Broadband Antireflective Microstructures on Diamond Fabricated by Femtosecond Laser and Selective Wet Etching

Abstract

Diamond antireflection techniques are of high interest for optical windows operating at extreme conditions. Herein, diamond antireflective microstructures in mid-infrared (MIR) spectral range were theoretically designed and experimentally fabricated. Finite difference time domain (FDTD) simulations were used to optimize the transmission performance of the diamond microstructures. Based on the simulation results, the optimized microstructures were fabricated by femtosecond (fs) laser direct writing (1030 nm, 300 fs, 25 kHz) followed by wet etching. After wet etching, the laser-modified zones and the accumulated graphitized clusters were effectively removed, thereby achieving the desired depth. The influences of laser power and scanning strategy on the morphology evolution of diamond microstructures were investigated. It was found that at the optimal conditions, the transmittance of the diamond increased from 70.9% to 81.4% (single-side) over a broad spectrum from 8 to 22 μm. This work demonstrates a promising hybrid fs laser/wet etching technique for diamond antireflective microstructures in MIR spectral range.

1. Introduction

In mid-infrared (MIR, 3–25 µm) spectral range, various applications have been demonstrated in fields such as environmental monitoring, security and defense, and pollutant detection [1]. Broadband antireflection diamond windows are in high demand for high-power or ultra-broadband MIR coherent light sources [2,3,4]. Single-layer quarter-wavelength (λ/4) coatings can be used to reduce reflection, but they are limited to narrow wavelengths and specific incidence angles. Multi-layer or gradient index coatings can be employed, but suffer from drawbacks such as adhesion, thermal stress mismatch, and low environmental stability [5,6]. Meanwhile, the design and preparation of these coatings are costly and complicated, especially for broadband antireflection applications [2,7].

An alternative technique to antireflection is employing subwavelength-structured surfaces [8,9]. In some cases, commonly used MIR transparent materials possess a high refractive index (typically >2, even up to 4). The high refractive index of these infrared materials (e.g., Ge, ZnSe, and ZnS) typically leads to high Fresnel loss, hindering light transmission and deteriorating imaging quality [2]. By designing the microstructures, a graded refractive index profile can be created, thus providing excellent broadband antireflection performance [10,11].

Due to the drawbacks in mechanical or thermal properties (typically low thermal conductivity, <30 W m⁻¹·K⁻¹), it was difficult for traditional infrared materials (e.g., MgF₂ and ZnS) to survive under extreme operating conditions [12]. In contrast, diamond has received widespread attention as an infrared material given its distinct strengths, such as high thermal conductivity (2200 W m⁻¹·K⁻¹), low coefficient of thermal expansion (1.2 × 10⁻⁶ K⁻¹), and superior mechanical hardness (10 on the Mohs scale) [12,13,14], especially for the case of high-power laser applications (e.g., CO₂ lasers, wavelength of 10.6 µm) [3,15]. In addition, diamond possesses a broad transmission window spanning from deep ultraviolet to far infrared [16]. However, the maximum theoretical transmittance of diamond in the MIR is 71.4% (based on a refractive index n = 2.38) due to Fresnel reflection losses [17]. Therefore, it is necessary to improve the transmittance performance of diamond for practical applications.

Processing diamond arbitrarily is a challenge due to its high hardness, brittle property, and chemical stability [18,19]. There have been various approaches for the microstructuring of diamond. Karlsson et al. [20] fabricated square-shaped structures (period of 4 µm, depth of 1.8 µm) on diamond using electron-beam lithography and plasma etching, and achieved a transmission of 97% at a wavelength of 10.6 µm for double-sided treated diamond. Kononenko et al. [21] used laser ablation (248 nm KrF laser or 1078 nm 9 ns Nd: YAP laser) to form antireflection microstructures on diamond, and achieved a 10% increase in transmission at the wavelength of 10.6 μm for the double-sided treated diamond. Furthermore, this antireflection effect was observed across a broad spectral range from 10 to 20 μm.

Femtosecond (fs) laser direct writing enables maskless, one-step fabrication of a wide variety of surface microstructures in an ambient atmosphere [22]. Because of its ultrashort pulse duration and ultra-high peak power density, a femtosecond laser can rapidly deposit energy into hard and brittle materials with minimal thermal effects, which has been widely used to fabricate antireflection structures on various substrates [23].

Table 1. Fs laser fabricated structures on diamond and their optical performances.

