Doppler-free resolution near-infrared spectroscopy at 1.28~μm with the noise-immune cavity-enhanced optical heterodyne molecular spectroscopy method
RResearch Article Optics Letters 1
Doppler-free resolution near-infrared spectroscopy at1.28 µ m with the noise-immune cavity-enhancedoptical heterodyne molecular spectroscopy method T ZU -L ING C HEN AND Y I -W EI L IU Department of Physics, National Tsing Hua University, Hsinchu 30013, Taiwan * Corresponding author: [email protected] May 16, 2017
We report on the Doppler-free saturation spectroscopy of the nitrous oxide (N O) overtone transition at1.28 µ m. This measurement is performed by the noise-immune cavity-enhanced optical heterodyne molec-ular spectroscopy (NICE-OHMS) technique based on the quantum-dot (QD) laser. A high intra-cavitypower, up to 10 W, reaches the saturation limit of the overtone line using an optical cavity with a highfinesse of 113,500. At a pressure of several mTorr, the saturation dip is observed with a full width at half-maximum of about 2 MHz and a signal-to-noise ratio of 71. To the best of our knowledge, this is the firstsaturation spectroscopy of molecular overtone transitions in 1.3 µ m region. The QD laser is then lockedto this dispersion signal with a stability of 15 kHz at 1 sec integration time. We demonstrate the potentialof the N O as markers because of its particularly rich spectrum at the vicinity of 1.28-1.30 µ m where liesseveral important forbidden transitions of atomic parity violation measurements and the 1.3 µ m O-bandof optical communication. © 2017 Optical Society of America OCIS codes: (140.3490) Lasers, distributed feedback; (060.2420) Fibers, polarization-maintaining;(060.3735) Fiber Bragg gratings. http://dx.doi.org/10.1364/ao.XX.XXXXXX
High-resolution spectroscopy plays an essential role inwidespread applications in sciences including astronomy,physics, chemistry and biology. The Doppler-free spectrumprovide a significant improvement in spectroscopic resolutionrelative to Doppler-broadened line shapes. Few results were re-ported in near-infrared (NIR) region by means of saturation spec-troscopy, because a relatively high power is required. However,NIR is an important spectral range not only offer laboratory com-parisons for the research on astrophysical phenomena [1, 2], butalso to provide a powerful analytical tool to the studies of molec-ular rotation-vibration structure [3]. While most atmospheremolecules such as CO , CO, CH , C H , H O, HF, and N Oall present strong fundamental vibrational transitions beyond2 µ m, the overtone transitions falling in the 1-2 µ m range areone to two orders of magnitude less intense [3–5]. They are oftenwith integrated line strengths ( < − cm − /molecule cm − ).Although it can be observed using highly sensitive techniques,such as cavity-ring down spectroscopy, most of the results areDoppler-limited. So far, the Doppler-free saturation spectrum inthe 1.5 µ m region has been successfully observed using cavity- based spectroscopy [6, 7], but only the Doppler-limited spectrumhas been reported in the 1.1 µ m-1.3 µ m band [8–10], which iswith even smaller dipole moments .On the other hand, considerable progress in QD lasers engi-neering development results in the broad availability of efficientInGaAs QD diode applied for NIR laser sources. Particularly inthe 1.1-1.3 µ m band, QD diode based lasers have demonstratedtheir superior performances in several respects in comparisonwith traditional quantum-well based diode lasers [11]. Many im-portant applications in this spectral range require stabilized lasersources, such as the coarse wavelength division multiplexing(CWDM) technology at 1.3 µ m [12], the parity non-conservation(PNC) measurements with