1. Introduction

Electronic and optical phenomena utilizing plasmonic effects have attracted considerable attention due to their promising applications to functional plasmonic nanodevices of the next generation [1–8]. In particular, surface plasmon polaritons (SPPs) propagating along dielectric/metal interfaces are necessary for the miniaturization of optical devices beyond the diffraction limit because of the remarkable capabilities of subwavelength field confinements and strong field enhancements [9]. At present, various functional elements based on SPPs have been developed, such as switcher [10–12], beam splitters [13], and waveguides [14,15]. Particularly, to achieve functional plasmonic components with broad bandwidth, multiplexing technologies have been widely used in traditional optical communication and computer processing systems [16–19].

As the main candidate for the multiplexing technologies, polarization multiplexing schemes have been proposed for applications in several plasmonic elements. When the coupling structure is illuminated by the light with different polarization conditions, the information carried by the excited SPPs from the coupling structure is different. That is, under different incident light polarization conditions, the same coupling structure produces different SPPs modes. Such coupling structures are defined as SPPs polarization multiplexing devices [20]. A protruded silver spherical cap structure and grating couplers were designed to launch SPPs with the linearly polarized pulse, which can be utilized for polarization multiplexing to achieve plasmonic interconnection with broad bandwidth [13,20]. Moreover, a plasmonic achiral nano-coupler integrated with germanium-based plasmonic waveguide photodetectors was designed to couple circularly polarized pulses for realizing both spin-dependent routing and electrical detection of plasmons based on polarization multiplexing [21]. Among these devices, the utilization of linearly polarized pulse or circularly polarized pulse enabled more efficient processing of SPPs. However, the coupling structures used are relatively complicated. For protruded silver spherical cap structures, it brings a challenge to fabricate such delicate structures. Besides, grating structures have frequency limitations. Only when multiple gratings are integrated can the polarization multiplexing function be realized, and the larger structure size of the grating is not conducive to the miniaturization requirements of plasmonic devices. Basically, SPPs multiplexing schemes require versatile couplers capable of efficient coupling for arbitrary polarized input fields. As far as we know, a polarization multiplexing scheme that is based on a simple structure and suited to both linearly and circularly polarized laser pulses with a broad bandwidth has been rarely explored.

In this paper, we first demonstrate that the symmetrical double-groove structure can act as a polarization multiplexing device by imaging the noncollinear mode using photoemission electron microscopy (PEEM). It is found that the interfered SPPs field distributions of the symmetrical double-groove structure are different under different polarized incident laser irradiations. Besides, the simulation results show that the symmetrical positions on the upper and lower sides of the double-groove structure are in the same phase when irradiated with p-polarized light, but are in the anti-phase when irradiated with s-polarized light. By imaging the mode distribution of the SPPs field under different incident laser polarization conditions with two-color PEEM, we characterize the functionality of the symmetrical double-groove structure as a polarization multiplexing device. The resulting PEEM images are supported by FDTD simulations. The proposed multiplexer can improve the information transmission capacity and is a promising building block for plasmonics nanocircuits.

