THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
PROGRAM
oEPA
M-S. Ennramiifal Ptvtoction Agmty
Bameiie
The Business of Innovation
TECHNOLOGY TYPE: ISOTOPIC CARBON DIOXIDE ANALYZERS
APPLICATION:
CARBON SEQUESTRATION MONITORING
TECHNOLOGY NAME: Cavity Ring-Down Spectroscopy Analyzer for Isotopic CO2
Model GllOl-i
COMPANY:
ADDRESS:
WEB SITE:
Picarro, Inc.
480 Oakmead Parkway
Sunnyvale, CA 94085
http://www.picarro.com/
PHONE: 408-962-3900
ETV Joint Verification Statement
The U.S. Environmental Protection Agency (EPA) has established the Environmental Technology Verification
(ETV) Program to facilitate the deployment of innovative or improved environmental technologies through
performance verification and dissemination of information. The goal of the ETV Program is to further
environmental protection by accelerating the acceptance and use of improved and cost-effective technologies.
ETV seeks to achieve this goal by providing high-quality, peer-reviewed data on technology performance to
those involved in the design, distribution, financing, permitting, purchase, and use of environmental
technologies. Information and ETV documents are available at www.epa.gov/etv.
ETV works in partnership with recognized standards and testing organizations, with stakeholder groups
(consisting of buyers, vendor organizations, and permitters), and with individual technology developers. The
program evaluates the performance of innovative technologies by developing test plans that are responsive to the
needs of stakeholders, conducting field and 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 Advanced Monitoring Systems (AMS) Center, one of six verification centers under ETV, is operated by
Battelle in cooperation with EPA's National Risk Management Research Laboratory. The AMS Center
evaluated the performance of isotopic carbon dioxide analyzers for carbon sequestration monitoring. This
verification statement provides a summary of the test results for Picarro, Inc.'s Cavity Ring-Down Spectroscopy
Analyzer for Isotopic Carbon Dioxide - Model Gl 101-/'.
VERIFICATION TEST DESCRIPTION
This verification test of the Model Gl 101-/ was conducted from July 9 through August 17, 2010 at Battelle
laboratories in Columbus, OH, Battelle's Ambient Breeze Tunnel (ABT) in West Jefferson, OH, and a geological
carbon sequestration site located at a coal-fired power plant in West Virginia. Performance of the Model
-------
Gl 101-/ was verified for carbon dioxide (CO2) concentration and the stable isotope ratio of carbon in CO2.
Deviations in the ratio of 13C to 12C (13C/12C) in atmospheric CO2 relative to that in ambient air can be used to
identify input from other carbon sources, such as fossil fuel combustion, since atmospheric, carbonate, and plant-
derived carbon differ in their 13C/12C relative to the Pee Dee Belemnite (PDB) standard. The relative difference
in stable carbon isotope from the PDB standard, referred to as 513C, is calculated as shown in Equation 1 and
expressed in per mil (%o), or part per thousand.
i3iZe 1
— -
_
Sample - i3r/i2r
\ L/ i-
- PDB
Since the PDB standard was highly enriched in 13C, most naturally occurring carbon sources have a negative
513C value. For example, ambient air CO2 has a global average 513C close to -8%o and fossil fuels are typically
in the range of -28%o. Reference analyses of CO2 concentration and 513C by nondispersive infrared (NDIR)
analysis and isotope ratio mass spectrometry (IRMS), respectively, were performed between August 1 1 and
August 20, 2010 by the National Oceanic and Atmospheric Administration (NOAA) and the Stable Isotope Lab
(SIL) at the Institute for Arctic and Alpine Research (INSTAAR) in Boulder, CO.