Laser Parameters	Structure Formation	Geometric Parameters	Optical Performance	Refs
1030 nm, 216 fs, 10 kHz	Square frustum microstructure	Period: 3 μm, depth: 1 μm	Single-side 8–14 μm, 77.5% 9 μm, 78.2%	[13]
515 nm, 390 fs, 25 kHz	Pyramidal array	Period: 3.5 µm depth: 1.6 μm	Double-side 8–12 μm, 92% 10.6 μm, 94.5%	[19]
515 nm, 390 fs, 25 kHz	LIPSSs	Period: ~λ/5	Single-side 625–750 nm, 76%	[24]
515 nm, 390 fs, 25 kHz	Pyramidal array	Period: 4 µm depth: 2 μm	Single-side 9–12 μm, 83%	[25]

summarizes fs laser fabricated structures on diamond and their optical performances.

Previous research on diamond antireflective microstructures has mainly focused on laser ablation. However, this approach often suffers from two major drawbacks: limited processing efficiency and the introduction of residual stress into the fabricated structures. Fs laser modification followed by selective chemical etching can enhance the processing efficiency and surface quality. However, due to its exceptional chemical inertness, the selective wet etching of diamond remains challenging and has not been widely reported. In this work, we propose an effective method to fabricate diamond MIR antireflective microstructures with broadband antireflection performance (8–22 μm). Firstly, FDTD simulation is used to optimize the geometry (i.e., period and depth) for obtaining high MIR transmittance of diamond microstructures. Then, the optimal microstructures are fabricated by fs laser direct writing followed by wet etching, which selectively removes the laser-modified zones to achieve the desired depth. The influences of laser power and scanning strategy (i.e., single scanning or multi-scanning) on the morphology evolution of diamond microstructures are investigated. Finally, the Raman and infrared spectra of fabricated diamond microstructures are discussed.

2. Materials and Methods

2.1. Experimental Setup

A Yb: YAG fs laser (Femto YL-20, Wuhan Yangtze Soton Laser Co., Ltd., Wuhan, China) with a central wavelength of 1030 nm, a pulse duration of 300 fs, and a repetition rate of 25 kHz was applied as the irradiation source. The experimental setup for fabricating diamond microstructures is shown in Figure 1a. The beam was focused onto the sample surface through an objective lens (50×, numerical aperture, NA =  0.80) mounted on an air-bearing Z-stage. A chemical vapor deposition (CVD) diamond slab (10 mm ×  10 mm × 0.16 mm, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences) was mounted on an XY motorized translation stage, and the scanning speed was set to 0.9 mm/s. Given the large-range translation stages, this method is suitable for fabricating bulk diamond surfaces. The laser power was tuned using a combination of a half-wave plate and a polarizing beam splitter. The structuring process was monitored in real time with a charge-coupled device (CCD) camera.

As shown in Figure 1b, after fs laser processing, the as-prepared microstructures (single-side-structured diamond) were treated with wet etching and processed with a mixed acid solution consisting of concentrated sulfuric acid, concentrated phosphoric acid, concentrated nitric acid and hydrogen peroxide in a volume ration of 5:1:1:3 at 120 °C for 2 days. Finally, large-area antireflective microstructures were obtained, as illustrated in Figure 1c.

2.2. Characterizations

Optical images of the diamond microstructures were taken on a digital microscope (VHX-S600E, Keyence, Osaka, Japan) with a 1000× objective (VHX-ZM1000, Keyence, Osaka, Japan). The surface morphologies of the samples were investigated with a scanning electron microscope (Scios2, Thermo Fisher Scientific, Waltham, MA, USA). Atomic force microscope (AFM) images were recorded on a scanning probe microscope (Dimension-3100, Veeco, Plainview, NY, USA). Raman spectroscopy (Witec alpha 300, Witec, Ulm, Germany) was conducted using a 532 nm excitation laser and a 10× objective (NA = 0.25) at room temperature. The transmission spectra were measured by a Fourier Transform Infrared Spectroscopy (Nicolet iS50 FT-IR, Thermo Fisher Scientific, Waltham, MA, USA). It is worth noting that the diamond microstructures, which were fabricated after fs laser fabrication but before the wet etching, were ultrasonically cleaned in ethanol for 15 min and then in acetone for 15 min to remove the debris before measurement.

