Open Access
Add to BibSonomyAbstract
We demonstrate for the first time, a portable multimodal coherent anti-Stokes Raman scattering microscope (exoscope) for minimally invasive in-vivo imaging of tissues. This device is based around a micro-electromechanical system scanning mirror and miniaturized optics with light delivery accomplished by a photonic crystal fibre. A single Ti:sapphire femtosecond pulsed laser is used as the light source to produce CARS, two photon excitation fluorescence and second harmonic generation images. The high resolution and distortion-free images obtained from various resolution and bio-samples, particularly in backward direction (epi) successfully demonstrate proof of concept, and pave the path towards future non or minimally-invasive in vivo imaging.
©2013 Optical Society of America
1. Introduction
As a label-free imaging modality coherent anti-Stokes Raman scattering (CARS) has proven over the last decade to be a crucial diagnostic tool within the vibrational spectroscopic imaging domain [1]. While CARS microscopy has been demonstrated in preliminary research performed in many fields including cancer identification [2], drug therapy selection [3], biological lipid droplet imaging [4], myelinated axon structure [5], spinal cord injuries and demyelination [6], determining the oxygenation state of blood within individual vessels [7], arterial atherosclerosis detection [8], determination of hepatic fat content of liver tissue [9],its expense, due to high equipment costs, and the need for highly qualified personnel for operation has limited its propagation into mainstream medicine. Based on this broad list of applications, it is apparent that the functionality of CARS is unique, and its proliferation into mainstream medicine could most certainly benefit from the development of a cost-effective, miniaturized and portable, imaging platform [10, 11].
The CARS process is based on the fundamentals of Raman scattered photons. However, Raman scattering is inherently weak; typical photon conversion efficiencies for Raman are lower than 1 in every 1018 [1]. CARS exploits the generation of inelastically scattered photons by stimulating this typically spontaneous event. This is accomplished through the interaction of three input photons: a pump (e.g. at 800 nm), a Stokes (e.g. at 1030 nm) and a probe (e.g. at 800 nm) that together generate the anti-Stokes photons (since the pump and the probe are the same beam in degenerate CARS, they will be referred to as simply the pump). The difference between the pump and Stokes wavelengths can be tuned to resonantly excite and stimulate the emission of anti-Stokes photons, allowing the CARS process to monitor the presence of specific molecular bond energies. In the case of the previously mentioned wavelengths, the CH-rich lipid region is strongly excited in a resonant manner.
When CARS is combined with other nonlinear imaging and spectroscopic techniques such as two photon excitation fluorescence (TPEF) and second harmonic generation (SHG), the utility of the acquired images greatly increases; considerable chemical information can be extracted from biological tissues [5, 8, 12]. Imaging these tissues ex vivo is useful but in order to make nonlinear imaging amenable to clinical use, endoscopic variants of current microscopes must be developed. Currently, extensive research has been performed on integrating fibre-delivered SHG and TPEF probes [13–22] but fibre-delivered CARS endoscopic research hasn’t been as prolific [22–24]. The difficulty with CARS as an imaging modality is that it is absolutely critical that the pump and Stokes beams overlap both temporally and spatially at the sample. Without this temporal and spatial overlap, there is no vibrational resonance created within the bonds of the molecule being probed, as the pump and Stokes photons cannot interact in tandem with the molecule. This fact causes many complications. First, a pair of ultrashort pulses separated by more than 200 nm must retain high peak powers and temporally overlap at the sample, overcoming dispersions induced by the delivery fibre. Second, size-constrained focusing optics that have little chromatic aberrations to ensure spatial overlap of both pump and Stokes beams at the sample, while maintaining high CARS resolution must be implemented. Lastly, the fibre delivery of a pair of ultrashort pulses can produce unwanted nonresonant effects including four-wave mixing (FWM) within a delivery fibre which can overwhelm any sample-generated CARS signals.
The vast majority of endoscopic CARS research has centred on the delivery of femtosecond and picosecond light. Significant work characterizing delivery fibres has been performed analyzing temporal and spectral characteristics, beam quality, damage thresholds and efficiencies of various fibres types, including single-mode fibres, fibre bundles and PCFs [19, 24–26]. Fibre based investigations have also been performed reducing unwanted four-wave mixing that is generated within the delivery fibres, through means of polarization control [27]. This work represents significant milestones in CARS endoscopy, although macro optics, or galvanometric scanners are used. While there has been recent research into piezo tube fibre scanning probes in conjunction with gradient index (GRIN) lenses for CARS and stimulated Raman scattering [28], the high costs associated with the development of a custom GRIN lens led to the use of a MEMS scanner. The high driving frequencies of a micro-electromechanical system (MEMS) mirror still allows for video rate imaging. To the best of our knowledge, we have developed the first MEMS scanner based portable CARS miniaturized microscope with the unique capability of collecting CARS, SHG, and TPEF photons in the epi direction, with the entire exoscope demonstrated in Fig. 1(a). Our previously published work [29] was based on a bench-top setup that neither included fibre delivery, nor the integration of the MEMS scanner into a miniaturized portable platform. In this paper, we demonstrate for the first time, an optical probe that not only allows capture of nonlinearly generated photons in the epi direction through an exoscope-connected multi-mode fibre, but in addition offers device portability. Moreover, the laser source for this exoscope is based on a single femtosecond laser, which reduces the required footprint for the entire system. Integrating this microscope with an entirely fibre based CARS system [30] would further improve the robustness and compactness of the device, allowing real-time bedside, nonlinear imaging and patient diagnostics.
