1. Introduction

Plasmonic nanomaterials play a crucial role in the fields of metamaterials, metasurfaces, solar cells, plasmonics based heating, and optical sensing techniques like surface enhanced Raman spectroscopy (SERS) and localized surface plasmon resonance (LSPR) sensing [1–5]. To meet specific applications, the plasmonic nanomaterials need to be designed so that they possess tunable optical properties. In general, the plasmonic properties are tuned by changing the size, shape, and electromagnetic coupling of nanostructures [6,7]. These methods are effective, but are limited by the plasmonic materials and fabrication techniques used. Noble metals such as Ag, Au, or Cu can only cover limited optical wavelength ranges due to their intrinsic optical responses. Nanocomposites composed of more than one metal component have attracted the interest of researchers due to their potentially improved physical and chemical performances, in addition to their optical properties as compared to the single-component nanostructures [8–14]. For example, Anodilla et al [15] fabricated Au-Cu nanoparticles with tunable photoluminescence. When varying the molar concentration of Cu from 0% to 100%, the emission intensity redshifts from 947 nm to 1024 nm. Evans et al [16] fabricated Ag nanorods coated with a Ag-Au alloy film at different compositions. The resonance frequency of the nanorods linearly shifts with Ag composition. Similarly, Liu et al [17] showed that the LSPR of Ag-Au alloy nanoparticles could be tuned by alloy compositions. The LSPR wavelength redshifts with increasing Au content. Gong et al [18] fabricated Ag-Cu, Au-Cu, and Ag-Au thin-film alloys. The Ag-Cu, Au-Cu or Ag-Au films at specific compositions possess higher surface plasmon polarizations than those of the corresponding pure metals. Whitney et al [19] fabricated mixed-phase Ag-Cu triangular nanopatterns and films by combining shadow nanosphere lithography with a glancing angle co-deposition technique. They verified that all compositions of the nanopatterns had the same shape, but showed tunable localized surface plasmon resonance (LSPR) properties. The LSPR of the nanopatterns redshifted with decreasing composition. Wang et al [20] fabricated Ni-P nanorod arrays with a series of aspect ratios in the pores of anodic aluminum oxide membranes. With the increase of the nanorod aspect ratio, their optical absorbance decreased in the ultraviolet region but increased in the visible region, whereas the photoluminescence decreased. Titanium, a kind of transition nontoxic metal, has good corrosion resistances and is not affected by atmosphere and sea water. It has also good compatibility with human tissues and blood. Hence, it is nessary to investigate titanium mixed with other metals. For the Ag-Ti system, Murray et al [21] made a review on the equilibrium phase diagram, metastable phases, and crystal structures of the equilibrium phases.

For most of above novel nanocomposites, they were fabricated generally by chemical and physical routes, including sol-gel [22], ball milling [23], ion implantation [24], ultrasound radiation [25], electrodeposition [26], chemical vapor deposition [27], physical vapor co-deposition [28] and hybrid growth techniques [8]. The predominating morphologies of the nanocomposites obtained using these methods are in the form of powders or films. The oblique angle co-deposition technique (OACD) is another simple technique to form composite nanostructures [19]. It is a combination of co-deposition and oblique angle deposition (OAD) technique in a physical vapor deposition system [29]. The co-deposition enables the simultaneous vaporization of two materials, allowing the exact composition of the fabricated nanorod array to be dictated by their relative deposition rates [30]. The self-shadowing effect in an OAD process causes the formation of a well-aligned nanorod structure [31–33].

In this work, well-aligned Ag-Ti composite nanorods with different Ag to Ti ratios and different deposition angles were systematically fabricated by the OACD method. The optical transmission spectra of the composite nanorod structures are tuned by composition ratios of Ag and Ti, polarization directions, and deposition angles. Such tunable optical properties have potential applications in optoelectronics.

