Open Access
Add to BibSonomy
Abstract
A popular and inexpensive experimental approach to broadband CARS microscopy involves the generation of a supercontinuum pulse from a microstructured fibre. Despite such fibres being used for CARS microscopy for over a decade, concerns about excessive spectral and polarization noise have prevented their wider adoption for this technique. In this work, the applicability of two commercially-available supercontinuum generating fibres, the NL-1.4-775-945 and the NL-PM-750 in packaged module form, are compared for spectral-focusing CARS (SF-CARS) imaging. The former has been a mainstay of SF-CARS implementations while the latter has largely found application in a related but distinct broadband CARS technique and has not been used for SF-CARS applications. We first characterize the intensity and polarization behaviour of the supercontinuum output of the fibres and demonstrate that while both can be used to produce high-quality CARS images and spectra, key operational differences between the fibres must be taken into consideration when optimizing their use for SF-CARS applications. In particular, we show via spectroscopy of a solvent sample and hyperspectral imaging of a pharmaceutical sample that the NL-1.4-775-945 is inferior for SF-CARS applications when used under fixed-power supercontinuum generation conditions. However, its operation can be significantly improved under power-tuned supercontinuum generation conditions known as spectral-surfing, an approach that does not confer the same advantage with the NL-PM-750. Furthermore, we identify a latent polarization-dependent spectral artifact in the NL-PM-750, which can degrade the CARS spectrum in time-gated techniques such as SF-CARS.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Corrections
8 January 2019: A typographical correction was made to Ref. 41.
1. Introduction
Coherent anti-Stokes Raman scattering (CARS) is a four-wave mixing (FWM) technique used for nonlinear vibrational spectroscopy [1]. In CARS, a pump pulse is temporally and spatially overlapped with a lower-energy Stokes pulse in a sample, generating a higher-energy anti-Stokes pulse. While some anti-Stokes light is always created by electronic FWM processes, when the frequency difference between the pump and Stokes matches a Raman resonance in the sample, the anti-Stokes signal is significantly boosted such that it can be efficiently used for rapid label-free imaging. Beyond simple contrast-based imaging at a fixed CARS (Raman) frequency, the ability to map a broad vibrational spectrum in an imaging modality lends CARS hyperspectroscopy its greatest potential for biomedical diagnostics [2,3] and materials characterization [4–6].
Experimental implementations that rely on broadband laser pulses provide the most spectrally dense and rapid CARS hyperspectroscopy. What are generally referred to as “broadband CARS” techniques can be separated into two categories, based on how the anti-Stokes signal is detected: “multiplex” CARS techniques detect the CARS signal with a spectrometer, and can thus be thought of as spectroscopy-forward techniques that build up a microscopy image one pixel-spectrum at a time [7–10]. By contrast, spectral-focusing (SF) CARS (and various associated methods) uses single-element detectors such as PMTs or photodiodes to create images at fixed CARS frequencies and then scans the CARS frequency to create hyperspectral maps [11–14]. Most often, multiplex-CARS techniques are spectrally superior; SF-CARS techniques are faster for imaging. Femtosecond excitation pulses have bandwidths far greater than typical vibrational linewidths and therefore achieving useful spectral resolution in either multiplex-CARS or SF-CARS requires pulse shaping in the pump and Stokes beams. For example, in SF-CARS, significant dispersion is added to the pump and Stokes pulses such that their chirps are matched at the sample position; focusing their frequency difference to a narrow range at any particular pulse overlap [11,15].
To access the widest vibrational spectrum, broadband CARS techniques require synchronized pulses with sufficient spectral intensity across a wide wavelength span. Dual-laser sources and single ultrabroadband oscillators are straightforward but expensive approaches that can satisfy the needs of broadband CARS [13,16,17]. In most cases, however, these sources do not produce sufficiently broad Stokes (or pump) pulses to span the vibrational spectrum from deep in the fingerprint (<1600 cm−1), through the CH-stretch region (>2600 cm−1), and to the OH-stretch region (<3500 cm−1) in a single scan. A powerful and inexpensive means of producing a synchronized pulse with sufficient bandwidth to span the CARS spectrum is via supercontinuum generation (SCG) in a microstructured photonic-crystal fibre (PCF). The use of PCFs for broadband CARS microscopy was demonstrated over a decade ago [8,18], and they have been used since for multiplex-CARS [5,8,19,20] and SF-CARS [2,11,12,14,21–23] implementations, as well as others [24,25].
Over the past decade, two commercially-available PCFs designed for 800-nm-pumped supercontinuum Stokes generation have seen continued use. These are the NL-1.4-775-945 (“PCF-CARS”) and the NL-PM-750 (“PCF-800”) fibres; commonly found in a 12-cm format mounted in a fused-edge module for easy fibre coupling. The PCF-800 uses a polarization-maintaining fibre with two zero-dispersion-wavelengths (ZDWs) at 750 nm and 1260 nm [26], and was the first of the two to be used for broadband CARS microscopy [19]. The PCF-CARS fibre is not polarization maintaining and has closely-lying ZDWs at 775 nm and 945 nm. Both fibres are pumped at a wavelength within their anomalous dispersion range, typically in the vicinity of 800 nm. Although the precise spectral characteristics of the supercontinua generated from each of the fibres depends on the SCG pump-pulse characteristics [12,18,24,27,28], there are general characteristics to the outputs of each of the two fibres that provide insight into how they have been traditionally used for CARS microscopy. When pumped at 800 nm, the PCF-800 fibre generates a relatively uniform supercontinuum spanning 600 nm to 1200 nm [26,29]. Increasing the SCG pump power largely increases the spectral intensity everywhere, and thus many labs opt to maximize the SCG power, subject to noise considerations and damage thresholds [21,29–31]. On the other hand, the PCF-CARS fibre, with its closely lying ZDWs, generates highly variable red-shifted supercontinua with most of the light at wavelengths beyond the 945 nm ZDW [18,32]. Because the output of the PCF-CARS is brightest at pump/Stokes frequency differences corresponding to the ubiquitous CH-stretch region, the fibre was identified as being particularly applicable for CARS microscopy. Thus, the broad spectral characteristics of the PCF-800 have largely led to its adoption as best for multiplex-CARS [5,19,20,25], while the brightness of the output of the PCF-CARS near 1050 nm has led to its adoption for SF-CARS applications. There are no reports of the PCF-CARS fibre being used for multiplex-CARS applications.
