Most optoelectronic components and consumer display devices require glass or plastic covers for protection against the environment. Optical reflections from these encapsulation layers can degrade the device performance or lessen the user experience. Here, we use a highly scalable self-assembly based approach to texture glass surfaces at the nanoscale, reducing reflections by such an extent so as to make the glass essentially invisible. Our nanotextures provide broadband antireflection spanning visible and infrared wavelengths (450–2500 nm) that is effective even at large angles of incidence. This technology can be used to improve the performance of photovoltaic devices by eliminating reflection losses, which can be as much as 8% for glass encapsulated cells. In contrast, solar cells encapsulated with nanotextured glass generate the same photocurrent as when operated without a cover. Ultra-transparent windows having surface nanotextures on both sides can withstand three times more optical fluence than commercial broadband antireflection coatings, making them useful for pulsed laser applications.

Optically transparent materials, such as glass and clear plastics, are visible because the refractive index discontinuity at the material surface causes reflection of a small amount of light. In some situations, this is undesirable. To reduce unwanted reflections, surfaces are typically coated with an anti-reflective layer, whose refractive index and thickness are chosen so that the reflected waves from each interface interfere destructively. Because this approach relies on interference, it is optimized only for a single wavelength and a single incident light angle. A solution for more complete antireflection is to eliminate the discontinuity at the interface altogether by gradually tapering the refractive index between the two materials. This can be accomplished by appropriately texturing the surface at the nanoscale.

Examples of surface nanotextures in nature that reduce optical reflections include some insects' eyes, such as those of the polygonia c-aureum moth, which are decorated with dense arrays of sub-wavelength protrusions.1,2 These structures provide the animal with an evolutionary advantage by aiding vision under low-light conditions. Other insects, such as cicadas and glasswing butterflies, have evolved similarly nanotextured and highly transparent wings, which make them less visible and help them avoid detection by predators.3–5 

Nanotextures that mimic these morphologies can impart broadband and omni-directional antireflection to material surfaces. For example, silicon nanotextures reduce the average reflectance of visible light from ∼34% to less than 1%.6 When integrated into silicon solar cells, they improve light coupling into the device and thereby increase the photocurrent produced, out-performing conventional multi-layer antireflection coatings.7 

Most of today's electronic technologies, including solar cells, computer displays, phones, TVs, and LEDs, require transparent encapsulation layers for protection against the environment while also allowing light into and/or out of the device. These protective covers are commonly made of glass or plastic, which can degrade performance or lessen the user experience. For example, a glass window suffers from 8% optical loss at normal incidence (4% from each interface), a value that worsens off-axis, at larger angles of incidence. Implementing biomimetic nanotextures on transparent substrates could lead to more efficient photonic devices by, for example, increasing the collection efficiency of solar cells.8 Additionally, this technology could help to eliminate parasitic reflections from computer screens, smartphone displays, windows, and goggles.9 For a number of years, display manufacturers have been pursuing different versions of “moth-eye” nanotextures to reduce glare and reflections, with varying success.10 

Nanotextured glass surfaces are of particular interest because materials with a bulk refractive index of n=nglass1.2, the ideal index for thin-film antireflective coatings, are not readily available. Effective refractive indices neff<nglass can be achieved in coatings made of nanoscale glass-air composites, such as porous glasses11,12 or lithographically defined arrays of glass nanopillars,13 both of which have been shown to reduce optical reflections from glass surfaces. Tapered nanotextures composed of densely packed glass nanocones provide an effective-index gradient and therefore maintain their antireflective properties over a broad range of wavelengths and incidence angles.14–16 However, since these nanotextures must be deeply sub-wavelength, the challenge is finding a scalable fabrication method for imparting nanoscale patterns over macroscopic areas. To avoid time-consuming serial lithographic patterning processes, optical interference lithography and nanoimprint lithography are commonly employed.17 Most recently, different self-assembly based approaches have been developed that promise truly wide-area nanotexturing, with pattern densities not achievable by other methods.18–24 

