Open Access
Add to BibSonomy
Abstract
The optical dispersion effect in ultrafast pulse laser systems broadens the laser pulse duration and reduces the theoretical peak power. The present study proposes an adaptive ultrashort pulse compressor for compensating the optical dispersion using a direct optical-dispersion estimation by spectrogram (DOES) method. The DOES has fast and accurate computation time which is suitable for real time controller design. In the proposed approach, the group delay dispersion (GDD) and its polarity are estimated directly from the delay marginal of the trace obtained from a single-shot frequency-resolved optical gating (FROG). The estimated GDD is then processed by a closed-loop controller, which generates a command signal to drive a linear deformable mirror as required to achieve the desired laser pulse compression. The dispersion analysis, control computation, and deformable mirror control processes are implemented on a single field programmable gate array (FPGA). It is shown that the DOES dispersion computation process requires just 0.5 ms to complete. Moreover, the proposed pulse compressor compensates for both static dispersion and dynamic dispersion within five time steps when closed-loop controller is performed at a frequency of 100 Hz. The experimental results show that the proposed pulse compressor yields an effective fluorescence intensity improvement in a multiphoton excited fluorescence microscope (MPEFM).
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
In ultrafast laser systems, the spectral components interfere constructively to form a laser pulse train with a high optical energy concentration in ultrashort temporal duration. The high spectral bandwidth of ultrafast lasers yields a short transform-limited pulse width [1]. Moreover, the instantaneous high peak power of ultrafast lasers induces various optical nonlinear phenomena which are useful for many applications. For example, multiphoton excited fluorescence and harmonic generation signals in optical microscopy make possible the reconstruction of three-dimensional images of deep bio-tissue [2–4]. Furthermore, through the use of fast laser scanning techniques or higher pulse energies to produce widefield excitation [5,6], volumetric fluorescence imaging enables many in vivo applications, such as neural activity imaging [7], cell interaction observation [8], and clinical diagnosis [9]. The ultrafast lasers have also been successfully used to accomplish the precise machining of many materials, including hard biotissues [10], metals [11,12], and dielectrics [13]. The plasma-induced ablation and the accompanied photodisruption phenomena with short laser pulses are widely studied. In addition, ultrashort pulse lasers have been employed to probe the instantaneous optical response in material with strong-field nonlinear phononics [14] and to realize high-speed optical transmission [15].
The evolution of ultrafast laser technology has had a significant impact in many scientific and industrial fields. However, in practical applications, the optical components and materials have refractive index which varies with the wavelength. Thus, as the laser propagates through the system, the group velocity dispersion (GVD) and higher-order dispersion effects cause the different spectral components of the laser to have different phase velocities and hence vary the spectral phase distribution of the electric field. The change in the spectral phase broadens the laser pulse shape and reduces the concentration of the optical energy distributed in the pulse. The reduced laser peak power seriously degrades the laser excitation efficiency for optical nonlinear phenomena and may result in thermal damage in machining and imaging applications.
The optical dispersion effect is particularly severe if the laser has a wider spectral bandwidth, and hence some form of pulse compression technique is required to restore the pulse to a transform-limited condition and improve the laser performance accordingly [1]. However, the optical dispersion effect often varies dynamically. For example, the group delay dispersion (GDD) in optical fiber is both wavelength- and temperature-dependent [16,17]. Moreover, different types of fibers have different dispersion characteristics, and hence it is difficult to develop a comprehensive and robust understanding of all the possible dispersion effects which may appear in a fiber system. The optical parametric oscillator exhibits intracavity GDD oscillations with different wavelength [18]. If the output wavelength is tuned, the dispersion varies as well. Free space optical (FSO) communications also suffer from dispersion effects due to temperature and pressure variations [19–21]. The resulting broadening of the temporal pulse increases the bit error rate (BER) and reduces the transmission throughput accordingly. Similarly, in temporal focusing microscopy, specimen-induced spectral phase distortion degrades the system axial resolution and fluorescence image contrast, and hence adversely affects the imaging quality [22].
One of the most common methods for compensating the dispersion in optical systems is the use of prism-based compressors, in which the laser pulse is passed through two prisms to produce a negative GDD with a value proportional to the distance between them [23,24]. Several other methods have been proposed for dispersion compensation, including gratings and acousto-optic deflectors (AOD) [25–27]. However, for all three methods (including prisms), the compensation speed is limited by the mechanical translation response of the components involved. Hence, while such methods are suitable for static dispersion correction, they cannot be applied for real-time compensation. Chirp mirrors provide a simple approach for correcting dispersion by using multiple coating layers to vary the group delays of different incident wavelengths. However, multiple reflections are required to compensate for large dispersions and achieving a precise control of the GDD poses a significant challenge [1]. Several authors have exploited the high pixel density SLMs to realize complex spectral phase manipulation for pulse-shaping applications [28,29]. However, while the desired phase pattern can be applied directly on the SLM, the overall phase has to be wrapped due to the limit phase retardation. The larger number of phase wraps may reduce the SLM modulation efficiency [30]. Electrically-tunable lenses (ETLs) offer a response of just several milliseconds and have a good GDD tuning ability [31]. However, the ETL performance is readily affected by gravity and temperature. Accordingly, deformable mirrors (DMs), which also have a millisecond response, have found increasing use for pulse compression applications in recent decades [32,33]. DMs modulate the spectral phase through the application of a high voltage to an array of actuators, which then deform the reflective membrane as required [34]. Notably, DMs are not only suitable for rapid adaptive pulse compression, but also have the potential to achieve higher-order dispersion compensation.
