Open Access
Add to BibSonomy
Abstract
By precisely managing fiber-optic nonlinearity with anomalous dispersion, we have demonstrated the control of generating plural few-optical-cycle pulses based on a 24-MHz Chromium:forsterite laser, allowing multicolor two-photon tissue imaging by wavelength mixing. The formation of high-order soliton and its efficient coupling to dispersive wave generation leads to phase-matched spectral broadening, and we have obtained a broadband continuum ranging from 830 nm to 1200 nm, delivering 5-nJ pulses with a pulse width of 10.5 fs using a piece of large-mode-area fiber. We locate the spectral enhancement at around 920 nm for the two-photon excitation of green fluorophores, and we can easily compress the resulting pulse close to its limited duration without the need for active pulse shaping. To optimize the wavelength mixing for sum-frequency excitation, we have realized the management of the power ratio and group delay between the soliton and dispersive wave by varying the initial pulse energy without additional delay control. We have thus demonstrated simultaneous three-color two-photon tissue imaging with contrast management between different signals. Our source optimization leads to efficient two-photon excitation reaching a 500-µm imaging depth under a low 14-mW illumination power. We believe our source development leads to an efficient and compact approach for driving multicolor two-photon fluorescence microscopy and other ultrafast investigations, such as strong-field-driven applications.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The development of fluorescent tags allows tracking molecules and biological signals with a high spatial and temporal precision [1], and multicolor microscopy has thus become a powerful tool in biomedical research. For example, simultaneous imaging of blood vessels and blood cells in mice has contributed to the reveal of cancer cell dynamics [2], and imaging circulatory systems in mouse kidneys with multicolor labeling also leads to the advanced investigation of the immune systems [3]. Moreover, visualizing metabolic processes with different dyes paves the way for drug delivery studies [4].
The capability of deep-tissue imaging is also critical in translating the multicolor approaches to intravital imaging. Two-photon fluorescence microscopy (TPFM) is of particular interest as that allows for achieving three-dimensional subcellular imaging with sub-millimeter penetration depth [5]. Realizing multicolor TPFM requires exciting plural fluorophores via two-photon processes while avoiding photodamages [6–9]. Several multi-photon excitation strategies have been investigated, such as single-wavelength or broadband excitation of multiple chromophores [10,11] or sequential illumination at three wavelengths [6–9,12,13]. Optical parametric oscillators (OPOs) pumped by a Ti:sapphire laser has been widely employed for covering the two-photon excitation range of most fluorescent proteins (FPs). Furthermore, the spatiotemporal overlap of the outputs from a Ti:sapphire laser and an OPO provides a sum-frequency excitation route called wavelength mixing. The wavelength mixing approach allows efficient and independent control over the strength of the different signals [6–9] by tuning the temporal overlap. Nevertheless, the Ti:sapphire-based laser system is costly to implement, making this approach hardly available for widespread use. An alternative source pumped by a compact fiber laser has been proposed for multicolor TPFM, enabling the excitation of green to red dyes using 1050 and 1250-nm spectrum [12]; however, an optical delay line is still required to overlap the excitation pulses and manage sum-frequency excitation.
Instead of using OPOs, femtosecond sources enabled by fiber-optic nonlinearity have attracted much attention as cost-effective tools for multicolor TPFM [10,11,13–15]. Supercontinuum generation (SCG) ranging from 750 nm to 1070 nm enables simultaneous excitation for 4-color Brainbow imaging [10]. However, despite the broad spectral coverage, SCG may involve complicated nonlinear interactions, exhibiting a complex spectral phase profile with severe pulse-to-pulse fluctuation, and the resulting pulse is hardly compressible for efficient two-photon excitation. Although active programable pulse shapers, such as spatial light modulators, can be used to modify the spectral phases of the broadband spectra [10,11], their use increases the system complexity and power loss.
Optimizing the spectral conversion while avoiding unwanted nonlinear interactions is thus of great importance in developing fiber-optic laser tools for ultrafast applications. Several mechanisms based on fiber-optic nonlinearity have been applied to generate compressible femtosecond pulses, including self-phase modulation (SPM) dominated spectral broadening [16–22], soliton self-frequency shift (SSFS) [13,23–26], and dispersive wave (DW) generation [27–30]. The employment of each nonlinear process alone leads to the generation of transform-limited femtosecond pulses without the need for active pulse shaping [28–30]. SPM-dominated spectral broadening features an oscillating spectral profile, and the shifted spectral peak and power can be easily controlled by managing the initial pulse’s power and width [16–22,31]. The nonlinear pulse shaping accompanied by Raman effects with anomalous dispersion causes SSFS, and it becomes a powerful mechanism to generate energetic femtosecond pulses above the wavelength of 1550 nm for deep-tissue three-photon imaging [23]. DW generation allows spectral enhancement around a specific wavelength, resulting from fiber-optic soliton perturbation [27,32]. Simultaneously managing any two of the above nonlinear processes may lead to generating plural compressible pulses with tunable delay. In this work, we aim to realize wavelength-mixing-enabled multicolor TPFM by controlling the generation of SPM-dominated spectral broadening and DW.
We experimentally demonstrate our source development based on a 24-MHz home-built Cr:forsterite (Cr:F) laser and a piece of large mode area (LMA) photonic crystal fiber (PCF). We have obtained a broadband continuum ranging 830-1200 nm with a high pulse energy of 5 nJ, and the resulting pulse can be easily compressed close to its transform-limited (TL) duration using double-chirped mirrors (DCMs) with negligible loss. We also simultaneously generate two close-to-TL pulses by SPM broadening and DW generation, and we are capable of controlling the delay and power ratio between the two pulses by varying the fiber input without an additional optical delay line. Thus, we have obtained three-color images from a mouse kidney slide by wavelength mixing, reaching a deep imaging depth of 500 µm under low 0.6-nJ pulse energy (i.e., 14-mW average power) on the sample surface. Our results indicate the potential of the demonstrated source system as a versatile and compact solution for driving multicolor TPFM.
