High-throughput in-volume processing in glass with isotropic spatial resolutions in three dimensions

We report on fabrication of three dimensional (3D) microstructures in glass with isotropic spatial resolutions. To achieve high throughput fabrication, we expand the focal spot size with a low-numerical-aperture lens, which naturally results in a degraded axial resolution. We solve the problem with simultaneous spatial temporal focusing which leads to an isotropic laser-affected volume with a spatial resolution of ~100 micron.


Introduction
Femtosecond laser provides a unique tool for processing inside transparent materials [1][2]. The enabling approach is three dimensional (3D) femtosecond laser direct writing (FLDW) which is conducted by scanning tightly focused spots in various transparent materials including glass, crystals, polymers, bio-tissues, etc to produce arbitrary 3D structures [3][4][5][6][7][8][9]. Interaction of the focused femtosecond laser pulses with transparent materials leads to modifications of the materials properties ranging from optical refractive index, density, and Young's module to compositional distribution and chemical stability. Importantly, the space-selective modifications enabled by the 3D FLDW provides a spatial resolution comparable to or even exceed that allowed by diffraction limit, thanks to the nonlinear dependence of the modifications in materials on the peak intensity of the femtosecond laser pulses as well as the suppressed thermal diffusion with the ultrashort pulse durations [10,11]. Therefore, 3D FLDW offers very high spatial resolution to construct fine 3D structures through two-photon polymerization or internal processing of glasses or crystals.
One of the challenging issues in 3D FLDW is its relatively low throughput as compared to the conventional photolithography, because it relies on scan of the tiny focal spots produced by high numerical aperture (NA) objective lenses. For some 3D processing applications such as building mechanical molds, high throughput is highly in demand to bring down the fabrication cost at a reasonably sacrificed spatial resolution. However, if one attempts to accomplish this merely by using focal lenses of low NAs, the focal spot will become highly asymmetric, as the focal depth increases quadratically with the increasing NA, whereas the transverse size of the focal spot only increases linearly with the increasing NA [12]. This causes a severe issue in achieving high throughput 3D FLDW which is in favor of using low-NA focal lenses whereas maintaining a nearly isotropic resolution in the modification by the focused femtosecond laser pulses.
In this work, we attempt to overcome the above-mentioned problem with a simultaneous spatial temporal focusing (SSTF) scheme [13][14][15]. The working mechanism of SSTF is to first spatially chirp the incident pulse before the focal lens, forming an array of beamlets at different frequencies. This leads to the elongation of the pulse duration due to the reduction of the bandwidth of each beamlet. After passing through the focal lens, all of the beamlets will be recombined at the focal plane, where the original broadband spectrum of the incident femtosecond pulse is restored to give rise to the shortest pulses. In such a manner, the peak intensity of the spatiotemporally focused pulses is strongly localized near the geometrical focus, i.e., the longitudinal resolution is improved in the interaction of the spatiotemporally focused pulses with transparent materials. In the first implementation of the SSTF scheme in 3D FLDW, F. He et al has demonstrated an isotropic fabrication resolution of a few micron using an objective lens of NA=0.46 [13]. In the current work, we demonstrate high throughput FLDW of complex 3D structures in glass with an isotropic resolution of ~100 m. Although our 3D structure is produced in glass, this technique can generally be extended to 3D processing of other types of materials such as two-photon polymerization or ablation of bio-tissue.

Experimental
The glass material used in the experiment is a well which is composed of a lithium aluminosilicate doped with trace amounts of silver and cerium [16]. During the exposure to the femtosecond laser, free electrons multiphoton ionization to treatment, silver atoms diffuse to form nanoparticles. Due to the plasmon resonance scattering of the silver nanoparticles, the laser modified area appears bro 3D intensity distribution of femtosecond pulses at the focus can be recorded and inspected under the microscope [15]. I recording medium. In this mm × 5 mm × 1.5 mm coupons with the all six sides polished of simultaneously spatio  Figure 1 schematically illustrates t (Libra-HE, Coherent Inc.) used in this experiment consists of a Ti: sapphire laser oscillator and amplifier, and a grating with a spectral bandwidth of In the SSTF, the compressor was bypassed first reduced to ∼0.55 mm ( mm) and a concave lens ( grating compressor, consisting of two angle of 53°. The distance between the two compensate for the temporal dispersion of pulses. After being dispersed by the grating pair, the laser beam was measured to be ~40 mm (1/e the y-axis. The spatially disper an objective lens (Leica, 2 ×, NA =0.35) with a focal length of had reduced our beam diameter to under filled. The effective NA estimated in this circumstance was only 0.013. experiment, the average power of laser beam was controlled using a variable neutral density (ND) filter VF. The glass samples can be arbitrarily translated three dimensionally by PC-controlled XYZ stage with a resolution of 1 µm. by a CCD camera.

