Surface Plasmon-Based Pulse Splitter and Polarization Multiplexer

Oct 11, 2018 - Physical Sciences Division, Pacific Northwest National Laboratory, P.O. ... be utilized for either polarization multiplexing or as a pu...
0 downloads 0 Views 3MB Size
Download PDF
Letter Cite This: J. Phys. Chem. Lett. 2018, 9, 6164−6168

pubs.acs.org/JPCL

Surface Plasmon-Based Pulse Splitter and Polarization Multiplexer Alan G. Joly,* Yu Gong, Patrick Z. El-Khoury, and Wayne P. Hess Physical Sciences Division, Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99352, United States

Downloaded via UNIV OF SUNDERLAND on October 11, 2018 at 19:52:13 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Surface plasmon polaritons (SPPs) launched from a protruded silver spherical cap structure using s-polarized femtosecond laser excitation are investigated using photoemission electron microscopy. The resulting SPP is comparable in intensity to SPPs launched with p-polarized excitation but propagates with a distinct spatial profile. The spatial and temporal properties of the nascent SPP are determined by splitting the femtosecond pulse into a spatially separated pump−probe pair of orthogonal polarizations. The s-polarized pump pulse initiates the SPP, which is then visualized by the photoelectron emission induced by a spatially and temporally separated p-polarized probe pulse. The s-polarization launched SPP displays a bifurcated spatial structure with an antisymmetric mirror plane and may be regarded as two spatially distinct, temporally phase-locked wave packets. Significantly, the wave packets are onehalf period out of phase with each other governed by the phase of the driving laser field. Finite difference time domain calculations corroborate the experimental results. The resulting SPP can be utilized for either polarization multiplexing or as a pulse splitter in nanophotonic circuits.

S

effectively increase transmission and hence efficiency in nanoplasmonic circuits.14 Here, we demonstrate that protruding hemispherical structures efficiently couple s-polarized light-forming SPPs with two spatially and temporally distinct channels. These channels are spatially separate from SPPs launched following ppolarization from the same structure, in essence providing a 50% SPP pulse splitter in addition to allowing polarization multiplexing. We utilize photoemission electron microscopy (PEEM) in conjunction with 15 fs laser pulses to visualize SPP propagation2,15−21 in real space and time. Detailed analysis of the resulting time-resolved PEEM images is consistent with finite-difference time-domain (FDTD) calculations and predicts that other coupling structures may also efficiently couple s-polarized excitation. Figure 1A displays an SEM image of a 9 μm diameter selfassembled hollow hemispherical cap structure of height 940 nm. Figure 1B displays a schematic illustration of a single pulse PEEM experiment on the silver cap. A 780 nm 15 fs laser pulse is incident on the sample at 75 degrees to the surface normal, generating an SPP from the hemispherical structure. As noted previously,15 the nascent SPP interferes with the residual laser pulse at and beyond the generation point, leading to an interference pattern in the direction of SPP propagation. Figure 1C displays a laser-excited PEEM image following irradiation using 15 fs, p-polarized 780 nm pulses, and Figure 1D displays an analogous image except using s-polarized pulses. Under 780 nm fs laser excitation, absorption of at least three photons is required to induce photoemission. The SPP field dramatically increases the probability of multiphoton

urface plasmon polaritons (SPPs) have received a great deal of attention largely because of their promise for use in next-generation, miniaturized electronic circuits and devices. SPPs are attractive information carriers as they comprise propagating electromagnetic fields coupled to charge density fluctuations at the surface of metals, with group velocities near the speed of light in a vacuum.1−3 An overarching goal is to design and construct subcomponents operative below the diffraction limit with high transmission speeds and low thermal loads. Designing nanocircuit elements such as waveguides,4 amplifiers,5 and demultiplexers6 requires the ability to launch, manipulate, and guide SPP waves to remote locations. Most often, this is achieved via tailored coupling structures lithographically etched into flat metal surfaces, although we have recently shown that extruded structures can also be used to achieve efficient coupling following p-polarized excitation.7 The utility of polarization control has previously been an important method for directing SPPs along specific paths.7−13 SPPs contain both a transverse field component perpendicular to the surface and a longitudinal component that oscillates along the propagation direction. Typical non-normal incidence excitation utilizes p-polarized laser pulses near grazing incidence, producing both transverse and longitudinal field components that may couple to the nascent SPP, whereas spolarized light does not result in efficient SPP excitation. Although commonly employed, this excitation geometry is limited to single-bit binary logic (on/off). This limits the ultimate information transmission bandwidth as polarization modulation cannot, in principle, be used as a form of multiplexing. SPP multiplexing schemes require versatile couplers capable of efficient coupling for arbitrary polarized input fields. Recent experiments using defects to allow spolarization coupling underscore this important point and demonstrate one way to use multiple polarizations to © XXXX American Chemical Society

