Different laser beam sources are used for the experiments, emitting sech² (picosecond laser) and gaussian (nanosecond lasers) shaped pulses with pulse widths between 12 ps and 220 ns and pulse energies between 30 μJ and 10 mJ, optionally at 1064 nm and 532 nm wavelength. Here, the nanosecond lasers are used as reference laser sources for state-of-the-art results. For all beam sources, the pulse energy can be adjusted manually with a polarizing beam splitter and a wave-plate up to the specific maximum. The following list describes the laser sources in more detail:

Laser 1: USP laser source with 12 ps pulse width: USP laser system consisting of a passively mode-locked Nd:YVO₄ (neodymium-doped yttrium orthovanadate) oscillator and an Nd:YVO₄ regenerative amplifier (37). The laser system emits a 1,064 nm pulse train with an adjustable pulse repetition rate between 1 Hz and 300 kHz, a pulse energy of up to 164 μJ, and a pulse width of 12 ps. The laser output can additionally be frequency doubled with a KTP (potassium titanyl phosphate—KTiOPO₄) nonlinear crystal, leading to a wavelength of 532 nm with a pulse energy conversion efficiency of up to 70%.

Laser 2: ns laser source with 1.6 ns pulse width: 1.6 ns, 20 Hz quasi-continuous-wave (QCW) passively Q-switched (PQS) Nd:YAG (neodymium-doped yttrium aluminum garnet) laser source with pulse energies up to 7 mJ at 1064 nm from MONTFORT Laser GmbH, Götzis, Austria.

Laser 3: ns laser source with 8 ns pulse width: 8 ns, 10 Hz actively Q-switched laser source with pulse energies up to 50 mJ at 1064 nm from MONTFORT Laser GmbH, Götzis, Austria. The laser output can additionally be frequency doubled with a KTP nonlinear crystal, leading to a wavelength of 532 nm.

Laser 4: ns laser source with adjustable pulse width: 20 W, 1062 nm pulsed fiber master oscillator power amplifier (MOPA) laser from JENOPTIK Laser GmbH, Jena, Germany (‘JenLas^®fiber ns 20 advanced single mode’). It emits in continuous wave (cw) operation or pulsed operation with pulse widths between 15 ns and 220 ns with maximum pulse energies between 68 and 570 μJ, respectively.

In contrast to ophthalmic surgical methods with special eye attachment optics/system, the different laser beams are focused on the samples via the following beam delivery system: This system consists of a 60 mm plano convex focusing lens, an x-y-z-translation stage for sample positioning, two laser diodes for laser focus alignment (FP-D-532-1-C-F from Laser Components GmbH, Olching, Germany), and a stereo microscope as visualization device for real-time observation and magnification (Stemi 508 trino from Carl Zeiss Microscopy GmbH, Oberkochen, Germany). This set-up is based on slit-lamp laser beam delivery systems such as used in clinics and is similar in function. For part of the capsulotomies and SLT, a laser scanner set-up is used, consisting of a two-axis laser scanner head (SS-II-15 [Y] D2 from Raylase GmbH, Wessling, Germany) with the F-Theta flat field objective S4LFT3046/328 (Sill Optics GmbH & Co. KG, Wendelstein, Germany) with a focal length of 50 mm. Since these components of the set-ups are designed for the near infrared (NIR), a second set-up for 532 nm was built, using another 60 mm plano convex focusing lens and the scanner head SS-IIE-15 [DY] D2 (Raylase GmbH, Wessling, Germany) with the F-Theta flat field objective S4LFT4066/292 (Sill Optics GmbH & Co. KG, Wendelstein, Germany) with a focal length of 67.2 mm. Focusing results in a focal spot diameter of about 25 μm for all focusing set-ups. This leads to fluences of 8 J/cm² (12 ps, 50 μJ) and 800 J/cm² (8 ns, 5 mJ), and peak intensities of 587 GW/cm² and 94 GW/cm², respectively.