2. Fabrication

Figure 1 shows a deposition process of the OACD by a dual-source electron beam evaporation system (DE500, DE Technology Inc.). Ag (99.999%) and Ti (99.995%) were evaporated separately by two electron beam sources. Two separate quartz crystal microbalances (QCMs) were used to monitor deposition rates, r_Ag and r_Ti, and thicknesses, $d_{\text{Ag}}^{\text{QCM}}$ and $d_{\text{Ti}}^{\text{QCM}}$, of Ag and Ti, independently, facing directly to the vapor in the near-normal direction of the substrates. Two kinds of substrate were used: glass microscope slides for optical properties and x-ray diffraction characterization, and Si (111) wafers for scanning electron microscopy observation. At the initial stage of deposition, random nucleation centers formed on the substrates acted as shadowing seeds. Preferential growth of nanorods towards the direction of deposition flux occurred. The Ag-Ti composite nanorods could thus be obtained. Before the substrate preparation, the chamber was evacuated to a base pressure of 5 × 10⁻⁷ Torr. During deposition, the pressure varied from 10⁻⁶ to 10⁻⁷ Torr. To fabricate composite nanorod structures, the substrate normal was tilted at α = 87.5°, 85°, 82.5°, 80°, 77.5°, and 75° relative to the incident Ag and Ti vapor flux. As a reference, composite films were obtained during the same deposition by installing additional pieces of substrates below the substrate holder with α = 0°. A total of seven different compositions of nanostructured samples, which have a Ag to Ti ratio of 100:0, 90:10, 80:20, 60:40, 40:60, 20:80, and 0:100, respectively, were fabricated by the OACD. The nominal mass ratio of Ag to (Ag + Ti) or Ag mass composition, $C_{\text{Ag}}^{\text{M}}$, in each sample was determined and controlled by the deposition rate ratio $r_{\text{Ag}}/r_{\text{Ti}}$ according to the relationship $r_{\text{Ag}}/r_{\text{Ti}}=d_{\text{Ag}}^{\text{QCM}}/d_{\text{Ti}}^{\text{QCM}}=C_{\text{Ag}}^{\text{M}}/\left(1-C_{\text{Ag}}^{\text{M}}\right)\cdot \left({\rho }_{\text{Ti}}/{\rho }_{\text{Ag}}\right)$, where ${\rho }_{\text{Ti}}$ and ${\rho }_{\text{Ag}}$ are the bulk density of the corresponding materials [34]. In the experiment, the nanorod arrays grew until the sum of the QCM thicknesses the Ag and Ti reached 200 nm, i.e., $d_{\text{Ag}}^{\text{QCM}}+d_{\text{Ti}}^{\text{QCM}}$ = 200 nm. The detailed deposition parameters, such as the deposition rates r_Ag and r_Ti, final thickness readings $d_{\text{Ag}}^{\text{QCM}}$ and $d_{\text{Ti}}^{\text{QCM}}$, nominal Ag mass composition $C_{\text{Ag}}^{\text{M}}$, and symbol for each sample are summarized in Table 1.

 

Fig. 1 Schematic diagram of deposition of nanorods by double-source oblique-angle co-deposition technique.

Download Full Size | PDF

Table 1. Detailed deposition parameters to control the compositions for all Ag-Ti samples.

View Table

3. Crystal structure evolution analyses

The Ag-Ti thin films with different $C_{\text{Ag}}^{\text{M}}$ were used to analyze the composition and crystal properties. These films were deposited simultaneously with the nanopatterns for each $C_{\text{Ag}}^{\text{M}}$ [19]. The crystal structures of the thin films on the glass substrates with α = 0° were characterized by an x-ray diffractometer (XRD, D2, Bruker). The XRD scans of the thin films were recorded with a Cu Kα1 radiation (λ = 1.5406 Å) in the 2θ range from 20° to 80° with a step of 0.010°.

Because Ag and Ti are immiscible, the resulting mixed phase films were expected to coexist both Ag and Ti polycrystals or nanocrystals [35, 36]. Figure 2 shows the measured XRD patterns of the thin films. In general, the trend from Fig. 2 indicates that as $C_{\text{Ag}}^{\text{M}}$ decreases, the XRD peak intensities of Ag decrease, while the Ti peak intensities increase. For the pure Ag sample, two distinct Ag peaks are present at 2θ = 38.1° and 44.3°, which correspond to the (111) and (200) planes of Ag, respectively (JCPDS Ref. No. 04-0783). With the introduction of Ti in the Ag90 sample, the (200) peak of Ag vanishes, and the (111) peak decreases significantly. In the Ag90, Ag80, and Ag60 samples, because of the doping of Ti, the (111) Ag peak gradually becomes weak. This means the crystallinity of Ag decreases. In the Ag60, Ag40, and Ag20 samples, with the increase of Ti composition, the diffraction peak at 2θ = 38.5°, which corresponds to the (110) plane of Ti (JCPDS Ref. No. 44-1288), becomes strong. This means that the crystallinity of Ti increases. Clearly, the Ag-Ti composite films do not follow Vegard’s law of a solid solution, which states that there is a linear dependence of the d-spacing versus the atomic concentration. This law implies that the lattice parameters of a solid solution can be approximated by the mixtures of the two constituents’ lattice parameters. Instead, our measurements show two distinct crystal patterns [37, 38], i.e. the face center crystal pattern of Ag seen from the Ag100, Ag90, Ag80, and Ag60 samples, and the hexahedral crystal pattern of Ti from the Ag40, Ag20, and Ag0 samples. Thus, the resulted Ag-Ti films or nanorods are mixture of Ag and Ti.

