Effect of Pore Thickness and the State of Polarization on the Optical Properties of Hexagonal Nanoarray of Au / Nanoporous Anodic Alumina Membrane

Hexagonal nanoarrays of Au particles were deposited on nanoporous anodic alumina membrane (NAAM) utilizing r.f. magnetron sputtering. The thickness of the NAAMs is adjusted by changing the second anodization time from 5min to 20min. The surface morphology, composition, and optical properties are characterized by using SEM, EDX, and spectrophotometer, respectively. The effects of theNAAM thickness and state of polarization on themorphological changes and on the optical properties of the fabricated nanoarrays were addressed. According to the measured optical spectra, the rate of decrease of NAAMs refractive index was found to be 3.825 × 10 nm. Using themodified Kubelka-Munk radiative transfermodel, the energy gap of NAAMswas calculated from diffused reflectance and was decreased from 1.682 to 1.376 as the anodization time increased from 5 to 20min. Also, the saturation of interference fringes is substantially enhanced, and field enhancement can be achieved due to the excitation and constructive interference of surface plasmon waves by coating NAAMs with the hexagonal nanoarrays of Au. Based on the advantages of the fabrication approach and the enhanced and controlled properties, this new generation of samples can be used as promising building blocks for nanophotonic and nanoelectronics devices.


Introduction
In recent years, porous materials have attracted much interest because of their potential use in a variety of applications [1][2][3].Among them, nanoporous anodic alumina membranes (NAAMs) have been widely used as membranes in the fabrication of various nanostructures.Two-step anodization was widely used to fabricate NAAMs because of the high controllability of the process, cheap equipment, and facile technology [4].Metal nanostructures are usually loaded to extend the application of porous membranes [5], so the conducting, mechanical, and optical properties of porous membranes could be improved [2].Up to now, different kinds of nanotubes, nanorods, nanoparticles, and nanowires have been successfully synthesized inside the pores of NAAMs [6][7][8].Homogenous deposition of metal nanostructures inside the pores of NAAMs has gained much attention in recent years.This attention is ascribed to their extensive use as substrates for sensors based on surface-enhanced Raman scattering (SERS), laser induced fluorescence (LIF), and electrochemical measurements [3,[9][10][11].
For an instant, the metal nanostructures on the outer surface of NAAMs have more possibility of touching the molecules that will be detected.Additionally, many reports have been done on optical properties of NAAMs [5,12,13].NAAMs showed bright colors, but their saturation is very tiny.Colored nanostructures always have to be sealed with the nanopores of NAAMs to increase and modify the color saturation [12,13].Also, the optical properties of gold nanostructures have been extensively studied [5,[13][14][15].Surface plasmon resonances interactions characterize these studies.These surface plasmon interactions can be used to enhance, guide, and modify the optical fields of novel nanoarrays such as photonic crystals [16], sensors based on SERS [10,17], laser induced fluorescence, and nearfield microscopy [3].In particular, Au nanoarrays possess features of both nanostructured metals that exhibit localized surface plasmon (LSP) excitations and planar metal films that exhibit propagating surface plasmon (PSP) excitations [5,13,18,19].Additionally the optical response of nanoparticle arrays depends on their size and shape.The interparticle coupling leads to energy shifts and splitting of degenerate modes, an effect that is prominent for regular nanoarrays.
In our previous studies, different metals have been loaded on the outer surface of NAAMs such as Au, Cr, Pt, and Sn [11,13,20,21].A controlled methodology of utilizing r.f.magnetron sputtering to coat NAAMs with hexagonal nanoarrays of Au nanoparticles was discussed [13].Also, experimental studies of Au particles size, pore diameter, and angle of the incident were addressed.These studies showed the surface plasmon enhancement of NAAMs interference and illuminated the relationships between resonance position and structures (pore diameter and thickness) of porous gold layer and angle of incident.
However, there are other factors that can affect the optical properties of these nanostructures including the thickness of NAAM and the state of polarization of light.All of these factors make it important to have experimental data that can correlate the thickness of NAAMs and state of polarization with the optical properties.Here, we studied the optical properties of NAAMs of different thicknesses before and after decoration with hexagonal nanoarrays of Au particles under illumination with s-and p-polarized light.

