1. Introduction

Femtosecond micromachining has been utilized to create highly localized microscopic structures inside materials, including three-dimensional microfabrication within silica glasses [1,2], buried tubular waveguides written in bulk poly(methyl methacrylate) [3,4], and channel waveguides or diffraction gratings formed in cross-linkable PMMA-based copolymers [5,6], owing to its unique characteristics, such as rapid and precise energy deposition into the materials, elimination of thermal diffusion to the surrounding area, structural modification within the focal volume and etcetera [7,8]. Over the last decade, this technique has been applied to alter the optical properties of ophthalmic materials, such as hydrogel-based contact lenses [9,10], intra-ocular lenses (IOL) [11], and cornea tissues [12] by creating different refractive index shaping structures. Gandara-Montano et al. successfully wrote arbitrary Zernike polynomials in hydrogel-based contact lenses [9]. A phase-wrapped lens was written directly inside an IOL to alter the hydrophilicity of targeted areas [11], and also high-quality gradient-index (GRIN) Fresnel lenses with wide power range from −3.0 to + 1.5 diopters were written in plano contact lens materials [13]. NIR/Blue femtosecond lasers have been used for noninvasive Intra-Tissue Refractive Index Shaping (IRIS) inside corneal tissue and even in the eyes of living cats [14]. There are many factors that can determine the effectiveness of the laser-induced refractive index change (LIRIC), including both material properties and laser exposure parameters. Material properties including chemical composition, water content, dopant type or dopant concentration have been studied so as to extend the limit of maximum achievable refractive index change [15–17]. The Raman spectrum signature arising from O-H stretching vibration of water centered at 3420 cm⁻¹ shows that larger phase change is associated with higher water content in the modified region and the lower refractive index of water compared to the studied hydrogels may account for the negative phase change induced in the laser-treated area [16]. The function of water in the femtosecond micromachining process may been considered as increasing the heat-induced breakdown threshold due to the high heat capacity, slowing down the heat dissipation from the laser-treated area into the surrounding due to low heat diffusivity [17], or facilitating the photo-induced hydrolysis of polymeric materials in aqueous media [18]. As for the chemical composition needed for femtosecond writing, there are two active components, a dopant and a quencher, to increase the energy deposited via nonlinear absorption and to transfer the energy absorbed for the chemical reaction to take place [16]. Doped hydrogels have been reported to produce much larger RI change than undoped hydrogels because of enhanced nonlinearity absorption coefficient [15].

Apart from material properties, the induced RI change is also highly dependent on the system parameters and the effects of many system parameters, such as the scan speed, the laser average power, the numerical aperture (NA), the pulse width, the laser beam wavelength, the number of layers written in the same area, have been widely studied so as to optimize the femtosecond micromachining process [17]. It has been reported that larger refractive index change can be induced in hydrogel polymers with a lower scan speed, a higher average power, and a shorter laser wavelength. However, to apply the femtosecond micromachining process in the field of vision correction, we need to care about the laser safety issue and thus the power used in the micromachining process should be properly chosen due to the limitation of the maximum permissible exposure (as power or energy density) at the cornea. Savage et al. have demonstrated the successful inscription of refractive structures into the stromal layer of living cat eyes by using femtosecond pulses at 400 nm, with a pulse duration of 100 fs and a power at 60 mW [14]. However, it is worth noting that the scan speed used in the writing was as low as 30 mm/s and only a small GRIN lens of 2.5 by 2.5 mm² was written into the cat eye, which render the technique less practical in clinic application. In order to enhance the feasibility and practicability of the femtosecond micromachining in vision correction, we rely on the laser pulse repetition rate to reduce the average power and to enhance the scan speed needed for inducing detectable RI change inside ophthalmic materials. There are several research groups that have reported the effect of the repetition rate on the microstructure formation inside fused silica and metals with the range of Hz and KHz. Most of the studies are focused on the effect of repetition rate on ablation efficiency [19,20], laser drilling of metals [21,22], defects formation [23], surface texture of the laser induced microstructures [24]. A comparison between kilohertz and megahertz laser systems made by Reichman et al. showed that megahertz repetition rate modification resulted in better quality waveguides with a greater refractive index increase than kilohertz repetition rate modification [23]. However, to our best knowledge, no studies so far have illustrated the effect of repetition rate on the RI change induced by femtosecond micromachining in hydrogel-based polymers. The typical repetition rate currently used in femtosecond micromachining hydrogel polymers is larger than 80 MHz. However, it is uncertain whether the high repetition rate (> 80 MHz) is the optimal repetition rate for writing ophthalmic materials without inflicting any optical damage. Given the same single pulse energy, the accumulated temperature rise induced by a high-repetition-rate pulse train is higher than a low-repetition-rate pulse train [25–27], indicating that the high repetition rate of the pulse train should be able to induce larger refractive index change if we assume a pyrolytic degradation of polymers. However, it is possible that the laser-induced RI change results from a combination of both thermal-induced decomposition process and direct photochemical changes made by the nonlinear multiphoton absorption process. Although the accumulated thermal effect might be more significant in high repetition rate region, the photon energy that can be absorbed by the material is diminished due to a smaller single pulse energy for the same amount of average power; on the other hand, the low-repetition-rate pulses may easily reach the optical breakdown threshold and cause gross damage at smaller average power due to a larger single pulse energy. Therefore, it is important to investigate the effect of repetition rate on both the induced RI change and the optical damage threshold. We also expect the optimal repetition rate can be taken advantage of to reduce the amount of power required and speed up the femtosecond machining process.