A recent comparison of the use of both fibres for SF-CARS focused on the intensity noise characteristics of supercontinua from each fibre [30]. Most of the studied noise manifests at the repetition rate of the laser (usually tens of MHz). However, small differences in relative intensity noise (RIN) do not impact the quality of CARS imaging at μs pixel dwell times, and thus both fibres are viable for SF-CARS hyperspectroscopy applications [29]. Nonetheless, depending on the details of the experimental approach, variability in polarization and intensity across the supercontinuum are likely to have a large effect on the usability of the fibres.
While the PCF-CARS has traditionally been used to access mostly the CH/OH spectral region in SF-CARS [11,22–24], we’ve recently demonstrated a power-tuning approach that extends its use deep into the fingerprint [33]. The supercontinuum generated by the PCF-CARS was found to be sensitive to both the SCG pulse duration and pump power [14]. Utilizing longer, non-transform limited SCG pump pulses allows the generation of supercontinua with a strong soliton peak below the 945 nm ZDW, which enables access to fingerprint frequencies. Furthermore, the spectral maximum of the soliton was found to be power-dependent. Synchronizing the power-tuned generation of the supercontinuum with the pump/Stokes overlap (i.e. the frequency difference) significantly boosts the CARS signals deep in the fingerprint region. We call this approach “spectral surfing” of the supercontinuum [33]. The ability to spectrally-surf the PCF-800 has not previously been explored.
In this work, we show that the PCF-800 can be used for SF-CARS hyperspectroscopy that spans the fingerprint-to-OH region. We further compare the two fibres in terms of polarization output, CARS brightness, and the opportunities for spectrally-surfing the PCF-800 fibre. We demonstrate that in addition to well-known differences in the generated supercontinua intensities of the two fibres, that significant differences exist in their polarization spectrum. Furthermore, we identify an important complication with the polarization-maintaining aspect of the PCF-800 fibre. Specifically, pumping the PCF-800 at non-optimal polarization angles results in the generation of a secondary-soliton that gives rise to pseudo-CARS peaks, and may significantly degrade the spectral-resolution with SF-CARS implementations. We ultimately find that with careful and informed operation, either fibre can be used advantageously for inexpensive and agile SF-CARS hyperspectroscopy across a wide range of CARS frequencies.
2. Experiment
The SF-CARS microscopy setup uses a Ti:Sapph oscillator that generates 800-nm laser pulses with bandwidths of 100 cm−1 (FWHM ~7 nm). While the oscillator can output much shorter pulses, utilizing laser bandwidths around 100 cm−1 (> 150 fs pulse duration) is favorable for SCG with the PCF-CARS fibre as it enables the generation of a dominant soliton peak [24]. The spectral location of the main soliton can then be positioned through dynamic control of the soliton-self frequency shifts and then utilized as a strong Stokes beam for the CARS process, enabled by the experimental algorithm of spectral surfing [33]. To improve experimental repeatability, flip mirrors were used to alternately route the SCG pump between the PCF-CARS and the PCF-800. The Stokes generated by the PCF-CARS is further chirped using a total of 158 mm of high-dispersion glass (SNPH-2; Ohara), while the Stokes from the PCF-800 is further chirped with an additional 12.5 mm of highly dispersive glass (SNPH-1; Ohara) to compensate for its largely negative dispersion profile [26]. The blocks of high-dispersion glass effectively lengthen the duration of the red-shifted portion of the supercontinuum from both PCFs to more than 20 ps at the sample location. The pump is chirped using a total of 101 mm of high-dispersion glass (SNPH-2; Ohara) yielding a pulse duration of ~1 ps with a chirp parameter of approximately 120 cm−1/ps. The spectral surfing algorithm from our previous work [33] has been improved from piecewise-linear function fits to cubic-interpolation fits, which reduces erratic behavior of the computer-controlled variable attenuator. This improves experimental repeatability and the effectiveness of spectral surfing. More details regarding our experimental setup can be found in ref [14]. A key modification made for the current study was the establishment of epi-CARS detection by use of a dichroic beamsplitter (Semrock FF776) before the objective and use of a GaAs cathode PMT (Hamamatsu H7422-50) for improved detection of longer-wavelength anti-Stokes signals from deep in the fingerprint region (λAS<770 nm).
The polarization spectrum of the supercontinuum Stokes is detected after the microscope objective through the condenser and analyzer (Thorlabs LPNIRE-100-B). The analyzer was rotated in 5° increments, where the resulting polarized Stokes is routed to a spectrometer (StellarNet BLACK Comet CXR SR-50). This setup, without the analyzer and using spectrometers instead of PMTs, can be used for both system diagnostics and CARS frequency–delay-stage calibration of the CARS spectrum.
The samples used in this work are a glass coverslip (to demonstrate non-resonant FWM and efficacy of spectral surfing), benzonitrile (for CARS spectroscopy), and a Tylenol pill (for CARS hyperspectral imaging). The pump power used to measure the FWM from glass was approximately 300 mW (~225 mW at the sample), while 20 mW (~15 mW at the sample) was used for both benzonitrile and Tylenol. The SCG pump power used to generate the Stokes for spectral surfing was varied from 0 mW to 300 mW (with 55% fiber coupling efficiency) by a computer-controlled half-wave plate and polarizer pair acting as a variable attenuator. The hyperspectral image stacks were processed with freely-available software coded in Python [34], and with ImageJ [35]. Noise reduction was performed by applying an Anscombe transform and singular value decomposition on the hyperspectral image stack [36].