Results and discussions
The glass material used in the experiment is a well-known photosensitive glass, Foturan, which is composed of a lithium aluminosilicate doped with trace amounts of silver and . During the exposure to the femtosecond laser, free electrons are multiphoton ionization to reduce the silver ions to silver atoms. By a subsequent heat treatment, silver atoms diffuse to form nanoparticles. Due to the plasmon resonance scattering of the silver nanoparticles, the laser modified area appears brown [17,18]. With this glass, the 3D intensity distribution of femtosecond pulses at the focus can be recorded and inspected [15]. In this sense, Foturan glass can be regarded as a 3D photographic recording medium. In this experiment, commercially available Foturan glass mm × 5 mm × 1.5 mm coupons with the all six sides polished to record the 3D spatial profile of simultaneously spatio-temporally focused spot. Figure 1 schematically illustrates the experimental setup. The femtosecond laser system HE, Coherent Inc.) used in this experiment consists of a Ti: sapphire laser oscillator and amplifier, and a grating-based stretcher and compressor that delivers 3.5 mJ, 50 fs pulses bandwidth of ∼26 nm centered at 800 nm wavelength at 1 kHz repetition rate. , the compressor was bypassed and the diameter of the amplified laser beam was 0.55 mm (1/e 2 ) using a telescope system consisting of a convex lens ( mm) and a concave lens (f = −50 mm). The beam were then directed through the single grating compressor, consisting of two σ = 1500 grooves/mm gratings, blazed for the incident . The distance between the two gratings was adjusted to be ~730 mm to compensate for the temporal dispersion of pulses. After being dispersed by the grating pair, the laser beam was measured to be ~40 mm (1/e 2 ) along the x-axis and ~0.55 mm (1/e The spatially dispersed laser pulses were then focused into the glass an objective lens (Leica, 2 ×, NA =0.35) with a focal length of f = 40 mm. However, since we had reduced our beam diameter to ~0.55 mm (1/e 2 ), the back aperture of the lens was severely illed. The effective NA estimated in this circumstance was only 0.013. the average power of laser beam was controlled using a variable neutral density he glass samples can be arbitrarily translated three dimensionally by controlled XYZ stage with a resolution of 1 µm. The machining process

Results and discussions
known photosensitive glass, Foturan, which is composed of a lithium aluminosilicate doped with trace amounts of silver and are generated by reduce the silver ions to silver atoms. By a subsequent heat treatment, silver atoms diffuse to form nanoparticles. Due to the plasmon resonance scattering . With this glass, the 3D intensity distribution of femtosecond pulses at the focus can be recorded and inspected , Foturan glass can be regarded as a 3D photographic ommercially available Foturan glass was cut into 5 to record the 3D spatial profile variable neutral density filter. G1-2: which are described in the main he experimental setup. The femtosecond laser system HE, Coherent Inc.) used in this experiment consists of a Ti: sapphire laser oscillator based stretcher and compressor that delivers 3.5 mJ, 50 fs pulses 26 nm centered at 800 nm wavelength at 1 kHz repetition rate. the diameter of the amplified laser beam was ) using a telescope system consisting of a convex lens (f = 800 The beam were then directed through the single-pass = 1500 grooves/mm gratings, blazed for the incident gratings was adjusted to be ~730 mm to compensate for the temporal dispersion of pulses. After being dispersed by the grating pair, axis and ~0.55 mm (1/e 2 ) along glass sample using However, since we , the back aperture of the lens was severely illed. The effective NA estimated in this circumstance was only 0.013. In our the average power of laser beam was controlled using a variable neutral density he glass samples can be arbitrarily translated three dimensionally by a The machining process was monitored We inscribed parallel lines in both the X SSTF scheme. The writing from 60 to 80 mW (measured before the grating pair). Under the above irradiation conditions, no visible modifications of the glass could be observed under the optical microscope after the femtosecond laser irradiation. Then the coupon was subjected to a programmed heat treatment. The temperature was first ramped from room temperature to 500 held at 500 • C for 1 hr; then it was raised to 605 the samples were naturally cooled down to room temperature, the laser modified area appears brown. We polished the samples and examined the modified areas using the optical microscope. Figure 2(a-c) show the optical micro setting the average