We present a simple and easily implementable scheme for multiplexed Coherent Anti-Stokes Raman Scattering (CARS) spectroscopy and microscopy using a single femtosecond pulse, shaped with a narrow spectral notch. We show that a tunable spectral notch, shaped by a resonant photonic crystal slab, can serve as a narrowband, optimally time-delayed probe, resolving a broad vibrational spectrum with high spectral resolution in a single-shot measurement. Our single-source, single-beam scheme allows the simple transformation of any multiphoton microscope with adequate bandwidth into a nearly alignment-free CARS microscope.
©2010 Optical Society of America
The oscillation frequencies of molecular vibrations reflect the chemical structure and are widely used as a spectroscopic fingerprint for chemical detection and identification. One of the most efficient techniques to acquire the vibrational spectrum is Coherent Anti-Stokes Raman Scattering (CARS) spectroscopy . CARS is a third-order nonlinear process in which ‘pump’ and ‘Stokes’ photons at frequencies ωp and ωs, respectively, coherently excite molecular vibration at the frequency Ωvib = ωp-ωs. The excited vibrational level is subsequently probed by interaction with ‘probe’ photons at frequency ωpr, generating blue-shifted Anti-Stokes photons at a frequency ωAS = ωpr + Ωvib (Fig. 1a ). The vibrational spectrum is resolved by measuring the blue-shift of the Anti-Stokes photons from the probe frequency. Since CARS spectroscopy is a nonlinear technique, it benefits from high resolution sectioning capability together with high efficiency at low average laser powers. CARS has been demonstrated useful in a variety of applications including noninvasive biomedical imaging, combustion analysis and remote sensing [1-14].
CARS spectroscopy is typically performed as a multi-beam, multi-source technique [1–3], which is experimentally challenging to implement due to the strict requirement of spatial and temporal overlap of the excitation beams. This paradigm changed with the introduction of coherent control pulse-shaping techniques which employ a single femtosecond pulse for CARS spectroscopy and microscopy [4–7]. In these schemes, a single femtosecond pulse simultaneously provides the necessary pump, Stokes and probe photons. The main difficulty with implementing CARS using femtosecond pulses is that the pulse bandwidth is much broader than the width and spacing of the vibrational lines, thus limiting the spectroscopic resolution [4,8]. However, through careful spectral-phase or polarization shaping, a single vibrational level can be selectively excited , or narrowly probed [5,6], yielding spectroscopic resolution orders of magnitude better than the pulse bandwidth. Such single-pulse techniques are particularly attractive as only a single laser source is required and the spatiotemporal overlap of the pump, Stokes, and probe photons is inherently maintained. Moreover, because of their high peak intensity, femtosecond pulses are favorable for a variety of nonlinear processes such as second- and third-harmonic generation (SHG, THG) and multi-photon fluorescence, an advantage for label-free multimodal microscopy . Since first demonstrated , single-pulse CARS schemes have been utilized for vibrational imaging [4,10,11], time-resolved chemical micro-analysis , and remote detection of hazardous materials [13,14]. However, a major drawback of these techniques is the necessity of a programmable dynamic pulse shaper apparatus, typically a relatively complex and costly experimental setup which requires careful calibration . In addition, conventional pulse shapers suffer from limited refresh rates, spectral coverage and pulse repetition rates.
In this work, we demonstrate a new approach and an easily implementable scheme for performing single-pulse single-beam CARS spectroscopy, without the necessity of a complex pulse shaping apparatus. We achieve this by shaping a femtosecond pulse with a tunable narrowband spectral notch using a commercially available resonant photonic crystal slab (RPCS) [16,17] as a simple and robust pulse shaping element. We utilize the tunable notch in the RPCS transmission spectrum as a narrowband, temporally extended and time-delayed probe to perform multiplex single-shot CARS spectroscopy with spectral resolution two orders of magnitude better than the pulse bandwidth. The simplicity and robustness of this single-beam, shaper-free scheme, together with its wide spectral coverage, rapid modulation capability, and straightforward alignment, are attractive for practical spectroscopic applications.