forbidden transitions in atomic thal-lium at 1283 nm [13, 14], ytterbium at 1280 nm [15], lead at1279 nm [16], and iodine at 1315 nm [17]. There is consequentlyimportant to find a general molecular reference lines for thefrequency stabilization of such light sources [5, 18]. The lackof strong frequency references, however, remains a challengein the NIR spectral region. N O could become potential wave-length calibration references in this band. Based on the HITRAN a r X i v : . [ phy s i c s . a t m - c l u s ] M a y esearch Article Optics Letters 2 database, its particular rich spectrum in the 1.2 µ m-1.3 µ m canbe found in [19]. However, all N O transitions in this regionare the overtone transitions with the line strengths typically at10 − cm − /molecule cm − and the dipole moments in the levelof 0.2 mDebye, which is smaller than most of the molecules re-ported in the previous Doppler-free NICE-OHMS experiments[7, 20, 21]. A higher optical power is required for the Doppler-free saturation spectroscopy. I /I π ( ν FSR ) ∆ t e -20 0 20 40 60 80-4-202 C a v i t y t r an s m i ss i on ( V o l t ) Time ( µ s) CRD time~15 µ s C a v i t y R e f l e c t i on ( V o l t ) BB4-1/4"-20 (cid:13539)(cid:16897)
A A V REV
DATE DESCRIOTION APPD
ABDCE
Q'ty AB SIZE : A 3
SUS304 ZHN001A0411100 C (cid:15116)(cid:16482)(cid:14855)(cid:11118)(cid:13615)(cid:7838) , (cid:26374)(cid:19771)(cid:13104)(cid:15026)(cid:18729)(cid:15114)(cid:22703)(cid:22693)(cid:20911)(cid:12057)(cid:8984) C O P Y I N G I N T H I S D O C U M E N T , G I V I N G I T T O O T H E R S AN D T H E U S E O R C O MM U N I C A T I O N O F T H E C O N T E N T S T H E R O F AR E F O R B I DD E N W I T H O U T E X P R E SS A U T H O R I T Y . ROUGHNESS SYMBOL
R DE
HEAT TREATMENT (HRc)
UNLESS OTHERWISE STATED
CF6" (cid:16929)(cid:22693)(cid:20732)(cid:27260)
UNIT: mm (cid:26275)(cid:22923)(cid:12915)(cid:8433) (cid:24789)(cid:13858)(cid:7785) (cid:24789)(cid:13858)(cid:7785)
TOLERANCE UNLESS OTHERWISE STATED
NO.
LENGTH
SCALE:
SURFACE TREATMENT (cid:73)(cid:74)(cid:72)(cid:73)(cid:77)(cid:74)(cid:72)(cid:73)(cid:85)(cid:33)(cid:85)(cid:70)(cid:68)(cid:73)(cid:33)(cid:68)(cid:80)(cid:83)(cid:81)(cid:47)
TITLEPARTSNO.MODEL E SHARP CORNERSRANGE DIAMETER
MATERIALASSEMBLYNAMEDRAWING R C (cid:8702)(cid:26378)(cid:9918) A-A . ° ° . I D = * t (cid:8702)(cid:26378)(cid:9918) B-B
PZT1PZT2 (a)
Freq-1283704000 (Hz)
T i m e ( s e c ) v a c u u m : 1 0 - 7
T o r r (b)
Fig. 1. (a) Left: Cavity mechanics. Right: Typical ring downcurves of transmission (PD3) and reflection of the cavity (PD1).The finesse of the cavity ( F =113,500) was derived from the lin-ear fitting of ring down time measurements. (b) The frequencydrift of the cavity FSR within 400 sec.Since the optical power required for saturation on overtonetransitions is at the level of several watts, an optical cavity is anadequate method to enhanced the optical power. NICE-OHMStechnique, the state-of-the-art spectroscopy, combines the cavityenhancement and the frequency modulation (FM) technique toreduce frequency noises [22]. Also, by matching the modula-tion frequency to the cavity free spectral range (FSR), an addi-tional advantage of the immunity to the residual laser-to-cavityfrequency noise results in the realization of close to shot-noise-limit sensitivity [23]. In this letter, we report a NICE-OHMSinterrogating a narrow overtone transition of N O, where theDoppler-free signal around 2 MHz was obtained by the satu-ration absorption spectroscopy performed in an optical cavitywith a finesse over 10 . To our best knowledge, this is the firsttime that the Doppler-free resonances of molecular overtone transitions had been observed in the 1.3 µ m band. Furthermore,the frequency of the laser system was locked to the saturated ab-sorption resonances. We demonstrate N O overtone transitionscan be candidate of frequency markers in the 1.2 µ m-1.3 µ mregion of the NIR.The NICE-OHMS experimental setup includes a high finessecavity and a quantum-dot external cavity diode laser (QD-ECDL)operated in a Littrow configuration. The details of the QD-ECDLcan be referred in our previous work [24]. The left figure ofFig. 