To achieve adaptive pulse compression, it is necessary to integrate a rapid dispersion estimation method with closed-loop control system in addition to the active phase modulator. Optical autocorrelators can estimate the laser pulse width from the nonlinear signal (e.g., second harmonics signal (SHG) or two-photon excited fluorescence signal) excited by the incident laser pulse and its replica with different group delays [1,35]. However, while several real-time optical autocorrelators have been successfully developed [36,37], obtaining accurate estimates of the GDD is difficult if the laser does not possess an ideal Gaussian pulse shape. Several advanced laser pulse characterization methods based on auxiliary spectrum information of the excited nonlinear signal have thus been developed, including spectral phase interferometry for direct electric-field reconstruction (SPIDER) [38], multiphoton intrapulse interference phase scan (MIIPS) [39], and frequency-resolved optical gating (FROG) [40]. In SPIDER, the electric field of the laser pulse is reconstructed and the optical dispersion is analyzed based on its spectral phase distribution. However, to maximize the interference contrast, the laser pulses in the two interferometry arms must be carefully shaped and optically aligned. Furthermore, MIIPS and FROG both require several scans for each GDD measurement, and are thus unsuitable for real-time applications. In addition, although single-shot FROG based on the GRENOUILLE (grating-eliminated no-nonsense observation of ultrafast incident laser light e-fields) technique can be used to achieve single laser pulse analysis, the reconstruction algorithm requires multiple iterations, which increases the dispersion computation time, particularly if the laser pulse is complex [40–42]. Many machine learning algorithms have been proposed in recent years for electric field reconstruction and optical dispersion estimation [33,43]. However, the performance of such methods is dependent on the robustness of the training process and the optimality of the neural network structure.
Accordingly, the present study utilizes a direct optical-dispersion estimation based on spectrogram (DOES) method to perform fast and accurate GDD measurement [44]. In particular, the frequency marginal of the spectrogram obtained by a single-shot GRENOUILLE FROG is used to obtain the laser spectrum information and transform-limit pulse shape for various GDDs [40]. A lookup table (LUT) is then constructed to map the GDD values to the corresponding FWHM (full width at half maximum) of the laser pulse intensity autocorrelation (IAC) signal. The delay marginal of the spectrogram is equivalent to the IAC, and hence, in the pulse compression process, the measured FWHM of the delay marginal can be used to obtain the GDD directly from the LUT without iteration. Notably, the DOES algorithm enables the GDD to be calculated within just 0.5 ms and is thus suitable for real-time dispersion measurement and control. The DOES algorithm is integrated with a digital PI (proportional-integral) controller implemented on a FPGA with a sampling rate of 100 Hz in order to realize DM-based adaptive ultrashort pulse compression. Compared with previous study [44], the proposed DOES has improved the calculation time by around 38 fold and the GDD polarity can be estimated which is ideally applied as feedback signal in real time controller. It is shown that the proposed pulse compressor enables both static and dynamic dispersion to be compensated within five time steps. Moreover, when integrated with a self-built point scanning-based multiphoton excited fluorescence microscope (MPEFM), the pulse compressor yields an effective improvement in the fluorescence intensity. Hence, the practical feasibility of the proposed pulse compressor for restoring the laser pulse width to the transform-limited condition is confirmed.