2. Phase-matched continuum generation
We employ a home-built mode-locked Cr:F laser to drive nonlinear fiber-optic processes, and we provide a schematic of our experimental setup as shown in Fig. 1(a). The center wavelength of a Cr:F laser is close to the zero-dispersion wavelength (ZDW) of fused silica, and thus the control of the nonlinear pulse propagation in different dispersion regimes does not require strong waveguide dispersion at the expense of a small mode-field area. The Cr:F laser configuration is similar to previous work [33–35]. We implement a Yb:fiber laser (YLM-20-LP, IPG) as the pumping source and employ a Z-fold linear cavity for astigmatism compensation. We use a lens with a focal length of 100 mm to focus the pump into a 13-mm-length Cr:F crystal. A thermoelectric (TE) cooler is used to stabilize the temperature of the laser crystal at 278K. We drive the laser under a pump power of 13 W with a 10% output coupler (OC). To obtain stable mode-locking, we use a pair of SF14 prisms with a 27-cm tip-to-tip distance for intracavity dispersion compensation, and the total cavity length is 6.22 meters, leading to a 24.1-MHz repetition rate. We possibly miniaturize the size of the oscillator, and the geometric size of the home-built Cr:F oscillator is 110 cm in length and 45 cm in width. We have characterized the laser spectrum via a spectrometer (Sidewinder, OTO Photonics) with a full width at half maximum (FWHM) of 48 nm at a center wavelength of 1260 nm with a 50-fs pulse width, shown as the yellow line in Fig. 1(c), and the output pulse energy is 20 nJ.
Fig. 1. (a) Schematic of phase-matched continuum generation for a multicolor two-photon imaging system. (b) Dispersion profile of PCF LMA-10 with the ZDW at 1150 nm. The yellow line indicates the center wavelength of a home-built Cr:F laser, and the laser output experiences anomalous dispersion in the PCF. (c) The measured output spectra from Cr:F laser (yellow) and LMA PCF (blue) in response to the calculated coherent length of the DW (dotted orange), estimated by using a 50-fs Gaussian initial pulse with a center wavelength at 1260 nm and 11-nJ coupled pulse energy. (d) The measured and reconstructed SHG FROG traces of the pulse with the blue-shifted spectrum (i.e., the fiber output covered in the gray area in (c)). (e) The retrieved temporal intensity (blue) and phase (orange), as well as the TL pulse profile (green dotted). HWP: Half-wave plate; PBS: Polarizing beam splitter; ISO: Isolator; SPF: Shortpass filter; DCM: Double-chirped mirror; ND filter: Neutral density filter; TL: Transform-limited.
We couple the laser pulses into a piece of LMA PCF for nonlinear spectral conversion. We control the input pulse energy with a polarizing beam splitter (PBS) and a half-wave plate (HWP). The size of our current fiber module, including the coupling lens, the fiber, and the collimating lens, is within 10 × 10 × 10 cm3, and we are working on designing a miniaturized module with the size less than 5 × 5 × 5 cm3. We place an isolator (ISO) before the PCF to prevent the backward coupling of the stray light into the laser cavity. After the PCF, we use a short-pass filter (SPF) to isolate the spectrum below 1200 nm for driving TPFM. To obtain compressed pulses with negligible loss, we employ a DCM pair (DCM7, Laser Quantum) for dispersion pre-compensation, which features over 99.5% reflectivity of a single bounce, and we may fine adjust the pulse chirp by a pair of fused-silica Brewster wedges. A neutral density (ND) filter is also used to control the illumination power on the sample.
We explore nonlinear fiber-optic processes for generating compressible femtosecond pulses covering multiple excitation bands. To manage the nonlinear spectral evolution inside the fiber, we have to control the propagation length precisely: SPM-dominated broadening exhibits strong spectral oscillations, and using a shorter fiber length is beneficial for suppressing unwanted nonlinear interactions [36]. Stimulated Raman scattering causes SSFS, and the soliton redshift requires a long-enough fiber length to realize sequential Raman amplification [37]. Efficient DW generation is a phase-matching (PM) process; we will explain the PM condition later, and using a short fiber length leads to broadband DW generation [38]. As a result, by using a short fiber length, we may simultaneously shape SPM-dominated spectral broadening and broadband DW generation, obtaining two pulses at different wavelength ranges.
To investigate the spectral coverage for plural-band excitation, we start by estimating the PM condition for efficient DW generation. Soliton formation occurs when ultrafast pulses experience negative group delay dispersion in the fiber, resulting in the balance between the negative dispersion and SPM. However, the perturbation to the solitary pulse during the propagation, mainly from the HOD of the fiber, sheds some of the soliton energy into DW generation, and the DW experiences nearly no nonlinear phase during the propagation. As a result, efficient DW occurs when the linearly propagating DW constructively interferes with the solitary wave. That is, the linear phase accumulated from HOD, ${\varphi _{HOD}}$, equals to the nonlinear soliton phase, ${\varphi _{soliton}}$. We may express the phase difference between the two pulses as the function of input nonlinearity, dispersion profile (i.e., propagation constant $\beta (\omega )$), and the fiber length $ L$, to obtain the following PM condition [27]:
As a result, we may obtain efficient DW generation at a chosen wavelength by matching the laser parameters with the fiber dispersion profile. We thus employ a piece of LMA PCF to enhance the DW at around 920nm for GFP excitation. Figure 1(b) depicts the dispersion profile of the LMA PCF (PCF LMA-10, NKT photonics) with the ZDW at 1150 nm, and we also indicate the center wavelength of the Cr:F laser as the orange line. The calculated coherent length at different wavelengths is shown as the dotted orange line in Fig. 1(c). The phase-matched wavelength, where the coherent length becomes infinity, is around 940 nm.