1(a) depicts the cavity design, which is composed of a flatand a R=1 m curved ultra-low loss mirrors. The average spotsize in the cavity is calculated to be 748.0 µ m, yielding a transit-time broadening 330 kHz at room temperature. The mirrors areattached to two PZTs (Noliac NAC2125-A01 and PiezomechanikHpst150/20-15/12), then mounted on Glue-In Mirror Mounts(Thorlabs, POLARIS-K1C4). Instead of using ULE material inour previous work, which is well-known for its ultra low ther-mal expansion coefficient ( ∼ × − ), we use the fused-quartzsubstrate as the cavity spacer. Although the thermal expansioncoefficient of fused-quartz ( ∼ × − ) is higher than which ofULE, it can has advantages of being relatively low-cost and eas-ily accessible. The entire cavity setup is placed on a sorbothanesheet for passively vibration isolation. The empty-cavity finesseis measured to be 113,500 by a series of ring-down curves withvarious cavity-sweeping rates [25], corresponding to a round-trip cavity loss of 56 ppm. One of ring-down signal is shown inthe right of Fig. 1(a), where the upper and lower traces are thecavity transmission and the reflection signals, respectively.The simplified experimental setup for NICE-OHMS de-tection is shown in Fig. 2. An electro-optical modulator(EOM, EOSPACE, PM-0K5-10-PFU-PFU-130) simultaneouslygenerates two pairs of sidebands and is coupled by a sin-gle mode polarization-maintained fiber for efficient coupling(spatial matching) into the high finesse cavity. The lower fre-quency modulation at ν PDH =20.8 MHz with a modulation index β ∼ ν FSR =1283.7 MHz generated by a synthesizer (HP, 8648B) is witha modulation index β ∼ ν FSR - ν PDH =1262.9 MHz)enables locking of the FM-modulation frequency to the FSRof the cavity by implementing the DeVoe–Brewer method [26].This signal is feedback to the external modulation port of the HPsynthesizer through a servo loop, to match the FM-modulationfrequency to the FSR of the cavity during the cavity scanning.Both of the PDH and FSR error signals are acquired from de-modulating the reflection light at PD1.In order to lock the laser on the high finesse cavity with a10 kHz narrow linewidth, the whole QD-ECDL system and thefiber EOM are placed in a protection case on a granite slab toreduce environmental noises to improve passive laser frequencystability [24]. Then the active locking is achieved by the PDHerror signal fed through the locking servo loop with a servobandwidth wider than 650 kHz. The loop is composed of threestages: an integrator for slow PZT (DC-1 kHz), a proportional-integral (PI) low-pass filter for ac current feedback ( <
100 kHz),and a direct current feed forward from the error signal. Theresultant laser linewidth evaluated from the Allan deviation is ∼
130 Hz at 1 ms integration time. The optical output powerfrom the fiber-EOM is ∼ >
10 W, which is derived from the cavity finesse and the cavity esearch Article Optics Letters 3 transmission power ( ∼ λ /4LP PD2Amp PD3 CRD timemeasurement flip
High finesse cavity containing N O phase ν FSR B NICE-OHMSsignal
DBM
Pol Fiber EOM PBS ν PDH
OI PD1
Pol λ /2PDH servo lock to high finesse cavity (FM) EXT DC ν FSR ∿ PZT current fast current ∿ B Vibrational isolation board
DBM
DeVoe-Brewer servo lock to cavity FSRcoupler microwave counter tunable BP ∿ CEAS (DC)(AC)
Fig. 2.