The RPCS notch-shaped single-pulse CARS technique is analogous to conventional multiplex CARS [1,5,6] (Fig. 1b). In conventional multiplex CARS a band containing several vibrational levels is coherently excited by broadband pump and Stokes beams and subsequently probed by a narrowband probe beam, producing blue-shifted spectral peaks in the CARS spectrum (Fig. 1b). In the notch-shaped single-pulse CARS scheme, a narrowband probe is defined within the pulse spectrum by filtering out a narrow spectral band (Fig. 1c). The notch-shaped excitation pulse generates a CARS spectrum containing narrow spectral interference features which are blue-shifted from the original notch frequency by the molecular vibrational frequencies (Fig. 1c). The vibrational spectrum is easily resolved by measuring the blue-shift of these features from the shaped notch frequency. The spectroscopic resolution is therefore dictated by the notch spectral width rather than the full pulse bandwidth. The exact shape of the resonant spectral features is related to the spectral notch phase relative to the nonresonant background signal, as is further explained in section 5.2.
Although intuitively depicted in the frequency domain, the notch-shaped single-pulse CARS is also simply portrayed in the time-domain picture (Fig. 1c): The notch-shaped excitation pulse is a result of destructive interference of the unshaped femtosecond pulse and a temporally extended narrowband probe pulse at the notch wavelength, which is generated by the RPCS resonance. Since the RPCS resonance buildup and filtering is a causal process (not necessarily the case with pulse shapers ), the resulting probe pulse is time-delayed to follow the femtosecond excitation. The overall outcome is an impulsive femtosecond excitation of coherent molecular motion by the unshaped pulse, followed by time-delayed narrowband probing. The time delayed probing is effective as it does not probe at times t<0, i.e. before the vibrational coherence is excited [8,19]. The resulting vibrational features in the CARS spectrum are the coherent interference of the CARS field induced by the time-delayed probe and the nonresonant-dominated four-wave mixing field induced by the unshaped part of the excitation pulse. The optimal probe pulse duration and delay are set by an RPCS notch spectral width equal to the vibrational linewidth , producing a ‘matched-filter’ probe for the vibrational motion (Fig. 1c).
2. Experimental setup and results
The experimental setup for RPCS notch-shaped single-pulse CARS technique is illustrated in Fig. 2a . A wideband excitation pulse from a femtosecond oscillator (Fig. 2a-i) is spectrally notch filtered by transmitting through the RPCS. The shaped pulse is focused into the sample after the shorter wavelength part is blocked by a long-pass filter, as it spectrally overlaps the CARS signal. The CARS signal from the sample is then collected, short-pass filtered and measured by a spectrometer.
The RPCS shaping element used in this work is a commercially available grating waveguide structure, comprised of a thin waveguide layer with an etched sub-wavelength grating (Fig. 2b) . When a light beam illuminates the RPCS at a given angle and polarization most wavelengths are fully transmitted. However, a narrow spectral band is “on resonance” with the grating waveguide; namely, the light diffracted from the grating at these wavelengths is effectively coupled to a guided mode in the waveguide layer. The guided light is diffracted again by the grating and interferes destructively with the transmitted wave, resulting in full reflection of the incident wave. As a result, practically no light within the resonant band is transmitted through the RPCS, thus creating a notch in the transmission spectrum which is manifested as a delayed narrowband pulse in the time-domain picture. The resonant wavelength can be tuned over a wide spectral range, covering the entire source bandwidth, by changing the angle of the RPCS with respect to the incoming beam (Fig. 3a ) .
Figures 3-5 show representative experimental results. Figure 3a presents several RPCS notch-shaped excitation pulse spectra. Figure 3b shows raw measured CARS spectra from toluene for two slightly shifted notch spectral locations. Each spectrum contains several narrow spectral features which correspond to toluene vibrational lines. The vibrational spectrum can be easily resolved by taking the normalized difference between two such spectral measurements corresponding to slightly shifted notch locations (Fig. 3c), effectively compensating for any spectral artifacts of the system, excitation spectrum and the nonresonant background contribution (see materials and methods).