2. System and method
2.1 System setup
Figure 1 illustrates the experimental setup considered in the present study. The ultrafast laser (Chameleon Ultra II, Coherent) had a maximum power of 3.5 W, a pulse width of 140 fs for a wavelength of 800 nm, and a repetition rate of 80 MHz. In the pulse-compression process, the half-wave plate (HWP1) was rotated such that only the horizontal portion of the laser light passed through the polarization beam splitter (PBS1) behind. The laser power can be well controlled while maintaining the polarization state to the optical system. The optical dispersion source, such as N-SF11 glass rod, introduced static dispersion behind the PBS to disturb the laser spectral phase and broaden the laser pulse width. In addition, a prism-based pulse compressor (BOA-800, Swamp Optics) can precompensate dispersion for system and was able to generate dynamic dispersion to verify the real-time compensation performance of the proposed adaptive pulse compressor. The dispersed beam was incident on a 1200-grooves/mm blazed grating (GR25-1208, Thorlabs) in order to separate the spectral components of the beam into different angles in accordance with the standard grating equation. A bi-convex lens with a focal length of 400 mm was placed at the Fourier plane of the grating to decouple the spectral components in a horizontal direction. The decoupled beam was then incident on a linear deformable mirror (MMDM linear 11 × 39mm 39 ch, OKO Tech) with a gold coating and a maximum deformation of 9.4 µm. The DM was tilted slightly such that the modulated beam passed back through the bi-convex lens and grating and was incident on a GRENOUILLE-based FROG (8-50-USB, Swamp Optics). The resulting optical spectrogram was captured by a CMOS camera (UI-3370CP-M-GL, IDS) of GRENOUILLE-based FROG acting under the control of a FPGA (PXIe-7971 + NI-1483, National Instruments) with an exposure time of 4 ms. Prior to incidence on the FROG, the modulated beam was passed through a second HWP and PBS to distribute the laser power between the FROG and a self-built MPEFM system. (Note that the FROG required an incident power of around 30 mW to achieve a trace with adequate contrast for analysis purposes.) The SHG-based FROG is unable to distinguish the dispersion polarity (which is required for the controller design), and hence a TeO2 bias dispersion crystal (23080-3-1.06-LTD, NEOS Technologies) with a known GDD bias of 11000 fs2 [44] was placed before the MPEFM.
In order to implement a real-time PI controller on the FPGA, the linear DM driver was customized to be driven by digital signals. In particular, the FPGA output was interfaced to a 40-channel digital-to-analog converter (DAC; EVAL-AD5380, Analog Devices) with 14-bit resolution and a settling time of approximately 3 µs. The signals produced by the DAC were then fed to a 40-channel high-voltage amplifier (OKO Tech) with a voltage gain of around 81 and a maximum drive voltage output of 179 V for maximum DM deformation. The amplified signals were finally passed to the actuators of the linear DM to deform the membrane surface as required to restore the laser pulse to a transform-limited condition. The DAC was set to have a maximum output control voltage of 2 V (i.e., around 162 V after amplification). The control voltage can set to 1.4 V indicating around half of the DM maximum deformation. (Note that the DM deformation is linearly proportional to the square of the applied voltage.) Only three FPGA digital I/Os were required to perform serial communication with the DAC and drive the linear DM.
The MPEFM system comprised a pair of x-y galvanometer scanners (6215H, Cambridge Technology), a 10x objective lens (MSPlan10 NA 0.3, Olympus), and a photomultiplier tube (PMT; H10721-210 + C6438-02, Hamamatsu) with a single photon counting technique for high signal-to-noise ratio (SNR) fluorescence signals. The two-photon absorption efficiency of the MPEFM was inversely proportional to the laser pulse width. Hence, changes in the intensity of the detected fluorescence signal indicated changes in the dispersion of the incident laser pulse during the compensation process.
2.2 Direct optical-dispersion estimation by spectrogram (DOES)
To achieve a real-time controller for adaptive pulse compression schemes, the feedback dispersion sensor must have a fast, responsive, and stable computation time. The DOES method proposed in the present study thus utilizes a GRENOUILLE-based FROG to evaluate the phase dispersion in the system due to its ability to produce a single-shot spectrogram (i.e., FROG trace) and without the need for iterative computation and electric field reconstruction [44]. The detailed steps in the proposed DOES algorithm are shown in Fig. 2. In general, the laser spectrum can be obtained directly by a spectrometer or derived from the frequency marginal of the spectrogram [40]. In the present study, the latter approach is employed, and the laser spectrum is obtained from the frequency marginal of the FROG trace. The spectrum contains the spectral distribution of the laser pulses, but provides no indication of the phase information. However, it indicates the Fourier transform of the transform-limited electric field $\mathbb {E}$(ω), for which the spectral phase is zero. Different amounts of GDD are then added to the spectral phase as follows:
In the compensation process, the FPGA triggers the CMOS camera to capture the SHG FROG trace and then measures the delay marginal of the spectrogram. The delay marginal is equivalent to the intensity autocorrelation, IAC(τ). Thus, the FWHM of the IAC(τ) (i.e., delay marginal) is measured and input to the LUT to determine the corresponding |GDDLUT|. Due to the symmetry of the IAC, the dispersion polarity cannot be estimated, i.e., the FPGA cannot be sure whether the estimated dispersion is positive or negative. Thus, as described above, a bias dispersion crystal with a known GDDbias of 11000 fs2 was placed before the MPEFM. During the calibration stage, the DM was driven in such a way as to provide negative GDDs, and the DOES measured the corresponding GDD quantity from −3000 to −19000 fs2 (i.e., −|GDDLUT|), corresponding to actual dispersions of the incident laser pulse at the MPEFM of 8000 to −8000 fs2. In the estimation process, the true GDD with polarity in the system was then obtained simply by adding GDDbias to the negative value of |GDDLUT|.
