When ultrafast laser intensity approaches the ablation threshold of the target material, it can induce surface nanostructures with periods considerably smaller than the wavelength. This phenomenon, known as laser-induced periodic surface structures (LIPSSs)^[1,2], occurs in nearly all instances of ultrafast laser exposure. As a straightforward and effective surface texturing technique, LIPSSs have attracted extensive interest due to their unique features, formation mechanisms, and diverse potential applications^[3–10]. LIPSSs form at the focal point of a polarized laser beam and exhibit periodic variations in surface morphology. Different types of induced stripe structures can be formed on the surface of diverse materials with changing laser parameters, including low-spatial-frequency LIPSSs (LSFLs) and high-spatial-frequency LIPSSs (HSFLs), each exhibiting notable differences in their periods and orientations. Recent research over the past decade has concentrated on understanding the formation mechanisms of LIPSSs. For LSFLs, the prevailing consensus attributes their creation to the interference of lasers with surface-scattered electromagnetic waves, as described by the Sipe model^[11]. In contrast, the formation mechanism of HSFL remains contentious. Various theoretical frameworks have been proposed, based on the Sipe model, including concepts such as Coulomb explosion^[12], surface plasmon excitations^[13], second harmonic generation^[14], liquid-phase surface driving^[15], and self-organization of unstable materials^[16]. Currently, the dominant theory concerning HSFL formation suggests that it results from the coherent interaction between the surface-scattered near field and the electromagnetic field produced by an incoming laser beam^[17,18].

Diamond exhibits unique characteristics, including a high refractive index (n=2.42), a significant breakdown electric field (107V/cm), and the highest thermal conductivity among solids (22 W/cm·K)^[19,20]. Diamond-based LIPSSs have been successfully exploited in optoelectronic devices, surface material treatments, and various other fields^[21]. Diamond-based LIPSSs have already been fabricated and morphologically investigated in recent studies. Mastellone achieved the fabrication of two-dimensional LIPSSs with an 80 nm period on diamond surfaces using time-delayed, cross-polarized femtosecond laser pulses^[22]. This configuration significantly boosts the light-trapping capability of diamond, enhancing visible light absorption by 50 times. Calvani fabricated LIPSSs with a period of 170 nm on the surface of polycrystalline diamond using an ultra-short titanium sapphire pulsed laser, enabling the absorption rate of diamond in the visible and infrared regions to exceed 80%^[23]. Martinez similarly produced LIPSSs with varying periods and aspect ratios on diamond surfaces, inspired by the antireflective properties of moth eyes^[24]. Simulations for LIPSSs with a 500 nm period and 0.4 aspect ratio yielded a transmittance of 99%. However, prior research encounters constraints regarding the processing efficiency of diamond-based LIPSSs, which hinders their practical applications. Producing large-area diamond LIPSSs remains a significant challenge. Current challenges involve ambiguous formation mechanisms and significant debris buildup during preparation, hindering the production of high-uniformity and high-fidelity LIPSSs.

In this paper, the evolution of diamond LIPSSs is systematically investigated; and the mechanism of scattering near-field-driven nanostripe formation is demonstrated in conjunction with the surface electric field distribution. Meanwhile, large-area diamond LIPSSs with a period of 105 nm are efficiently fabricated using a femtosecond laser line spot shaped through a cylindrical lens. A low-energy defocused laser is employed for a secondary scan over the processed area. This step removes debris produced during processing, significantly enhancing the fidelity of the LIPSSs, which in turn demonstrates notable anti-reflective properties. Optical performance results confirm the ability of LIPSSs to enhance the optical transmittance of diamond, boosting its transmittance from 65% to 76% at 750 nm.

For the modification of samples, a femtosecond laser operating at a wavelength of 515 nm, generated by a second harmonic generation system, was employed. This laser operated at a repetition frequency of 25 kHz and featured a pulse width of 390 fs. The thickness of a single crystal diamond is 300 µm. The laser beam was focused onto the sample sequentially using a cylindrical lens and a 60× objective lens with a numerical aperture of 0.7, producing a line spot length of 24 µm. The production of diamond-based LIPSSs was facilitated by regulating the motion of a displacement stage using auxiliary software. Following the laser modification, a low-energy defocused laser was utilized to rescan the processed area, effectively removing the surface debris. Common post-treatment processes, such as ultrasound and acid corrosion, were not employed here. Field emission scanning electron microscope (FE-SEM, JSM 7610F, Japan) facilitated the visualization of the surface morphology of the resulting LIPSSs. The optical properties of diamond LIPSSs were evaluated by a transmittance testing system operating in the visible spectrum. The light source of the test system is a white-light continuum with a wavelength range of 450 to 750 nm, produced by the action of a femtosecond laser on a sapphire crystal.

