The influences of deposition rate and alloying with aluminum (0 to 65 at.%) on the structure and optical properties of magnesium films on quartz are examined. Results show that increasing the deposition rate improves the plasmonic figure of merit (FOM) values, and decreases the surface roughness and grain size. Addition of Al to Mg blue shifts the localized surface plasmon (LSP) figure of merit peak position and improves LSP FOM values over portions of the ultraviolet (UV) region. Nanotriangles (NTs) of Mg and Mg-Al on quartz were fabricated using colloidal lithography. Mg-Al NTs show improved LSP resonance (LSPR) response compared to Mg particles due to improved structural integrity.
© 2016 Optical Society of America
The use of plasmonics in the UV spectral range has often been overlooked, but is of interest as many important chemical/biochemical reaction steps are associated with these energies. UV plasmonic structures are therefore very attractive for use in enhanced chemical synthesis and other photochemical processes in addition to UV spectroscopy for sensing and detection as well as energy harvesting applications . Al is often considered the best plasmonic metal across the UV range due to its high plasma frequency and absence of interband transitions in this region . However, Al suffers from drawbacks including a large electron damping factor and the formation of a very stable oxide layer that is of concern in both fabrication and applications. In our earlier work [3,4], it was demonstrated that Mg also has potential as an alternative plasmonic material in the UV region from 280 to 400 nm. For instance, the figure of merit of localized surface plasmons using Mg films is more than 50% higher than with Al films at near-UV wavelengths.
Researchers [5,6] have proposed that alloying metals (like Cu, Ag, Au) with a small volume fraction of another metal modifies the metal’s reflection and absorption properties as observed by spectral characterization. Alloying of monovalent metals such as Ag and Au with bi- or tri-valent metals increases the free electron density. This raises the Fermi level which in turn increases the d→sp transition energy. Using Finite-Difference Time-Domain (FDTD) calculations, Chowdhury et al.  showed shifting of the interband transition in Ag by the addition of trivalent metal Al. Bobb et al.  have shown shifts in interband transitions in Au by alloying with bivalent transition metals (e.g. Cd, Zn) . Cd and Zn contribute one extra electron to the free-electron plasma in Au thereby increasing the electron density. In the case of noble metal alloy thin-films, the dielectric function can be manipulated by tuning the concentration of two individual metals, which in turn modifies the strength of the polarization (ε1) induced by an external electric field and the losses (ε2) in the material due to absorption . In addition, researchers [10–12] have shown that by adding small amounts of Al into Ag, the morphology and optical properties of ultra-thin Ag films can be significantly improved. In this work, the influence of alloying Mg with Al on the structural characteristics and the optical behavior of Mg is examined. This work also examined the localized surface plasmon resonance (LSPR) response of Mg-Al alloy nanotriangles on quartz substrates compared to pure Mg and pure Al nanotriangles.
2. Experimental Section
Mg-Al alloy thin films were deposited on fused quartz substrates by co-sputtering of Mg and Al using magnetron sputtering in an ultra-high vacuum (UHV) chamber. Air Products Research grade argon gas (>99.9995% purity) was used for the sputtering process. The Mg target was machined from 99.99 + % purity Mg discs that had a diameter of 50 mm and thickness of 6.25 mm. The Al sputtering target, with a diameter of 50 mm, thickness of 6.25 mm, and a purity of 99.9995%, was obtained from Alfa Aesar.
The sputtering chamber was evacuated using a turbo pump to a base vacuum level of < 3 × 10−7 Torr. Argon gas was then introduced at a flow rate of 3.0 × 10−2 cc/min and maintained for 3 hours with the turbo pump on. The sputtering was then carried out at a pressure of 2.5 x 10−3 Torr. Al and Mg films were co-sputtered using an Advanced Energy 500 W DC sputtering power supply and a 500 W RF power supply, respectively. The targets were sputter cleaned for 5 minutes prior to each deposition to remove the oxide film on the target surface. Mg-Al films were deposited on optically smooth surfaces, 25 mm × 25 mm polished fused quartz substrates. The compositions of the alloys were controlled by varying the power for Al and Mg sputtering.