To verify the accuracy of the proposed DOES algorithm, positive and negative dispersions were applied individually to the system, and the corresponding GDD was estimated, as shown in Table 1. In the absence of any dispersion medium in the system, the DOES estimated the GDD to be −10500 ± 200 fs2, which is provided by DM in order to compensate the bias dispersion of TeO2 and it corresponds to a GDD of 500 ± 200 fs2before the MPEFM system. Given the introduction of a 40-mm N-SF11 glass rod (SF11G0400, Newlight Photonics) with a GDD of 7500 fs2 after prism-based compressor, the GDD was estimated to be −3600 ± 100 fs2, corresponding to a value of 7400 ± 100 fs2 before the MPEFM. The estimation error was thus determined to be 1.3%. Finally, following the introduction of a prism-based compressor with a negative dispersion of −4700 fs2, the DOES was estimated to be −15600 ± 600 fs2, i.e., −4600 ± 600 fs2 prior to the MPEFM. In other words, the estimation error was equal to 6%. Overall, the results in Table 1 confirm that the proposed DOES algorithm yields accurate estimates of the GDD value in the system together with its polarity, and is hence suitable for integration with the FPGA controller to control the pulse compression in the considered MPEFM system.
Table 1. The GDD estimation results obtained by DOES algorithm for different dispersive media
2.3 Controller design
As shown in Fig. 1, pulse width compensation was performed using a DM-based approach. To select an appropriate control system model, voltages ranging from 0 to 162 V after voltage amplification were first applied to the DM and the induced GDD was measured by DOES accordingly. Figure 3 shows the measured values of the GDD in each case (see blue circles). The fitting results (shown by the red line) confirm that the GDD is ideally linear proportional to the square of the control voltage. In other words, the DM-based pulse compressor is a linear system and can thus be driven by a closed-loop PI controller.
Fig. 3. Relationship between measured GDD and DM voltage squared. Note that the blue circles show the experimental data, while the red line shows the linear fitting curve.
Figure 4(a) presents a block diagram of the overall control system. As shown, the GDD in the system is estimated using the DOES method described above and is taken as a feedback signal for comparison with the target. The error between the estimated GDD and the target GDD is then provided as an input to the digital PI controller implemented on the FPGA to compute the actuation command required to drive the DM in such a way as to restore the measured GDD to the target value [34,45]. In the proposed controller, the proportional gain (P) and integral gain (I) parameters were set as 0.1 and 50, respectively. Note that a larger P value was deliberately avoided since, while it yields a faster response, it may also induce an overshoot effect that might damage the DM. In addition, the I value was chosen in such a way as to eliminate the steady state error and ensure that the final output traced the target. The loop rate of the PI controller was set as 100 Hz in accordance with the minimum exposure time of the CMOS camera for an appropriate image quality.
Fig. 4. (a) Block diagram of closed-loop PI control system; (b) timing diagram of single control loop.
The FROG trace analysis, DOES algorithm, controller calculation, and multichannel DAC communication were all implemented on the FPGA in order to facilitate a real-time digital controller performance. The overall timing diagram for each control loop is shown in Fig. 4(b). Once the camera is triggered, the exposure and spectrograph image acquisition process require approximately 6 ms. Thereafter, the analyzed spectrogram data is sent to the FPGA via DMA (direct memory access) FIFO (first-in-first-out), where the GDD is estimated by the DOES algorithm. The data transfer and DOES computation consume approximately 0.5 ms given an FPGA operating clock speed of 40 MHz. The PI controller calculation and subsequent DM driver response (including serial communication via the multichannel DAC) consume a further 0.74 ms and 0.835 ms, respectively. Finally, the DM deformation response has a duration of around 1 ms. Thus, the overall computation time is around 9 ms, indicating that the closed-loop controller can operate at a stable 100 Hz including both dispersion detection and DM correction.
3. Experimental results
3.1 Sampling rate consistency of dispersion estimation in controller
To confirm the ability of the digital controller to operate stably at a sampling speed of 100 Hz, dispersion disturbances with known frequencies of 10, 30 and 50 Hz were applied to the system shown in Fig. 1 under the assumption that the measured dispersion would exhibit a frequency drift if the sampling rate were not consistent. To ensure the objectivity of the measurement results, the dispersion disturbances were applied by an independent FPGA (myRIO-1900, National Instruments) rather than the FPGA used to implement the DOES estimation scheme and closed-loop controller. For each of the considered dispersion frequencies, a sinusoidal control voltage ranging from 0 to 162 V was applied to the DM and the generated dispersion was estimated by the DOES algorithm. In performing the experiments, the controller computation and multi-channel DAC communications were left running on the FPGA in order to maintain the timing consistency as shown in Fig. 4(b).