The tissue samples used are post mortem porcine eyes, which are processed about 4 h after excision. The eyes are moistened with water before, during and after all surgeries. The iridotomy and capsulotomy surgeries were performed as open sky surgeries (with prior cornea removal) for better comparability without the influence of the cornea. SLT surgeries were performed in two ways: First, on a bisected porcine eye (removed lens) with the slit-lamp set-up to focus the laser beams directly on the trabecular meshwork (TM) without the necessity of a clinically used gonioscope. Second, on an exposed trabecular meshwork with the scanner set-up. For this, the eyes were prepared in such a way that the cornea, about 2 mm of the sclera around the cornea, and the TM were preserved. This is comparable to pieces of human donor corneas, where a part of the sclera and the TM are also still attached. The lasers can then be focused directly on the TM from the posterior side. Visualization, examination and measurement of the results are carried out using high-resolution light microscopy and scanning electron microscopy (SEM). For the latter, a special sample preparation and drying method is applied to the tissue samples that avoid water inside the tissue to boil and expand under vacuum in the plasma coating machine and the SEM. This would lead to warpage/deformation/distortion and destruction of the tissue. The preparation method starts with 1,5-pentanedial (glutaraldehyde solution 50% in water (5.6 mol); Sigma-Aldrich, St. Louis, United States) for protein linkage and tissue fixation. The tissue is then buffered with a phosphate buffer of pH = 7.0 (Bernd Kraft, Duisburg, Germany) and subsequently placed in an alcoholic solution in ascending concentration (start: 20%/80% ethanol/water; end: 100% ethanol; gradation: 10% ethanol—water: purified water; ethanol: Ethanol absolute ≥99.5% from VWR International, Radnor, United States). The last step before air drying is soaking the tissue in 1,1,1-Trimethyl-N-(trimethylsilyl)silanamine (hexamethyldisilazane, HMDS; Sigma-Aldrich, St. Louis, United States). The samples of the exposed TM for SLT are soaked in pure ethanol only and subsequently air dried, because the cornea and sclera lead to a high stiffness even without fixation. Some of the capsule samples are strictly air dried during compression to avoid warpage/deformation/distortion.

All samples can then be metal-coated (metal: gold/palladium—SC502-314B from Quorum Technologies, Laughton, United Kingdom) and visualized by SEM. For this, the plasma coating machine Q150R ES (Quorum Technologies, Laughton, United Kingdom) and the SEM JSM 6510 (Jeol Ltd., Tokyo, Japan) are used. Light microscope images were taken using the incident light microscope Axio Lab.A1, the transmitted light microscope Primo Star and the stereo microscope Stemi 508 trino, all from Carl Zeiss AG, Oberkochen, Germany.

In total, the research was performed on about 400 porcine eyes with over 1,000 surgeries (approx. 800 iridotomies, 25 capsulotomies, 200 SLTs).

Shock waves generated by laser pulses were investigated using the following method and set-up (see Figure 1): The different laser sources are focused via a laser mirror and an f = 40 mm aspherical plano-convex lens in a basin (74 mm × 74 mm × 33 mm; l × w × h) filled with distilled water. Focusing results in a diameter of about 10 μm, leading to fluences of 50 J/cm² (12 ps, 50 μJ) and 5 kJ/cm² (8 ns, 5 mJ), and peak intensities of 5.5 TW/cm² and 587 GW/cm², respectively. The laser input results in an optical breakdown and shock waves, which are visualized using a streak camera system. The field of observation has a diameter of approximately 20 mm, roughly corresponding to the distance between laser focus and retina.