 

Fig. 2 XRD patterns of all thin film samples with different Ag compositions.

Download Full Size | PDF

4. Morphology characterization

For all samples deposited on silicon substrates, their morphology and compositions were characterized by a field-emission scanning electron microscope (SEM) (SU8010, Hitachi) integrated with an energy dispersive x-ray spectroscopy (EDX). SEM images were then analyzed by an Image J software (NIH).

Figure 3 shows the top (in the left column) and cross-sectional view (in the right column) SEM images of the Ag20 and Ag80 samples while the deposition tilted angles are 80° and 85°, respectively [34]. The SEM images reveal that the samples consist of arrays of aligned and tilted nanorods with different dimensions and tilting angles. The measured C_Ag values for the samples with α = 85° are 100, 85 ± 10, 75 ± 9, 56 ± 7, 37 ± 5, 18 ± 3, and 0 wt.% corresponding to the mass concentrations of Ag100, Ag90, Ag80, Ag60, Ag40, Ag20, and Ag0, respectively. The C_Ag measured values are close to the nominal $C_{\text{Ag}}^{\text{M}}$ values estimated from the deposition rate ratios, but deviate the predicted values. This is because that Ag has a lower melting point than Ti (T_Ag = 962 °C and T_Ti = 1668 °C, respectively), and also has a smaller sticking coefficient than Ti. Consequently, even if the deposition fluxes were the same, less Ag would be deposited due to its lower sticking coefficient [29, 39, 40]. To quantitatively analyze how the morphology of the Ag-Ti nanorod arrays evolves with $C_{\text{Ag}}^{\text{M}}$, we plot the average nanorod width w, the nanorod thickness δ, the height h, and the tilting angle β (defined in Fig. 3(a)) with respect to $C_{\text{Ag}}^{\text{M}}$ in Fig. 4 [41, 42]. The average parameters were obtained by measuring more than 30 values in given domains from different SEM photos. Here α is fixed in 85°. Both the width w and nanorod thickness δ are almost a constant, while the height h decreases monotonously with $C_{\text{Ag}}^{\text{M}}$, and the tilting angle β increases with $C_{\text{Ag}}^{\text{M}}$. Such a trend is also consistent with the sticking coefficents of Ag and Ti.

 

Fig. 3 Top view (in the left column) and cross-sectional view (in the right column) SEM images of nanorod arrays with different deposition tilted angles and different Ag contents of (a) Ag20, α = 80°; (b) Ag20, α = 85°; (c) Ag80, α = 80°; (d) Ag80, α = 85°.

Download Full Size | PDF

 

Fig. 4 Relationships between nanorod width w, thickness δ, height h, and tilting angle β and Ag concentration for the case of the same deposition angle α = 85°.

Download Full Size | PDF

Figure 5 plots w, δ, h, β as a function of incident angle α for the Ag20 samples. The nanorod width w becomes narrower as α increases. The thickness δ and height h decreases slightly with α. The tilting angles β are almost unchanged with the α. This means that for the same $C_{\text{Ag}}^{\text{M}}$, the morphologies of Ag-Ti nanorods have similar morphological parameters when α is within 77.5°-87.5°.

 

Fig. 5 Relationships between nanorod width w, thickness δ, height h, and tilting angle β of the Ag20 nanorod arrays and deposition tilted angle.

Download Full Size | PDF

5. Optical transmission characterization

The optical properties of the nanorod samples were studied by a UV–vis spectrophotometer (Lambda 950, PerkinElmer) with a beam size of 1.8 mm in length and 0.6 mm in width. To obtain polarized transmission spectra, one rotatable Glan Taylor polarizer (DGL10, Thorlabs Inc.) was placed in the path of the incident light beam. For convenience, we defined p-polarization as the direction of the incident electric field parallel to the major long axis of the nanorods (φ = 0°, corresponding to the case of E_p shown in Fig. 2(b)) and s-polarization as the direction of the incident field perpendicular to the long axis of the nanorods (φ = 90°). In the measurement, the corresponding transmission spectra from a glass substrate were used as references. The resulted spectrum is the average spectra taken from five different random locations at the substrate.