In this work, we have used a highly scalable self-assembly based approach to texture glass surfaces, reducing reflections from both the front and rear interfaces to less than 0.2% over the entire visible and near-infrared spectrum, which renders the glass essentially invisible [Fig. 1(a)]. We accomplish this by leveraging the pattern-forming abilities of block copolymers–an important class of industrial polymers whose physical properties can be straightforwardly tuned. The block copolymer macromolecular architecture facilitates their robust self-assembly into periodic nanopatterns with characteristic length scales in the range of nanometers that are uniform over areas as large as ∼m2.25 A significant advantage of block copolymer-based patterning is its compatibility with thin-film manufacturing processes, which is a main reason it is widely used among academic nanoscience researchers as well as in the microelectronics industry. Block copolymer assembly schemes based on lamellar forming copolymers19 and block copolymer micelles18,22 have previously been used with good success in patterning anti-reflection subwavelength nanotextures in glass. Here, we detail an approach using cylindrical phase materials, which provides high performance nanotextures, texture profile tenability, and process simplicity/robustness. Our nanopatterning technique is readily integrated into semiconductor wafer-processing because all materials used are compatible with semiconductor electronics manufacture. Additionally, the deeply sub-wavelength feature size of our nanotextures enables accurate theoretical modeling of their optical properties based on the effective medium picture. Previously, we applied a similar technique to texture silicon surfaces.26 To texture fused silica surfaces, a modified fabrication procedure is required.

FIG. 1.

Surface nanotextures lead to reduced reflectance. (a) Comparison of treated (left) and untreated (right) fused silica slides illuminated by a desk lamp, showing greatly reduced surface reflections from the treated sample. (b) Cross-sectional Scanning Electron Micrograph (xSEM) of a cylindrical phase PS-b-PMMA film on a silicon substrate after phase separation. The PMMA has been chemically removed to provide SEM contrast. (c) xSEM image of alumina nanostructures fabricated on a silicon substrate after infiltration and polymer removal. (d) xSEM of a fused silica surface textured with 170-nm-tall cones.

FIG. 1.

Surface nanotextures lead to reduced reflectance. (a) Comparison of treated (left) and untreated (right) fused silica slides illuminated by a desk lamp, showing greatly reduced surface reflections from the treated sample. (b) Cross-sectional Scanning Electron Micrograph (xSEM) of a cylindrical phase PS-b-PMMA film on a silicon substrate after phase separation. The PMMA has been chemically removed to provide SEM contrast. (c) xSEM image of alumina nanostructures fabricated on a silicon substrate after infiltration and polymer removal. (d) xSEM of a fused silica surface textured with 170-nm-tall cones.

Close modal

To produce these highly transmissive surface nanotextures, we begin by coating a fused silica substrate with a cylindrical-phase polystyrene-b-poly(methyl methacrylate) (PS-b-PMMA) block copolymer thin film, which undergoes thermally driven phase separation to self-organize into locally ordered hexagonal arrangements of PMMA cylinders within a PS matrix [Fig. 1(b), also see the details at the end of this letter.]. Block-selective infiltration of the organometallic precursor tri(methyl aluminium) into PMMA followed by reaction with H2O vapor, a process known as sequential infiltration synthesis,27 preferentially load aluminum oxide within the PMMA domains. Subsequent removal of all organics from the surface by oxygen plasma leaves behind fused silica substrates decorated with ∼20 nm tall aluminum oxide nanoparticles arranged in a hexagonal pattern with a ∼52 nm separation [Fig. 1(c)]. This pattern is then transferred to the fused silica substrate by anisotropic inductively coupled plasma etching using fluoroform chemistry. With the increasing etching time, the height of the transferred pattern increases while gradually removing the alumina nanostructures, which serve as a hard mask for the plasma etching. The resulting nanotextured fused silica surface contains an extremely high density of nanoscale cones (4.6×1014 m−2) having sharp tips with average diameter <10 nm [Fig. 1(d)]. These patterns are uniform over large areas, limited only by the size of the equipment used for processing. In the results reported here, we have worked exclusively with round, 5 cm-diameter fused silica wafers. Texturing the glass surface in this manner creates an air-glass interface region where the effective refractive index changes continuously between that of air and that of the glass. For properly optimized nanotexture dimensions, the texture provides broadband and omnidirectional antireflection spanning the entire visible and near-infrared spectrum and effective over a wide range of angles of incidence.

