Anti-reflective surfaces structures (ARSS) have been successfully fabricated on fused silica windows, lenses and fibers, and spinel ceramics. The reflection loss for spinel was reduced from 7% per surface to 0.9%. For fused silica with ARSS, the reflection loss was reduced to 0.02% near 1 µm. Pulsed laser damage thresholds at 1.06 µm were measured and thresholds as high as 100 J/cm2 were obtained for fused silica windows of up to 10 cm in diameter with ARSS and 850 J/cm2 for silica fibers with ARSS on the end faces. Spinel samples with ARSS showed damage thresholds more than two times higher than that of spinel with traditional AR coatings.
© 2014 Optical Society of America
The direct nano-patterning on the surface of an optical element to achieve reduced Fresnel reflections is an attractive alternative to traditional AR coatings. Unlike coatings, the anti-reflective surface structures (ARSS) processing does not involve applying additional materials on the surface of the optics, which often results in coating delamination under thermal cycling and laser damage to the coating at lower thresholds than the window. In contrast, state-of-the-art processing has resulted in antireflective performance of ARSS comparable to that of the traditional AR coatings, while adding significant advantages such as higher laser damage thresholds, large acceptance angles and ease of cleaning, since there is no foreign material on the surface. Typically, periodic anti-reflective surface structures can be designed to work over large bandwidths [1–3] with a variety of materials  and have been shown to exhibit high laser damage thresholds . The period of the pattern is designed to be on a sub-wavelength scale in order to avoid undesired diffraction effects, while the height of the individual features is on the order of one-half the wavelength, in order to simulate a graded index variation between air and the optical substrate. Modeling of these structures is typically done using rigorous coupled-wave analysis (RCWA) methods, although the effective-index method provides reasonable insight (although it cannot correctly model the diffraction edge region) .
There are different designs of ARSS that show some variance in performance for a given material. In the general case where the individual features have a variable cross-section (for example, all features are pyramidal in shape and of fixed height) we obtain the so-called “moth eye” structure, where the shape and the height of the individual features will determine the performance over a certain spectral bandwidth. Another type of ARSS is where the individual features have a constant cross-section, such that all features are cylindrical and with fixed height, and the structure is simply called sub-wavelength structure (SWS), providing best performance at and around a single wavelength, which is determined by the height of the features. Finally, there is also the very important case where the features are randomly dispersed and have variable heights, which is called random ARSS (rARSS).
For moth eye types of surface structures, nano-structuring is typically done using lithography followed by dry-etching-based methods. In the case of the rARSS, the lithographical step is not needed and the optical surface is typically processed with reactive ion etching, using plasma and gas mixtures appropriate to the substrate material.
In this paper we present the latest results utilizing the ARSS technique as applied to windows and optics needed for applications in the 1 – 2 µm spectral range for high energy lasers, including spinel ceramic and silica windows, and silica lenses and fibers. We also show progress in obtaining windows of large size (up to 10 cm) for practical applications.
2. Experimental procedures
A variety of optics were utilized for this study, including fused silica windows and lenses (Corning 7980 glass), single-mode telecommunications optical fiber (Corning SMF-28), and transparent MgAl2O4 spinel ceramic windows fabricated at NRL . Most of the samples were 1” diameter with thickness of 1 to 3 mm, with the exception of the large windows that were 10 cm diameter. The samples were polished prior to ARSS fabrication with commonly accepted levels of optical surface finish: 10/5 scratch/dig and average surface roughness less than 1 nm.
2.2. AR surface structures fabrication
Surface nanostructuring was performed using two principal methods. The first method, which yields a periodic ARSS, required the surface to be resist-patterned through UV lithography, followed by dry-etching, which transfers the resist pattern into the substrate surface [7,8]. Fused silica windows and spinel ceramic windows were processed with these methods using a holography-based patterning system followed by the dry-etching process.
The second method, which yields a rARSS, consisted of direct reactive ion etching of the optical element surface using process parameters suited to the respective substrate material [8, 9]. Excellent results were obtained on fused silica windows, fibers and lenses.
Another method, specifically designed for the spinel ceramic, was also investigated where the spinel powder is hot-pressed with a mold which contains the negative of the surface structure desired to be created . Vitreous carbon discs were initially patterned, using lithography and dry-etching, with the required periodic structure and then used as the hot-press mold for shaping the spinel powder. During the hot-press process, the pattern was transferred directly into the surface of the spinel window as it formed .
