 summarizes the reference method QC requirements. Both reference methods were
required to analyze continuing calibration verifications (CCV), QC samples (QCS), and field
blanks. The time-integrated H2S reference method was also required to repeat analysis of 10% of
the samples to verify method precision.

4.1.1   Time-Integrated Reference Method Quality Control Results

It was determined that the USDA laboratory GC-PFPD system required calibration each day
before analysis of reference samples. This eliminated the need for running CCV samples, so
there was no expectation for agreement to previous calibration results. QCSs were not run as
frequently as stated in Table 4-1, but often were included at the end of the analysis run.
Approximately half of the analysis  runs had at least one QCS that passed the requirement listed
in Table 4-1. The other half either had failed QCSs or none were included in the run. Replicate
H2S precision was not determined for the same injection volume. However, the results for four
out of 13 comparisons of variable-volume injections from the same sample were within 30% of
one another by percent difference (%D). Measurement accuracy results are discussed in more
detail in Section 4.2.1. Briefly, four performance evaluation (PE) samples were submitted to the
USDA laboratory; reference method results for two of the samples were within the acceptance
criterion for measurement accuracy. The other two results were 38% and undetectable H2S
levels. Finally, two field blank samples were submitted to the USDA laboratory for analysis, and
both resulted in undetectable H2S levels by the GC-PFPD system. Since the QC requirements for
the time-integrated reference method were not satisfied and only two quantitative time-integrated
reference results were available for the period during which the APSA-360 was operational at
                                           13

-------
Table 4-1. Reference Method Quality Control Requirements and Target Acceptance
Criteria
    QC Parameter
            Addressed By
                                          Required Performance
 CCV


 QCS


 Replicate H2S
 precision

 Measurement
 accuracy

 Field blanks
CCV run before analysis of reference
samples each day

QCS run every 4 hours and after analysis
of reference samples each day

Analyze 10% of all samples twice(a)
Analyze H2S standard from independent
source(b)

Analyze canisters filled with zero air
recovered from the test site (weekly)(a)
Analyze zero air passed through sample
manifold (weekly)(c)
                                       %D of CCV result within 30%
                                       of expected value

                                       %D of QCS result within 30%
                                       of expected value

                                       %D within 30% of one another
                                       Results within 30% of expected
                                       value

                                       If blank >30% of sample, H2S,
                                       data must be flagged
                                                    OO
(a) Time-integrated H2S reference method only.
(b:i This standard was provided as part of the PE audit.
^ In situ H2S reference method only.

the field site, the results were not quantitatively compared to the APSA-360 data. The time-
integrated reference method and APSA-360 data were analyzed to determine whether they were
significantly different from each other at the 95% confidence level (see Section 5.8,
Comparability).

4.1.2  In Situ Reference Method Quality Control Results

CCV samples were run each day when the in situ reference method was conducting ambient
measurements. If results were not within 30% of the expected value by %D, a multilevel
calibration curve was generated. At least once daily, a QCS or measurement accuracy sample
was analyzed.  Six QCS samples were analyzed, and all were within 30% of the expected
concentration by %D. QCS samples from a second gas standard (110 ppb H2S, Air Liquide) were
analyzed six times. Two results fell  outside of the calibration curve, and one was outside of the
acceptance criterion; three results met the acceptance criterion. One QCS from a third gas
standard (4.78 ppm H2S, Scott Marrin) was made and was within 30% of the expected value. The
measurement precision of four analyses of a 10-ppb H2S standard was 8.1% relative standard
deviation (RSD). Three out of four measurement accuracy samples delivered as PE audit
samples were within the acceptance criterion. Once during the verification test, the in situ
reference method sampled zero air delivered through the ambient air inlet. The measurement
result was 3.1 ppb, which is approximately the same as the method quantitation limit for a
50-cubic-centimeter sample (200 pg/sample). Since the QC requirements for the in situ reference
method were not all satisfied, the results were compared to the APSA-360 data both
quantitatively and qualitatively [i.e., to determine whether they were significantly different from
each other at the 95% confidence level (see Section 5.8, Comparability)].
                                            14

-------
4.2 Audits

4.2.1   Performance Evaluation Audits

A PE audit was conducted to assess the quality of the H2S reference method measurements. In
the PE audit, key aspects of the reference measurements were checked by comparing them with
an independent National Institute of Standards and Technology- (NIST-) traceable standard. The
PE audit of the H2S reference methods was performed by supplying to each reference method a
blind, independent, NIST-traceable H2S standard provided by Battelle. The output of a certified
H2S permeation tube (VTCI Metronics, held at 30°C) was diluted in ultra-high purity (UHP) zero
air (approximately 2.7 to 3.9 Lpm) to produce H2S concentrations between 60 and 90 ppb. The
PE samples were analyzed in the same manner as the ambient air samples, and the analytical
results for the PE samples were compared to the nominal concentration. The target criterion for
the PE audit was agreement of the analytical result within 30% of the nominal H2S concentra-
tion. If the PE audit results did not meet the tolerances required, they were repeated.  PE audits of
the reference methods were required to be performed once prior to the start of the test and two
times during the test, at a minimum. A total of four PE audit samples each were submitted to the
USDA laboratory and to the in situ reference method for analysis. The USDA time-integrated
reference method results for the first and last PE audit samples met the acceptance criterion,
while the other two did not. The in situ reference method result met the acceptance criterion for
the first, third,  and fourth PE audit sample.

A PE audit of the ambient air sample flow rate for the time-integrated reference method was
performed by comparing it to an  independent flow measurement device. The target criterion for
this PE audit was agreement within the expected range (i.e., 2 to 3 seem). The PE audit of the
canister air sampling rate revealed that the actual flow rates for the Entech Flow Controller used
for this verification test ranged from 1.01 to 2.52 seem over 7.5 hours. The flow controller was
not adjusted to increase the flow rates since this would have the undesirable effect of shortening
the time-integrated sample duration. This deviation from the test/QA plan(1) was filed. This
deviation did not impact the quality of this verification test since the actual flow rate is not used
in the reference method analysis. However, variability in the canister sampling flow rate over the
7.5-hour collection time would impact the comparability of the air collected in the canister and
that sampled by the APSA-360.

A PE audit of the programmable dilution system was performed by comparing its output to an
independent flow measurement device.  One mid-range flow rate was audited for each flow
controller (i.e., 0.03, 0.3, and 5 Lpm) within the dilution system. The target criterion for this PE
audit was agreement within 5% of the flow readings; all measured flows agreed within 5%.
These audits were performed once during the verification test.