Surface nanotextures with increasing heights impart progressively higher levels of antireflection to the interface, with 300-nm-tall structures reflecting less than 0.15% of the incident light at a wavelength of 633 nm [symbols in Fig. 2(a)], in agreement with theoretical calculations based on numerically solving Maxwell's equations in a stratified medium with an anisotropic effective refractive index [dashed line in Fig. 2(a), see details at the end of this letter.]. The evolution of the shape of the fused-silica cones (blue) and alumina nanostructures (gray) with the increasing nanotexture height is also shown schematically, based on measurements from cross-sectional scanning electron micrographs. The calculated in-plane refractive index profiles for short (60 nm), medium (150 nm), and tall nanotextures (280 nm) [Fig. 2(b)] illustrate that taller nanotextures offer a more gradual transition from the refractive index of air to that of glass, in correspondence with the reduced reflections from the interface.

FIG. 2.

Antireflective properties of glass nanostructures. (a) Measured and simulated reflectance of fused silica surfaces decorated with nanotextures of different heights at λ = 633 nm. (b) Effective refractive index profiles calculated by our model for 60-nm, 150-nm, and 280-nm-tall nanotextures compared to an air-glass interface.

FIG. 2.

Antireflective properties of glass nanostructures. (a) Measured and simulated reflectance of fused silica surfaces decorated with nanotextures of different heights at λ = 633 nm. (b) Effective refractive index profiles calculated by our model for 60-nm, 150-nm, and 280-nm-tall nanotextures compared to an air-glass interface.

Close modal

Creating surface nanotextures on both sides of a piece of glass results in extremely high light transmission over a broad wavelength range and across a wide range of angles of incidence (Fig. 3). Our optical simulations predict well the normal-incidence transmission spectra for a fused silica substrate with both surfaces structured into 260-nm-tall textures, which transmits on average 99.7% of visible light (390–700 nm). The glass is almost invisible, having a maximum transmission at 656 nm of 99.8%. These substrates retain enhanced transmission even at long optical wavelengths (we measured out to 2.5 μm) and at large incident angles. The average transmission of visible light remains higher than 90% even at 70° from the surface normal [Fig. 3(b)]. Nanotextured glass textured in this manner can find use as an ultra-transparent cover for consumer display devices, for glare-free operation over a wide range of viewing angles, and for reducing power consumption. Note that the refractive index gradient provided by the nanotextured surface is not affected by particulate contamination; dust particles are typically 10s of μm in size and therefore cannot occupy the space between neighboring cones. Additionally, such surfaces can be made to be superhydrophobic and self-cleaning.11,13,14,26,28–31

FIG. 3.

Glass slides treated on both sides become extremely transparent. Comparison of experimental (points) and simulated (dashed) transmissivities of nanotextured (orange) and flat (cyan) fused silica slides shows improved performance across the entire visible and near-IR region (a) and for all incidence angles (b).

FIG. 3.

Glass slides treated on both sides become extremely transparent. Comparison of experimental (points) and simulated (dashed) transmissivities of nanotextured (orange) and flat (cyan) fused silica slides shows improved performance across the entire visible and near-IR region (a) and for all incidence angles (b).