Received: August 27, 2018 Accepted: October 8, 2018

6164

DOI: 10.1021/acs.jpclett.8b02643 J. Phys. Chem. Lett. 2018, 9, 6164−6168

Letter

The Journal of Physical Chemistry Letters

Figure 1. (A) SEM image of a 9 μm diameter, 940 nm high silver spherical cap structure. (B) Schematic of the experimental setup for single-beam excitation. The laser is incident from the left at a 75° angle of incidence relative to the sample surface normal and irradiates the surface region denoted by the white dashed circle. (C) Hemispherical cap PEEM image after excitation from a p-polarized 780 nm fs laser. (D) PEEM image of the same cap following illumination by s-polarized 780 nm fs laser pulses. The red dashed circles in (C) and (D) denote the location of the spherical cap.

Figure 2. (A) Schematic illustration of the spatially separated pump−probe experiment. The p- or s-polarized pump pulse launches the SPP, and the p-polarized probe pulse interrogates the SPP at a remote point. (B) PEEM image following excitation with 780 nm p-polarized spatially separated femtosecond pulse pairs. (C) PEEM image of the same hemisphere but recorded following excitation with s-polarized laser pulses and probed using p-polarized pulses. Red dashed circles mark the location of the hemisphere in (B) and (C), and white dashed ovals indicate the approximate positions of the spatially separated pump and probe beams.

to increase the photoelectron yield, and the pump pulse can be set to any polarization. In Figure 2B, the pump pulse is set to p-polarization, and the resulting pattern displays strong circular interference beats that gradually spread to nearly crescentshaped patterns. The initial region is primarily a manifestation of the pump and SPP interference pattern because the probe pulse has little overlap as indicated by the dashed white circles in the figure. Beyond this region, the interference pattern is determined primarily by the probe-SPP interference as there is little pump overlap. In contrast, Figure 2C displays the identical configuration except that the pump pulse in now spolarized. There are two important results obtained by comparing Figure 2B and C. First, s-polarized excitation launches an SPP analogous to p-polarized excitation except that the photoelectron yield is approximately a factor of 2.5

absorption, and therefore, the photoemission intensity allows direct visualization of the SPP field. The relative electron yield of p-polarized light excitation to s-polarized excitation displayed in Figure 1 is over ten-to-one, reflecting two factors. The first factor is that s-polarized light does not usually couple efficiently to the SPP. The second factor is that s-polarized light induces only weak photoemission as observed previously in PEEM measurements.8,22 For the second issue to be overcome, the illuminating pulse can be split into a pump−probe pulse pair, and the spatial position, polarization, and timing of each can be independently controlled as displayed in the Figure 2A schematic. Here, the original pulse is split into two nearly equal intensity pulses and spatially separated to observe the SPP propagation at positions remote from the hemisphere.2,22 The probe pulse is p-polarized 6165

DOI: 10.1021/acs.jpclett.8b02643 J. Phys. Chem. Lett. 2018, 9, 6164−6168

Letter

The Journal of Physical Chemistry Letters

Figure 3. (A, B) PEEM images following excitation with 780 nm s-polarized femtosecond pump pulses and p-polarized probe pulses separated by approximately one-half period (1.24 fs). (C) Time trace of electron yield from the red boxed areas in (A) and (B) as a function of relative delay between the pulses. The resulting trace demonstrates that the SPP in the upper plane is π radians out-of-phase with that generated in the lower plane. Zero delay is defined as the delay in which the pump and probe pulses are overlapped in time. The traces are offset vertically for clarity.