We firstly couple the 20-nJ laser pulse into the LMA PCF with a 55% coupling efficiency, and the corresponding soliton order of the initial pulse is around 6. According to the coherent length estimation, we use a fiber length of 30 mm to obtain a broadband continuum while realizing PM enhancement. We show the experimental spectrum as the blue line in Fig. 1(c), and the spectral enhancement at around 930 nm agrees well with the calculation according to Eq. (1). The blue-shifted broadband continuum, as covered in the gray area in Fig. 1(c), ranges from 830 nm to 1200 nm, and it may be suitable for two-photon green to red fluorescence excitation.
We characterize the temporal characteristic of the spectrum via frequency-resolved optical gating (FROG). We employ the same second harmonic generation (SHG) FROG setup and retrieval algorithm as a previous study [22], allowing broadband measurements from 700 nm to 1200 nm. After 1-mm-long fused silica and two bounces from the surface of the DCMs, we compress the blue-shifted spectrum below the wavelength of 1200 nm in the entire grey-shaded spectral region in Fig. 1(c), and the corresponding FROG traces and the retrieved pulse information are shown in Fig. 1(d) and 1(e), respectively. The compressed pulse duration is 10.5 fs, close to its TL duration of 9.8 fs. We thus obtain 2.9-optical-cycle pulses centered at 1030 nm. It is worth noting that using Er:fiber laser [28,29] and Yb:fiber laser [30] as femtosecond pumping sources also led to similar broadband spectra ranging 900-1400 and 750-950 nm, supporting TL durations of 8 fs and 14 fs, respectively. However, the need for highly nonlinear fibers in these cases resulted in limited output pulse energies within the sub-nJ level, which may not be enough for nonlinear microscopy applications. In contrast, using LMA fibers allows the delivery of higher-energy pulses (i.e., 5 nJ in our case).
3. Control of optical delay between SPM and DW for wavelength mixing-enabled multicolor TPFM
To better understand the fiber-optic soliton dynamics, we simulate the pulse propagation by solving the generalized nonlinear Schrödinger (GNLS) equation based on the split-step method [39], using 50-fs initial pulses centered at 1260 nm, corresponding to the frequency of 238 THz. To show the spectrotemporal behavior under different initial pulse energy, we plot the output spectrograms, as shown in Fig. 2, using a Gaussian gate function with a temporal FWHM of 25 fs [40].
Fig. 2. Simulated spectrotemporal evolution of the output pulses with different coupled pulse energy in a LMA PCF. (a-f) Spectrograms corresponding to coupled pulse energy from 8 to 13 nJ. The peak of the non-phase-matched pulse is around 270 THz (i.e., 1110 nm in wavelength), and the DW one is around 320-330 THz, which is determined by the PM condition. (g) Animation of the energy-dependent spectrogram only considering group delay dispersion and SPM (see Visualization 1).
As the initial pulse energy increases, we may obtain higher-order soliton compression with significant spectral broadening [41]. When the spectrum reaches the matched wavelength around 940 nm (i.e., around 320 THz in frequency, as shown in Fig. 2(b) and 2(c)), PM leads to an effective energy transfer from the solitary wave to the DW. The more initial pulse energy results in more energy transferred to the DW with a slightly larger frequency detuning. On the other hand, the phase mismatch $\Delta \varphi $ also increases at the neighboring wavelengths due to the larger nonlinear phase. As a result, with even more initial pulse energy (i.e., above 11 nJ, as shown in Fig. 2), the phase mismatch leads to destructive interference and spectral isolation between the spectrum centered at around 270 THz (i.e., 1110 nm in wavelength) and the DW at 330 THz. In contrast, Fig. 2(g) shows an animation (see Visualization 1) of the energy-dependent spectrogram only considering group delay dispersion and SPM, and we may observe the high-order soliton evolution without DW radiation above the frequency of 320 THz.
Moreover, the DW and solitary pulse travel along the fiber with different propagation speeds. Therefore, we can precisely control the optical delay and the power ratio between the two pulses by simply tuning the initial pulse energy. We want to note that we may filter the redshift in the 200-240 THz range, as shown in Fig. 2, as the third pulse. The spectrotemporal behavior of the third pulse should be similar to previous works [16–22]. One may thus realize precise spectrotemporal control of the three pulses without the need for active pulse shaping, which may open more opportunities for multicolor applications.
To confirm the simulation result experimentally, we couple the femtosecond pulses from the Cr:F laser into a 33-mm-long LMA PCF. We experimentally characterize the output pulses under different initial pulse energy, as shown in Figs. 3(a) to 3(d): We measure the output FROG traces and retrieve the spectral intensity and phase ranging from 830 nm to 1200 nm. The retrieved spectra (dotted) are almost identical to the measured ones (solid), implying the reliability of our phase retrieval and pulse characterization.
Fig. 3. (a-d) Measured and retrieved spectral information after a 1200-nm short-pass filter in response to different launched pulse energy into the PCF. The solid black lines, dotted black lines, and solid green lines indicate measured spectra, retrieved spectra, and retrieved spectral phase, respectively. The labeled blue and red areas show the spectral regions of non-phase-matched broadening (non-PM) and DW generation, respectively. (e-h) Retrieved pulses (solid), as well as their TL pulses (dotted), corresponding to the spectra covered in the blue and pink regions in (a-d). We also denote the TL durations and the relative delays between the peaks of the two pulses.