4.2.2   Technical Systems Audits

The Battelle Quality Manager performed a technical systems audit (TSA) on April 28 and 29,
2005, to ensure that the verification test was being performed in accordance with the AMS
Center QMP,(3) the test/QA plan,(1) ASTM method D5504-01,(2) and any standard operating
procedures used by USDA or Applied Measurement Science. In the TSA, the Battelle Quality
Manager toured the test site and the USDA laboratory, observed the H2S reference method
sampling and sample recovery, inspected documentation of H2S sample chain of custody, and

                                           15

-------
reviewed APSA-360-specific record books. The Battelle Quality Manager also reviewed the
reference methods used, compared actual test procedures to those specified by the test/QA
plan,(1) and reviewed data acquisition and handling procedures.

Observations and findings from this audit were documented and submitted to the Battelle
Verification Test Coordinator for response. No findings were documented that required any
corrective action. The records concerning the TSA are stored for at least seven years with the
Battelle Quality Manager.

4.2.3  Audit of Data Quality

At least 10%  of the data acquired during the verification test were audited. Battelle's Quality
Manager or his designee traced the data from the initial 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.
4.3 Quality Assurance/Quality Control Reporting

Each assessment and audit was documented in accordance with Sections 3.3.4 and 3.3.5 of the
QMP for the ETV AMS Center.(3) Once the assessment report was prepared, the Battelle
Verification Test Coordinator ensured that a response was provided for each adverse finding or
potential problem and implemented any necessary follow-up corrective action. The Battelle
Quality Manager ensured that follow-up corrective action was taken. The results of the TSA
were sent to the EPA.
4.4 Data Review

Records generated in the verification test were reviewed before these records were used to
calculate, evaluate, or report verification results. Table 4-2 summarizes the types of data
recorded. The review was performed by a technical staff member involved in the verification
test, but not the staff member who originally generated the record. The person performing the
review added his/her initials and the date to a hard copy  of the record being reviewed.
                                            16

-------
Table 4-2. Summary of Data Recording Process
Data to Be
Recorded
Dates, times, and
details of test
events, APSA-360
maintenance,
down time, etc.
APSA-360
calibration
information

APSA-360 H2S
readings
Where Recorded
ETV laboratory
record books or
data recording
forms
ETV laboratory
record books or
electronically

Recorded
electronically by
How Often
Recorded
Start/end of test
procedure, and at
each change of a
test parameter or
change of APSA-
360 status
At APSA-360
calibration or
recalibration

Recorded
continuously
By Whom
Battelle if on -site;
USDAifBattelle
not on-site
Electronic data by
vendor; Battelle if
on-site; USDA if
Battelle not on-
site
APSA-360
vendor, for
Disposition of
Data
Used to organize
and check test
results; manually
incorporated in
data spreadsheets
as necessary
Incorporated in
verification report
as necessary

Converted to
spreadsheet for
 Reference sample
 collection
 procedures,
 reference method
 procedures,
 calibrations and
 QA data, etc.
each APSA-360
and then
downloaded to
computer at least
weekly
Laboratory record
books and
electronically by
analytical method
Throughout
sampling and
analysis processes
transfer to
Battelle if on-site;
transfer to USDA
if Battelle not on-
site
USDA and
Applied
Measurement
Science
                                                                            statistical analysis
                                                                            and comparisons
Retained as
documentation of
reference method
performance
Reference method
H2S analysis
results





Electronically
from H2S
analytical method

Hard-copy
printouts and data
sheets

Every sample
analysis


Every sample
analysis


Applied
Measurement
Science

USDA



Entered into or
converted to
spreadsheets for
calculation of
ambient H2S
results and
statistical analysis
and comparisons
                                              17

-------
                                     Chapter 5
                  Statistical Methods and Reported Parameters
The statistical methods presented in this chapter were used to verify the performance parameters
listed in Section 3.1.
5.1 Accuracy

Accuracy of the H2S APSA-360 with respect to the individual H2S gas standards was assessed as
the percent recovery (%R), using Equation 1:
                              %R =
1 +
Y-X
  X
x 100                              0)
where Y is the average measured APSA-360 response (as defined in Section 3.3) and X is the
nominal H2S gas standard concentration. The average, minimum, and maximum %R values are
reported for each series of multilevel H2S challenges. A %R value of 100% indicates perfect
agreement between the averaged measured APSA-360 response and the nominal H2S gas
standard concentration.
5.2 Bias

Bias of the APSA-360 was defined as a systematic error in measurement that resulted in
measured error that was consistently positive or negative compared to the true value. The bias
was calculated as the average %D of the APSA-360 compared to the nominal H2S gas standard
concentration and was calculated for each series of multipoint H2S challenges, using Equation 2:

                                        Y -
                                                «100
where k is the number of valid comparisons, and Y and X are the same as in Equation 1.
5.3 Precision

The precision of the APSA-360 was evaluated from the triplicate responses to each H2S gas
standard supplied during the multipoint H2S standard challenges (outlined in Table 3-2). The

                                          18

-------
precision was defined as the %RSD of the averaged triplicate measurements and calculated for
each H2S concentration listed in Table 3-2, using Equations 3 and 4:

                                        SD,
                             %RSD  =  —L x  100
                                        Y,                                           (3)
                                     I (Y- Y)2
                                     ^V
                                        n - 1
                                                                                      (4)

where Y is the average APSA-360 response calculated from the last 5 data points (5 minutes) of
each gas standard delivery period, Y. is the overall average of the Y values at H2S concentration /
(/' = 30,  90, 150, and 300 ppb), and n is the number of measurements (3). The overall average
%RSD was calculated for each series of multipoint H2S challenges and included the %RSD for
all H9S concentrations tested.
5.4 Linearity

Linearity was assessed by a linear regression analysis using the diluted H2S standard gas
concentrations as the independent variable and results from the APSA-360 being tested as the
dependent variable. Linearity was expressed in terms of slope, intercept, and coefficient of
determination (r2).
5.5 Span and Zero Drift

The baseline response of the APSA-360 to zero air and the 30-ppb H2S standard was established
on the first day of testing (Week 4). The overall average (Y) and SD of the APSA-360 response
to zero air and the 30-ppb H2S standard were calculated from the average APSA-360 responses
from each of the five replicate measurements conducted during the first week of testing. From
these values, a control chart was constructed, and the  Y+ 2 SD "warning limit" and the Y+ 3 SD
"action limit" were calculated. Span drift was defined as having occurred if three consecutive
span checks fell either above or below the warning limit. Zero drift was defined as having
occurred if three consecutive zero checks fell either above or below the warning limit.
5.6 Response Time

Response time was assessed in terms of both the rise and fall times of the APSA-360 when
sampling the 30-ppb H2S gas standard and zero air on the first day of testing. Rise time (i.e., 0%
to 95% response time for the change in H2S concentration) was determined from the APSA-360
response to a rapid increase in the delivered H2S concentration. Once a stable response was
achieved with the H2S standard, the fall time (i.e., the 100% to 5% response time) was
determined in a similar way, switching from the H2S standard back to zero air.
                                           19