2. Methods

2.1 Experiment description

Two rectangular 5µm×1µm grooves were milled into ∼a 120-nm thick gold film by a focused ion beam (FIB) lithography. These two grooves are perpendicular to each other and the spacing between the end of the two grooves is 1 µm, as shown in the inset of Fig. 1. The gold film was deposited on a glass substrate coated with an approximate 180-nm-thick ITO layer via sputtering. The thickness of the fabricated gold film is much larger than the skin depth of Au, thereby the SPPs field at the Au/vacuum interface is considered in the experiment [22]. Figure 1 shows the schematic diagram of the experimental setup. A Ti: sapphire laser oscillator (Coherent, Mira 900), which provides ∼130 fs duration pulses at an 800 nm center wavelength with 76 MHz repetition rate is utilized. A beam-splitting mirror (BS1) is used to divide the laser pulse into two paths with equal intensity. One path entered the Mach–Zehnder type interferometer to generate two pulses pair with the same fundamental frequency of 800 nm, whereas the other path is employed for a second harmonic generation (SHG) by a β-BaBO₃ (BBO) crystal. The multiphoton photoemission signal from a superposition of SPPs and laser field is recorded using a photoemission electron microscope (PEEM, Focus GmbH). The incident laser is focused onto the sample surface using a 20 cm focal length lens at an incident angle of 65° from the surface normal, which is determined by the PEEM instrument. A typical spot size for the red laser is 50µm×25µm at the sample. Under these conditions, the spot sizes of the separate beams are adjusted such that the red pulse spot size is roughly 50% smaller than the blue pulse spot size at the sample position. For time-resolved PEEM experiments, the pulses are split into a stabilized Mach-Zehnder interferometer with an adjustable time delay, specifically, the incident pulse is split into two laser pulses through a beam splitter, and then they are spatially combined through a beam combiner. The time delay of the two laser pulses can be accurately obtained by adjusting the optical path difference of the two arms of the interferometer. In our experiment, the power range of the two fundamental frequency pulses is 30∼ 50 mW (pump light) and 10 mW (probe light), respectively, whereas the power of the second harmonic pulse is fixed at 5 mW. To achieve a better contrast interference pattern between SPPs and probe pulse in the detection region, the power of the probe pulse is slightly lower than that of the pump pulse. The details of the experimental scheme of investigation on the interference field distribution from the symmetrical grooves are shown in Fig. 2(a) and Fig. 3(a).

 

Fig. 1. Experimental setup of the two-color three beams scheme. The laser pulse is split by BS1 after passing through the silver reflective mirrors of M0 - M4. Part of the light reflected by the beam splitting enters the Mach-Zehnder interferometer. BS2 and BS3 correspond to a beam splitter and a combiner, respectively. M5 - M8 are silver mirrors. The partial beam passing through the BS1 is converted into SHG by a β-BaBO₃ (BBO) crystal. L1 and L2 are convex lenses with a focal length of 200 mm and 125 mm, respectively. After that, the two silver mirrors of M9 - M10 are used to reflect the 400nm light which is then combined with the 800nm pulse through the combiner BS4. Finally, the two-color pulse will enter into the PEEM chamber via a 20 cm focal length lens L3 and illuminate the sample with 65° to the normal of the sample surface. The inset in Fig. 1 shows the SEM image of a double-groove structure, which is consists of two 1 µm width and, 5 µm length grooves. The angle between the two grooves is 90°.

Download Full Size | PDF

 

Fig. 2. (a) Schematic for investigation of interference field from the excitation of the symmetrical double-groove via a pump-probe method. The SPPs generated by the p- or s-polarized pump pulse are detected by the p-polarized probe pulse at a remote point away from the double-groove structure. The red ellipse on the left and right sides represents the zones irradiated by the pump pulse (left side) and the probe pulse (right side), respectively. The near-field distributions (E_Z)² of the symmetrical double-groove after excitation with p-polarized laser pulses (b), and with s-polarized laser pulses (c), both are probed using p-polarized pulses. Photoemission interference patterns following excitation with 800 nm p-polarized spatial-separated femtosecond pulse pairs (d), and with s-polarized laser pulses (e), in the probe pulse area. The wave vectors of the incident light in-plane (k_l) and SPPs (ks) are displayed in (a). k_l is directed toward y.

Download Full Size | PDF

 

Fig. 3. (a) Schematic of the experimental setup for the two-color scheme. The SPPs generated by the 800 nm s-polarized pump pulse are detected by the 400nm-assisted 800-nm p-polarized probe pulse at a remote point away from the symmetrical double-groove structure. The red ellipse on the left and right sides represents the zones irradiated by the pump pulse (left side) and the probe pulse (right side), respectively. The second harmonic pulse irradiates the entire field of view for opening the two-color quantum channel in multiphoton photoemission to increase the photoemission yields. Photoemission interference patterns following excitation with p-polarized light (b), and with s-polarized light (c), both probed by a 400nm-assisted 800-nm p-polarized spatially separated femtosecond pulse pairs in the probe pulse area.