The light source of the system, a collimated pulsed green laser diode PLT5 510 B1-3 (OSRAM GmbH, München, Germany), is synchronized to the pulse repetition rate of the breakdown laser. The phase delay between the laser pulse and the illumination pulse can be adjusted manually with a ns-increment to illuminate different shock wave propagation points. Both synchronization and delay adjustment are performed with a waveform generator (33500B from Agilent Technologies, Inc., Santa Clara, United States). Recording is done with a digital camera (RX100 from Sony, Tokyo, Japan). In addition to the photographs, a 2 mm diameter needle hydrophone with attenuator, pre-amplifier and DC-coupler (Precision Acoustics Ltd., Dorchester, United Kingdom) is placed next to the breakdown below the water surface. It serves as a receiver of the shock wave fronts. For visualization and measurement, it is connected to an oscilloscope (MSO64 from Tektronix, Inc., Beaverton, United States, 6 GHz bandwidth). The result of this measurement is a quantitative intensity profile of the shock waves as a function of the laser parameters.

For cavitation bubble investigation, the streak camera set-up was modified (see Figure 2): The water bath was replaced by a 2.5 mL polystyrene macro cuvette and the illumination source was replaced by a 1,062 nm cw laser source (Laser 4, see Sec. 2.1), triggered and modulated with a Pockels cell. The camera was changed to the system camera VEXU-24 M (Baumer Electric AG, Frauenfeld, Switzerland). Both the Pockels cell and the camera are triggered by the waveform generator to the pulse repetition rate of the breakdown laser. This set-up provides a minimum illumination time of 5 ns, which is limited by the switching time of the Pockels cell.

The heat input of the different laser sources was investigated using the thermal imaging camera optris Xi 400 (Optris GmbH, Berlin, Germany). The different lasers were focused via the laser scanner set-up either on black anodized aluminum plates or on porcine eye irises. The scanner scribed a 5 mm long line with 10 mm/s. The emissivity was set to to ε_eye = 0.98 for the iris tissue (equal to water and skin) and ε_alu = 0.67 for anodized aluminum (38). The measurements are then compared to simulations computed with MATLAB (The MathWorks, Inc., Natick, United States) following the approach of Ansari et al. (39).

The study revealed a major variation in size, shape, quality and reproducibility of iridotomies with different pulse widths. A comparison of four different pulse widths (12 ps, 8 ns, 15 ns, 220 ns) is shown in Figure 3. The images show typical results for each pulse width. In each case, the SEM magnification factor is 500 and the scale bars are 50 μm. While results (A) and (B) for 12 ps, 1 kHz at 1064 nm and 532 nm are very good, showing small and well-defined holes, the nanosecond results (C)–(F) with different pulse lengths, pulse energies, repetition rates, and wavelengths show frayed edges, ejection of ablated material, and large areas of altered surrounding tissue. The latter is attributed to the higher temperatures incurred through these pulse widths. Furthermore, Figure 4G shows an ideal result, which was achieved with the 12 ps USP laser source, with 40 μJ pulse energy, 1 kHz repetition rate, and 1 s irradiation time. Here, the quality is almost perfect: The ablated hole has very sharp edges and appears almost as if punched. The scale bar is also 50 μm but the SEM magnification factor is 230.

In addition to the far better defined laser processing, the pulse energy necessary for piercing the iris could be reduced to several 10 of microjoules (between 30 μJ and 70 μJ) when using USP laser pulses compared to several millijoules (1 mJ to 10 mJ) when using ns laser sources. To achieve complete penetration of the iris, approximately 1,000 USP pulses, dependent on the pulse energy, are required (and thus an exposure time of 1 s at 1 kHz repetition rate). The ablation per pulse is considerably lower than with nanosecond pulses, where a few pulses (with 5 mJ pulse energy) are sufficient to achieve complete penetration of the iris.

With ps pulses, well-defined capsulotomies can be performed in addition to iridotomies. As seen in Figure 4, precise cutting (left) and piercing (right) of the anterior lens capsule of porcine eyes is possible with 12 ps laser pulses—without tearing the remaining capsule. Similar to iridotomy, pulse energies between 30 μJ and 70 μJ prove to be suitable.

