Transcript 2015 ISMS TALK Okumurax

Breathing Easier Through
Spectroscopy: Studying Free Radical
Reactions In Air Pollution Chemistry
by Frequency Comb Spectroscopy
Mitchio Okumura
California Institute of Technology
Days of Ozone Exceedances,
Unhealthy for Sensitive Groups; 2001-05 vs. 2006
US EPA
Photochemical Smog Cycle:
Catalytic Free Radical Chain Reactions
of Trace Species
+ O2
RH
R
+ O2
RO2
RO
R' HO
+
OH

NO2
NO

HO2
hv
Aerosol
HO2
O
O2
O3
Time-resolved Spectroscopy Of Gas Phase Reactions
Norrish and Porter's Flash Photolysis Experiment
time
NO (v) vs time
Time-Resolved Free Radical Kinetics
Caltech Group
1. Pulsed and cw Cavity Ringdown Spectroscopy
2. VUV Photoionization Mass Spectrometry (Sandia, ALS)
3. Frequency Comb Laser Spectroscopy (Ye group, JILA)
PROOF OF PRINCIPLES EXPERIMENT:
Detection of the DOCO Radical by
Time Resolved Frequency Comb Spectroscopy (TRFCS)
J. Phys. Chem. Letters, 5, 2241-1146 (2014).
Adam Fleisher, Bryce Byork, Kevin Cossel, Bryan Changala, Benjamin Spaun,
Oliver Heckl , and Jun Ye
JILA, NIST and University of Colorado
Thinh Bui, Mitchio Okumura
California Institute of Technology
Mid-IR Cavity Ringdown Spectrometer with
Variable T Flow Cell
Frequency
Doubled
Nd-YAG
Dye Laser
Digital Delay
Generator
Photolysis Length 4 cm
OPA
Mid-IR Light
N2 Purge
Gas Exit
N2 Purge
Cooled
Cell
InSb
Detector
Mirror
R > 99.98%
Mirror
R > 99.98%
Gas Entrance
Excimer Laser
The HO2 + HCHO Reaction
Can we detect the HMP intermediate?
H
O
H
H
+
O
O
- 29 kJ/mol
IR
- 62 kJ/mol
NIR
hydroxymethylperoxy (HMP) radical
IR-CRD Spectra of Hydroxy-Methyl Peroxide
Matthew Sprague, Laura Mertens, Heather Widgren (CALTECH)
Anne McCoy (OSU), Stanley P. Sander (NASA JPL)
J. Phys. Chem. A, 117 10006 (2013)
v1 OH Stretch band
mid-IR
AX Electronic Transition
near-IR
Pros and Cons of "Low Resolution" Pulsed CRDS
for Free Radical Kinetics
• broadly tunable
• good sensitivity
IR-UV
nmin ~ 1011-1013 /cc)
– can require high concentrations of precursors, radicals; fast rates
• time resolution
1-100 us
– 'SKAR' deconvolution required for fast processes (t ≲ τringdown)
• Linewidth issues
Γlaser > 40 x Γline
– Intensity and lineshape are problems
• Low rep rate
10's Hz to 1 kHz
– one time-point per excimer shot in time-dependent kinetics
Mid-Infrared Frequency Comb System
Collaboration with Jun Ye
Experiments at JILA, University of Colorado, Boulder
F. Adler et al. Opt. Lett.
34, 1330 (2009).
Motivation:
The OH + CO ➝ H + CO2 Reaction
IN THE ATMOSPHERE:
• Rate Constant k (298,1 atm) ≈ 2 × 10−13 cm3 s−1
• CO concentrations
Free Troposphere
Polluted urban air
[CO] ≈ 50-200 ppb
[CO] ≈ 100-2000 ppb
• OH + CO is the major tropospheric sink for OH globally
The OH + CO ➝ H + CO2 Reaction
• A key reaction in combustion and the atmosphere
• Extensive studies on the kinetics, dynamics and
spectroscopy of this system
• Anomalous Arrhenius Plot: Eact becomes nonlinear at low T
Nguyen et al. J. Phys. Chem. Lett. 2012, 3, 1549
High Pressure
Low Pressure
Reaction path proceed via the HOCO Intermediate
OH + CO ➞
H + CO2
⇄ HOCO*
➞ H + CO2
+M
➞ HOCO
OH + CO
c-HOCO
t-HOCO
-50 kcal/mol
H + CO2
Thanh Lam Nguyen; Bert C. Xue; Ralph E. Weston Jr.; John R. Barker; John F. Stanton; J. Phys. Chem.