Download Full Size | PDF

2.2 Numerical simulation

Numerical simulations were performed using a commercial FDTD package (Lumerical FDTD Solutions). The simulations employ a total field-scattered plane wave source that allows monitoring pure instantaneous phase information of SPPs (does not contain the electric field of incident light). The time evolution of the SPPs near-field was captured by the point time monitor. The monitor was placed 5nm above the gold film. In the pump-probe scheme, two laser pulses with relative time delays are used, and the delay, as well as the polarization direction of the two laser pulses, can be flexibly adjusted in the FDTD simulation. The incident light parameters (central wavelength, pulse duration, and incident angle) were consistent with the experiment. The dielectric permittivity of gold is taken from the experimental data of Johnson and Christy [23]. The ITO-covered glass substrate was assumed to behave as a dielectric with an average refractive index of n = 1.55. The surrounding medium is the vacuum. The boundary conditions of FDTD are set with perfectly matched layers. To ensure an accurate description of the edges of the grooves, we used a 3D override mesh with a unit of size of 5×5×5nm³.

3. Results and discussion

Figure 2(a) shows a schematic representation of our experimental geometry following the irradiation with the pump and probe pulses. Excitation of the double-groove structure with ultrafast pump laser pulses launches surface plasmon wave packet propagating in the Y-direction. The probe pulse is spatially and temporally offset from the pump pulse, affording imaging of launched SPPs at distances about 100µm away from the coupling groove structure. Note that the direction of propagating SPPs (K_S) is not aligned with the direction of incident light (K_L), as shown in Fig. 2(a). This results from the fact that the relation of wave vectors of the laser (K_L) and SPPs (K_S) satisfies Snell’s law of refraction [24]. Conventional photoemission experiments show that the measured multiphoton signal is predominantly governed by the E_Z component of the near-field [25]. Figure 2(b) shows the simulated near-field distributions (E_Z)² of the symmetric double grooves after excitation with p-polarized pump and probe pulses. It is found that the near-field distributions (E_Z)² of SPPs are symmetrical and the resulting pattern displays strong circular interference beats that gradually spread to nearly crescent-shaped patterns. The pattern is caused by the interference between the SPPs excited by the incident light irradiating the symmetrical double-groove structure and the probe pulse. When the incident light polarization turned to s-polarization as shown in Fig. 2(c), it is found that the pattern induced by the interference between SPPs and probe pulse changes from a crescent shape to a staggered, herringbone shape along the central axis of the two groove structure as indicated by a dashed line in Fig. 2(a). By revealing the simulated interference mode distribution of the SPPs field under different polarization conditions, it shows that the symmetric double grooves have the potential to serve as plasmonic polarization multiplexing devices. In addition, we have also simulated the field distribution with different time delays under p-polarized pump and probe light excitation (see Supplement 1 Figure S1). The results show that SPPs mode gradually propagates to the right side with increasing of the time delays. Besides, according to our previous research, it is found that the groove structure has broadband coupling characteristics [24]. By simulating the broadband coupling characteristics of this structure under different excitation wavelengths (see Supplement 1 Figure S2), it suggests that the symmetric double-groove structure can achieve broadband coupling, which shows that the symmetric double-groove structure we proposed can be acted as a broadband polarization multiplexing device.