The origin of the blue-shifted features in the resonant CARS spectrum can be analyzed using the expression for the nonlinear polarization producing the CARS signal. For a singly resonant Raman transition, the vibrationally resonant nonlinear polarization spectrum, Pr(ω), driven by the pulse electric field, E(ω), can be approximated by [5,7]:
Since the spectral notch is narrow compared with the pulse bandwidth, it has a negligible effect on the total pulse energy and temporal shape, and consequently on the vibrational excitation spectrum A(Ω). However, the notch shaping has a significant effect on the resonant CARS spectrum given by Pr(ω) (Eq. (1). Due to the resonant term in the denominator in Eq. (1), the main contribution to Pr(ω) originates from E(ω-Ωvib). As a result, the resonant CARS spectrum has a shape that is similar to the excitation spectrum but blue-shifted by the vibrational resonant frequency. Therefore, the vibrationally resonant CARS spectrum contains narrow spectral dips at locations dictated by the vibrational frequencies and the notch location.
Experimentally resolved vibrational spectra from various samples in the 300cm−1-1000cm−1 spectral range with <20cm−1 spectral resolution are shown in Figs. 3 and 4 . The resolved spectra are in good agreement with the Raman spectra reported in the literature. We also verified that the raw measured CARS spectra are in good agreement with numerical simulations of the total CARS signal calculated by substituting in Eqs. (1) and 2 the excitation pulse spectrum and the Raman lines found in literature (Fig. 3b, see materials and methods).
Finally, we demonstrate single-pulse shaper-free vibrational imaging of compound samples, by scanning the samples using a piezo-controlled translation stage. The vibrational spectrum at each spatial location in the sample is resolved from two single-shot measurements with slightly different notch locations, and images of chemically specific contrast are generated (Fig. 5).
The implementation of single-pulse CARS using an RPCS yields an easy to implement, robust, and virtually alignment free experimental system, employing a single beam and a single source of light. Several other single source CARS techniques have been recently demonstrated, employing spectral focusing of two chirped femtosecond pulses [20–22] or double-pulse excitation [23,24]. Yet, all of these schemes rely on a multi-beam experimental apparatus. A recent technique introduced by Xu et al.  utilizes a single beam of noise-shaped pulses, but requires averaging over many realizations for obtaining adequate signal to noise.
An additional advantage of the RPCS as a shaping element is in rapid pulse shaping. Modulation at kilohertz rates are obtained by using a galvanometric mirror mounted RPCS, while megahertz rates have been demonstrated by electro-optical modulation of the waveguide refractive index . These high rates surpass the current refresh rate of the common LC-SLMs by one to four orders of magnitudes and hold a great potential for microscopy and remote-sensing applications, allowing for the use of lock-in detection and fast scanning. Another advantageous coherent-control feature is accessible by using multiple notch filters. By spectral positioning of several notches spaced by the vibrational lines of a known substance, tailored selective chemical detection and identification is possible. This creates a large coherent addition of the contributions from several vibrational levels, generating a coherent interference signal which is significantly larger than the linear sum of the separate contribution .
Both the accessible vibrational spectral range and spectroscopic resolution can be easily improved. A dramatic increase in the spectral range allowing for an upper limit of >4500cm−1 can be achieved by an increased excitation bandwidth . The lower spectral limit can be extended down to 100cm−1 using carefully tuned filters . Higher spectroscopic resolution is possible as the RPCS notch bandwidth can be designed to be as narrow as 0.1nm (~1.5cm−1) .
It is worth noting that a host of shaping techniques could be employed to generate a narrow probe spectral notch-like feature. Examples for such techniques include interference filters, Fabry-Perot interferometers, fiber Bragg gratings and even a thin wire in the spectral plane of a pulse compressor , a common apparatus in amplified femtosecond sources. Although these techniques do not possess all of the advantages and simplicity of the RPCS element, the resultant CARS spectrum will be of similar nature and will follow the same theoretical analysis presented in Eqs. (1-4) (see materials and methods).
