1. Introduction

In high power laser systems there are a large number of transmissive optics, for which AR coatings are indispensable for realizing high transmittance and thereby the required power transfer. One necessary property of the AR coating is the ability of resisting laser-induced damage, and this requirement has resulted in the wide application of sol-gel SiO₂ coatings produced via the Stöber process [1,2]. This kind of coating is composed of nano-sized SiO₂ particles, generating lots of pores inside and between them. It is the porous structure that determines the low refractive index for antireflection and the relaxed texture for the high LIDT. Although having some defects, such as the weak mechanical strength and the susceptivity to environmental contamination, this coating is now still employed in the most powerful laser systems after its serving for more than 20 years. People have kept making effort to improve the coating’s quality, including increasing its LIDT and decreasing its adsorption of contaminants [3–13].

However, an alternative choice – an ordered mesoporous SiO₂ AR coating – has emerged in the area of high power laser, which owns the same chemical composition, comparable porosity but different physical structure in contrast to the traditional one introduced above, and which has been found to feature both strong mechanical strength [14] and high LIDT [15]. There is no doubt that it is meaningful to contrast the two kinds of coatings for future study and application, but there is no clear impression of the comparison by now, for the performances of them had been studied separately all the time. As known, the value of LIDT is impacted not only by the coatings themselves, but also by many other factors such as the glass substrate, the testing equipment, the scheme of the test, the operation during the test, and so on. It makes sense to compare the LIDTs in one experiment, and this is what was addressed in this work.

2. Sample preparation

Both of the two coatings are derived from SiO₂ sols. Generally, there are two kinds of SiO₂ sols distinguished by acid and base catalysis. In the case of acid catalysis, the speed of hydrolysis is faster than that of condensation, while it is the opposite in the case of base catalysis [16]. Consequently, there forms fiber chains structure in the acid-SiO₂ sol so that the resulting coating features a dense structure and a tough texture with a low porosity [17]. To get the ordered mesoporous coating, the acid-SiO₂ sol system is usually used with the addition of block copolymers as pore formers [14,15,18–20]; Pluronic F127 – a triblock copolymer composed of a hydrophobic chain in the middle and hydrophilic chains at both ends – is commonly adopted and was also used in this work considering its compatibility with silica systems. While the traditional porous coating is derived from the base-SiO₂ sol, in which nano-sized particles are formed, with some of them monodispersed and the rest connected with each other, so that the resulting coating is skeletonized by nanoparticles and is characterized by an amorphous porous structure [2].

In this work, we prepared F127-SiO₂, base-SiO₂, and acid-SiO₂ coatings, respectively, making their thickness as λ/4n (λ is the center wavelength, i.e. 1064 nm in this work, and n is the refractive index of the coating) in the dip-coating process by adjusting the withdraw speed. The evolvement of the acid-SiO₂ coating is based on such a consideration: although it is not suitable for being a single-layer AR coating due to its relative high refractive index, it can provide reference for understanding the factors that affect LIDT. The preparations were almost the same as what we described in the previous papers [11,14,17] and are charted in Fig. 1. In brief, three SiO₂ sols were prepared all in the ethanol systems with tetraethyl orthosilicate (TEOS) as the precursor, and HCl or ammonia as the catalyst. The mole ratios were TEOS: EtOH: H₂O: NH₃ = 1: 38: 2: 0.54 (base-SiO₂ sol), TEOS: EtOH: H₂O: HCl = 1: 38: 2: 0.25 (acid-SiO₂ sol), and TEOS: EtOH: H₂O: HCl: F127 = 1: 38: 4.48: 0.46: 9.95 × 10⁻³ (F127-SiO₂ sol). These ratios were found feasible to get coatings with the greatest quality (including ideal refractive indices and low-density defects) according to our experience. Furthermore, an ammonia treatment of 20 minutes of the base-SiO₂ coating was carried out which has been proved to be very effective for the increase of LIDT [5,11]. The F127-SiO₂ and the acid-SiO₂ coatings were heated at 500 °C and 150 °C, respectively.