3. Polarization spectrum of supercontinua generated from the PCF-CARS and PCF-800 modules
Unintentional birefringence due to fabrication imperfections in a non-polarization-maintaining fibre such as the PCF-CARS may generate supercontinua with complicated polarization properties [37,38]. On the other hand, a polarization-maintaining fibre, such as the PCF-800, is expected to provide a fixed polarization across the output supercontinuum when pumped along the main polarization axis [37–39]. The PCF-800’s polarization behavior when pumped at a polarization angle away from its polarization axis has not been reported previously. The presence of polarization anisotropy in the PCF-800 may significantly affect its usefulness for CARS/FWM applications as most implementations of CARS microscopy have co-polarized pump and Stokes pulses in order to maximize CARS signal intensities. Use of non-copolarized beams will result in weaker overall CARS intensity due to depolarization ratios being less than or equal to 0.75 [40,41]. The presence of polarization anisotropy in the Stokes may yield non-copolarized pump and Stokes beams resulting in CARS microscopy systems generating less than ideal CARS signal intensities. Polarization anisotropy may also be detrimental to the application of spectral surfing, since its calibration procedure largely depends on the characterization of the power-dependence of the supercontinuum [33]. Furthermore, with polarization anisotropy, the non-resonant FWM, also known as non-resonant background (NRB) in CARS [42,43], may misrepresent the spectral shape of the supercontinuum Stokes. The simplest means of assuring co-polarization of both pulses is with the use of a combination of a half-wave-plate and a polarizer in each of the Stokes and pump arms. However, a study of the polarization state of the supercontinuum is important to the understanding of the optimal operating conditions of each of these fibres in its application to CARS/FWM microscopy.
Fig. 1 Polarization-dependent spectral intensity of the supercontinuum output from the PCF-CARS module, with low- and moderate-SCG pump powers. (a) The raw spectral-intensity polarization map of the supercontinuum with low incident SCG pump power condition (35 mW). (b) Supercontinuum polarization intensity map obtained by normalizing (a) at each wavelength to highlight the polarization state. (c) The raw spectral-intensity polarization map of the supercontinuum with moderate incident SCG pump power condition (220 mW). (d) Supercontinuum polarization intensity map obtained by normalizing (c) at each wavelength to highlight the polarization state. (e) The raw and polarization-filtered supercontinuum from PCF-CARS at the two power conditions. The polarization angle is set to 0°, as denoted along the dashed lines in (a)–(d).
Fig. 2 Polarization-dependent spectral intensity of the supercontinuum output from the PCF-800 module, with low- and moderate-SCG pump powers. (a) The raw spectral-intensity polarization map of the supercontinuum with low incident SCG pump power condition (35 mW). (b) Supercontinuum polarization intensity map obtained by normalizing (a) at each wavelength to highlight the polarization state. (c) The raw spectral-intensity polarization map of the supercontinuum with moderate incident SCG pump power condition (220 mW). (d) Supercontinuum polarization intensity map obtained by normalizing (c) at each wavelength to highlight the polarization state. (e) The raw and polarization-filtered supercontinuum spectra from PCF-800 at the two power conditions. The polarization angle is set to 0°, as denoted along the dashed lines in (a)–(d).
Pumping the PCF-800 along the polarization axis, not only assures that the output supercontinuum is uniformly polarized, but also prevents the formation of closely-timed dual solitons, which can degrade the SF-CARS spectrum, as described in the section on CARS spectroscopy below.
4. Non-resonant four-wave mixing with PCF-generated supercontinua
We use non-resonant four-wave-mixing [44] to measure the efficiency of CARS excitation throughout the vibrational spectrum. Glass is chosen as the sample because it demonstrates a relatively uniform electronic hyperpolarizability and doesn’t have strong vibrational resonances over the frequency region of interest. In terms of the CARS-relevant fields, the four-wave mixing signal can be written as:
where 𝜒(3) is the third-order electronic susceptibility containing vibrationally-resonant 𝜒R(3) and the non-resonant 𝜒NR(3) contributions, and IP(ωP) and IS(ωS) are the pump and Stokes intensities (frequencies), respectively. The signal intensity depends linearly on the Stokes intensity, and thus the FWM spectrum is reflective of the shape of the Stokes primarily due to its highly undulating spectrum [44]. A broad Stokes supercontinuum that is sufficiently bright to access the ~500 cm−1 to ~3500 cm−1 vibrational spectrum is desired for most CARS hyperspectral imaging applications to enable chemical identification. The absence of sufficient Stokes intensity in some frequency range will consequently limit the generation of anti-Stokes at these regions and this has been the most prominent limitation of the PCF-CARS [12,22,24,32] which was mainly utilized for its intensity at wavelengths >945 nm (ωP - ωS >1920 cm−1). Here, we enhance the Stokes generation of PCF-CARS by using a SCG pump with a relatively narrow bandwidth (~100 cm−1, > 150 fs transform-limited duration) and by implementing spectral surfing [33]. Furthermore, the pump beam and Stokes supercontinuum are co-polarized along 0°, as outlined in Figs. 1 and 2.Fig. 3 Nonresonant FWM from a glass coverslip, at three different incident SCG pump power conditions: moderate power (160 mW), high power (300 mW) and swept (surfed). (a) FWM with a SCG Stokes from the PCF-CARS and (b) from the PCF-800. (c) Ratio of surfed-to-static FWM from PCF-CARS and (d) from PCF-800. Unity ratio is shown as a dotted black line.
