Crabtree_2012

FRONTIERS ARTICLE
Sub-Doppler mid-infrared spectroscopy of molecular ions
Kyle N. Crabtree a,1
, James N. Hodges a
, Brian M. Siller a
, Adam J. Perry a
, Joseph E. Kelly a
,
Paul A. Jenkins II a
, Benjamin J. McCall a,b,⇑
a
Department of Chemistry, University of Illinois, Urbana, IL 61801, USA
b
Departments of Astronomy and Physics, University of Illinois, Urbana, IL 61801, USA
a r t i c l e i n f o
Article history:
Available online 15 September 2012
a b s t r a c t
The technique of velocity modulation spectroscopy has recently been combined with cavity enhancement
and frequency modulation methods into a technique called noise-immune cavity-enhanced optical het-
erodyne velocity modulation spectroscopy (NICE-OHVMS). We have implemented NICE-OHVMS with a
cw-optical parametric oscillator (OPO) tunable from 3.2 to 3.9 lm, and used it to record spectra of the
R(1,0) and R(1,1)u
transitions of the m2 fundamental band of Hþ
3 . The high optical power and cavity
enhancement enable saturation of rovibrational transitions, which allows for line center frequencies to
be measured with a precision of 70 kHz.
Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction
Molecular ions play a key role as intermediates in chemical
reactions, and a detailed understanding of their structure and
intramolecular dynamics in the gas phase, generally obtained by
spectroscopy, is a critical first step toward understanding their
behavior in more complicated systems. The primary technique
used for ion spectroscopy over the past 30 years has been velocity
modulation spectroscopy (VMS) [1,2]. In VMS, ions are produced in
an AC positive column plasma whose polarity is alternated at fre-
quency fvm. The average drift velocity of the ions in the plasma is
shifted toward the cathode from the applied electric field, while
neutral molecules are generally unaffected. As the polarity is re-
versed, the average ion drift velocity also reverses, resulting in a
periodic oscillation in the ion velocity distribution at fvm. By inter-
rogating the ions with a laser beam passing in one direction
through the plasma, the absorption profiles are alternately red-
and blue-shifted with respect to their rest frequencies, and
phase-sensitive detection at fvm allows for selective retrieval of io-
nic signals. VMS therefore addresses one of the main challenges of
ion spectroscopy, i.e., detection of ionic species that are only
$10À5
—10À6
as abundant as neutral molecules. Well over 40 un-
ique molecular ions (not including isotopologues) have been de-
tected with VMS; these have been extensively reviewed by
Stephenson and Saykally [3].
Recently, the use of a Fabry–Perot optical cavity to enhance the
optical path length was demonstrated by Siller et al. (cavity-en-
hanced velocity modulation spectroscopy, CEVMS), who locked a
Ti:Sapphire laser to an optical cavity surrounding a velocity mod-
ulated positive column cell and detected the transmitted light [4].
Because the light in the cavity is bidirectional, red and blue Dopp-
ler shifts are simultaneously superimposed, encoding the velocity
modulation (VM) signal at 2f vm. Initially, this was believed to be
problematic because any neutral molecules produced or excited
by the discharge are concentration modulated (CM) at 2fvm; that
is, the population of the excited species varies with the magnitude
of the applied voltage, but not the sign of the voltage. By also
encoding the ion signal at 2f vm, it was thought the concentration
modulation signal of neutral molecules would overwhelm the ion
velocity modulation signal. Siller et al. showed that the ion and
neutral signals occurred at different phases with respect to the
plasma voltage, thereby preserving ion-neutral discrimination
through phase-sensitive detection. Additionally, the power
enhancement from the optical cavity enables saturation spectros-
copy and precise line-center determination [5]. A related tech-
nique has also been employed using an optical frequency comb
as the light source and a unidirectional ring cavity surrounding
a plasma cell, effectively converting VMS to a broadband tech-
nique while preserving the high resolution of laser spectroscopy
[6].
Cavity enhanced absorption spectroscopy suffers from the fact
that frequency noise in the laser is directly converted into intensity
noise as a result of reduced cavity transmission. This limitation
was overcome by Ye et al. with noise-immune cavity-enhanced
0009-2614/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cplett.2012.09.015
⇑ Corresponding author at: Department of Chemistry, University of Illinois,
Urbana, IL 61801, USA.
E-mail addresses: kcrabtr2@illinois.edu (K.N. Crabtree), jnhodge@illinois.edu
(J.N. Hodges), bsiller2@illinois.edu (B.M. Siller), aperry10@illinois.edu (A.J. Perry),
bjmccall@illinois.edu (B.J. McCall).
URL: http://bjm.scs.illinois.edu (B.J. McCall).
1
Present address: Harvard-Smithsonian Center for Astrophysics, Cambridge, MA
02138, USA.
Chemical Physics Letters 551 (2012) 1–6
Contents lists available at SciVerse ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett

optical heterodyne molecular spectroscopy (NICE-OHMS) [7]. In
this technique, the laser is phase modulated at fh (typically
$100s of MHz), effectively generating an FM triplet consisting of
a carrier (at the optical frequency fo) and a pair of sidebands with
opposite phase at fo Æ fh. The triplet is coupled into the optical cav-
ity by setting fh equal to an integer multiple of the cavity free spec-
tral range (FSR). In the absence of any intracavity absorption or
dispersion, the beat notes between each sideband beating with
the carrier are balanced in both amplitude and phase, so there is
no net signal at fh. However, in the presence of an intracavity ab-
sorber or disperser, the beat notes are unbalanced in amplitude
and/or phase, yielding a net signal. By encoding absorption/disper-
sion information at a high frequency, 1=f technical noise is re-
duced. Another advantage is that any laser frequency noise
affects the cavity coupling efficiency of the carrier and both side-
bands equally, which keeps the beat notes of each sideband with
the carrier balanced, and eliminates direct conversion of laser fre-
quency noise to noise in the final spectrum. The disadvantages are
the complexity of the technique, particularly the demands of main-
taining the laser-cavity lock, and the requirement of a detector
whose bandwidth is at least fh. An extensive review of the NICE-
OHMS technique has been published by Foltynowicz et al. [8].
CEVMS and NICE-OHMS have been combined into a technique
called noise-immune cavity enhanced optical heterodyne velocity
modulation spectroscopy (NICE-OHVMS) in the near-infrared with
a Ti:Sapphire laser [9]. Because fh and fvm are at significantly differ-
ent frequencies (typically $100 MHz and $10 kHz, respectively),
the detector signal is first demodulated at fh, and then sent on to
further phase-sensitive detection at 2f vm. NICE-OHVMS preserves
the ion-neutral discrimination afforded by VMS, and takes advan-
tage of the ultra-high sensitivity, saturation, and noise immunity
of NICE-OHMS.
However, the technical demands of the NICE-OHMS technique
have largely precluded its use in the mid-infrared spectral region
in which VMS has been successfully exploited. The high bandwidth
detectors and phase modulators required for NICE-OHMS are not
as readily available in the mid-IR compared with the visible/
near-IR. The only published mid-IR NICE-OHMS work was done
with a quantum cascade laser near 8.5 lm, and was limited by
the detector bandwidth and the phase modulation characteristics
of the device [10]. Work is currently underway in our laboratory
to extend NICE-OHMS into the mid-IR using a difference frequency
generation (DFG) source in the 3–5 lm region [11], which is partic-
ularly attractive for a general-purpose ion spectrometer because
the vast majority of molecules have at least one fundamental
vibrational band in that portion of the spectrum.
Here we report the first mid-IR NICE-OHVMS spectrometer,
which uses a commercially available cw-optical parametric oscilla-
tor (OPO) tunable from 3.2 to 3.9 lm. The high optical power of the
OPO ($1 W) allows for use of high bandwidth mid-IR detectors
that are not sensitive enough to be used with many other lower-
power cw lasers in this region. This technique enables all of the
advantages of the NICE-OHVMS technique to be brought to bear
on fundamental vibrational transitions of molecular ions, including
high precision sub-Doppler spectroscopy.
Our initial demonstration of this instrument focuses on the
study of Hþ
3 , which is the simplest polyatomic molecular ion and
serves as the primary initiator of ion–molecule chemistry in inter-
stellar clouds [12,13]. Its infrared spectrum was first observed by
Oka in 1980 [14], and since then it has been extensively studied
spectroscopically (see, for instance, the review in Ref. [15]). More
recent research on Hþ
3 focuses on spectroscopy above the so-called
‘‘barrier to linearity,’’ at which point the molecule adopts a linear
geometry that induces a singularity in the Hamiltonian, complicat-