-------
5.7 Interference Effects

The interference effects of the APSA-360 were calculated in terms of the ratio of the response of
the APSA-360 to the interferant relative to the actual concentration of the interfering species. For
example, if 100 ppb of an interfering species resulted in a 1-ppb change in the response of the
APSA-360, the interference effect was reported as 1% (i.e., 1 ppb/100 ppb). Interference effects
are reported separately for each interferant both in the absence and in the presence of H2S in zero
air.
5.8 Comparability

The comparability of the APSA-360 and reference method results with respect to ambient air
was assessed by linear regression using the reference method H2S concentrations as the
independent variable and the results from the APSA-360 as the dependent variable. The
APSA-360 H2S measurements were averaged over the appropriate sample collection period for
each reference method (i.e., approximately 7.5 hours or 15 minutes). Comparability was
evaluated by linear regression analysis only for the in situ reference methods and was expressed
in terms of slope, intercept, and r2; only two quantitative time-integrated reference results were
available for the period during which the APSA-360 was operational at the field site. The linear
regression analysis was repeated for the in situ reference method, including only the reference
method results that were not significantly different from the APSA-360 average results at the
95% confidence level.  The 95% confidence interval (CI) was calculated for each APSA-360
average, using Equation 5:
                                                                                      (5)
where Y is the average APSA-360 response over the sample collection period, SD is the
standard deviation of the APSA-360 data over the sample collection period, n is the number of
APSA-360 readings used in the average, and t is the t-value of the Student's t-distribution for
95% confidence level and the degrees of freedom («-l). The calculated 95% CI for each
APSA-360 average was compared to the corresponding reference measurement value to
determine whether the results were statistically significantly different at the 95% confidence
level. For comparison to the time-integrated reference method, the APSA-360 readings used for
each average (approximately 7.5 hours) were plotted as a histogram to determine whether they
were normally distributed. Most of the samples (3 out of 4) were best represented by a log-
normal distribution.  For those samples, the natural logarithms (In) of the APSA-360 and
reference measurements were used to calculate the 95% confidence level and to determine
whether the results were significantly different at the 95% confidence level. This approach was
also applied to the in situ H2S reference method, using the APSA-360 averages over 15-minute
intervals, centered on the in situ reference measurement times. The APSA-360 readings used to
compare to the in situ reference method were assumed to be normally distributed.
                                           20

-------
5.9 Data Completeness

Data completeness was calculated as the percentage of the total possible data return achieved
over the entire field period. This calculation used the total hours of data recorded from each
APSA-360, divided by the total hours of data in the entire field period. The field period was
defined as beginning at 8:00 a.m. on April 25, 2005 and ending at 9:00 a.m. on June 3, 2005. No
distinction was made in this calculation between data recorded during a specific test activity
(e.g., data recorded for comparison to H2S reference method data) and that recorded during
routine ambient air monitoring.
                                            21

-------
                                      Chapter 6
                                      Test Results
The results of the verification test of the APSA-360 are presented in this section. The values
presented in this section are based on one-minute averages of the APSA-360 readings sampled
every 10 seconds by the data logger. The APSA-360 was factory-calibrated and the calibration
verified with UHP zero air and a 400-ppb dilution from a certified compressed gas cylinder
standard (100 ppm H2S, Scott Specialty Gases) that was independent of the gas standard used for
performing this verification test (5 ppm H2S, Scott Specialty Gases). Since the APSA-360
readings were -0.9 ppb and 393 ppb for zero air and the  400-ppb H2S calibration check,
respectively, the APSA-360 calibration was not adjusted. Although the calibration check
standard and the standard used for performing the verification testing were certified by the
manufacturer to have accuracy better than ±5%, differences between the actual H2S
concentration in the two cylinders may exist. Any differences between the gas standards used for
calibration and testing would be manifested in the accuracy and bias performance parameters
evaluated during this test; other performance parameters  such as linearity, precision, and
interference effects would not be impacted by differences in the gas standards because of the
nature of these calculations. Gas standard dilutions for calibration and testing activities were
prepared using the same dynamic dilution system. All APSA-360 measurement data were
analyzed and included in this report as output by the APSA-360. Any negative H2S concentration
values should be considered to indicate measurements of H2S concentrations less than those in
the zero air used for the APSA-360 factory calibration and/or drift in the APSA-360 response.
The APSA-360 was not recalibrated over the duration of this verification test.  As discussed in
Sections 3.3 and 6.10, the APSA-360 operated at the test site from May 16 through June 3, 2005
(Weeks 4, 5, and 6 of the  verification test).

Meteorological  conditions collected by a nearby (less than 2 miles)  meteorological station are
presented in Figure 6-1. The ambient data set collected by the APSA-360 is shown (in the
bottom panel), along with the wind direction, wind speed, and ambient temperature data. The
average ambient H2S concentration measured by the APSA-360 during the verification test was
11.9 ppb, with a range of  -1.8 to 522.9 ppb. The  meteorological conditions, which were
recorded as 1-hour averages, varied widely over the duration of the verification test. The average
ambient temperature was  14.3 °C, with a range of -4.9 to 29.0°C. The average APSA-360
ambient H2S concentrations are shown in Figure 6-2 plotted on polar coordinates as a function of
wind direction.  When winds were from the south, the APSA-360 was exposed to emissions from
the nutrient lagoons. As shown in Figure 6-2, the highest H2S concentrations were observed
during southwesterly winds, which passed across the primary nutrient lagoon before reaching the
instrument trailer. During northerly winds, the APSA-360 sampled  barn emissions and measured
much lower H2S concentrations. Winds were most frequently from  the northwest and southeast,
as shown by the diamonds in Figure 6-2. Under southerly winds, spikes in the measured H2S
concentration were often observed at the start of rain,  as shown in Figure 6-1.

                                           22

-------
to
5/1/2005      5/6/2005      5/11/2005      5/16/2005     5/21/2005
                                                                                           5/26/2005
5/31/2005
          Figure 6-1. Meteorological Conditions and APSA-360 Ambient H2S Measurements

-------
                                         N
                   W
                            O  Wind Direction Frequency (hourly averages)
                           	 Average APSA-360 Hydrogen Sulfide Data
                               Number of Observations = 349

                Figure 6-2. Hourly Averaged APSA-360 H2S
                Measurements Plotted as a Function of Wind
                Direction
6.1 Accuracy

Accuracy checks were conducted during Week 4 and Week 5 of the verification test. The
APSA-360 was challenged with compressed H2S gas standards diluted in zero air at several
concentrations (30 ppb to 300 ppb H2S). The H2S gas standards were diluted in zero air and
delivered to the Teflon manifold at a flow rate of 3 to 4 Lpm, with a vent to ambient pressure.