In this paper, we present both the qualitative and quantitative results showing how the repetition rate affects the magnitude of the induced phase change, which is related to RI change via the following equation,

2. Experimental setup

The system configuration is depicted in Fig. 1. The light source used for the writing process is a KM Lab Y-Fi (Ytterbium fiber) laser, which delivers femtosecond laser pulses with a pulse duration of ~120 fs, a central wavelength at 1035 ± 5 nm. The pulse duration is measured by a commercially available autocorrelator (Newport, PSCOUT2-NIR-PMT) and can be adjusted by an internal grating pair working as a pulse compressor. This powerful Y-Fi laser contains an all-normal-dispersion (ANDi) Yb doped fiber oscillator, which enables tunable repetition rates in the range from 1 MHz to 60 MHz via the usage of a pulse picker. The pulse repetition rate from the oscillator is fixed at 60 MHz, and a pulse picker extracts certain pulses from the fast pulse train with an identical temporal separation, resulting in a different repetition rate from the oscillator. The available repetition rates are 60/N MHz, where N is an integer number from 4 to 60. An amplifier is employed after the pulse picker to amplify the average power up to 30 W, corresponding to a maximum pulse energy of 0.5 μJ at 60 MHz. A HWP and a PBS are placed after the Y-Fi laser to split the incident linearly polarized beam into two beams with orthogonal polarization and controllable power. The split ratio of the HWP and PBS unit is set to be 90:10, where 90% of the incoming beam transmits through the PBS and 10% of the beam is reflected onto a power meter for monitoring the power used in the writing process. A beam expander, consisting of two positive lenses, is placed after the HWP and PBS unit to amplify the beam size so that the effective numerical aperture (NA) for writing the samples is 0.47, and the corresponding diffraction-limited spot size is 2.75 μm. Three low GDD mirrors (Thorlabs, UM10-45B) are inserted in the system to fold the optical beam path and then the beam is sent through a microscope objective (Olympus, LCPLN50XIR) to form a diffraction-limited spot inside a sample. The microscope objective is attached to a motorized vertical stage (Newport, GTS30V) which provides Z axis step control. A sample is sandwiched between a microscope glass slide and a cover slip and soaked with solution to maintain hydrated. The sample is then mounted on a 2D linear translation stage (Aerotech, PRO115LM) which allows XY axial scanning with a speed up to 3 m/s. A back reflection monitor consisting of a focusing lens and a CCD camera is used to locate the writing depth which is usually set to be in the middle of the polymer sample.

 

Fig. 1 Femtosecond writing system diagram. The Yb doped fiber laser is able to deliver femtosecond pulses at a series of adjustable repetition rates and with a maximum output power up to 30 W.