Figure 5(a) shows the square root of the measured dispersion (|GDDLUT|) change over time for each of the three dispersion disturbances. As discussed earlier, the GDD is proportional to the square of the control voltage, and hence Fig. 5(a) shows the square root of the dispersion in order to reveal the applied sinusoidal control voltage. Figure 5(b) shows the spectra of the measured dispersion curves after Fourier transformation and the removal of the zero-frequency components. High contrast peaks are observed at frequencies close to 10, 30, and 50 Hz, respectively. Thus, the stable operation of the controller at the specified sampling rate of 100 Hz is confirmed. (Note that, in accordance with the Nyquist theorem, 50 Hz is the maximum frequency the system can detect with a 100 Hz sampling rate.)
Fig. 5. Dispersion disturbances with different frequencies (red: 10 Hz; blue: 30 Hz; and magenta: 50 Hz). (a) Variation over time of measured square root of GDD (i.e., |GDDLUT|). (b) Fourier transformed curves in frequency domain with red, blue, and magenta peaks located at 10.03, 29.99, and 49.94 Hz respectively.
3.2 Dispersion correction response
The prism-based pulse compressor was used to introduce a GDD of approximately −6000 fs2 into the optical system for compensating MPEFM system in additional to bias dispersion. Thus, a DOES estimate of −17000 fs2 (i.e., a bias dispersion of 11000 fs2 and an MPEFM GDD of 6000 fs2) was taken to indicate the ability of the proposed pulse compressor to properly compensate the dispersion within the system. To evaluate the dispersion correction response time of the pulse compressor, a 40-mm N-SF11 rod with a known GDD of 7500 fs2 was placed after the prism-based compressor in Fig. 1. The GDD command of the controller was set to −17000 fs2 in order to compensate for the static dispersion in the system. The MPEFM was then used to scan a homogeneous fluorescence slide (FSK3, Thorlabs) with a pixel scanning rate of 200 Hz and to construct a fluorescence-based image with a size of 80×80 pixels. The controller was turned on midway through the scanning process in order to compensate the dispersion within the system and enhance the quality of the reconstructed image as a result.
Figure 6(a) shows that, following the activation of the controller, the GDD converges to the target value of −17000 fs2 within 50 ms (i.e., five time steps). Furthermore, Fig. 6(b) shows that the fluorescence intensity detected by the MPEFM rapidly increases by around 1.4 fold within 10 pixels after the controller is turned on. Figure 6(c) shows the corresponding improvement in the quality of the reconstructed 2D MPEFM image, in which the upper region with a weaker fluorescence intensity and lower region with an enhanced fluorescence intensity are constructed before and after the controller is activated, respectively.
Fig. 6. Static dispersion correction response. (a) GDD compensation curve converges to GDD command of −17000 fs2. (b) Variation of MPEFM fluorescence intensity over time while scanning fluorescence slide with and without dispersion control. (c) Reconstructed 2D MPEFM image with and without dispersion control.
3.3 Dynamic dispersion compensation
The results presented in the preceding sections confirm that the proposed ultrashort pulse compressor achieves a rapid compensation of static dispersion. A further investigation was thus performed to evaluate its ability to perform the dynamic compensation of low-frequency dispersion. The prism-based compressor was driven by a sinusoidal signal with a frequency of 0.2 Hz and used to vary the dispersion within the system dynamically between −18000 fs2 and −23000 fs2. As shown in Fig. 7(a), the dynamic dispersion was eliminated almost immediately (within 50 ms) once the controller was turned on, and the GDD converged to the command value of −17000 fs2, indicating the ability of the pulse compression scheme to compensate for both dynamic disturbance in the MPEFM system and static GDD. The results presented in Figs. 7(b) and 7(c) show that the activation of the controller yields a rapid and stable improvement of around ∼1.3 fold in the fluorescence intensity of the MPEFM system and significantly improves the quality of the reconstructed 2D image as a result.
Fig. 7. Dynamic dispersion correction response: (a) GDD compensation curve converges to GDD command of −17000 fs2. (b) Variation of MPEFM fluorescence intensity over time while scanning fluorescence slide with and without dispersion control. (c) Reconstructed 2D MPEFM image with and without dispersion control.
4. Conclusion and discussion
Ultrashort laser pulse systems suffer an optical dispersion problem, in which the spectral components of the laser pulse are temporally disturbed, resulting in a broadening of the pulse shape and a reduction in the energy concentration. Optical dispersion has a detrimental effect on many applications, including, nonlinear optical microscopy, optical communications, laser ablation, and so on. As a result, the development of effective methods for compensating the optical dispersion effect is of great practical importance.