The samples were retained in the deposition chamber under vacuum until their transfer to the ellipsometry measurement system. For transfer, the samples were removed from the deposition chamber and sealed in polyethylene bags under an argon atmosphere. The transfer time was typically ~30 minutes. The optical measurements were then carried out using a Variable Angle Spectroscopic Ellipsometer (VASE) made by J.A. Woollam Co [13,14]. Since the samples were briefly exposed to atmosphere before and during the VASE measurement, we expect the formation of a few nm thick MgO-Al2O3 on the thin metal film surface. We used a Drude model for fitting the optical properties and considered the Mg alloy film and substrate in the model. Measurements in this work were made over the wavelength range of 250 to 1700 nm at incident angles of 70° and 75°.
The surface morphology of the films was analyzed using a high resolution FEI Nova Nano 630 Scanning Electron Microscope (Nova NanoSEM). The surface roughness of the films was analyzed using a Bruker Dimension ICON-PT atomic force microscope (AFM), which was used in the contact mode with a nominal tip radius of curvature of 2 nm. The average compositions of the alloy thin films were analyzed using Energy Dispersive Spectroscopy (EDS) in a Nova NanoSEM. The standardless quantification routine of the EDAX Genesis software was used for the analysis. Three measurements were made on each of the samples in adjacent location to obtain standard deviation of the measured composition. The crystalline structure and crystallographic texture were analyzed using x-ray diffraction measurements using a Siemens D5000 x-ray diffractometer.
2.1 Fabrication of nanostructures
From the optical constant measurements, it was observed that Mg-Al alloy films have improved LSP figure of merit values compared to Al in some portions of the UV range. To examine the localized surface plasmon (LSP) response of Mg-Al alloy films, nanotriangles were fabricated using nanosphere lithography , also commonly known as colloidal lithography. Unlike electron beam lithography (EBL), it is inexpensive and is easily applied over large substrate areas. This fabrication method also involves a very limited number of steps, lowering the risk of oxidation of the Mg film. Nevertheless, the alloy films would be compatible with other patterning methods making use of a lift-off step, such as electron beam lithography.
2.1.1 Preparation of close packed polystyrene (PS) beads
Water with a resistivity of about 18 MΩ obtained after passing through a Barnstead NANOpure Diamond filtration system was used. Chemicals used were reagent grade and purchased from Fisher Scientific unless otherwise noted. Fused silica quartz slides were first cleaned in piranha acid, 3:1 concentrated sulfuric acid (Macron Fine Chemicals), 30% H2O2, for 45 min followed by sonication in a basic solution (5:1:1 water, concentrated ammonium hydroxide, 30% H2O2) for 1 hour at 60°C. Slides were then placed under water in a petri dish that was elevated by 5°. Aldehyde functionalized polystyrene (PS) macrospheres (Life Technologies) were diluted 50:50 in 200 proof ethanol (Decon) and introduced into the petri dish using a microliter syringe pump. Once a monolayer of PS beads was formed at the air-water interface, the water was pumped out of the petri dish using a peristaltic pump, leaving behind a close packed monolayer on the clean quartz slides and the slides were left to dry in an ambient environment [16,17].
(Caution: Piranha solution is a strong oxidizing agent, which has been known to detonate spontaneously upon contact with organic material, and should be handled with extreme care.)
2.1.2 Fabrication of nanotriangles (NTs)
After drying, a thin film of required material (Mg or Al) of thickness 50 nm was magnetron sputter deposited (described in section 2) on the quartz slides containing a monolayer of close packed PS beads of desired size. The PS beads were selectively removed using adhesive tape. The resulting size of the nanotriangles was varied by using different sizes of the PS beads .
3. Influence of deposition rate
Mg films were deposited at different deposition rates (1.8, 5.2, and 7.0 Å/s) on fused quartz substrates. The nominal thicknesses of the Mg films were 100 nm as measured using stylus profilometry. Various deposition rates were achieved by changing the RF sputtering power for Mg deposition from 30 to 84 W. Figure 1 shows AFM images of Mg films prepared with different deposition rates of (a) 1.8, (b) 5.2, and (c) 7.0 Å/s. The RMS roughness (Rq) values observed for Mg films with different deposition rates (1.8, 5.2, and 7.0 Å/s) are 3.7, 2.7, 2.3 nm, respectively. Rq values are the standard deviations from the mean plane and represent the roughness. The result suggests that surface roughness values tend to decrease as the deposition rate increases.