Download Full Size | PDF

3. Experimental results

In order to illustrate the effect of repetition rates on the induced phase change, we conducted both qualitative and quantitative experiments at different repetition rates, with different powers and different scan speeds. Samples used in our qualitative experiments are hydrogel-based contact lenses (Acuvue2, Johnson & Johnson) made of a soft hydrophilic material known as “etafilcon A”, a copolymer of 2-hydroxyethyl methacrylate and methacrylic acid cross-linked with 1, 1, 1-trimethylol propane trimethacrylate and ethylene glycol dimethacrylate [28]. Periodic grating structures were fabricated ~50 μm below the top surface of a contact lens using a raster scanning method. Figure 2(a) and (b) show the DIC images of the gratings written with a power of ~500 mW, at three different scan speeds, 5 mm/s, 50 mm/s and 100 mm/s and at two different repetition rates of 15 MHz and 60 MHz. Each grating consists of 10 grating lines with a line spacing of 5 μm, resulting in a total width of 50 μm. In order to maintain a constant velocity inside the sample, the length of each grating is set to be 20 mm, which is ensured to be long enough to compensate for the acceleration and deceleration travel distance of the stage. Given the same average power and the same repetition rate, the grating lines became fainter as the scan speed was increased, indicating that a larger RI change can be induced using a smaller scan speed. Distinct difference can be perceived between the DIC image obtained at 15 MHz (Fig. 2(a)) and 60 MHz (Fig. 2(b)). Grating lines micro-machined at 15 MHz can be seen at all the three scan speeds while the grating lines are difficult to be detected at 60 MHz even at the lowest scan speed. Results obtained at 100 mm/s using three different powers, 500 mW, 1000mW and 1500 mW and two different repetition rates are shown in Fig. 2(c) and (d), respectively. Similar results were obtained from the power scaling experiments as from the scan speed scaling experiments. Grating lines became brighter as the power was increased until the optical breakdown threshold was reached while the repetition rate and the scan speed were kept the same. The optical damage was indicated by the formation of dark carbon spots or material distortion with melting traces and porosities highly localized along the grating lines. The results obtained at the same repetition rate are consistent with the conclusions drawn in Ding et al.’s paper [17], implying that larger RI change can be achieved by increasing the average power and lowering the scan speed. However, comparing the DIC images at two different repetition rates while keeping other irradiation parameters the same leads to several new findings. As shown by the bright central grating lines written at 15 MHz (Fig. 2(c)) and the fairly faint grating lines at 60 MHz (Fig. 2(d)), the usage of a lower repetition rate pulse train can result in a much larger refractive index change at the same average power. Lower-repetition-rate pulses also cause optical damage at a smaller average power, as indicated in the images that for a given average power at 1500 mW, carbonized grating lines could be observed at 15 MHz while minor modification was induced at 60 MHz. The noteworthy difference between 15 MHz and 60 MHz can be explained from the single pulse energy point of view. Single pulse energy as a product of the average power and pulse interval is 4 times larger at 15 MHz than at 60 MHz for the same average power, meaning more energy per pulse can be absorbed by the material through multiphoton absorption at a lower repetition rate to induce more significant amount of RI change.

 

Fig. 2 DIC image of the grating lines written at an average power of ~500 mW, at scan speeds of 5 mm/s, 50 mm/s and 100 mm/s (from top to bottom), and at repetition rates of 15 MHz (a) and 60 MHz (b), respectively; DIC images of the grating lines written at average powers of ~500 mW, ~1000 mW and ~1500 mW (from top to bottom), at a scan speed of 100 mm/s, and at repetition rates of 15 MHz (c) and 60 MHz (d), respectively.

Download Full Size | PDF

More systematic experiements were performed to generate quantitative results. We inscribed phase bars instead of grating lines inisde an Acuvue J + J contact lens with + 4 diopters and a base curvature of 8.7 mm. The travel distance of the 2D linear stage is still set to be 20 mm so as to keep the scan speed constant at 200 mm/s inside the sample. Each phase bar is made up of 30 grating lines separated by 1 μm, which is smaller than the focused spot size, thus creating a continuous phase carpet of dimensions 30 μm by 20 mm. As shown in Fig. 3(a), there are 4 phase bars written at two different exposure conditions. The top two phase bars were fabricated at 1075 mW while the bottom ones at 1105 mW. The difference between two writing conditions (as long as no damage occurs) cannot be detected directly from the DIC image but can be well spotted by measuring the phase change induced inside the written area with respect to the surroundings under a home built two-wavelength Mach-Zehnder interferometer (MZI) as described in Ref [9]. The scanning method used in the experiments is a rectangular loop scanning as shown in Fig. 3(b) rather than a raster scanning, which allows us to write side by side two same phase bars with a separation of 100 μm. Additionally, the rectangular loop scanning minimizes the spatial thermal effect induced by two adjacent grating lines since the separation between two adjacent grating lines is 100 times larger than that in the raster scanning. The amount of phase change measured with a reference light at 543 nm is extracted and unwrapped via the Goldstein’s branch cut unwrapping algorithm and Fourier-transform method of fringe pattern analysis [29,30]. The interferogram showing two phase bars written at the same irradiation parameters (used for creating the two bottom phase bars in Fig. 3(a)) and the corresponding retrieved phase map are displayed in Fig. 3(c) and (d), respectively. Obvious phase shift can be observed from the interferogram and phase change inside the written region differs from the pristine region with sharp discontinuity at the edges. The average phase change was calculated to be −0.54 waves and the variation inside one phase bar was less than 0.05 waves. The difference of the induced phase change between the top phase bar and the bottom one in Fig. 3(d) is roughly 0.02 waves, indicating that consistent and stable phase change can be achieved by the femtosecond micromachining process. The longitudinal thickness of each scanning line is estimated to be around 8 μm by taking a cross section image of one grating structure consisting of 10 scanning lines and therefore the corresponding RI change to a phase change of −0.54 waves at 543 nm is calculated to be −0.037 from Eq. (1).