Figure 5 shows a comparison of capsulotomies with (A) 8 ns, 3 mJ and (B) 12 ps, 50 μJ. Both capsulotomies are performed on dissected anterior lens capsules. The lasers were focused on the capsules via the laser scanner head set-up. The capsulotomy diameter was set to 4 mm and the laser processing speed was set to 2.5 mm/s, resulting in a spot spacing of 5 μm for the 1 kHz ultrashort pulse train and 250 μm for the 10 Hz, 8 ns pulse train. Due to high energy input with accompanying high shock impact of the ns laser pulses, the processing speed could not be reduced to achieve the same spot spacing as for the USP pulses, since the capsule would have been torn right away. A comparison of the images reveals a big difference in capsules shape and cut quality. While the ns capsulotomy is torn and has frayed edges due to the high pulse energy required for these pulse widths, the USP capsulotomy is almost perfectly round with sharp and smooth cutting edges. The lower images (C) and (D) show higher magnifications of the respective capsulotomies cutting edges, visualized by light microscopy. In (C) the cutting edge is marked with a black arrow, since the capsule is folded over due to the high shock input. Additionally, Figure 5E shows an SEM visualization of the cutting edge of (B) with a magnification factor of 1,500. Here, the edge is very sharp and flat, which is comparable to state-of-the-art femtosecond and manual capsulotomies (19, 23).

Similar to the results of the iridotomy and the capsulotomy, SLT is possible with the 12 ps USP laser pulses at much lower pulse energies compared to ns pulses, namely 30 μJ to 70 μJ instead of 1 mJ (40). A result with 12 ps, 50 μJ at 1064 nm is shown in Figure 6, where “champagne bubbles” can be seen to the left of the green target laser (marked with black arrow). These bubbles appeared during the surgery and are considered an indicator of successful treatment. All tested pulse energies between 30 μJ and 70 μJ led to comparable results.

Figure 7 shows four typical results of SLTs in exposed trabecular meshwork with (A) picosecond ultrashort laser pulses and (B)–(D) nanosecond laser pulses. All SEM visualizations were taken with a magnification factor of 200. As with the other surgeries, the differences between USP and ns pulses are significant. Regardless of wavelength and pulse energy, 12 ps produce a small hole with sharp edges and clear boundaries. In contrast, all results with nanosecond pulses show large holes with poor quality. Furthermore, the heat affected zones are strongly pronounced around the holes.

Figure 8 shows the visualization of a laser induced shock wave in water—real image on the left and false color image on the right. The laser induced breakdown leads to a shock wave, which is reflected on a steel target. Here, the breakdown was induced with 147 μJ, 12 ps at 1064 nm from the top. Figure 9 shows a color-inverted photograph of a laser induced breakdown (dark spot) with 161 μJ, 12 ps at 1064 nm together with a cavitation bubble (brighter cycle) and a typical jet formation to the left of the bubble.

The shock wave pressure was measured with a hydrophone. Figure 10 shows a typical hydrophone measurement visualized with an oscilloscope. Peak 1 is the initial breakdown with expansion of the cavitation bubble—this is the shock wave of the laser induced breakdown (LIB). Peak 2 is the signal from the first bubble collapse and peak 3 from the second bubble collapse. Dependent on the breakdown energy, and therefore on the pulse energy, the cavitation bubble oscillation of expansion and collapse repeats more often. Here, the pulse energy of 5 mJ at a pulse width of 1.6 ns led to three (detectable) shock waves.

The needle hydrophone measurements are shown in Figure 11 (logarithmic scale due to major differences in amplitude). Here, the normalized hydrophone signal (voltage) U_hyd is plotted for 12 ps and 1.6 ns pulse width as a function of the pulse energy E_pulse and the hydrophone distance d_hyd to the focal spot. The pulse energies were chosen accordingly to the results from the ophthalmic surgeries, as 50 μJ are typical for USP and 3 mJ and 6 mJ are typical for ns laser sources. 100 μJ at 12 ps are measured to gain information about energy dependency at a fixed pulse width. As can be seen, there are large differences of a factor of 25 in average for ns/USP lasers between the different parameters.