Quantification of the grain size was difficult because of the change in morphology of the film structure at the higher deposition rate, but a decrease in the grain size of the Mg film with an increase in the deposition rate could be qualitatively seen (Fig. 1). Other work on metallic thin films has shown that a decrease in grain size can be attained by (i) faster deposition rate of the metallic film which reduces the time that metal atoms can diffuse on the substrate and leads to smaller grains, especially when the substrate is at ambient temperature conditions [19,20], and (ii) the presence of residual gas molecules present in the deposition chamber, especially for metals with high reactivity like Al and Mg, which tend to reduce grain growth. Here the depositions of Mg films were carried out in a UHV chamber with a base pressure of < 2 × 10−7 Torr, so the probability of the formation of an oxide layer and pinning of the grain boundaries is very low .
Figure 2(a) and 2(b) shows the optical constants ε1 and ε2 for the Mg films as a function of incident wavelength, where ε1 (Eq. (1)) and ε2 (Eq. (2)) are the real and imaginary parts of the dielectric function (Eq. (3)) [13,14,22].23], while for surface plasmon polaritons (SPP) applications, it is given by FOMSPP = β1/β2, where β (Eq. (4)) is the complex propagation constant of the SPP at the metal/air interface, and β1 and β2 are the real and imaginary components, respectively .25,26]. On the contrary, in the case of Fig. 2(b), the Mg films with finer grain size (5.2 and 7.0 Å/s) seem to have lower losses when compared to the film with bigger grain size (1.8 Å/s). However, there is a reduction in surface roughness of the films which contributes to the improvement in the FOMLSP values.
Figure 2(c) shows a plot of FOMLSP versus wavelength and Fig. 2(d) shows a plot of FOMSPP versus wavelength observed in the case of Mg films deposited at different rates. The Mg film prepared with a 7.0 Å/s deposition rate gives higher FOMLSP of about 8.3 at 370 nm, when compared to 7.9 at 380 nm (5.2 Å/s) and 6.6 at 360 nm (1.8 Å/s). FOMSPP values are also significantly improved for films with higher deposition rates (5.2 and 7.0 Å/s) when compared to the lower deposition rate (1.8 Å/s) especially at wavelengths beyond 300 nm (Fig. 2(d)).
4. Optical response of Mg-Al alloy thin films
The optical responses of Mg-Al alloy films are presented in this section. For the purpose of analysis, the Mg-Al alloy thin films were divided into two categories: (i) Mg-rich (Mg > 50 at.%) and (ii) Al-rich (Al > 50 at.%). The nominal thicknesses of all the Mg alloy thin films were 100 nm.