Dispersion compensation can be achieved by maximizing the nonlinear excited signal, e.g., through second-harmonic-generation (SHG) methods. However, the enhanced signal provides no indication of any residual dispersion quantity which may not have been compensated. Several advanced dispersion estimation techniques based on electric field reconstruction have been developed, including FROG, SPIDER, and MIIPS [1]. However, these methods have several drawbacks, such as the need for mechanical scanning, iterative reconstruction algorithms, or complex computations. As a result, they are infeasible for dynamic dispersion compensation.
Accordingly, this study has proposed an adaptive ultrashort pulse compressor scheme consisting of a DOES algorithm for GDD estimation, a DM-based compressor system for pulse correction, and a digital closed-loop controller implemented on a FPGA. In the DOES estimation method, a LUT is prepared in advance to record the relationship between the GDD and the FWHM value of the corresponding IAC. In the dispersion estimation process, the FWHM of the delay marginal obtained from a FROG trace (equivalent to the IAC) is used to interrogate the LUT and extract the corresponding GDD directly without the need for iteration. Having estimated the magnitude of the GDD, the polarity of the dispersion is determined by means of a bias dispersion crystal GDD placed in the system immediately before the MPEFM. Given the estimated magnitude and polarity of the dispersion, a closed-loop controller implemented on the FPGA is used to drive the DM as required to restore the GDD to the target value.
It has been shown that the overall computation time of the FPGA (including both dispersion estimation and DM correction) is approximately 9 ms. Hence, the closed-loop controller can be operated stably at a frequency of 100 Hz. The experimental results have shown that the proposed pulse compressor enables both static and dynamic dispersion to be rapidly compensated within just five time steps. Moreover, the MPEFM observation results have confirmed that the pulse compressor yields a 1.3∼1.4 fold improvement in the intensity of the fluorescence signal. If the ultrafast laser has shorter transform-limited pulse width, the dispersion compensation would result higher fluorescence intensity improvement because the shorter laser pulses are more rigorously affected by dispersion. For instance, adding a GDD of 1000 fs2 to transform-limited Gaussian pulses of 140 fs and 40 fs would broaden the laser pulses to 141 fs and 80 fs, respectively.
The present study has adopted a single-shot GRENOUILLE FROG to analyze the laser pulse. However, future studies may usefully consider the use of other single-shot autocorrelation techniques [36,37] to perform rapid IAC measurement and synchronization with the controller. Furthermore, while the DM used in the present study to compensate for the dispersion within the system has a rapid response, the laser beam is reflected twice from the associated grating structure during the dispersion estimation process. Thus, the diffraction efficiency of the grating may be an issue for laser systems with a limited power. Accordingly, future studies can design a high diffraction efficiency grating superiorly optimized for laser wavelength of 800 nm [46]. The loop rate of the present system is 100 Hz. However future studies may attempt to increase the loop rate further by replacing the CMOS camera with a high-sensitivity scientific camera for better SNR images with a shorter exposure time. Finally, in the present study, the pulse compressor has been implemented using a simple linear closed-loop model. Future studies may consider the use of a more advanced controller design in order to extend the frequency bandwidth over which the pulse compressor can feasibly operate [34,45].
Funding
Ministry of Science and Technology (110-2622-E-006-035, 110-2636-E-006-018).
Disclosures
The authors declare no conflicts of interest.
Data availability
No data were generated or analyzed in the presented research.
References
1. F. Träger, Handbook of Lasers and Optics (Springer, 2007).
2. M. Rumi and J. W. Perry, “Two-photon absorption: an overview of measurements and principles,” Adv. Opt. Photonics 2(4), 451–518 (2010). [CrossRef]
3. D. S. James and P. J. Campagnola, “Recent Advancements in Optical Harmonic Generation Microscopy: Applications and Perspectives,” BME Frontiers 2021, 1–24 (2021). [CrossRef]
4. K. Choe, Y. Hontani, T. Wang, D. Ouzounov, K. Lai, A. Singh, W. Béguelin, A. M. Melnick, and C. Xu, “Intravital three-photon microscopy allows visualization over the entire depth of mouse lymph nodes,” Nat. Immunol. 23(2), 330–340 (2022). [CrossRef]
5. C.-Y. Chang, Y. Y. Hu, C.-Y. Lin, C.-H. Lin, H.-Y. Chang, S.-F. Tsai, T.-W. Lin, and S.-J. Chen, “Fast volumetric imaging with patterned illumination via digital micro-mirror device-based temporal focusing multiphoton microscopy,” Biomed. Opt. Express 7(5), 1727–1736 (2016). [CrossRef]
6. J. Wu, N. Ji, and K. K. Tsia, “Speed scaling in multiphoton fluorescence microscopy,” Nat. Photonics 15(11), 800–812 (2021). [CrossRef]
7. D. M. Chow, D. Sinefeld, K. E. Kolkman, D. G. Ouzounov, N. Akbari, R. Tatarsky, A. Bass, C. Xu, and J. R. Fetcho, “Deep three-photon imaging of the brain in intact adult zebrafish,” Nat. Methods 17(6), 605–608 (2020). [CrossRef]
8. L. Perrin, B. Bayarmagnai, and B. Gligorijevic, “Frontiers in intravital multiphoton microscopy of cancer,” Cancer Rep. (N. Y., NY, U. S.) 3(1), e1192 (2020). [CrossRef]
9. J. Yan, Y. Zheng, X. Zheng, Z. Liu, W. Liu, D. Chen, X. Dong, K. Li, X. Liu, G. Chen, J. Lu, J. Chen, S. Zhuo, and G. Li, “Real-time optical diagnosis of gastric cancer with serosal invasion using multiphoton imaging,” Sci. Rep. 6(1), 31004 (2016). [CrossRef]