The power-tuning behavior of the PCF-800 is fundamentally different from that of the PCF-CARS, as shown in Figs. 3(b,d). As reported by Naji et al. [30], increasing the SCG pump power into the PCF-800 increases the output spectral power at all wavelengths beyond 830 nm, and extends the supercontinuum further into the IR. At moderate SCG powers, the supercontinuum has more power between 800 nm and 830 nm (providing access to the deep fingerprint), but considerably less power across the rest of the supercontinuum. At high SCG pump powers, the supercontinuum extends well beyond 1100 nm, allowing for considerably more CARS signal beyond 3500 cm−1. Ultimately, unlike the PCF-CARS, there is little advantage in implementing spectral surfing with the PCF-800. This is made clear in Fig. 3(d), where the ratio of FWM signal under the surfed condition to that of the high-power condition is essentially unity for all frequencies above 500 cm−1.
5. CARS spectroscopy with PCF-generated supercontinua
Unlike the non-resonant FWM process, the need to extract vibrational resonances through CARS leads to key challenges that have to be overcome by any Stokes source. Here, we demonstrate CARS spectroscopy of benzonitrile as a reference material, having sharp vibrational resonances from the deep fingerprint (450 cm−1) up to the CH/OH frequencies (3100 cm−1). The CARS spectrum from benzonitrile is used to demonstrate aspects of spectral resolution, as well as an anomalous spectral artifact that arises when pumping the PCF-800 away from the polarization axis.
As shown with FWM, above, pumping the PCF-CARS with large powers optimizes access to vibrational frequencies > 2000 cm−1 (Fig. 4(a)
Fig. 4 Spectral-focusing CARS spectroscopy of benzonitrile with Stokes supercontinuum generated by PCF-CARS and PCF-800 modules. (a) CARS spectrum obtained using the PCF-CARS, (b) the PCF-800 with dual soliton when pumped away from the fibre axis, and (c) the PCF-800 with single soliton when pumped along the polarization axis. In (a)-(c) three incident SCG power conditions are presented: 160 mW (red); 300 mW (green), and spectral surfing (blue). For clarity of display, the CARS spectra have different vertical scaling below and above 2,000 cm−1. While panels (a)–(c) use the same scale, variability in the sequential coupling between the two fibre modules is significantly large to caution against quantitative comparisons between the spectra. (d) When the PCF-800 is pumped with polarization away from the preferred axis, the anti-Stokes spectrogram from benzonitrile shows the appearance of dual-solitons separated in time; emphasized at the 2,220 cm−1 CARS resonance (anti-Stokes frequency of 14,720 cm−1). (e) When the PCF-800 is pumped with the preferred polarization, the anti-Stokes spectrogram from benzonitrile shows that the dual soliton shown in (d) is nearly extinguished. The inset shows the magnified 2,220 cm−1 resonance.
The PCF-800, on the other hand, easily covers a large range of the vibrational spectrum without the need for spectral surfing (Figs. 4(b,c))—a property that has led to its adoption for multiplex-CARS applications [19,20,47–50]. The polarization-maintaining aspects of this fibre are touted, in part, as a reason for its enhanced spectral generation stability. However, the advantages of polarization-maintenance can easily become detrimental when the fibres are (perhaps mindlessly) pumped away from the polarization axis. As shown in Fig. 4(b), supercontinuum generation with non-optimal polarization conditions degrades the CARS spectrum by seemingly broadening or doubling resonant peaks. Such peak doubling becomes increasingly prominent with higher SCG pump powers (Fig. 4(b)).
To determine the source of the degradation in the CARS spectrum that arises when the PCF-800 is pumped away from the preferred axis, we collected the transmitted anti-Stokes spectrum as a function of pump-delay stage position. This time-gated CARS spectrum, shown in Fig. 4(d), reveals the presence of a duplicate supercontinuum that is delayed by ~400 fs from the main pulse. Fundamentally, spectral-focusing CARS correlates pump-Stokes time-delays with CARS frequencies, the 400-fs delay of the secondary supercontinuum appears as a ~35 cm−1 offset in the spectrum, ultimately manifesting as a combination of peak doubling and/or peak broadening, as seen in Fig. 4(c). Recent works have referred to this generation of two supercontinua in highly-birefringent polarization-maintaining PCFs, as “dual-soliton generation” [51–53]. The generation of dual-solitons in the PCF-800 is similar with the observation by Chen et al. [52,53], albeit with a different PCF. Chen et al. utilized the presence of the dual-solitons, separated in time by more than a picosecond, to perform background reduction for multiplex CARS. The solitons generated by PCF-800, however, are too close (~400 fs, 130 µm path length difference) to be useful for background subtraction methods such as in dual-soliton CARS [52] or differential CARS [13,23,54]. Furthermore, because SF-CARS time-gates the anti-Stokes signal, the advantages of dual-soliton spectra that exist for multiplex-CARS become hindrances for SF-CARS schemes. In the case that the secondary-soliton could not be utilized, eliminating the generation of the secondary soliton from the PCF-800 is particularly important for SF-CARS.
Under our experimental conditions, the dual-solitons were found to have non-orthogonal polarizations, and thus cannot simply be eliminated with a post-module polarizer. Instead, it was found that the generation of one of the solitons can be minimized either by using weaker PCF pump powers, which lessens the non-linear interactions along the path of the fiber, or by making sure to pump the module along a primary polarization axis. Figure 4(e) shows the anti-Stokes spectrogram when the PCF is pumped along this axis. The generation of the secondary soliton is clearly minimized, although it can still be faintly observed at higher anti-Stokes frequencies close to the 3050 cm−1 peak (15,550 cm−1 anti-Stokes frequency). As shown in Fig. 4(c), the CARS spectrum of benzonitrile is improved under the single-soliton condition, and most of the double-peaks that are seen in Fig. 4(b) are eliminated. Nonetheless, a comparison with the CARS spectrum obtained with the PCF-CARS under spectral-surfing conditions suggests that the latter yields a brighter and better-resolved CARS spectrum across the entire fingerprint-to-CH vibrational frequency range.