To investigate the relationship between the evolution of LIPSSs and the laser parameters, two sets of experiments were conducted. In the first set of experiments, the number of pulses applied to the diamond surface was kept constant at 20, while the laser energy irradiating the sample varied from 0.09 to 0.26 µJ, as shown in Fig. 1(a). When the laser energy (E=0.09μJ) was below the ablation threshold for diamond, no ablation occured in the irradiated region. Instead, a non-destructive phase transition took place in the center region. As the laser energy increased within the near-threshold range, the nanostripes widen and the grooves became uniformly shaped. Moreover, the surrounding structure remained intact since there was no damage caused by ablation. The induced structures exhibited irregularities due to the homogeneity of the material and the inherent fluctuations of the laser. At a laser energy of 0.26 µJ, a distinct ablation crater was observed on the surface, accompanied by additional damage to the streak structure within the crater. This damage was considered to be the formation of LIPSSs on diamond after surface ablation. In the first set of experiments, a direct evolution of LIPSSs was demonstrated from an undamaged surface to multiple striped structures. This phenomenon could be attributed to a higher number of accumulated pulses. In the second set of experiments, the generation of each individual point depended on the number of accumulated pulses.

Figure 1(b) shows the evolution of nanostripes under the influence of varying pulse numbers as observed through SEM. At a pulse number of 10, only a single trench structure with a width of 130 nm exists in the irradiated region. As the number of pulses increases, the nanostripe structure broadens to both sides and shows a clear periodicity. To further explore the role of the optical near field during the evolution of the nanostripes, based on the prepared nanogroove structure [Fig. 1(b)], the localized electric field redistributed on the surface of the nanostripes when subjected to subsequent pulsed irradiation is calculated as shown in Figs. 1(c)–1(e). In particular, the bottom layer of the constructed laser irradiation model is a diamond crystal (simulated thickness of 500 nm, refractive index of 2.42), and the top layer is an air layer (simulated thickness of 1500 nm, refractive index of 1). In order to exclude the effect of interfacial reflections, the side and bottom surfaces of the model structure have been set up with perfectly matched boundary conditions. The Gaussian beam is incident from the topmost interface of the model, with a femtosecond laser wavelength of 515 nm and an electric field polarization direction along the x-axis. When the pulse number is 10, the majority of the incident light is confined within a single nanostripe. The electric field along both sides of the trench in the x-direction forms two secondary peaks, while along the y-direction, the electric field gradually becomes weaker and weaker from the inner strength of the trench. This indicates that the nanostripe structure is prone to further damage in the y-direction. At this point, melting and subsequent re-solidification occur on both sides of the single groove [Fig. 1(b1)]. When the pulse is continuously applied [Fig. 1(b3)], secondary groove structures have been generated on both sides of the primary groove. The growth of the main groove and the generation of secondary grooves are consistent with the effects of the electric field distribution illustrated in Fig. 1(c). The near-field intensity at both ends of the y-direction is less than that of a single trench due to the influence of the longer primary trench structure and the secondary trench. This implies that the growth of the y-direction nanostripe slows down in the subsequent stage. The secondary nanostripe in the x-direction is observed to be bent toward the center due to the near-field effect. Additionally, the two weaker secondary peaks at 400 nm suggest that the subsequent damage to the inward-bending tertiary nanostripe structure has occurred. However, due to the relatively low intensity of this peak, the length of the tertiary nanostripe structure is found to be shorter than that of the secondary nanostripe. In an ideal situation, the initially damaged nanopore acts as a seed structure, which is irradiated by subsequent pulses to generate multiple nanostripes. This process is then repeated with irradiating pulses to damage secondary and tertiary nanostripes, resulting in the formation of a striped structure with a defined periodicity.

To enhance the fabrication efficiency of LIPSSs and improve the consistency of energy density in the laser irradiation area, we constructed a femtosecond laser line spot processing system utilizing columnar lens shaping, as depicted in Fig. 2(a). The laser power was maintained at 16 mW, employing femtosecond laser line spots with varying scanning speeds (v=0.05 to 0.13 mm/s, corresponding to pulse counts N=90 to 234) for expedited fabrication. Figure 2(b) presents the SEM images of LIPSSs produced at different scanning speeds. When the laser scanning speed is 0.05 mm/s, the density of excited free electrons is saturated due to the accumulation of a large number of pulses. The subsequent pulses cause significant thermal accumulation, leading to the ablation and destruction of the nanostructures. Conversely, when the scanning speed is increased to 0.13 mm/s, the LIPSSs are not fully modified because the excited free electron density does not reach a certain threshold, resulting in insufficient modification of the LIPSSs. Overall, we observe high-quality LIPSS formation within a range where modification is minimal yet prevents destruction. Accordingly, operating within a scanning speed range of 0.08 to 0.1 mm/s can produce LIPSSs with superior quality.

