Open Access
Add to BibSonomy
Abstract
We report an all-fiber free-running bidirectional dual-comb laser system for coherent anti-Stokes Raman scattering spectroscopy based on spectral focusing. The mode-locked oscillator is a bidirectional ring-cavity erbium fiber laser running at a repetition rate of $\sim{114}\;{\rm MHz}$. One output of the bidirectional laser is wavelength-shifted from 1560 to 1060 nm via supercontinuum generation for use as the pump source. We have been able to record the Raman spectra of various samples such as polystyrene, olive oil, polymethyl methacrylate (PMMA), and polyethylene in the C–H stretching window. We believe that this all-fiber laser design has promising potential for coherent Raman spectroscopy and also label-free imaging for a variety of practical applications.
© 2020 Optical Society of America
Coherent anti-Stokes Raman scattering (CARS) spectroscopy is a powerful technique that detects resonant vibrational properties of chemical compounds. This technique can be used to perform chemically selective measurements and imaging to distinguish species that have different Raman spectra [1]. The advantage of CARS is that it is a label-free imaging technique, which can be applied to species that do not fluoresce [2,3]. In Raman spectroscopy, the “C–H window” is one of the interesting spectral regions that covers from $\sim{2800}$ to ${3100}\;{{\rm cm}^{- 1}}$. In this region, the stretching motion of hydrogen bonds has unique features on the Raman spectrum. Complex laser systems have been developed to perform CARS spectroscopy in the C–H window [4]. Thus, a simple and compact fiber laser source working in this region are useful for studying cells, lipids, and some proteins that have hydrocarbon chains [3–6].
Since CARS is a resonant four-wave mixing process [7], two light fields (pump and Stokes) with frequency detuning matching the Raman resonance are usually required. Traditionally, narrowband and tunable lasers are used for CARS spectroscopy [4,8]. In order to scan multiple Raman resonances, a mechanical scanning stage is often implemented for sweeping the wavelength of the laser, which can be complex, bulky, and sensitive to the surrounding environment.
Supercontinuum (SC) generation [9] laser sources, a mature technology for obtaining a broad spectrum, have been used for broadband multiplexed CARS spectroscopy without the need for tunable lasers [10–12]. Selm et al. demonstrated a broadband Raman source covering from 0 up to 4000 cm−1using an Er-fiber-laser-driven SC source [12]. The experiment can be done motionlessly if the optimal delay of the probe pulse is known. However, in order to excite the molecules impulsively [13,14], the pulses were compressed externally with a compressor before they were delivered to the sample, which reduces robustness and compactness of the fiber laser system. Moreover, in multiplexed CARS spectroscopy, to acquire the broadband Raman signal, a spectrometer or monochromator has to be integrated into the system, which further increases the complexity and cost of the setup.
As an alternative approach, dual-comb-based Fourier transform (FT) CARS spectroscopy has been proposed [15–17]. Dual-comb CARS uses two femtosecond lasers: one acts as the pump, and the other acts as the probe. The probe is probing the vibrational coherence from the molecules excited impulsively by the pump. In the time domain, the modulated oscillation, due to Raman gain induced frequency shift, can be detected with a photodiode. Since this approach is a FT-based spectroscopic technique, the scaled Raman spectrum can be obtained via Fourier transforming the detected time signal. Thus, dual-comb CARS has the advantage of using a single photodiode for detection. In addition, because the dual-comb laser has two pulse trains with slightly different repetition frequencies, the process of detecting the time-resolved signal can be entirely motionless. Nevertheless, there are a few drawbacks for this approach. The experiment usually requires two femtosecond laser combs that are electronically locked to each other; the system is complicated and sensitive to the environment. In addition, because of using impulsive Raman scattering, many free-space components are used for compressing the pump pulses to close to transform-limited. For the bandwidth, dual-comb CARS measurements have been done mostly with Ti:sapphire lasers [15,16] or Yb-fiber laser combs [17], which only covered the fingerprint window (${600 - 1800}\;{{\rm cm}^{- 1}}$). In order to cover the C–H window, the laser’s spectrum had to be nonlinearly broadened with a photonic crystal fiber (PCF), and then the pulses have to be compressed by chirp mirrors [18].