2.3. Optical characterization
A Cary spectrophotometer was used to measure window transmission before and after the ARSS fabrication process. However, the Cary data is not reliable when transmission loss approaches very low values, such as 0.01% per surface. Thus a broadband reflectivity measurement system was used in order to carefully determine the very low reflection loss from a single surface. A 6 + 1 fiber bundle was coupled to a white-light source, such that 6 of the fibers irradiate the surface studied, while the distal end of the 7th fiber was coupled to a spectrophotometer for the analysis of the reflected light (with great care taken to block the response from the back surface). A NIST-calibrated reference surface was used to calibrate the experimental setup. Surface analysis was also done by measuring the surface profile and roughness with SEM and a 405-nm Keyence confocal laser microscope.
We also investigated the back-scattering from a variety of substrates in order to compare the ARSS process with traditional AR coatings. We used the bi-directional reflection distribution function (BRDF), which is the ratio of the reflected radiance over the incident radiance, as a way to characterize a given surface. For a given incidence angle (i), observation angle (θ) and wavelength (λ), the BRDF can be determined as given in Eq. (1) from the incident power Pinc that irradiates the optical surface as follows:Eq. (2) as:
2.4. Laser damage testing
Laser damage testing was performed at 1.06 µm wavelength, using the following parameters: linear polarization, incidence angle 0°, TEM00 beam profile, 10 ns pulsewidth, 300 μm diameter spot (at 1/e2), 20 Hz repetition rate, and 200 shots/site over a total of 90 to 100 sites. The onset of damage was determined with visible observation, at 150x with a Nomarski darkfield-microscope or when plasma or increased scatter was detected using a helium neon laser.
In addition, ATFilms applied high quality thin film AR coatings onto samples to provide comparative data between ARSS substrates and traditional AR coatings.
3. Experimental results and discussion
3.1. Fused silica windows
Three types of ARSS structures were tested and compared: motheye ARSS, random ARSS, and SWS fabricated on the surface of the 1” diameter silica windows. Broadband spectral transmission was measured before and after ARSS fabrication, in order to quantify enhancement (using Fresnel formulae and the index of refraction values for Corning’s 7980 glass). The results are shown in Fig. 1.For comparison, the transmission for a silica sample with one side AR coated is shown in Fig. 1(c). It is notable that the rARSS shows high transmission as well as a better broadband AR performance than the other types of surface structures.
We investigated the surface morphology and uniformity of the patterns by using scanning electron microscopy (SEM) with typical images shown in Fig. 2(SWS) and Fig. 3 (rARSS). The uniformity of the SWS etching is evident in Fig. 2, with the features exhibiting straight walls with a well-defined rectangular cross-section. Feature depth was estimated to be 220 nm.
Typical cross-sectional SEM images for the silica sample with random ARSS are shown in Fig. 3 with feature depth of about 400 nm.
Some of the best-performing silica samples with rARSS were also tested in reflection mode, using the broadband reflectivity system with the fiber bundle. As shown in Fig. 4, reflection loss as low as 0.02% at 1.06 µm was measured, a value that rivals the best AR coatings provided by traditional thin-film coating technology.
Several samples, including an untreated fused silica window, two fused silica windows with random ARSS, and a silica window with traditional AR coating were tested for laser damage at 1.06 µm. The results, shown in Table 1, and illustrated in Fig. 5 confirm that the random ARSS has up to five times higher laser damage threshold than that for a traditional AR coating, and the threshold of 100 J/cm2 is near the value for surface damage on an untreated window. This result is more than twice as high as the pulsed laser damage thresholds at 1.06 µm previously reported for random ARSS on silica .
We should note that there was little to no transmission decrease through the samples with ARSS after irradiation at the power level of the laser damage threshold, while there was a 55% decrease in transmission for the traditional AR coated sample. This is very significant from an applications standpoint, since it shows that an optical element with ARSS will continue to perform in situations where a sudden loss of transmission is unacceptable, while in contrast, an AR coated (ARC) optic will experience catastrophic damage, with extensive if not total loss of functionality. Figure 6(a) shows an example of an SEM image of the surface after irradiation and (b) the smoothing of the surface as exhibited by confocal microscope measurement.
Of interest from an optical system viewpoint is the BRDF or amount of light back-scattered by windows with each type of surface treatment. The results of measurements at different wavelengths for several of the fused silica windows are shown in Fig. 7.
From these results we obtained interpolated data at 1.06 µm, and calculated the estimated power an observer would receive to the eye (assuming the observer is 1 meter away from the window, with pupil diameter of 5 mm, beam spot size of 2.5 mm at the window and a 100 kW power level from the laser). As summarized in Table 2, the estimated exposure to the eye is extremely small and also comparable to that from an AR coated or uncoated window.
3.2. Large silica windows
Significant progress has been made to scale up the random ARSS processing of silica windows for sizes used in practical applications of high-energy laser systems. Techniques for reactive ion etching of the random ARSS, including choice of gases, etching time, and plasma power were optimized for windows up to 10 cm diameter.