ing its theoretical treatment [16]. Nevertheless, for the lowest-ly-
ing energy levels, the agreement between experiment and ab
initio theory is good to spectroscopic accuracy [17]. Measuring
the energy level spacings experimentally with higher precision
and accuracy would present a greater challenge for ab initio theory,
possibly spurring new developments. The NICE-OHVMS technique
opens the possibility of measuring such energy level spacings in
Hþ
3 .
As a demonstration of the capabilities of this instrument, we
present in this Letter spectra of the m2 fundamental band of Hþ
3 at
3.67 lm. In Section 2 we describe the instrumental details, and
in Section 3 we discuss NICE-OHVMS lineshapes and present spec-
tra acquired with the instrument. Finally, in Section 4, the perfor-
mance of the instrument is assessed, and future directions are
discussed.
2. Experimental details
Our NICE-OHVMS instrument is outlined in Figure 1. A Ytter-
bium-doped fiber laser (YDFL, Koheras Adjustik Y-10) is sent
through a fiber EOM (EOSPACE PM-0K5–00-PFU-PFU-106-S),
amplified (IPG Photonics YAR-10 K-1064-LP-SF), and used to pump
a singly-resonant OPO (Aculight Argos 2400 SF). The pump
(1.064 lm) and signal (1.5–1.6 lm) beams are sent to a wavemeter
(Burleigh WA-1500) for frequency calibration. The idler (3.2–
3.9 lm) is locked with the Pound-Drever-Hall (PDH) technique to
a 1.9-m-long optical cavity consisting of two 1 m radius of curva-
ture concave Si mirrors dielectric coated for 99.7% reflectivity over
3.1–3.4 lm surrounding a plasma cell. Cavity reflection and trans-
mission are monitored by photodiodes (Boston Electronics Vigo
PVM-10.6-1x1) with an effective bandwidth of $125 MHz.
Heterodyne and PDH sidebands are generated by applying volt-
ages at their respective frequencies to the fiber EOM on the seed
laser simultaneously. The resultant frequency spectrum of the
pump laser is imprinted on the idler beam. A PDH error signal used
for cavity locking is generated by demodulating the cavity reflec-
tion signal at the frequency RF1 (typically 2–20 MHz). The error
signal is sent to feedback electronics which stabilize the length
of the cavity on slow timescales via a piezoelectric transducer
(PZT) attached to one of the cavity mirrors, and apply fast correc-
tions to the idler frequency by a PZT mounted to one of the signal
cavity mirrors inside the OPO head. The cavity transmission detec-
tor signal is sent to a pair of mixers $90° out of phase with one an-
other, and each is demodulated at frequency RF2 (equal to the
cavity FSR of 79.12 MHz). The overall phase of the heterodyne
detection is adjusted by phase shifting the RF signal driving the
EOM using cables of appropriate lengths. The demodulated signal
from each mixer is then sent to a lock-in amplifier referenced to
twice the frequency used to drive the plasma, and the in-phase
and quadrature outputs of each amplifier are digitized and stored
on a computer.
Ions are produced in a liquid-nitrogen-cooled multi-inlet multi-
outlet positive column discharge cell, which is placed between the
two free-standing mirrors of the optical cavity [18]. Intracavity
radiation is admitted into the cell by means of CaF2 windows
aligned at Brewster’s angle. The plasma is driven by a 40 kHz sine
wave voltage produced by amplifying the output of an arbitrary
waveform generator with an audio amplifier (Techron 7780) and
a step-up transformer. Hþ
3 was produced at a pressure of 200 mTorr
and a discharge current of 170 mA. The outputs of the two mixers
were each demodulated at 80 kHz with a lock-in amplifier set to a
10 ms time constant (16 Hz detection bandwidth).
Typical operation begins by filling the cooling jacket of the cell
with liquid nitrogen and igniting the plasma. The cavity length is
adjusted to bring it into resonance with the laser, and the laser-
cavity lock is established. The idler frequency is tuned by applying
a voltage to an internal PZT on the YDFL, and the cavity length is
2 K.N. Crabtree et al. / Chemical Physics Letters 551 (2012) 1–6