The spectral broadening is enhanced via PM at around 920 nm, shown as the spectra covered in the pink region in Figs. 3(a) to 3(d), and we also specify the spectral region without PM enhancement in the blue. The spectral phases in the different regions feature linear trends with distinct slopes, as shown in Figs. 3(a) to 3(d). We also show the corresponding temporal characteristics in Figs. 3(e) to 3(h). We have obtained few-optical-cycle pulses (i.e., 15-20 fs) supported by the spectra in both 830-980 nm and 980-1200 nm ranges. The pulse durations (solid) are close to their transform limits (dotted), which are also implied by the linear spectral phase trends shown in Figs. 3(a) to 3(d). In addition, increasing the initial pulse energy leads to the growth of the DW with a considerable delay. Varying the launched pulse energy from 8.9 to 11 nJ allows the control of the delay between the non-phase-matched (non-PM) pulse and DW by 45 fs. In Fig. 3, the DW is always behind the non-PM pulses: Due to the phase-matching condition, the DW pulse is only generated in the normal dispersion regime, and the DW generation in the shorter wavelength range makes the soliton pulse lead in time. On the other hand, the DW pulse will lead in time when its wavelength is longer than the soliton pulse. If we modify the dispersion behavior of the nonlinear medium (e.g., using highly nonlinear PCF or custom-designed waveguide), we are able to generate the DW pulse in the longer wavelength range. In fact, such DW generation has been realized in the previous works [42,43].
Therefore, we can temporally scan the DW over the non-PM pulse by quickly adjusting the initial power, and the pulse durations are still close to their transform limit during the adjustment. These features show great potential for demonstrating multicolor TPFM with wavelength mixing. We use a customized scanning microscope (JadeBio-MP, Southport) and a 60X 0.9 NA water immersion objective lens (LUMPlanFl 60X/0.90 W, Olympus). We separate the excitation and emission light with a 735-nm long-pass dichroic mirror (735 nm edge Brightline, Semrock), and the signals are further purified by a 750-nm SPF (FESH750, Thorlabs). We detect the fluorescence with photomultiplier tubes (H7422, Hamamatsu). The image size is 200 × 200 µm2 with 1024 × 1024 pixels, and the frame rate is around 15 Hz. We demonstrate our system using commercially available fluorescent tissues of a mouse kidney (FluoTissue mouse kidney section, SunJin Lab).
Our sample is labeled with AF488 for blood vessels, SYTOX Orange for the nucleus, and AF647 for neurons. We can excite AF647 using the pulse with the spectra in the non-PM region and AF488 using the DW alone; the spatiotemporal overlap of the two driving pulses results in wavelength mixing for SYTOX Orange excitation. We show the image results in Fig. 4: The blue, green, and red pseudo colors in the tissue images represent the blood vessels, nucleus, and neurons contrasts, respectively, and we isolate the signals using corresponding bandpass filters (525/39 nm BrightLine, 607/36 nm BrightLine, and 660/52 nm BrightLine, Semrock, respectively). The tissue images in Fig. 4 show clear contrasts of renal structures: Blood vessels are aggregated into spheres forming the glomerulus, which is used for filtering the waste from blood to the urine; the nucleus of different cells are labeled and uniformly dispersed in the tissues; the neurons and the axons are randomly distributed, and the amounts of neural cells are typically less than the amount in other tissues, leading to the insensitivity of renal systems to external stimulation. Imaging kidney tissues allows for investigating immune cell dynamics in renal systems [3] and metabolic phenomena in response to drug delivery or disease [44].
Fig. 4. (a-f) Multicolor two-photon tissue images of a mouse kidney sample, excited by the blue-shifted spectra with different pulse energy coupled into the PCF. The blue, green, and red pseudo colors represent the blood vessels labeled with AF488, the nucleus labeled with SYTOX Orange, and the neurons labeled with AF647, respectively. (g) The ratios of the signal strengths from the three fluorescent dyes to the AF647 one under different coupled pulse energy. (h) Z-stack projection of a mouse kidney sample with a 500-µm imaging depth, using a low 14 mW illumination power on the sample surface. The size of the scale bar is 50 µm.
We show the two-photon images excited by the fiber sources with initial power management in Figs. 4(a) to 4(f). The pulse energies of the two spectra from the non-PM and DW pulses are listed in Table 1. The inter-pulse delay and the DW pulse energy grow as the input pulse energy increases, while the pulse energy of the non-PM part is barely changed. However, we want to note that only 7% of the input power change (i.e., 9.25-9.95 nJ of the coupled pulse energy) is required to scan through the temporal overlapping for efficient sum-frequency excitation, and the corresponding change of the DW pulse energy is around 23%. According to Table 1, the total energy of the two pulses is directly linked with the input pulse energy, and the total energy is thus coupled with the delay tuning. However, we may also control the input peak power by varying the input pulse width (e.g., varying the initial chirp via prism pairs) instead of adjusting the input pulse energy, which potentially leads to smaller changes in the total energy during the delay control. Although the inter-pulse delay is always correlated to the DW pulse energy, the change in the DW pulse energy can be reduced under the condition: If the DW pulse is located in the spectral regime with a larger group delay dispersion, the inter-pulse delay can be principally varied more rapidly when tuning the input peak power. In this case, the change of the DW pulse energy may be acceptable over the required delay tuning range. To show the relative intensity between different signals, we equalize the brightness of the AF647 signals in different images, and we fix the ratios of the signal strengths from the other two contrasts to the AF647 one. The signal ratios under different initial pulse energy are shown in Fig. 4(g) to demonstrate the feasibility of controlling sum-frequency excitation. The DW becomes weaker with lower initial pulse energy, diminishing the contrast from the green dye, AF488. When the initial pulse energy reaches around 9.6 nJ, we obtain maximized contrast from SYTOX Orange, indicating an optimized wavelength mixing. The further increase of the initial pulse energy results in a larger delay between the non-PM pulse and DW. Although the DW becomes more pronounced, the lower temporal overlapping of the two pulses leads to the less efficient sum-frequency excitation. As a result, we may individually enhance the contrast of different fluorescent signals, and we can control the temporal overlap of the excitation pulses for efficient sum-frequency excitation by varying the input pulse energy before the nonlinear fiber-optic propagation. Besides, AF488, SYTOX Orange, and AF647 share similar excitation conditions with GFP, tdTomato, and miRFP, respectively, implying the potential of our approach for widespread use. Moreover, we can drive multicolor TPFM using other fluorophores by matching the excitation wavelengths with the spectral peaks of the phase-matched continuum. We can determine the locations of the spectral peaks by adequately selecting the fiber type and the femtosecond driving source.
