In this paper we demonstrate the realization of an autocorrelator for the characterization of ultrashort surface plasmon polariton (SPP) pulses. A wedge shaped structure is used to continuously increase the time delay between two interfering SPPs. The autocorrelation signal is monitored by non-linear two-photon photoemission electron microscopy. The presented approach is applicable to other SPP sensitive detection schemes that provide only moderate spatial resolution and may therefore be of general interest in the field of ultrafast plasmonics.
© 2013 OSA
Surface plasmon polaritons, propagating electromagnetic waves which are bound to a metal-dielectric interface, are considered as a main component in future broadband and ultrafast nano-photonic devices [1–4]. Motivated by this perspective, several techniques have been successfully applied in the recent past to particularly track the propagation of SPP pulses. Examples include time-resolved scanning near-field optical microscopy (SNOM) [5, 6], two-photon luminescence (TPL) [7, 8], interferometric time-resolved photoemission electron microscopy (ITR-PEEM) [9–11], far-field spectral interferometry  and interfermetric cavity ring-down spectroscopy . Intrinsically, such experiments as well as potential applications ask for methods that enable one to determine in-situ the temporal and spectral characteristics of the SPP pulses. Autocorrelation techniques are common tools used in conventional optics to perform these tasks . A main challenge in transferring such schemes to the sub-micron scale is the implementation of an interferometer that allows for controlled adjustment of the temporal delay between the correlation pulses. Mach-Zehnder interferometers, which are typically used in autocorrelators, have been integrated into SPP devices in the past [15, 16]. These approaches are, however, restricted to a fixed time delay, due to the lack of a movable stage. Alternative interferometer schemes in conventional optics make use of tilted pulse correlation geometries projecting the time delay onto a spatial coordinate. Single-shot autocorrelators are for instance constructed in this manner . In plasmonics a micro-interferometer based on this principle has been demonstrated by using a tilted slit-groove geometry . In the present paper we use a simple wedge shaped structure of gold as a plasmonic interferometer to realize a plasmonic autocorrelator. A scanning electron microscope image of the gold wedge, which is deposited on top of a homogeneous gold film, is shown in Fig. 1(a). In this device, the increase in the edge–to-edge distance from the bottom to the top is translated into the required SPP pulse-to-pulse time delay variation, as illustrated in Fig. 1(c). The actual SPP autocorrelation signal is detected at the gold film to the right of the wedge. We will particularly show the sensitivity of this autocorrelation signal to the SPP pulse width demonstrating the principle capability of the presented device.
The present study uses photoemission electron microscopy (PEEM) to detect the SPP autocorrelation signal. However, we would like to emphasize that the presented approach can also be used in combination with other SPP detection schemes, such as TPL, SNOM and leakage radiation microscopy, and particularly does not rely on sub-diffraction resolution capabilities or the availability of an optical pump-probe setup.
The samples used for the experiments consist of 140 nm thick gold films which were evaporated by electron beam metal deposition onto silicon substrates. The wedge shaped structures were fabricated on top of the gold layer by electron beam lithography, 140 nm gold deposition, and subsequent lift-off. The sample was covered with an additional 100 nm gold layer to avoid spurious artefacts on the PEEM signal from residuals from the lift-off-process.
Two different samples with the following dimensions have been used in this study. The wedge structure A has an overall lateral extension of L = 98.5 μm and an opening angle of α = 2.93°. For preparational reasons the sharp tip of structure A was replaced by a straight part with a length of 10 μm and a width of 400 nm. For structure BL = 96.5 μm and α = 2.85° were used.
The photoemission electron microscope experiments were performed using a microscope (IS PEEM, Focus GmbH ) mounted in an ultrahigh vacuum μ-metal-chamber, providing a lateral resolution of better than 40 nm. PEEM imaging was performed in conventional threshold PEEM mode with a mercury lamp (hν = 4.9 eV) as well as in two-photon PEEM mode (2P-PEEM) [11, 20–24]. In order to facilitate the two-photon photoemission process with near-infrared light pulses, the work function of the gold surface was lowered from 5.5 eV to below 3 eV by covering it with a small amount of Cesium prior to the 2P-PEEM experiments.