is exactly one-half period (π radians) out-of-phase with its counterpart launched in the lower half plane. The implications of Figures 2 and 3 are somewhat remarkable. First, s-polarized excitation can launch an SPP from a hemispherically shaped cap structure with comparable (2−3× lower) intensity relative to p-polarized excitation. This occurs even though the electric field vector of the s-polarized pulse is perpendicular to both the transverse and longitudinal field vectors of the SPP. Second, the resulting SPP may be decomposed into two spatially separate SPPs launched from the upper and lower planes, respectively. These two SPPs are launched with exactly one-half period phase shift. In other words, when the electric field vector of the excitation pulse oscillates toward the upper plane, the launched SPP field oscillation undergoes a maximum in the upper plane. When the electric field vector is reversed, the upper plane SPP oscillation is at a minimum. The reverse process occurs in the lower plane leading to a π radian phase shift between the two wave packets. This interpretation assumes there is no phase shift between the s-polarized driving field and the nascent SPP. However, any phase shift between the driving field and the SPP would be symmetric about the hemispherical cap center line such that the two SPP wave packets are always one-half period out-ofphase. To supplement our experimental results, we have performed FDTD calculations for s-polarized excitation of a 9 μm diameter hollow silver hemisphere on a surrounding silver film as well as other similar structures. The calculations employ a total field scattered field plane wave source that allows monitoring of either the total field (laser plus SPP) or scattered field (SPP) only. Figure 4 displays the results of these calculations in the region between 20 and 26 μm beyond the coupling structure where the field is dominated by the scattered SPP wave. Figure 4A displays the results for single plane wave excitation impinging on the hollow 9 μm diameter hemisphere. Figure 4A displays the E field z component (normal to the surface) enhancement monitored at the surface

weaker. Second, the probe pulse SPP interference pattern changes from a crescent shape to a staggered, herringbone shape with an antisymmetric mirror plane along the center. Panels B and C in Figure 2 display images at a time delay between the pump and probe pulses adjusted such that the SPP has traveled to the probe region. Scanning the pump− probe delay allows a sequence of PEEM images (Supplemental Movie) to be obtained where each image represents a different time delay. Figure 3 displays two selected images from the movie in which the delay difference is 1.24 fs, which represents nearly one-half period of phase evolution (3.0 radians). The spatial profiles show a dramatic difference. In Figure 3A, the SPP-laser interference pattern is manifested mostly in the lower plane, whereas in Figure 3B, the pattern is primarily in the upper plane. In both cases, the SPP propagates radially away from the center line as the distance from the hemispherical coupling structure increases. Figure 3C displays a trace of the photoelectron yield as a function of delay between the s-polarized pump and ppolarized probe pulses. The yield is determined by integrating the photoelectron emission from a symmetric region above and below the center line as indicated by the red boxes in Figure 3A and B. The resulting trace displays the photoelectron intensity resulting from the probe laser/SPP interference in time; thus, regularly spaced fringes result whenever the polarization fields overlap constructively, and valleys occur when the fields overlap destructively. Time zero is defined as the delay that maximizes the photoelectron yield from overlap between the probe pulse and the far edge of the pump pulse (pump−probe overlap).2 Therefore, the ∼25 fs delay in photoemission yield maximum noted in Figure 3C represents the delayed arrival of the SPP wave packet due to the slower SPP group velocity relative to the speed of light. Previous measurements have established that the SPP velocity is 0.93c on silver following 780 nm excitation23 consistent with the measurements recorded here. Figure 3C demonstrates that the SPP launched in the upper half plane (as shown in Figure 3B) 6166