Table 1. Pulse energy of non-phase-matched (non-PM) and DW pulses and the energy ratio used for two-photon imaging in response to different coupled pulse energy
Additionally, using few-optical-cycle pulses leads to efficient two-photon excitation for deep-tissue imaging, as shown in Fig. 4(h). With low 0.6-nJ pulse energy (i.e., 14 mW) on the sample surface, we have achieved up-to-500-µm penetration depth, currently limited by the sample thickness, and the signal contrasts remain clear at the bottom. The illuminated pulse energies of the DW pulse and the non-PM pulse are as low as 0.16-nJ and 0.44-nJ, respectively. In comparison, the typical average power for the two-photon excitation applied in previous work is varied from 5 to 20 mW [45]. For deep-tissue imaging with the penetration depth more than 500 µm, the two-photon excitation would require at least 100-mW average power [46,47]. In Fig. 4(h), the distribution of neural tissues and blood vessels varies from different depths, which is related to the physiological structures in kidneys. According to our observation, the neural tissue is less distributed around the glomerulus, as shown in Fig. 4(h), with fewer neurons and axons around the depth of 270 µm. However, the density of neurons and axons is still not that sparse around the non-aggregated blood vessels below the depth of 300 µm in Fig. 4(h). We have thus shown the capability of multicolor tissue imaging with efficient and manageable excitation, indicating the strong potential of the resulting sources for biomedical TPFM applications.
The use of shorter pulses often contributes to more efficient two-photon excitation, but the efficiency would decrease if the excitation spectrum is wider than the absorption cross-sections of the used fluorophores. Therefore, in Fig. 5(a), we plot the absorption cross-section of commonly used green to red fluorophores, in comparison to the spectrum shown in Fig. 3(b). A broadband DW pulse may lead to more efficient two-photon excitation of GFP and GCaMP, which exhibit broader absorption cross-sections than the DW spectrum. Using red indicators with two-photon absorption longer than 980 nm, such as AF647, allows the excitation only by the non-PM pulse. Although the non-PM spectrum is not around the absorption peak of AF647, using few-optical-cycle pulses supported by broadband spectra helps to increase the overlap between the excitation spectrum and the absorption cross-section. Furthermore, we may optimize the sum-frequency excitation using yellow indicators such as tdTomato and YFP. Additionally, we have compared the two-photon fluorescence signal of individual dyes using the non-PM pulse and the DW pulse, as shown in Fig. 5(b). We use the pulse with the spectrum shown in Fig. 3(b), and we employ 1000-nm longpass and shortpass filters to isolate the non-PM pulse and the DW pulse, respectively. The fluorescent signals from AF647 and AF488 using the unfiltered pulse are comparable with the signals from the separated excitations using the non-PM pulse and the DW pulse, respectively. Moreover, the two-photon signals from SYTOX Orange using the pulse with the whole spectrum are significantly higher than the sum of using individual pulses, implying the realization of wavelength mixing.
Fig. 5. (a) Two-photon absorption cross-sections of commonly used fluorescent dyes, in comparison to an excitation spectrum (gray). (b) The fluorescent signal intensity of AF488, SYTOX Orange, and AF647 in response to the two-photon excitation using the gray spectra shown in (a), including the part below 1000 nm (DW), above 1000 nm (non-PM), and without spectral filtering (DW + non-PM).
4. Conclusion
We have precisely managed fiber-optic nonlinearity to develop versatile laser tools for TPFM, allowing efficient multicolor excitation by wavelength mixing. We first introduce the mechanisms of different nonlinear fiber-optic processes, and we can control the onset of nonlinear phenomena by selecting the driving source and the employed fiber with a proper length. Here, we aim to simultaneously generate multiple pulses with manageable temporal overlap for two-photon tissue imaging. In the anomalous dispersion regime, the balance between SPM and negative group delay dispersion forms optical soliton, and the formation of high-order soliton leads to nonlinear pulse compression with a significant spectral broadening. The presence of HOD may lead to efficient DW generation when the spectral broadening reaches the phase-matched wavelength.