Figure 6-3 presents the H2S concentrations recorded by the APSA-360 during each accuracy
check gas challenge, along with the nominal H2S concentration levels supplied to the APSA-360
for Week 4 and Week 5. The averages of the last five minutes (5 data points) of the measure-
ments at each nominal H2S concentration and the calculated %R are presented in Table 6-1,
along with the average %R for each week. The SD for each average measured concentration is
also reported in Table 6-1 for reference  purposes. As shown in Table 6-1, the APSA-360 %R
values ranged from 106% to 133%, with an average of 128% for the Week 4 check. The
APSA-360 %R values for the Week 5 check ranged from 120% to 135%, with an average of
131%. Except for measurements of zero air, all of the APSA-360 concentrations reported for
Week 5 were higher than for Week 4.
                                           24

-------
                     APSA-360 Measurement Data
                     Nominal Hydrogen Sulfide Concentration
400-
                                              1   '   r
         00:00        01:00        02:00        03:00
                                Elapsed Time (hours)

Figure 6-3. APSA-360 Accuracy Results
                                                   04:00
05:00
                               25

-------
Table 6-1. Accuracy Results
Week 4
Measurement
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Average
Minimum
Maximum
Bias(%D)
H2S Gas
Standard
Concentration
(ppb)
0
30
90
150
300
90
0
300
150
30
300
150
90
30
0




Average
Measured
Concentration
(ppb)
-1.1
31.9
117.6
197.4
399.3
118.5
-0.8
398.5
197.2
36.2
395.8
196.0
117.1
36.7
1.3




SD
(ppb)
0.0
0.2
0.2
0.2
0.1
0.3
0.1
0.2
0.2
0.1
0.3
0.2
0.4
0.2
0.2




%R
NA
106
131
132
133
132
NA
133
131
121
132
131
130
122
NA
128
106
133
+28
WeekS
Average
Measured
Concentration
(ppb)
-1.6
35.9
119.1
200.2
402.4
119.4
-1.3
403.3
200.3
37.0
404.2
200.4
119.5
36.9
-0.6





SD
(ppb)
0.1
0.0
0.1
0.3
0.2
0.1
0.1
0.4
0.1
0.1
0.3
0.2
0.3
0.1
0.1





%R
NA
120
132
133
134
133
NA
134
134
123
135
134
133
123
NA
131
120
135
+31
NA = not applicable.
                                                     26

-------
6.2 Bias

Bias in the APSA-360 response to H2S gas standards was assessed for each of the accuracy
checks presented in Section 6.1 and calculated separately for each sequence of multilevel H2S
challenges. The APSA-360 bias observed during the Week 4 and Week 5 accuracy checks were
+28% and +31%, respectively. The consistently high bias is indicative of systematic error, which
would also affect the APSA-360 accuracy and could be caused by a number of factors,
including, but not limited to, differences in H2S gas standards used for calibration and testing
activities, the gas standard dilution system, and APSA-360 instrumental errors. The APSA-360
bias values are presented in Table 6-1.
6.3 Precision

Table 6-2 presents the calculated precision of the APSA-360 determined from the average
APSA-360 responses to the triplicate challenges at each H2S concentration level during the
Week 4 and Week 5 accuracy checks. The precision of the APSA-360 reading varied from 0.4%
to 7.5% during the Week 4 accuracy check and from 0.1% to 1.6% during the Week 5 accuracy
check. For both weeks, the highest RSD values were observed for the lowest concentration
standard (30 ppb). The average precision  calculated from each check was 2.2% and 0.5%.

Table 6-2. Calculated Precision of the APSA-360

H2S Gas Standard
Concentration (ppb)
30
90
150
300
Average
Week 4
Average Measured
Concentration (ppb)
34.9
117.8
196.9
397.9


%RSD
7.5
0.6
0.4
0.5
2.2
WeekS
Average Measured
Concentration (ppb)
36.6
119.3
200.3
403.3
%RSD
1.6
0.2
0.1
0.2
0.5
6.4 Linearity

Figure 6-4 shows the linearity results for the Week 4 and Week 5 accuracy checks. For each
check, a linear regression was calculated from the results presented in Table 6-1 (average
APSA-360 response versus the nominal H2S gas standard concentration) over the range of 0 to
300 ppb.  The 95% CI for the slope and intercept of the regression line were also calculated
(shown in the following text within parenthesis). For Week 4, the slope of the regression line
was 1.33  (± 0.02), with an intercept of -2.56 (± 3.62) and r2 value of 0.9998.
                                           27

-------
             O  Average Response
                        Linear Fit     —  1:1 Line
    400-
         0
 \
50
        \       \       \      \       \
100    150   200    250    300    350
    Nominal Hydrogen Sulfide (ppb)
400    450
Figure 6-4. APSA-360 Linearity Results
                                     28

-------
During Week 5, the linear regression showed a slope of 1.35 (± 0.01), an intercept -2.47
(± 1.76), and an r2 of 1.000. Over the range of concentrations tested (0 to 300 ppb H2S), the
APSA-360 demonstrated a high degree of linearity.
6.5 Span and Zero Drift

The baseline response of the APSA-360 to zero air and a 30-ppb H2S dilution was determined
during Week 4 of the verification test (on the first day of testing for the APSA-360). Figure 6-5
shows the APSA-360 response and nominal H2S  concentrations for the Week 4 check. The
average responses of the APSA-360 during each replicate delivery of zero air and 30 ppb H2S
are shown in Table 6-3. Each average utilized the last five data points for each zero air or H2S
standard delivery. The warning ( Y± 2SD ) and action ( Y± 3SD ) limits were calculated for
zero air and 30 ppb H2S and also are shown in the table.

Span and zero drift checks were performed at least twice each week during the verification test,
for a total of seven drift checks. The gas standard dilution system was not flushed with the H2S
gas standard before performing two of the span checks. Results from these span checks are
included in this report, but were not used to evaluate drift. The results of the span and zero drift
checks are shown in Table 6-4. Each average utilized the last five data points for each zero air or
H2S standard delivery. A control chart was prepared from the data shown in Tables 6-3 and 6-4
to demonstrate graphically whether drift occurred over the duration of the verification test. The
control chart is shown in Figure 6-6.
                  APSA-360 Measurement Data
                                              Nominal Hydrogen Sulfide Concentration
   _a
   Q.
   OT
   D)
   O
   T3
   >,
   I
       40 H
       30-
       20-
       10-
        0-
             Week4
                   I          I          I          I
       00:00      00:30      01:00      01:30      02:00
                                   Elapsed Time (hours)
Figure 6-5. APSA-360 Baseline Response Results
                                                              I          I
                                                            02:30      03:00
                                           29

-------
Based on the data presented in Table 6-3 and Figure 6-6, drift in the APSA-360 zero response
did not occur during the verification test; zero drift is defined as three consecutive drift check
results that fell above or below the warning limit of -1.3 ppb to -0.9 ppb established during the
first week of the testing for the APSA-360.