controlled with the locking electronics to maintain the resonance
condition. When the cavity PZT reaches the end of its travel, the la-
ser-cavity lock is electronically interrupted, the cavity length is re-
set to the other end of its travel, a new resonance is found, and the
lock is reestablished. In this manner, the spectrometer can scan
without manual intervention over the entire range of the YDFL
PZT (around 100 GHz), although in practice a scan is generally
much shorter. It is in principle possible to extend the automated
tuning range further by electronic control of the intracavity etalon
of the OPO and the nonlinear crystal position/temperature, but the
practical utility of such efforts would likely be minimal.
3. Results and analysis
3.1. Lineshapes
The overall Doppler lineshape for NICE-OHVMS in both absorp-
tion and dispersion has odd symmetry, and qualitatively appears
similar to the third derivative of a Gaussian absorption profile. A
detailed analysis of the lineshape is beyond the scope of this Letter
and will be the subject of a future work, but a qualitative descrip-
tion follows. Consider a general NICE-OHMS lineshape such as that
shown in panels a and b of Figure 2 in Ref. [8]. If the signal belongs
to an ion, then the AC voltage of the plasma causes velocity mod-
ulation (VM), Doppler shifting the lineshape at the plasma fre-
quency. As a result of the bidirectional nature of light in our
optical cavity, the lineshape is simultaneously Doppler shifted to
the red and to the blue by the same amount at each point in time
along the plasma voltage cycle. Consequently, the time-dependent
signal repeats itself every plasma half-cycle, or at twice the AC
plasma frequency (2fvm). In addition to VM, an ion also experiences
concentration modulation (CM) at 2fvm, and CM may be phase
shifted with respect to VM. Thus, in addition to a periodic Doppler
shift at 2fvm from VM, the lineshape amplitude varies at 2f vm from
CM. The net signal observed comes in 4 channels corresponding to
the even and odd second order Fourier coefficients of the absorp-
tion and dispersion profiles affected by VM and CM.
The sub-Doppler lineshape is more straightforward. As has been
discussed in regard to previous NICE-OHMS setups [19,20,9], the
carrier and sidebands can each act as pumps and probes for satura-
tion spectroscopy. The Lamb dips arising from the Bennet holes
burned in the population appear at half-integer multiples of the
carrier-sideband spacing fh, i.e., at all frequencies at which forward-
and reverse-propagating beams sample the same velocity compo-
nent of the Doppler distribution. At the line center m0, when the
Figure 1. Block diagram of NICE-OHVMS instrument. Details of its operation are given in the main text. YDFL: Ytterbium-doped fiber laser; EOM: electro-optic modulator;
OPO: optical parametric oscillator with pump (P, blue), signal (S, green) and idler (I, red) beams; PZT: piezoelectric transducer, PS: phase shifter; PSD: phase sensitive
detector; : signal splitter.
6
4
2
0
-2
Mixer1Signal(AU)
2726.32726.22726.12726.02725.9
Frequency (cm
-1
)
6
4
2
0
-2
Mixer2Signal(AU)
2726.32726.22726.12726.02725.9
Frequency (cm
-1
)
Figure 2. NICE-OHVMS spectrum of the R(1,0) and R(1,1)u
transitions of the m2
fundamental band of Hþ
3 . Each panel shows the in-phase (black, bottom) and
quadrature (red, top) outputs of a lock-in amplifier demodulating the indicated
mixer’s signal.
K.N. Crabtree et al. / Chemical Physics Letters 551 (2012) 1–6 3

