Chirped lasers dispersion spectroscopy
implemented with single- and dual-sideband
electro-optical modulators
Michal Nikodem,1,2,3 Genevieve Plant,1 Zhenxing Wang,1 Paul Prucnal,1
and Gerard Wysocki1,*
1
Electrical Engineering Department, Princeton University, Princeton, 08540 NJ, USA
2
Current address: Wroclaw Research Centre EIT + , 54-066 Wrocław, Poland
3
michal.nikodem@eitplus.pl
*
gwysocki@princeton.edu
Abstract: We report new approaches for signal generation in Chirped Laser
Dispersion Spectroscopy (CLaDS). Two optical arrangements based on
electro-optical modulators significantly reduce CLaDS system complexity
and enable optimum performance when applied to detection of GHz-wide
molecular transitions. Proof-of-principle experiments in the near-infrared
spectral range are presented and potential strategies for application in the
mid-infrared are discussed.
©2013 Optical Society of America
OCIS codes: (300.6260) Spectroscopy, diode lasers; (300.6390) Spectroscopy, molecular;
(260.2030) Dispersion.
References and links
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
A. A. Kosterev and F. K. Tittel, “Chemical sensors based on quantum cascade lasers,” IEEE J. Quantum
Electron. 38(6), 582–591 (2002).
M. R. McCurdy, Y. Bakhirkin, G. Wysocki, R. Lewicki, and F. K. Tittel, “Recent advances of laserspectroscopy-based techniques for applications in breath analysis,” J Breath Res 1(1), 014001 (2007).
E. Kerstel and L. Gianfrani, “Advances in laser-based isotope ratio measurements: selected applications,” Appl.
Phys. B 92(3), 439–449 (2008).
R. Lewicki, J. H. Doty 3rd, R. F. Curl, F. K. Tittel, and G. Wysocki, “Ultrasensitive detection of nitric oxide at
5.33 microm by using external cavity quantum cascade laser-based Faraday rotation spectroscopy,” Proc. Natl.
Acad. Sci. U.S.A. 106(31), 12587–12592 (2009).
G. B. Rieker, J. B. Jeffries, and R. K. Hanson, “Calibration-free wavelength-modulation spectroscopy for
measurements of gas temperature and concentration in harsh environments,” Appl. Opt. 48(29), 5546–5560
(2009).
R. F. Curl, F. Capasso, C. Gmachl, A. A. Kosterev, B. McManus, R. Lewicki, M. Pusharsky, G. Wysocki, and F.
K. Tittel, “Quantum cascade lasers in chemical physics,” Chem. Phys. Lett. 487(1-3), 1–18 (2010).
S. Lundqvist, P. Kluczynski, R. Weih, M. von Edlinger, L. Nähle, M. Fischer, A. Bauer, S. Höfling, and J.
Koeth, “Sensing of formaldehyde using a distributed feedback interband cascade laser emitting around 3493
nm,” Appl. Opt. 51(25), 6009–6013 (2012).
B. Brumfield, W. Sun, Y. Ju, and G. Wysocki, “Direct In Situ Quantification of HO2 from a Flow Reactor,” J.
Phys. Chem. Lett. 4(6), 872–876 (2013).
G. Wysocki and D. Weidmann, “Molecular dispersion spectroscopy for chemical sensing using chirped midinfrared quantum cascade laser,” Opt. Express 18(25), 26123–26140 (2010).
M. Nikodem and G. Wysocki, “Chirped Laser Dispersion Spectroscopy for Remote Open-Path Trace-Gas
Sensing,” Sensors (Basel) 12(12), 16466–16481 (2012).
M. Nikodem, D. Weidmann, C. Smith, and G. Wysocki, “Signal-to-noise ratio in chirped laser dispersion
spectroscopy,” Opt. Express 20(1), 644–653 (2012).
M. Nikodem and G. Wysocki, “Molecular dispersion spectroscopy--new capabilities in laser chemical sensing,”
Ann. N. Y. Acad. Sci. 1260(1), 101–111 (2012).
G. C. Bjorklund, “Frequency-modulation spectroscopy: a new method for measuring weak absorptions and
dispersions,” Opt. Lett. 5(1), 15–17 (1980).
G. C. Bjorklund, M. D. Levenson, W. Lenth, and C. Ortiz, “Frequency modulation (FM) spectroscopy,” Appl.
Phys. B 32(3), 145–152 (1983).