Close modal

Nanotextured glass can provide an immediate benefit as a solar cell encapsulation layer that does not degrade device performance, as these textures perform best at visible and UV wavelengths. Here, we compared the performance of a commercial silicon solar cell without a cover (Fig. 4, black), covered with a piece of conventional glass (cyan), and covered with a piece of nanotextured glass (orange). The uncovered device short circuit current of 27.1 mA cm−2 under simulated AM1.5G illumination (calibrated against a certified silicon reference cell) was essentially unchanged when covered by nanotextured glass but reduced to 25.6 mA cm−2 under the same conditions when covered by conventional glass. The corresponding solar cell power conversion efficiency for an uncovered device is 8.12%, which becomes 8.07% when covered with nanotextured glass and reduces to 7.68% when covered with conventional glass.

FIG. 4.

Nanotextured glass as a solar cell encapsulation layer. Illuminated and unilluminated current-voltage characteristics of a commercial polycrystalline silicon solar cell operated without a cover (black line), covered with an untreated glass slide (cyan), and covered with nanotextured glass (orange).

FIG. 4.

Nanotextured glass as a solar cell encapsulation layer. Illuminated and unilluminated current-voltage characteristics of a commercial polycrystalline silicon solar cell operated without a cover (black line), covered with an untreated glass slide (cyan), and covered with nanotextured glass (orange).

Close modal

Endowing optical windows with high transmissivity is useful for pulsed laser applications as well. However, conventional multi-layer antireflective coatings are easily damaged by high fluence laser light, and the porous silica sol-gel coatings historically used for laser line applications suffer from environmental degredation.32 Surface relief anti-reflective structures etched into the glass are promising alternatives to sol-gel coatings and have been shown to withstand high fluence laser pulses.22,33 The broadband and omnidirectional nature of the antireflection imparted by our method makes nanotextured glass an attractive option for high-power laser applications that operate over a broad wavelength range. Measurements of the laser-induced damage threshold of a nanotextured glass slide (performed by Quantel, see details at the end of this letter) indicate that these nanotextures can withstand up to 14 J/cm2, or 1633.0 MW/cm2, of 7-ns-long laser pulses at a wavelength of 532 nm. This is approximately three times more optical fluence than commercial multi-layer broadband antireflection coatings, whose damage thresholds under similar conditions are typically in the range of 2–4 J/cm2.

We have demonstrated a nanotexturing approach to rendering glass substrates extremely transparent, which reduces surface reflections to almost zero. We have endowed antireflective properties to fused silica windows not by coating the substrate with thin layers of different materials but by changing the geometry of the glass surface at the nanoscale. The resultant structure is composed entirely of fused silica. We therefore expect our nanotextures to be more tolerant to extreme environments (elevated temperature, high vacuum, etc.) than traditional thin-film antireflective coatings. Nanotextured glass outperforms commercially available broadband antireflective coatings with regard to both the optical reflections and the laser damage threshold (see Table I). The self-assembly based production of surface nanotextures is scalable and compatible with thin-film manufacturing processes. Additionally, the antireflective properties of sub-wavelength structures are largely insensitive to imperfections in the self-assembled template, and therefore, our nanotextures can be implemented with current block copolymer processing technology. Nanotextured glass can be used as an ultra-transparent encapsulation layer for solar cells and other optoelectronic devices with significant improvement of performance or as an optical window for broad-band laser applications and wide-angle (high numerical aperture) optical systems.

TABLE I.

Comparison of the nanotextured glass reported here with commercially available broad-band antireflective coatings.34Ravg(0°) is the reflectance of visible light, averaged over the wavelength range indicated, and measured at or near normal incidence, while Ravg(45°) is the reflectance of unpolarized light incident at 45° to the surface normal, where available. The threshold fluence for laser-induced damage (LIDT) of these coatings is also quoted.

Ravg(0°)Ravg(45°)LIDTa
Newport AR.14 (430–700) <0.5% … 2 J/cm2 
CVI HEBBARTM (415–700) <0.4% <1.5% 3.8 J/cm2 
Thorlabs A coat (390–700) <0.4% <2.2% 7.5 J/cm2 
Nanotextured glass (390–700) <0.2% <0.4% 14 J/cm2 
Ravg(0°)Ravg(45°)LIDTa
Newport AR.14 (430–700) <0.5% … 2 J/cm2 
CVI HEBBARTM (415–700) <0.4% <1.5% 3.8 J/cm2 
Thorlabs A coat (390–700) <0.4% <2.2% 7.5 J/cm2 
Nanotextured glass (390–700) <0.2% <0.4% 14 J/cm2 
a

With 10 ns pulses at 532 nm.