laser power lines are all approximately the samples and polishing their facets lines, which are almost round the fabrication parameters unchanged. The optical micro 2 (g-i) whilst the corresponding observed clearly that all circular cross section with a diameter of 85 was set at 60 mW, a perfectly isotropic resolution in all XYZ axes was achieved To further examine the maximum laser direct writing speed fabrication resolution of various speeds ranging from 4 mm/s to 8 mm/s mW. Figure 3(a-d) show the writing speeds at 4 mm/s, 5 mm/s, 6 mm/s an cross-sectional profiles of the lines in lines all appear symmetric approximately 85 m. However is higher than 6 mm/s ( overlap between consecutive pulses during the laser direct writing femtosecond laser operates at 1 kHz repetition rate. Therefore, higher writing speeds can be i) Top-view and (d-f, j-l) cross-sectional-view optical micrographs of lines inscribed in Foturan glass along x axis (a-f) and z axis (g-i) at different laser power 80 mW in the first row (a, d, g, j), 70 mW in the second row (b, e, h, k) and 60 in the third row (c, f, i, l). Figure 2 shows various lines inscribed in Foturan glass along different scan directions at different laser intensities for individually examining the spatial resolutions along XYZ axes. We inscribed parallel lines in both the X-and Y-directions in Foturan glass samples with the SSTF scheme. The writing speed was fixed at 5 mm/s and the average laser power varied from 60 to 80 mW (measured before the grating pair). Under the above irradiation conditions, no visible modifications of the glass could be observed under the optical microscope after the cond laser irradiation. Then the coupon was subjected to a programmed heat treatment. The temperature was first ramped from room temperature to 500 • C at 5 ; then it was raised to 605 • C at 5 • C/min and again held for the samples were naturally cooled down to room temperature, the laser modified area appears brown. We polished the samples and examined the modified areas using the optical c) show the optical micrographs of lines oriented along x direction average laser power at 80 mW, 70 mW and 60 mW, respectively. approximately 85 um. We examined the cross sections of these lines s and polishing their facets. Figure 2(d-f) show the cross sectional , which are almost round-shaped. Next we fabricated three lines along y direction with all parameters unchanged. The optical micrographs of the lines are shown in Fig.  responding cross sectional profiles are shown in Figure 2(j all the lines oriented in both the x and y directions exhibit a nearly circular cross section with a diameter of 85 m. Particularly, when the average laser power , a perfectly isotropic resolution in all XYZ axes was achieved To further examine the maximum laser direct writing speed achievable at the resolution of ~85 m, we fabricated four parallel lines in Foturan glass samples at from 4 mm/s to 8 mm/s. The femtosecond laser power was d) show the top-view micrographs of the lines inscribed by 4 mm/s, 5 mm/s, 6 mm/s and 8 mm/s, respectively. Figure 3(e of the lines in Fig. 3(a-d). It can be seen that the cross sections of the all appear symmetric at the different writing speeds, and the diameters of the . However, the lines become obviously inhomogeneous when the speed (e.g., see Fig. 3(c) and (d)), which should be caused by insufficient overlap between consecutive pulses during the laser direct writing. It should be noted that our femtosecond laser operates at 1 kHz repetition rate. Therefore, higher writing speeds can be s of several at different laser powers. The (b, e, h, k) and 60 glass along different scan directions at different laser intensities for individually examining the spatial resolutions along XYZ axes.
directions in Foturan glass samples with the speed was fixed at 5 mm/s and the average laser power varied from 60 to 80 mW (measured before the grating pair). Under the above irradiation conditions, no visible modifications of the glass could be observed under the optical microscope after the cond laser irradiation. Then the coupon was subjected to a programmed heat C at 5 • C/min and C/min and again held for 1 hr. After the samples were naturally cooled down to room temperature, the laser modified area appears brown. We polished the samples and examined the modified areas using the optical direction written by 80 mW, 70 mW and 60 mW, respectively. The widths of 85 um. We examined the cross sections of these lines by cutting al profiles of the direction with all are shown in Fig. are shown in Figure 2(j-l). It can be directions exhibit a nearly when the average laser power , a perfectly isotropic resolution in all XYZ axes was achieved.