Fig. 1 (a) Portability demonstration of the exoscope (b) Schematic of the experimental setup for the multimodal CARS imaging using the exoscope in (a). Various components are as follows: (1) Titanium sapphire laser source (2) Isolator (3) Half wave plate (4) Polarizing beam splitters (5) Prism compressor (6) 40X microscope objective lens (7) Supercontinuum generation PCF (8) Aspheric 5 mm focal length lens (9) 1040 nm central wavelength band-pass filter (10) Grating compressor (11) Short pass dichroic mirror (12) 5X microscope objective lens (13) LMA 20 PCF (14) Exoscope (15) 40X water immersion lens (16) Multimode collection fibre (17) Short-pass Filter (18) Hamamatsu PMT (19) Discriminator and field programmable gate array.
2. Materials and methods
2.1 Experimental setup
Our CARS system is based on a single Ti:sapphire femtosecond laser Fig. 1(b1) (Tsunami, Spectra-Physics, USA). The source is tuneable (between 700 and 1000 nm) and produces a transform limited ~70fs pulse train at an 80 MHz repetition rate. This light is split into pump and Stokes arms.
As described in detail in our earlier work [29, 31], 250 mW of this light at ~800 nm is directed into a FemtoWhite photonic crystal fibre (PCF) (NKT Photonics, Denmark) for creating the Stokes beam. The remainder (~600 mW) is the pump beam used for CARS, TPEF and SHG imaging. The supercontinuum output from the PCF Fig. 1(b7) is band-pass filtered Fig. 1(b9) (D1040/60 M, Chroma Technology, USA) so that it consists of wavelengths between 1014 and 1067 nm. The pump beam is sent through a computer-controlled delay stage to maintain temporal overlap of pump beam, as well as a grating compressor to compensate for the dispersion of the delivery fibre. The pump is then recombined with the Stokes beam at a dichroic mirror Fig. 1(b11). The spatially overlapped beams are coupled to the delivery fibre Fig. 1(b13) (LMA-20, NKT Photonics) by a 5x 0.09 numerical aperture (NA) microscope objective lens Fig. 1(b12). The fibre-coupled light is delivered to the vertically mounted exoscope Fig. 1(b14). The sample is mounted onto a three axis computer controllable stage. The generated signals can be collected in the forward direction by a 40x, 0.8 NA water immersion microscope objective Fig. 1(b15) (Olympus, Japan) or in the epi direction by the optical head of the exoscope. It was observed that upon exit, the excitation beams had a slight off axis tilt, measured to be ~8°. This was compensated by tilting the exoscope body itself to maintain a level imaging plane. The implications of this tilt for the imaging performance of the exoscope are discussed in the following sections.
In both forward and epi configurations, light is collected by a large core (1.0 mm) multi-mode fibre Fig. 1(b16) (Thorlabs, USA). For epi collection, the fibre is mounted on the body of the exoscope while for forward collection the fibre is mounted above the exoscope. Before entering the photomultiplier tube (PMT) Fig. 1(b18), a 680 nm short pass filter Fig. 1(b17) (ET680SP-2P8, Chroma Technology) is used to remove the excitation light. Additional filters appropriate to the imaging modality can be introduced in the signal collection path as needed.
2.2 Miniaturized optics
The miniature objective lens was designed to enable in-vivo imaging of rat spinal cord. This required that the distal end of the exoscope must not have an outer diameter greater than 3.0 mm. Based on this prerequisite, the internal optics of the exoscope could have a maximum diameter of 2.0 mm while still maintaining a high NA of 0.6 and a long enough working distance of 400 µm while still attaining sub-micron resolution. The reasoning behind the choice to design a miniaturized set of optics is discussed in our previous work [29].
Designing small-diameter bulk optics to achieve high NA with as little chromatic aberrations as possible for widely-spaced wavelengths was non-trivial in such a constrained mounting volume. The chosen design includes several lenses with varying focal lengths ranging from 3.0 mm diameter BK7 glass for input light collimation down to a combination of SF4 and FK51 glass of 1.8 mm diameter for a 5x post-MEMS beam diameter expansion to fill the back aperture of the focusing optics. All of these optical elements are coated with broadband anti-reflection coatings with a reflection of less than 1% over 652-1036 nm on both air/glass surfaces. This low reflection per surface is made possible by a custom four layer deposition of alternating hafnium oxide and silicon dioxide (BMV Optical Technologies, Canada). This entire optical system is capped off with a 100 µm glass window to protect the optics and allow for water/specimen immersion. Simulations indicate that the optical system will give a diffraction-limited focus for both wavelengths. The ray diagram of the system is shown in Fig. 2(a).