Simplified layout of the experimental setup for NICE-OHMS. OI, optical isolator; Pol, Glan-Taylor calcite polarizer;PBS, polarizing beam splitter; EOM, electro-optical phase mod-ulator; LP, low-pass filter; BP, band-pass filter; DBM, doubledbalanced mixer; PZT, piezoelectric actuator; PD, photodetec-tor.The FSR locking error signal generated by DeVoe-Brewertechnique is used to feedback control the synthesizer to matchthe FM side frequency ( ∼ ν FSR . The resulting NICE-OHMS signal is alsooptimized for dispersion signal using a phase shifter (Hittite,131046-HMC934LP5E) to adjust the reference phase.The frequency of the modulation sideband ν FSR is measuredusing a microwave counter (EIP, Model 575B) with a resolu-tion of 1 Hz and referenced to the external 10 MHz frequencyreference from atomic clock. The FSR of the empty-cavity is mea-sured to be ∼ ∼ ∼
400 Hzwithin 400 secs, corresponding to a fractional instability of cav-ity length of 3 × − . The averaged fractional instability within1 sec is better than 5 × − . It should be noticed that the mea-sured fluctuation of FSR frequency could also be resulted fromother sources such as FSR lock performance, residual amplitudemodulation (RAM), and refractive index fluctuation. Our exper-iment shows that the new cavity design made of fused quartzis reliable enough for the laser stabilization and NICE-OHMSdetection.The NICE-OHMS dispersion profile is obtained from thesignal demodulated from the cavity transmission on the detectorwith a bandwidth of 1 GHz. Shown in Fig. 3, the observedtransition performed at 1.283 µ m is R e (17), which is just 9 GHz - 3 0 0 - 2 0 0 - 1 0 0 0 1 0 0 2 0 0 3 0 0- 4- 3- 2- 10123 NICE-OHMS dispersion signal (a.u.)
R e l a t i v e f r e q u e n c y ( M H z )
D o p p l e r b r o a d e n i n g s i g n a l N O R e ( 1 7 ) o v e r t o n e t r a n s i t i o n - 4 0 - 2 0 0 2 0 4 0 - 0 . 50 . 00 . 5
R e la t iv e f r e q u e n c y ( M H z )
Cavity transmission (Volt) (a) - 6 - 4 - 2 0 2 4 6
D o p p l e r - f r e e s i g n a lR e l a t i v e f r e q u e n c y ( M H z )
NICE-OHMS dispersion signal (a.u.)
R e s i d u a l ( x 1 0 ) (b)
Fig. 3.
Observed NICE-OHMS dispersion signal of the R e (17)absorption line for intra-cavity power of 10.7 W, which yieldsa degree of saturation of 0.4 for the carrier. (a) Upper part:cavity transmission signal detected by a 125 kHz PD and Gaus-sian fit. Lower part: Doppler-broadened dispersion signal at23 mTorr intra-cavity pressure with a detection band-widthof 30 Hz. Inset shows the saturation dip signal at variouspressure conditions. (b) NICE-OHMS signal demodulatedat 5.8 mTorr intra-cavity pressure with a detection band-widthof 1 Hz, where Doppler-free features and theoretical derivativeLorentzian fit are shown with fit residuals.away from one of the thallium PNC hyperfine transitions. Thecavity enhanced absorption spectroscopy (CEAS) and NICE-OHMS are both shown in Fig. 3(a). The Doppler-broadenedcavity transmission is with a peak absorption at the pressureof 23 mTorr, and a FWHM of 200 MHz that is comparable tothe 215 MHz theoretical prediction at the temperature of 25 ◦ C.The integrated absorption area of the Doppler broadened CEASprofile is measured to be 0.002 cm − . The absorption rate givenby HITRAN is 3.1 × − cm − . Therefore, the number of passesin the cavity is equivalent to 63000, which is in a good agreementwith the CRD time measurement of the cavity finesses.The saturation intensity at 20 mTorr is 6.4 × W/m , basedon the 0.175 mDebye transition dipole moment. In our experi-ment, the 10.7 W intra-cavity power reaches a degree of satura-tion of 0.4 for the carrier. In