The warning limit established during Week 4 for the APSA-360 response to the 30-ppb H2S span
gas was 34.6 to 35.5 ppb. The last three span drift check results fell above the warning limit,
indicating a drift in the APSA-360 span response. The final span check response was 1.4 ppb
greater than the baseline response.

Table 6-3. Span and Zero Baseline Response

Drift
Check Date
Week 4
Tuesday
Week 4
Tuesday
Week 4
Tuesday
Week 4
Tuesday
Week 4
Tuesday

Average
(Ppb)
-0.9
-1.0
-1.1
-1.2
-1.1
Zero
SD
(Ppb)
0.0
0.0
0.0
0.0
0.0
Response'3'
Min-
imum
(Ppb)
-1.0
-1.1
-1.2
-1.2
-1.2

Max-
imum
(Ppb)
-0.9
-1.0
-1.1
-1.1
-1.1

Average
(Ppb)
35.2
35.3
35.0
35.0
34.7
30-ppb Span
SD
(Ppb)
0.1
0.2
0.0
0.1
0.2
Response(a)
Min-
imum
(Ppb)
35.1
35.1
35.0
34.8
34.4

Max-
imum
(Ppb)
35.3
35.5
35.0
35.1
34.9
    Baseline
    Response

   Overall SD

 Warning Limit(a)

   Action Limit
   -1.1

   0.1

-1.3 to -0.9

-1.4 to -0.8
   35.0

   0.2

34.6 to 35.5

34.3 to 35.8
(a) Statistics calculated from the last 5 data points (5 minutes) for each zero air or H2S standard challenge (n=5).
6.6 Response Time

Response time was determined during Week 4 from the amount of time required for the
APSA-360 to reach 95% of the change in response during the zero air and 30-ppb H2S span gas
replicate deliveries shown in Figure 6-5.  Table 6-5 presents a summary of the response time
determinations for the APSA-360. The average rise time was 5 minutes, and the average fall
time was 4 minutes. APSA-360 readings were recorded as one-minute averages, so the average
rise and fall times represent 5 and 4 readings, respectively.
                                            30

-------
Table 6-4. Span and Zero
Drift Check Results
Zero Check(a)
Check Average SD
Number (ppb) (ppb)
Week4 -10 01
r-pi i i . v/ vy . i
Thursday
?6?4 -0.6 0.1
Friday
Week 5 , - _ ,
A/T A -1-5 al
Monday
™*k5 -1.5 0.0
Friday
fek6 -1.1 o.i
Sunday
feek6 -1.5 0.1
Tuesday
Week 6
-1.3 0.1
Inursday
.„ . ._ . Within Within
Minimum Maximum ,,, . . ,.
, , . , , . Warning Action
(ppb) (ppb) Lim.t?8 Lim.t?
-1.1 -0.9 Yes Yes
-0.6 -0.5 No No
-1.5 -1.4 No No
-1.6 -1.5 No No
-1.1 -1.0 Yes Yes
-1.5 -1.4 No No
-1.4 -1.2 Yes Yes
30-ppb Span Check(a)
. __ -„. . -„ . Within Within
Average SD Minimum Maximum ,,, . ,.
, , f , , . , , . , , . Warning Action
(ppb) (ppb) (ppb) (ppb) UmKf Lim.t?
30.5(b) 0.4 30.1 31.1 (b) (b)
34.5 0.2 34.3 34.7 No Yes
23.1(b) 0.5 22.6 23.9 (b) ^
35.5 0.1 35.3 35.7 Yes Yes
36.4 0.0 36.4 36.5 No No
36.2 0.2 36.0 36.4 No No
36.4 0.1 36.3 36.6 No No
(a) Statistics calculated from the last 5 data points (5 minutes) for each zero air or H2S standard challenge (n=5).
(^ Gas standard dilution system was not flushed before this span check was performed.

-------
      40
                 O  Span Drift Check Measurement Data
                 X  Dilution System Not Flushed
                 O  Zero Drift Check Measurement Data
                                                      Average Response
                                                      Warning Limit
                                                      Action Limit
      35
        &
                                                                         0
                                                                                  0
 _a
  Q.
  &.
 V)
  ro
  
-------
6.7 Interference Effects

The effect of potential interferant gases on the response of the APSA-360 was assessed by
supplying the APSA-360 with a series of seven gases (listed in Table 6-6) in zero air and a
100-ppb H2S standard. The response of the APSA-360 during the introduction of these gases is
summarized in Table 6-6.

No interference effect was observed in the APSA-360 response to SO2, a blend of Cl to C6
alkanes, and ammonia. The APSA-360 showed an interference effect for carbonyl sulfide in zero
air of 31% and in 100 ppb H2S of 10%. The interference effect of methyl mercaptan on the
APSA-360 was 59% in zero air and 63% in 100 ppb H2S. Carbon disulfide and dimethyl sulfide
resulted in a 2% to 5% interference effect.
Table 6-6. Interference Effect Evaluation
                          Approximate      	Interference Effect (%)
                           Interferant
     Interferant      Concentration (ppb)     Zero Air Matrix     100-ppb H2S Matrix
Sulfur dioxide
Carbonyl sulfide
Carbon disulfide
Methyl mercaptan
Dimethyl sulfide
Hydrocarbon blend
Ammonia
100
100
100
100
100
500 (total)
500
0
31
2
59
O
0
0
0
10
5
63
5
0
0
6.8 Comparability

As stated previously, the APSA-360 was factory-calibrated with an H2S gas standard that was
independent of the standard used in the verification testing. The instrumentation for both
reference methods was calibrated using the same gas standard used in the verification testing of
the APSA-360. To reduce the potential impact on the comparability results due to differences in
calibration gases, the APSA-360 data were corrected using the results of the linearity checks
(Section 6.4) closest in time to the reference sample collection date. Thus, both reference method
and APSA-360 calibrations were referenced to the same H2S gas standard for the comparability
evaluations,  and any differences observed between the APSA-360  and reference method data can
be attributed to the analytical approach rather than the calibration source.

It should be  noted that the reference method quality control requirements  were not fully satisfied,
and, therefore, the accuracy of the reference method results could not be verified. In addition, the
                                           33

-------
swine finishing farm ambient air, which can contain high levels of ammonia and other small,
polar molecules, was very challenging analytically and may have caused measurement artifacts
resulting from contact of H2S and other gases with non-pas si vated surfaces in the air sampling
system. The comparability results presented here should be considered cautiously in light of the
reference method quality control results and the challenges associated with the complex ambient
air matrix.