4.1. Mg-Al alloy thin films
Two intermetallic compound phases are observed in the Al-Mg binary system, namely Al3Mg2 (β), and Al12Mg17 (γ). Al3Mg2 has a Cd2Na type structure while the Al12Mg17 phase has an α-Mn (A12) type structure (Table 1) . Preparation of Al3Mg2 and Al12Mg17 films of exact stoichiometry using the co-sputtering approach required several trials and was challenging. In this work, Mg-Al films with Al content in the 0 to 65 at.% range were examined. The films were prepared by varying relative sputtering powers used for Mg and Al sputtering. Table 2 shows the compositions of various films obtained and the sputtering powers used for Mg and Al. Figure 3 shows the XRD patterns of both Mg-rich and Al-rich Mg-Al alloy films on quartz substrates.At room temperature, the Mg-65 at.% Al film composition lies in the Al (α) and Al3Mg2 (β) two phase region and the film is expected to contain predominantly the Al3Mg2 (β) (~90%). The Mg-59 at.% Al film composition lies in the Al3Mg2 (β) single phase region. The XRD patterns (Fig. 3) for both Mg-65 at.% Al (35Mg-65 Al) and Mg-59 at.% Al (41 Mg- 59 Al) alloys do not show any peak corresponding to the Al3Mg2 (β) phase likely due to preferred orientation present in the sample where the grain orientations are such that the Bragg condition is not satisfied in the θ-2θ scan configuration. While it is possible that growth is preferred along distinct directions, even with small deviation of this preferred growth direction from film normal, peak intensity can be dramatically reduced with relatively small film thickness levels. The other possibility is that the film is amorphous. For films examined with Mg content from 48 to 59 at.%, a two phase mixture containing Al3Mg2 (β) and Al12Mg17 (γ) is expected. In this composition range, the amount of Al12Mg17 (γ) phase increases to ~95% as Mg content increases to 58 at.%. In the XRD patterns for alloys in this composition range of 48 to 59 at.% Mg, a peak at 2θ value of 35.36° is observed. This peak corresponds to (400) peak from the Al12Mg17 (γ) phase grains with its intensity increasing as Mg content is increased. This suggests strong (400) preferred orientation of the γ phase which is based on bcc type Bravais lattice. For films examined with Mg contents from 70 to 97 at.%, two phase mixture of Al12Mg17 (γ) and Mg (α) phase is expected, with the amount of γ phase decreasing and consequently the decrease in the intensity of (400) peak of the γ phase as Mg content increases. At these high Mg contents, peaks at 34.71° which corresponds to the (0002) peak of Mg phase and its position is close to that of the (400) peak of the γ phase is observed. At Mg contents close to 97 at.%, the film mainly consists of pure Mg phase and peak count rates for the (0002) are similar to the observed values for  textured pure Mg films.
4.1.1. Structure and optical responses of Mg-rich Mg-Al films
Figure 4 shows the AFM images of Mg-rich Mg-Al alloy films on fused quartz substrates. The RMS roughness values observed for Mg-rich alloy films are also shown in Fig. 4. The surface roughness does not seem to vary significantly. For the purpose of analysis, Mg-rich Mg-Al alloy films are grouped into two categories (group 1 and 2) based on their deposition rates as highlighted in Table 2. Group 1 has 51Mg-49Al, 58Mg-42Al, and 70Mg-30Al and group 2 has 81Mg-19Al, 86Mg-14Al, 91Mg-9Al, and 97Mg-3Al.
Figure 5(a) and 5(b) show the optical constants ε1 and ε2 as a function of incident wavelength for group 1 Mg-rich Mg-Al alloy thin films. Figure 5(a) shows that the ε1 values for group 1 Mg-rich Mg-Al alloy films lie between those of pure Al and Mg films. The same trend is seen for the ε2 values for wavelengths above 300 nm (Fig. 5(b)). Figure 5(c) shows a plot of the FOMLSP versus wavelength and Fig. 5(d) shows a plot of the FOMSPP versus wavelength. FOMLSP values increases with an increase in Mg content and the peak in the FOMLSP curve blue shifts from 360 nm for pure Mg (1.8 Å/s) to 320 nm (a shift of about 40 nm) for 70Mg-30Al film. Both 58Mg-48Al and 70Mg-30Al alloy films have slightly higher FOMLSP than pure Mg film at the short wavelengths (below about 300 nm) due to the blue shift even though the FOMLSP peak is lowered. FOMSPP values for the group 1 Mg-rich alloy films show no significant improvement and the value decreases after about 340 nm. Both 51Mg-48Al and 58Mg-42Al thin films show a strong (400) peak corresponding to the γ phase (Al12Mg17) (Fig. 3). The results suggest that these intermetallic phases also improve FOMLSP values at the short wavelengths (below about 290 nm) when compared to either pure Al or pure Mg (Fig. 5(c)).
