Transcript Columbus-2009 - University of Virginia

Broadband Cavity Enhanced
Absorption Spectroscopy With a
Supercontinuum Source
Paul S. Johnston
Kevin K. Lehmann
Departments of Chemistry & Physics
University of Virginia
In the past decade, the use of low loss optical
cavities have become widely used to
achieve high sensitivity absorption
spectroscopy. Multiple variants, including
Cavity Ring-Down Spectroscopy
NICE-OHMS
Cavity Enhanced Absorption Spectroscopy
Dielectric Super Mirrors ( < 100 ppm reflection
loss) have been key to the sensitivity
enhancements of these methods.
Limitations of Super Mirrors
• High reflectivity bandwidth of dielectric mirrors
limited to a few % in wavelength
– They are 1-dimensional photonic crystals.
• One can extend bandwidth by chirping the
coatings of the layers but this increases loss and
dramatically reduces damage threshold
• High reflectivity (>99.9%) is only available for far
less than an octave in the spectrum.
• Need to use reflectors that do not depend upon
interference.
Brewster Angle Prism Retroreflector
Ring-down Resonator
Output
Input
qb
P- polarization
qb
6 meter radius
of curvature
G. Engel et al., in Laser Spectroscopy XIV International Conference,
Eds. R. Blatt et al. pgs. 314-315 (World Scientific, 1999).
Advantages of Prism Cavity
• Wide spectral coverage - Simultaneous
detection of multiple species
• Compact ring geometry (no optical
isolation required)
• No dielectric coatings (harsh
environments)
• Coupling can be optimized for broadband
• Analysis: Paul S. Johnston & KKL,
Applied Optics 48, 2966-2978 (2009)
What are loses of Prism Cavity?
• Deviation from Brewster’s Angle
R

 q
n 4 1
4n 6
2
2
 96 ppmq / deg
2
• Surface Scattering at optical surfaces

– Need super polishing
• Bulk Absorption and Scattering Losses
– Rayleigh Scattering Dominates for fused silica prisms
• Birefringence which converts P -> S polarization
– Strain must be minimized.
dq B
 1 n 2
dT


1
dn 
  40 rad / K
dT 
Surface Scattering Losses
• Surfaces super polished to s = 1 Å rms
• Loss for each total internal reflection:
– (4  n cos(q) s / 0)2 = 0.15 ppm for 0 = 1 m
• Loss for each Brewster Surface:
 s n 2 1 2


2 
  0.8 ppm / surface
2

0
n 1 
• Total < 2 ppm/prism at 1 m.
• Finite angular spread of beam leads to <0.02 ppm

loss
Bulk Loss
• For fused silica, scattering loss dominates
absorption for  < 1.8 m.
– small residual [OH] absorption near 1.4 m.
– Prisms made of Suprasil 3001 which has [OH]
< 1 ppm
• Rayleigh scattering loss ~ 0-4
– loss of ~ 1 ppm/cm @ 1 m.
– Intracavity pathlength of 3.8 cm for our prisms
(16 mm length on short side)
Loss of Prism Cavity in near-IR
(Tiger Optics)
Fused Silica Prisms (built in 43cm cavity)
ring down tau/ppm loss vs. wavelength
Tau Trend
measure Tau
70
150
ppm loss Trend
60
130
50
110
40
90
30
70
20
50
10
30 1310, 1368,
ppm loss
Ringdown(tau-microsecond)
Measure PPM Loss
Tau measurenment at
0
1300
1350
1400
1450
1500
1550
Wavelength(nm)
1600
1650
1377,1392,1522,1531,
1578,1635,1671nm.
10 Every Diode laser
1700 Temp scan from (40
or ) 35~0 Celsisus
degree.
Source for Broad Bandwidth Coherent Radiation:
Supercontinuum Photonic Crystal Fibers
•
•
•
•
•
Material: Pure Silica
Core diameter: 4.8 + 0.2 µm
Cladding diameter: 125 + 3 µm
Zero dispersion wavelength: 1040 + 10 nm
Nonlinear Coefficient at 1060 nm: 11 (W·Km)-1
www.crystal-fibre.com
Supercontinuum Generation
• Fiber
– Length = 12 m
• Input
–
–
–
–
Average power: 1.0 W @ 1064 nm
Rep rate: 29.41 kHz
Pulse energy: 34 J
Peak power: 3.4 kW
• Output
– Average output power: 0.270 W (at
input polarizer)
– Loss of ~50% power through
polarizer
Higher Power Supercontinuum
from mode lock Nd:YAG laser
• Input:
– Spectra Physics Vanguard.
– 80 MHz/ 30 psec pulse train
– Average power: 9.5 W @ 1064
nm
– Peak power: ~10 kW
• Supercontinuum Output
– Average output power: 3.2 W
– With optimized fiber, we expect
higher conversion
Broadband system using white light
from photonic crystal fiber
Paul S. Johnston and KKL, Optics Express, 16, 15013-23 (2008)
Observed Cavity Loss
Model:
Loss  scattering  Brewster' s angle loss
 4
2
A
(n
1)
2
2
Loss 
 2
q

q
1
2

6
4
4n








Cavity enhanced spectroscopy
• Measure time integrated intensity
Io ( ) 1 R
 ( )  
1
 I( )  l
I( )  time integrated intensity with absorbing species
I o ( )  time integrated intensity of empty cavity

• Advantages
– Relatively high
sensitivity
– Simpler set up
• Sensitivity limitations
– Residual mode structure
– Laser noise
Berden, G.; Peeters, R.; Meijer, G. Int. Rev. Phys. Chem. 2000, 19, 565.
O2 SPECTRUM IN AIR
Fifth Overtone Spectrum of C2H2
Allan Variance
• Read CCD every 10 sec
for ~8 hrs
• Successive CCD readings
were binned for time
intervals of Dt.
• Variance calculated for
ratio of spectra for each Dt
pair.
• Minimum noise point: 90
min
650 nm
Current Status
• Absorption Sensitivity 5.88x10-9 cm-1
– Equivalent to 1.6 x 10-9 cm-1 Hz-1/2
– Shot noise limited
– Shot noise limit extends to ~90 min
integration.
• Resolution ~0.05 cm-1 (2 GHz)
– Close to diffraction limit for 25 cm grating
used
• Bandwidth vs. resolution limited by CCD
Improvements....
• Expand simultaneous spectral coverage
– Plan to use FTIR to cover entire spectral range
of super continuum.
– Considering construction of Echelle
Spectrograph which will allow efficient use of
most of CCD pixels.
• CaF2 prisms should allow extension into the
UV, BaF2 prisms into the mid-IR.
I admit it; I’ve got comb envy!
• As already discussed by Jun Ye, a vastly higher
power can be coupled in with a frequency comb
• Hansch’s group has shown how a vernier principle
can be used to get single comb resolution with a
modest resolution spectrograph
• Dispersion limits the spectral width that can be
simultaneously coupled into the cavity
1
D dFSR 
D d 

 ~ 100 GHz
8  FSR  d 
2
c
Acknowledgments
• Dr. Paul Rabinowitz
• Tiger Optics research team
• University of Virginia, National Science
Foundation, and the Petroleum Research
Fund.
White light sources
http://www.crystal-fibre.com