In summary, we have demonstrated an easily implementable scheme for rapid single-beam coherent vibrational spectroscopy. The femtosecond CARS scheme is a promising technique for practical spectroscopic applications, particularly for the straightforward transformation of any multiphoton microscope with adequate bandwidth into a sensitive CARS microscope. The RPCS shaping scheme can be directly extended to the UV spectral range where conventional LC-SLMs are limited by absorption.
5. Materials and methods
5.1 Experimental setup
The RPCS is a commercial h124RE BioChip from Unaxis Balzers, Liechtenstein (optics.unaxis.com) . The h124RE consists of a glass substrate (1.1mm thick Schott AF45, refractive index, n = 1.52 at 800nm) with a uniformly etched sub-micron diffraction grating having a period of 360nm and depth of 40nm. A subsequent thin (150nm) film layer of Ta2O5 (n = 2.09 at 800nm) and a second identical etched grating. We illuminated the RPCS at classical incidence (i.e. the plane of incidence is perpendicular to the grating grooves) using TM (‘p’) polarization. The advantages of TM polarization are the narrower resonant spectral bandwidth (notch width) in comparison to the TE resonance, and a larger effective spectral tuning range (approximately 300-900nm) as the contrast of the notch is higher due to the better off-resonance transmittance for TM polarization around the Brewster angle. The laser source is a home-built prism-compensated Ti:Sapphire oscillator with a bandwidth of 60nm FWHM (940cm−1), central wavelength of 800nm, and 200mW average power at 80MHz repetition rate. All filters used are Omega optical (AELP779, AGSP770, 3RD720-760). Microscope objective is a 20X, 0.4 NA from Newport. The spectrometer is a Jobin Yvon Triax 320 equipped with liquid Nitrogen cooled CCD. The CARS signal is coupled to the spectrometer using a fiber-bundle and a 0.5NA condenser. For vibrational imaging samples were point scanned at 1μm steps with 100ms pixel dwell time. The pixel dwell time was limited by the readout time of the spectrometer CCD and software. Considerably shorter acquisition (<1ms) and exposure times (<100μs) are attainable by using frame transfer CCD architecture . Average laser power was ~30mW in all measurements except for the potato results of Fig. 5, where the power was lowered to <10mW. Figure 3 spectra were obtained using 120ms integration time. Figure 4(a-c) spectra were obtained with 500ms to 2s integration time, Fig. 5(c) spectra with 1s integration time, and Fig. 5(f) with 4s integration time.
5.2 Spectral analysis
The measured CARS signal is a coherent sum of the vibrationally resonant CARS spectrum Pr(ω) (Eq. (1), and a broad featureless nonresonant polarization, Pnr(ω). Using femtosecond excitation Pnr(ω) is typically orders of magnitude larger than Pr(ω), thus the measured spectral intensity is [5,10]:
Due to the relative spectral phase between the resonant and nonresonant signals, ϕ(ω), which originates from the vibrational resonance lineshape and the phase structure of the RPCS notch resonance, the interference dips in the measured CARS spectrum appear at the longer wavelength side of each vibrational feature and an interference peak manifest on its blue-side (Fig. 3b).
Previous works have demonstrated the ability of resolving the vibrational spectrum from a single-shot measurement of the CARS spectrum [19,23,29,30]. In a single-shot acquisition, a signal proportional to the resonant signal, Pr(ω), can be obtained from the measured CARS spectrum by subtracting and normalizing by the nonresonant background (Eq. (2):19,23]. In this work, we utilize the rapid tunability of the RPCS notch wavelength to implement a differential measurement scheme, eliminating spectral artifacts caused by the spectral transfer function of the system and the excitation spectrum. Utilizing the RPCS notch tunability we effectively compensate for irregularities in the CARS spectrum arising from the various filters’ spectral transmissions, spectrometer response and non-smooth excitation spectrum, by taking the difference between two spectral measurements corresponding to slightly shifted notch locations (shift by the notch width). After further normalization by the vibrational excitation amplitude A(ω), we obtain a signal clean from the experimental system spectral artifacts with peaks corresponding to the Raman lines strengths (Fig. 3c):
The resolved spectral line-shapes obtained with this simple differential method are similar to the line-shapes acquired in other differential techniques , and are not identical to the Raman line-shapes. Other, more involved spectral analysis techniques may be used to retrieve the true Raman line shapes from the measured CARS spectrum [10,29,30].