Fig. 2 (a). Optical ray diagram of the exoscope featuring the lenses and scanning mechanism. (b). Computer-aided design SolidWorks image of the miniaturized multimodal CARS exoscope model including (1) the barrel of the exoscope (2) the body of the exoscope.
All of these optical components are mounted into a two-piece unit. The first section labeled as (1) in Fig. 2(b) has been characterized in the forward signal collection direction using a benchtop mounted MEMS mirror [29]. This portion is a tapered cylinder where the beam expander, chromatic correction and focusing lenses are contained. It is 41 mm long, 5.5 mm (outer diameter) at the proximal end and 3.0 mm (outer diameter) at the distal end.
The recently attached second piece (body) labeled as (2) in Fig. 2(b) houses the dichroic mirror, MEMS mirror, collimating lens and fibre mounting receptacles. As we are using a mirror to scan the beam, the beam path could not be entirely straight; a folding mirror had to be introduced. This, in conjunction with the need for physical fibre mounts, necessitated a larger housing for the body. The dimensions of this piece are 31 mm by 18.4 mm which also contains the set of four precision alignment screws to aid in the positioning and angle of the MEMS mirror and to ensure efficient light throughput.
In addition to the miniaturized lenses, a dichroic mirror is used to provide a discrete path into the collection fibre for the sample emission wavelengths. A reflection of greater than 95% over a wavelength range of 400-700 nm and transmission of greater than 99% from 800 to 1036 nm is accomplished by a highly customized 105 layer design of alternating Ta2O5 and SiO2, with a total thickness of 88,495 Å (BMV Optical Technologies).
2.3 MEMS mirror
In order to create a minimally invasive, portable scanning CARS exoscope, a laser scanning system that is extremely small, robust, and produces high frame rates is crucial. Without a high frame rate scanner, the ability to produce high quality images from living samples would not be practical due to the continuous displacement of the specimen being imaged. A 2D, gimballed scanning mirror fulfilling these requirements was obtained from the Institut Photonische Mikrosystem (Germany). Both axes of the 500 µm mirror are driven as close as possible to their natural resonance frequency to maximize the mirror’s scan angle and desired field of view. The slow and fast axis electrical driving frequency is ~4.0 KHz and ~32 KHz respectively. The electrical driving square wave has an amplitude of 50 V for the slow axis and 70 V for the fast axis. This corresponds to a sinusoidal mechanical oscillation frequency of ~2.0 kHz for the slow axis and ~16 kHz for the fast axis. When combined these parameters result in a mechanical oscillation of +/−8.5 degrees and a corresponding Lissajous pattern scanning a ~70x70 µm field of view.
2.4 Photonic crystal fibre delivery
An integral element in this nonlinear exoscope is the fibre delivery of single-mode, low noise, femtosecond pump and Stokes beams. Standard fused silica single-mode fibre cannot be used to deliver femtosecond pulses to the sample for CARS imaging due to the inherent high nonlinear signal generation and spectral broadening effects associated with the small mode area of standard fibres. PCFs circumvent these issues for ultrashort pulses [24] as they have customizable dispersive properties, and single mode operation for a broad range of wavelengths. Moreover, the large mode area makes them ideal for ultrashort pulse delivery, as the energy density of the wave is kept as low as possible. Pre-compression is still performed prior to the fibre to compensate for the dispersion that is encountered within the PCF to maintain a high peak power pulse at the sample.
A 1 metre long, large mode area PCF (LMA-20, NKT Photonics) fibre was used for the delivery of the two excitation wavelengths while a multi-mode fibre was used for signal collection. Using independent fibres for the delivery and collection allowed us to isolate and eliminate the four-wave mixing components in the fibre, through the insertion of a dichroic mirror between the delivery, and collection fibres. Using LMA-20 fibre allows single mode operation for a wide spectral band from 600 nm to greater than 1.0µm which includes both pump and Stokes wavelengths.
The low dispersion characteristic of LMA-20 fibre has allowed other groups [24, 28] to obtain CARS signal while using picosecond sources. Without compensation, and an input pulse of ~300 fs, an output pulse width of 1.331 ps was measured. An even stronger CARS signal could be obtained if the peak powers at the sample were maximized by using femtosecond pulses. This was accomplished by using a grating compressor setup for the pump arm at the input to the LMA-20 fibre, delivering a near-transform-limited pulse to the exoscope. Based on the LMA-20 output pulse width, the total dispersion of all the normally dispersive elements in the pump beam path was calculated. A negative 45521 fs2 chirp was introduced to the pump beam, which, after fine tuning of the grating compressor geometry, resulted in a measured ~90 fs pulse width out of the LMA-20. This was accomplished while maintaining a Gaussian beam profile (M2 = 1.34). The pulse width out of the exoscope itself was measured to be ~150 fs. Comparisons of the spectrum and autocorrelation traces of the pump and Stokes before and after the LMA-20 and exoscope did not show any abnormal spectral or temporal anomalies. The LMA-20 did introduce a polarization shift of 11°, although the ellipticity remained relatively unchanged. The difference between the input and the output ellipticity was determined to be less than 0.004. Bending loss measurements were also performed on the LMA-20, and very little additional losses were experienced at radii greater than 15cm. The maximum bending radius used in these experiments was never less than 20cm, maintaining high coupling efficiency.