In this Letter, we present an all-fiber, single-cavity, free-running dual-comb laser system for CARS spectroscopy based on spectral focusing (SF) technique [19,20]. The principle of dual-comb based SF CARS is shown in Fig. 1. Instead of using very short pulses as in FT dual-comb CARS [15–18], SF uses chirped pulses as the pump and Stokes, so a free-space pulse compressor is not required.
For SF-based CARS, because the pump field and the Stokes field are linearly chirped with an identical chirp-rate (Fig. 1), a narrowband instantaneous frequency difference (IFD) is created when the pump and Stokes are overlapped spatially and temporally. For a dual-comb laser, the pump field and Stokes field are asynchronous with repetition frequencies $ f_r $ and ${f_r} + \Delta f$, and the relative time delay between two fields changes at $\Delta \tau = \frac{{\Delta f}}{{{f_r^*}({{f_r} + \Delta f})}}$ for every $\frac{1}{f_r}$ [21], which changes the corresponding IFD of the two matched linearly chirped pulses as well. Therefore, multiple Raman states can be excited at the varying relative time delays [22], and a full scan is completed every $\frac{1}{f_r}$. Our design is similar to the design reported in Ref. [22]. There, Chen et al. have demonstrated a SF-based dual-comb CARS working in the fingerprint region with two Yb-fiber laser combs, and a free-space coupled PCF for Stokes generation. In contrast, our laser system was constructed based on a single-cavity bidirectional mode-locked fiber laser enabled by fiber-taper carbon nanotube saturable absorber technology [23–25]. This design has been used in a free-running dual-comb spectroscopy setup for the measuring absorption spectrum of hydrogen cyanide [25]. Our design of the cavity is based on erbium fiber technology and operates at $\sim{1.56}\;\unicode{x00B5}{\rm m}$. Here, one of the laser outputs is frequency shifted to $\sim{1}\;\unicode{x00B5}{\rm m}$ via SC generation using a short piece of a highly nonlinear fiber (HNLF). The HNLF was fabricated by OFS, and has a conventional circular core-cladding profile, which can be directly spliced to another single mode fiber without using free-space coupling tools [26]. The 1 µm arm is used as the pump field, while the 1.56 µm output is used as the Stokes field. The optical frequency difference between the pump and Stokes can excite the C–H stretching modes. In order to cover the entire C–H window ($\sim{2800}$ to ${3100}\;{{\rm cm}^{- 1}}$), the laser pulses are spectrally broadened and acquire a linear chirp via parabolic generation in normal-dispersion gain fiber [27–30]. This parabolic generation technique has been used in another SF-based stimulated Raman scattering microscopy for generating broadband linearly chirped pulses [31]. In addition, compared to other dual-comb-based CARS setups [15–18,22], our laser design can be built in an all-fiber format owing to the use of spliceable HNLF for pump generation and an all-fiber amplifier for parabolic linear chirped pulse generation.
Fig. 2. All-fiber dual-comb spectral focusing CARS system. (a) Free-running bidirectional fiber laser. WDM, wavelength division multiplexer; SA, saturable absorber; PC, polarization controller; EDF, Er-doped fiber; OC, output coupler; ISO, isolator; EDFA, Er-doped fiber amplifier; PD, photodiode; SMF, single mode fiber; HNLF, highly nonlinear fiber; YDFA, Yb-doped fiber amplifier; FC, fiber collimator; AFC, adjustable fiber collimator; HWP, half-wave plate; DM, dichroic mirror; PL, polarizer; M, mirror. (b) CARS spectroscopic and multiphoton microscopic system. SM, scanning mirrors; PM, parabolic mirror; OBJ1, OBJ2, microscope objectives; F, filters; PMT1, PMT2, photomultiplier tube; MPM, multiphoton image; OSC, oscilloscope.