6.8.1  Time-Integrated Comparability

Four of the time-integrated reference method measurements were collected while the APSA-360
was  operating at the test site. Two of the time-integrated reference results were below the
quantitation limit (2.2 ppb).  The maximum preanalytical holding time stated in the test/QA
plan(1) was 24 hours; however, holding times exceeded 24 hours for three of the four time-
integrated reference measurements. The long holding times may have resulted in degradation of
H2S  in the canisters. The two quantitative measurements were compared with the time-averaged
APSA-360 responses over the same periods (approximately 7.5 hours, n=456 data points) to
determine the time-integrated comparability. One of these measurements was a grab sample («=3
data  points), collected by allowing the canister to fill rapidly without a flow controller on the
inlet. The other three samples were collected as described in Section 3.3.5.1. The reference
method measurements were compared with the APSA-360 data by determining whether the
measurements were significantly different at the 95% confidence level; a linear regression
analysis could not be performed.

Figure 6-7 shows the time-integrated reference H2S measurements (red traces), the APSA-360
raw H2S data  (blue trace), and the APSA-360 averages for the reference measurement sample
periods (black trace). The grab sample is shown by individual symbols (red triangle for reference
measurement, black circle for APSA-360  average). Both of the time-integrated reference
measurements were statistically significantly different from the APSA-360 averages. The
APSA-360 and time-integrated reference method data are presented in Appendix B.
6.8.2   In Situ Comparability

The results of 41 in situ reference method results were compared with 15-minute averages
(w=15) calculated from the APSA-360 data that were centered in time on the in situ reference
measurement times. The 95% CI was calculated for each APSA-360 average and compared with
the in situ reference measurement to determine whether the results were significantly different at
the 95% confidence level. Figure 6-8 shows selected in situ reference measurements (red and
green diamonds), the APSA-360 H2S data, and the APSA-360 15-minute averages. Any upper
and lower limits reported for the in situ reference method are also shown in Figure 6-8. As
demonstrated by the green diamonds in Figure 6-8, 32% (13 of 41) of the quantitative in situ
reference values were not significantly different from the corresponding APSA-360 15-minute
averages. The APSA-360 and in situ reference method data are presented in Appendix C.

A linear regression analysis of the APSA-360 averages during the reference sampling periods
versus the H2S concentration determined by the in situ reference method is presented as a scatter
plot in Figure 6-9 to illustrate the correlation between the reference results and the APSA-360
data. The scatter plot includes reference method results that were within (green diamonds) and

                                           34

-------
    100-
     80-
APSA-360 Measurements:
    Measurement Data
	Averaged Data
 ©  Individual Result
Time-integrated Reference Measurements
	Results
 A  Grab Sample Result
	Quantitation Limit
 •  Canister Sampling Period
   1	r
  00:00
5/19/2005
                            12:00
   ~	1	r
       00:00
      5/20/2005
12:00
  00:00
5/21/2005
 Figure 6-7. Comparison of Time-Integrated Reference Measurements with Averages
 from the APSA-360

outside (red diamonds) the APSA-360 95% CI. The slope of the regression line including all
available quantitative results was 0.15 (±0.5), with an intercept of 26 (± 22) and an r2 value of
0.0325. When only the 13 results that were not significantly different at the 95% CI were
included in the linear regression analysis, the slope was 0.99 (± 0.34), with an intercept of 1.8
(± 13), and an r2 value of 0.9374.
6.9 Data Completeness

The APSA-360 operated for only 45% of the available time during the verification test (April 25
through June 3, 2005). During the time the APSA-360 was operating at the test site (May 16
through June 3, 2005), the data set was 98% complete. The data logger battery lost power due to
a faulty power cord, resulting in a 2% loss of data.
6.10 Operational Factors

The APSA-360 was first installed at the test site (on April 22) by the vendor representative, and
the installation was completed in less than one day. However, the APSA-360 was not supplied
with the 50-pin connector needed to collect data using the data logger provided by USD A. A
                                            35

-------
.0
Q.
 CD
T3
 (D
 O)
 O

-D
 >s
I
    300-
    250-
    200-
    150-
    100-
     50-
                                                       APSA-360 Measurements:
                                                         — Measurement Data
                                                       	15 minute average
                                                         — Range
                                                       In-situ Reference Measurements
                                                           Within APSA-360 95% Cl
                                                           Outside APSA-360 95% Cl
                                                        V  Upper Limit
                                                        A  Lower Limit
        18:00
      6/1/2005
            20:00
22:00
 14:00
6/2/2005
  r
16:00
  \
18:00
Figure 6-8. Comparison of Selected In Situ Reference Measurements with APSA-360
Averages and Measurement Data
      JD
      Q.
      Q.

      CD
      CO
      o
      CD
      CO
90


80 -


70 -


60 -


50 -


40 -


30 -


20 -


10 -


 0
                »  Results outside 95% Cl
                o  Results within 95% Cl
               	1:1 Line
                                                                   Results within 95% Cl
                                                                      y = 0.99x+ 1.8

                                                                        r2 = 0.9374
                  oo
           0       10       20      30       40      50       60

                                In Situ Hydrogen Sulfide Concentration (ppb)
                                                              70
                                             80
    Figure 6-9. Scatter Plot of APSA-360 Results versus In Situ Reference

    Measurements                           »,
                                              JO
                                    90

-------
connector was fabricated at the field site by USDA staff and used with a resistor to connect the
APSA-360 output to the data logger. It was then apparent that the data output range was not set
up as indicated on the Horiba Test Report for the APSA-360. The Horiba Test Report indicated
that the analog data output range was set to -4 to 20 milliamps, but the analog data output range
was actually set to 0 to 1 volt. After consulting the maintenance manual for the APSA-360, the
internal jumpers were set to the correct position for the output signal range of -4 to
20 milliamps. A program for the data logger was written by USDA staff. The data logger
sampled the APSA-360 data every 10 seconds and recorded one-minute averages.