The final issue with utilizing a large mode area PCF for the propagation of two broadband pulses, is the generation of in-fibre non-resonant FWM at the CARS frequency [24]. This signal was observed exiting the LMA-20, with the spectrum of this signal depicted in Fig. 3. However, this non-resonant signal completely extinguished by the internal dichroic mirror and was not observed coming out of the exoscope, or the separate collection fibre.
Fig. 3 Non-resonant FWM spectrum out of the LMA-20 VS the output spectrum of the exoscope (both with temporal overlap of pump and Stokes). The lack of the FWM signal at ~635 nm demonstrates the effectiveness of the internal dichroic mirror.
3. Results
3.1 Resolution measurements
A United States Air Force 1951 (USAF) (Edmund Optics, USA) target was used to determine the resolving power and to calibrate the exoscope in the forward direction. The negative pre-chirped 800 nm pump beam from the titanium sapphire laser was coupled to the LMA-20 fibre and then to the exoscope. A water drop was placed on the tip of the exoscope, and the glass USAF target was placed with the marking side down (as the microscope is inverted) and making contact with the water. The smallest features of the resolution target were brought into focus for transmission imaging. The average power at the sample was 16 mW, which was attenuated after the target by a 680 nm short-pass filter (ET680SP-2P8, OD5, Chroma Technology), to avoid saturating the PMT. The smallest features (group 7, element 6) are very well resolved at the left side of Fig. 4.
Fig. 4 Transmission image of the USAF target captured in the forward direction: The line spacing of this element is 228 line pairs per mm, which results in a 2.18 µm separation distance between lines and a calibration value of 0.118 µm per pixel.
Following the USAF target calibration, the performance of the exoscope was further examined by determining the axial and lateral resolutions. TPEF sensitive 1.0 µm diameter fluorescent microspheres (Polysciences Inc., USA) were spin coated onto a #1 cover slip. 800 nm light from the femtosecond laser was negatively chirped for dispersion pre-compensation, coupled to the LMA-20 fibre, and delivered to the exoscope, with an average power of 15 mW at the sample. The transmission efficiency through the exoscope was measured as 47%. Figure 5(a) shows the TPEF signal from the microspheres in the epi direction.
Fig. 5 (a) TPEF excitation of 1.0 µm Fluorescebrite microspheres in the epi direction. (b) Lateral intensity profile of a single 1.0 µm Fluorescebrite microsphere. (c) Axial profile the same bead.
For a bead located 14.7 µm from the image centre, the intensity profile was plotted. A Gaussian curve was fit to the profile, shown in Fig. 5(b), and an observed bead size was found to be 0.94 µm full width at half maximum (FWHM), which is an improvement compared to the measured ~1.3 µm demonstrated in the free-space delivered version [29].The exoscope design goal was to achieve a lateral resolution better than 1.0 µm at the sample, and this measurement demonstrates that the exoscope has at least 1.0 µm resolution for single wavelength illumination. 500 nm TPEF sensitive beads were also imaged, but they could not be resolved. Axial resolution measurements were also performed with these 1.0µm TPEF beads, resulting in a value of 24 µm, with the profile demonstrated in Fig. 5(c). This is much larger than the anticipated design value of 3.0 µm and is mainly attributed to the aberration caused inside the miniature objective due to the slight off-axis propagation of the excitation beam as mentioned in Section 2.1.
The final TPEF calibration measurement was undertaken to characterize the inherent chromatic aberrations of the exoscope. Ultra rainbow fluorescent particles (Polysciences Inc.) that are TPEF sensitive for a broad wavelength range were used. The polychromatic sensitivity is due to the affixation of multiple fluorochromes, allowing efficient excitation for both pump and Stokes wavelengths. As there was not enough peak power in the 1040 nm Stokes beam to generate a TPEF signal, the Ti:sapphire laser was tuned to 960 nm which is as close to the Stokes wavelength as could be attained while maintaining mode locking. 1.0mW of 800 nm light and 2.5 mW of 960 nm light were independently used to image these polyfluorescent beads, with the specific bead imaged located 16 µm from the image centre. To determine the lateral chromatic aberrations, the z axis was scanned until the beads were in focus. The centres of the beads were located by analyzing the intensity profile in both the X and Y directions. The lateral offset caused by chromatic aberrations was found to be 0.1 µm in the X direction, and 0.73 µm in the Y direction. This relatively large Y- offset could be a result of the tilt in the optical output of the exoscope. The difference in the z focus was used to discern the strength of the axial chromatic aberrations, which ended up shifting the focus of the 960 nm light with respect to the 800 nm light by 1.3 µm. It is important to note that the axial and lateral resolutions are not constant throughout the field of view (for all resolution measurements). It was determined that the resolutions decreased as a function of distance from the centre of the field of view.