Figure 8 shows measured transmission results on a typical window, as obtained with a Cary spectrometer, as well as a photograph of the window, showing that negligible reflection is observed in the visible. The window transmission with random ARSS on both sides was measured with the spectrophotometer to be > 99.5%, in the range from 775 to 1350 nm. Using a laser at 1.06 µm, the transmission was measured to be 99.75% (+/− 0.18%). We note that the transmission is broadband and featureless over the 500 – 2100 nm region, which is a significant improvement over typical broadband AR coatings performance.
The laser damage threshold was also measured for the 10 cm diameter windows with random ARSS using the same pulsed laser parameters as for the 1” diameter silica samples. The measured thresholds were 90 – 100 J/cm2 for the large windows, which are similar to that for the previous samples. Thus we have shown that the ARSS fabrication process can be successfully scaled up, resulting in not only reduced reflectivity but also high laser damage thresholds for the treated surfaces.
3.3. Fused silica lenses
We have demonstrated increase in transmission through curved silica lens surfaces using the random ARSS fabrication processes developed for flat windows. Fused silica lenses of 0.5” diameter and 3 cm focal length were dry etched only on the curved, convex surface. Transmission through the curved ARSS surface increased to 99.6% at 1.06 µm from 96.6% for the uncoated surface. In addition, we also processed random ARSS on the curved surface of 5 mm diameter hemispherical lenses, to test the process capability for high-aspect ratio curved surfaces. These samples are shown in Fig. 9, where the unprocessed lens is shown on the left and the two lenses with ARSS fabricated on the facing surface are shown on the right. As shown in the photograph, the visible front surface Fresnel reflection of the overhead lights does not appear on the processed lenses. Measurements showed an increase in transmission to 99.3% through the highly curved surface after ARSS treatment. These results show that the process is suitable for highly curved or conformal windows, which are often needed for high energy laser systems and other applications.
3.4. Fused silica single mode fibers
Successful fabrication of rARSS on the end faces of single mode silica optical fibers (SMF28) has also been achieved. For processing in the reactive ion etching chamber, the ends of several fibers were mounted with epoxy in a v-groove assembly as shown Fig. 10.
Figure 11 shows a low-magnification SEM image of two unetched fiber end faces in the v-groove assembly.
As illustrated in Fig. 12, the surface morphology after etching the random ARSS on the end faces indicates a similar appearance for the core and clad areas of the fiber.
The performance of the fibers with random ARSS was excellent, whereby the measured transmission per end face increased to 99.3% at 780 nm and 99.4% at 1550 nm.
Laser damage testing was performed at 1.06 µm on the end faces of untreated silica fibers and silica fibers with ARSS on the end faces. The laser parameters were: 20 nsec pulsewidth, 20 Hz pulse repetition rate and spot size 8.7 µm (at 1/e2) which nearly matches the fiber core diameter (8.2 µm). A total of 600 laser shots irradiated the fiber end faces at increasing fluence until damage occurred. The results obtained, as summarized in Table 3, show remarkably high laser damage thresholds, up to 850 J/cm2 for silica fiber end faces with ARSS, which approaches that of the untreated fiber.
For launching into optical fiber, it is important to determine whether the reflectance for an end face with ARSS would vary with the angle (from normal incidence) of the incoming laser beam to the fiber, where for launch angle θ, the fiber numerical aperture is defined as N.A. = sinθ. We measured the transmission at 1.06 µm of a silica window with random ARSS as a function of incident angle and saw no measurable change in transmission up to N.A. = 0.4 (corresponding to launch angle of 23.5°) which is typically the largest practical N.A. for a silica optical fiber.
3.2. Spinel ceramic windows
Spinel ceramic windows for operation in the near IR (1 – 2 µm) spectral band had the ordered moth eye ARSS processing applied to the surface for reducing reflectivity in the 1-2 µm region. As shown in Fig. 13, there was significant transmission enhancement at 1.06 µm, with bulk transmission reaching 92.3% at the peak, which is nearly equal to the theoretical maximum of about 93% and corresponds to a surface transmission of about 99.1% with the moth eye ARSS. The ARSS pattern designed for 1.06 µm operation in the confocal microscope image shown in Fig. 13(b) had feature separation of about 580 nm and a height around 550 nm.
Spinel samples with and without the moth eye ARSS, as well as with AR coating (ARC) were tested for laser damage at 1.06 µm. These test results are summarized in Fig. 14, whereby a significantly improved damage threshold for spinel with moth eye surface (ARSS) was observed (10 J/cm2) compared with that of a traditional AR coated surface (4.5 J/cm2) and untreated surface (3.5 J/cm2).