 

Fig. 1 Flow charts of preparations of the base-SiO₂ (a), the acid-SiO₂ (b), and the F127-SiO₂ (c) coatings.

Download Full Size | PDF

3. Results and discussions

Figure 2 shows microscopy images of the three coatings. Unlike the base-SiO₂ coating, the porous structure of the F127-SiO₂ one would not occur until after the heat treatment was applied, which burned out the micelle template that formed via the EISA (evaporation-induced self-assembly) process. As shown in Fig. 2(a), the coating structure is highly ordered with uniform hexagonal mesopores of ~10 nm. The texture of its SiO₂ skeleton is dense fiber-like due to the acid catalysis. In contrast, the base-SiO₂ coating (Fig. 2(b)) is also porous but amorphous, and the pores have a wide range from micropores (<2 nm) to macropores (>50 nm). Its SiO₂ skeleton is composed of particles of ~40 nm. The two coatings performed as good as each other in terms of antireflection – the transmittance of the coated glasses are both of ~99.9% (Table 1).Their refractive indices are 1.23 and 1.20, respectively, resulted from modeling of the transmittance spectra using Film Wizard32 software. Accordingly, their porosities were determined as 46% and 53%, respectively, using Lorentz-Lorenz relationship [21]:

 

Fig. 2 TEM image of the F127-SiO₂ coating (a) and SEM images of the base-SiO₂ (b) and the acid-SiO₂ (c) coatings.

Download Full Size | PDF

Table 1. Properties of the bare substrate and the three coatings.

View Table

The damage testing was performed in the R-on-1 regime (i.e. multi pulses per location on the sample with increasing laser energy until the occurring of damage) according to ISO standard 11254-2.1, using a Q-switched Nd:YAG pulsed laser with a laser wavelength of 1064 nm and a pulse length of 10 ns. 20 different locations (5 × 4 array) were irradiated, with the distance between two close spots of ~5 mm. The damage was estimated from the visual inspection of plasma flash and detected in situ with a microscope. The LIDT of each location was defined as the average between the maximum incident pulse energy density for which the damage didn’t happen, and the minimum one for which the damage happened. Eventually, the sample’s LIDT was obtained by averaging each location’s.

The testing results are displayed in Table 1. The fused silica substrate (Φ35 × 2 mm, etched in 6% NH₄HF₂ aqueous solution for 17 min for removing the subsurface of ~500 nm that was rich in defects [22–26]) has a LIDT of 62.5 J/cm², higher than all the coated samples. The LIDTs of the F127-SiO₂ and the base-SiO₂ coatings are close to each other, and much higher than that of the acid-SiO₂ coating. Usually, the damage of the substrate or the coating is begun with the absorption of laser energy, so we measured the absorptions of 1064 nm laser by the samples with the surface thermal lens technique. As listed, the values of absorptions for all are just several ppm, and increase successively from the substrate, the F127-SiO₂, the base-SiO₂ to the acid-SiO₂ coatings. However, the small gaps between these absorptions might be not considerable for the differences among the LIDTs, even if they did have effect. Nevertheless, it implies that the absorptive defects were very limited in all the three coatings, as the intrinsic absorption of pure silica was supposed to be very small. The absorptive defects could be the residual organic substances in the coating which originated from the reagents in the sol preparations, and the residual defects in the substrates. The heating temperature of the F127-SiO₂ coating was as high as 500 °C, probably removing the organic substances the most thoroughly compared with the other two, and thus resulting in the lowest absorption.

The overwhelming superiority of the F127-SiO₂ and the base-SiO₂ coatings should be attributed to their high porosities. The dense acid-SiO₂ coating, although loading less laser energy due to the relative low transmittance, presented the weakest ability of resisting laser damage. This is reminiscent of the traditional PVD (physical vapor deposition) AR coatings, which also feature dense structure and relative low LIDTs, and this is actually the main reason that the sol-gel coatings were developed for the high power laser system.