The schematic diagram of the free-running dual-comb laser system is presented in Fig. 2. The essential part of the dual-comb system is the bidirectional mode-locked ring laser cavity in the left side of Fig. 2(a). The cavity consists of a fiber-taper saturable absorber (based on carbon nanotubes) that permits bidirectional femtosecond pulse generation [25]. A polarization controller (PC) is used to adjust the birefringence of the fiber to optimize the operation of the laser. The repetition rate of the laser is approximately 113.8 MHz. The counter-pump (CCW) direction has output average power $\sim{1.25}\;{\rm mW}$; the co-pump (CW) direction has output average power $\sim{0.75}\;{\rm mW}$. The output spectra of the bidirectional laser are plotted in Fig. 3(a). The difference of repetition frequencies ($\Delta f$) between the two directions can be tuned from $\sim{12}$ to 41 Hz by adjusting the PC in the cavity. At the laser output, a ${2} \times 2$ 90/10 fiber coupler is used to couple out 10% power from each direction. After the output coupler, another 90/10 fiber coupler is used in each direction. The 10% ports from the couplers are spliced to a 50/50 fiber coupler to generate an interferogram signal to trigger the time-resolved SF-based dual-comb CARS signal. In Fig. 3(b), an interferogram corresponding to $\Delta f$ of $\sim{17}\;{\rm Hz}$ is plotted. A zoom-in graph of the interferogram is shown in Fig. 3(c). The 90% port of the CW arm is spliced to a lab-built erbium-doped fiber amplifier (EDFA). The EDFA is made with $\sim{7}\;{\rm m}$ (Coractive EDF-L900) normal-dispersion low gain fiber for parabolic pulse amplification [28,29]. Approximately $\sim{3}\;{\rm m}$ MetroCor fiber is spliced to the EDFA for further propagation and matching the chirp-rate of the pump. An FC/APC fiber connector has been made directly at the end of the MetroCor fiber. After the connector, $\sim{120}\;{\rm mW}$ of average output power is obtained. The output is collimated with an adjustable fiber collimator (Thorlabs, CFC11A) to ensure that the 1.56 and 1 µm beams are overlapped on same focal plane after the high NA objective (OBJ1) shown in Fig. 2(b).
Fig. 3. (a) Output spectra of the bidirectional laser cavity (red, CW; blue, CCW); (b) interferogram resulting from the beating of the bidirectional laser outputs on a photodetector; (c) a single interferogram (inset, the zoom-in of the center part of the averaged interferogram with eight measurements).
In Fig. 2(a), the output from the 90% port of the CCW arm is spliced to another lab built EDFA to be amplified to $\sim{100}\;{\rm mW}$. Then, the EDFA output is spliced to $\sim{1.5}\;{\rm m}$ SMF-28 fiber for soliton compression. After the fiber compressor stage, the fiber output is spliced to $\sim{7}\;{\rm cm}$ HNLF to generate 1 µm via SC generation [26]. The splice loss is optimized to be $\lt{20}\%$ between the HNLF and the SMF-28. This approach has been used to build a high-power 1 µm femtosecond laser from a 1.55 µm mode-locked fiber laser in the past [26]. The HNLF output is directly spliced to a lab-built ytterbium-doped fiber amplifier (YDFA), and the 1 µm part is amplified to $\sim{100}\;{\rm mW}$. After the YDFA, $\sim{70}\;{\rm cm}$ Hi1060-flex fiber is used for matching the chirp-rate of the 1.56 µm pump pulses.
The output spectra for both pump and Stokes are measured and plotted in Figs. 4(a) and 4(c), respectively. According to the measurements, the spectral full width at half-maximum (FWHM) bandwidth is calculated to be $\sim{9.60}\;{\rm THz}$ for the 1 µm pulses and $\sim{4.78}\;{\rm THz}$ for the 1.56 µm pulses. For SF, the temporal pulse width of both arms is optimized to have the same ratio of their spectral FWHM bandwidths (${9.6}\;{\rm THz}/4.78\;{\rm THz} \,=\, \sim 2$). After the optimization process, the autocorrelation (AC) traces, for 1 µm and 1.56 µm, are measured and plotted in Figs. 4(b) and 4(d), respectively. The measured traces show the ratio of their pulse widths is ${8}\;{\rm ps}/4\;{\rm ps} = 2$. The AC traces are not deconvolved because we have assumed that the pulses have similar pulse shapes, so the deconvolution factor will be canceled out in the calculation.