We employ a 24-MHz Cr:F laser with an energetic pulse energy of 20 nJ to demonstrate the control of these nonlinear processes. Using a Cr:F laser as a driving source is beneficial for generating energetic pulses via nonlinear fiber-optic conversion. The center wavelength of a Cr:F laser is close to the ZDW of fused silica, and thus controlling the nonlinear pulse propagation in different dispersion regimes does not require strong waveguide dispersion at the expense of a small mode-field area. The compatibility of using LMA fibers allows the delivery of higher-energy pulses than previous works based on Er:fiber, Yb:fiber, and Ti:sapphire lasers [27–30,32], and the boost of the pulse energy is desired for driving TPFM with the consideration of imaging system loss. We can control the DW generation by selecting the employed fiber and its length, and we intentionally realize the spectral enhancement at around 920 nm for the two-photon excitation of green fluorophores. We have demonstrated a broadband spectrum ranging from 830 nm to 1200 nm and achieved PM enhancement at about 920 nm using a 30-mm-long LMA PCF. We have obtained high pulse energy of 5 nJ and compressed the resulting pulse down to 10.5 fs, corresponding to 2.9 optical cycles centered at 1030 nm. We have avoided unwanted nonlinear interactions during the fiber propagation, and thus the broadband spectrum can be fully compressed based on a dispersion compensation scheme with negligible loss.
We manage the soliton formation and DW generation by varying the initial pulse energy, demonstrating the feasibility of controlling the optical delay and power ratio between the solitary pulse and DW. Our investigation paves the way for simultaneously generating multiple few-optical-cycle pulses without needing active pulse shaping and additional delay control. We have simulated and experimentally characterized the spectrotemporal behaviors of the fiber output under different initial pulse energy. Increasing the initial pulse energy leads to the growth of the DW pulse, and varying the launched pulse energy from 9 to 11 nJ allows the control of the delay between the non-PM pulse and DW by 45 fs. We have thus obtained few-optical-cycle pulses supported by the spectra in both ranges of 800-980 nm and 980-1200 nm, and the resulting pulse durations are around 15-20 fs. Despite of the ease and simplicity of our approach, the change in the relative delay and the total output pulse energy are both entangled with the input pulse energy, which would be inappropriate for applications requiring consistent output power during the delay control, such as pump-probe experiments. However, we can also realize the inter-pulse delay tuning by varying the input pulse width (e.g., varying the initial chirp via prism pairs) instead of adjusting the input pulse energy, which potentially leads to smaller changes in the total energy during the delay control.
Furthermore, we demonstrate the capability of our fiber source for multicolor TPFM use, and we have obtained clear images labeled with AF488, SYTOX Orange, and AF647 in a mouse kidney sample slide. By varying the input power before the PCF, we have also shown the availability of wavelength mixing to manage the sum-frequency excitation, enabling us to adjust the image contrasts and optimize the excitation efficiencies. The temporal overlapping between the two pulses is very sensitive to the input power change. Only 7% of the input power change (i.e., 9.25-9.95 nJ of the coupled pulse energy) is enough to scan through the temporal overlapping for efficient sum-frequency excitation, and the corresponding change of the DW pulse energy is around 23%. Additionally, by using few-optical-cycle pulses with a low 14-mW illumination power on the sample, we have achieved deep-tissue images reaching a penetration depth up to 500 µm. Moreover, we can drive multicolor TPFM using other fluorophores by adequately selecting the fiber type and the driving parameters to match the required pulse energy and excitation wavelengths with the spectral peaks of the phase-matched spectral broadening. We believe our demonstrations provide a compact and cost-effective approach for driving multicolor TPFM and the investigation of strong-field-driven phenomena [48].
Funding
Ministry of Science and Technology, Taiwan (MOST 111-2636-E-A49-013).
Acknowledgment
The authors thank Prof. Chi-Kuang Sun for the support of the Cr:forsterite laser system.
Disclosures
The authors declare no conflicts of interest.
Data availability
No data were generated or analyzed in the presented research.
References
1. C.-D. Hu and T. K. Kerppola, “Simultaneous visualization of multiple protein interactions in living cells using multicolor fluorescence complementation analysis,” Nat. Biotechnol. 21(5), 539–545 (2003). [CrossRef]
2. K. Yamauchi, M. Yang, P. Jiang, M. Xu, N. Yamamoto, H. Tsuchiya, K. Tomita, A. R. Moossa, M. Bouvet, and R. M. Hoffman, “Development of real-time subcellular dynamic multicolor imaging of cancer-cell trafficking in live mice with a variable-magnification whole-mouse imaging system,” Cancer Res. 66(8), 4208–4214 (2006). [CrossRef]
3. A. R. Kitching and M. J. Hickey, “Immune cell behaviour and dynamics in the kidney—insights from in vivo imaging,” Nat. Rev. Nephrol. 18(1), 22–37 (2022). [CrossRef]
4. M. Boulina, H. Samarajeewa, J. D. Baker, M. D. Kim, and A. Chiba, “Live imaging of multicolor-labeled cells in Drosophila,” Development 140(7), 1605–1613 (2013). [CrossRef]
5. F. Helmchen and W. Denk, “Deep tissue two-photon microscopy,” Nat. Methods 2(12), 932–940 (2005). [CrossRef]
6. C. Stringari, L. Abdeladim, G. Malkinson, P. Mahou, X. Solinas, I. Lamarre, S. Brizion, J.-B. Galey, W. Supatto, and R. Legouis, “Multicolor two-photon imaging of endogenous fluorophores in living tissues by wavelength mixing,” Sci. Rep. 7(1), 3792 (2017). [CrossRef]