To pattern the surface of fused silica substrates, we first spin-coat a thin film of cylindrical-phase polystyrene b-poly(methyl methacrylate) (PS-b-PMMA) block copolymer (PS:PMMA 64:35, PDI = 1.09) and anneal it at 220 °C for 5 min. To facilitate the orientation of the cylindrical copolymer domains perpendicular to the substrate surface, prior to block copolymer coating, we form a PS-r-PMMA-OH random copolymer brush [61 mol. % styrene, determined by 13C NMR, Mn = 9.2 kg/mol and MW/Mn = 1.35 (determined by gel permeation chromatography relative to PS standards), provided by The Dow Chemical Company] on the glass by spin casting at 600 RPM for 45 s, annealing for 5 min, and rinsing the substrate with toluene. PS-b-PMMA templates were converted into aluminum oxide nanostructures by sequential infiltration synthesis using a commercial atomic layer deposition system (Cambridge Nanotech Savannah S100). The samples were exposed to eight successive cycles of trimethylaluminium (TMA) (300 s, >5 Torr) and water vapor (300 s, >5 Torr) at 85 °C. Between TMA and water vapor exposures, the sample chamber is purged with N2 (20  sccm, 5 m) to remove any unattached molecules. After infiltration, all remaining organic material was removed by 3 min of oxygen plasma (20 W RF power, 100 mTorr) (March Plasma CS1701), leaving fused silica substrates decorated with ∼20-nm-tall aluminum oxide nanodots, locally ordered into a hexagonal array with a ∼52 nm pitch. These structures serve as hard masks during the anisotropic etching of the fused silica substrates, which was performed using a CHF3 and O2 gas mixture (3:1) using an inductively coupled plasma reactive ion etching machine (Oxford Plasmalab 100 ICP etcher) at 15 mTorr, with a RF power of 40 W and 700 W ICP at room temperature.

To model the optical properties of nanotextured glass, we utilize a 4 × 4 matrix approach that numerically solves Maxwell's equations in a stratified medium. We approximate the nanotextured surface by a stack of 1-nm-thick strata, each characterized by an anisotropic effective dielectric tensor. The fill fraction of glass in air within each stratum is calculated from measurements of scanning electron micrographs. The in-plane components of the effective dielectric tensor are described by the Maxwell-Garnett effective-medium model, while the out-of-plane component is given by a volume average. For short nanotextures, the surviving alumina nanostructures are treated in a similar manner.

We measured the optical transmission and reflection of nanotextured glass slides using a broad-band CW laser-driven light source (Energetiq EQ-99) coupled to a commercial Fourier-transform spectrometer (Bruker v80). For transmission measurements, the spectrometer's internal sample chamber was used. For measurements of reflectance, light was focused onto the sample using an upright optical microscope (Nikon FN-1). A glass prism and index matching fluid were used to suppress reflectance from the rear (untextured) glass surface. The optical power reflected by the textured surface was measured using a silicon detector (Thorlabs, DET100A) and calibrated against a broad-band silver mirror (Thorlabs PF10-03-P01). We measured the current-voltage characteristics of a 1.5″ × 0.75″ Si solar cell in the dark and under simulated AM1.5G illumination at normal incidence using a Keithley 2401 source meter. Laser-induced damage threshold tests of a nanotextured glass slide were performed by Quantel USA. The sample was exposed to 7-ns-FWHM, 532-nm optical pulses, with a repetition rate of 20 Hz. The laser had a TEM00 profile with a spot diameter of 0.43 mm (FW/e2) and was directed at normal incidence to the sample surface. 120 sites on the sample were tested, with 200 shots per site. The damage threshold was determined by the axis-intercept of a linear least-squares fit to the damage frequency (see supplementary material).