6. CARS hyperspectral imaging
To demonstrate the qualitative performance of the two PCFs, we performed CARS hyperspectral imaging on a Tylenol oral-dose tablet. Having established that spectrally-surfing the PCF-800 does not confer a significant advantage over static operation, we compare the use of PCF-CARS under spectral surfing conditions with PCF-800 pumped at 300 mW. The 500 × 500 pixel hyperspectral image stack comprises of 960 spectral points, collected with a 2 µs pixel dwell time for a total collection time of 16 minutes. As shown in Fig. 5
Fig. 5 Hyperspectral CARS imaging of an acetaminophen oral form. (a) False-colour raw CARS image obtained using the PCF-CARS at the 790 cm−1 (red) and 850 cm−1 (green) resonances of Tylenol. (b) Contrast-enhanced image derived from (a); specifically, a CARS image from the PCF-CARS at 790 cm−1 (red) and 850 cm−1 (green), and at (c) 2920 cm−1 (red) and 3050 cm−1 (green). The lower panels (d-f) show the corresponding (d) raw CARS image, and (e,f) contrast-enhanced CARS images obtained using the PCF-800 showing the (d,e) 790 cm−1 and 850 cm−1, and (f) 2920 cm−1 and 3050 cm−1 resonances. (g) Raw CARS spectra using each of the two PCFs at regions of interest 1 and (h) 2, as denoted by squares in (f). The leftmost pair of red and green vertical lines in (g) and (h) represent the frequencies 790 cm−1, and 850 cm−1; the rightmost pair represents 2920 cm−1 and 3050 cm−1, all of which are used to obtain the CARS images (a-f). Images (a) and (d) show raw (unprocessed) CARS signal intensities, while images (b-c, e-f) are contrast-enhanced through image subtraction of the off-resonance “trough” CARS image from the on-resonance “peak” CARS image similar to the procedure used by Ryu et al. in [49]. The colour-maps of the colour-merged images (a,d) were individually intensity normalized (min/max), while the contrast-enhanced images (b,c,e,f) are all intensity-scaled with respect to image (c), and are thus directly comparable. The scale bar represents 25 µm.
The spectroscopic information obtained from the PCF-CARS shows that sharp resonances at the fingerprint can be easily observed while providing qualitatively better spectral resolution than from the PCF-800, as shown in Figs. 5(g) and 5(h). The differences in spectral resolution are most pronounced near 1500 cm−1, where three peaks are resolved in acetaminophen using the PCF-CARS, but only one broad peak is seen with the PCF-800. It is likely that remaining dual-soliton character is degrading the CARS spectrum when using this fibre. Nonetheless, the spectral resolution of the PCF-800 can be improved with increased dispersion and/or improved chirp-matching [44]. In terms of CARS spectral strength, the PCF-CARS can clearly generate strong sub-900 cm−1 fingerprint signals that are comparable in strength to the CARS intensities in the 2000 cm−1–3000 cm−1 region. The PCF-800 on the other hand, has CARS intensities that gradually decrease from 3000 cm−1 down to ~350 cm−1, excluding that of particularly strong resonances between 1200 cm−1–1700 cm−1. From the standpoint of CARS spectroscopy, the spectrally-surfed PCF-CARS outperforms the PCF-800 in the sub-900 cm−1 region, as shown by comparison of Figs. 5(g) and 5(h).
The composite images generated from the hyperspectral data show the most notable difference between the performance of the fibres, specifically in the fingerprint (Figs. 5b, e). With the PCF-CARS, a highly contrasting composite image can be generated from the 790 cm−1 resonance of acetaminophen (red) and the 850 cm−1 resonance of sucrose (green) as shown in Fig. 5(b). This can be explained by the spectral contrast obtained from the two regions-of-interests (ROIs) located at each respective chemical’s location. Acetaminophen (ROI-1) was found to have a strong 790 cm−1 but a weak 850 cm−1 CARS response (Fig. 5(g)), while sucrose (ROI-2) has an oppositely weak 790 cm−1 but a strong 850 cm−1 CARS response (Fig. 5(h)). The PCF-800 on the other hand, was able to resolve a resonance at ~850 cm−1 (ROI-2, Fig. 5(h)) but fails to resolve a resonance around 790 cm−1 (ROI-1, Fig. 5(g)) and was thus unable to generate adequate contrast for a composite image (Fig. 5(e)). The performance of the PCF-800 in this case may be attributed to the combined effect of sub-optimal spectral resolution, which causes the non-resonant four-wave mixing to overwhelm resonant signals [58], and the generation of weaker Stokes due to its relatively “flatter” but broader SC, as outlined above.
In the CH-stretch region, at 2920 cm−1 and 3050 cm−1, both fibres are able to obtain similar CARS images (Figs. 5(c,f)) albeit with poorer spectral resolution and weaker signal intensity with the PCF-800. Overall, we find that while static high-SCG pump power operation of the PCF-800 works adequately well for hyperspectral SF-CARS imaging in the CH/OH and silent regions, though somewhat less well in the fingerprint region, a spectrally-surfed PCF-CARS is a superior Stokes supercontinuum source for these applications.