We thank Silvia Soria (CNR, Italy) for generously providing the RPCS filters used in this work, and the RPCS AFM image . This work was supported by grants from the Israel Science Foundation, the Israel Ministry of Science and FASTQUAST (Marie Curie Framework 7).
References and links
1. B. Schrader, Infrared and Raman Spectroscopy (VCH, Weinheim, 1995).
2. C. L. Evans and S. X. Xie, “Coherent Anti-Stokes Raman Scattering Microscopy: Chemical Imaging for Biology and Medicine,” Annu. Rev. Anal. Chem. 1(1), 883–909 (2008). [CrossRef]
3. A. Volkmer, “Vibrational imaging and microspectroscopies based on coherent anti-Stokes Raman scattering microscopy,” J. Phys. D Appl. Phys. 38(5), R59–81 (2005). [CrossRef]
5. D. Oron, N. Dudovich, and Y. Silberberg, “Single-pulse phase-contrast nonlinear Raman spectroscopy,” Phys. Rev. Lett. 89(27), 273001 (2002). [CrossRef]
6. D. Oron, N. Dudovich, and Y. Silberberg, “Femtosecond phase-and-polarization control for background-free coherent anti-Stokes Raman spectroscopy,” Phys. Rev. Lett. 90(21), 213902 (2003). [CrossRef] [PubMed]
7. N. Dudovich, D. Oron, and Y. Silberberg, “Single-pulse coherent anti-Stokes CARS spectroscopy in the fingerprint spectral region,” J. Chem. Phys. 118(20), 9208–9215 (2003). [CrossRef]
8. M. O. Scully, G. W. Kattawar, R. P. Lucht, T. Opatrny, H. Pilloff, A. Rebane, A. V. Sokolov, and M. S. Zubairy, “FAST CARS: engineering a laser spectroscopic technique for rapid identification of bacterial spores,” Proc. Natl. Acad. Sci. U.S.A. 99(17), 10994–11001 (2002). [CrossRef] [PubMed]
9. R. Carriles, D. N. Schafer, K. E. Sheetz, J. J. Field, R. Cisek, V. Barzda, A. W. Sylvester, and J. A. Squier, “Invited review article: Imaging techniques for harmonic and multiphoton absorption fluorescence microscopy,” Rev. Sci. Instrum. 80(8), 081101 (2009). [CrossRef] [PubMed]
10. B. C. Chen and S. H. Lim, “Optimal laser pulse shaping for interferometric multiplex coherent anti-stokes Raman scattering microscopy,” J. Phys. Chem. B 112(12), 3653–3661 (2008). [CrossRef] [PubMed]
11. K. Isobe, A. Suda, M. Tanaka, H. Hashimoto, F. Kannari, H. Kawano, H. Mizuno, A. Miyawaki, and K. Midorikawa, “Single-pulse coherent anti-Stokes Raman scattering microscopy employing an octave spanning pulse,” Opt. Express 17(14), 11259–11266 (2009). [CrossRef] [PubMed]
12. B. von Vacano and M. Motzkus, “Time-resolving molecular vibration for microanalytics: single laser beam nonlinear Raman spectroscopy in simulation and experiment,” Phys. Chem. Chem. Phys. 10(5), 681–691 (2008). [CrossRef] [PubMed]
13. O. Katz, A. Natan, Y. Silberberg, and S. Rosenwaks, “Standoff detection of trace amounts of solids by nonlinear Raman spectroscopy using shaped femtosecond pulses,” Appl. Phys. Lett. 92(17), 171116 (2008). [CrossRef]