Note that the skeletons of the F127-SiO₂ and the acid-SiO₂ coatings are basically the same; this directly proves the significance of the high porosity in realizing high LIDT. On the other hand, despite the weaker mechanical strength, the base-SiO₂ coating also overwhelmed the acid-SiO₂ one, implying that the strength of the coating skeleton may not have important effect on the LIDT. Overall, the porous structure should be the key factor for high LIDT, no matter how the coating skeleton is.

Comparing the F127-SiO₂ and the base-SiO₂ coatings, the former was no more susceptible to laser damage than the latter, and was even slightly better according to the data. The ordered mesoporous structure makes the F127-SiO₂ coating similar to a crystal of anisotropy, which would lead to non-uniform heat transfer when absorbing laser energy, and therefor the coating could be easier to damage as a result of the non-uniform heat distribution. In this context, the base-SiO₂ coating with an amorphous structure has an advantage. However, there is another important factor, structural defect, affecting the LIDT. For the base-SiO₂ coating, there may exist large-size clusters derived from the sol [8,11], damaging the homogeneity of the coating. In the worst case, the coating can present macroscopical spots owing to these clusters. Hence, one should try to decrease the number and size of the clusters during the sol preparation (by adjusting the reaction speed, adding organic binder and so on). In contrast, the F127-SiO₂ coating is more homogeneous on the whole, almost with no obvious structural defects.

Other differences between the two coatings like the surface roughness and the coating thickness are not considered important, because microroughness of several nanometers (measured by AFM) is too small, compared with the laser wavelength (1064 nm), to generate considerable scattering and the resulting laser absorption. As for the difference in thickness of several nanometers, the possible extra defects embedded in the coating are also negligible (usually thick coating has more defects).

The damage morphologies of the bare and the three coated substrates are displayed in Fig. 3.The four samples were all shot with the same laser energy of 65 J/cm² which was greater than all the samples’ LIDTs. As shown, the sizes of the damage spots increase successively from the bare substrate, the F127-SiO₂, the base-SiO₂, to the acid-SiO₂ coatings, in accordance with the relationship among their LIDTs. The damage of the substrate presented as a melting spot, and there were no pits for all the samples that usually arise from the polishing compound, such as ceria, in the substrate’s subsurface [23,25], thanks to the etching treatment. The morphologies of the coated substrates indicate that the coatings were molten in the fusion damages. The areas of substrates’ damage in the base and the acid-SiO₂ samples (black areas) are much larger than that of the bare substrate; in contrast, the one in the F127-SiO₂ sample is the smallest even compared with the bare substrate. These mean that the former two coatings exacerbated the substrates’ damage while the F127-SiO₂ coating did conversely. In view of the basically same distribution of electromagnetic fields in the base-SiO₂ and the F127-SiO₂ coatings during laser irradiation, the skeleton of particles of the base-SiO₂ coating probably resulted in more powerful impact on the substrate when the coating was damaged in the process of thermal mechanical coupling. This conjecture could be proved in some sense by the damage morphology of the acid-SiO₂ coating. Because of the lowest LIDT, its damage was the most serious and thus did harm to the substrate more badly than what the F127-SiO₂ coating did, however, its fiber-like skeleton may account for such a fact that the damage area of its substrate was not larger than that in the base-SiO₂ sample.

 

Fig. 3 Damage morphologies of the bare substrate (a) and the three coated samples: F127-SiO₂ (b), base-SiO₂ (c), and acid-SiO₂ (d) coatings.

Download Full Size | PDF

4. Conclusions

The F127-SiO₂ and the base-SiO₂ coatings, both possessing porous structures with close porosities, have comparable abilities of resisting laser damage. High porosity rather than the skeleton texture, is the dominant factor for high LIDT. When the damage occurs, the F127-SiO₂ coating results in the minimum damage of the substrate, indicating the superiority of the fiber skeleton compared with the particle one. Relatively, the base-SiO₂ coating is thought to have greater room for the structure optimization, which can help increase its LIDT. The choice of the two coatings in future application should also rely on other properties, such as the optical stabilities when working in the environment with contaminations.