To experimentally verify the polarization multiplexing capability of the symmetric double-groove structure, we used PEEM to image the mode distribution of SPPs field excited in the symmetric double-groove structure under different polarization conditions. The experimental conditions are consistent with the simulated conditions. Figure 2(d) displays the photoemission interference patterns following excitation with 800-nm p-polarized spatially separated femtosecond pulse pairs in the probe pulse area. The resulting pattern displays circular interference beats that gradually spread to nearly crescent-shaped patterns in the center area. However, the interference formed around the circular interference is marginally distinguishable. Figure 2(e) displays the PEEM image of s-polarization excited SPPs with the pump-probe scheme. It shows that the mode distribution of SPPs field excited by s-polarized light is fuzzy and hardly recognizable, even if we applied a p-polarized probe pulse to interfere with the SPPs. Moreover, it is noted that the location of the fringes with maximum intensity for Fig. 2(b) and Fig. 2(d) is different. In Fig. 2(b), it can be seen that the most left strip corresponds to the maximum intensity and after that the fringe intensity gradually reduced. This is due to the attenuation characteristics of the SPPs field. However, in our experiment, as shown in Fig. 2(d), the maximum intensity fringe locates in the second strip rather than the first one. We attributed the difference to the fact that an elliptical light spot with Gaussian distribution for the laser focus is employed, the highest intensity area is mainly concentrated in the center of the focus spot and this leads to the field distribution as shown in Figs. 2(d), in which the middle stripe is brighter.

To securely confirm the SPPs field distribution from the two grooves, a two-color scheme as shown in Fig. 3(a) has been employed in the imaging of the SPPs field [26,27], which reduces the order of non-linearity in the multi-photon emission process and increases the photoemission yield to improve the imaging clarity. In the two-color scheme, at a distance of about 80 microns from the structure, a beam of 800-nm p-polarized light assisted with a beam of 400-nm light is combined to form probe pulses to image the SPPs field. Figure 3(b) displays the photoemission interference patterns following excitation with 400nm-assisted 800-nm p-polarized spatially separated femtosecond pulse pairs in the probe pulse area. Compared with the single-color experimental results as shown in Fig. 2(d), it is found that the interference fringes formed by SPPs and the probe light are not only distributed in the central area but also observed around the central area and the measured photoemission yields are increased by about 5.6 times. The sharpness of SPPs field imaging has been improved because of the increase in photoemission yields. Figure 3(c) displays the photoemission interference patterns following excitation with 800-nm s-polarized femtosecond pulse and detected with 400nm-assisted 800-nm p-polarized in the probe pulse area. The resulting pattern displays a staggered, herringbone shape mode distribution along the center, it is in agreement with the FDTD simulation as shown in Fig. 2(c). With the help of the two-color scheme, we have successfully revealed the mode distribution of the SPPs field excitation under different polarization directions.

Insights into the origin of the symmetric double grooves work as plasmonic polarization multiplexing devices can be obtained by considering the destructive and constructive interference of the SPPs modes excited by incident light with different polarization directions. Figures 4(a) and 4(b) show the mode description (E_Z) of the symmetric double grooves following irradiation with different polarization directions (p-polarized in Fig. 4(a) and (s)-polarized in Fig. 4(b)). It is found that when the incident light is p-polarized, the near-field of SPPs presents a symmetrical mode along the axis as indicated with dash lines in Figs. 4(a) and 4(b). The temporal evolution of the SPPs field captured on symmetrical positions of the upper and lower sides (the positions are depicted in the inset of the figure) are shown in Fig. 4(c). It is found that the SPPs field has the same phase at the symmetrical positions on the upper and lower sides of the double-groove structure. This is the reason why the photoemission interference pattern following p-polarized excitation displays symmetrical distribution, as shown in Fig. 3(b). When the incident light is s-polarized, the near-field of SPPs presents an anti-symmetrical mode, as shown in Fig. 4(b). At this time, it is found that the SPPs field has the phase delay ΔT= π at the symmetrical positions on the upper and lower sides of the double-groove structure, as shown in Fig. 4(d). When the probe pulse interferes with the SPPs excited by s-polarized, the interference fringes produced at the upper and lower sides are different. Due to the phase delay ΔT= π existing at the symmetrical positions on the upper and lower sides of the double-groove structure, when the probe pulse interferes with the SPPs field, both constructive interference, and destructive interference will occur in anti-symmetric positions. This causes the probe pulse tangled SPPs interference pattern to be a staggered, herringbone shape distribution along the central axis, as shown in Fig. 3(c) when the incident light is s-polarized. We should mention that even though the designed multiplexing structure is relatively larger for real application, the results we obtained here are scalable to miniaturization devices that can be integrated with the real nanocircuits. We simulated the near-field distribution of the symmetrical double-groove structure with a smaller size of 600nm×120nm under p/s-polarized light illumination (see Supplement 1 Figure S3). The results show that the symmetrical and anti-symmetric SPPs modes will still be excited respectively when p- and s-polarized light illuminate the smaller size double-groove structure. These results are consistent with those on a larger size of double-groove structure experimentally demonstrated in this work.