DOI: 10.1021/acs.jpclett.8b02643 J. Phys. Chem. Lett. 2018, 9, 6164−6168

Letter

The Journal of Physical Chemistry Letters

protruded hemispherical structure in both intensity and surface field pattern. Figure 4D displays calculations from a 9 μm diameter, 900 nm deep cylindrical hole. Although the cylindrical hole has some similarities to the spherical cap, the surface field enhancement pattern is more dispersed laterally with weak central intensity. We have also run calculations on a 2 μm × 8 μm trench etched into silver. The results indicate that the z-component of the field enhancement is approximately a factor of 10 less than for the hemispherical cap and does not produce a strong bifurcated spatial field pattern. Overall, the calculations display more intense, bifurcated SPP fields for curved structures such as the dimple and the protruded hemisphere. Coupling structures such as circles and ellipses can focus SPPs efficiently,25,26 indicating that rounded structures can generate distinctive, concentrated SPP field patterns. Thus, other curved structures are likely to show appreciable s-polarization-launched SPP field intensity as well. For these structures, it is clear from Figure 1C that the ppolarization launched SPP propagates directly along the center line, whereas from Figure 3A and B, the s-polarizationlaunched SPP propagates radially away from the center line. In other words, farther propagation distance results in better separation of the two channels (s- and p-input polarization), facilitating use of these structures for polarization multiplexing or pulse splitting. In conclusion, s-polarized femtosecond excitation of silver spherical cap structures launches strong SPPs with a spatial pattern distinct from the p-polarized analogue. The spatial pattern may be decomposed into two nearly identical SPP wave packets with only a π phase shift differentiating the two. FDTD calculations demonstrate that the z-field component of the s-polarization-launched SPP displays a similar pattern with an antisymmetric mirror plane through the center. Such a construct can therefore be used as a polarization multiplexer or a 50% SPP pulse splitter in nanophotonic circuits.

■

Figure 4. Results from FDTD calculations displaying the zcomponent enhancement of the SPP electric field in the region between 20 and 26 μm beyond the coupling structures depicted on the left: (A) p-polarized excitation of a hemispherical cap, (B) spolarized excitation of a hemispherical cap, (C) s-polarized excitation of an inverted hemispherical dimple, and (D) s-polarized excitation of a 9 μm diameter, 900 nm deep cylindrical hole. The 780 nm laser pulses are incident from the left at an angle of 75 degrees to the surface normal. The black scale bars indicate 1 μm distance.

METHODS

The hemispherical cap structure is produced upon deposition of ∼100 nm of silver on a silicon substrate cleaned in organic solvent as described previously.7 A mode-locked Ti:sapphire laser produces 15 fs (fs) pulses centered at 780 nm at a repetition rate of 90 MHz. For dual-beam experiments, the laser beam is split into a “pump” and “probe” pair with roughly 50 mW average power per beam. The pulses are interferometrically locked using a Mach-Zender interferometer with each arm having separate polarization control. The incident laser sources are focused onto the sample surface at an incident angle of 75 degrees with respect to the surface normal, leading to highly elliptical laser spots on the sample surface (∼100 μm × 30 μm Gaussian fwhm). The pump and probe beams are spatially offset to facilitate visualization of the nascent SPP.2 Numerical simulations were performed using a commercial FDTD package (Lumerical FDTD Solutions). The calculations employ a total field scattered field plane wave source that allows monitoring of either the total field (laser plus SPP) or scattered field (SPP) only. This configuration allows field enhancement determination of the SPP (scattered) field as it propagates away from the spherical cap. Model laser pulses of 80 fs are used to ensure that the angular dispersion in the injected wave vector is minimized. The dielectric constants for silver were obtained from Yang et al.27

for p-polarization, and Figure 4B displays analogous results for s-polarization. We choose to display the z component as it readily couples to the p-polarized probe field to achieve efficient ionization. It is important to note that the FDTD calculations do not directly correspond to, nor replicate, the PEEM data. For calculational efficiency, the FDTD calculations display the z-component of the field over a limited region, particularly in the axis perpendicular to propagation. The actual PEEM signal is the time-integrated polarization response to both input pulses as has been described previously.24 According to Figure 4, the p-polarized zcomponent displays fringes at the laser wavelength that are symmetric with respect to the plane bisecting the sphere. In contrast, the s-polarized z-component displays antisymmetric fringes that are π radians out-of-phase. Figure 4C displays FDTD calculations on a 9 μm diameter inverted hemispherical dimple of depth 900 nm. The results from the inverted hemisphere demonstrate similarities to the 6167