 

Fig. 3 (a) DIC image of four phase bars written in a J + J contact lens at two different powers, 1075 mW (top two) and 1105 mW (bottom two); (b) Schematic diagram showing two different scanning method, raster scanning (left) and rectangular loop scanning (right); (c) Interferogram of the two phase bars written at a power of 1105 mW, a scan speed of 200 mm/s and a repetition rate of 15 MHz; (d) The phase map with tilt moved showing phase difference between the laser-treated area and the surroundings.

Download Full Size | PDF

4. Quantitative proposed photochemical model of phase changes

To map out the induced phase change at different repetition rates, the power tested started from a small value at which a small phase change can be measured by the MZI and ended at a value where the sample damage occurs. The scanning pattern was the rectangular loop scanning and the writing speed was fixed at 200 mm/s. Two materials were laser-treated for collecting quantitative data at four different repetition rates, 5 MHz, 10 MHz, 15 MHz and 60 MHz. One of the materials is an Acuvue J + J Contact lens as described in the previous section and the other one is a plano hydrogel sample named Contaflex GM Advance 58 (Contamac Inc.), which is made of “Acofilcon A”, a synonym for “2-Butenedioic acid (2Z)-, di-2-propenyl ester, polymer with 2,3-Dihydroxypropyl 2-methyl-2-propenoate, 1-Ethenyl-2-pyrrolidinone, 2-Hydroxyethyl 2-methyl-2-propenoate and Methyl 2-methyl-2-propenoate” [31]. The magnitude of the measured phase change at four repetition rates as a function of average power and single pulse energy is shown respectively in Fig. 4(a) and 4(b) for the J + J contact lens as well as in Fig. 4(c) and 4(d) for the Contaflex GM Advance 58 sample. Each experimental data was obtained by averaging the phase change values from 8 to 12 phase bars that were written with the same laser irradiation parameters. The variance assigned to each individual data shows the standard deviation obtained from these rectangles.

 

Fig. 4 Below the damage threshold, quantitative results showing the phase change magnitude measured at 543 nm for the phase bars written in a J + J Acuvue contact lens as a function of power (a) and as a function of single pulse energy (b); phase bars written in a Contaflex GM Advance 58 sample as a function of power (c) and as a function of single pulse energy (d) at different repetition rates, 5 MHz, 10 MHz, 15 MHz and 60 MHz.

Download Full Size | PDF

The fitting curves are derived from a photochemical model involving multiphoton absorption process and overwriting factor. RI change is assumed to be induced by the density change of polymer matrix in the excitation volume through multiphoton absorption. In our specific case where near infrared light at 1035 nm was used, a four photon absorption process was expected to occur during the writing process due to a strong UV linear absorption cut off wavelength around 300 nm. Then the excited molecular density change D can be expressed in the following form assuming a small absorption limit for simplicity,

Following Eq. (2) and Eq. (3), the overall excited molecule density change can be expressed as the product of the molecule density change caused by a single pulse times the overlapping effect N. After mathematica manipulation and expressing the single pulse energy in terms of average power and repetition rate, we are able to propose the final result of the phase change induced by a single scanning line,