10. M. H. Niemz, Laser-Tissue Interactions: Fundamentals and Applications, 4th edition (Springer, 2019.
11. K. C. Phillips, H. H. Gandhi, E. Mazur, and S. K. Sundaram, “Ultrafast laser processing of materials: a review,” Adv. Opt. Photonics 7(4), 684–712 (2015). [CrossRef]
12. J. Winter, M. Spellauge, J. Hermann, C. Eulenkamp, H. P. Huber, H. P. Huber, M. Schmidt, and M. Schmidt, “Ultrashort single-pulse laser ablation of stainless steel, aluminium, copper and its dependence on the pulse duration,” Opt. Express 29(10), 14561–14581 (2021). [CrossRef]
13. S. Schwarz, S. Rung, C. Esen, and R. Hellmann, “Ultrashort pulsed laser backside ablation of fused silica,” Opt. Express 29(15), 23477–23486 (2021). [CrossRef]
14. A. von Hoegen, R. Mankowsky, M. Fechner, M. Först, and A. Cavalleri, “Probing the interatomic potential of solids with strong-field nonlinear phononics,” Nature 555(7694), 79–82 (2018). [CrossRef]
15. T. Hirooka, R. Hirata, J. Wang, M. Yoshida, and M. Nakazawa, “Single-channel 10.2 Tbit/s (2.56 Tbaud) optical Nyquist pulse transmission over 300 km,” Opt. Express 26(21), 27221–27236 (2018). [CrossRef]
16. T. Kato, Y. Koyano, and M. Nishimura, “Temperature dependence of chromatic dispersion in various types of optical fiber,” Opt. Lett. 25(16), 1156–1158 (2000). [CrossRef]
17. S. Tiwari, A. Dixit, and P. C. Pandey, “Temperature-based dispersion compensating ability of a photonic crystal fiber,” Opt. Eng. 61(01), 016105 (2022). [CrossRef]
18. K. Stankevičiūtė, M. Vengris, S. Melnikas, S. Kičas, R. Grigonis, and V. Sirutkaitis, “Tuning characteristics of femtosecond optical parametric oscillator with broadband chirped mirrors,” Opt. Eng. 54(12), 126111 (2015). [CrossRef]
19. M. Hippke, “Interstellar communication: Short pulse duration limits of optical SETI,” J. Astrophys. Astron. 39(6), 74 (2018). [CrossRef]
20. M. Bouhadda, F. M. Abbou, M. Serhani, F. Chaatit, A. Abid, and A. Boutoulout, “Temporal pulse broadening due to dispersion and weak turbulence in FSO communications,” Optik 200, 163327 (2020). [CrossRef]
21. K. Sunilkumar, N. Anand, S. K. Satheesh, K. K. Moorthy, and G. Ilavazhagan, “Enhanced optical pulse broadening in free-space optical links due to the radiative effects of atmospheric aerosols,” Opt. Express 29(2), 865–876 (2021). [CrossRef]
22. C.-Y. Chang, L.-C. Cheng, H.-W. Su, Y. Y. Hu, K.-C. Cho, W.-C. Yen, C. Xu, C. Y. Dong, and S.-J. Chen, “Wavefront sensorless adaptive optics temporal focusing-based multiphoton microscopy,” Biomed. Opt. Express 5(6), 1768–1777 (2014). [CrossRef]
23. R. L. Fork, O. E. Martinez, and J. P. Gordon, “Negative dispersion using pairs of prisms,” Opt. Lett. 9(5), 150–152 (1984). [CrossRef]
24. S. Akturk, X. Gu, M. Kimmel, and R. Trebino, “Extremely simple single-prism ultrashort-pulse compressor,” Opt. Express 14(21), 10101–10108 (2006). [CrossRef]
25. S. Bucht, D. Haberberger, J. Bromage, and D. H. Froula, “Methodology for designing grism stretchers for idler-based optical parametric chirped-pulse-amplification systems,” J. Opt. Soc. Am. B 36(8), 2325–2337 (2019). [CrossRef]
26. W. Zhang, W. Kong, G. Wang, F. Xing, F. Zhang, H. Zhang, and S. Fu, “Review of pulse compression gratings for chirped pulse amplification system,” Opt. Eng. 60(2), 020902 (2021). [CrossRef]
27. A. Yamaguchi, D. Karagyozov, and M. H. Gershow, “Compact and adjustable compensator for AOD spatial and temporal dispersion using off-the-shelf components,” Opt. Lett. 46(7), 1644–1647 (2021). [CrossRef]