To determine the actual pressure of the shock waves, the hydrophone voltages are converted to pressure values via spectrum analysis. In Figure 12, the LIB shock wave and the first collapse shock wave pressure are plotted for different pulse energies. The pulse width is 12 ps and the distance between the hydrophone and the breakdown area is 6 mm. This measurement results in higher pressures for higher pulse energies with a maximum pressure of 0.6 MPa at 161 μJ pulse energy.

A similar measurement is made for pulse energies between 0.2 mJ and 5 mJ at 1.6 ns pulse width (Figure 13). The distance between the needle hydrophone and the laser induced breakdown is 17 mm. The measurement was done for the LIB, the 1^st and the 2^nd collapse shock wave. It results in much stronger shock wave pressures compared to 12 ps pulses—even for longer distance. For example, a 50 μJ, 12 ps laser pulse leads to a maximum pressure of 0.25 MPa while a 5 mJ, 1.6 ns laser pulse generates 37 MPa.

Large differences between treatment with picosecond and nanosecond pulses also exist with respect to heat input and transfer around the laser focus (see Figure 14). On the left, the test is shown on porcine eyes for pulse widths of 12 ps, 8 ns, 15 ns and 220 ns as used for ophthalmic surgeries (Figures 14A,C,E,G), on the right, for comparison on black anodized aluminum (Figures 14B,D,F,H). As mentioned in the methods (see Sec. 2.4), the lasers scribed a 5 mm long line with 10 mm/s. The respective temperature before irradiation P, maximum temperature during irradiation M, and heating difference H in °C are indicated. Results (A) and (C), as well as (B) and (D) are the same, with different temperature scales for both better visibility and comparability to the nanosecond results. While 12 ps only lead to a marginal heating of a few degrees which is not even recognizable in the images (C) and (D), both 15 ns and 220 ns lead to a heating of more than 150°C. Again, the pulse energies for the various pulse widths were selected as required for the treatment.

Figure 15 presents simulations of the temperature change on a water surface for different pulse widths. The simulations are calculated for one single pulse each, so that the heat accumulation for several pulses has no effect. As displayed in Figure 15 left (temperature change as function of pulse energy), there is a large difference for pulse widths of 12 ps, 100 ps, 1 ns, 15 ns and 220 ns: The longer the pulse width and the higher the pulse energy, the higher the temperature change. As expected, utilization of low energy ultrashort pulses results in much lower heat inputs than high energy nanosecond pulses (2.5 K change for 12 ps, 50 μJ and 250 K change for 1 ns, 570 μJ). This simulation in general agrees with the experimental results. It should be noted that the simulation reflects the maximum temperature at the center of the laser focus, while the measurement averages over a larger area of space (about 0.15 mm²). On the right in Figure 15, the temperature change is plotted as a function of distance to the laser spot. Here, 12 ps, 50 μJ pulses lead to a heat-affected zone of about 300 μm, while it is about 900 μm for 220 ns, 50 μJ and even larger for 220 ns, 570 μJ. Pulses with 8 ns, 3 mJ show a similar heat-affected zone than 12 ps. However, the much higher temperature change of 1800 K (8 ns) compared to 2.5 K (12 ps) would affect the surrounding tissue, e.g., due to the aforementioned stronger shock waves and instant evaporation.

Whereas the impact of a single ps laser pulse is significantly lower compared to nanosecond pulses, the question remains whether or not this holds true for the required pulse train of some hundreds to thousand pulses. The thermo mechanical response of tissue to pulsed irradiation was investigated by, e.g., Vogel et al. (44) and Yakovlev et al. (45). According to these studies typical thermal relaxation times are in the range of a few microseconds. In comparison to the temporal distance of consequent pulses, which is in the millisecond range for ps lasers, this relaxation time ensures temperature and stress confinement in any case.