Now beginning with CARS experiments, we used a selection of polystyrene (PS) microspheres with diameters of 20 µm, 4.5 µm, 2.0 µm and 1.0 µm. These beads were spin coated onto #1 coverslips. The aromatic CH bonds in these beads have a strong peak at 3070 cm−1, which is suitable for excitation with the available pump and Stokes wavelengths. The CARS signal of the polystyrene beads was generated by focusing the temporally and spatially overlapped pump (800 nm) and Stokes (1014-1067 nm) beams on the sample. The generated CARS signal was collected by the exoscope and detected in the epi direction. Temporal overlap is achieved by maximizing the CARS signal in the forward direction from a bulk oil sample, while fine-tuning the delay between the pump and the Stokes beams.
CARS lateral and axial resolution measurements were performed in the epi direction, on a 2.0 µm PS bead sample, shown in Fig. 6. The FWHM of the intensity profile of a single 2.0 µm bead located 23.3 µm from the centre of the field of view was determined to be 2.02 µm. The 1.0 µm PS beads were not resolvable. For the axial resolution measurement, a stack of images was collected with 1.0µm steps along the optical axis (z). For the entire data set, a line profile of the bead’s intensity was taken. These values were then plotted against their corresponding z position and a Gaussian curve was fit. The FWHM of this curve which corresponds to the axial resolution of the miniaturized CARS microscope was determined to be 13 µm, similar to the 12.74 µm value reported in the free-space iteration [29]. Although this is much larger than the anticipated and designed value of 3.0 µm, it still allows for some 3D sectioning ability. The aberrations caused by the off-axis propagation of the pump and Stokes excitation beams are mainly responsible for this relatively large axial resolution for CARS. For these PS beads, multiple axial and lateral resolution measurements were performed, to better understand the relationship of the resolution to the radial distance from the centre of the field of view. It was determined that beads located near the edges of the field of view have a decreased lateral and axial resolution by upwards of 0.3 µm, and 3.0 µm respectively.
Fig. 6 (a) CARS-excited 2.0 µm polystyrene microspheres in the epi direction. (b) Lateral profile curve (c) Axial profile curve.
3.2 Structural and biological multimodal imaging
Beginning with the mouse lung tissue, a section was placed between a standard microscope slide, and a #1 coverslip. The slide was placed on the microscope stand, coverslip down, with a water drop bridging the gap between the exoscope tip and the coverslip. The pre-chirped 800 nm beam was delivered to the exoscope via the LMA-20 fibre, and 18 mW of power was scanned on the sample. The generated TPEF signal was collected in the epi direction and the excitation wavelengths were filtered out using a short pass filter (ET680SP-2P8, Chroma Technology). The fluorescein, which had propagated throughout the mouse’s vasculature, can be seen in Fig. 7(a) with the vasculature rich alveolar walls visible surrounding the alveolar cavities.
Fig. 7 (a) Epi - Fluorescein stained mouse lung tissue. (b) Epi - SHG image of KDP crystal. (c) For comparison: White light transmission microscopic image of KDP crystal at 100X magnification.
In order to determine the full multimodality of the exoscope, an SHG signal must also be collected, and imaged. To accomplish this task, a sample of potassium dihydrogen phosphate (KDP) crystal was exposed to 3.5 mW of 800 nm laser light. An additional short pass filter (Z485SP, Chroma Technology) was included to allow only the SHG signal to pass while the generated photons were still being captured in the epi direction. The longitudinal structure of the KDP crystal captured in Fig. 7(b) demonstrates the ability to observe a strong backscattered SHG signal in the epi direction. A transmission image of the same sample is shown for reference in Fig. 7(c).
Finally the exoscope was used to demonstrate its biological epi-CARS imaging capabilities. Unstained, unfixed, fresh mouse sciatic nerves, rich in C-H bonds were mounted on #1 coverslips. The vibrational mode of the C-H bonds at ~2840 cm−1 contained within the myelin and surrounding fat cells is resonantly excited by the frequency difference between the pump and the Stokes wavelengths, with powers of 12.5 mW (~0.156 nJ) and 1.2 mW (~0.015 nJ) respectively. The pump and Stokes wavelengths are extinguished with a 65 nm band pass filter with central peak at 645 nm (ET645/65 M, OD7, Chroma Technology) while allowing the epi-collected photons to reach the PMT. This unlabeled sample is pictured in Fig. 8(a) with the myelin surrounding the central axons clearly visible. To confirm the signal is indeed from the CARS process, the Stokes beam is blocked in Fig. 8(b). The images demonstrated in Figs. 8(c) and Fig. 8(d) are of fat cells surrounding the sciatic nerve, with both pump and Stokes present, and then with the Stokes beam blocked respectively.