One of the goals of this verification test was to provide information on the potential use of the Model Gl 101-/'
for monitoring at or near facilities utilizing geologic carbon sequestration (CS) for captured CO2. Since the 513C
of CO2 from fossil fuels can be distinguished from ambient air, technologies that detect 513C in addition to CO2
concentration may be able to identify leakage more quickly than by monitoring for CO2 concentration alone and
therefore are likely to be of interest to CS site operators and regulators. To accomplish this monitoring goal, the
experimental design included a combination of controlled gas challenges in an indoor laboratory environment
and a sheltered ambient breeze tunnel, survey measurements for above-ground leak detection, and continuous
ambient monitoring to provide performance data under a variety of simulated and real -world conditions. The
Model Gl 10 1-/' was evaluated in terms of:
• Accuracy and bias - comparison of analyzer CO2 concentration and 513C response to dilutions from
certified gas standards to nominal levels and at variable temperature and relative humidity (RH)
conditions
• Linearity - linear regression analysis of analyzer CO2 concentration and 513C response to dilutions from
certified gas standards versus nominal levels
• Precision - average analyzer response to triplicate challenges at each of 1 1 CC>2 concentrations
• Response time - 95% rise and fall times calculated from transitions between CO2 gas standard dilutions
of increasing and decreasing concentration, respectively
• Minimum detectable leak rate - the minimum flow rate of 12CO2 detected by the Model Gl 101-/ above
ambient 513C variability under controlled field conditions
• Comparability - analyzer CO2 concentration and 513C response to ambient air compared to NDIR and
IRMS methods, respectively
• Data completeness - assessment of data return by the Model Gl 101-/ during the verification test
• Operational factors - such as general operation, data acquisition, set up, and consumables.
The Model Gl 10 1-/' was installed and operated according the vendor's instructions and manual by Battelle staff;
the vendor representative was available to answer questions and provide support, but no formal training was
conducted. Phase 1 of this verification test was conducted in Battelle laboratories in Columbus, OH to evaluate
the analytical performance of the Model Gl 10 1-/' under controlled laboratory conditions from July 9 through July
23 and August 1 1 through August 17, 2010. The Model Gl 10 1-/ was challenged with gas standards of known
isotopic composition and concentration to generate test samples over a range of CO2 concentrations and isotopic
compositions. CO2 concentration and 513C data were first corrected for an error in the factory-set calibration and
for water vapor interference then used to calculate accuracy, bias, linearity, precision, and response time, where
appropriate. Bias with respect to ambient temperature and relative humidity (RH) was also assessed.
The ability of the Model Gl 10 1-/ to detect CO2 leaks was evaluated during Phase 2 of this verification test, which
was conducted at Battelle 's Ambient Breeze Tunnel (ABT) facility in West Jefferson, OH. The ABT was used to
simulate leaks of 13C-depleted CO2 (i.e., 12CO2) in ambient air under simulated field conditions. The Model
Gl 10 1-/ was installed inside the ABT, where ambient air was drawn through the tunnel at approximately 4 miles
-------
per hour (mph) and a stream of pure 12CO2 at a fixed flow rate was periodically introduced. By varying the 12CO2
flow rate, the minimum detectable CO2 leak rate was determined. In addition, ambient air reference samples were
collected to determine the comparability of the Model Gl 101-/ to CO2 concentration and 513C reference methods.
Testing for Phase 2 was conducted from July 28 through July 30, 2010.
The utility of the Model Gl 101-/ for monitoring at GCS sites was evaluated during Phase 3, which was conducted
at a coal-fired power plant in West Virginia. The analyzer was installed in a shed near the sequestration wells and
sampled ambient air drawn from near the main wellhead over a one-week period from August 2 through August
6, 2010. During that period, ambient air reference samples were collected to determine the comparability of the
Model Gl 101-/' to CO2 concentration and 513C reference methods. The Model Gl 101-/' was also installed in a
hybrid sedan vehicle and operated using battery power to conduct mobile surveys of GCS site transmission lines
and infrastructure. Finally, CO2 from the sequestration operation was intentionally released to simulate a high-
risk area above-ground leak, and the leak-rate response time was determined.
QA oversight of verification testing was provided by Battelle and EPA. Battelle QA staff conducted technical
systems audits of both the laboratory and field testing, and Battelle QA staff conducted a data quality audit of at
least 10% of the test data. This verification statement, the full report on which it is based, and the test/QA plan
for this verification test are available at www.epa.gov/etv/centers/centerl.html.
TECHNOLOGY DESCRIPTION
The following is a description of the technology, based on information provided by the vendor. The information
provided below was not verified in this test.
The Model Gl 101-/ is a low-drift, high-precision analyzer designed to measure the stable isotope ratio of carbon
in CO2 and CO2 concentration. This analyzer is based on cavity ring-down spectroscopy (CRDS), which is a
technique in which a gas sample is introduced into a high-finesse optical cavity, and the optical absorbance of the
sample is determined, thus providing concentration or isotopic ratio measurements of a particular gas species of
interest.
The analyzer continuously scans a laser through a sample, and determines individual carbon dioxide rovibrational
resonant absorption lines for 12CO2 and 13CO2. Each spectrum is comprised of absorption loss as a function of
optical frequency. The concentration is proportional to the area under each measured spectral feature.