the NICE-OHMS, the Doppler-freesaturation dip is clearly observed with a 5% contrast to the back-ground Doppler broadened signal. Shown in the inset of Fig. 3(a)is the dip under varies pressure from 1.4 to 12 mTorr. Despitethe Doppler-broadened profile becomes difficult to distinguishfrom the background at a relatively low pressure of 1.4 mTorr,the Doppler-free signal remains apparent.The particular of the saturation dip is depicted in Fig. 3(b),together with the line fitting to a Lorentzian dispersion func-tion (red curve). The observed width of the Doppler-free fea-ture is about 2 MHz (FWHM) that is slightly larger than theexpected 0.5 MHz saturated homogeneous width includingthe calculation from the self-broadening coefficient under asaturation parameter of 0.4, the 330 kHz transit-time broad-ening and the 60 kHz pressure broadening at 23 mTorr. Thecorresponding fitting residual is shown in the lower part ofFig. 3(b) (gray line). The signal-to-noise ratio is 71 at a detec- esearch Article Optics Letters 4 tion band-width of 1 Hz, corresponding to a detection sensi-tivity of ∼ × − cm − Hz − . It is still two order largerthan the shot-noise-limited sensitivity. In our spectrometer,the shot-noise-limited fractional absorption ( α L) min , is givenby 1.2 × − . With the cavity length of 11.7 cm, it corre-sponds to a noise-equivalent bandwidth-reduced sensitivity of1.0 × − cm − Hz − . Currently the optimum bandwidth ofthe LP after the DBM (shown in Fig. 2) for the best signal to noiseratio is 1 ∼
10 Hz. Potential signal-to-noise ratio limitations couldbe from the limited bandwidth of the PD2 (only 1 GHz), lockingperformance of laser and FSR stabilization, and RAM. Whilethe noise dominated in the Doppler-broadened NICE-OHMSis mainly from RAM, the Doppler-free NICE-OHMS suffers arandom drift noises. Both of the noises were observed in theempty chamber without gas input. The further improvementcan be made by the use of a WM dither [22] or the active controlof RAM [28].The Doppler-free dispersion signal is then used as an errorsignal to actively feedback control the length of the high-finessecavity, then to stabilize the laser. The slope of the error signalis measured to be 0.2 V/MHz. While the cavity and the laserare locked to the center of the Doppler-free line, the error signalis presented in Fig. 4, where the feedback loop bandwidth is300 Hz. The laser frequency stability is evaluated using theAllan variance. The laser stability can down to 15 kHz at 1 secintegration time, corresponding to a stability of 6.4 × − . After τ > τ − ) limit and the laser stability is mainly limited by the slowdrift of the NICE-OHMS signal baseline. A D E V A l l o m e t r i c 1 F i t o f A D E V ( t - 1 / 2 ) Allan variance (Volt) t ( s e c ) - 0 . 2- 0 . 10 . 00 . 10 . 2 T i m e ( s e c )
NICE-OHMS (Volt)
Fig. 4.
Upper panel: The error signal of Doppler-free disper-sion NICE-OHMS signal while the cavity and the laser arelocked. Lower panel: The square root of the Allan varianceof the error signal. The red solid line shows the τ − whitenoise limit.In summary, we have presented the NICE-OHMS withDoppler-free resolution in a molecular overtone transition ofN O. It was achieved by an optical cavity with finesse of 135,000.The sensitivity of 4.1 × − cm − at 1 Hz bandwidth is reached.The laser frequency stabilization on the Doppler-free resonanceshas been also demonstrated with a stability of 15 kHz. Our resultsuggest that N O is a promising candidate for future frequencyreferences at the wavelength of 1.2-1.3 µ m. FUNDING INFORMATION
National Science Council of Taiwan (103-2112- M-007-007-MY3).