Figure 6(a) and 6(b) show the optical constants ε1 and ε2 as a function of wavelength for group 2 Mg-rich Mg-Al alloy thin films. Figure 6(a) shows that the ε1 values for group 2 Mg-rich Mg-Al alloy films lie between those of pure Al and Mg films, and the same trend is seen for the ε2 values (Fig. 6(b)). Figure 6(c) presents a plot of the FOMLSP versus wavelength and Fig. 6(d) shows a plot of the FOMSPP versus wavelength. FOMLSP values increase with an increase in Mg content and the peak in the FOMLSP curve blue shifts from 380 nm for pure Mg (5.2 Å/s) to about 330 nm (a shift of about 40 nm) for 81Mg-19Al. Like group 1 Mg-rich Mg-Al alloy films, Group 2 alloy films also have slightly higher FOMLSP than pure Mg film at the short wavelengths due to the blue shift even though the FOMLSP peak is lowered. FOMSPP values for the group 2 Mg-rich alloy films lie between those of pure Al and Mg films but 91Mg-9Al and 97Mg-3Al alloy films have higher FOMSPP than either Mg or Al beyond 600 nm (Fig. 6(c)).
Figure 7 shows the plot of FOMLSP value and peak position as a function of composition for Mg-rich Mg-Al alloy films. The FOMLSP value of Mg-rich Mg-Al alloy films increases from 4.1 to 7.9 (pure Mg) with increase in Mg concentration. As discussed earlier, the addition of Al to Mg blue shifts the FOMLSP peak position of Mg. However, increase in Mg content beyond about 70 at.% red shifts the FOMLSP peak position towards the peak position for pure Mg.
4.2. Al-rich Mg-Al alloy thin films
Figure 8(a) and 8(b) show the optical constants ε1 and ε2 as a function of incident wavelength for Al-rich Mg-Al alloy thin films. Figure 8(a) shows that the ε1 values for Al-rich Mg-Al alloy films lie between pure Al and Mg films. However, for wavelengths from ~350 to 650 nm, Al-rich Mg-Al alloy films suffer higher losses (higher ε2 values) when compared to both pure Al and Mg as shown in Fig. 8(b). Figure 8(c) shows a plot of the FOMLSP versus wavelength and Fig. 8(d) shows a plot of the FOMSPP versus wavelength. FOMLSP values increases with the increase in Al content and the FOMLSP curve of pure Al blue shifts from 330 nm to 280 nm (a shift of about 50 nm) for 65Al-35Mg film when compared to pure Al film but the values are still lower than pure Mg. FOMSPP values for the alloy films show no improvement and value decreases after about 340 nm
5. Transmission measurements of nanotriangles
Transmission measurements of nanotriangle arrays were performed using a Perkin Elmer Lambda 750 UV/Vis/NIR spectrophotometer with a cleaned quartz slide as a reference over a wavelength range of 200−2500 nm in air with a spot size of 1 cm2. Plots of ε1, ε2, and FOMLSP as a function of wavelength (250 to 1700 nm) for pure Mg, pure Al, 81Mg-19Al, and 91Mg-9Al alloy films are shown in Fig. 9(a), 9(b), and 9(c). For the fabrication of NTs, large PS bead templates were used to investigate the effect that alloying has on the LSPR response. These particles exhibit multiple LSPR modes from the UV to the NIR wavelengths as shown in Fig. 9(d). As expected, Mg has poor plasmonic properties at these long wavelengths compared to pure Al. The Mg nanotriangles exhibit a broad, low intensity LSPR peak around 1250 nm attributed to a dipole mode as well as a broad shoulder in the visible wavelength that may be associated with a higher order quadrupole mode. Both a lack of structural integrity and LSPR coupling also may contribute to the broadness of the dipole peak. Pure Al nanotriangles exhibit multiple sharp LSPR peaks in the visible and NIR (near-infrared) wavelength range attributed to both dipole and quadrupole resonance modes . The spectral response for the Al nanotriangles is relatively featureless in the UV.
By alloying Mg with Al, considerable improvement is seen in the LSPR response versus pure Mg particles, especially at longer wavelengths. In addition to the sharp dipole resonance at 1200 nm, the alloyed structures exhibit a smaller response at 700 nm (assigned to a quadrupole mode) and longer wavelength response centered at 1750 nm attributed to LSPR coupling between NTs.