DOI: 10.1021/acs.jpclett.8b02643 J. Phys. Chem. Lett. 2018, 9, 6164−6168

Letter

The Journal of Physical Chemistry Letters

■

(11) Sukharev, M.; Seideman, T. Coherent Control of Light Propagation via Nanoparticle Arrays. J. Phys. B: At., Mol. Opt. Phys. 2007, 40, S283−S298. (12) Lin, J.; Mueller, J. P.; Wang, Q.; Yuan, G.; Antoniou, N.; Yuan, X. C.; Capasso, F. Polarization-Controlled Tunable Directional Coupling of Surface Plasmon Polaritons. Science 2013, 340, 331−334. (13) Chen, J.; Li, Z.; Yue, S.; Gong, Q. Efficient Unidirectional Generation of Surface Plasmon Polaritons with Asymmetric SingleNanoslit. Appl. Phys. Lett. 2010, 97, 041113. (14) Chen, J.; Sun, C.; Rong, K.; Li, H.; Gong, Q. Polarization-Free Directional Coupling of Surface Plasmon Polaritons. Laser Photonics Rev. 2015, 9, 419−426. (15) Zhang, L.; Kubo, A.; Wang, L.; Petek, H.; Seideman, T. Imaging of Surface Plasmon Polariton Fields Excited at a NanometerScale Slit. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 84, 245442. (16) Kahl, P.; Wall, S.; Witt, C.; Schneider, C.; Bayer, D.; Fischer, A.; Melchior, P.; Horn-von Hoegen, M.; Aeschlimann, M.; Meyer zu Heringdorf, F.-J. Normal-Incidence Photoemission Electron Microscopy (NI-PEEM) for Imaging Surface Plasmon Polaritons. Plasmonics 2014, 9, 1401−1407. (17) Shibuta, M.; Eguchi, T.; Nakajima, A. Imaging and Characterizing Long-Range Surface Plasmon Polaritons Propagatin in a Submillimeter Scale by Two-Color Two-Photon Photoelectron Emission Microscopy. Plasmonics 2013, 8, 1411−1415. (18) Lemke, C.; Leissner, T.; Jauernik, S.; Klick, A.; Fiutowski, J.; Kjelstrup-Hansen, J.; Rubahn, H.-G.; Bauer, M. Mapping Surface Plasmon Polariton Propagation via Counter-Propagating Light Pulses. Opt. Express 2012, 20, 12877−12884. (19) Dabrowski, M.; Dai, Y.; Argondizzo, A.; Zou, Q.; Cui, X.; Petek, H. Multiphoton Photoemission Microscopy of High-Order Plasmonic Resonances at the Ag/Vacuum and Ag/Si Interfaces of Epitaxial Silver Nanowires. ACS Photonics 2016, 3, 1704−1713. (20) Yamagiwa, K.; Shibuta, M.; Nakajima, A. Two-Photon Photoelectron Emission Microscopy for Surface Plasmon Polaritons at the Au(111) Surface Decorated with Alkanethiolate Self-Assembled Monolayers. Phys. Chem. Chem. Phys. 2017, 19, 13455−13461. (21) Frank, B.; Kahl, P.; Podbiel, D.; Spektor, G.; Orenstein, M.; Fu, L.; Weiss, T.; Horn-von Hoegen, M.; Davis, T. J.; Meyer zu Heringdorf, F.-J.; Giessen, H. Short-Range Surface Plasmonics: Localized Electron Emission Dynamics from a 60-nm Spot on an Atomically Flat Single-Crystalline Gold Surface. Sci. Adv. 2017, 3, No. e1700721. (22) Podbiel, D.; Kahl, P.; Meyer zu Heringdorf, F.-J. Analysis of the Contrast in Normal-Incidence Surface Plasmon Photoemission Microscopy in a Pump-Probe Experiment with Adjustable Polarization. Appl. Phys. B: Lasers Opt. 2016, 122, 1−6. (23) Joly, A. G.; El-Khoury, P. Z.; Hess, W. P. Spatiotemporal Imaging of Surface Plasmons Using Two-Color Photoemission Electron Microscopy. J. Phys. Chem. C 2018, 122, 20981−20988. (24) Gong, Y.; Joly, A. G.; El-Khoury, P. Z.; Hess, W. P. Enhanced Propagating Surface Plasmon Signal Detection. ACS Photonics 2016, 3, 2413−2419. (25) Liu, Z.; Steele, J. M.; Lee, H.; Zhang, X. Tuning the Focus of a Plasmonic Lens by the Incident Angle. Appl. Phys. Lett. 2006, 88, 171108. (26) Liu, Z.; Steele, J. M.; Srituravanich, W.; Pikus, Y.; Sun, C.; Zhang, X. Focusing Surface Plasmons with a Plasmonic Lens. Nano Lett. 2005, 5, 1726−1729. (27) Yang, H. U.; D’Archangel, J.; Sundheimer, M. L.; Tucker, E.; Boreman, G. D.; Raschke, M. B. Optical Dielectric Function of Silver. Phys. Rev. B: Condens. Matter Mater. Phys. 2015, 91, 235137.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.8b02643. Time-resolved PEEM movie displaying SPP propagation from a hemispherical cap following s-polarized femtosecond pulse excitation at 780 nm (AVI)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Alan G. Joly: 0000-0003-2931-4524 Patrick Z. El-Khoury: 0000-0002-6032-9006 Wayne P. Hess: 0000-0002-3970-9282 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge support from the US Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences & Biosciences. This work was performed in EMSL, a national scientific user facility sponsored by DOE’s Office of Biological and Environmental Research and located at PNNL. PNNL is operated by Battelle Memorial Institute for the United States Department of Energy.