zero-velocity distribution is both pumped and probed by the car-
rier, a signal only appears in dispersion as the heterodyne detection
scheme employed by NICE-OHMS is insensitive to absorption of the
carrier. In the current work, the heterodyne modulation index (b) is
small enough ($0.63) that only the carrier has sufficient power to
saturate transitions, while the sidebands can only act as probes.
The dispersion signal therefore contains Lamb dips at m0 and
m0 Æ fh=2, while the absorption signal contains them at m0 Æ fh=2.
While VM and CM have a strong influence on the lineshape of
the Doppler profile, they do not affect the sub-Doppler features
in the same way. Because Bennet holes are only burned in the pop-
ulation at or spaced evenly around the zero-velocity component of
the ion distribution, VM effectively changes the abundance of ions
with the appropriate velocity. In that way, VM effectively behaves
like CM, and the net effect is to influence the amplitude of the
Lamb dip lineshape and the relationship between the amplitudes
of the even and odd Fourier coefficients within absorption or dis-
persion. Neglecting any change of the ions’ collision rate with
changing velocity, VM and CM do not affect the sub-Doppler profile
beyond its amplitude. The sub-Doppler lineshape function [9] is
vðmdÞ ¼ A1 va md À
fh
2

À va md þ
fh
2

sin hh
þ À2A0vdðmdÞ þ A1 vd md À
fh
2

þ vd md þ
fh
2

cos hh; ð1Þ
where md is the frequency detuning from the transition center fre-
quency, hh is the heterodyne detection phase, A0 is the effective
amplitude of the central (carrier-carrier) dispersion Lamb dip, and
A1 is the effective amplitude of the carrier-sideband Lamb dips for
absorption and dispersion. vaðxÞ is a Lorentzian lineshape function
for absorption, and vdðxÞ is a lineshape function for dispersion re-
lated to vaðxÞ by the Kramers–Kronig relations. These are defined as
vaðxÞ ¼
1
1 þ c2ðx À x0Þ2
and vdðxÞ ¼
Àðx À x0Þc
1 þ c2ðx À x0Þ2
;
where x0 is the center and c is the inverse of the half-width at half-
maximum. When using this fit function, fh is held at the cavity FSR
(79.12 MHz), the amplitudes are constrained such that A0 A1, and
the Doppler profile near the line center is approximated by a third-
order polynomial with the quadratic term set to 0.
3.2. Hþ
3 Spectra
A sample spectrum of the R(1,0) and R(1,1)u
transitions of the m2
fundamental band of Hþ
3 is shown in Figure 2. The signals in the top
and bottom panels are the demodulated in-phase (black) and
quadrature (red, offset) components of the two mixers, which in
our setup we measure to be 96° out of phase with one another.
The in-phase components of the R(1,0) transition are shown in
greater detail in Figure 3. The overall lineshape is slightly asymmet-
ric; the blue side of the transition is stronger than the red side, par-
ticularly in mixer 2. The origin of this asymmetry is unknown, and
its impact on the spectroscopic accuracy will be discussed below.
A simultaneous fit of the R(1,0) sub-Doppler features in all four
detection channels to Eq. (1) is shown in Figure 4. A number of con-
straints were employed to ensure that the fit parameters were all
internally consistent. The line center frequency and Lamb dip width
were forced to be equal for all four data channels. Mixers 1 and 2
were held at 96° apart, and the sideband spacing was held equal
to the cavity FSR of 79.12 MHz. Because the in-phase and quadra-
ture components of each mixer sample different blends of CM and
VM, the Lamb dip amplitudes were allowed to be different for the
in-phase channels and the quadrature channels. However, the two
in-phase channels were forced to have equal Lamb-dip amplitudes,
and likewise for the two quadrature channels. After all of these
constraints, there are a total of 19 fit parameters for the entire data
set: the line center frequency m0, the Lamb dip full width at half
maximum (2=c), the overall detection phase hh, Lamb dip ampli-
tudes A0 and A1 for the in-phase channels and the quadrature
channels, and baseline terms of the form c0 þ c1x þ c3x3
to approx-
imate the central portion of the Doppler profile in each channel.
The overall quality of the fit is quite good; the residuals are com-
posed primarily of fringing evident on top of the spectrum (espe-
cially mixer 2) with the notable exception of the quadrature
channel of mixer 1, which has some small systematic deviations
on the Lamb dips. The line center m0 derived from the fit is
2725:89401954 Æ 0:0000023 cmÀ1
, but the absolute accuracy is
limited by the 2 Â 10À3
cmÀ1
accuracy of the wavemeter, so the
exact frequency should not be trusted. More important is the uncer-
tainty of the line center determination, which is $70 kHz; this rep-
resents the maximum potential accuracy of the technique provided
suitable frequency calibration is made (but see the discussion
3
2
1
0
-1
-2
-3
Signal(AU)
2725.962725.922725.882725.84
Frequency (cm
-1
)
Figure 3. NICE-OHVMS spectrum of the R(1,0) transition of the m2 fundamental
band of Hþ
3 . The black trace is the in-phase output of mixer 1, and the red is the in-
phase output of mixer 2.
-2
-1
0
1
2
Signal(AU)
-100 -50 0 50 100
Relative Frequency (MHz)
-0.2
0.0
0.2
M1Resid.
-0.2
0.0
0.2
M2Resid. Mixer 1 In-phase
Mixer 1 Quadrature
Mixer 2 In-phase
Mixer 2 Quadrature
Figure 4. A simultaneous fit of the central sub-Doppler region of the Hþ
3 R(1,0)
transition from all four data channels to Eq. (1). The symbols in the central portion
of the graph are the data, and the solid lines are the fit results. The residuals of the
fits are shown in the upper panel for mixer 1 and in the lower panel for mixer 2; in
each case the symbols correspond to those in the central panel.