5.1. Influences of Ag contents on transmission spectra

Firstly, we show the influence of Ag contents on UV-Vis transmission spectra. Figure 6 is the p-polarization (φ = 0°) and s-polarization (φ = 90°) transmission spectra of the Ag-Ti nanorod samples with different $C_{\text{Ag}}^{\text{M}}$ for α = 85°. Evidently, the two polarization types of transmission spectra are closely related to Ag contents and are tuned by polarization. The Ag20 sample has the highest transmission for both p-polarization and s-polarization in the measured wavelength range, which indicates that there exists an optimum ratio of Ag and Ti to obtain a high transmission. Figure 6 also shows that the transmission in general decreases with increasing Ag contents. The reason is that Ag has a higher free-electron density than Ti and the resulting electron resonances in the samples with more Ag contents occur more easily. That is to say, external light fields couple with the free electrons in these samples more easily. As a result, their transmission is much lower.

 

Fig. 6 (a) P- polarization (φ = 0°) and (b) s-polarization (φ = 90°) transmission spectra of the Ag-Ti composite nanorod samples with different composition ratios under α = 85°.

Download Full Size | PDF

For further comparison, the two polarization types of transmission spectra at λ = 400 nm, λ = 600 nm and λ = 800 nm as functions of C_Ag are plotted in Fig. 7. The polarization transmission spectra at λ = 400 nm have a similar trend with the increase of C_Ag. The transmission spectra are affected slightly by polarization. However, the transmission spectra at λ = 600 nm and λ = 800 nm are deeply affected especially for more than 60% of C_Ag for the two different polarizations. Meanwhile, the s-polarization transmission at the two wavelengths is stronger than the p-polarization transmission.

 

Fig. 7 P-polarization and s-polarization transmission spectra $T_{400}^{\text{p}}$ and $T_{400}^{\text{s}}$ at λ = 400 nm, $T_{600}^{\text{p}}$ and $T_{600}^{\text{s}}$ at λ = 600 nm, and $T_{800}^{\text{p}}$ and $T_{800}^{\text{s}}$ at λ = 800 nm as functions of C_Ag.

Download Full Size | PDF

Because the Ag-Ti nanorods are the mixture of Ag and Ti analyzed from the XRD patterns. The average permittivity for a composite nanorod material can be expressed as

Then a finite-difference time-domain (FDTD) software was employed to simulate transmission spectra of the Ag20, Ag80, and Ag100 samples with a deposition tilted angle of α = 85° for p-polarization (φ = 0°). The used geometric parameters of nanorods are generally chosen in the measurement error value range mentioned above. The nanorods are regarded as hexagonal lattice arrangement. The simulation results are plotted in Fig. 8. It can be found that the experimental results are in rough agreement with numerical simulations except for the transmission spectra of the Ag100 sample. The main reason is that the Ag100 sample has more disorderly nanorod arrangement, which was seen from additional SEM images.

 

Fig. 8 Simulated as well as experimental transmission spectra of the Ag20, Ag80, and Ag100 samples with the deposition tilted angle α = 85° for p-polarization (φ = 0°). The lengths of nanorods for the Ag20, Ag80, and Ag100 samples are set to 158 nm, 154 nm, and 52 nm, respectively. The widths of nanorods for the Ag20, Ag80, and Ag100 samples are set to 10 nm, 11 nm, and 40 nm, respectively. The core distances of nanorods for the Ag20, Ag80, and Ag100 samples are set to 70 nm, 66 nm, and 110 nm, respectively. The tilting angles for the Ag20, Ag80, and Ag100 samples are set to 42°, 66.3°, and 68°, respectively.

Download Full Size | PDF

5.2. Influences of polarization directions on transmission

Because the composite nanorods arrange in specified directions, their optical properties should change for different polarization fields. To explore the influences of polarization directions on transmission, we measured transmission spectra of the Ag80 and Ag20 composite nanorod samples under α = 85° for different polarization directions. The corresponding results are shown in Fig. 9. From this figure, we can find that the transmission for the Ag80 composite nanorod sample changes much bigger than that of the Ag20 composite nanorod sample for different polarization directions. Each of the transmission curves varies evidently near the wavelength of 400 nm, indicating that resonances occur near this wavelength. Interestingly, there is a crossover point of the transmission curves of the Ag80 composite nanorod sample for the different polarization directions. The corresponding wavelength at the crossover point is about λ = 409 nm. At this wavelength, the transmission has nothing to do with the polarization directions, which means that the composite nanorod sample can be regarded as an isotropic one. From Fig. 3(d), it can be seen that the arrangement of the composite nanorods is not greatly orderly. Under different polarization directions, resonances in the composite nanorods can be excited. The average length of the composite nanorods is certain, which decides that their resonance wavelength is unchanged. Therefore, the transmission spectra under different polarization directions converge at one point.

