Enabling Continuous, Field-Based Isotope and Greenhouse Gas Measurements with
WS-CRDS-based Analyzers
C. Rella and A. Van Pelt Picarro, Inc., Sunnyvale, CA, United States ([email protected])
Optical
components
1550.4
Time (micro-seconds)
Calculate loss
Photodetector
δ=1/cτringdown
Data collection &
analysis electronics
Laser control
electronics
Gas concentration is proportional
to the area under the curve
Tiny temperature and pressure instabilities cause BIG
concentration errors!
Once the spectroscopic feature is stable, measuring it
accurately requires determination of its horizontal axis
(wavelength) and vertical axis (absorption) with high-precision.
Where:
I: Light intensity in the cavity
c: Speed of light
d: Loss per unit length
WS-CRDS analyzer schematic
• When the optical frequency matches the resonance frequency
of the cavity, energy builds up in the cavity.
• When the build-up is complete, the laser is shut off.
• Light circulates in the cavity ~100,000 times, traveling >20km
The high precision of WS-CRDS comes from this long
interaction pathlength providing parts-per-trillion detection
levels for some gases.
• The energy decays from the cavity (through a partially-reflective
mirror) exponentially in time, or “rings down,” with a
characteristic decay time. This energy decay is measured, as a
function of time, on a photodiode.
• The ring down time measurement is continuously repeated
(>100 times per second) at several different well-controlled
points in wavelength as the laser is tuned across the molecular
signature of the analyte gas.
• WS-CRDS is a measurement of time not of absorbance. When
the laser is at a wavelength where the gas in the cavity is
strongly absorbing, the ring down time is short; when the
wavelength is such that the gas does not absorb, the ring down
time is long. WS-CRDS has complete immunity to laser noise
since the laser is actually off during the measurement.
• The gas concentration or isotope ratio is determined by a
multi-parameter fit to this lineshape (red curve) and is
proportional to the area under the curve. The vertical axis is loss
(or absorption, measured with cavity ring down); the horizontal
axis is wavelength (measured with Picarro Wavelength meter)
Choosing a spectral line
Hundreds of lines for any given species are available for analysis (only need one line per species to measure its concentration).
How do we choose which line to use?
Zooming in: choose the strongest possible molecular feature that is as free from interference as possible (example of CO2in ambient air).
CO2
CO2
CO
CH4
CH4
H 2O
CO2
Line may still have
interferences.
(example at right
shows how WS-CRDS
addresses this)
CO
Use multiple lasers and wide-bandwidth optics to allow multi-species operation in a single analyzer.
–
Vertical axis precision given by wavelength monitor
–
Horizontal axis precision given by long pathlength and
inherent sensitivity of ring down measurement itself
High-accuracy wavelength control –
critical to accurate measurements
In an ideal world, there are no nearby interfering lines, so you
could just measure the peak and the baseline to find
concentration. But in the real world, interfering lines are
everywhere
Measuring the peak is easy, but…
How do you measure the baseline under the peak?
1 ppmv CO @ T2 and P2
1.0
0.8
0.6
0.4
0.2
0.0
• Measure the baseline here
(off-resonance) or empty the
cavity
• The concentration is
proportional to the difference
between peak and baseline
B
Stability with Picarro’s
temperature control
383
382
381
Picarro
wavelength
precision
380
379
Without
temperature control
378
44.0
45.0
46.0
47.0
Ambient Temperature (ºC)
1.2
100 ppbv CO
1.0
4.0
3.5
3.0
H 2O
In the real world, you
only get to measure
the blue circles, not
the red line
CO2
2.5
2.0
CO
1.5
1.0
0.5
???
0.0
-0.20 -0.15
-0.10 -0.05
0.00
4.0
3.5
3.0
• Stabilized to environmental fluctuations
• Non-operational hot/cold extreme temperature soaks and
start-up testing
• Precision of 8 femtometers in the near infrared
• >1 week operational burn-in testing
• 8 femtometers compared to the laser wavelength is like
comparing the size of a penny to the width of the US!
• 100% performance testing for all key specifications
High-precision, high-stability data
CH4 5 Minute Average
C02 5 Minute Average
0.6
366.5
0.4
0.2
366.4
0.0
-0.2
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4
Relative Frequency (Wavenumbers)
2.083
Mean = 366.285 ppmv
SD = 11.2 ppbv
Range = 76 ppbv
2.082
366.3
366.1
90
31.5 Days
2.0
0.10
0.15
0.25
1.0
0.5
Relative Frequency (Wavenumbers)
To measure the baseline under the CO peak, you can’t move off
resonance, because there is no spot in the spectrum that
corresponds to the spectral baseline with ONLY the CO removed
(the H2O and CO2 are still present). And you can’t remove just
the CO gas. What do you do?...
