Detection of methane with mid-infrared polarization spectroscopy

Abstract

We report on the application of polarization spectroscopy in the mid-infrared spectral region for the detection of methane by probing its asymmetric ro-vibrational transitions. Tunable infrared laser radiation, near 3.4m, was obtained from difference-frequency generation in a LiNbO3 crystal. Infrared polarization spectroscopy (IRPS) spectra of the P, Q and R branches of the 3 band, recorded with both linearly and circularly polarized pump beams, are presented. The experiments were performed in an atmospheric pressure gas jet with methane diluted with Ar. An IRPS spectrum with signal-to-noise ratio better than 104 was observed. The dependence of the IRPS signal intensity on the methane mole fraction and on the pumping laser power density was investigated.

Full text

DOI: 10.1007/s00340-004-1475-9
Appl. Phys. B 79, 135–138 (2004)
Rapid communication
Lasers and Optics
Applied Physics B
z.s. li
1,u
m. rupinski
1
j. zetterberg
1
z.t. alwahabi
2
m. ald
´
en
1
Detection of methane with mid-infrared
polarization spectroscopy
1
Division of Combustion Physics, LTH, Lund University, Box 118, 221 00 Lund, Sweden
2
School of Chemical Engineering, University of Adelaide, SA 5005, Australia
Received: 9 January 2004/Revised version: 22 April 2004
Published online: 17 May 2004 • © Springer-Verlag 2004
ABSTRACT We report on the application of polarization spectroscopy in the mid-
infrared spectral region for the detection of methane by probing its asymmetric
ro-vibrational transitions. Tunable infrared laser radiation, near 3.4 µm, was obtained
from difference-frequency generation in a LiNbO
3
crystal. Infrared polarization spec-
troscopy (IRPS) spectra of the P, Q and R branches of the ν
3
band, recorded with both
linearly and circularly polarized pump beams, are presented. The experiments were
performed in an atmospheric pressure gas jet with methane diluted with Ar. An IRPS
spectrum with signal-to-noise ratio better than 10
4
was observed. The dependence of
the IRPS signal intensity on the methane mole fraction and on the pumping laser power
density was investigated.
PACS 33.20.Ea; 52.35.Mw
Polarization spectroscopy (PS) was
firstly introduced by Wieman and Hän-
sch [1] as a Doppler-free spectroscopic
method, which related to saturation
spectroscopy but offered a consider-
ably better signal-to-background ratio.
In a typical PS setup [2], a strong pump
beam and a weak probe beam, tuned to
optical transitions of the target species
with a common ground state or excited
state, are crossed at the interrogated
region. The optical pumping of the tar-
get species with the polarized pump
beam produces birefringence and in-
duces detectable polarization changes
in the weak probe beam. The PS tech-
nique has the unique property of allow-
ing identification of spectral transitions
belonging to
P, R or Q branches by
an evaluation of the
J-dependence of
the absorption cross sections [3, 4]. As
a coherent technique with the signal
generated in a laser-like beam, PS has
obvious advantages for efficient collec-
tion of signal and discrimination against
u Fax: +46-46/222-4542, E-mail: zhongshan.li@forbrf.lth.se
background from scattered light and
chemiluminescence. With almost all the
merits of other laser diagnostic tech-
niques, like high temporal, spatial and
spectral resolution, species-specific de-
tection and low detection limits, PS has
been widely used in combustion and
plasma diagnostics, e.g. with single-
photon excitation for
OH [5], C
2
[6],
NH [7] and NO [8] and with two-photon
excitation for
NH
3
[9], CO [9] and
N
2
[10, 11]. Two-dimensional PS de-
tection of
OH [12, 13] has also been
demonstrated. Comprehensive calcula-
tions based on direct numerical integra-
tion [14] have been performed aimed at
a quantitative detection with PS.