Fig. 8 (a) Epi – CARS image from unlabeled mouse sciatic nerve, showing a few axons in the field of view. (b) Epi image of the same region demonstrated in (a), with the Stokes beam blocked. (c) Epi-CARS image of fat cells surrounding the sciatic nerve. (d) Epi image of the same region demonstrated in (c) with the Stokes beam blocked.
4. Discussion and conclusion
We have successfully demonstrated a portable, miniaturized, fibre delivered, multimodal CARS imaging device. The device was made possible by the inclusion of many specialized parts, including a MEMS scanning mirror, with a 500 µm diameter reflective surface being driven at very high frequencies, a custom built dichroic mirror, a large mode area fibre and a custom built miniaturized objective lens. Due to the portability, the applications in which this exoscope can be applied to clinical situations are numerous. The demonstrated TPEF, SHG and CARS capabilities could potentially allow serial monitoring of morphological and chemical changes in vivo, in the same animal. Applications of this in vivo imaging can range from cancer identification, drug therapy selection, detection of spinal cord injuries and demyelination to arterial atherosclerosis, all potentially enabled by the use of this device. Specific medical applications of this exoscope include, but are not limited to detecting squamous cell carcinoma [10], performing nanosurgery [11], and studying myelination [5].
The design, lens manufacture, and assembly of the exoscope provided a significant hurdle. A slight misalignment in one of the nine different glasses would have caused catastrophic aberrations in the exoscope barrel. Spatial overlap of both pump and Stokes beams would have been impossible, rendering the exoscope useless for CARS imaging. Substantial attention, effort, and time ensured that each component was manufactured to specifications, and installed precisely according to the design.
The end result of this effort was an optical system that completely blocked unwanted fibre-generated non-resonant FWM, successfully separated the excitation light from the emitted light while maintaining a 3.0 mm probe diameter at the distal end. Unfortunately, all design specifications were not as successfully met. The chromatic aberrations caused substantial lateral and axial offsets limiting the three-dimensional sectionability, although these aberrations were not severe enough to inhibit the generation of CARS photons. Further improvements to the exoscope manufacture and assembly will most certainly benefit the lateral and axial offsets, and increase system performance.
Together, all of these components resulted in the successful resonant excitation of CH bonds at 2845 cm−1 and 3070 cm−1 Raman shifts, generating a strong CARS signal, whether it is from oil, beads or biological tissue. While the fundamentals of a multimodal nonlinear exoscope have been successfully verified, there were some issues in this first prototype that caused the quality of the images demonstrated to be less than ideal. The optical path of the excitation beams out of the exoscope’s barrel was off axis, with an observable tilt of 8°. This slight tilt in the optical path was impossible to correct for, even during the initial integration of the body and barrel. It is expected that better control of the longitudinal and lateral position of the input fibre tip with respect to the collimating lens, as well as finer control of the angular placement of the MEMS mirror, will enable elimination of this off-axis tilt. It is most likely that this tilt is responsible for the majority of the aberrations and for the resulting degradation of the axial resolution measurements. The lack of available Stokes beam power also negatively contributed to the quality of the images reported. A system with ample Stokes power would have a more intense CARS signal, and consequently, images with a higher signal to noise ratio.
The lateral resolution measurements taken confirm that the focus is near diffraction limited for TPEF measurements, as designed. The lateral resolution measurements are larger for CARS, and this can be attributed to the chromatic aberrations, reducing the overlap in the X-Y plane. The discrepancy between the chromatic aberration in the X and Y directions can be attributed to the off axis propagation of the beams due to the optical tilt that was reported. The axial resolution measurements, however, were larger than anticipated. This could be due to a number of reasons, most likely caused by various types of aberrations experienced within the miniaturized optics. The decrease seen in the axial resolution for the two wavelength CARS process is due to the mismatch in the chromatic overlap of the two beams. The larger axial resolution of the TPEF signal can be attributed to the previously mentioned aberrations caused by the off-axis propagation of the pump beam, extending the axial focal volume.
Further improvements on the manufacturing processes of exoscope will reduce the optical aberrations of the system in the next prototype. This will increase pump and Stokes beam overlap, reduce the elongated active focal volume, increasing the image quality and resolution capabilities. Signal collection will be improved by using a more sensitive PMT, as well as locating more suitable samples. As the goal of this project was to image lipids using CARS imaging, a broadband pump and Stokes was suitable to excite the strong vibration resonances of the CH bonds in the lipid molecules as demonstrated here. A technique such as spectral focusing based CARS microscopy using a single femtosecond laser and PCF-generated supercontinuum [31], will enable extension of the application range of the exoscope to the Raman fingerprint region of ~800 – 1800 cm−1 [32]. Integration with a femtosecond fibre laser [33] is also being considered to reduce the footprint of the entire setup, bringing multimodal nonlinear imaging closer to a clinical application.