While Mg has excellent plasmonic properties as a film, it may not be an ideal metal for structuring at the nanoscale. However, the film studies show that alloying Mg with Al improves the plasmonic properties in the UV-Vis range, the addition of Al may also have an added benefit of helping the Mg to form better defined structures at nanometer sizes. Figure 10 shows SEM images comparing nanotriangles of Al, Mg, and 91Mg-9Al fabricated using 1400 nm PS templates. These images clearly show the Mg particles are much less defined and have a considerable variation in shape when compared to Al NTs fabricated in a similar fashion. As expected, this lack of definition leads to a severe broadening of the LSPR response of pure Mg as shown in Fig. 9(d). However, alloying Mg with Al results in well-defined triangular shaped particles (Fig. 10(c)), when compared to pure Mg. In addition, relatively sharp LSPR peaks are observed in the extinction spectra (Fig. 9(d)). Addition of Al to Mg would tend to form MgO-Al2O3 spinel oxide which also stabilizes the shape of nanotriangles. In addition to rigorous structural studies, a thorough theoretical analysis would also be required to elucidate the contributions that shape and metal composition have on the observed LSPR response.
In summary, the results show that on increasing the deposition rate from 1.8 to 7.0 Å/s, the FOMLSP of Mg films was found to increase. This could be attributed to the reduced surface roughness of the film which dominates the optical response when compared to electron scattering by grain size. The addition of Al to Mg tends to blue shift the FOMLSP peak, resulting in small regions in the UV where an increased FOM can be realized. In the case of nanotriangle fabrication, addition of Al to Mg leads to considerable improvement in the LSPR response versus pure Mg particles, especially at longer wavelengths. Alloying results in well-defined triangular shaped particles when compared to pure Mg, which produces relatively sharp LSPR peaks in the extinction spectra.
The authors are grateful for the support of this work by the National Science Foundation under NSF-MRSEC program (Grant DMR-1121252) and by the University of Utah. This work made use of University of Utah shared facilities of the Micron Technology Foundation Inc. Microscopy Suite sponsored by the College of Engineering, Health Sciences Center, Office of the Vice President for Research, and the Utah Science Technology and Research (USTAR) initiative of the State of Utah.
References and links
1. S. A. Maier and H. A. Atwater, “Plasmonics: Localization and guiding of electromagnetic energy in metal/dielectric structures,” J. Appl. Phys. 98(1), 011101 (2005). [CrossRef]
2. D. Gérard and S. K. Gray, “Special issue on aluminium plasmonics,” J. Phys. D Appl. Phys. 48(18), 180301 (2015). [CrossRef]
3. K. Appusamy, S. Blair, A. Nahata, and S. Guruswamy, “Low-loss magnesium films for plasmonics,” Mater. Sci. Eng. B 181, 77–85 (2014). [CrossRef]
4. K. Appusamy, X. Jiao, S. Blair, A. Nahata, and S. Guruswamy, “Mg thin films with Al seed layers for UV plasmonics,” J. Phys. D Appl. Phys. 48(18), 184009 (2015). [CrossRef]