3. Results and Discussion

3.1. Design of Diamond Microstructures and FDTD Simulation Results

Firstly, we analyzed the influence of key geometric parameters of the diamond microstructures on the MIR transmittance by FDTD simulation, in which the refractive index n was set to 2.38 and the extinction coefficient k was set to 0. As illustrated in Figure 2a, a square frustum array was chosen as the geometry of the microstructures, with the key parameters of period (P) and depth (D) being the primary considerations. Figure 2b displays the simulated transmission spectra of diamond microstructures within a 8–25 μm wavelength range at different periods but a fixed depth of 1.5 μm, as well as the Fresnel losses of a pristine diamond for comparison. Compared to the pristine diamond, fabricating microstructures on diamond surfaces can significantly enhance the MIR transmittance in the 8–25 µm spectral range. The highest transmittance occurred at the periodic of 3.5 µm, which was chosen as the optimal feature parameter. Figure 2c shows the simulated transmission spectra of diamond microstructures with various depths at a fixed period of 3.5 μm. The highest transmittance was observed for depths exceeding 1.5 μm. These simulation results are consistent with previous findings [19], indicating that for optical windows operating at a typical wavelength of 10 μm, a period of 3.5 μm and a depth greater than 1.5 μm are required to achieve optimal antireflective performance.

3.2. Fabrication of the Diamond Microstructures

Figure 3 shows the optical images of diamond microstructures before and after chemical treatment at various laser powers and scanning times (i.e., single or multi-scanning). These samples were ultrasonically cleaned (in ethanol for 15 min and then in acetone for 15 min) before the measurement. The microstructures were observed to be uniformly distributed over a large area. When the laser power was lower than 7.6 mW, a well-depicted structure was observed. However, when the laser power continued to increase (e.g., 11.2 mW), chips were clearly observed, which is attributed to the hard and brittle properties of diamond [19]. In addition, the top side length of the square frustum (illustrated by W in Figure 3a) dramatically decreased at high laser power (e.g., 14.3 mW).

Subsequently, we investigated the effect of the number of scans on the morphology of the microstructures. As can be seen in Figure 3b, a well-defined microstructure array was maintained after one or two scans at a laser power of 7.6 mW. However, when the number of scans was increased to three, it became severely damaged. Furthermore, a substantial accumulation of graphitic layers, which could not be removed by ultrasonic cleaning, was observed on the diamond surface. The extent of this graphitization increased with the number of scans, which is attributed to the pronounced thermal effects of enhance graphitization in the fs laser-modified region [19].

Figure 3c shows the diamond microstructures after chemical treatment. At the laser power of 7.6 mW, high-fidelity structures were obtained for single or double scanning. However, when the number of scans exceeded two, uncontrollable surface damage and chip accumulation occurred. Meanwhile, the graphite layers were thoroughly removed, confirming the effectiveness of the wet etching cleaning process.

The diamond microstructures were further characterized by AFM. Figure 4 displays the topological change before and after chemical treatment. The surface morphology was consistent across various locations, featuring a well-defined square frustum array. Under the conditions of 7.6 mW laser power and double scanning, the structure depth increased from 0.9 µm to 1.8 µm after chemical treatment, indicating an effective approach for increasing the depth of the microstructures. The root mean square (RMS) roughness increased from 353 nm to 576 nm.

As shown in Figure 5, debris was effectively removed after chemical treatment. Furthermore, the presence of LIPSSs on the edge of the square frustum (Figure 5b) confirms that the laser-modified area was selectively etched. The period of LIPPSs was ~275 nm, close to the λ/2n (λ is the wavelength of 1030 nm and n is the refractive index of the diamond at 1030 nm), which can be explained by the interference between the incident laser and a surface electromagnetic wave [26]. A direct comparison between the unmodified and laser-structured area (i.e., those formed by chemical etching after laser modification) is provided in Figure S1.

Raman spectroscopy is versatile tool for distinguishing sp³ bonding (a three-dimensional tetrahedral covalent network) from sp² bonding (planar hexagonal ring layers interconnected by van der Waals forces) [27,28,29]. There are various forms of sp² bonded carbons, exhibiting different degrees of graphitic ordering (from microcrystalline graphite to glassy carbon). Generally, amorphous carbon may contain a mixture of sp³, sp², and even sp¹ bonding [27]. The pristine diamond sample (blue curve in Figure 6) exhibited a well-defined sharp peak centered around 1331 cm⁻¹. At the Γ point in the Brillouin zone, the diamond’s triply degenerate optical phonon modes displayed a single prominent Raman peak at ~1332 cm⁻¹ at room temperature [29,30]. In our case, this peak downshifted slightly from 1332 cm⁻¹ to 1331 cm⁻¹. This shift may be attributed to the presence of mechanical stresses [31].