Fig. 4. Spectra and autocorrelation traces of the pump (1 µm) and Stokes (1.55 µm) pulses for the CARS spectroscopic experiment. (a) Output spectrum of the 1 µm arm; (b) autocorrelation trace of the 1 µm arm (without deconvolution); (c) spectrum of the 1.56 µm arm; (d) autocorrelation trace of the 1.56 µm arm (no deconvolution).
A frequency step for this SF CARS can be obtained by multiplying the chirp-rate and the relative time delay $\Delta \tau$. The resolution of the CARS measurement is approximately equal to the frequency step times the number of pulse pairs that can be resolved by the detection bandwidth (${f_r}/{f_{{\rm detection}}}$). The optical resolution is determined by the pump’s and Stokes’ chirp-rate [20]. According to SF CARS calculation [20], the optical resolution with the current pulse duration can be $\sim{8}\;{{\rm cm}^{- 1}}$ by assuming a Gaussian pulse shape.
Next, the optimized 1 and 1.56 µm pulses are combined with a dichroic mirror. Then, the combined beam is sent into a scanning microscope system as shown in Fig. 2(b). A polarizer is inserted into the combined beam path to ensure that both 1 and 1.56 µm beams are linearly polarized and oriented in the same polarization direction. An oil immersion objective, OBJ1 in Fig. 2(b) (Olympus, 1.35 NA), is employed for the collinear CARS spectroscopic/imaging experiment [32]. The average powers on the sample are $\sim{25}\;{\rm mW}$ for pump and $\sim{15}\;{\rm mW}$ for Stokes, respectively. The forward emitted CARS signal is collected by a 0.65 NA objective. The collected signal is filtered with an optical filter (FF01-850/310, Semrock) and then detected by a photomultiplier tube (PMT) (H10720-20, Hamamatsu). The detected CARS signal is amplified in the electrical domain and sent to an oscilloscope to be visualized. This imaging system is designed to be capable of collecting epi-detected multiphoton signals [third-harmonic generation (THG), second-harmonic generation, and multiphoton fluorescence] as well, which can be used to obtain a multiphoton image at the same time as the CARS measurement.
To demonstrate that this laser system is capable of distinguishing different Raman spectra, two polystyrene (PS) bead samples (LB30, Sigma Aldrich) have been prepared and studied as shown in Fig. 5. One sample had PS beads only; another sample featured PS beads immersed in olive oil. In Fig. 5(a), a multiphoton image of the PS bead only sample has been captured from its THG excited by the 1.56 µm pulses. The time-resolved CARS spectra are obtained for both dark background [green square in Fig. 5(a)] and the PS bead [red square in Fig. 5(a)], and are plotted in Fig. 5(b) in the same color code. For the PS bead only sample, the time-resolved CARS signal only appeared when the laser was focused on the bead while the signal disappeared as the laser was focused on the air background. As a comparison, we imaged PS beads immersed in olive oil and measured the time-resolved CARS signal. As shown in Fig. 5(d), the time-resolved CARS signals appeared for both PS bead and the olive oil background. The signals appeared at different time delays with respect to the trigger because of the different C–H stretching modes between PS and olive oil.
Fig. 5. Multiphoton images and time-resolved CARS signals. (a) Multiphoton image PS only sample; (b) time-resolved CARS signal from PS only sample; (c) multiphoton image PS beads immersed in olive oil sample; (d) time-resolved CARS signal from corresponding sample; (e) measured Raman spectra for polystyrene bead, olive oil, polyethylene and polymethyl methacrylate (PMMA) obtained from the fiber dual-comb laser system.