See supplementary material for the details of laser damage threshold tests on a nanotextured glass slide.

This research used resources of the Center for Functional Nanomaterials, which is a U.S. DOE Office of Science User Facility, at Brookhaven National Laboratory under Contract No. DE-SC0012704.

1.
P.
Vukusic
and
J. R.
Sambles
,
Nature
424
,
852
(
2003
).
2.
P. B.
Clapham
and
M. C.
Hutley
,
Nature
244
,
281
(
1973
).
3.
A.
Yoshida
,
M.
Motoyama
,
A.
Kosaku
, and
K.
Miyamoto
,
Zool. Sci.
14
,
737
(
1997
).
4.
G. S.
Watson
and
J. A.
Watson
,
Appl. Surf. Sci.
235
,
139
(
2004
).
5.
R. H.
Siddique
,
G.
Gomard
, and
H.
Hölscher
,
Nat. Commun.
6
,
6909
(
2015
).
6.
S.-Y.
Chuang
,
H.-L.
Chen
,
J.
Shieh
,
C.-H.
Lin
,
C.-C.
Cheng
,
H.-W.
Liu
, and
C.-C.
Yu
,
Nanoscale
2
,
799
(
2010
).
7.
A.
Rahman
,
A.
Ashraf
,
H.
Xin
,
X.
Tong
,
P.
Sutter
,
M. D.
Eisaman
, and
C. T.
Black
,
Nat. Commun.
6
,
5963
(
2015
).
8.
D.
Huh
,
J.-H.
Shin
,
M.
Byun
,
S.
Son
,
P.-H.
Jung
,
H.-J.
Choi
,
Y.-D.
Kim
, and
H.
Lee
,
Sol. Energy Mater. Sol. Cells
170
,
33
(
2017
).
9.
G.
Tan
,
J.-H.
Lee
,
Y.-H.
Lan
,
M.-K.
Wei
,
L.-H.
Peng
,
I.-C.
Cheng
, and
S.-T.
Wu
,
Optica
4
,
678
(
2017
).
10.
M.
Nishisaka
, www.eetimes.com/document.asp?doc_id=1262615 for Sharp tackles reflected light with ‘moth eye’ LCD;
D.
Katzmaier
, www.cnet.com/products/samsung-un55ks9500f for Samsung SUHD TVs chase OLED with quantum dots, moth's eyes;
S.
Miles
, www.pocket-lint.com/news/116974-philips-8000-series-amblilgiht-ifa-tv for Philips 9000 series TV copies moth's eye for screen tech.
11.
T.
Aytug
,
A. R.
Lupini
,
G. E.
Jellison
,
P. C.
Joshi
,
I. H.
Ivanov
,
T.
Liu
,
P.
Wang
,
R.
Menon
,
R. M.
Trejo
,
E.
Lara-Curzio
 et al,
J. Mater. Chem. C
3
,
5440
(
2015
).
12.
A.
Kumar
,
S.
Siddhanta
, and
H. C.
Barshilia
,
Sol. Energy
129
,
147
(
2016
).
13.
J.
van de Groep
,
P.
Spinelli
, and
A.
Polman
,
Nano Lett.
15
,
4223
(
2015
).
14.
K.-C.
Park
,
H. J.
Choi
,
C.-H.
Chang
,
R. E.
Cohen
,
G. H.
McKinley
, and
G.
Barbastathis
,
ACS Nano
6
,
3789
(
2012
).
15.
S.
Ji
,
K.
Song
,
T. B.
Nguyen
,
N.
Kim
, and
H.
Lim
,
ACS Appl. Mater. Interfaces
5
,
10731
(
2013
).
16.
H.
Xu
,
L.
Liu
,
F.