Acknowledgments
This work was made possible by the support of CIHR – NSERC grant 87490. We would like to thank Dr. Tito Scaiano and his students Chiara Fasciani and Madeline Simpson, as well as Dr. M. Parsanejad for the preparation of various biological samples and Dr. Ileana Micu and Thomas Kannanayakal for sample preparation consultation.
References and links
1. C. L. Evans and X. S. Xie, “Coherent anti-Stokes Raman scattering microscopy: chemical imaging for biology and medicine,” Annu Rev Anal Chem (Palo Alto Calif) 1(1), 883–909 (2008). [CrossRef] [PubMed]
2. T. T. Le, T. B. Huff, and J. X. Cheng, “Coherent anti-Stokes Raman scattering imaging of lipids in cancer metastasis,” BMC Cancer 9:42 (2009). BMC Cancer 9(1), 42 (2009), doi:. [CrossRef]
3. E. Kang, H. Wang, I. K. Kwon, J. Robinson, K. Park, and J. X. Cheng, “In situ visualization of paclitaxel distribution and release by coherent anti-Stokes Raman scattering microscopy,” Anal. Chem. 78(23), 8036–8043 (2006). [CrossRef] [PubMed]
4. X. Nan, J. X. Cheng, and X. S. Xie, “Vibrational imaging of lipid droplets in live fibroblast cells with coherent anti-Stokes Raman scattering microscopy,” J. Lipid Res. 44(11), 2202–2208 (2003). [CrossRef] [PubMed]
5. Y. Fu, T. B. Huff, H. W. Wang, H. Wang, and J. X. Cheng, “Ex vivo and in vivo imaging of myelin fibers in mouse brain by coherent anti-Stokes Raman scattering microscopy,” Opt. Express 16(24), 19396–19409 (2008). [CrossRef] [PubMed]
6. Y. Fu, H. Wang, T. B. Huff, R. Shi, and J. X. Cheng, “Coherent anti-Stokes Raman scattering imaging of myelin degradation reveals a calcium-dependent pathway in lyso-PtdCho-induced demyelination,” J. Neurosci. Res. 85(13), 2870–2881 (2007). [CrossRef] [PubMed]
7. H. A. Rinia, M. Bonn, E. M. Vartiainen, C. B. Schaffer, and M. Müller, “Spectroscopic analysis of the oxygenation state of hemoglobin using coherent anti-Stokes Raman scattering,” J. Biomed. Opt. 11(5), 050502-1– 050502-3 (2006).
8. +T. T. Le, I. M. Langohr, M. J. Locker, M. Sturek, and J. X. Cheng, “Label-free molecular imaging of atherosclerotic lesions using multimodal nonlinear optical microscopy,” J. of Biomedical Opt. 12 (5), 054007–1–054007–10 (2007).
9. Y. M. Wu, H. C. Chen, W. T. Chang, J. W. Jhan, H. L. Lin, and I. Liau, “Quantitative assessment of hepatic fat of intact liver tissues with coherent anti-Stokes Raman scattering microscopy,” Anal. Chem. 81(4), 1496–1504 (2009). [CrossRef] [PubMed]
10. H. G. Breunig, M. Weinigel, R. Bückle, M. Kellner-Höfer, J. Lademann, M. E. Darvin, W. Sterry, and K. König, “Clinical coherent anti-Stokes Raman scattering and multiphoton tomography of human skin with a femtosecond laser and photonic crystal fiber,” Laser Phys. Lett. 10(2), 1–5 (2013). [CrossRef]
11. A. Uchugonova, M. Lessel, S. Nietzsche, C. Zeitz, K. Jacobs, C. Lemke, and K. König, “Nanosurgery of cells and chromosomes using near-infrared twelve-femtosecond laser pulses,” J. Biomed. Opt. 17(10), 101502 (2012). [CrossRef] [PubMed]
12. H. Chen, H. Wang, M. N. Slipchenko, Y. Jung, Y. Shi, J. Zhu, K. K. Buhman, and J. X. Cheng, “A multimodal platform for nonlinear optical microscopy and microspectroscopy,” Opt. Express 17(3), 1282–1290 (2009). [CrossRef] [PubMed]
13. F. Helmchen, M. S. Fee, D. W. Tank, and W. Denk, “A miniature head-mounted two-photon microscope. high-resolution brain imaging in freely moving animals,” Neuron 31(6), 903–912 (2001). [CrossRef] [PubMed]
14. H. Bao, J. Allen, R. Pattie, R. Vance, and M. Gu, “Fast handheld two-photon fluorescence microendoscope with a 475 microm x 475 microm field of view for in vivo imaging,” Opt. Lett. 33(12), 1333–1335 (2008). [CrossRef] [PubMed]
15. C. J. Engelbrecht, R. S. Johnston, E. J. Seibel, and F. Helmchen, “Ultra-compact fiber-optic two-photon microscope for functional fluorescence imaging in vivo,” Opt. Express 16(8), 5556–5564 (2008). [CrossRef] [PubMed]