Additionally, we imaged polyethylene microspheres (CPMS-0.96, Cospheric) and polymethyl methacrylate (PMMA) microspheres (PMPMS-1.4, Cospheric) for this study. The measured Raman spectra are plotted in Fig. 5(e), and each spectrum is averaged with eight measurements. For clarity, the spectra are normalized to their peaks. The important features of the Raman spectra are well resolved with the dual-comb laser system. We compared our results with the Raman spectra from published literatures: PS [33], olive oil [34], polyethylene [35], and PMMA [36]. Our spectra show a great similarity with the spontaneous Raman spectra from the above listed references. The signal-to-noise ratio (SNR) of the measurement is $\sim{50}$, and the resolution is $\sim{30}\;{{\rm cm}^{- 1}}$. The resolution of the experiment is currently limited by the sensitivity of the PMT current amplifier (SR570, Stanford Research). In order to resolve the different Raman spectra, we used 10 µA/V gain setting and the low-noise mode of the amplifier. The corresponding bandwidth of the amplifier is $\sim{200}\;{\rm kHz}$, which limits the resolution of the system to $\sim{30}\;{{\rm cm}^{- 1}}$. If the amplifier can resolve 100 pulse pairs ($\gt{1.14}\;{\rm MHz}$) with the same noise level and sensitivity, the resolution could be improved to $\lt{10}\;{{\rm cm}^{- 1}}$. With current detection system, reducing the repetition frequency difference $\Delta f$ to few Hz level may also improve the resolution to close to ${10}\;{{\rm cm}^{- 1}}$. Furthermore, making the pulses longer (lower chirp-rate) will be beneficial for improving the resolution as well. However, the longer pulses will have less peak power to excite the sample resulting in lower stimulated Raman signal, which might be detrimental to the SNR.
For future improvement, normal-dispersion HNLF [37] can be used to generate a flatter SC, with which the bidirectional laser system will be able to measure Raman resonances within the fingerprint window as well ($\sim{600}$ to ${1800}\;{{\rm cm}^{- 1}}$).
In summary, an all-fiber, single-cavity, free-running dual-comb laser system has been constructed for SF CARS spectroscopy. The measurements for proof-of-concept have been made with PS, olive oil, polyethylene, and PMMA. The C–H stretching signals have been observed in wavenumber window $\sim{2800}$ to ${3100}\;{{\rm cm}^{- 1}}$. The laser system has been made without compromising the all-fiber format. We believe this approach has promising potential for Raman spectroscopy and label-free imaging in many practical applications.
Funding
Achievement Rewards for College Scientists Foundation; National Science Foundation (DGE-1143953); State of Arizona TRIF funding; National Institutes of Health (1R01EB020605).
Acknowledgment
The authors thank Prof. Rongguang Liang for providing the polystyrene beads.
Disclosures
The authors declare no conflicts of interest.
REFERENCES
1. A. Zumbusch, G. R. Holtom, and X. S. Xie, Phys. Rev. Lett. 82, 4142 (1999). [CrossRef]
2. G. de Vito, I. Tonazzini, M. Cecchini, and V. Piazza, Opt. Express 22, 13733 (2014). [CrossRef]
3. J. Trägårdh, T. Pikálek, M. Šerý, T. Meyer, J. Popp, and T. Čižmár, Opt. Express 27, 30055 (2019). [CrossRef]
4. C. L. Evans, E. O. Potma, M. Puoris’haag, D. Cote, C. P. Lin, and X. S. Xie, Proc. Natl. Acad. Sci. USA 102, 16807 (2005). [CrossRef]
5. Y. Ozeki, W. Umemura, Y. Otsuka, S. Satoh, H. Hashimoto, K. Sum-imura, N. Nishizawa, K. Fukui, and K. Itoh, Nat. Photonics 6, 845 (2012). [CrossRef]