Figure 2(c) shows the influence of laser power on the morphology of LIPSSs produced by femtosecond laser irradiation on a diamond substrate. The scanning speed remained constant at 0.1 mm/s while laser power increased from 13.2 to 14.7 mW. Notably, when laser power exceeded 14.5 mW, an accumulation of debris appeared on the treated surface, correlating positively with the applied laser power. This phenomenon arises from pronounced ablation effects during LIPSS formation at higher laser power, resulting in the ejected material adhering to the modified area as debris. At a laser power of 14.2 mW, the treated surface exhibited fewer debris attachments, revealing a fine structure depicted in Fig. 2(d). Although minor granular attachments persisted on the treated surface, a clear stripe structure with a periodicity of approximately 100 nm was obtained. In contrast, at laser powers below 14 mW, a small quantity of fragments reemerged on the treated surface. Unlike areas processed with higher laser power, these regions displayed a coating of nanoscale particles. During the femtosecond-laser-induced production of LIPSSs, several nanometer-sized particulate debris forms without significant ablation. These particles adhere stubbornly to the LIPSS surface, resisting removal through ultrasonic cleaning. To mitigate this issue, the implementation of laser cleaning techniques has been introduced for the effective removal of the generated debris and particles.

To confirm that the nanoparticles on the processed areas can be effectively removed by a high-speed scanning femtosecond laser, we investigated the impacts of laser power, scan count, defocusing distance, and scanning velocity on debris removal during laser cleaning. Figure 3(a) presents SEM images of a single scan with a focused femtosecond laser at different powers in the LIPSS region. As laser power increases, surface debris in the processed area diminishes gradually. At a power of 2.4 mW, LIPSSs in the processed area sustain damage, as highlighted by the yellow region, whereas the regularity of LIPSSs in the red region appears less defined. Figure 3(b) displays the SEM image of the LIPSSs after two scanning sessions with the focused femtosecond laser. This image reveals numerous fragments alongside black deposits (graphite layer) on the surface of LIPSSs, which significantly deteriorate surface quality and fail to enhance debris removal efficiency. Consequently, debris persists on the processed surface even after targeted cleaning efforts.

Figure 4(a) shows the SEM images of LIPSSs after a scan of the processed area with a femtosecond laser at different defocus distances. To avoid any influence on the subsequent optical performance of the LIPSSs, the selected focus positions were all on the diamond surface. This is because when the laser focuses inside the diamond, it can cause ablation or modification of the diamond interior. It can be seen that at a defocus distance of 7 µm, LIPSSs with minimal surface debris attachment are obtained. By setting the laser power to 14 mW and the laser defocus distance to 7 µm, the effect of scanning speed on debris removal by the defocusing laser is further investigated [Fig. 4(b)]. When the scanning speed is slow, the deposition of energy causes slight damage to LIPSSs, as shown in the red region in Fig. 5(b). When the scanning speed is 0.1 mm/s, the LIPSSs have good morphology and good periodicity, and there is no extensive debris in the processing area. If the scanning speed is too high, the surface debris cannot be wiped out.

Large-area diamond-based LIPSSs were fabricated at a scanning speed of 0.1 mm/s with a laser power of 14.2 mW, while the fabrication process was completed in less than 1 s. The defocusing laser scans occurred once, and the cleaning procedure finished within 1 s. The cleaning settings included a laser power of 8.7 mW, a scanning speed of 0.1 mm/s, and a defocus distance of 7 µm. Utilizing these parameters, LIPSSs were produced on the diamond surface within an area of 12μm×24μm, as illustrated in Fig. 5(a). The LIPSS surface displays minimal debris accumulation, showcasing excellent connectivity and pronounced periodicity with a period of 105 nm. The inset illustrates the fast Fourier transform spectrum linked to the stripe structure, confirming the homogeneity of the nanostripe structure. Furthermore, as a unique one-dimensional grating structure, the optical characteristics of LIPSSs within the visible spectrum are evaluated in Figures 5(b) and 5(c). Notably, the transmittance of diamonds exhibiting LIPSSs demonstrates a substantial increase spanning the wavelength from 625 to 750 nm, while the reflectance shows a pronounced decrease. At a wavelength of 750 nm, the optical transmittance of the diamond LIPSSs achieves 76%, representing a 10% enhancement compared to the intrinsic surface area of diamond.

In summary, we report a cylindrical lens method for shaping femtosecond laser line spots to create large-area LIPSSs on diamond substrates. This process involves removing debris from the LIPSS surface using a laser secondary defocusing scanning technique. Ultimately, we achieve the fabrication of sub-wavelength (Λ≈λ/5≈100nm) LIPSSs that exhibit high homogeneity and continuity on single-crystal diamond. Our method can generate LIPSSs over an area of 12μm×24μm in just 3 s, suitable for texturing bulk diamond surfaces. The quality of these nanostripes improves through adjustments in scanning speed and pulse energy. Additionally, we can effectively clean the LIPSSs by modifying the focus position during the secondary scan. Notably, we experimentally demonstrate that LIPSSs can enhance the optical properties of diamond, boosting its transmittance within the 625–750 nm range by as much as 10%. This advancement paves the way for new applications of diamond in integrated photonics, microfluidics, and photovoltaics.