Concentration measurements are provided approximately every second, corresponding to a total of 100 ring-down
and wavelength monitor measurements, and the isotope ratio (13C/12C) is derived from the ratio of the 13CO2 and
12CO2 concentrations in the sample volume.
The Model Gl 101-/ weighs 26 kg (58 Ibs), has dimensions of 43 x 25 x 59 cm (17" x 9.75" x 23") including the
feet, and can be rack mounted or operated on a bench top. The approximate purchase price of the Model Gl 101-/'
is U.S. $60,500.
VERIFICATION RESULTS
The verification of the Model Gl 101-/' is summarized below. The estimate of uncertainty in the nominal CO2
concentrations for dilutions used to evaluate the Model Gl 101-/' was approximately 7%. It is not possible to
determine from these measurements alone whether the observed inaccuracies and biases relative to gas standard
challenges are due to errors in the instrument response or the gas preparation.
Concentration Accuracy, Bias, Precision, and Response Time. The accuracy of the Model Gl 101-/ was
assessed over the range of 100 ppm to 5,000 ppm in terms of %R, which ranged from 90 to 113%, with an
average of 96%. Bias, or the average percent difference between the Model G1101-/ response and the known
value, was -4.0%. Precision of the Model Gl 101-/ was determined from the average responses to triplicate
challenges at each of 11 CO2 concentrations. The relative standard deviation values ranged from 0.10% to 1.2%,
with an average of 0.30%. The average 95% response time was 142 seconds for rise time and 152 seconds for fall
time.
-------
Concentration Linearity. Linearity was evaluated in terms of slope, intercept, and R2. Over the 0 to 400 ppm
range, the slope of the regression line was 0.935 (±0.036), with an intercept of 11.3 (±8.90) and R2 value of 0.996.
Over 0 to 5,000 ppm, the slope of the regression line was 0.938 (±0.006), with an intercept of-1.32 (±13.6) and
R2 value of 0.999. (The 95% confidence interval for the slope and the intercept of each line is shown in
parenthesis.)
Isotope Ratio Accuracy, Bias, and Linearity. The accuracy of the Model Gl 101-/ 513C response was assessed at
-3.60 %o, -10.4%o, and -40.8%o at three concentration levels: 259 ppm, 370 ppm, and 740 ppm. Values for 513C
differed from the expected value by between 1.1 to 2.7%o, with an average of 1.7%o. The lowest absolute
differences were observed for the -40.8%o standard and at the higher CO2 concentrations. Isotope ratio linearity
was assessed in terms of slope, intercept, and R2 at -3.60%o, -10.4%o, and -40.8%o. The strongest correlation
between concentration and measured isotope ratio was observed for the -3.60%o standard, with an R2 value of
0.947, aslope of-0.0025 and intercept of 3.24. (The 95% confidence interval for the slope and the intercept of
each line is shown in parenthesis.)
Temperature and Relative Humidity Bias. Temperature and RH bias were assessed by comparing the Model
Gl 101-/ CO2 concentration and 513C response to dilutions from a certified CO2 standard at five temperature/RH
conditions to its response at 20°C and 0% RH. During this evaluation, the Model Gl 101-/' was installed in a
temperature/RH-controlled chamber; humidified zero air was added to the CO2 gas standard dilution to achieve
the desired RH. The following test conditions were evaluated: 20°C/50% RH; 20°C/90% RH; 32°C/50% RH;
32°C/90% RH; 4°C/50% RH. In general, variability in ambient temperature and RH conditions resulted in bias
values of 3.0% or less for the Model G1101-/' concentration measurements and 0.70%o or less for isotope ratio.
The maximum concentration bias value, 3.0%, was observed for CO2 concentration at 4°C/50% RH. The largest
isotope ratio average difference of 0.7%o was observed for 32°C/90% RH.
Minimum Detectable Leak Rate. The ability of the Model G1101 -/' to identify CO2 leaks above ambient air
variability was evaluated by simulating leaks under controlled field conditions in the Ambient Breeze Tunnel.
Pure 12CO2 was periodically released into a constant flow of ambient air and the flow rate adjusted until the
Model Gl 101-/' difference in the 513C response during the "leak" compared to ambient air was greater than 2
times the ambient air 513C variability. Under conditions that simulated 1.8 m/s winds, the minimum detectable
leak rate was 0.423 LPM 12CO2, which resulted in a 0.90%o decrease, on average, in the Model Gl 101-/ 513C
readings compared to ambient air (approximately -6.4). This result was extrapolated to determine the equivalent
leak rates for CO2 sources of with 513C values -35%o, -20%o, and -3.5%o. The equivalent leak rates were 14.62
LPM, 31.84 LPM, and 198.2 LPM, respectively.