Sample Fabrication and Characterization
Two-step anodization combined with a pore widening process was utilized to fabricate NAAMs with homogeneous pore diameters and high-quality pore walls [5,22].The 1st anodization process was performed at 40 V in 0.3 M oxalic acid at 10 ∘ C for 3 h.After the removal of the nanoporous Al 2 O 3 , the second anodization was performed at the same conditions for varying lengths of time followed by the barrier-thinning process.The barrier layer was thinned by a successive drop of the DC voltage from 40 to 15 V at a rate of 0.1 V/s and then maintained at 15 V for 5 min.The pore widening was carried out for 70 min pore widening after 10 min cathodic polarization [22].Au nanoparticles were sputtered on the NAAMs using r.f.magnetron sputtering for 40 s [13].The distance between the target and the substrate was 100 mm.Magnetron sputtering was carried out in pure Ar with a radiofrequency power of 65 W. The total pressure during deposition was 5.0 mTorr.The substrate temperature during the sputtering was set at room temperature.
Atomic force microscopy (AFM), field emission scanning electron microscope (FE-SEM, JSM-7500F/JEOL), and energy dispersive X-ray spectrometer (EDX, Oxford Link ISIS 300 EDX) were used to study the morphological and structural properties and chemical composition of the fabricated nanoarrays.Optical reflectance in the spectral range from 300 to 1000 nm was measured using UV/VIS/NIR 3700 double beam Shimadzu spectrophotometer at RT.

Morphological and Chemical Composition
Studies.FE-SEM characterized the surface properties of the blank NAAMs.Figure 1(a) illustrates a typical FE-SEM top view of NAAM.Hexagonal aligned nanopores were formed in NAAM.The pore diameter is ∼79 nm, the interpore distance is ∼100 nm, and the pore density is ∼1.1 × 10 10 cm −2 .The inset of Figure 1(a) illustrates the defect map of this sample.The ratio of defects is 0.17.As shown in cross-sectional views of NAAM anodized for (b) 5 min and (c) 10 min, around each nanopore there is hexagonal arrangement from six active nanodots, which is consistent with the hexagonal packing structure of the pores.Because the farthest distances from the nanopores induced the weakest intensity of the electric field in the electrochemical process, then the oxidation rate is slowest resulting in the protuberances at special sites [23].However, the number of defects or cracks decreased and homogeneity of the hexagonal arrangement increased as the second anodization increased.Then, the percentage of defects decreases and the averaged regularity ratio increases by increasing the anodization time, which is consistent with the literature data [24].Additionally, from these images the rate of growth of the pores was found to be 1.3 nm/sec.
Figure 2 shows FE-SEM and AFM images of NAAMs of different thicknesses sputtered with Au under the same conditions for 40 sec.As shown from this figure, the homogeneity and hexagonal arrangement of the Au nanoparticles are improved by increasing the thickness of the starting NAAM.This may be attributed to the better arrangement of nanopores of the start NAAM and their active nanodots that was obtained with the longer second anodization time.As the Au particles grown around the active dots, the averaged pore diameters are decreased to 59 and 56 nm for Au nanoarrays that deposited on the 5 and 10 min anodized NAAMs, respectively.Also, the average Au particle diameters are increased from ∼44 and 58 nm.Moreover, one can observe a slight increase in the averaged Au nanoparticles heights from ∼39.5 to ∼44.5 nm.This may be a result of the shorter distance between the surface of the substrate (NAAM) and the surface of Au target, due to the difference in the heights of the starting NAAMs.Then, for a longer anodization time, the Au nanoparticles are grown faster around the active dots, and the rate of lateral growth is faster than the rate of vertical growth.Moreover, the subgaps (holes) between the hexagonal Au nanoparticles grown on 5 min anodized NAAM are greater than that of the 10 min anodized NAAM.
To study the chemical composition of the fabricated sample, the Au/NAAMs were analyzed by EDX as shown in Figure 3.The EDX pattern exhibits the signals of Au, Al, and O elements.The quantitative results were 89.34%Al, 9.11% O, and 1.55% Au for the NAAMs anodized for 5 min.By increasing the anodization duration of the starting NAAMs from 5 to 10 min, the ratio of the Au is increased to 2.05%, consistent with the results of the analysis of AFM images in Figure 2.  The optical measurement geometry is shown in the inset schematic diagram of Figure 4(a).By increasing the membrane thickness, the number of interference ripples increases, which is consistent with Bragg's law.Increasing the thickness of NAAMs shifts the dips in the reflectance that occurs to longer wavelengths; these dips are originated from the destructive interference between the reflected waves from Al/NAAM and NAAM/air interfaces.As shown in the inset, the absorption band below 400 nm is associated with the decrease in the absorption coefficient.The average reflectivity of NAAMs, ∼80%, is almost unchanged.This means that the scattering coefficient and hence the effect of scattering are minuscule and negligible for these NAAMs for wavelengths >400 nm.It is well known that the strength of interference ripple increases, as the wavelength increased; however, the interference ripple around 430 nm is much stronger.This may be a result of the blue emission band of NAAM which can be attributed to mixed emission from F and F + centers as discussed by Nasir et al. [25].Using Bragg's law and interference fringes in the reflectance spectra of Figure 4(a), the effective refractive indices,  eff , can be calculated using [13]