28. D. Yelin, D. Meshulach, and Y. Silberberg, “Adaptive femtosecond pulse compression,” Opt. Lett. 22(23), 1793–1795 (1997). [CrossRef]
29. A. Weiner, “Femtosecond pulse shaping using spatial light modulators,” Rev. Sci. Instrum. 71(5), 1929–1960 (2000). [CrossRef]
30. Yuhan Yang, Wei Chen, Jiang Lan Fan, and Na Ji, “Adaptive optics enables aberration-free single-objective remote focusing for two-photon fluorescence microscopy,” Biomed. Opt. Express 12(1), 354–366 (2021). [CrossRef]
31. M. E. Durst, A. Turcios, C. Laurence, and E. Moskovitz, “Dispersion compensation by a liquid lens (DisCoBALL),” Appl. Opt. 58(2), 428–435 (2019). [CrossRef]
32. E. Zeek, K. Maginnis, S. Backus, U. Russek, M. Murnane, G. Mourou, H. Kapteyn, and G. Vdovin, “Pulse compression by use of deformable mirrors,” Opt. Lett. 24(7), 493–495 (1999). [CrossRef]
33. C. A. Farfan, J. Epstein, and D. B. Turner, “Femtosecond pulse compression using a neural-network algorithm,” Opt. Lett. 43(20), 5166–5169 (2018). [CrossRef]
34. Yu-C. Wu, J.-C. Chang, and C.-Y. Chang, “Adaptive optics for dynamic aberration compensation using parallel model-based controllers based on a field programmable gate array,” Opt. Express 29(14), 21129–21142 (2021). [CrossRef]
35. J. B. Guild, C. Xu, and W. W. Webb, “Measurement of group delay dispersion of high numerical aperture objective lenses using two-photon excited fluorescence,” Appl. Opt. 36(1), 397–401 (1997). [CrossRef]
36. K. Kondo and T. Baba, “On-chip autocorrelator using counter-propagating slow light in a photonic crystal with two-photon absorption photodiodes,” Optica 4(9), 1109–1112 (2017). [CrossRef]
37. F. Billard, A. Dubrouil, E. Hertz, S. Lecorné, E. Szmygel, O. Faucher, and P. Béjot, “BOAR: Biprism based optical autocorrelation with retrieval,” Rev. Sci. Instrum. 90(6), 063110 (2019). [CrossRef]
38. C. Iaconis and I. A. Walmsley, “Spectral phase interferometry for direct electric-field reconstruction of ultrashort optical pulses,” Opt. Lett. 23(10), 792–794 (1998). [CrossRef]
39. V. V. Lozovoy, I. Pastirk, and M. Dantus, “Multiphoton intrapulse interference. IV. Ultrashort laser pulse spectral phase characterization and compensation,” Opt. Lett. 29(7), 775–777 (2004). [CrossRef]
40. R. Trebino, Frequency-Resolved Optical Gating: The Measurement of Ultrashort Laser Pulses (Kluwer Academic, 2002).
41. D. J. Kane, “Recent progress toward real-time measurement of ultrashort laser pulses,” IEEE J. Quantum Electron. 35(4), 421–431 (1999). [CrossRef]
42. R. Jafari and R. Trebino, “High-speed “multi-grid” pulse-retrieval algorithm for frequency-resolved optical gating,” Opt. Express 26(3), 2643–2649 (2018). [CrossRef]
43. G. Genty, L. Salmela, J.M. Dudley, D. Brunner, A. Kokhanovskiy, S. Kobtsev, and S. K. Turitsyn, “Machine learning and applications in ultrafast photonics,” Nat. Photonics 15(2), 91–101 (2021). [CrossRef]
44. J.-C. Chang, S.-Yu Chang, Yu-C. Wu, and C.-Y. Chang, “Fast and direct optical dispersion estimation for ultrafast laser pulse compression,” Rev. Sci. Instrum. 92(11), 113702 (2021). [CrossRef]
45. G. F. Franklin, J. D. Powell, and M. Workman, Digital Control of Dynamic Systems, 3rd ed. (Ellis-Kagle, 1998).
46. D. A. Alessi, H. T. Nguyen, J. A. Britten, P. A. Rosso, and C. Haefner, “Low-dispersion low-loss dielectric gratings for efficient ultrafast laser pulse compression at high average powers,” Opt. Laser Technol. 117, 239–243 (2019). [CrossRef]