Lett. 2012, 3, 1549-1553. DOI: 10.1021/jz300443a
Frequency Comb Lasers
• Leading to in advances AMO Physics (Hansch, Hall, Physics Nobel Prize 2005)
• Metrology and spectroscopy
• Combines advances in laser stabilization, ultrafast lasers, and coherent and
nonlinear optics
CAN FREQUENCY COMB LASERS BE APPLIED TO
FREE RADICAL KINETICS AND DYNAMICS?
Frquency combs are generated by mode-locking
Laser
Crystal
mode locking
mirror
Output
Coupler
TIME
FREQUENCY (TEM00 modes)
Generating a Frequency Comb with
a Mode-Locked Laser
Single femtosecond pulse
Train of pulses
stabilized mode-locked laser
Time domain
Δϕ
2Δϕ
E(t)
1/frep
Frequency domain
frep
FREQUENCY COMB
COHERENT: Laser cavity length & phase must be stabilized
S. T. Cundiff, J. Ye, and J. L. Hall, Scientific American, April 2008
Generating a Frequency Comb with
a Mode-Locked Laser
A train of pulses in phase
at repetition rate frep
𝜏 = 1/ frep
Frequency Comb
Δϕ
E(t)
frep
FREQUENCY COMB
S. T. Cundiff, J. Ye, and J. L. Hall, Scientific American, April 2008
JILA Mid-Infrared Cavity-Enhanced Frequency-Comb
Spectrometer
1. Mid-IR Frequency Comb Source
2. Optical Enhancement Cavity
3. Spatially Dispersive VIPA Spectrometer
Mid-IR Camera
VIPA etalon
(3)
(2)
(1)
Mid-IR Time-Resolved Frequency Comb System
Imaging
Detection
MODE-LOCKED
FEMTO-SECOND
INFRARED LASER
Yb:fiber laser
MIR
Comb
Probe
Wavelength
Dispersion
Optical CavityEnhanced
Absorption
Mid-IR OPO
PZT
Pulsed-Laser UV
Photolysis in
Flow Cell Reactor
ArF excimer laser
BS
InSb
array
VIPA +
grating
1. Mid-Infrared Frequency Comb Laser
Comb-Laser-Pumped Optical Parametric Oscillator
Mid-Infrared Radiation tunable between 2.8-4.8 μm
10 W, 1 μm
Yb Fiber
Comb
> 500 mW
idler power
2.8 – 4.8 μm
Synchronously
Pumped PPLN OPO
F. Adler et al. Opt. Lett.
34, 1330 (2009).
Mid-Infrared Frequency Comb Laser
TUNING RANGE
2.7 – 4.8 microns
ϕo+nϕo
ϕo
FREQUENCY COMB
FSR = 136 MHz
Bandwidth
~150- 300 nm
7.35 ns
Current limitation: Best if
λ > 2.9 microns
due to H2O absorption in OPO Cavity
CAN OBSERVE
OD stretches but not OH stretches
Frequency
MODE SPACING (FSR)
frep = 136 MHz
order (104 modes)
SINGLE MODE LINEWIDTH
~ 40 kHz
f
2. Comb beam is injected into an external cavity
mid-IR Laser
frep=136 MHz
Reaction Cell/
High Finesse Optical Cavity
FSR = 273 MHz
Dispersion/
Detection
VIPA spectrometer/
InSb Camera
Mid-IR
Mode-locked laser
99.7% < R < 99.95%
Finesse 1000-6000
M.J. Thorpe et al. Opt. Express 16, 2387 (2008).
A. Foltynowicz et al. Appl. Phys. B. doi:10.1007/s00340-012-5024-7(2012).
Cavity Enhanced Spectroscopy
narrow line-width
diode laser
beam
High Reflectivity
Mirrors
R>> 99%
gas cell
High Resolution Spectroscopy
Continuous Wave cw-Cavity Ringdown Spectroscopy
Ringdown Signal seen
when on resonance
Tuning high resolution laser
Ideal: Comb laser excites every resonance
simultaneously
LASER
CAVITY
2. Comb beam is injected into an external cavity
mid-IR Laser
frep=136 MHz
Reaction Cell/
High Finesse Optical Cavity
FSR = 273 MHz
VIPA spectrometer/
InSb Camera
Mid-IR
Mode-locked laser
Feedback Loop
Dither Lock
of Cavity and Laser
Modes
50 kHz
Dispersion/
Detection
99.7% < R < 99.95%
Finesse 1000-6000
3. Dispersive Imaging Detector:
Virtually Imaged Phase Array (VIPA)