7. Conclusion
We have compared the use of two commercially-available microstructured-fibre-based supercontinuum sources for spectral-focusing CARS hyperspectroscopy. Both the PCF-CARS and the PCF-800 can be used for high-quality hyperspectral imaging without much consideration of supercontinuum noise at the repetition rate of the laser. In terms of experimental application, the PCF-800 is less complex to operate because its supercontinuum output is both broader and more uniform in polarization and intensity than that from the PCF-CARS. However, spectral surfing—the synchronized power-tuning of the supercontinuum generation—uniquely and significantly improves the usefulness of the PCF-CARS for spectral-focusing CARS applications, providing superior spectral and imaging characteristics over the PCF-800. By contrast to the operation of the PCF-CARS, increasing the SCG pump power in the PCF-800 broadens and overall brightens the output—rather than shifting the wavelength location of supercontinuum peaks. Thus, spectral-surfing is not motivated with the PCF-800. Operation of the PCF-800 can be complicated by the latent generation of dual-solitons, particularly when pumped at polarizations away from its primary polarization-maintaining axis. Such dual-solitons represent a time-delayed duplication of the supercontinuum that gives rise to spurious peak doubling and broadening in SF-CARS. This fibre has previously been touted for use in multiplex-CARS applications that do not rely on time-gating of a supercontinuum and thus are impervious to (or perhaps even make advantage of) the supercontinuum duplicate. Our results contrast with the details of past comparisons of the two PCF modules’ applicability to CARS microscopy [30], but confirms the fact that either PCF can be used to advantage in inexpensive approaches to SF-CARS hyperspectroscopy.
Considering that the non-polarization-maintaining PCF-CARS was used with significant attenuation due to the presence of polarization optics, a polarization-maintaining version of this fibre [28] should yield stronger supercontinuum intensities that are even more useful for broadband SF-CARS applications.
Funding
Canada Research Chair-Natural Sciences and Engineering Research Council of Canada (CRC-NSERC-231086); NSERC Discovery Award (RGPIN-2018-04491).
Acknowledgments
Jeremy Porquez acknowledges funding from the Ontario Trillium Scholarship program. We thank Joel Tabarangao for technical assistance with the CARS microscope. We thank Andrew Ridsdale and NRC Canada for loan of the PCF-800 module.
References
1. W. M. Tolles, J. W. Nibler, J. R. McDonald, and A. B. Harvey, “A Review of the Theory and Application of Coherent Anti-Stokes Raman Spectroscopy (CARS),” Appl. Spectrosc. 31(4), 253–271 (1977). [CrossRef]
2. H. Tu, Y. Liu, D. Turchinovich, M. Marjanovic, J. Lyngsø, J. Lægsgaard, E. J. Chaney, Y. Zhao, S. You, W. L. Wilson, B. Xu, M. Dantus, and S. A. Boppart, “Stain-free histopathology by programmable supercontinuum pulses,” Nat. Photonics 10(8), 534–540 (2016). [CrossRef] [PubMed]
3. M. Weinigel, H. G. Breunig, M. Kellner-Höfer, R. Bückle, M. E. Darvin, M. Klemp, J. Lademann, and K. König, “In vivo histology: Optical biopsies with chemical contrast using clinical multiphoton/coherent anti-Stokes Raman scattering tomography,” Laser Phys. Lett. 11(5), 055601 (2014). [CrossRef]
4. Y. J. Lee, C. R. Snyder, A. M. Forster, M. T. Cicerone, and W. L. Wu, “Imaging the molecular structure of polyethylene blends with broadband coherent Raman microscopy,” ACS Macro Lett. 1(11), 1347–1351 (2012). [CrossRef]
5. C. M. Hartshorn, Y. J. Lee, C. H. Camp Jr., Z. Liu, J. Heddleston, N. Canfield, T. A. Rhodes, A. R. Hight Walker, P. J. Marsac, and M. T. Cicerone, “Multicomponent chemical imaging of pharmaceutical solid dosage forms with broadband CARS microscopy,” Anal. Chem. 85(17), 8102–8111 (2013). [CrossRef] [PubMed]
6. R. C. Burruss, A. D. Slepkov, A. F. Pegoraro, and A. Stolow, “Unraveling the complexity of deep gas accumulations with three-dimensional multimodal CARS microscopy,” Geology 40(12), 1063–1066 (2012). [CrossRef]
7. J. Cheng, A. Volkmer, L. D. Book, and X. S. Xie, “Multiplex coherent anti-stokes raman scattering microspectroscopy and study of lipid vesicles,” J. Phys. Chem. B 106(34), 8493–8498 (2002). [CrossRef]
8. T. W. Kee and M. T. Cicerone, “Simple approach to one-laser, broadband coherent anti-Stokes Raman scattering microscopy,” Opt. Lett. 29(23), 2701–2703 (2004). [CrossRef] [PubMed]
9. Y. J. Lee, D. Moon, K. B. Migler, and M. T. Cicerone, “Quantitative Image Analysis of Broadband CARS Hyperspectral Images of Polymer Blends,” Anal. Chem. 83(7), 2733–2739 (2011). [CrossRef] [PubMed]
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, “High-speed coherent Raman fingerprint imaging of biological tissues,” Nat. Photonics 8(8), 627–634 (2014). [CrossRef] [PubMed]