 

Fig. 9 Transmission spectra of the (a) Ag80 and (b) Ag20 composite nanorod samples under α = 85° for different polarization directions of φ = 0°, φ = 30°, φ = 45°, φ = 60°, and φ = 90°.

Download Full Size | PDF

In order to further explain the relationship between transmission and polarization, angular-dependent polarization transmission spectra of the samples with the deposition tilted angle of α = 85° at the three wavelengths of λ = 400 nm, λ = 600 nm, and λ = 800 nm were plotted in Fig. 10. For the samples at λ = 400 nm, the polarization dependence of transmission spectra is low, indicating that the samples exhibit optical isotropy behaviors. While for the samples at λ = 600 nm, and λ = 800 nm, the polarization dependence of transmission spectra is strong, which means that the samples have strong anisotropy.

 

Fig. 10 Transmission spectra of the seven different samples with the deposition tilted angle of α = 85° as functions of polarization angle of incident light at (a) λ = 400 nm, (b) λ = 600 nm, and (c) λ = 800 nm.

Download Full Size | PDF

5.3. Influences of deposition angles on transmission

In this section, we investigate influences of different deposition tilted angles on transmission. Figures 11 and 12 show the transmission spectra of the Ag80 and Ag20 composite nanorod samples obtained from different deposition angles α for different polarization directions, respectively. Both two figures show that the greater the deposition angles are, the higher the transmission is for the same polarization direction. The transmission spectra of the two kinds of composite nanorod samples are greater than 90% for α = 87.5°, while their transmission spectra for α = 75° are less than 20%. With the decrease of the deposition angle, the fabricated composite nanorod samples tend to thin films. As a result, their transmission becomes much lower. Utilizing such a property, high or low transmission lenses can be designed and fabricated simply.

 

Fig. 11 Transmission spectra of the Ag80 composite nanorod samples obtained from different deposition angles of α = 87.5°, 85°, 82.5°, 80°, 77.5°, and 75° for different polarization directions of (a) φ = 0°, (b) φ = 45°, and (c) φ = 90°.

Download Full Size | PDF

 

Fig. 12 Transmission spectra of the Ag20 composite nanorod samples obtained from different deposition angles of α = 87.5°, 85°, 82.5°, 80°, 77.5°, and 75° for different polarization directions of (a) φ = 0°, (b) φ = 45°, and (c) φ = 90°.

Download Full Size | PDF

In order to understand the influences of deposition angle on transmission spectra more intuitively, the transmission spectra of the Ag80 and Ag20 samples with different deposition angles at λ = 400 nm, λ = 600 nm, and λ = 800 nm as functions of polarization angle were plotted in Figs. 13 and 14, respectively. It can evidently be seen from the two figures that the transmission changes with increasing polarization angle as well as increases with the increase of deposition angle.

 

Fig. 13 Transmission spectra of the Ag80 samples for the six different deposition tilted angles as functions of polarization angle of incident light at (a) λ = 400 nm, (b) λ = 600 nm, and (c) λ = 800 nm.

Download Full Size | PDF

 

Fig. 14 Transmission spectra of the Ag20 samples for the six different deposition tilted angles as functions of polarization angle of incident light at (a) λ = 400 nm, (b) λ = 600 nm, and (c) λ = 800 nm.

Download Full Size | PDF

6. Conclusion

In summary, we fabricated various Ag-Ti composite nanorod structures by the oblique angle co-deposition technique using a dual-source electron beam deposition system. The effects of increasing Ag contents during the deposition on morphology, structure and optical property were investigated. Morphologically, the nanorods disperse in the direction perpendicular to the vapor flux, resulting in a separation of the nanorods in this direction smaller than that along the vapor flux. X-ray diffraction reveals crystal structures of Ag-Ti alloys. The composite nanorod structures have different transmission spectra for different composition ratios of Ag and Ti, different polarization directions, and different deposition angles. Interestingly, for the Ag80 composite nanorod structures, there exists a wavelength, where it is isotropic. The reason is that the arrangement of the composite nanorods is not greatly orderly. Under different polarization directions, resonances in the composite nanorods can be excited. Their resonance wavelength is decided by the average length of the composite nanorods. We also demonstrated that the transmission spectra of the Ag80 composite nanorod structure for α = 87.5° are greater than 90% under different polarization directions, while their transmission spectra for α = 75° are lower than 20%. Utilizing such a property, high or low transmission lenses can be designed simply.