 

Fig. 4. (a-b) The mode description (E_Z) of the symmetric double grooves following laser irradiation with different polarization directions (p-polarized and s-polarized, (a and b)). (c and d) The simulated temporal evolution of SPPs near-field at the upper side (P₁) and lower side (P₂) of the symmetric double grooves following irradiation with different laser polarization directions (p-polarized (c) and s-polarized (d)). θ is the initial phase of the incident light.

Download Full Size | PDF

SPPs multiplexing schemes require versatile couplers being capable of excited by arbitrary polarized input fields [13]. To this end, we further explore the mode distribution of SPPs excited by a circularly polarized light. Figures 5(a) and 5(b) display the PEEM images following irradiation with circular polarization (left and right-hand circularly polarized) femtosecond pulses (λ = 800 nm, 1.55 eV) on the double grooves, respectively. The interference pattern is induced by the interaction of the incident femtosecond laser with the laser-induced SPPs. The asymmetric photoemission distributions can be observed for the different laser polarizations. Specifically, the interference pattern mainly appears on the lower side of the groove under the left-handed circular polarization light, indicating SPPs are preferentially excited on the lower side of the field. Inversely, when the right-handed circular polarization is used, the preferentially directional launch of the SPPs shift to the upper side of the field. Note that extremely bright hot spots have appeared on the groove edges and corners under femtosecond laser illumination due to strong localized surface plasmon resonance, and those hot spots will easily cause CCD saturation and affect the observation of the relatively weak SPPs field [24]. Therefore, for a clear presentation of the propagation of SPPs, we removed the double-groove structure out of PEEM’s field of view (FOV). The black dotted frames on the left of the PEEM images indicate the position of the double grooves as shown in Fig. 5(a).

 

Fig. 5. PEEM image of SPPs launching using the 800 nm femtosecond laser with left-handed circular polarization (a), and right-handed circular polarization (b). The black dotted frame on the left of PEEM images indicates the position of the symmetrical double-groove structure. The near-field distributions (E_Z)² of the symmetrical double-groove following irradiation of left-handed circular polarized light (c), and right-handed circular polarized light (d). The simulated temporal evolution of SPP near-field at P₁ (e) and P₂ (f) following irradiation with polarization states.