Optical Properties of NAAM and Au/NAAM.
where  1 and  2 are the wavelengths of two adjacent maxima or minima,  = 30 ∘ is the angle of incident, and  is the thickness of NAAM.The values of the order of interference,   , positions of reflectance dips or minima,  res , and effective refractive index are shown in Table 1. Figure 4(c) displays the linear decrease of an effective refractive index versus thickness of NAAM.From the linear fitting, the rate of decrease of refractive index is 3.825 × 10 −4 nm −1 .Using the diffused reflectance spectra, Figure 4(a), the band gap energy can be calculated by applying the modified Kubelka-Munk radiative transfer model [26].From the measured reflectance spectra, the absorbance  =  max −, where  max is the maximum value of reflectance for wavelengths longer than that of the dip in the diffuse reflectance and  is the reflectance.The calculated absorbance spectra are shown in Figure 4   To emphasize the role of Au nanoparticles in the modulation of the reflective spectroscopic response of the studied NAAMs, the samples were sputter-coated with Au.
The reflectance spectra of 40 s Au-coated NAAMs anodized in oxalic acid for 5, 10, and 15 min are shown in Figure 5(a).As shown in the figure, all three samples exhibit strong oscillations; these oscillations result from surface plasmon enhancement of optical interference between the light beams that reflected from the top and bottom surfaces of the sample.With increasing thickness of NAAMs, the number and the position of interference bands increase, while their bandwidth and reflectivity decrease.The distance between interference bands increases as the wavelength is increased.Additionally, the oscillation amplitude decreases with decreasing wavelength, and the rate of decrease is sample dependent.The data shows that a reduced anodizing time (NAAM thickness) will lead to more pronounced interference; this may be ascribed to a more efficient SPR propagation at the metal/dielectric interface in addition to the decrease of the dielectric layer thickness.The strong enhancement of the plasmon field occurs at contacting surfaces between Au nanoparticles and leads to superfocusing of electromagnetic waves.Then, the reflected waves of NAAMs are strongly modulated by hexagonal arrays of Au nanoparticles and the oscillation strength (the difference between the maximum and minimum) of interference fringes is increased.the characteristic minima as a result of the coupling between interference fringes and Au surface plasmon resonances (SPRs).When the coincidence between the position of interference fringe and SPR occurs, the absorbance will be enhanced at that position and strong absorption band is noted in the reflectance spectra [5,13].Figure 5(c) shows the minimum reflectance versus  min for the characteristic minima.This figure clearly illustrates that the strongest absorption band is located around 550 nm for 40 s Au/15 min NAAM sample.Then, this minimum is affected by LSPR more than other minima.This indicates that the position of the SPR for this sample is located very close to 550 nm.By decreasing the anodization time to 5 min, LSPR is redshifted.This redshift is ascribed to the negative dipole-dipole coupling energy contribution due to the lack of the direct conductive interlink between Au nanoparticles [13].Figure 5(d) shows reflectance spectra measured for sand the p-polarized light incident at an angle of 45 ∘ for 40 s Au-coated NAAM anodized in oxalic acid for 5 min.As shown in the figure a broad minimum centered at 642 nm for p-polarized light and 635 nm for s-polarized light; however, the effect was more pronounced for p-polarized light, which shows an increasing asymmetry.The broad minimum appears to be due to the existence of two minima: the first is similar to that seen in the s-polarized light and the second is sharper than the first one and occurs very close to it.To illustrate this point, the p-polarized spectrum has been normalized to the s-polarized spectra in the inset of Figure 5(d), revealing a minimum at a wavelength of 638 nm.For nanoparticles that are in direct contact or a few nanometers apart, direct electronic coupling occurs as a result of the overlap of electronic wave functions [27].These electronic exchange reactions are resulting in redshift and broadening of the plasmon resonance compared to that of noninteracting particles.For s-polarized light, the electric field vector had only one component parallel to the surface of the Au array and excited the electron oscillation in the same plane.For p-polarized light, an additional normal component appears at an oblique angle and excited electron oscillations normal to the plane of the Au array on the top surface of NAAM in addition to the parallel electron oscillations.Therefore, the redshift and broadening were attributed to the plasmon coupling of these normal oscillations.
Recently, structural colors have attracted much interest because their applications have been rapidly progressing in many fields related to vision and decoration purposes [28].Most of the structure colors originate from fundamental optical processes including thin-film interference, multilayer interference, diffraction grating effect, photonic crystals, and light scattering [28].We noticed that the coated Au-NAAMs used in this study have a brilliant color with high saturation caused by the surface plasmon enhancement of optical interference, as we discussed in a previous report [13].This is evident from Figure 6, where the effect of angle of incident and normal incident-photographs of decorated membranes are shown.As shown in the figure, the samples possess obvious structural color and have strong light mirroring in a certain wavelength.It has been found that the sample color is mainly affected by the thickness of NAAM (anodizing time).Note that the color of the samples changes with different viewing angles because they are derived from interference.Also, as the angle of incident increases, the number of fringes is increased but the oscillation strength and the fringe width are decreased.The variation of the resonance minima wavelength with the angle of incident is shown in Figure 6(b).As the angle of incident increased,  res is blueshifted.