However, most of the PS studies
heretofore were limited to the ultra-
violet
/visible spectral region by ex-
citing electronic transitions. Probing
the molecular ro-vibrational transition
by infrared (IR) excitation has always
been attractive to the combustion di-
agnostic community. Many important
combustion species such as
CO
2
, CO,
H
2
O, CH
4
, etc., which have no con-
veniently accessible electronic transi-
tions, are detectable in the mid-IR spec-
tral range. Due to the poor availabil-
ity of proper tunable IR laser sources,
low sensitivity of the practically avail-
able infrared detectors and the rela-
tively low fluorescence quantum yields,
only limited spatially resolved laser-
based combustion diagnostic experi-
ments in the mid-IR spectral region
via ro-vibrational transitions have been
reported. Laser-induced fluorescence
detection of
CO and CO
2
via IR ex-
citation of the overtone and combina-
tion bands was demonstrated by Kirby
and Hanson [15, 16]. Settersten et al.
investigated the ultraviolet
/IR double
resonance in detection of
CH
3
[17]
and
OH [18]. IR degenerate four-wave-
mixing spectroscopy has been applied
in the detection of
C
2
H
2
and CH
4
in
low-pressure chambers by probing the
C−H asymmetric vibrations [19–22].
Recently, detection of
CO
2
via IR po-
larization spectroscopy (IRPS) has been
investigated by probing rotational lines
belonging to different overtone and
combination bands [23, 24]. We re-
port, for the first time to your know-
ledge, on the detection of
CH
4
with
IRPS making use of the
C−H asymmet-
ric stretching ro-vibrational transitions
near
3000 cm
−1
. Methane is an import-
ant fuel and combustion intermediate.
The applicability of IRPS as a sensitive
optical technique in spatially resolved
CH
4
concentration measurements is
presented in this communication.
The sample-gas mixture of
CH
4
and
Ar was prepared in a 10-mm-diameter
atmospheric pressure gas jet with an
Ar

136 Applied Physics B – Lasers and Optics
co-flow in the coaxial tube. The CH
4
and Ar flows were controlled separately
with mass-flow controllers (Bronkhorst
HIGH-TECH). The complete mixing of
the two gases was ensured by sending
the gases through a more than
10-m-
long,
6-mm-diameter plastic tube from
the flow meters to the optical interro-
gating region. The mole fraction of
CH
4
in the gas jet was varied by changing
the relative flow speed of the mass-
flow meters. A schematic view of the
experimental setup is shown in Fig. 1.
The experiments employed an injection-
seeded single-longitudinal-mode
Nd :
YAG
laser (Spectra Physics, PRO 290-
10) operated at a repetition rate of
10 Hz
and a pulse length of 8ns. The sec-
ond harmonic at
532 nm from the Nd :
YAG laser was used to pump a tunable
dye laser (Sirah, PRSC-D-18) operated
with styryl 9 dye. The residual funda-
mental beam after frequency doubling
at
1064 µm was difference-frequency
mixed in a
LiNbO
3
crystal with the
dye-laser output centered at
805 nm,
and a tunable IR laser beam was gen-
erated at
3.4 µm with a pulse energy
of approximately
1mJ. The bandwidth
of the IR laser was estimated to be
less than
0.04 cm
−1
from the line width
of the dye laser (
0.03 cm
−1
, speci-
fied by the manufacturer) and from
the injection-seeded
Nd : YAG laser
(about
100 MHz). The horizontally po-
larized IR beam was combined and
made collinear with a
HeNe laser beam
by the use of a
CaF
2
plate. The reflected
IR beam from the
CaF
2
plate was dir-
ected to a liquid-
N
2
-cooled HgCdTe
photovoltaic infrared detector (Infrared
Associates, HCT-100C) to monitor the
laser pulse-to-pulse jitter. The co-pro-
pagating geometry PS setup was utilized
in this experiment. The probe beam,
a
0.5% reflection from a CaF
2
plate, was
focused with a
90-cm fused-silica lens.