Ambient Air Monitoring. The Model Gl 101-/ monitored ambient air at the GS site between August 2 and
August 6, 2010. During this period, the average ambient CO2 concentration was 411 ppm, with a range of 365 to
488 ppm. The average measured stable isotope ratio was -6.42 and values ranged from -9.50 to -4.28. The
relationship between CO2 concentration and 513C was investigated by producing a Keeling Plot. The value of the
intercept, which represents the 513C of the CO2 source, is -23. l%o and is consistent with the value for captured
CO2 at this site. An intentional release of captured CO2 was detected by the Model Gl 101-/ in less than 60
seconds. A Keeling Plot of the data from the intentional release period had an R2 value of 0.939 and an intercept
of-24.0 (± 0.2)%o, that is similar to the intercept found for ambient measurement data and is consistent with the
isotope ratio of the CO2 injected at the site.
Mobile Surveys.
During Phase 3, the Model Gl 101-/' was transported to road-accessible features of the GS, such as transmission
lines and monitoring wells, to evaluate the ease of use and operational factors of the analyzers during use in a
mobile survey mode. The Model Gl 101-/ surveyed 16 features at the GS while installed in the back seat of a
Nissan Altima hybrid sedan and operating on power from a marine deep cycle/RV battery and power inverter.
Once the first installation in the vehicle had been completed, it generally took approximately 15 minutes with two
testing staff to shut down the analyzer, move all the components into the vehicle, and conduct data collection on
-------
battery power. Some additional time was then needed for the analyzer response to stabilize. At least one hour of
monitoring data could be collected on a single battery.
Comparability to Reference Methods. Comparability was determined as the accuracy (%R) and bias (average
percent difference) of the Model Gl 101-/' response compared to CO2 concentration (NDIR) and 513C (IRMS)
reference method results for 10 duplicate grab samples of ambient air. The average accuracy for CO2
concentration was 98% with a range of 94% to 101%. For 513C, the average difference was -3.0% and values
ranged from -3.5% to -2.1%. Bias was calculated separately for each site. At the ABT and GS site,
concentration bias was -0.20% and -2.5%, respectively.
Data Completeness. The Model Gl 101-i operated for 100% of the available time during Phase 1 and Phase 2 of
the verification test. During Phase 3, internal calibrations took place during 7.0% of the available testing time (6
hours) and the analyzer was shut down for 2.6 hours because ambient temperatures in the shed where the analyzer
was operated exceeded operating limits identified by the vendor. The internal calibrations and temperature-
related downtime resulted in a 91% data return during Phase 3 of this verification test. When supplied with the
necessary power (i.e., a fully charged battery), the Model Gl 101-/' data return during mobile survey testing was
100%.
Operational Factors. The Model Gl 101-/ was installed in the laboratory and at both field sites by Battelle testing
staff; the installation was completed in less than one hour; no formal training by the vendor was necessary.
Instructions in the user manual for the installation were clear and easy to follow. A checklist was provided by the
vendor representative to establish whether the analyzer was in proper working order during the test. No
maintenance was performed on the analyzer. Data were downloaded on a daily basis to a USB memory stick or
expansion drive. In general, the Model Gl 101-/ software was easy to use. Battelle staff found the zoom and
other features on the graphical display to be somewhat cumbersome and not especially intuitive. Ease of use of
the software improved with practice. Batteries used to operate the Model 1101-/ during mobile surveys were
reusable and rechargeable. The Model Gl 101-/' did not generate any waste or use consumable supplies.
Signed by Tracy Stenner June 15. 2011 Signed by Sally Gutierrez June 17. 2011
Tracy Stenner Date Sally Gutierrez Date
Manager Director
Environmental Solutions Product Line National Risk Management Research Laboratory
Energy, Environment, and Material Sciences Office of Research and Development
Battelle U.S. Environmental Protection Agency
NOTICE: ETV verifications are based on an evaluation of technology performance under specific,
predetermined criteria and the appropriate quality assurance procedures. EPA and Battelle make no
expressed or implied warranties as to the performance of the technology and do not certify that a technology
will always operate as verified. The end user is solely responsible for complying with any and all applicable
federal, state, and local requirements. Mention of commercial product names does not imply endorsement.
------- |