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14649
15. B. Hraimel, X. Zhang, Y. Pei, K. Wu, T. Liu, T. Xu, and Q. Nie, “Optical Single-Sideband Modulation With
Tunable Optical Carrier to Sideband Ratio in Radio Over Fiber Systems,” J. Lightwave Technol. 29(5), 775–781
(2011).
16. W. C. Swann and S. L. Gilbert, “Line centers, pressure shift, and pressure broadening of 1530-1560 nm
hydrogen cyanide wavelength calibration lines,” J. Opt. Soc. Am. B 22(8), 1749–1756 (2005).
17. D. Richter, A. Fried, and P. Weibring, “Difference frequency generation laser based spectrometers,” Laser
Photonics Rev. 3(4), 343–354 (2009).
18. M. Nikodem, K. Krzempek, G. Plant, K. Abramski, G. Wysocki, “Methane sensing at 3.4um using Chirped
Laser Dispersion Spectroscopy with DFG source,” in CLEO/Europe-IQEC 2013 Conference Digest, OSA
Technical Digest (CD) (Optical Society of America, 2013), paper CH-1.3.
19. A. Hangauer, G. Spinner, M. Nikodem, and G. Wysocki, “Chirped Laser Dispersion Spectroscopy with Directly
Modulated Quantum Cascade Laser,” in CLEO: Science and Innovations 2013, OSA Technical Digest (online)
(Optical Society of America, 2013), paper CW1L.5.
1. Introduction
Laser spectroscopy is a powerful tool for sensitive and selective trace gas detection in a large
number of applications including environmental monitoring, medical diagnostic, industrial
process control, and many others [1–8]. Majority of spectroscopic techniques currently in use
rely on measurement of light that is absorbed by the target analyte. However, detection of the
absorbed light, a measurement that is relatively simple to perform, has some shortcomings
that limit its use especially in more challenging applications. These include measurements of
strongly absorbing samples (absorption larger than 10-20%) that require correction of the
nonlinear (exponential) dependence of the Beer-Lamberts law, or measurements with large
baseline (transmitted power) fluctuations that require reliable normalization of the power
received by the photodetector to assure accuracy of the concentration retrieval process [5].
These limitations can be mitigated by applying a chirped laser dispersion spectroscopy
(CLaDS) [9]. In CLaDS, the molecular concentration is retrieved from the measurement of
optical dispersion related to the refractive index changes in the vicinity of the target
absorption line (information about optical absorption is also available, but will not be
discussed in the context of the work presented here). Immunity of the CLaDS signal to
intensity variations and/or amplitude noise, as well as its linear response within the full range
of accessible concentration levels (from the minimum detection limit up to above 99% of
absorption) make CLaDS an interesting alternative to absorption-based techniques. A
prototype remote sensing instrument based on CLaDS has already been tested in the field
[10], where the unique capabilities of this technology have been successfully verified.
However, due to some technological limitations the performance of the open-path CLaDS
systems presented to-date was far from optimal. In this paper we present new optical
arrangements that simplify CLaDS optical setup, and enable optimal performance of CLaDS
spectrometers targeting molecules at ambient conditions that typically exhibit transition
linewidths of a few GHz. These new approaches are tested in the near-IR with hydrogen
cyanide (HCN) used as a test analyte. Applications of these new CLaDS configurations in the
mid-IR, a spectral region particularly suitable for sensitive molecular spectroscopy, are
briefly discussed.
2. New optical configurations for CLaDS
The signal in previously reported CLaDS systems [9–12] was produced through interaction of
two co-propagating electromagnetic waves with the gas sample. The frequencies of the two
waves are separated by Ω and after transmission through the sample both waves are focused
onto a square-law photodetector, which allows for detection of a heterodyne beatnote
produced by the wave interference. The two waves can be conveniently produced using a
single laser source and an optical frequency shifter. A molecular dispersion in the sample
results in different propagation velocities for the two waves which affects phase/frequency of
the heterodyne beatnote. By applying a fast chirp to the laser frequency, these changes in
heterodyne beatnote frequency are significantly enhanced and the molecular dispersion
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14650
spectrum can be retrieved through straightforward frequency demodulation performed at the
carrier frequency Ω (details on CLaDS signal generation and its comparison with other
techniques, including FM-spectroscopy [13, 14], can be found in [9] and [12] respectively).