11. 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]
12. A. F. Pegoraro, A. D. Slepkov, A. Ridsdale, D. J. Moffatt, and A. Stolow, “Hyperspectral multimodal CARS microscopy in the fingerprint region,” J. Biophotonics 7(1-2), 49–58 (2014). [CrossRef] [PubMed]
13. I. Pope, W. Langbein, P. Watson, and P. Borri, “Simultaneous hyperspectral differential-CARS, TPF and SHG microscopy with a single 5 fs Ti:Sa laser,” Opt. Express 21(6), 7096–7106 (2013). [CrossRef] [PubMed]
14. J. G. Porquez, R. A. Cole, J. T. Tabarangao, and A. D. Slepkov, “Spectrally-broad coherent anti-Stokes Raman scattering hyper-microscopy utilizing a Stokes supercontinuum pumped at 800 nm,” Biomed. Opt. Express 7(10), 4335–4345 (2016). [CrossRef] [PubMed]
15. T. Hellerer, A. M. K. Enejder, and A. Zumbusch, “Spectral focusing: High spectral resolution spectroscopy with broad-bandwidth laser pulses,” Appl. Phys. Lett. 85(1), 25–27 (2004). [CrossRef]
16. A. Zumbusch, G. R. Holtom, and X. S. Xie, “Three-Dimensional Vibrational Imaging by Coherent Anti-Stokes Raman Scattering,” Phys. Rev. Lett. 82(20), 4142–4145 (1999). [CrossRef]
17. J. Cheng, A. Volkmer, L. D. Book, and X. S. Xie, “An Epi-Detected Coherent Anti-Stokes Raman Scattering (E-CARS) Microscope with High Spectral Resolution and High Sensitivity,” J. Phys. Chem. B 105(7), 1277–1280 (2001). [CrossRef]
18. K. M. Hilligsøe, T. Andersen, H. Paulsen, C. Nielsen, K. Mølmer, S. Keiding, R. Kristiansen, K. Hansen, and J. Larsen, “Supercontinuum generation in a photonic crystal fiber with two zero dispersion wavelengths,” Opt. Express 12(6), 1045–1054 (2004). [CrossRef] [PubMed]
19. Y. J. Lee, Y. Liu, and M. T. Cicerone, “Characterization of three-color CARS in a two-pulse broadband CARS spectrum,” Opt. Lett. 32(22), 3370–3372 (2007). [CrossRef] [PubMed]
20. J. Rehbinder, C. Pohling, T. Buckup, and M. Motzkus, “Multiplex coherent anti-Stokes Raman microspectroscopy with tailored Stokes spectrum,” Opt. Lett. 35(22), 3721–3723 (2010). [CrossRef] [PubMed]
21. A. F. Pegoraro, A. Ridsdale, D. J. Moffatt, J. P. Pezacki, B. K. Thomas, L. Fu, L. Dong, M. E. Fermann, and A. Stolow, “All-fiber CARS microscopy of live cells,” Opt. Express 17(23), 20700–20706 (2009). [CrossRef] [PubMed]
22. B.-C. Chen, J. Sung, X. Wu, and S.-H. Lim, “Chemical imaging and microspectroscopy with spectral focusing coherent anti-Stokes Raman scattering,” J. Biomed. Opt. 16(2), 021112 (2011). [CrossRef] [PubMed]
23. B. Li, P. Borri, and W. Langbein, “Dual/differential coherent anti-Stokes Raman scattering module for multiphoton microscopes with a femtosecond Ti:sapphire oscillator,” J. Biomed. Opt. 18(6), 066004 (2013). [CrossRef] [PubMed]
24. S. Murugkar, C. Brideau, A. Ridsdale, M. Naji, P. K. Stys, and H. Anis, “Coherent anti-Stokes Raman scattering microscopy using photonic crystal fiber with two closely lying zero dispersion wavelengths,” Opt. Express 15(21), 14028–14037 (2007). [CrossRef] [PubMed]
25. P. Groß, L. Kleinschmidt, S. Beer, C. Cleff, and C. Fallnich, “Single-laser light source for CARS microscopy based on soliton self-frequency shift in a microstructured fiber,” Appl. Phys. B Lasers Opt. 101(1–2), 167–172 (2010). [CrossRef]
26. N. K. T. Photonics, “FemtoWHITE 800,” https://www.nktphotonics.com/wp-content/uploads/sites/3/2015/03/femtoWHITE-800.pdf.
27. J. M. Dudley, G. Genty, and S. Coen, “Supercontinuum generation in photonic crystal fiber,” Rev. Mod. Phys. 78(4), 1135–1184 (2006). [CrossRef]
28. P. Klarskov, A. Isomäki, K. P. Hansen, and P. E. Andersen, “Supercontinuum generation for coherent anti-Stokes Raman scattering microscopy with photonic crystal fibers,” Opt. Express 19(27), 26672–26683 (2011). [CrossRef] [PubMed]
29. J. G. Porquez, E. R. Korfanty, A. D. Slepkov, and W. Bank, “Ultra-broadband coherent anti-Stokes Raman scattering microscopy with a dynamically power-tuned Stokes supercontinuum,” Proc. SPIE 10069, 1–10 (2017).
30. M. Naji, S. Murugkar, and H. Anis, “Determining optimum operating conditions of the polarization-maintaining fiber with two far-lying zero dispersion wavelengths for CARS microscopy,” Opt. Express 22(9), 10800–10814 (2014). [CrossRef] [PubMed]
31. Y. Liu, Y. Zhao, J. Lyngsø, S. You, W. L. Wilson, H. Tu, and S. A. Boppart, “Suppressing Short-term Polarization Noise and Related Spectral Decoherence in All-normal Dispersion Fiber Supercontinuum Generation,” J. Lightwave Technol. 33(9), 1814–1820 (2015). [CrossRef] [PubMed]
32. N. K. T. Photonics, “FemtoWHITE CARS,” http://www.nktphotonics.com/wp-content/uploads/sites/3/2015/03/femtoWHITE-CARS.pdf.
33. J. G. Porquez, R. A. Cole, J. T. Tabarangao, and A. D. Slepkov, “Brighter CARS hypermicroscopy via “spectral surfing” of a Stokes supercontinuum,” Opt. Lett. 42(12), 2255–2258 (2017). [CrossRef] [PubMed]