■

REFERENCES

(1) Kahl, P.; Podbiel, D.; Schneider, C.; Makris, A.; Sindermann, S.; Witt, C.; Kilbane, D.; Horn-von Hoegen, M.; Aeschlimann, M.; Meyer zu Heringdorf, F. Direct Observation of Surface Plasmon Polariton Propagation and Interference by Time-Resolved Imaging in NormalIncidence Two Photon Photoemission Microscopy. Plasmonics 2018, 13, 239−246. (2) Gong, Y.; Joly, A. G.; Hu, D. H.; El-Khoury, P. Z.; Hess, W. P. Ultrafast Imaging of Surface Plasmons Propagating on a Gold Surface. Nano Lett. 2015, 15, 3472−3478. (3) Kubo, A.; Pontius, N.; Petek, H. Femtosecond Microscopy of Surface Plasmon Polariton Wave Packet Evolution at the Silver/ Vacuum Interface. Nano Lett. 2007, 7, 470−475. (4) Maier, S. A.; Kik, P. G.; Atwater, H. A.; Meltzer, S.; Harel, E.; Koel, B. E.; Requicha, A. A. G. Local Detection of Electromagnetic Energy Transport Below the Diffraction Limit in Metal Nanoparticle Plasmon Waveguides. Nat. Mater. 2003, 2, 229−232. (5) De Leon, I.; Berini, P. Amplification of Long-Range Surface Plasmons by a Dipolar Gain Medium. Nat. Photonics 2010, 4, 382− 387. (6) Zhao, C.; Zhang, J. Plasmonic Demultiplexer and Guiding. ACS Nano 2010, 4, 6433−6438. (7) Gong, Y.; Joly, A. G.; Zhang, X.; El-Khoury, P. Z.; Hess, W. P. Surface Plasmon Coupling and Control Using Spherical Cap Structures. J. Phys. Chem. Lett. 2017, 8, 2695−2699. (8) Gong, Y.; Joly, A. G.; El-Khoury, P. Z.; Hess, W. P. PolarizationDirected Surface Plasmon Polariton Launching. J. Phys. Chem. Lett. 2017, 8, 49−54. (9) Fang, Y.; Li, Z.; Huang, Y.; Zhang, S.; Nordlander, P.; Halas, N. J.; Xu, H. Branched Silver Nanowires as Controllable Plasmon Routers. Nano Lett. 2010, 10, 1950−1954. (10) Sukharev, M.; Seideman, T. Phase and Polarization Control as a Route to Plasmonic Nanodevices. Nano Lett. 2006, 6, 715−719. 6168

DOI: 10.1021/acs.jpclett.8b02643 J. Phys. Chem. Lett. 2018, 9, 6164−6168