In order to maximize CLaDS signal amplitude the frequency shift Ω should equal the
linewidth of the target absorption line, which is typically around 3 GHz for gases at
atmospheric conditions. In the first CLaDS experiments the two frequency-shifted waves
were created using an acousto-optical modulator (AOM) that provided frequency-shifts of 50
to 100 MHz, causing instrument performance to be far from optimal for measurements at
atmospheric pressure [10]. Commercially available AOM frequency shifters are generally
limited to frequency shifts of up to 1GHz achievable only at shorter wavelengths (i.e. in the
near-IR), and their performance (especially diffraction efficiency) deteriorates at high
modulation frequencies. Moreover, the two beams (the fundamental and the frequency-shifted
one) emerge from the AOM at different angles, thus an additional optical setup is needed to
combine them into a single dual-color beam that is used for CLaDS sensing. This increases
complexity of the instrument and can reduce its opto-mechanical stability.
In this paper we demonstrate a new approach to perform CLaDS. The method uses
telecom-based electro-optical modulators (EOMs) and mitigates the technical limitations of
AOM-based CLaDS. Two experimental arrangements have been developed and successfully
tested with CLaDS: 1) a single side-band (SSB) modulation based on a dual-parallel MachZehnder modulator (DPMZM), as well as 2) a conventional dual side-band (DSB) intensity
modulation performed with a standard Mach-Zehnder modulator (MZM). The optical
configurations of both approaches are shown in Fig. 1(a) and Fig. 1(b), respectively.
Fig. 1. a) CLaDS system with DPMZM for dual-color beam (SSB spectrum) generation. 3dB
hybrid coupler is used to drive modulator with two orthogonal signals; b) CLaDS system with
MZM for triple-color beam (DSB spectrum) generation (LD – laser diode, SG – signal
generator, PC – polarization controller, PD – photodetector, OS – optical spectrum).
CLaDS with SSB modulator
DPMZM consists of two Mach-Zehnder modulators embedded into parallel branches of the
third Mach-Zehnder structure (shown in Fig. 1(a)). SSB optical spectrum (carrier +
frequency-shifted side-band) is created when modulators are driven with orthogonal signals
and when appropriate bias levels are applied [15]. DPMZM generating the SSB spectrum can
simply replace the AOM in CLaDS setup, while providing an enhanced performance because:
1) the dual-color beam is created directly at the output of the DPMZM, which reduces the
complexity of the optical setup, and 2) EOMs can deliver frequency-shifts up to tens of
gigahertz, which provides optimal conditions for the CLaDS signal generation with broad
absorption lines (e.g. collisionally broadened lines at elevated pressures).
Figure 2(b) shows an example optical spectrum of the light emerging from the output of
the DPMZM used in this proof-of-concept experiment. The optical spectrum is a quasi SSB
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14651
spectrum with two distinct frequencies (a carrier and a side-band with ~17dB or 50 ×
suppression of other sidebands). Due to limitations in the available RF power from the signal
generator, side-band suppression better than 17dB below the carrier and/or the primary
sideband could not be achieved (typically, suppression of >25 dB is desired). Nevertheless
this performance was sufficient to reliably generate CLaDS signals and to perform initial
studies. CLaDS spectra were acquired by chirping the laser frequency across the P15
transition at 1553.755 nm in the 2ν3 ro-vibrational band of HCN (see [16], Swann et al. for
details) with the rate of approximately S = 250 MHz/μs. This was accomplished by applying
10 kHz triangular modulation to the injection current. A dual-color beam generated by the
DPMZM was directed through a 5.5-cm-long gas cell containing pure HCN (transmission of
the gas cell in the vicinity of the target transition is shown in Fig. 2(a). The CLaDS spectrum
acquired at Ω = 1.45 GHz is presented in Fig. 2(c). Its shape agrees well with the numerical
model (Eq. (1) and Eq. (16) from [9] and HCN line parameters from [16], Swann et al. were
used in this simulation). A dependence of the CLaDS signal amplitude on Ω was also
measured and it is shown in Fig. 3(b). For the conditions used in this experiment (pressure in
the cell of 25 Torr and room temperature) the spectral linewidth of the transition is
approximately 1.5 GHz. The optimum frequency shift Ω = ~1.46 GHz has been estimated
from the measured data, which agrees very well with the numerical model (also shown in Fig.
3(b). It should be noted that despite the reduced pressure in the cell the Doppler linewidth of
the target transition is >1 GHz, which is reasonably close to pressure broadened linewidths
observed at atmospheric conditions (typically 3 to 6 GHz). Thus, conditions used in the
present work serve as an acceptable simulator for the spectral lines typically observed at
ambient pressures (different lineshape, Gaussian vs. Lorentzian, has no fundamental impact
on this proof-of-principle experiment).