6. C. H. Camp Jr. and M. T. Cicerone, Nat. Photonics 9, 295 (2015). [CrossRef]
7. H. Lotem, R. T. Lynch, and N. Bloembergen, Phys. Rev. A 14, 1748 (1976). [CrossRef]
8. T. Hellerer, C. Axäng, C. Brackmann, P. Hillertz, M. Pilon, and A. Enejder, Proc. Natl. Acad. Sci. USA 104, 14658 (2007). [CrossRef]
9. J. M. Dudley, G. Genty, and S. Coen, Rev. Mod. Phys. 78, 1135 (2006). [CrossRef]
10. C. H. Camp Jr., Y. J. Lee, J. M. Heddleston, C. M. Hartshorn, A. R. H. Walker, J. N. Rich, J. D. Lathia, and M. T. Cicerone, Nat. Photonics 8, 627 (2014). [CrossRef]
11. H. Tu, Y. Liu, D. Turchinovich, M. Marjanovic, J. K. Lyngsø, J. Lægsgaard, E. J. Chaney, Y. Zhao, S. You, W. L. Wilson, B. Xu, M. Dantus, and S. A. Boppart, Nat. Photonics 10, 534 (2016). [CrossRef]
12. R. Selm, M. Winterhalder, A. Zumbusch, G. Krauss, T. Hanke, A. Sell, and A. Leitenstorfer, Opt. Lett. 35, 3282 (2010). [CrossRef]
13. Y. X. Yan and K. A. Nelson, J. Chem. Phys. 87, 6240 (1987). [CrossRef]
14. S. Ruhman, A. G. Joly, and K. A. Nelson, IEEE J. Quantum Electron. 24, 460 (1988). [CrossRef]
15. T. Ideguchi, S. Holzner, B. Bernhardt, G. Guelachvili, N. Picqué, and T. W. Hänsch, Nature 502, 355 (2013). [CrossRef]
16. K. Mohler, B. J. Bohn, M. Yan, T. W. Hänsch, and N. Picqué, Opt. Lett. 42, 318 (2017). [CrossRef]
17. N. Coluccelli, C. R. Howle, K. McEwan, Y. Wang, T. T. Fernandez, A. GambettA, P. Laporta, and G. Galzerano, Opt. Lett. 42, 4683 (2017). [CrossRef]
18. P.-L. Luo, M. Yan, T. W. Hänsch, and N. Picqué, Light, Energy and the Environment (Optical Society of America, 2016), paper FW2E.2.
19. T. Hellerer, A. M. K. Enejder, and A. Zumbusch, Appl. Phys. Lett. 85, 25 (2004). [CrossRef]
20. W. Langbein, I. Rocha-Mendoza, and P. Borri, J. Raman Spectrosc. 40, 800 (2009). [CrossRef]
21. I. Coddington, N. Newbury, and W. Swann, Optica 3, 414 (2016). [CrossRef]
22. K. Chen, T. Wu, T. Chen, H. Wei, H. Yang, T. Zhou, and Y. Li, Opt. Lett. 42, 3634 (2017). [CrossRef]
23. K. Kieu and M. Mansuripur, Opt. Lett. 32, 2242 (2007). [CrossRef]
24. K. Kieu and M. Mansuripur, Opt. Lett. 33, 64 (2008). [CrossRef]
25. S. Mehravar, R. Norwood, N. Peyghambarian, and K. Kieu, Appl. Phys. Lett. 108, 231104 (2016). [CrossRef]
26. K. Kieu, J. Jones, and N. Peyghambarian, Opt. Express 18, 21350 (2010). [CrossRef]
27. M. E. Fermann, V. I. Kruglov, B. C. Thomsen, J. M. Dudley, and J. D. Harvey, Phys. Rev. Lett. 84, 6010 (2000). [CrossRef]
28. V. I. Kruglov, A. C. Peacock, J. M. Dudley, and J. D. Harvey, Opt. Lett. 25, 1753 (2000). [CrossRef]
29. V. I. Kruglov, A. C. Peacock, J. D. Harvey, and J. M. Dudley, J. Opt. Soc. Am. B 19, 461 (2002). [CrossRef]
30. C. Finot, J. M. Dudley, B. Kibler, D. J. Richardson, and G. Millot, IEEE J. Quantum Electron. 45, 1482 (2009). [CrossRef]
31. B. Figueroa, W. Fu, T. Nguyen, K. Shin, B. Manifold, F. Wise, and D. Fu, Opt. Express 9, 6116 (2018). [CrossRef]
32. J. X. Cheng, A. Volkmer, and X. S. Xie, J. Opt. Soc. Am. B 19, 1363 (2002). [CrossRef]
33. J. X. Cheng and X. S. Xie, J. Phys. Chem. B 108, 827 (2004). [CrossRef]
34. A. M. Herrero, P. Carmona, F. J. Colmenero, and C. R. Capillas, Food Hydrocolloids 36, 374 (2014). [CrossRef]
35. C. Brideau, K. W. C. Poon, P. Colarusso, and P. K. Stys, J. Biomed. Opt. 24, 046502 (2019). [CrossRef]
36. V. K. Thakur, D. Vennerberg, S. A. Madbouly, and M. R. Kessler, RSC Adv. 4, 6677 (2014). [CrossRef]
37. N. Nishizawa and J. Takayanagi, J. Opt. Soc. Am. B 24, 1786 (2007). [CrossRef]