below about the effects of the asymmetry). The overall detection
phase hh was found to be 132°, indicating a blend of absorption
and dispersion in each mixer. Because the Lamb dip widths (full
width at half maximum of $110 MHz as indicated by our fit) are
much broader than the Lamb dip spacing of 39.56 MHz (fh=2), the
individual Lamb dips for absorption and dispersion cannot be re-
solved at any RF detection phase. Rather than tuning the phase to
separate absorption from dispersion in the two mixers, a phase of
132° was chosen because it was found to isolate the fringing as
much as possible in a single detection channel (mixer 2 in-phase),
thus minimizing the fringing in the other 3 channels.
4. Discussion
The most interesting aspect of the NICE-OHVMS technique is
the presence of Lamb dips which enables high precision
measurements of line center frequencies. Prior to the development
of NICE-OHVMS, the only methods capable of routine sub-Doppler
molecular ion spectroscopy utilized fast ion beams which, due to
kinematic compression, give linewidths on the order of 10 to
120 MHz [21–23]. The linewidths from ion beams are comparable
to the Lamb dips presented here, though the current work with a
positive column has the advantage of much higher ion density than
that within an ion beam.
With the Hþ
3 Lamb dip linewidths of $110 MHz demonstrated
here, the precision of the line center determination is on the order
of 70 kHz. In its present implementation, the technique’s accuracy
is limited by the wavemeter, and also by slow drifts in the fre-
quency of the signal and idler beams caused by thermal fluctua-
tions of the OPO cavity. Use of an optical frequency comb to
stabilize and measure the frequencies of the pump and signal
beams would reduce the accuracy uncertainty to 100 kHz. Ulti-
mately, the total uncertainty of the technique will be determined
by the reproducibility of line center determinations once appropri-
ately calibrated.
The asymmetry observed in the overall NICE-OHVMS lineshape
can adversely affect the overall accuracy. As mentioned above, the
origin of this asymmetry is unknown, although it varies with het-
erodyne detection phase; similar effects were not observed in the
near-IR implementation of NICE-OHVMS [9]. Nevertheless, we
have performed simulations of the effects of the asymmetry by
synthesizing skewed profiles and comparing the results of our fit
function to the actual location of the Lamb dips. Based on the fit-
ting of our simulations, we estimate the magnitude of this line cen-
ter shifting to be less than a few MHz, even for Doppler profiles
that are much more asymmetric than those shown in this Letter.
Further study of this phenomenon will be possible with an optical
frequency comb, and such work is envisioned in the near future.
The width of the Lamb dips ($110 MHz FWHM from the fitting)
is fairly broad. We have varied the intracavity laser power and the
cell pressure, but any differences in the linewidth were not obser-
vable. However, the ranges of the power and pressure measure-
ments were limited: the intracavity power could only be changed
by a factor of 2 before the laser-cavity lock was adversely affected,
and the plasma could only give stable operation over 200–600
mTorr. Such wide Lamb dips were also observed in the NICE-
OHVMS experiment performed on Nþ
2 in the near-IR [9]; in that
study, the authors were able to observe a change in linewidth with
pressure, but extrapolating to zero pressure still gave a linewidth
of $30 MHz. Assuming that the linewidth is related to the time
an ion spends at zero velocity, it is perhaps unsurprising that a less
massive ion like Hþ
3 has a broader linewidth than Nþ
2 , as its velocity
may be more easily altered by weak long-range interactions.
The fringing apparent in the figures above limits the sensitivity
of the present measurements in 2 of the 4 detection channels. The
origin of the fringing is not fully understood; however, it appears
to have a definite phase with respect to both the heterodyne detec-