0.95
0.90
CLaDS signal / Chirp rate
0.85
c)
b)
1.00
Optical power
(dBm)
Transmission
a)
-8 -4 0 4 8
Frequency (GHz)
2x10
-11
1x10
-11
-1x10
-11
-2x10
-11
Ω =1.5GHz
-20
>17dB
-40
-60
-10
-5
0
5
10
Frequency (GHz)
measurement
simulation
0
-10
-5
0
5
Frequency (GHz)
10
Fig. 2. a) Normalized transmission spectrum of the target HCN transition and b) a quasi SSB
spectrum at the output of DPMZM when driven at 1.5 GHz, both measured with highresolution spectrum analyzer; c) CLaDS spectrum of the P15 transition of HCN recorded using
SSB modulation with spacing of Ω = 1.45 GHz. Frequency axes show detuning of the laser
frequency from the center of the P15 transition at 1553.755 nm (~193.86 THz). Spectrum in c)
is normalized by the chirp rate to facilitate comparison with DSB-based signals recorded at
different chirp rate.
CLaDS with DSB modulator
The second configuration studied in this work is based on a standard Mach-Zehnder
modulator that can produce pure intensity modulation of the optical carrier. Therefore, an
optical spectrum that consists of a carrier and two side-bands with equal amplitudes is
generated. An approximation of the generated CLaDS signal can be performed with the
following assumptions: a pure intensity modulation with no phase difference between
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14652
sidebands, and negligible absorption effects introduced by the gas sample. With these
assumptions the retrieved CLaDS dispersion spectrum (after FM-demodulation at Ω)
generated with a triple-color beam can be approximately analyzed as a superposition of two
CLaDS signals generated by the two pairs of frequency components: 1) the carrier and the +
1st side-band and 2) the carrier and the −1st side-band. In this idealized case, the
superposition of the two co-existent heterodyne beatnote signals at frequency Ω that have
equal amplitudes but different phases (differences are related to molecular dispersion
experienced by all DSB frequency components) can be interpreted as a signal at Ω with its
phase being an average of the individual beatnote signals phases. As a result, while the laser
frequency is chirped the instantaneous frequency determined by the FM detector will be an
average of the instantaneous frequencies expected for the two individual beatnotes. Hence,
using Eq. (16) in [9], we can express the CLaDS signal measured with a DSB modulator as:
f DSB (ω ) = Ω + ω ⋅

1 SL  dn
dn 
1 SL  dn
dn
⋅
−
⋅
−
 + (ω − Ω) ⋅
 (1)
2 c  d ω ω +Ω d ω ω 
2 c  d ω ω d ω ω −Ω 
where ω is the optical frequency, S is the chirp rate, n is the sample’s refractive index, and L
is the sample/light interaction length. Assuming ω >> Ω we obtain:
f DSB (ω ) = Ω +

1 SLω  dn
dn
⋅
−

2 c  d ω ω +Ω d ω ω −Ω 
(2)
Equation (2) indicates that CLaDS signal recorded using DSB modulation at Ω has the same
shape with one half of the magnitude of the signal measured using SSB modulation [9] at 2 ×
Ω. To confirm this property a series of measurements were performed at different Ω and for S
= 400 MHz/μs. Figure 3(a) shows sample CLaDS spectra recorded using triple-color beam
generated by MZM at three different Ω. A very good agreement between model (Eq. (2)) and
experimental data is observed. The dependence of the DSB CLaDS signal amplitude on
modulation frequency Ω varied between 100 MHz and 2.3 GHz is compared with similar plot
for SSB in Fig. 3(b). As expected from Eq. (2) for DSB the maximum amplitude is obtained
for Ω ≈0.73 GHz which is half of the 1.46 GHz obtained for SSB spectrum.
CLaDS signal / Chirp rate
-11
1x10
0
-11
-1x10
750MHz
1600MHz
measurement
simulation
-11
-2x10
b)
100MHz
CLaDS signal / Chirp rate
a)
-15
-10
-5
0
5
10
Frequency - 193.86e12 (GHz)
15
1.5x10
-11
1.0x10
-11
5.0x10
-12
max@1.46GHz
DSB
simulation
SSB
simulation
max@0.73GHz
0.0
0.0
0.5
1.0
1.5
2.0
2.5
Modulation frequency (GHz)
Fig. 3. a) Experimental CLaDS spectra recorded using setup shown in Fig. 1(b) for three
different modulation frequencies (spectra are shifted vertically and horizontally for viewing
purposes only) shown together with simulations calculated using Eq. (2). The changes of the
CLaDS signal shape at different Ω is in good agreement with the prediction shown in Fig. 7 of
Ref [9]; b) The amplitude of the CLaDS signal as a function of modulation frequencies Ω
obtained with SSB modulator shown in Fig. 1(a) and DSB modulator in Fig. 1(b). Excellent
agreement with the model is observed.