14. H. Li, D. A. Harris, B. Xu, P. J. Wrzesinski, V. V. Lozovoy, and M. Dantus, “Standoff and arms-length detection of chemicals with single-beam coherent anti-Stokes Raman scattering,” Appl. Opt. 48(4), B17–B22 (2009). [CrossRef] [PubMed]
15. A. M. Weiner, “Femtosecond pulse shaping using spatial light modulators,” Rev. Sci. Instrum. 71(5), 1929 (2000). [CrossRef]
16. D. Rosenblatt, A. Sharon, and A. A. Friesem, “Resonant Grating Waveguide Structures,” IEEE J. Quantum Electron. 33(11), 2038 (1997). [CrossRef]
17. A. Thayil, A. Muriano, J. P. Salvador, R. Galve, M. P. Marco, D. Zalvidea, P. Loza-Alvarez, T. Katchalski, E. Grinvald, A. A. Friesem, and S. Soria, “Nonlinear immunofluorescent assay for androgenic hormones based on resonant structures,” Opt. Express 16(17), 13315–13322 (2008). [CrossRef] [PubMed]
19. D. Pestov, R. K. Murawski, G. O. Ariunbold, X. Wang, M. Zhi, A. V. Sokolov, V. A. Sautenkov, Y. V. Rostovtsev, A. Dogariu, Y. Huang, and M. O. Scully, “Optimizing the laser-pulse configuration for coherent Raman spectroscopy,” Science 316(5822), 265–268 (2007). [CrossRef] [PubMed]
20. W. Langbein, I. Rocha-Mendoza, and P. Borri, “Single source coherent anti-Stokes Raman microspectroscopy using spectral focusing,” Appl. Phys. Lett. 95(8), 081109 (2009). [CrossRef]
21. T. Hellerer, A. M. Enejder, and A. Zumbusch, “Spectral focusing: High spectral resolution spectroscopy with broad-band laser pulses,” Appl. Phys. Lett. 85, 25 (2004). [CrossRef]
22. A. F. Pegoraro, A. Ridsdale, D. J. Moffatt, Y. Jia, J. P. Pezacki, and A. Stolow, “Optimally chirped multimodal CARS microscopy based on a single Ti:sapphire oscillator,” Opt. Express 17(4), 2984–2996 (2009). [CrossRef] [PubMed]
23. B. von Vacano, L. Meyer, and M. Motzkus, “Rapid polymer blend imaging with quantitative broadband multiplex CARS microscopy,” J. Raman Spectrosc. 38(7), 916–926 (2007). [CrossRef]
25. X. G. Xu, S. O. Konorov, J. W. Hepburn, and V. Milner, “Noise autocorrelation spectroscopy with coherent Raman scattering,” Nat. Phys. 4(2), 125–129 (2008). [CrossRef]
26. D. Oron, N. Dudovich, and Y. Silberberg, “All-optical processing in coherent nonlinear spectroscopy,” Phys. Rev. A 70(2), 023415 (2004). [CrossRef]
27. B. von Vacano, W. Wohlleben, and M. Motzkus, “Single-beam CARS spectroscopy applied to low-wavenumber vibrational modes,” J. Raman Spectrosc. 37(1-3), 404–410 (2006). [CrossRef]
28. A. Natan, O. Katz, S. Rosenwaks, and Y. Silberberg, “Single-pulse standoff nonlinear Raman spectroscopy using shaped femtosecond pulses,” in Proceedings of Ultrafast Phenomena XVI, (Springer Series in Chemical Physics, Vol. 92, 2009), pp. 985–987.
30. Y. Liu, Y. J. Lee, and M. T. Cicerone, “Fast extraction of resonant vibrational response from CARS spectra with arbitrary nonresonant background,” J. Raman Spectrosc. 40(7), 726–731 (2009). [CrossRef]