tion and the plasma modulation. When the plasma is turned off, the
fringing does not appear in a scan, and if the cavity transmission
detector is blocked while a signal originating from a single fringe
is present on a lock-in channel, the signal vanishes. Thus, it appears
that the fringing is the result of the plasma interacting with the la-
ser light rather than a purely electronic effect. One possibility is that
residual amplitude modulation (RAM) in the heterodyne sidebands
is being modulated by the plasma. RAM is an imbalance in the
amplitude and/or phase of the sidebands with respect to one an-
other. When demodulated, RAM appears as a DC offset in the het-
erodyne signal; because of our detection scheme using velocity
modulation and 2f detection, the NICE-OHVMS would ordinarily
be insensitive to such an offset. However, if the refractive index of
the plasma varies at 2f, the DC signal from RAM will be modulated
at 2f as well, resulting in a net NICE-OHVMS signal. Because RAM is
also affected by the presence of etalons in the optical system and
the optical frequency, a fringing pattern could possibly result as a
function of laser frequency. Testing whether this is truly the origin
of the fringing is difficult; however, it is probable that the fringing
would be reduced by employing a RAM compensation scheme via
temperature and voltage control of the fiber EOM [24].
The sensitivity of the technique at the experimental detection
bandwidth of 16 Hz, as determined from the noise-equivalent
absorption in the baseline of the in-phase component of mixer 1
(which has the least fringing of the four detection channels), is
3:4 Â 10À9
cmÀ1
. which is about two orders of magnitude above
the shot noise limit of 3:9 Â 10À11
cmÀ1
calculated from
amin ¼
p
2F
ffiffiffiffiffiffiffiffi
eB
gP0
s
1
J0ðbÞJ1ðbÞL
; ð2Þ
where F is the cavity finesse (120), e the fundamental electric
charge, B the detection bandwidth (16 Hz), g the detector respon-
sivity, P0 the power incident on the detector, JnðbÞ the nth order Bes-
sel function for modulation index b (0.63), and L the cavity length
(190 cm). While NICE-OHMS has been able to achieve a noise level
within a factor of 2 of the shot noise limit in one implementation
[7], the performance achieved by NICE-OHVMS relative to the shot
noise limit is already comparable to a number of other NICE-OHMS
setups (see the extensive discussion in Section 4 of [8]).
Ultimately, the absolute sensitivity can be improved by identi-
fying and eliminating noise sources and by increasing the cavity fi-
nesse. An increase in cavity finesse leads to additional technical
challenge in maintaining the laser-cavity lock, and may make the
system even more susceptible to the fringing effects that have al-
ready been observed. Such challenges can likely be overcome by
improving the bandwidth of the laser frequency corrections (cur-
rently limited to the 10 kHz bandwidth of the signal cavity PZT),
and correcting for RAM as discussed above.
5. Conclusions
In this Letter, we have demonstrated sub-Doppler spectroscopy
of molecular ions in the mid-infrared spectral region using the
NICE-OHVMS technique with a cw-OPO. By phase modulating the
seed laser with a fiber EOM prior to amplification and optical para-
metric oscillation, the mid-infrared idler is also phase modulated
without requiring a mid-IR EOM. The high optical power of the idler
beam allows use of high-bandwidth detectors, which in turn make
ultra-sensitive spectroscopy via NICE-OHMS possible. Velocity
modulation spectroscopy is then combined with NICE-OHMS to af-
ford ion-neutral discrimination, and the intracavity laser power is
sufficient for saturating fundamental rovibrational transitions as
demonstrated by spectroscopy of Hþ
3 . By fitting the sub-Doppler
spectral features, the center frequencies of individual rovibrational
lines can be measured with a precision of 70 kHz, and the maximum
K.N. Crabtree et al. / Chemical Physics Letters 551 (2012) 1–6 5