16. W. Piyawattanametha, E. D. Cocker, L. D. Burns, R. P. Barretto, J. C. Jung, H. Ra, O. Solgaard, and M. J. Schnitzer, “In vivo brain imaging using a portable 2.9 g two-photon microscope based on a microelectromechanical systems scanning mirror,” Opt. Lett. 34(15), 2309–2311 (2009). [CrossRef] [PubMed]
17. J. T. C. Liu, M. J. Mandella, N. O. Loewke, H. Haeberle, H. Ra, W. Piyawattanametha, O. Solgaard, G. S. Kino, and C. H. Contag, “Micromirror-scanned dual-axis confocal microscope utilizing a gradient-index relay lens for image guidance during brain surgery,” J. Biomed. Opt. 15(2), 026029 (2010). [CrossRef] [PubMed]
18. H. Bao, A. Boussioutas, R. Jeremy, S. Russell, and M. Gu, “Second harmonic generation imaging via nonlinear endomicroscopy,” Opt. Express 18(2), 1255–1260 (2010). [CrossRef] [PubMed]
19. F. Légaré, C. L. Evans, F. Ganikhanov, and X. S. Xie, “Towards CARS Endoscopy,” Opt. Express 14(10), 4427–4432 (2006). [CrossRef] [PubMed]
20. F. Knorr, D. R. Yankelevich, J. Liu, S. Wachsmann-Hogiu, and L. Marcu, “Two-photon excited fluorescence lifetime measurements through a double-clad photonic crystal fiber for tissue micro-endoscopy,” J Biophotonics 5(1), 14–19 (2012). [CrossRef] [PubMed]
21. D. R. Rivera, C. M. Brown, D. G. Ouzounov, I. Pavlova, D. Kobat, W. W. Webb, and C. Xu, “Compact and flexible raster scanning multiphoton endoscope capable of imaging unstained tissue,” Proc. Natl. Acad. Sci. U.S.A. 108(43), 17598–17603 (2011). [CrossRef] [PubMed]
22. M. Weinigel, H. G. Breunig, P. Fischer, M. Kellner-Höfer, R. Bückle, and K. König, “Compact clinical high-NA multiphoton endoscopy,” Proc. SPIE 8217, 821706, 821706-8 (2012), doi:. [CrossRef]
23. H. Wang, T. B. Huff, and J. X. Cheng, “Coherent anti-Stokes Raman scattering imaging with a laser source delivered by a photonic crystal fiber,” Opt. Lett. 31(10), 1417–1419 (2006). [CrossRef] [PubMed]
24. M. Balu, G. Liu, Z. Chen, B. J. Tromberg, and E. O. Potma, “Fiber delivered probe for efficient CARS imaging of tissues,” Opt. Express 18(3), 2380–2388 (2010). [CrossRef] [PubMed]
25. P. S. Hsu, A. K. Patnaik, J. R. Gord, T. R. Meyer, W. D. Kulatilaka, and S. Roy, “Investigation of optical fibers for coherent anti-Stokes Raman scattering (CARS) spectroscopy in reacting flows,” Exp. Fluids 49(4), 969–984 (2010). [CrossRef]
26. P. S. Hsu, W. D. Kulatilaka, J. R. Gord, and S. Roy, “Single-shot thermometry using fiber-based picosecond coherent anti-Stokes Raman scattering (CARS) spectroscopy,” J. Raman Spectro . (2013). [CrossRef]
27. Z. Wang, L. Gao, P. Luo, Y. Yang, A. A. Hammoudi, K. K. Wong, and S. T. C. Wong, “Use of multimode optical fibers for fiber-based coherent anti-Stokes Raman scattering microendoscopy imaging,” Opt. Express 19(9), 7960–7970 (2011). [CrossRef] [PubMed]
28. B. G. Saar, R. S. Johnston, C. W. Freudiger, X. S. Xie, and E. J. Seibel, “Coherent Raman scanning fiber endoscopy,” Opt. Lett. 36(13), 2396–2398 (2011). [CrossRef] [PubMed]
29. S. Murugkar, B. Smith, P. Srivastava, A. Moica, M. Naji, C. Brideau, P. K. Stys, and H. Anis, “Miniaturized multimodal CARS microscope based on MEMS scanning and a single laser source,” Opt. Express 18(23), 23796–23804 (2010). [CrossRef] [PubMed]
30. 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]
31. 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]
32. A. F. Pegoraro, A. D. Slepkov, A. Risdale, D. J. Moffatt, and A. Stolow, “Hyperspectral multimodal CARS microscopy in the fingerprint region,” J. Biophot. (2012). [CrossRef]
33. K. J. Bock, H. E. Kotb, M. A. Abdelaim, and H. Anis, “Increasing energy in an ytterbium femtosecond fiber laser with a longer gain medium and lower doping,” Proc. SPIE 8237, 823731-1–823731-7 (2012), doi:. [CrossRef]