7. P. Mahou, M. Zimmerley, K. Loulier, K. S. Matho, G. Labroille, X. Morin, W. Supatto, J. Livet, D. Débarre, and E. Beaurepaire, “Multicolor two-photon tissue imaging by wavelength mixing,” Nat. Methods 9(8), 815–818 (2012). [CrossRef]
8. A. Rakhymzhan, R. Leben, H. Zimmermann, R. Günther, P. Mex, D. Reismann, C. Ulbricht, A. Acs, A. U. Brandt, and R. L. Lindquist, “Synergistic strategy for multicolor two-photon microscopy: application to the analysis of germinal center reactions in vivo,” Sci. Rep. 7(1), 7101 (2017). [CrossRef]
9. L. Abdeladim, K. S. Matho, S. Clavreul, P. Mahou, J.-M. Sintes, X. Solinas, I. Arganda-Carreras, S. G. Turney, J. W. Lichtman, and A. Chessel, “Multicolor multiscale brain imaging with chromatic multiphoton serial microscopy,” Nat. Commun. 10(1), 1–14 (2019). [CrossRef]
10. Y.-T. Hsiao, Y.-F. Huang, B. J. Borah, S.-K. Chen, and C.-K. Sun, “Single-laser-based simultaneous four-wavelength excitation source for femtosecond two-photon fluorescence microscopy,” Biomed. Opt. Express 12(8), 4661–4679 (2021). [CrossRef]
11. Y. Liu, H. Tu, W. A. Benalcazar, E. J. Chaney, and S. A. Boppart, “Multimodal nonlinear microscopy by shaping a fiber supercontinuum from 900 to 1160 nm,” IEEE J. Sel. Top. Quantum Electron. 18(3), 1209–1214 (2012). [CrossRef]
12. E. P. Perillo, J. W. Jarrett, Y.-L. Liu, A. Hassan, D. C. Fernée, J. R. Goldak, A. Bonteanu, D. J. Spence, H.-C. Yeh, and A. K. Dunn, “Two-color multiphoton in vivo imaging with a femtosecond diamond Raman laser,” Light: Sci. Appl. 6(11), e17095 (2017). [CrossRef]
13. K. Wang, T.-M. Liu, J. Wu, N. G. Horton, C. P. Lin, and C. Xu, “Three-color femtosecond source for simultaneous excitation of three fluorescent proteins in two-photon fluorescence microscopy,” Biomed. Opt. Express 3(9), 1972–1977 (2012). [CrossRef]
14. X. Liang and L. Fu, “Enhanced self-phase modulation enables a 700–900 nm linear compressible continuum for multicolor two-photon microscopy,” IEEE J. Sel. Top. Quantum Electron. 20(2), 42–49 (2014). [CrossRef]
15. K.-C. Li, L. L. Huang, J.-H. Liang, and M.-C. Chan, “Simple approach to three-color two-photon microscopy by a fiber-optic wavelength convertor,” Biomed. Opt. Express 7(11), 4803–4815 (2016). [CrossRef]
16. W. Liu, S.-H. Chia, H.-Y. Chung, R. Greinert, F. X. Kärtner, and G. Chang, “Energetic ultrafast fiber laser sources tunable in 1030–1215 nm for deep tissue multi-photon microscopy,” Opt. Express 25(6), 6822–6831 (2017). [CrossRef]
17. H.-Y. Chung, R. Greinert, F. X. Kärtner, and G. Chang, “Multimodal imaging platform for optical virtual skin biopsy enabled by a fiber-based two-color ultrafast laser source,” Biomed. Opt. Express 10(2), 514–525 (2019). [CrossRef]
18. H.-Y. Chung, W. Liu, Q. Cao, F. X. Kärtner, and G. Chang, “Er-fiber laser enabled, energy scalable femtosecond source tunable from 1.3 to 1.7 µm,” Opt. Express 25(14), 15760–15771 (2017). [CrossRef]
19. H.-Y. Chung, W. Liu, Q. Cao, L. Song, F. X. Kärtner, and G. Chang, “Megawatt peak power tunable femtosecond source based on self-phase modulation enabled spectral selection,” Opt. Express 26(3), 3684–3695 (2018). [CrossRef]
20. W. Liu, C. Li, Z. Zhang, F. X. Kartner, and G. Chang, “Self-phase modulation enabled, wavelength-tunable ultrafast fiber laser sources: an energy scalable approach,” Opt. Express 24(14), 15328–15340 (2016). [CrossRef]
21. R. Chen and G. Chang, “Pre-chirp managed self-phase modulation for efficient generation of wavelength-tunable energetic femtosecond pulses,” J. Opt. Soc. Am. B 37(8), 2388–2397 (2020). [CrossRef]
22. L.-T. Chou, Y.-C. Liu, D.-L. Zhong, W.-Z. Lin, H.-H. Hung, C.-J. Chan, Z.-P. Chen, and S.-H. Chia, “Low noise, self-phase-modulation-enabled femtosecond fiber sources tunable in 740-1236 nm for wide two-photon fluorescence microscopy applications,” Biomed. Opt. Express 12(5), 2888–2901 (2021). [CrossRef]
23. N. G. Horton, K. Wang, D. Kobat, C. G. Clark, F. W. Wise, C. B. Schaffer, and C. Xu, “In vivo three-photon microscopy of subcortical structures within an intact mouse brain,” Nat. Photonics 7(3), 205–209 (2013). [CrossRef]
24. J.-Y. Huang, L.-Z. Guo, J.-Z. Wang, T.-C. Li, H.-J. Lee, P.-K. Chiu, L.-H. Peng, and T.-M. Liu, “Fiber-based 1150-nm femtosecond laser source for the minimally invasive harmonic generation microscopy,” J. Biomed. Opt. 22(03), 1 (2017). [CrossRef]
25. B. Li, M. Wang, K. Charan, M.-J. Li, and C. Xu, “Investigation of the long wavelength limit of soliton self-frequency shift in a silica fiber,” Opt. Express 26(15), 19637–19647 (2018). [CrossRef]
26. G. Krauss, T. Hanke, A. Sell, D. Träutlein, A. Leitenstorfer, R. Selm, M. Winterhalder, and A. Zumbusch, “Compact coherent anti-Stokes Raman scattering microscope based on a picosecond two-color Er: fiber laser system,” Opt. Lett. 34(18), 2847–2849 (2009). [CrossRef]
27. G. Chang, L.-J. Chen, and F. X. Kärtner, “Fiber-optic Cherenkov radiation in the few-cycle regime,” Opt. Express 19(7), 6635–6647 (2011). [CrossRef]
28. A. Sell, G. Krauss, R. Scheu, R. Huber, and A. Leitenstorfer, “8-fs pulses from a compact Er:fiber system: quantitative modeling and experimental implementation,” Opt. Express 17(2), 1070–1077 (2009). [CrossRef]
29. W. P. Putnam, P. D. Keathley, J. A. Cox, A. Liehl, A. Leitenstorfer, and F. X. Kärtner, “Few-cycle, carrier–envelope-phase-stable laser pulses from a compact supercontinuum source,” J. Opt. Soc. Am. B 36(2), A93–A97 (2019). [CrossRef]
30. H.-W. Chen, Z. Haider, J. Lim, S. Xu, Z. Yang, F. X. Kärtner, and G. Chang, “3 GHz, Yb-fiber laser-based, few-cycle ultrafast source at the Ti: sapphire laser wavelength,” Opt. Lett. 38(22), 4927–4930 (2013). [CrossRef]
31. L.-T. Chou and S.-H. Chia, “Widely Tunable Femtosecond Sources with Continuously Tailorable Bandwidth Enabled by Self-Phase Modulation,” in 2022 Conference on Lasers and Electro-Optics (CLEO), IEEE (2022).