achieved sensitivity is within a factor of $90 of the shot noise limit.
Improvements to the technique, such as addition of an optical fre-
quency comb for accurate wavelength calibration, technical modiﬁ-
cations to improve its sensitivity, and expanding the frequency
coverage of the OPO from 3.2–3.9 lm to 2.8–4.8 lm, are envisioned.
The authors thank Takeshi Oka for providing us with the liquid
nitrogen cooled plasma cell and its associated pumps and plasma
electronics. KNC and BMS acknowledge support from a NASA Earth
and Space Science Fellowship. JNH acknowledges support from a
Springborn Fellowship and a National Science Foundation Gradu-
ate Research Fellowship (DGE 11-44245 FLLW). This work has been
supported by the National Science Foundation (PHY 08-55633), the
NASA Laboratory Astrophysics program (NNX08AN82G), and a Da-
vid and Lucile Packard Fellowship.
References
[1] C.S. Gudeman, M.H. Begemann, J. Pfaff, R.J. Saykally, Phys. Rev. Lett. 50 (1983)
727.
[2] C.S. Gudeman, M.H. Begemann, J. Pfaff, R.J. Saykally, J. Chem. Phys. 78 (1983)
5837.
[3] S.K. Stephenson, R.J. Saykally, Chem. Rev. 105 (2005) 3220.
[4] B.M. Siller, A.A. Mills, B.J. McCall, Opt. Lett. 35 (2010) 1266.
[5] A.A. Mills, B.M. Siller, B.J. McCall, Chem. Phys. Lett. 501 (2010) 1.
[6] L.C. Sinclair, K.C. Cossel, T. Coffey, J. Ye, E.A. Cornell, Phys. Rev. Lett. 107 (2011)
093002.
[7] J. Ye, L.-S. Ma, J.L. Hall, J. Opt. Soc. Am. B 15 (1998) 6.
[8] A. Foltynowicz, F. Schmidt, W. Ma, O. Axner, Appl. Phys. B 92 (2008) 313.
[9] B.M. Siller, M.W. Porambo, A.A. Mills, B.J. McCall, Opt. Exp. 19 (2011) 24822.
[10] M.S. Taubman, T.L. Myers, B.D. Cannon, R.M. Williams, Spectrochim. Acta 60
(2004) 3457.
[11] M.W. Porambo, B.M. Siller, J.M. Pearson, B.J. McCall, Opt. Lett., in press.
[12] W.D. Watson, Astrophys. J. 183 (1973) L17.
[13] E. Herbst, W. Klemperer, Astrophys. J. 185 (1973) 505.
[14] T. Oka, Phys. Rev. Lett. 45 (1980) 531.
[15] C.M. Lindsay, B.J. McCall, J. Mol. Spectrosc. 210 (2001) 60.
[16] C.P. Morong, J.L. Gottfried, T. Oka, J. Mol. Spectrosc. 255 (2009) 13.
[17] L. Velilla, B. Lepetit, A. Aguado, J.A. Beswick, M. Paniagua, J. Chem. Phys. 129
(2008) 084307.
[18] C. Lindsay, E.T. White, T. Oka, Chem. Phys. Lett. 328 (2000) 129.
[19] A. Foltynowicz, W. Ma, O. Axner, Opt. Exp. 16 (2008) 14689.
[20] O. Axner, W. Ma, A. Foltynowicz, J. Opt. Soc. Am. B 25 (2008) 1166.
[21] C.S. Gudeman, R.J. Saykally, Ann. Rev. Phys. Chem. 35 (1984) 387.
[22] J.V. Coe, J.C. Owrutsky, E.R. Keim, N.V. Agman, D.C. Hovde, R.J. Saykally, J.
Chem. Phys. 90 (1989) 3893.
[23] A.A. Mills, B.M. Siller, M.W. Porambo, M. Perera, H. Kreckel, B.J. McCall, J. Chem.
Phys. 135 (2011) 224201.
[24] I. Silander, P. Ehlers, J. Wang, O. Axner, J. Opt. Soc. Am. B 29 (2012) 916
.
Kyle Crabtree graduated with a B.S. in Chemistry from
Ball State University in 2006, and received his Ph.D. in
Chemistry from the University of Illinois in 2012. He is
currently a CfA Postdoctoral Fellow at the Harvard-
Smithsonian Center for Astrophysics.
James Hodges received a B.S. in Chemistry and a B.S. in
Polymer Fiber Chemistry in 2010 from Clemson Uni-
versity. He currently attends the University of Illinois
pursuing a PhD in Chemistry as a Springborn Fellow and
an NSF Graduate Research Fellow.
Brian Siller graduated with a B.A. in Chemistry, Com-
puter Science, and Mathematics from Ohio Wesleyan
University in 2007, and is currently pursuing a Ph.D. in
Physical Chemistry at the University of Illinois.
Adam Perry graduated with a B.S. in Chemistry from
Oregon State University in 2011, and is currently a
graduate student in Physical Chemistry at the Univer-
sity of Illinois.
Joe Kelly graduated with a B.S. in Chemistry from the
University of Illinois in 2012.
Paul Jenkins II is currently pursuing his B.S. in Chemical
Engineering at the University of Illinois, and plans to
graduate in 2015.
Ben McCall received his Ph.D. from the University of
Chicago in Chemistry and Astronomy Astrophysics.
Following a Miller Research Fellowship at the University
of California at Berkeley, he joined the faculty of the
University of Illinois, where he is currently an Associate
Professor of Chemistry and Astronomy.

Crabtree_2012

More Related Content

What's hot (20)

Viewers also liked (12)

Similar to Crabtree_2012 (20)

Crabtree_2012