L. Nugent-Glandorf et al., Opt. Lett. 37, 3285 (2012).
VERTICAL
HIGH RESOLUTION
(ETALON – VIRTUALLY IMAGED PHASE ARRAY)
HORIZONTAL
LOW RESOLUTION
IR CAMERA
3. Dispersive Imaging Detector:
Virtually Imaged Phase Array (VIPA)
L. Nugent-Glandorf et al., Opt. Lett. 37, 3285 (2012).
2.0 GHz Comb of Cavity-Filtered 3.1 µ OPO Output
Image Plane of InSb Camera
Observed FWHM = 500 MHz
(0.018 cm-1)
VIPA
Dispersion
high
resolution
low resolution
Grating
Dispersion
Detection of N2O Spectrum
N2 reference
0.2% N2O in N2
1930 Elements
7072 Comb Teeth
4. Timing: Limited by IR InSb Camera
193 nm pulse
t=0
τ
~~
InSb
CAMERA
Integration
~~
4 ms
25 μs integration window
pre-photolysis
reference images
Laser:
InSb CAMERA
signal images
100 fs pulse
every 7.35 ns
INTEGRATION TIME
READOUT TIME
VARIABLE EXCIMER/CAMERA DELAY
> 10 μs (25, 50 μs)
4 ms
τ
K.C. Cossel et al., Appl. Phys. B 100, 917 (2010)
Preliminary Experiment: Detecting trans-DOCO
3 mJ/cm2
193 nm
d1-Acrylic Acid
C2H3C(O)OD
Vinyl Radical
C2H3
DOCO
J.T. Petty and C.B. Moore, J. Chem. Phys. 1993, 99, 47-55.
J.T. Petty, J. A. Harrison, and C. B. Moore, J. Phys. Chem. 1993, 97, 11194.
Experimental Conditions:
• Acrylic Acid in N2/CO2
• 10 Torr, 295 K
Chemistry in high-finesse cavity/flow cell
N2
pump
pump
N2
P gauge
M1
M2
N2 purge
Optical Length
Photolysis Region
54.88 cm
3 cm
Typical chemistry conditions:
10 Torr N2 buffer gas
10 mJ 193 nm excimer pulse
[AA] ∼ 3 x1014 cm-3
1-2% dissociation
N2 purge
Laser
range: 2400-2700 cm-1
signal-averaging:
rep rate 0.2 Hz
typ. 960 averages
VIPA bandwidth: 65 cm-1
7072 comb teeth
1930 Resolvable elements
Camera integration time:
25 µs or 50 µs
Normalized Transmission ΔI/I0
t=0
Ref
Signal
Wavenumber / cm-1
average
950 shots
Time / ms
Observed and modeled spectrum at τ = 0 μs
50 μs camera integration window, 950 averages
10-7 cm-1
DOCO Detection
2665
Frequency, cm-1
Minimum detectable DOCO concentration
(3 cm pathlength)
5 x 1010 cm-3 in Fe
960 shots
(using ab initio band strength with quartic f.f.)
VIPA
Bandwidth = 1.93 THz (65cm-1)
Resolution = 1 GHz (0.03 cm-1)
275
Spectra observed in 3665-2705 cm-1
t = 0 ms
t = 4 ms
trans-DOCO Simulation
HOD Simulation
t = 80 ms
D2O Simulation
Time-resolved decay of products
(4 ms time intervals)
Concentration (ppm)
Transient absorption following photolysis
D2O
HOD
DOCO
Time (ms)
Time-resolved decay of products
(25 μs time intervals)
HOD
trans-DOCO
2-3 x 1012 cc
Average
Concentration (ppm)
Incrementing τ, variable excimer-camera delay time
Time (µs)
D2O
Time evolution of spectra
(Fast timescale – excimer/probe delay)
A new product: Acetylene-d1
HCCD
PROMPT PRODUCT YIELDS
DOCO : HOD : HCCD ≈ 1 : 2 : 0.15
Previous Observations on
UV Photolysis of Acrylic Acid
C3
C1
D
C3-O and C2-C3 bond fission
C2
Why do we observe prompt HOD? HCCD??