Conclusion
In conclusion, we have successfully coated NAAMs of different thicknesses with nanoporous hexagonal Au particles.The morphologies, structures, and chemical composition of NAAMs and Au/NAAMs are studied by FE-SEM, AFM, and EDX.The defects percentage decreases and the averaged regularity ratio increases by increasing the anodization time.The average Au particles diameters are increased from ∼44 to 58 nm, and the Au nanoparticles heights are also increased from ∼39.5 to ∼44.5 nm as the anodization time increased from 5 to 10 min.From the reflectance of the blank NAAM, spolarized light shows a sharp minimum at 457 nm.However, p-polarized light showed a broad minimum centered at 470 nm that originally results from the existence of two minima at 415 and 488 nm.This indicates an increasing asymmetry for p-polarized light.The effective refractive index of NAAMs is linearly decreased with the thickness of NAAM, and the rate of decrease was found to be 3.825 × 10 −4 nm −1 .Using the modified Kubelka-Munk radiative transfer model, the energy gap of NAAMs is decreased from 1.682 to 1.376 as the anodization time increased from 5 to 20 min.For 40 s Au-coated NAAMs, the number and the position of the strong interference bands increase while their bandwidth and reflectivity decrease as the NAAM thickness increased.Additionally, the oscillation strength and oscillation width decrease with decreasing wavelength.The samples possess obvious structural color and have strong light mirroring in a certain wavelength.The advantages of the fabrication approach and the enhanced and controlled optical properties make these samples suitable for nanophotonic and nanoelectronics devices.
Figure 4(a)   shows reflectance spectra of NAAMs of different thicknesses.

Figure 1 :
Figure 1: FE-SEM images (a) top view and cross-sectional view of NAAMs anodized for (b) 5 min and (c) 10 min.The inset of (a) shows the defect map of this sample.The scale bar is 100 nm.

Figure 4 (
b) shows reflectance spectra measured for s-and the p-polarized light incident at an angle of 30 ∘ for NAAMs anodized in oxalic acid for 15 min.As shown in the figure a broad minimum centered at 470 nm for p-polarized light, which was originally located at 457 nm for s-polarized light.This indicates an increasing asymmetry for p-polarized light and the existence of two minima.To illustrate this point, the p-polarized spectrum has been normalized to the s-polarized spectra in the inset of Figure4(b), revealing the existence of two minima at 415 and 488 nm.

tFigure 5 :
Figure 5: (a) Reflectance spectra of 40 s Au-coated NAAMs anodized for different lengths of time, (b)  min versus the anodization time for NAAM (black spheres) and Au/NAAM (blue triangles), (c) minimum reflectance versus  min , and (d) reflectance spectra of Au/NAAM illuminated with s-and p-polarized light incident at an angle of 45 ∘ .The inset shows the dependence of  p / s on the wavelength.

Figure 5 (Figure 6 :
Figure 6: Reflectance spectra of 40 s Au/NAAM measured at different angles.The inset shows photographs captured at the normal incident.

Table 1 :
The values of the order of interference,   , positions of reflectance minima,  res , and the effective refractive index.
absorbance to the zero line of absorbance.The variation of the obtained energy gap values with the anodization time is shown in the inset of Figure4(d).The energy gap is decreased from 1.682 to 1.376 as the anodization time increased from 5 to 20 min.