achievable at the isotropic el lines in Foturan glass samples at he femtosecond laser power was fixed at 60 inscribed by setting the d 8 mm/s, respectively. Figure 3(e-h) are the d). It can be seen that the cross sections of the he diameters of the lines are all inhomogeneous when the speed , which should be caused by insufficient t should be noted that our femtosecond laser operates at 1 kHz repetition rate. Therefore, higher writing speeds can be expected at higher repetition rates with SSTF scheme from the results in Fig. 3. The homogeneous and continuous fabrication quality as shown in Fig   Fig. 3 (a-d) Top-view and (e at different speeds mm/s in (c, g), and For comparison, we also performed focal conditions and at different powers the amplified pulses were and then focused by the same objective lens writing speed was fixed at 100 µm lines inscribed in Foturan glass samples with the conventional focusing mW, 1 mW (measured before line-widths were measured to be 4(e)-(g) show the cross sectional the lines in Fig. (a)-(c) are with the conventional focusing asymmetric, leading to seriously expected at higher repetition rates with SSTF scheme from the results in Fig. 3. The homogeneous and continuous when the writing speed is under 5 mm/s, which ication quality as shown in Fig. 3(a) Fig. 3(a-c), showing that the depth of (c) are 164 m, 210 m, and 233 m, respectively. The results prove that with the conventional focusing, the focal spot produced by a low-NA lens is highly leading to seriously degraded axial resolution.
expected at higher repetition rates with SSTF scheme from the results in Fig. 3 Furthermore, we performed the laser writing of lines in Foturan glass with the conventional focusing scheme in which the effective NA maintains the same as that used in SSTF scheme. For this purpose, compressor was reduced to 0.55 mm (1/e The beam was then focused speed was fixed at 100 µm before objective lens). The higher laser power was chosen because of the looser focusing condition. The top-view and cross Fig. 4 (d) and (h), respectively depth of the lines along the optical axis of the lens was measured to be clearly show that it is impossible to of the conventional focusing To experimentally demonstrate the unique capability of large scale 3D structures EXPO 2010 in Foturan glass this demonstration, the experimental parameters are the same as that in Fig. 2( provides the most isotropic resolution in XY model are 3.5 mm, 3.5 mm, and writing of the whole structure relatively low repetition rate of 1 kHz. performed the laser writing of lines in Foturan glass with the conventional focusing scheme in which the effective NA maintains the same as that used in me. For this purpose, the diameter of the laser beam after the double compressor was reduced to 0.55 mm (1/e 2 ) with the same telescope system as shown then focused into the glass sample with the same objective lens. The writing speed was fixed at 100 µm/s and the average laser power was set to be 3 mW (measured The higher laser power was chosen because of the looser focusing view and cross-sectional view optical micrographs of the line , respectively. The lateral linewidth was measured to be 32  depth of the lines along the optical axis of the lens was measured to be 854 that it is impossible to obtain large voxel sizes of 3D isotropic resolution conventional focusing scheme. To experimentally demonstrate the unique capability of high-throughput fabrication of structures with SSTF scheme, we inscribed a 3D model of China Pavilion in Foturan glass, as shown by its digital-camera-captured image in Fig. 5 this demonstration, the experimental parameters are the same as that in Fig. 2( ost isotropic resolution in XYZ directions. The length, width, and mm, and 1.5 mm, respectively. It only took 6 minutes whole structure despite that our experiment employs a femtosecond laser at a relatively low repetition rate of 1 kHz. lines written (a-c) and was used in (d) and (h). The (b, f), 1mW in (c, g) performed the laser writing of lines in Foturan glass with the conventional focusing scheme in which the effective NA maintains the same as that used in he diameter of the laser beam after the double-pass as shown in Fig. 1. with the same objective lens. The writing s and the average laser power was set to be 3 mW (measured The higher laser power was chosen because of the looser focusing line are shown in m, whereas the 854 m. The results isotropic resolution by use throughput fabrication of China Pavilion of captured image in Fig. 5. In this demonstration, the experimental parameters are the same as that in Fig. 2(c), which Z directions. The length, width, and height of the to complete the despite that our experiment employs a femtosecond laser at a

Conclusion
To conclude, we have shown the potential femtosecond laser fabrication. The voxel size achieved in our experimenta 85 m  85 m  85  produced by a high NA objective lens. fabrication efficiency. It is flexible to the incident pulse at the back aperture of the lens resolution and the production polymerization for producing bio