3.5. Cleaning of surfaces with ARSS
One key issue often believed is that the surfaces with the ARSS cannot be cleaned due to the tiny features that could suffer mechanical damage. Thus both fused silica and spinel optical windows with ARSS were tested for ease of cleaning. The windows were mishandled on purpose (thumb fingerprint applied) and then cleaned with methanol followed by a rinse with isopropyl alcohol and de-ionized water. Transmission using an expanded-beam 1.06 µm fiber laser was measured before and after fingerprinting and cleaning, but no measureable change in transmission was observed, thus proving it is possible to clean surfaces with the AR surface structures using usual optical cleaning methods.
We have demonstrated the successful fabrication of surface AR nanostructures on a variety of fused silica optics as well as spinel ceramics for the 1-2 µm region, which are important for high-energy laser applications. Both ordered “moth eye” surface structures as well as random AR surface structures were applied. Increased transmission and laser damage thresholds were measured and compared to surfaces with traditional thin-film AR coatings. We have shown excellent results for fused silica windows with ARSS, where Fresnel losses as low as 0.02% per surface was measured. Record-high pulsed laser damage thresholds at 1.06 µm were measured on fused silica windows with random ARSS (100 J/cm2), and single mode fibers (850 J/cm2) with random ARSS on the end faces. The successful fabrication of random ARSS surfaces on highly curved 5 mm diameter silica hemispherical lenses was also demonstrated. In addition, progress for scale up of the random AR surface structures fabrication to large size silica windows (up to 10 cm diameter) has been demonstrated. For spinel, reflectivity as low as 0.9% (compared to 7% for untreated samples) was observed for surfaces with ordered moth eye ARSS. The laser damage threshold was measured to be 10 J/cm2 which was more than two times higher than that measured for an AR coated spinel window and three times higher than the untreated spinel window. Cleaning with conventional optics cleaning procedures was also demonstrated, with no change observed in the optical transmission for both silica and spinel windows with ARSS. These results significantly highlight the important utility of these AR surface structures as a highly attractive alternative to traditional AR coatings for practical applications including high-energy laser optics and windows.
The authors acknowledge support for this work from the Joint Technology Office for High Energy Lasers (JTO-HEL) and some of the ARSS fabrication as provided by TelAztec, Inc. and the laser damage testing conducted by Spica Technologies, Inc.
References and links
1. J. J. Cowan, “Aztec surface-relief volume diffractive structure,” J. Opt. Soc. Am. 7(8), 1529 (1990). [CrossRef]
3. R. J. Weiblen, C. Florea, A. Docherty, C. R. Menyuk, B. Shaw, J. Sanghera, L. Busse, and I. Aggarwal, “Optimizing motheye antireflective structures for maximum coupling through As2S3 optical fibers,” IEEE Photonics Conference, 824–825 (2012).
4. D. S. Hobbs, B. D. MacLeod, and J. Riccobono, “Update on the development of high performance anti-reflecting surface relief micro-structures,” Proc. SPIE 6545, 65450Y (2007). [CrossRef]
5. D. S. Hobbs, B. D. MacLeod, E. Sabatino III, T. M. Hartnett, and R. L. Gentilman, “Laser damage resistant anti-reflection microstructures in Raytheon ceramic YAG, sapphire, ALON, and quartz,” Proc. SPIE 8016, 801628 (2011). [CrossRef]
6. S. Bayya, G. Villalobos, W. Kim, J. Sanghera, G. Chin, M. Hunt, B. Sadowski, F. Miklos, and I. Aggarwal, “Recent developments in transparent spinel ceramic and composite windows,” Proc. SPIE 8837, 88370V (2013). [CrossRef]
7. D. S. Hobbs, B. D. McLeod, A. F. Kelsey, M. A. Leclerc, and E. Sabatino III, “Automated interference lithography systems for generation of sub-micron feature size patterns,” Proc. SPIE 3879, 124–135 (1999). [CrossRef]
8. T. Lohmueller, R. Brunner, and J. P. Spatz, “Improved properties of optical surfaces by following the example of the ‘moth eye’,” Biomimetics Learning From Nature, ed. by A. Mukherjee (Intech, 2010), Chap. 22.
9. D. S. Ruby, S. H. Zaidi, S. Narayanan, B. M. Damiani, and A. Rohatgi, “Rie-texturing of multicrystalline silicon solar cells,” Sol. Energy Mater. Sol. Cells 74(1-4), 133–137 (2002). [CrossRef]
10. G. Villalobos, S. Bayya, W. Kim, J. Sanghera, B. Sadowski, R. Miklos, C. Florea, and I. Aggarwal, (2012) “Polished Spinel Directly from the Hot Press,” in Advances in Ceramic Armor VIII (John Wiley & Sons, Inc. 2012).
11. J. P. Nole, “Novel micro-structures with high laser-induced-damage-thresholds,” SPIE Newsroom (2008). [CrossRef]