Download Full Size | PDF

Further, to explore the physical mechanism of the double-groove structure for coupling circularly polarized light, the temporal evolution of SPPs near-field at the positions P₁ (upper side of the groove) and P₂ (lower side of the groove) was calculated following irradiation with the left-hand circular polarization states, as shown in Figs. 5(e) and 5(f). It is noted that positions P₁ and P₂ are outside the range of incident light. As we all know, the light in any polarization state can be decomposed into two orthogonally polarized components. As the example under the left-handed circular polarized state, it can be formed through the superposition of p- and s- polarization with an initial phase difference of -π/2. We find that the time evolution of the near electric field of the two SPPs modes at the position P₁ is close to the anti-phase phase, but approximately the same at P₂, as shown in Figs. 5(e) and 5(f). The results imply that when the above two SPPs modes are superimposed, destructive interference occurs on the upper side of the field and constructive interference occurs on the lower side of the field, which will result in constructive SPPs on the lower side of the field, as shown in Figs. 5(a) and 5(c). Therefore, the directional launching of the SPPs caused by the left-handed circular polarized light in Fig. 5(a) can be attributed to the superposition of the SPPs modes excited by the incident light with the polarization of p- and s-, respectively. The same rule applies to explain the asymmetrical SPPs field distribution in Figs. 5(b) and 5(d) under irradiation of right-hand circular laser pulse.

There are some related studies on the polarization-controlled plasmons field distribution [28–32], but our work is different from their reports. Aeschlimann et.al achieved the adaptive control of the localized surface plasmon field in the two dimensions of time and space by shaping the polarization of the incident laser pulse [28]. They studied the localized surface plasmon rather than propagating surface plasmon polaritons as we did in this work. In addition, our main purpose is to design and characterize a broadband plasmon polarization multiplexing device via imaging the mode distribution of the SPPs field excited by differently polarized light using PEEM, whereas their purpose is to control the field distribution of localized surface plasmon. C. Rewitz et.al. and E. Krauss et.al. designed the Y-shaped protrusion structure and adjusted the polarization of the incident laser to achieve selective routing of SPPs [29,30]. Their scheme cannot access the SPPs mode distribution in the near field and lacks broadband coupling capabilities that we have demonstrated here. In addition, the structure they used is not easy to process. D. Tyagi et.al uses the cleverly designed crossed Babinet-inverted nanoantennas structure to realize the omnidirectional control of the SPPs field in the first quadrant via adjusting the incident laser polarization. However, it can only work under a specific wavelength excitation, which means a broadband coupling could not be achieved in this crossed Babinet-inverted nanoantennas structure [31]. Y. Dai et al. reported to analyze the physical mechanism of the directional excitation of SPPs on both sides of the silver convex islands under different polarization conditions and demonstrated that the directional excitation of SPPs under circular polarization conditions is caused by the matching spin angular momenta (SAM) of light and the transverse SAM of SPPs [32]. However, in their scheme, due to limitation of structure size, it is impossible to study the mode distribution of SPPs field under the condition of broadband and to eliminate the influence by the reflected SPPs at the rim of the island. Additionally, our work is also different from that of A. G. Joly’s [13]. In our work, we demonstrated that the polarization multiplexer has femtosecond broadband response characteristics and realize the function of polarization multiplexing under the excitation of circularly polarized light, and those properties were not reported in their work. The polarization multiplexing device we designed is composed of symmetrical double grooves, which is easier to process than protruded silver spherical cap structure as reported by A.G. Joly. More importantly, we applied a two-color PPEM technique to visualize the weakly excited SPPs field of this multi-multiplexing device in the near field. The results show that the two-color excitation method can significantly optimize the PEEM imaging contrast compared to the monochromatic scheme.

4. Conclusions

In summary, we demonstrated that the symmetrical double-groove structure can act as a polarization multiplexer by investigating the mode distributions with a PEEM. We find that the field from the double-groove structure has different mode distributions under different polarized incident laser irradiation. The simulation results show that the symmetrical positions on the upper and lower sides of the double-groove structure are in the same phase when irradiated with p-polarized light, but are in the anti-phase when irradiated with s-polarized light. The resulting PEEM images are supported by the FDTD simulations. These findings prove that the symmetrical double-groove structure can be used as a broad bandwidth polarization multiplexer, which supports the frequency and polarization multiplexing function simultaneously. This study paves a way for accurately characterizing the functionality of the plasmonic functional elements and the proposed multiplexer can be a promising building block for highly integrated optoelectronic integrated circuits.