The transmitted part of the IR beam was
reflected by an aluminum mirror and fo-
cused with a
55-cm fused-silica lens to
serve as the pump beam. The pump and
probe beams were crossed in the middle
of the interrogation region with an angle
of
6.3
◦
. Fused silica has a transmis-
sion window in the 3- to
3.5-µm spec-
tral region (with absorption
< 1% per
mm [25]). This useful property makes
fused silica an obvious choice for op-
tics to detect the
C−H stretching band
of hydrocarbons. A quarter-wave or
FIGURE 1 Experimental setup. BS, beam splitters; M, mirror; F, narrow-band filter; WP, wave plate;
P, polarizer; L, lens
FIGURE 2 IRPS spectra of the ν
3
band of CH
4
. a With the pump beam linearly polarized and ori-
ented 45
◦
to the probe beam; b with the pump beam circularly polarized; c expanding of the P(5) line
to visualize the symmetric fine structures

LI et al. Detection of methane with mid-infrared polarization spectroscopy 137
half-wave plate was placed before the
focusing lens in the pump beam to ma-
nipulate the pump-beam polarization,
either circular or linear (oriented
45
◦
relative to the probe beam for the lin-
ear case). Two
YVO
4
infrared polarizers
(Newphotons, PGL0312) were utilized
crossed with each other over the inter-
rogation region in the probe beam. The
extinction ratio of the IR polarizer pair
was measured to be
6.6 × 10
−7
with
a previously published method [24].
The PS signal beam was focused with
a
30-cm CaF
2
lens through an aperture
to a liquid-
N
2
-cooled InSb photovoltaic
infrared detector (Judson, J10D). The
transient signals from the two IR detec-
tors were collected, time integrated and
stored in a
3-GHz bandwidth digital os-
cilloscope (Lecroy, WaveMaster 8300),
which was triggered by the
Q-switch of
the
Nd : YAG laser.
Typical IRPS spectra of the
P, R
and Q branches in the ν
3
band of CH
4
are shown in Fig. 2, which were ob-
tained with
0.5mJ per pulse in the
pump beam and with
2% CH
4
in the
gas mixture; 10 laser shots were aver-
aged for each data point. The spectra
in Fig. 2a and b were recorded with
linearly polarized and circularly po-
larized pump beams, respectively. The
ability of spectral line discrimination
of PS, as earlier described and experi-
mentally proved with electronic transi-
tions [3, 4], is also confirmed here by
IRPS with ro-vibrational transitions. Al-
though the unresolved
Q-branch lines
make the quantitative comparison with
predicted intensity ratios impossible,
it is obvious that the linear pumping
geometry favors the
Q-branch lines
while the circular pumping geometry
favors the
P-andR-branch lines. Me-
thane exhibits a complex spectrum with
each
J rotational level split into tetra-
hedral components labeled by sym-
metry species
A
1
, A
2
, E, F
1
or F
2
,
andanorderingindex
N [26]. A se-
lected part of the spectrum in Fig. 2b
was expanded and presented in Fig. 2c.
The symmetric fine-structure lines of
P(5) were partially resolved. The line
position and assignment in [27] were
adopted in line identification and nota-
tion in Fig. 2. Owing to the high extinc-
tion ratio of the polarizer pair adopted
in the experiments and the low light
scattering in the infrared region, the
background was almost undetectable.
The signal-to-noise (S
/N) ratio of the
spectrum in Fig. 2b was estimated to
be better than
10 000 : 1 and the prob-
ing volume was calculated to be
6.2 ×
0.7
2
mm
3
from the adopted experimen-
tal geometry.
The power dependence of the IRPS
signal was studied by varying the pump-
laser power with different neutral-den-
sity filters. IRPS signal line-center in-
tensity versus pumping power is plotted
on a log–log scale in Fig. 3 for the
R(3)
line with 1.7% CH
4
mole fraction. The
circular pumping geometry was adopted
in the studies of
R-branch lines. A lin-
ear fit to low-power data points up to
FIGURE 3 Dependence of the IRPS signal intensity on the pumping laser power density. The laser
frequency was fixed on the line center of the R(3) line. The solid line in the figure represents a linear fit
to the low-power experimental data points up to 1.32 MW/cm
2
and K is the slope of the line
FIGURE 4 Dependence of the line-integrated IRPS signal intensity on the CH
4
mole fractions. The
solid line in the figure represents a linear fit to the low CH
4
mole fraction experimental data points up to
3.56% and K is the slope of the line
1.32 MW/cm
2
yields a slope of 1.5. The
divergence from quadratic dependence
might be due to a slight saturation and
beam geometric effects. It is evident
from the figure that strong saturation
starts with pump powers higher than
3MW/cm
2
.