32. I. Cristiani, R. Tediosi, L. Tartara, and V. Degiorgio, “Dispersive wave generation by solitons in microstructured optical fibers,” Opt. Express 12(1), 124–135 (2004). [CrossRef]
33. S.-H. Chia, T.-M. Liu, A. A. Ivanov, A. B. Fedotov, A. M. Zheltikov, M.-R. Tsai, M.-C. Chan, C.-H. Yu, and C.-K. Sun, “A sub-100fs self-starting Cr: forsterite laser generating 1.4 W output power,” Opt. Express 18(23), 24085–24091 (2010). [CrossRef]
34. A. Ivanov, A. Voronin, A. Lanin, D. Sidorov-Biryukov, A. Fedotov, and A. Zheltikov, “Pulse-width-tunable 0.7 W mode-locked Cr: forsterite laser,” Opt. Lett. 39(2), 205–208 (2014). [CrossRef]
35. A. A. Ivanov, G. N. Martynov, A. A. Lanin, A. B. Fedotov, and A. M. Zheltikov, “High-energy self-mode-locked Cr: forsterite laser near the soliton blowup threshold,” Opt. Lett. 45(7), 1890–1893 (2020). [CrossRef]
36. J. M. Dudley, G. Genty, and S. Coen, “Supercontinuum generation in photonic crystal fiber,” Rev. Mod. Phys. 78(4), 1135–1184 (2006). [CrossRef]
37. F. M. Mitschke and L. F. Mollenauer, “Discovery of the soliton self-frequency shift,” Opt. Lett. 11(10), 659–661 (1986). [CrossRef]
38. D. R. Austin, C. M. de Sterke, B. J. Eggleton, and T. G. Brown, “Dispersive wave blue-shift in supercontinuum generation,” Opt. Express 14(25), 11997–12007 (2006). [CrossRef]
39. J. M. Dudley and J. R. Taylor, Supercontinuum generation in optical fibers, Cambridge University Press (2010).
40. R. Paschotta, “Spectrograms,”
41. G. Agrawal, Nonlinear Fiber Optics, 5th ed., Academic Press (2013).
42. T. Schreiber, T. Andersen, D. Schimpf, J. Limpert, and A. Tünnermann, “Supercontinuum generation by femtosecond single and dual wavelength pumping in photonic crystal fibers with two zero dispersion wavelengths,” Opt. Express 13(23), 9556–9569 (2005). [CrossRef]
43. M. H. Frosz, P. Falk, and O. Bang, “The role of the second zero-dispersion wavelength in generation of supercontinua and bright-bright soliton-pairs across the zero-dispersion wavelength,” Opt. Express 13(16), 6181–6192 (2005). [CrossRef]
44. J. Peti-Peterdi, K. Kidokoro, and A. Riquier-Brison, “Novel in vivo techniques to visualize kidney anatomy and function,” Kidney Int. 88(1), 44–51 (2015). [CrossRef]
45. W. R. Zipfel, R. M. Williams, R. Christie, A. Y. Nikitin, B. T. Hyman, and W. W. Webb, “Live tissue intrinsic emission microscopy using multiphoton-excited native fluorescence and second harmonic generation,” Proc. Natl. Acad. Sci. 100(12), 7075–7080 (2003). [CrossRef]
46. D. Kobat, N. G. Horton, and C. Xu, “In vivo two-photon microscopy to 1.6-mm depth in mouse cortex,” J. Biomed. Opt. 16(10), 106014 (2011). [CrossRef]
47. M. Oheim, E. Beaurepaire, E. Chaigneau, J. Mertz, and S. Charpak, “Two-photon microscopy in brain tissue: parameters influencing the imaging depth,” J. Neurosci. Methods 111(1), 29–37 (2001). [CrossRef]
48. W. P. Putnam, R. G. Hobbs, P. D. Keathley, K. K. Berggren, and F. X. Kärtner, “Optical-field-controlled photoemission from plasmonic nanoparticles,” Nat. Phys. 13(4), 335–339 (2017). [CrossRef]