Trial operations with the APSA-360 revealed that its performance did not satisfy the vendor
representative's experience-based expectations for response time and sensitivity. The vendor
representative, with assistance from Battelle and USDA staff, examined the APSA-360 in an
attempt to determine the cause of the poor performance, but no problems were found. Since the
APSA-360 was not working properly and could not be repaired at the field site,  it was shipped to
Horiba in California on April 25 for inspection and repair, as needed. Although no specific
problem was found, the APSA-360 was disassembled and reassembled, which appeared to
improve the APSA-360 performance, The APSA-360 was then shipped back to the test site (on
May 16), where it was installed by USDA staff. The installation was completed in less than one
day, and testing on the APSA-360 began the following day (May 17). Verification testing on the
APSA-360 was conducted over the next approximately three weeks. Since the verification test
had begun three weeks before the APSA-360 was operational at the test site, some reference
method measurements could not be compared with the APSA-360 results. All other planned
testing activities were performed, but condensed into a shorter time period than  was planned. No
maintenance was required for the APSA-360, which, once it was operational at the field site, was
easy to operate. The APSA-360 could be operated by a user with minimal  experience, once it
was working properly.
                                          37

-------
                                      Chapter 7
                               Performance Summary
The performance of the APSA-360 was evaluated for its accuracy, bias, precision, linearity, span
and zero drift, response time, interference effects, and comparability by evaluating the APSA-
360 response while sampling H2S and other gas standards at known concentrations and ambient
air. The APSA-360 was factory-calibrated prior to this verification test and verified with a
400-ppb dilution from an H2S gas standard (100 ppm) H2S that was independent of the gas
standard (5.12 ppm H2S) used for performing the verification test. All gas standard dilutions
were prepared using the same dynamic dilution system. The results of this evaluation are
described below.

The accuracy of the APSA-360 was assessed over the range of 30 ppb to 300 ppb in terms of
%R, which ranged from 106% to 133%, with an average of 128% for the Week 4 check. The
APSA-360 %R values for the Week 5 check ranged from 120% to 135%, with an average of
131%.

The APSA-360 bias observed during the Week 4 and Week 5 accuracy checks (30 ppb to
300 ppb) was +28% and +31%, respectively. The consistently high bias is indicative of
systematic error, which would also affect the APSA-360 accuracy and could be caused by a
variety of sources, including, but not limited to, differences in H2S gas standards used for
calibration and testing activities, the gas standard dilution system, and APSA-360 instrumental
errors.

The precision of the APSA-360 reading varied from 0.4% to 7.5% during the Week 4 accuracy
check and from  0.1% to 1.6% during the Week 5 accuracy check. The average precision
calculated from  each check was 2.2% and 0.5% for Weeks 4 and 5, respectively.

Linearity was evaluated in terms of slope,  intercept, and r2  over the range from 0 ppb to 300 ppb
H2S. For Week 4, the  slope of the regression line was 1.33  (± 0.02), with an intercept of -2.56
(± 3.62) and r2 value of 0.9998. During Week 5, the linear regression showed a slope of 1.35
(± 0.01), an intercept  -2.47 (± 1.76),  and an r2 of 1.000.

Drift was defined to have occurred if three consecutive  drift check results fell outside of the
warning limit (±2 standard deviations) calculated for zero (-1.3 ppb to -0.9 ppb) and a 30-ppb
span gas (34.6 to 35.5 ppb). Seven drift checks were conducted over a period of two weeks. Drift
was not observed in the APSA-360 response to zero air. The last three span drift checks fell
above the warning limit, indicating that drift in the APSA-360 response to the 30-ppb H2S span
gas did occur. The final span drift check value  was 1.4 ppb greater than the baseline response.

The average 95% rise time was 5 minutes, and the average 95% fall time was 4 minutes.

                                           38

-------
No interference effect was observed in the APSA-360 response to SO2, a blend of Cl to C6
alkanes, and ammonia. The APSA-360 showed an interference effect for carbonyl sulfide in zero
air of 31% and in 100 ppb H2S of 10%. Carbon disulfide and dimethyl sulfide resulted in an
interference effect of 2% to 5%. The interference effect for methyl mercaptan was 59% in zero
air and 63% in 100 ppb H2S.

Comparability was evaluated in terms of the slope, intercept, and r2 of a linear regression
analysis of the APSA-360 averages versus the reference measurements and was calculated
separately for the time-integrated and in situ reference methods. It should be noted that the
reference method quality control requirements, such as for preanalytical holding time and
analysis of quality control and performance evaluation standards, were not fully satisfied.
Therefore, the accuracy of the reference method results  could not be verified. In addition, the
swine finishing farm ambient air, which can contain high levels of ammonia and other small,
polar molecules, was very challenging analytically and may have caused measurement artifacts
resulting from contact of H2S and other gases with non-pas si vated surfaces in the air sampling
system. The comparability results presented here should be considered cautiously in light of the
reference method quality control results  and the challenges associated with the complex ambient
air matrix. Only two quantitative time-integrated reference results were available for the period
during which the APSA-360  was operational at the field site. Therefore, time-integrated
comparability could not be evaluated by linear regression analysis. Both of the two time-
integrated reference measurements were significantly different from the corresponding APSA-
360 averages at the 95% confidence level. The regression line slope for 41  quantitative in situ
reference measurements was 0.15 (± 0.5), with an intercept  of 26 (± 22) and an r2 value of
0.0325. Thirteen of the 41 quantitative in situ reference values (32%) were not significantly
different from the corresponding APSA-360 15-minute  averages. The regression analysis of
those 13 data points yielded a slope of 0.99 (± 0.34), an intercept of 1.8 (± 13), and an r2 value of
0.9374.

The APSA-360 was not functioning properly when first installed at the test site. The APSA-360
was sent to Horiba in California for repair and returned  to the test site, after which it was
successfully installed by USDA staff. A user with minimal experience and the instruction
manual could install and operate the APSA-360. No maintenance was required after the APSA-
360 was repaired by Horiba.  Daily checks of the APSA-360 were simple and quick.

The APSA-360 operated during 45% of the verification test because it was not running properly
for the entire test (April 25 to June 3). Once the APSA-360 was successfully installed at the test
site (May  16 to June 3), 98% of the data was collected and retrieved.
                                           39

-------
                                    Chapter 8
                                    References
1.    Test/QA Plan for Verification of Ambient Hydrogen Sulfide Analyzers at a Swine Finishing
     Farm, Battelle, Columbus, Ohio, April 2006.

2.    ASTM International. Standard Test Method for Determination of Sulfur Compounds in
     Natural Gas and Gaseous Fuels by Gas Chromatography and Chemiluminescence.
     Designation: D5504-01, 2001.