Taking advantage of the possibility
to precisely control the
CH
4
molecule
density, the dependence of the IRPS
signal on the
CH
4
mole fraction was
investigated using the
R(3) line with
3-MW/cm
2
pump power. Shown in
Fig. 4 is a curve of the line-integrated
IRPS signal intensity versus
CH
4
mole
fraction. A linear fit to the data points

138 Applied Physics B – Lasers and Optics
with a CH
4
mole fraction lower than
3.56% yields a slope of 1.1. The dis-
crepancy from the generally expected
quadratic dependence on number dens-
ity may be caused partially by the ab-
sorption of the probe beam and par-
tially by the relatively lower number
density in the probed volume due to
gas diffusion in the gas jet. A simu-
lation of the absorption of the
R(3)
line using the HITRAN database [28]
gives an absorption of
85% at the line
center through the
10-mm, 4% CH
4
mole-fraction jet. The PS signal for
a
CH
4
mole fraction of more than 4%
is in the optically thick region and an
extended model is needed for a cor-
rect interpretation of the experimental
results.
In order to test the detection limit,
a scan of the
R(4) line was performed
with
0.2% of CH
4
(this is the small-
est mole fraction that can be used for
the present setup with a reliable value
oftheflowspeed).TheS
/N ratio was
estimated to be better than
400 : 1 for
the recorded IRPS spectrum. This indi-
cates a detection limit of
100 ppm for
the
R(4) line with the present setup. The
noise originates mostly from electronic
noise, which may be improved by in-
creasing the signal intensity with either
a stronger probe beam or with a more
powerful pre-amplifier and proper elec-
trical shielding.
In the UV
/visible spectral range, PS,
as a sensitive coherent technique, has
been widely utilized in laser combustion
diagnostics. However, the applications
of PS in the IR spectral range, by prob-
ing molecular ro-vibrational transitions,
are still underdeveloped. There have so
far only been two reports on IRPS de-
tections of
CO
2
in this field [23, 24]
and detailed investigation of this tech-
nique was hindered by the low S
/N
ratio (
50 ± 25) achieved in the previous
work [23]. In the present communi-
cation, we have demonstrated the ap-
plication of mid-IR polarization spec-
troscopy for detection of
CH
4
in at-
mospheric pressure cold flows. To our
knowledge, this represents the first re-
port of the detection of a hydrocarbon
with polarization spectroscopy in the
mid-IR spectral region by probing ro-
vibrational transitions. An S
/N ratio
better than
10
4
has been observed and
a detection limit of
100 ppm in an at-
mospheric pressure cold flow was es-
timated from the experimental meas-
urements. These results indicate that
IRPS provides a sensitive optical diag-
nostic tool for the methane molecule,
and it holds great promise for the de-
tection of other polyatomic molecules
with IR-active ro-vibrational transi-
tions. Encouraged by the low detec-
tion limit obtained for methane, efforts
are being made to detect methyl radi-
cals in a methane
/air flame in order to
achieve spatially resolved
CH
3
detec-
tioninflames.
ACKNOWLEDGEMENTS This re-
search was supported by the Swedish Research
Council and the Swedish Energy Administration.
REFERENCES
1 C. Wieman, T.W. Hänsch: Phys. Rev. Lett. 36,
1170 (1976)
2 W. Demtröder: Laser Spectroscopy: Ba-
sic Concepts and Instrumentation, 2nd edn.