ACKNOWLEDGMENT
We would like to thank Professor Jow-Tsong Shy for the usefuldiscussions. We also thank Professor Li-Bang Wang for use ofhis instruments.
REFERENCES
1. T. Oka, Faraday discussions , 9 (2011).2. Nicholls, C. P., Lebzelter, T., Smette, A., Wolff, B., Hartman, H., Kaufl,H.-U., Przybilla, N., Ramsay, S., Uttenthaler, S., Wahlgren, G. M.,Bagnulo, S., Hussain, G. A. J., Nieva, M.-F., Seemann, U., and Seifahrt,A., Astronomy and Astrophysics , A79 (2017).3. M. Reichenbächer and J. Popp,
Challenges in molecular structuredetermination. (Chapter 2, Vibrational Spectroscopy) (Springer Science& Business Media, 2012).4. M. de Labachelerie, K. Nakagawa, and M. Ohtsu, Opt. Lett. , 840(1994).5. H. S. Moon, Appl. Opt. , 1097 (2008).6. A. A. Madej, A. J. Alcock, A. Czajkowski, J. E. Bernard, and S. Chep-urov, J. Opt. Soc. Am. B , 2200 (2006).7. S. Saraf, P. Berceau, A. Stochino, R. Byer, and J. Lipa, Opt. Lett. ,2189 (2016).8. Y. He, M. Hippler, and M. Quack, Chemical physics letters , 527(1998).9. S. Tashkun, V. Perevalov, E. Karlovets, S. Kassi, and A. Campargue,Journal of Quantitative Spectroscopy and Radiative Transfer , 62(2016).10. T. Asakawa, N. Kanno, and K. Tonokura, Sensors , 4686 (2010).11. S. Li, Q. Gong, C. Cao, X. Wang, J. Yan, Y. Wang, and H. Wang, Opticaland Quantum Electronics , 623 (2014).12. M. J. Adams and I. Henning, Optical fibres and sources for communi-cations (Springer Science & Business Media, 2013).13. P. A. Vetter, D. M. Meekhof, P. K. Majumder, S. K. Lamoreaux, andE. N. Fortson, Phys. Rev. Lett. , 2658 (1995).14. N. H. Edwards, S. J. Phipp, P. E. G. Baird, and S. Nakayama, Phys.Rev. Lett. , 2654 (1995).15. D. F. Kimball, Phys. Rev. A , 052113 (2001).16. D. M. Meekhof, P. Vetter, P. K. Majumder, S. K. Lamoreaux, and E. N.Fortson, Phys. Rev. Lett. , 3442 (1993).17. G. E. Katsoprinakis, L. Bougas, T. P. Rakitzis, V. A. Dzuba, and V. V.Flambaum, Phys. Rev. A , 040101 (2013).18. T. Dennis, E. A. Curtis, C. W. Oates, L. Hollberg, and S. L. Gilbert, J.Lightwave Technol. , 776 (2002).19. R. A. Toth, Journal of Molecular Spectroscopy , 158 (1999).20. A. Foltynowicz, W. Ma, and O. Axner, Opt. Express , 14689 (2008).21. H. Dinesan, E. Fasci, A. Castrillo, and L. Gianfrani, Opt. Lett. , 2198(2014).22. A. Foltynowicz, F. Schmidt, W. Ma, and O. Axner, Applied Physics B , 313 (2008).23. J. Ye, L.-S. Ma, and J. L. Hall, J. Opt. Soc. Am. B , 6 (1998).24. T.-L. Chen and Y.-W. Liu, Opt. Lett. , 4352 (2015).25. J. Poirson, F. Bretenaker, M. Vallet, and A. Le Floch, JOSA B , 2811(1997).26. R. G. DeVoe and R. G. Brewer, Phys. Rev. A , 2827 (1984).27. M. Aketagawa, T. Yashiki, S. Kimura, and T. Q. Banh, InternationalJournal of Precision Engineering and Manufacturing , 851 (2010).28. Y. Zhang, S. Qiao, L. Sun, Q. W. Shi, W. Huang, L. Li, and Z. Yang,Opt. Express22