3.    Quality Management Plan (QMP) for the ETV Advanced Monitoring Systems Center,
     Version 5.0, U.S. EPA Environmental Technology Verification Program, Battelle,
     Columbus, Ohio, March 2004.
                                         40

-------
   Appendix A
APSA-360 Checklist
       A-l

-------
   /)

9/
                         Horiba APSA-360 H2S
              ETV Verification of Ambient Hydrogen Sulfide Analyzers
                            at a Swine Feedin9 Farm                |H i* I
Observe Analyzer Front Panel

D   Check for alarms                            Alarm =
D   Check for unreasonable readings              Value =
D   Download data
D   Send data to Battelle (at least weekly) Most recent Date   _
D   Send data to Horiba (at least weekly) Most recent Date   _
Action:  If any of issues above fails, note in logbook and contact:
Operator Name:

Signature:  	

Date:
Comments:
Note: Please re member to sign and date this form in non-erasable ink.
                                     A-2

-------
                 Appendix B
APSA-360 and Time-Integrated Comparability Data
                     B-l

-------
Time-Integrated Reference Method
Start Date
and Time

5/18/05
15:11

5/19/05
15:00

5/20/05
14:13
5/21/05
10:05
Stop Date
and Time

5/18/05
15:14

5/19/05
22:30

5/20/05
21:43
5/21/05
17:35
Result
(Ppb)

34.3


<2.2


<2.2

9.5

Distri-
bution

normal


log-
normal

log-
normal
log-
normal
In
Result

NA


NA


NA

2.25

Holding
time (hr)

5


46


25

44

Final
QCS

Pass

(a)


(a)


Fail

Average
Result (ppb)

26.90


5.90


1.94

6.51

SD
(Ppb)

0.41


2.02


0.82

6.92

APSA-360
In
Average




1.72


0.58

1.53


InSD




0.31


0.38

0.75


n

3


451


449

456


95%


CI

25.87 - 27.95 ppb


5.88-


1.90-

1.46-



5.93


1.97

1.60

Reference
Result
Within
95% CI?

No


NA


NA

No

       (a) No final QCS data provided.
       (b:i APSA-360 data were normally-distributed, so the natural logarithm was not calculated.
td
to

-------
             Appendix C
APSA-360 and In Situ Comparability Data
                 C-l

-------
In Situ Reference Method
APSA-360 (ppb)
Result Final Result at Sample Average
Sample Midpoint (ppb) QCS Midpoint Result
5/30/05
5/30/05
5/30/05
5/30/05
5/30/05
5/30/05
5/30/05
5/30/05
5/30/05
5/30/05
5/31/05
5/31/05
5/31/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/1/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
14:04
14:20
14:36
15:26
15:49
19:08
19:24
21:42
21:56
22:12
11:20
11:48
22:17
17:55
18:02
18:20
18:35
19:05
19:21
19:42
20:03
20:24
20:45
21:05
21:26
21:56
22:43
23:16
11:53
12:08
12:25
12:46
13:08
13:23
13:39
13:55
14:11
16:06
16:22
1.51
<1.48
<1.48
2.69
<1.78
Pass
<1.48
<1.48
12.79
168.62
2.95
0.38
0.37 (a)
2.96
10.21
19.28
>42.49
17.05
3.37
14.59
5.04
35.18 (a)
>42.85
>42.85
>42.85
>42.85
3.09
0.44
1.85
59.55
9.40
60.56
7.83
45.65 Pass
41.57
46.60
88.07
7.99
7.83
47.66
52
.80
3334
4.
5.
4.
1.
4.
96
67
64
00
51
182.85
2.
6.
1.
1.
0.
12
22
19
90
58
58
.37
18.11
67
3.
39
10
22
40
.58
99
.39
.93
.14
.32
104.68
118.88
191.12
137.79
3.
1.
1.
6.
8.
64
8.
79
17
6.
1.
7.
6.
68
46
33
96
40
94
.32
56
.23
.69
09
92
65
76
.02
14
3.
o
J.
4.
o
5.
i.
i.
.66
04
58
09
17
28
91
153.88
7.
3.
1.
1.
0.
28
22
50
15
22
21
22
45
18
93
07
78
66
.83
.48
.25
.04
.60
.82
.28
.53
103.88
121.84
189.15
112.37
5.
1.
1.
47
35
61
34
61
41
30
35
29
17
31
90
36
83
.93
.01
.93
.79
.14
.25
.55
.64
.34
.14
.66
SD
19
1.
2
2.
1.
0.
1.
68
10
1.
0.
0.
0.
15
9.
24
12
18
5.
14
12
13
45
70
56
6.
0.
0.
34
27
25
30
33
18
18
47
18
14
19
.65
45
07
41
67
39
40
.46
.53
91
36
32
07
.62
92
.77
.92
.93
80
.33
.54
.24
.06
.95
.46
93
10
70
.71
.82
.18
.41
.03
.28
.06
.03
.08
.59
.06
]
95% CI
3.03
2.23
2.43
2.76
2.25
1.07
1.14
115.97
1.35
2.87
0.88
1.60
0.62
20.18
16.99
36.54
7.89
12.12
18.61
14.34
38.59
96.55
96.88
149.86
81.10
2.06
1.30
1.44
28.71
19.60
47.98
17.95
42.85
31.13
20.55
9.60
19.32
9.06
21.11
-26.29
-3.84
-4.73
-5.43
-4.10
-1.50
-2.69
-191.80
- 13.01
-4.99
- 1.27
-1.96
-0.70
-37.48
-27.98
-63.97
- 22.20
-33.09
-25.04
-30.21
- 52.48
-111.21
- 146.79
- 228.44
- 143.64
-9.74
- 1.41
-2.22
-67.15
- 50.42
-75.88
-51.64
-79.44
-51.37
-40.55
-61.69
-39.35
-25.22
- 42.22
Reference
Result
Within
95% CI?
No
NA
NA
No
NA
NA
NA
No
No
Yes
No
No
No
No
Yes
NA
Yes
No
No
No
No
NA
NA
NA
NA
Yes
No
Yes
Yes
No
Yes
No
Yes
Yes
No
No
No
No
No
C-2

-------
In Situ Reference Method
APSA-360 (ppb)
Result Final Result at Sample Average
Sample Midpoint (ppb) QCS Midpoint Result
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/2/05
6/3/05
16:38
16:54
17:11
17:28
17:44
18:00
18:16
18:31
19:05
19:20
19:35
19:51
4:33
<7.69
7.80
8.32
8.70
39.82
13.12 pass
62.63
18.32
18.01
45.34
7.87
21.67
1.56 (a)
11
9.
29
42
42
33
61
21
47
27
43
15
1.
.95
57
.06
.56
.69
.55
.20
.87
.18
.84
.35
.14
28
11
7.
27
57
46
50
64
40
38
25
27
23
1.
.42
20
.72
.00
.54
.38
.07
.90
.73
.64
.42
.98
39
SD
8.66
4.82
26.71
31.94
25.24
21.57
16.44
21.23
28.35
10.58
10.15
8.85
0.30
95% CI
6.
4
12
39
32
38
54
29
23
19
21
19
1
63-
.53-
.93 -
.31 -
.57-
.43-
.97-
.14-
.03 -
.78-
.79-
.08-
.22-
16.22
9.87
42.52
74.69
60.52
62.32
73.18
52.65
54.43
31.49
33.04
28.88
1.56
Reference
Result
Within
95% CI?
NA
Yes
No
No
Yes
No
Yes
No
No
No
No
Yes
No
' QCS not analyzed at end of sampling on this date because the liquid nitrogen supply ran out.
                                                  C-3

-------