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14653
3. Discussion
Both configurations presented in this paper are capable of CLaDS signal detection.
Application of DPMZM allows for straightforward generation of a dual-color beam (SSB
spectrum). The molecular dispersion is encoded in the frequency of the heterodyne beatnote
between these two frequency-shifted waves. Thus this configuration can be considered as an
equivalent version of the previously demonstrated prototype that used SSB generation with an
AOM [9–12]. The EOM based configuration allows for achieving much larger frequency
spacing between the two frequency components (up to 40 GHz using standard telecom
devices) which allows for optimal CLaDS performance even for significantly broader
transitions (>1 GHz). Moreover, DPMZM generates a dual-color beam directly at its fiber
coupled output, which significantly reduces the complexity of the optical system and allows
for improved stability and robustness needed in field applications. The second approach
generates CLaDS signals using DSB modulation based on a simple MZM. In this technique
molecular dispersion is encoded into instantaneous frequencies of not two, but three waves
transmitted through the sample. The most important benefit of using DSB spectrum for
CLaDS is that modulation frequency and photodetector bandwidth that guarantee optimum
performance are only half of what is needed when SSB modulator is applied. The amplitude
of the DSB-CLaDS signal is half of the SSB-CLaDS signal amplitude, but given the
simplicity of DSB spectrum generation that requires only one MZM instead of the complex
DPMZM, such a signal reduction represents a reasonable compromise.
The CLaDS systems demonstrated in this paper are based on standard telecom-based
modulators for which specified operating wavelength range is typically from 1530 to 1580
nm. Nonetheless, these instruments can be used also at longer wavelengths e.g. we have
tested operation of the presented configurations at ~1650 nm and no significant reduction in
performance was observed. This shows clear potential for methane or carbon dioxide
detection around 1650nm. However, since the strongest ro-vibrational transitions can be
targeted in the mid-IR spectral region (3-12 µm), it would be desired to perform EOM-based
CLaDS in this spectral range. Unfortunately, due to lack of commercially available mid-IR
EOM technology, implementations of the developed techniques at longer wavelengths
represents a true technological challenge. Therefore an immediate approach that could be
implemented in the mid-IR is through application of a nonlinear frequency-conversion
process such as differential frequency generation (DFG) [17]. Such implementation would
allow the optical SSB or DSB spectrum to be generated in the near-IR followed by a
frequency-conversion to the mid IR using DFG. This would allow application of the near-IR
modulator and waveguide technology, while targeting the strongest molecular transitions in
the mid-IR [18]. We are also investigating possibilities of generation of a DSB spectrum
through direct modulation of mid-IR laser sources. Among potentially promising approaches
that are being currently studied is a fast (>500 MHz) direct modulation of an injection current
in quantum cascade lasers that would result in further reduction of the complexity of CLaDS
system [19].
4. Conclusions
Two new CLaDS signal generation techniques were demonstrated. As compared to the
previous work, the new approaches enable optimal CLaDS performance for pressurebroadened transitions (linewidth >1 GHz), and allow for significant simplification of the
optical setup. The systems under study use telecom-based components and their applications
are limited to the near-IR spectral range, in which several chemical species can be targeted
(e.g. acetylene, carbon dioxide, water vapor, or methane). The implementation of this
technology can be extended to the mid-IR and potential strategies have been briefly
discussed.
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14654
Acknowledgements
The authors would like to acknowledge financial support by the NSF CAREER award
CMMI-0954897, by the NSF SECO award EEC-1128282 and partial support by the MIRTHE
NSF Engineering Research Center. Andreas Hangauer is acknowledged for reading the
manuscript and providing valuable comments.
#189314 - $15.00 USD
(C) 2013 OSA
Received 23 Apr 2013; revised 3 Jun 2013; accepted 5 Jun 2013; published 13 Jun 2013
17 June 2013 | Vol. 21, No. 12 | DOI:10.1364/OE.21.014649 | OPTICS EXPRESS 14655