SPP pulses were generated at illumination of the wedge structure under a fixed angle of incidence (65°) towards the surface normal by 100 fs laser pulses from a tunable Ti:Sapphire oscillator (Tsunami, Spectra Physics; tuning range: 710 to 890 nm). For the generation of ultrashort SPP pulses laser pulses from a Ti:Sapphire oscillator (Griffin, KM labs) with a center wavelength of about 800 nm and pulse lengths down to 10 fs were used. The pump-probe measurements presented at the end of this study were conducted with an optical Mach-Zehnder interferometer. In contrast to former ITR-PEEM experiments , no active stabilization of the interferometer was used.
Figure 1(a) shows a scanning electron microscope image of the wedge structure A. As will be shown later, the dimensions are related to a variation of the time delay between 3.0 and 32.7 fs from the bottom to the top of the wedge at a SPP group velocity of vg = 0.938 c and at λSPP = 783.5 nm. The group velocity vg and the SPP wavelength λSPP at a gold-vacuum interface are given by the SPP dispersion relation calculated under consideration of the permittivity data as reported in [27, 28]. Figure 2(a) is an image of the wedge recorded in conventional threshold PEEM at an excitation energy of hν = 4.9 eV. At this photon energy light-coupling to SPP modes at the gold-vacuum interface is not possible and the observed contrast reflects the geometry of the surface structure. The identical sample position, now monitored in 2P-PEEM mode, is shown in Fig. 2(b). The laser is incident from the right at a central wavelength of λL = 800 nm. The image contrast has completely changed and is now dominated by two distinct periodic photoemission intensity patterns emerging from the left and the right edge of the wedge. Both signatures arise from the emission of SPP pulses into the homogeneous gold film, which are excited by laser-coupling to the wedge edges. The periodic pattern to the left of the wedge is formed by the superposition of illuminating laser and co-propagating SPP (i. e. the SPP propagates in the same direction as the incident laser pulse). Its origin has been discussed and analyzed in numerous works before [7, 9–11, 24]. It will not be of relevance for the rest of this study. Instead, we will focus in this work onto the pattern located at the right-hand side of the wedge exhibiting a periodicity which is oriented almost perpendicular to the excitation edges of the wedge. This pattern arises from the interference of two SPP pulses, which are excited at the left- and the right-hand edge of the wedge, respectively, and which are propagating in opposite direction with respect to the incident laser pulse. An intensity pattern of alternating constructive and destructive interference emerges as the two SPPs meet at the right-hand edge with a phase delay which is continuously modulated by the increase in the edge-to-edge distance of the wedge. Signatures of the individual SPP pulses excited at the two edges become visible at an increased image magnification as shown in Fig. 2(c), which is a zoom into the rectangular area highlighted in red in Fig. 2(b). The two distinct, high periodicity patterns on top of the wedge structure as well as to the right of the wedge structure, marked in blue, are the result of the superposition between illuminating laser and the two SPP pulses, respectively, which are, in contrast to the former case, probed now in a counter-propagating manner . Note that also in Fig. 2(c) areas of constructive and destructive interference of the two SPPs are clearly visible in the right-hand area of the image. In the following, we will perform a detailed quantitative analysis of the SPP-SPP interference pattern, and finally, based on these results, show that this correlation signal can be used to monitor SPP pulse characteristics.
The interference pattern should substantially depend on the wavelength λSPP of the superimposed SPPs. Simple geometric considerations as illustrated in Fig. 1(c) provide a direct relation between the periodicity of the interference pattern, λint = 2π/kint, and the SPP wavelength for a given wedge opening angle α:26, 29]. Here, neff = λL/λSPP is the effective index of refraction for SPP excitation and φ is, as marked in Fig. 1(b), the laser incidence angle with respect to the surface normal. For λL = 800 nm, reference data from [27, 28] yield λSPP = 784 nm so that, for the given geometry (structure A, φ = 65°), β = 2.60°. Under these conditions Eq. (1) yields a periodicity λint = 8.15 μm. Note that, in the special case of normal incidence (φ = 0°) Eq. (1) simplifies to λint = λSPP/ sin(α) since β = 0.