Bulk experiments (UV photolysis)
HOCO (+ products) C-C bond fission
CO2 (+ products)
Decarboxylation
CO (+ products)
Decarbonylation
OH (+ products)
C-O bond fission
Molecular Beam Dissociation (Butler & co-workers)
HOCO + C2H3
OH + C3H3O
What are the sources of isotope exchange products?
C3
C1
C2
DOCO
HOD, HCCD ??
•
Secondary chemistry? No. Bimolecular reactions cannot occur
in first 50 us at these concentrations
•
Wall reactions? No. (probe volume 1 cm from walls)
•
Water impurities? AA-H2O clusters? No. All HOD converts to
D2O, system is passivated.
•
Impure precursors (exchange on walls) possibly!!
•
AA dimers or clusters? PAA = 20mTorr is too low, but Acrylic
acid is VERY STICKY
Possibly primary AA photochemistry??
Excited States of Acrylic Acid
C1
C3
PRODUCTS
S2
<148 kcal
30 Torr He
-Fang & Liu et al. (2000).
JACS. 112, 10886
193 nm
(148 kcal)
S1
88 kcal
T3
112 kcal
T2
87 kcal
T1
71 kcal
S0
Acrylic Acid Photolysis at 193 nm:
Ab Initio Dynamics on the
Excited Singlet S2 State
Dorit Shemesh and R. Benny Gerber
Hebrew University, UC Irvine and
University of Helsinki
Methodology:
Dynamics on S2 surface (only)
Semi-empirical OM2 and OM2/MRCI potentials
Ground State (300K) (initial wavepacket):
optimized with OM2
Molecular dynamics on OM2/MRCI surface (10 ps)
100 structures as initial configurations
Assume Vertical Transition
Excited State
Molecular dynamics on OM2/MRCI surface (100 ps)
Predicted Product Yields
*
*
*
*
* Observed products
Experimental yields are not in agreement
*
*
Synchronous Channel for Formation of HOD
C2H3C(O)OD ➞ HOD + HCCH + CO
• Synchronous formation of CO and acetylene C2H2
Sequential Formation of C2HD
(Limited number of trajectories)
• H2 elimination via
transient CH3-C-C(O)OD
• D-atom transfer in
transient CH-C-C(O)OD
• Dissociation of CO2
and C2HD
• H or D atom loss in
some trajectories
Caveats
•
•
•
•
Semi-empirical S2 surface, corrected
Classical trajectories
Dynamics on single excited state surface
No surface crossings to other states:
Internal conversion (IC) to S1, S0
Intersystem crossing (ISC) to T2, T1
• Experiment needs to be repeated
Summary: Mid-IR
Time-Resolved Frequency Comb Spectroscopy
• Broadband absorption spectra in the mid-IR
Tuning Range 2.8-4.8 µ
Single-shot bandwidth 65 cm-1 (VIPA-limited)
• High sensitivity approaching single mode cw-CRDS
DOCO sensitivity 5x1010/cm-3 in 960 shots (limited by flow
cell)
3 cm path length
• VIPA-limited spectral resolution
≥ 0.018 cm-1
• Good time-resolution
10 - 20 µs
Camera integration, Cavity response time
Full time
• Cavity-enhanced sensitivity
Caveats for MIR TRFCS
• Best sensitivity for narrow lines (rotationally resolved)
• Mid-IR experiments limited by detection system
– InSb Camera has limited sensitivity, range, pixels (and $$$$)
– Better performance in NIR, Visible (Faster, cheaper, better)
• Cavity Enhanced, not Cavity Ringdown
– need to measure ΔI / I0
• Cannot yet resolve single comb lines
Improvements
• Improved Mid-IR mirrors (throughput 20%)
• Flow cell (limited to < 1Hz photolysis rep rate)
Other Approaches
• Frequency Comb Velocity Modulation (Ye, Cornell)
• Dual Frequency Comb Spectroscopy
Acknowledgments
JUN YE GROUPJILA|NIST and CU
AJ Fleisher, Bryce J. Bjork, Bryan Changala, Ben Spaun, Oliver Heck, Kevin C. Cossel
Alexandra Foltynowicz, Piotr Masliowski
OKUMURA GROUP Caltech
Thinh Bui
UC IRVINE/HEBREW UNIVERSITY
Dorit Shemesh, R. Benny Gerber