5. J. H. Simmons and K. S. Potter, Optical Materials (Academic Press 2000).
6. H. Ehrenreich and H. R. Philipp, “Optical properties of Ag and Cu,” Phys. Rev. 128(4), 1622–1629 (1962). [CrossRef]
7. M. H. Chowdhury, S. Chakraborty, J. R. Lakowicz, and K. Ray, “Feasibility of using bimetallic plasmonic nanostructures to enhance the intrinsic emission of biomolecules,” J Phys Chem C Nanomater Interfaces 115(34), 16879–16891 (2011). [CrossRef] [PubMed]
8. D. A. Bobb, G. Zhu, M. Mayy, A. V. Gavrilenko, P. Mead, V. I. Gavrilenko, and M. A. Noginov, “Engineering of low-loss metal for nanoplasmonic and metamaterials applications,” Appl. Phys. Lett. 95(15), 151102 (2009). [CrossRef]
9. C. Gong and M. S. Leite, “Noble metal alloys for plasmonics,” ACS Photonics 3(4), 507–513 (2016). [CrossRef]
10. C. Zhang, D. Zhao, D. Gu, H. Kim, T. Ling, Y.-K. R. Wu, and L. J. Guo, “An ultrathin, smooth, and low-loss Al-doped Ag film and its application as a transparent electrode in organic photovoltaics,” Adv. Mater. 26(32), 5696–5701 (2014). [CrossRef] [PubMed]
11. D. Gu, C. Zhang, Y.-K. Wu, and L. J. Guo, “Ultrasmooth and thermally stable silver-based thin films with subnanometer roughness by aluminum doping,” ACS Nano 8(10), 10343–10351 (2014). [CrossRef] [PubMed]
12. D. Zhao, C. Zhang, H. Kim, and L. J. Guo, “High-performance Ta2O5/Al-doped Ag electrode for resonant light harvesting in efficient organic solar cells,” Adv. Energy Mater. 5(17), 1500768 (2015). [CrossRef]
13. J. A. Woollam, B. D. Johs, C. M. Herzinger, J. N. Hilfiker, R. A. Synowicki, and C. L. Bungay, “Overview of variable-angle spectroscopic ellipsometry (VASE): I. Basic theory and typical applications,” Proc. SPIE CR72, 3–28 (1999).
14. B. D. Johs, J. A. Woollam, C. M. Herzinger, J. N. Hilfiker, R. A. Synowicki, and C. L. Bungay, “Overview of variable angle spectroscopic ellipsometry (VASE), Part II: Advanced applications,” Proc. SPIE CR72, 29–58 (1999).
15. J. C. Hulteen and R. P. Van Duyne, “Nanosphere lithography: A materials general fabrication process for periodic particle array surfaces,” J. Vac. Sci. Technol. A Vacuum, Surfaces. Film. 13(3), 1553–1558 (1995).
16. P. Moitra, B. A. Slovick, W. li, I. I. Kravchencko, D. P. Briggs, S. Krishnamurthy, and J. Valentine, “Large-scale all-dielectric metamaterial perfect reflectors,” ACS Photonics 2(6), 692–698 (2015). [CrossRef]
18. G. H. Chan, J. Zhao, G. C. Schatz, and R. P. Van Duyne, “Localized surface plasmon resonance spectroscopy of triangular aluminum nanoparticles,” J. Phys. Chem. C 112(36), 13958–13963 (2008). [CrossRef]
19. C. R. M. Grovenor, H. T. G. Hentzell, and D. A. Smith, “The development of grain structure during growth of metallic films,” Acta Metall. 32(5), 773–781 (1984). [CrossRef]
20. H. Savaloni and M. A. Player, “Influence of deposition conditions and of substrate on the structure of uhv deposited erbium films,” Vacuum 46(2), 167–179 (1995). [CrossRef]
21. K. M. McPeak, S. V. Jayanti, S. J. P. Kress, S. Meyer, S. Iotti, A. Rossinelli, and D. J. Norris, “Plasmonic films can easily be better: rules and recipes,” ACS Photonics 2(3), 326–333 (2015). [CrossRef] [PubMed]
22. D. A. Bobb, G. Zhu, M. Mayy, A. V. Gavrilenko, P. Mead, V. I. Gavrilenko, and M. A. Noginov, “Engineering of low-loss metal for nanoplasmonic and metamaterials applications,” Appl. Phys. Lett. 95(15), 151102 (2009). [CrossRef]
24. A. V. Zayats, I. I. Smolyaninov, and A. A. Maradudin, “Nano-optics of surface plasmon polaritons,” Phys. Rep. 408(3–4), 131–314 (2005). [CrossRef]
25. R. C. O’Handley, D. K. Burge, S. N. Jasperson, and E. J. Ashley, “Residual gas and the optical properties of silver films,” Surf. Sci. 50(2), 407–433 (1975). [CrossRef]
26. K. Yoshihara and K. Nii, “The effect of oxygen potential on the surface self-diffusion coefficient of silver,” Trans. Japan Inst. Met. 20(9), 533–542 (1979). [CrossRef]
27. T. B. Massalski, J. L. Murray, L. H. Bennett, and H. Baker, Binary Alloy Phase Diagrams (American Society for Metals, 1986).