34. J. G. Porquez, “CRASHyperspectralAnalysis,” https://github.com/JeremyPorquez/CRASHyperspectralAnalysis.
35. J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, J. Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P. Tomancak, and A. Cardona, “Fiji: an open-source platform for biological-image analysis,” Nat. Methods 9(7), 676–682 (2012). [CrossRef] [PubMed]
36. C. H. Camp Jr., Y. J. Lee, and M. T. Cicerone, “Quantitative, comparable coherent anti-Stokes Raman scattering (CARS) spectroscopy: Correcting errors in phase retrieval,” J. Raman Spectrosc. 47(4), 408–415 (2016). [CrossRef] [PubMed]
37. M. J. Steel, T. P. White, C. Martijn de Sterke, R. C. McPhedran, and L. C. Botten, “Symmetry and degeneracy in microstructured optical fibers,” Opt. Lett. 26(8), 488–490 (2001). [CrossRef] [PubMed]
38. A. Ortigosa-Blanch, J. C. Knight, W. J. Wadsworth, J. Arriaga, B. J. Mangan, T. A. Birks, and P. S. J. Russell, “Highly birefringent photonic crystal fibers,” Opt. Lett. 25(18), 1325–1327 (2000). [CrossRef] [PubMed]
39. K. Okamoto, T. Edahiro, and N. Shibata, “Polarization properties of single-polarization fibers,” Opt. Lett. 7(11), 569–571 (1982). [CrossRef] [PubMed]
40. Y. Saito, T. Ishibashi, and H. Hamaguchi, “Polarization-resolved coherent anti-Stokes Raman scattering (CARS) spectroscopy: a new probe of molecular symmetry through accurate determination of the Raman depolarization ratio,” J. Raman Spectrosc. 31(8–9), 725–730 (2000). [CrossRef]
41. The anti-Stokes intensity can be written as , where is the angle between the pump and Stokes beams, and is the depolarization ratio (less than unity). Use of co-polarized beams will generate anti-Stokes, , while use of orthogonal pump and Stokes beams will generate (a weaker) anti-Stokes, .
42. H. A. Rinia, M. Bonn, M. Müller, and E. M. Vartiainen, “Quantitative CARS spectroscopy using the maximum entropy method: the main lipid phase transition,” ChemPhysChem 8(2), 279–287 (2007). [CrossRef] [PubMed]
43. Y. Liu, Y. J. Lee, and M. T. Cicerone, “Broadband CARS spectral phase retrieval using a time-domain Kramers-Kronig transform,” Opt. Lett. 34(9), 1363–1365 (2009). [CrossRef] [PubMed]
44. Y. Wang, C.-Y. Lin, A. Nikolaenko, V. Raghunathan, and E. O. Potma, “Four-wave mixing microscopy of nanostructures,” Adv. Opt. Photonics 3(1), 1 (2011). [CrossRef]
45. R. A. Cole and A. D. Slepkov, “Interplay of pulse bandwidth and spectral resolution in spectral-focusing CARS microscopy,” J. Opt. Soc. Am. B 35(4), 842 (2018). [CrossRef]
46. M. Mohseni, C. Polzer, and T. Hellerer, “Resolution of spectral focusing in coherent Raman imaging,” Opt. Express 26(8), 10230–10241 (2018). [CrossRef] [PubMed]
47. H. Kano and H. Hamaguchi, “Dispersion-compensated supercontinuum generation for ultrabroadband multiplex coherent anti-Stokes Raman scattering spectroscopy,” J. Raman Spectrosc. 37(1–3), 411–415 (2006). [CrossRef]
48. H. Kano, “Molecular vibrational imaging of a human cell by multiplex coherent anti-Stokes Raman scattering microspectroscopy using a supercontinuum light source,” J. Raman Spectrosc. 39(11), 1649–1652 (2008). [CrossRef]
49. I. S. Ryu, C. H. Camp Jr., Y. Jin, M. T. Cicerone, and Y. J. Lee, “Beam scanning for rapid coherent Raman hyperspectral imaging,” Opt. Lett. 40(24), 5826–5829 (2015). [CrossRef] [PubMed]
50. J. H. Lee, E. H. Cho, S.-M. Shin, M. Oh, and D.-K. Ko, “Comparative Study of Breast Normal and Cancer Cells Using Coherent Anti-Stokes Raman Scattering Microspectroscopy Imaging,” Appl. Phys. Express 5(8), 082401 (2012). [CrossRef]
51. F. R. Arteaga-Sierra, C. Milián, I. Torres-Gómez, M. Torres-Cisneros, G. Moltó, and A. Ferrando, “Supercontinuum optimization for dual-soliton based light sources using genetic algorithms in a grid platform,” Opt. Express 22(19), 23686–23693 (2014). [CrossRef] [PubMed]
52. K. Chen, T. Wu, H. Wei, and Y. Li, “Dual-soliton Stokes-based background-free coherent anti-Stokes Raman scattering spectroscopy and microscopy,” Opt. Lett. 41(11), 2628–2631 (2016). [CrossRef] [PubMed]
53. T. Wu, K. Chen, H. Zhao, W. Zhang, Y. Li, and H. Wei, “Flexible dual-soliton manipulation for coherent anti-Stokes Raman scattering spectroscopy,” Opt. Express 26(17), 22001–22010 (2018). [CrossRef] [PubMed]
54. I. Rocha-Mendoza, W. Langbein, P. Watson, and P. Borri, “Differential coherent anti-Stokes Raman scattering microscopy with linearly chirped femtosecond laser pulses,” Opt. Lett. 34(15), 2258–2260 (2009). [CrossRef] [PubMed]
55. J. F. Kauffman, L. M. Batykefer, and D. D. Tuschel, “Raman detected differential scanning calorimetry of polymorphic transformations in acetaminophen,” J. Pharm. Biomed. Anal. 48(5), 1310–1315 (2008). [CrossRef] [PubMed]
56. M. de Veij, P. Vandenabeele, T. De Beer, J. P. Remon, and L. Moens, “Reference database of Raman spectra of pharmaceutical excipients,” J. Raman Spectrosc. 40(3), 297–307 (2009). [CrossRef]
57. J. G. Porquez and A. D. Slepkov, “Application of spectral-focusing-CARS microscopy to pharmaceutical sample analysis,” AIP Adv. 8(9), 095213 (2018). [CrossRef]
58. M. Cui, B. R. Bachler, and J. P. Ogilvie, “Comparing coherent and spontaneous Raman scattering under biological imaging conditions,” Opt. Lett. 34(6), 773–775 (2009). [CrossRef] [PubMed]