For the diamond microstructures before chemical treatment (red curve in Figure 6), the characteristic Raman peak appeared at around 1331 cm⁻¹, but exhibited a slightly broader shape, with a full width at half-maximum (FWHM) of 5.1 cm⁻¹, compared to that of the pristine diamond of 5.0 cm⁻¹, suggesting a slight decline in crystal quality and/or due to the presence of mechanical stresses predominantly near the boundaries of the fabricated microstructures. A broad band was observed comprising modes at ~1350 cm⁻¹ (the D vibrational mode) and ~1582 cm⁻¹ (the G vibrational mode). D and G peaks originate from sp²-bonded carbon: the D peak arises from the breathing modes of sp² atoms in rings, while the G peak is attributed to the bond stretching of all pairs of sp² atoms in both rings and chains [27]. Diamond graphitization refers to the phase transition in which diamond undergoes atomic rearrangement under specific conditions, transforming from sp³ to sp² hybridization [30]. This transformation indicates a phase transition induced by graphitization of the diamond in the fs laser-irradiated area [32]. Notably, a distinct broad peak appeared at ~1420 cm⁻¹, which may be attributed to fluorescence from nitrogen-vacancy centers [28].

For the fs laser-fabricated diamond microstructures after chemical treatment (yellow curve in Figure 6), the characteristic Raman peak appeared at 1331 cm⁻¹ but the peak intensity was significantly higher, which may be attributed to the increased surface roughness after etching [33] (see AFM results for details). The disappearance of the D and G vibrational modes indicates the high removal efficiency and selectivity of the graphitic phases introduced by fs laser irradiation.

3.3. Optical Characterization

Figure 7 shows the experimentally obtained transmittance spectra of pristine diamond, and the diamond microstructures before and after chemical treatment. The transmission increased from 70.9% for the pristine diamond to 81.4% of diamond microstructures after wet etching treatment, indicating a significant enhancement in optical transmission in the MIR range. It is worth noting that the maximum theoretical transmission of intrinsic diamond is 71.4% (n = 2.38) [17], but the actual transmittance of the diamond is affected by the surface roughness and the fabrication technology [13]. That is why the value of transmission of the pristine diamond is lower than the theoretical one. The experimentally measured transmittance spectra displayed a decrease in MIR regions from the theoretical simulation. The relatively poor optical performances of the diamond microstructures before wet etching treatment may be due to the shallow structural depth (as shown in Figure 2c) as well as the accumulated graphitized layers (see details in Figure 6). For comparison, the simulated and experimental transmittance spectra of the diamond microstructures are presented in Figure S2. The experimentally measured transmittance was slightly higher in MIR regions than the theoretical simulation. One reason is that the depth of diamond microstructures after chemical treatment in some regions could reach 1.9 µm or even 2.0 µm, which would result in higher transmittance in the MIR range in simulations (see Figure 2c). Additionally, the natural gradient in surface roughness and structural size formed at the micro–nano scale may contribute to better refractive index matching [34].

Diamonds comprise the lowest mass element that forms a stable covalently bonded crystal lattice, which is highly symmetric and tightly bound [35]. The melting point of diamond is as high as 3550 °C. However, the oxidation activation energy (135.7 kJ mol⁻¹) of diamond is very low, and even a trace amount of oxygen can markedly reduce the graphitization temperature of diamond to ~700 °C [31]. Meanwhile, the diamond substrate possesses an ultrahigh thermal conductivity of 2200 W m⁻¹·K⁻¹, reducing the temperature gradients and eliminating the risk of structural failure due to heat accumulation. Herein, a combined process of fs laser and selective wet etching was adopted to prepare microstructures. No additional coatings, heterogeneous binders, or doped materials were introduced, which meant the excellent thermal stability characteristics of diamonds could effectively be inherited. In addition, the heat-affected zone of fs laser processing of diamond is extremely narrow, only forming micro–nano scale structures on the material surface, and will not damage the bulk thermal stability of the diamond substrate. Based on the intrinsic thermal stability of the above-mentioned diamond substrate and the preparation characteristics of this process, the antireflection microstructures prepared in this study can be adapted to the conventional high-temperature (at least 500 °C [19]) conditions in the MIR field, without obvious problems such as morphology collapse and transmission performance attenuation. However, special attention should be paid to the extreme conditions of short-term thermal shock (e.g., hypersonic flight) [36].

4. Conclusions

In summary, we have successfully fabricated anti-reflective microstructures in diamond by fs laser direct writing followed by wet etching. The morphological evolution and corresponding MIR optical performance were systematically investigated. A broadband MIR transmittance enhancement (8–22 μm) to 80–81.4% was achieved on a single microstructure surface using optimal laser parameters (7.6 mW, double scanning). This method offers distinct advantages, including higher processing efficiency and significantly lower residual surface stress. Benefiting from large-range translation stages and flexible trajectory design, this approach is suitable for large-area texturing of bulk diamond with designed patterns. These results may pave the way for the fabrication of diamond-based functional structures for applications in high-power laser systems.