-0.2
-0.1
0.0
-122
95
100 105 110
Time (Days)
115
120
90
95
31.5 Days
-15.4
-124
-15.6
-125
2.079
2.076
Stdev = 0.29‰
-123
2.080
2.077
Hydrogen
Oxygen
-15.8
-126
-127
-128
Standard Deviation 0.024%
100 105 110
Time (Days)
115
-16
120
-129
Stdev = 0.05‰
0
20
40
60
80
-16.2
100
Injection Number
The precision and long-term stability of these greenhouse gas and isotope
measurements is afforded by the temperature, pressure and wavelength control
systems of the WS-CRDS analyzers.
0.0
0.05
Mean = 2.0795 ppmv
SD = 0.509 ppbv
Range = 2.1 ppbv
-15.2
30 day test – single bottle – no calibration adjustment
Cause big baseline
errors here
1.5
2.081
-121
2.078
Standard Deviation 0.003%
Even small wavelength
errors here ...
2.5
• Operational temperature cycling, 5°– 45°C
• Broadband, fiber-coupled device
0.8
5000 ppmv H20
500 ppmv CO2
500 ppbv CO
total spectrum
4.5
• 0.5g vibration, 15 minutes, 2 axes
• Proprietary optical sensing technology
377
43.0
• MIL-STD810F bench-handling on three axes, 20 drops from
4” height onto a hard surface
Wavelength Monitor
366.2
total spectrum
spectrum without analyte
All instruments undergo field
ruggedness testing:
Nassau
In an ideal world:
• Measure the peak height
here (on-resonance)
ngth
San Antonio
384
376
-0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20
Relative Frequency (Wavenumbers)
vele
385
In the real world:
Optical Absorption (parts-per-billion/cm)
Wavelength-Scanned Cavity Ring Down Spectroscopy (WS-CRDS) – How it Works
• Light from a tunable semiconductor diode laser is directed into
an optical resonator cavity containing the continuously-flowing
analyte gas.
1.2
1 ppmv CO @ T1 and P1
r Wa
CH4 (ppmv)
0.0
Wavelength
monitor
1.4
• Temperature
• Pressure
1549.8 1550.0 1550.2
Wavelength (nm)
Lase
Juarez
1.0x10-6
1549.6
A
Los Angeles
CO2 (ppmv)
Tunable
diode laser
2.0x10-6
• Pressure control to better than 1 part in 1000
Indicated CO2 Concentration (ppmv)
Outlet gas
flow
(to pump)
3.0x10-6
In a given gas matrix, only two parameters affect the lineshapes
of these features:
A key feature of the WS-CRDS analyzer is the wavelength
monitor which allows the wavelength of the laser to be precisely
controlled and positioned accurately along the absorption
profile of the molecule. This enables fast data acquisition as
the entire peak profile does not need to be measured. Further,
this wavelength certainty ensures that the system is
interrogating the molecule of interest and not a nearby
absorption peak of an interfering species.
• Temperature control to better than 1 part in 15,000
Optical Absorption (parts-per-billion/cm)
Temperature
gauge
4.0x10
-6
Detector Signal
High-finesse optical
“Ring-down” cavity with
3 high-reflectivity mirrors
Absorbance (1/cm)
Pressure
gauge
Icirc(t) = Icirc(t0)exp -tτ
Accurate gas and isotope measurements require stable
spectroscopic features.
The Picarro wavelength monitor:
Spectroscopic GPS
δ18O vs. V-SMOW (‰)
[]
5.0x10-6
In the WS-CRDS analyzer, the temperature and pressure of the
gas are also tightly controlled enabling virtually drift and
calibration-free measurements over long periods of time.
Optical Absorption (parts-per-billion/cm)
Sample
gas inlet
Measure decay time
using CDRS
Temperature & Pressure control
Optical Absorption (parts-per-billion/cm)
Select wavelength
using wavelength monitor
What are the minimum requirements for
high-stability spectroscopic
measurements?
δD vs. S-SMOW (‰)
How can a measurement of time be used to quantify concentrations or isotope ratios?
0.1
Relative Frequency (Wavenumbers)
Use real-time nonlinear spectral analysis to quantify different
constituents in the sample. Using the shape of each line,
extrapolate the baseline under the analyte peak. This process
requires a highly accurate wavelength monitor
0.2
Conclusion:
This WS-CRDS technology has been pivotal in developing gas
an isotope analyzers capable of being deployed in the field,
unattended, for long periods of time, enabling measurements to
be done more simply, at lower cost, by a greater number of
scientists, moving information-rich, laboratory-quality
measurements out of the lab. The combination of this
technology and its validation has resulted in highly robust, easy
to use instruments that provide highly sensitive and stable data
even in challenging environmental conditions. No other gas or
isotope analyzer has, to date, been demonstrated to achieve
superior precision, accuracy and long-term stability in field
conditions without calibration.