Several conclusions can be drawn from Fig. 4. As shown in Fig. 4(a) and 4(c), the magnitude of the phase change measured at 543 nm rises as the power increases for a given repetition rate for both materials and lower repetition rates can induce the same amount of phase change at a much smaller power, which are consistent with the qualitative results shown in Fig. 2. Lower repetition rate, on the other hand, causes sample damage at a smaller power for both materials; however, the damage threshold in terms of the single pulse energy behaves differently for two materials. As shown in Fig. 4(b), the single pulse energy needed for initiating the optical breakdown in a J + J contact lens is ~82 nJ at 5 MHz, ~81 nJ at 10 MHz, ~77 nJ at 15 MHz and ~63 nJ at 60 MHz. Therefore, the experimental results indicate that the sample damage threshold of a J + J contact lens decreases as the repetition rate increases or equivalently the number of laser shots at the same spot increases. However, Fig. 4(d) shows that the single pulse energy damage threshold of a Contaflex sample is ~83 nJ at 5 MHz, ~82 nJ at 10 MHz, 84 nJ at 15 MHz and 82 nJ at 60 MHz, indicating a damage threshold independent of repetition rate. We assume the relationship between the optical breakdown threshold and the number of pulses per spot can be described by the incubation effect [33,34]. The topic about the optical breakdown is beyond the scope of this paper and we will present the details about the optical damage threshold as a function of repetition rates in our future work. Given the same single pulse energy, the magnitude of the phase change increases as the repetition rate rises, which could be attributed to the accumulation effect arising from the overlapping of two adjacent pulses. A higher-repetition-rate pulse train is able to produce a larger number of laser shots on the same spot, meaning more pulse energy could be absorbed by the material. Therefore, it is reasonable to conclude that the maximum achievable phase change just below the damage threshold also increases as the repetition rate increases by taking into account that the change in single pulse energy damage threshold is modest for the repetition rates from 5 MHz through 60 MHz. As shown in Fig. 4(b) and 4(d), the maximum achievable phase change for the J + J contact lens is ~0.38 waves at 5 MHz, ~0.6 waves at 10 MHz, ~0.7 waves at 15 MHz and ~1 wave at 60 MHz and for the Contaflex sample, ~0.17 waves at 5 MHz, ~0.26 waves at 10 MHz, ~0.33 waves at 15 MHz and ~0.56 waves at 60 MHz. The increasing pulse overlapping effect at higher repetition rates and the roughly same single pulse energy threshold at different repetition rates together contribute to the growth of the maximum achievable phase change just below the damage threshold.

5. Empirical model of phase changes

As shown in Fig. 4, the fitting curves derived from the photochemical model and the overlapping effect work well for the low repetition rates at 5 MHz, 10 MHz and 15 MHz but not for 60 MHz. There are multiple reasons to account for the deviation at 60 MHz. The deviation at high repetition rate may indicate a reduced nonlinearity order from 4 at low number of pulses per spot to around 1 at high number of pulses per spot due to the accumulated linearly absorbing color centers via the incubation reaction [35]. The formation of linearly absorbing defects were also confirmed by an increased absorbance of the laser irradiated area near the UV absorption edge after examining UV-visible transmission spectra [36]. In order to show the decrease in the nonlinearity order we plot the experimental data in a double logarithmic scale. As shown in Fig. 5, we plot the logarithm of the induced phase change at four different repetition rate versus the logarithm of the average power in the focal volume. The empirical model was obtained using linear fitting method and thus the order of the multiphoton absorption process was indicated by the linearly fitted trend line gradient. Similar to the results in Ref [36], we also observed a decreasing trend line gradient as the repetition rate increases. However, in contrast to the diffraction efficiency used in their paper, which should scale quardratically as the RI change, a direct relationship between the induced phase/RI change and the average power was obtained to determine the order of the multiphoton absorption process. For the Contaflex sample (as shown in Fig. 5(a)), the line gradient decreases from 4.15 at 5 MHz to 3.17 at 60 MHz, which indicates the multiphoton process evolves from a pure four photon absorption process at low number of pulses per spot to a lower order of multiphoton absorption process at high number of pulses per spot due to the formation of larger number of color centers. For the J + J contact lens (as shown in Fig. 5(b)), the trend line gradient follows the decreasing rule as well and slightly reduces from 3.4 at 5 MHz to 3.1 at 60 MHz. The order of multiphoton absorption process keeps decreasing and the overall power dependence is lowered by taking account into both the four photon absorption of the material and the linear absorption of the color centers.

 

Fig. 5 Plot of logarithm of the induced phase change at 543 nm versus logarithm of the average power in the focal volume at 5 MHz, 10 MHz, 15 MHz and 60 MHz for the Contaflex sample (a) and the J + J contact lens (b) for the purpose of the determination of the order of multiphoton absorption process.

