Sub-50-femtosecond gain-managed amplified pulses enhance nonlinear ablation efficiency

Abstract

Nonlinear femtosecond (fs) laser ablation enables highly localized energy deposition for cell microsurgery. Conventional systems operate at either low (∼1 kHz, amplified µJ pulse energy) or high (∼80 MHz, unamplified low nJ) repetition rates, but intermediate rates with amplified pulse energy offer a promising balance of ablation speed and thermal control. We custom-built a low-cost, 32 MHz femtosecond fiber laser system with gain-managed nonlinear amplification, boosting pulse energy from 5 to 90 nJ while compressing pulse duration to 46 fs. In this intermediate-repetition-rate regime, the use of sub-50-fs pulses enhances ablation efficiency by strengthening multiphoton absorption and lowering the effective ablation threshold, while also leveraging multi-pulse incubation effects that promote cumulative energy deposition at reduced per-pulse energies. Compared to 200-500 fs pulses typically used for ablation, shorter durations double ablation efficiency in silicon and yield a ∼10× increase in cell membrane damage. In 3D tumor models, this approach enables targeted subsurface ablation up to 400 µm depth with 6 nJ pulse energy. These results inform the design of next-generation femtosecond laser systems for microsurgery.

Fig. 1.

(a) Schematic of the complete optical system used for the nonlinear ablation experiments. A 32 MHz all-normal-dispersion (ANDi) laser oscillator is coupled to the input collimator (C1) and serves as the seed source for the gain-managed amplifier (GMA), which outputs pulses with sufficient energy and compressibility for ablation. A short pass (SP) filter is used to allow targeting of the ablation beam along with a camera for live-aiming. For silicon ablation, a low NA air objective was used for focusing. For cellular ablation, the beam was focused through a high-numerical-aperture (NA) microscope objective (20 $\times$ , 1.0 NA; Olympus XLUMPLFLN). (b) Autocorrelation trace showing the optimally compressed pulse duration ( $\sim$ 46 fs autocorrelation fit) used in ablation experiments. (c) Power spectral density of the pulse measured at the GMA output, showing a spectral bandwidth of approximately 140 nm.

Fig. 2.

(a) Cross-sectional areas depicting the varying morphologies of ablation grooves generated by ablation using pulse widths of approximately 46, 110, 214, 319, 431, 571, 714, and 1431 fs (i-viii). Arrows denote the grooves generated at 0.2 m/s scan speed. Scale bars are 20 µm. (b) One technical replicate acquired at 46 fs showing the cross-sectional areas that were measured. (c) Silicon ablation experiments reveal a clear increase in material ablation with decreasing laser pulse duration. The ablation laser delivered pulses at 32 MHz with a constant energy of 60 nJ per pulse. The blue ellipse highlights the typical pulse durations used in commercially available systems with similar pulse energies. Cross-sectional areas for each groove were measured as shown in panel (b) and the mean ablation rate for each pulse duration is reported here. Results are mean $\pm$ s.e.m. (n=4 technical replicates, ${\chi }_{red}^2=10.4)$ . Given the relative simplicity of this model and the relatively large reduced chi-squared value, the model may underfit the data. With constant scan speed and stable beam shape, this ablation rate serves as a reliable metric of ablation efficiency, demonstrating the advantages of ultrashort pulses in the mid-repetition-rate regime.

Fig. 3.

The dependence of laser ablation efficiency on pulse duration was evaluated in Ovcar5-GFP cell cultures using propidium iodide (PI) fluorescence as a marker of membrane disruption. All ablation experiments were performed using laser pulses delivered at 32 MHz with a constant pulse energy of 6 nJ . (a) The mean PI fluorescence intensity per ablation laser raster scan was quantified by fluorescence microscopy and plotted as a function of pulse duration ( $\tau$ ). A ${\tau }^{-2}$ trendline is overlaid to illustrate the approximate relationship between pulse duration and ablation efficiency. For each of these replicates the estimated energy deposited within each cell is 1.5 mJ per raster scan. Results are mean $\pm$ s.e.m. (n=3 biological replicates, ${\chi }_{\text{red}}^2=1.325$ ). (b–d) Representative fluorescence images acquired 1 hour post-ablation. Red rectangles in (b) and (c) indicate the $\sim$ 10,000 ${\mathrm{\mu m}}^2$ ablation region used for intensity quantification. (b) Exemplary image for a 50 fs pulse duration. (c) Representative image for a 170 fs pulse duration. (d) Negative control image with no ablation.

Fig. 4.

(a) Schematic representation of the 3D Ovcar5-GFP ablation experimental process. (b) Fluorescence imaging of 3D Ovcar5-GFP cell cultures demonstrates the depth-resolved localized ablation enabled by ultrashort laser pulses. All ablation was performed using pulses delivered at 32 MHz with a surface pulse energy of 6 nJ (n=4 replicates, 2 biological replicates). Ablation laser scan number was increased to maintain effective ablation at depth. Arrows pointing to the right indicate cells that were successfully ablated, as evidenced by PI uptake. Arrowheads pointing to the left identify PI-positive cells that were not targeted by the laser and represent basal levels of cell death in the model. Spheroids shown exhibited a mean diameter of 82.7 $\pm$ 4.4 µm (n=16, mean $\pm$ s.e.m.) (i) Zoomed-in maximum intensity projection of an ablated spheroid $\sim$ 270 µm deep, showing PI-positive cells extending to a depth of approximately 75 µm within an individual tumor nodule. The red square denotes the 100 µm $\times$ 100 µm ablation region. (ii) Zoomed-in projection of a neighboring non-ablated control spheroid showing no evidence of laser-induced damage.

Fig. 5.

Combined GFP + PI dye intensity profiles (left column) and the isolated PI dye intensity profiles (right column) along the depth of the ablated spheroid for a three of the spheroids seen in Fig. 4. The maximum one percent of PI dye intensities across each z-slice (slice width = 9.18 µm) were plotted and fit to the axial gaussian beam intensity profile described in Eq. (14). a) Z-profile of the GFP+PI intensity and isolated PI intensity along with the associated axial intensity fit for the spheroid starting at a depth of 200 µm. This spheroid was subject to two planar ablation scans. b) Z-profile of the GFP+PI intensity and isolated PI intensity along with the associated axial intensity fit for the spheroid starting at a depth of 230 µm. This spheroid was subject to a single planar ablation scan. c) Z-profile of the GFP+PI intensity and isolated PI intensity along with the associated axial intensity fit for the spheroid starting at a depth of 370 µm. This spheroid was subject to two planar ablation scans.