Wu
,
D.
Xu
, and
N.
Lu
,
RSC Adv.
5
,
28014
(
2015
).
17.
H.
Deniz
,
T.
Khudiyev
,
F.
Buyukserin
, and
M.
Bayindir
,
Appl. Phys. Lett.
99
,
183107
(
2011
).
18.
T.
Lohmüller
,
M.
Helgert
,
M.
Sundermann
,
R.
Brunner
, and
J. P.
Spatz
,
Nano Lett.
8
,
1429
(
2008
).
19.
B.
Päivänranta
,
P. K.
Sahoo
,
E.
Tocce
,
V.
Auzelyte
,
Y.
Ekinci
,
H. H.
Solak
,
C.-C.
Liu
,
K. O.
Stuen
,
P. F.
Nealey
, and
C.
David
,
ACS Nano
5
,
1860
(
2011
).
20.
S.
Ji
,
J.
Park
, and
H.
Lim
,
Nanoscale
4
,
4603
(
2012
).
21.
E.
Wang
and
Y.
Zhao
,
Opt. Lett.
39
,
3748
(
2014
).
22.
Z.
Diao
,
M.
Kraus
,
R.
Brunner
,
J.-H.
Dirks
, and
J. P.
Spatz
,
Nano Lett.
16
,
6610
(
2016
).
23.
P.
Buskens
,
M.
Burghoorn
,
M. C. D.
Mourad
, and
Z.
Vroon
,
Langmuir
32
,
6781
(
2016
).
24.
P.
Mokarian-Tabari
,
R.
Senthamaraikannan
,
C.
Glynn
,
T. W.
Collins
,
C.
Cummins
,
D.
Nugent
,
C.
O'Dwyer
, and
M. A.
Morris
,
Nano Lett.
17
,
2973
(
2017
).
25.
C. T.
Black
,
R.
Ruiz
,
G.
Breyta
,
J. Y.
Cheng
,
M. E.
Colburn
,
K. W.
Guarini
,
H.-C.
Kim
, and
Y.
Zhang
,
IBM J. Res. Dev.
51
,
605
(
2007
).
26.
A.
Checco
,
A.
Rahman
, and
C. T.
Black
,
Adv. Mater.
26
,
886
(
2014
).
27.
Q.
Peng
,
Y.-C.
Tseng
,
S. B.
Darling
, and
J. W.
Elam
,
Adv. Mater.
22
,
5129
(
2010
).
28.
K.
Ogawa
,
M.
Soga
,
Y.
Takada
, and
I.
Nakayama
,
Jpn. J. Appl. Phys., Part 2
32
,
L614
(
1993
).
29.
D.
Tulli
,
S. D.
Hart
,
P.
Mazumder
,
A.
Carrilero
,
L.
Tian
,
K. W.
Koch
,
R.
Yongsunthon
,
G. A.
Piech
, and
V.
Pruneri
,
ACS Appl. Mater. Interfaces
6
,
11198
(
2014
).
30.
P.
Mazumder
,
Y.
Jiang
,
D.
Baker
,
A.
Carrilero
,
D.
Tulli
,
D.
Infante
,
A. T.
Hunt
, and
V.
Pruneri
,
Nano Lett.
14
,
4677
(
2014
).
31.
X.
Men
,
X.
Shi
,
B.
Ge
,
Y.
Li
,
X.
Zhu
,
Y.
Li
, and
Z.
Zhang
,
Chem. Eng. J.
296
,
458
(
2016
).
32.
I. M.
Thomas
,
Proc. SPIE
2114
,
232
(
1994
).
33.
D. S.
Hobbs
,
B. D.
MacLeod
,
E.
Sabatino
,
J. A.
Britten
, and
C. J.
Stolz
,
Proc. SPIE
8885
,
88850J
(
2013
).
34.
See http://assets.newport.com/pdfs/e3885.pdf for Newport AR.14; See www.cvilaseroptics.com/file/general/Coating.pdf for CVI HEBBARTM: Thorlabs A (350–700 nm) Broadband AR Coating: Data Provided by Thorlabs.

Supplementary Material