Download Full Size | PDF

6. Discussion

Other than the reduced nonlinearity order at high number of pulses due to color center formation, the deviation at 60 MHz could be attributed to a much higher accumulated temperature rise achieved inside the material after a high-repetition-rate laser exposure [27,37]. The heat accumulation effect may play a more significant role at a high repetition rate domain and induce the phase change via a distinct photothermal effect which has a power dependency different from that of a photochemical effect. A third explanation for the deviation at 60 MHz could be a saturation effect at high intensities since we noticed the induced phase change cannot increase infinitively and tends to be saturated when the induced phase change reaches a high value. If the saturation factor is not introduced into the theoretical model, the fitting results would be higher than experimental results as the power and repetition rates increase. Therefore, a saturation factor can be incorporated into the photochemical model for the purpose of a better fitting. The saturation factor can be inserted under the material constant ε if we assume there is a possible limit on the finite number of available molecules, or the capability of the material to redirect the absorbed energy to induce phase change is limited or the depolymerization process is balanced by a recombination effect. On the other hand, the saturation factor can also be attached to the fourth photon absorption coefficient by analogy with the decreasing nonlinear absorption coefficient measured at high laser intensity. We then propose the modified photochemical model can be expressed in the following form by employing a classic saturation factor,

The exact mechanism associating the absorbed pulse energy to RI change is still controversial and requires further investigation. The accumulated pulse energy absorbed by the material via multiphoton absorption can excite electrons to a high electronic state to induce direct bond dissociation or the absorbed energy can be transferred from electrons to lattice to heat the polymer matrix and induce subsequent pyrolysis of the polymer. Depolymerization, either a photolysis reaction or a thermal-driven process, usually increases polymer chain volume by random chain scission and direct bond breaking, thus causing monomer formation, internal tensile stress propagation and volume expansion which is confirmed by the shift of the molecular weight to a lower value and the broadening of the polymer molecular weight distribution [36,38]. After the decomposition of polymer into smaller monomers or oligomers, we assume these smaller fragments may diffuse out from the laser-irradiated area into the surrounding and water molecules can then occupy the empty expanded area due to enhanced hydrolysis [18], thus resulting in an increasing water content in the laser treated area. Therefore, the negative sign of induced phase/RI change might be attributed to a lower refractive index of water than the sample itself and a reduced polymer density due to hydrodynamic expansion. Our group proposes taking advantage of Raman spectroscopy to gain insight into the chemical composition change in the laser-irradiated area and then determine whether the RI change in hydrogel polymers arises from a thermal-induced decomposition process via non-radiative decay or a direct chemical bond breaking procedure via multiphoton ionization or a combined photothermal and photochemical effect.

7. Conclusion

The effect of repetition rate has been demonstrated to play an important role in femtosecond micromachining process. The same amount of phase change can be achieved at a lower average power and a faser scan speed by taking advantage of low repetition rate pulses while the damage threshold in terms of single pulse energy maintians almost the same for all the tested repetition rates. Based on both the qualitative and quantitative results, we are able to conclude that the induced phase change depends on the amount of pulse energy absorbed by the material via both the multiphoton absorption and the overlapping effect induced by adjacent pulses. In order to fit the experimental data, we developed a calibration function based on a photochemical model. The induced phase change scales as the power to the fourth order as a consequence of four photon nonlinear absorption at 1035 nm. The calibration function works well for the three low repetition rates but more effort needs to be put into explaining the deviation at high repetition rates. The optimal repetition rate for femtosecond micromachining contact lenses seems to be around 15 MHz for the reason that the maximum achievable phase change just below the damage is a little bit lower than that obtained at 60 MHz but the average power required is roughly 3 times smaller. The employment of low repetition rate pulses can be useful in femtosecond micromachining corneas since we need to concern much about the maximum permissible exposure light intensity to the eye. If there is no constraint on the power, a pulse train with a repetition rate higher than 60 MHz can be used to produce even higher RI/phase change below the damage threshold. However, the optical damage energy threshold on the other hand may decrease once the repetition rate becomes too high since the heat-induced damage might play a dominant role and result in a smaller energy threshold than the damage caused by the multiphoton ionization. The optical damage mechanism for causing the hydrogel to crinkle and inducing unmeasurable phase change remains unclear and needs more investigation. The optimal repetition rate may vary as the sample changes. The results showing the repetition rate dependence on micromachining rabbit cornea tissues will be presented elsewhere. We expect that the usage of a low repetition rate pulsed laser would be a much safer light source for performing the intra-tissue refractive index shaping in corneal tissues.