(Springer, Berlin, New York 1996)
3 R. Teets, R. Feinberg, T.W. Hänsch,
A.L. Schawlow: Phys. Rev. Lett. 37, 683
(1976)
4 R.E. Teets, F.V. Kowalski, W.T. Hill, N. Carl-
son, T.W. Hänsch: Proc. SPIE 113, 80 (1977)
5 K. Nyholm, R. Maier, C.G. Aminoff, M. Kai-
vola: Appl. Opt. 32, 919 (1993)
6 K. Nyholm, M. Kaivola, G. Aminoff: Appl.
Phys B 60, 5 (1995)
7 A.A. Suvernew, A. Dreizler, T. Dreier, J. Wol-
frum: Appl. Phys. B 61, 421 (1995)
8 B. Löfstedt, R. Ritzon, M. Ald
´
en: Appl. Opt.
35, 2140 (1996)
9 K. Nyholm, R. Fritzon, N. Georgiev, M. Al-
d
´
en: Opt. Commun. 114, 76 (1995)
10 C.F. Kaminski, B. Löfstedt, R. Fritzon,
M. Ald
´
en: Opt. Commun. 129, 38 (1996)
11 C.F. Kaminski, T. Dreier: Appl. Opt. 39, 1042
(2000)
12 K. Nyholm, R. Fritzon, M. Ald
´
en: Opt. Lett.
18, 1672 (1993)
13 J. Reppel, Z.T. Alwahabi: Appl. Opt. 41,
4267 (2002)
14 T.A. Reichardt, R.P. Lucht: J. Chem. Phys.
109, 5830 (1998)
15 B.J. Kirby, R.K. Hanson: Appl. Phys. B 69,
505 (1999)
16 B.J. Kirby, R.K. Hanson: Appl. Opt. 41, 1190
(2002)
17 T.B. Settersten, R.L. Farrow, J.A. Gray:
Chem.Phys.Lett.369, 584 (2003)
18 T.B. Settersten, R.L. Farrow, J.A. Gray:
Chem.Phys.Lett.370, 204 (2003)
19 G.J. Germann, A. McIlroy, T. Dreier, R.L. Far-
row, D.J. Rakestraw: Ber. Bunsen-Ges. Phys.
Chem. 97, 1630 (1993)
20 G.J. Germann, R.L. Farrow, D.J. Rakestraw:
J. Opt. Soc. Am. B 12, 25 (1995)
21 Y. Tang, S.A. Reid: Chem. Phys. Lett. 248,
476 (1996)
22 D. Voelkel, Y.L. Chuzavkov, J. Marquez,
S.N. Orlov, Y.N. Polivanov, V.V. Smirnov,
F. Huisken: Appl. Phys. B 65, 93 (1997)
23 S. Roy, R.P. Lucht, A. McIlroy: Appl. Phys. B
75, 875 (2002)
24 Z.T. Alwahabi, Z.S. Li, J. Zetterberg, M. Al-
d
´
en: Opt. Commun. 233, 373 (2004)
25 J.H. Moore, C.C. Davis, M.A. Coplan: Build-
ing Scientific Apparatus, 2nd edn., 11th print-
ing (Addison-Wesley, New York 1997)
26 W.L. Barnes, J. Susskind, R.H. Hunt,
E.K. Plyler: J. Chem. Phys. 56, 5160 (1972)
27 A.S. Pine: J. Opt. Soc. Am. 66, 97 (1976)
28 HITRAN database, based on L.S. Rothman,
A. Barbe, D. Chris Benner, L.R. Brown,
C. Camy-Peyret, M.R. Carleer, K. Chance,
C. Clerbaux, V. Dana, V.M. Devi, A. Fayt,
J.-M. Flaud, R.R. Gamache, A. Goldman,
D. Jacquemart, K.W. Jucks, W.J. Lafferty,
J.-Y. Mandin, S.T. Massie, V. Nemtchinov,
D.A. Newnham, A. Perrin, C.P. Rinsland,
J. Schroeder, K.M. Smith, M.A.H. Smith,
K. Tang, R.A. Toth, J. Vander Auwera,
P. Varanasi, K. Yoshino: J. Quantum Spec-
trosc. Radiat. Transfer 82, 5 (2003)