Experimental wavelength scans confirm the applicability of Eq. (1) to the present study and support the interpretation in terms of the observation of an SPP-SPP interference pattern. Figures 3(a)–3(c) show PEEM images of the wedge recorded at three different SPP wavelengths. The change in the periodicity of the SPP-SPP interference pattern as a function of λSPP is evident. The results of a Fourier analysis of the interference patterns for eight different SPP wavelengths are shown in Fig. 3(d). For comparison the corresponding results from Eq. (1) for the three wavelengths shown in Figs. 3(a)–3(c) are marked by arrows. The experimental data match the calculated values almost perfectly.
As a side note, it should be mentioned that, in the normal-incidence geometry or for the case of constant neff, λSPP can be calculated directly from the experimentally obtained value λint without any knowledge of the laser used for excitation.
For a pulsed excitation, as in the present study, the observed interference signal can be interpreted in terms of a two-SPP-pulse correlation signal. The increase of the edge-to-edge distance provides a continuous adjustment of the pulse-to-pulse delay time τ(x) from the bottom to the top of the wedge as sketched in Fig. 1(c). In the case of normal incidence excitation and for small angles α the delay time is given by:
The actual situation of the experiment presented in this work is more complex due to the grazing-incidence excitation geometry. The excitation of the SPP pulse at the left hand edge of the wedge is delayed with respect to the excitation of the SPP pulse at the right hand edge by the propagation time of the laser pulse across the wedge. This contribution results in a modified expression for the time delay τ(x):
Note that for the given geometry the maximum error of the used approximation of a small wedge opening angle α is rather small (0.23%). Particularly, it is one order of magnitude smaller than the error in determining the dimensions of the wedge structure. The exact term for arbitrary angles α is derived in the supplementary information.
Contrary to the case of a conventional optical autocorrelator, it is in general necessary to distinguish between group delay τg(x) and phase delay τph(x) due to the dispersive character of SPP propagation. Correspondingly the group velocity vg and the phase velocity vph have to be used in Eq. (2) and Eq. (3). At a center wavelength of λSPP = 784 nm, vg = 0.938 c and vph = 0.979 c, respectively, yielding τg(x) = 0.327 · x fs/μm and τph(x) = 0.320 · x fs/μm for autocorrelation measurements at structure B. In the wavelength regime probed in the present study the difference in τg(x) and τph(x) is obviously rather small. For convenience we will in the following refer all data to τg(x).
The correlation signal as a function of time delay contains information on the temporal profile of the SPP pulses excited at the two excitation edges. Figure 4(a) shows PEEM data of the interference pattern recorded at illumination of the wedge with transform-limited 46 fs laser pulses at a center wavelength of 800 nm. The corresponding intensity profiles are additionally displayed as blue line in Fig. 4(c). As expected, the signal amplitude decreases as the time delay increases from the bottom to the top of the wedge, i.e. as the temporal overlap of the two SPP pulses decreases (see also Fig. 1). This behavior is qualitatively well reproduced by pattern simulations based on an analytic model as described in the supplementary information and shown in Fig. 4(b). For direct comparison with the experiment, intensity profiles deduced from these data are added as red line to Fig. 4(c). Note that the obvious distortions in the experimental data in particular visible at time delays < 10 fs arise from imperfections of the coupling edges. Figures 4(d)–4(f) show corresponding results recorded/simulated now at illumination with transform-limited 15 fs laser pulses at 800 nm center wavelength. The pulse duration effect is directly visible in the 2P-PEEM images. The amplitude of the interference signal decays significantly faster in the short pulse case indicative for the generation of corresponding short and broadband SPP pulses at the excitation edges. Once again the agreement with the simulations shown in Fig. 4(f) is rather satisfying. The simulations also show that the background signal observed for large delays (≈50% of the maximum signal amplitude) arises from the individual (non-interfering) SPP pulses. These data prove that the presented wedge device acts indeed as an autocorrelator for the characterization of ultrashort SPP pulses with the time delay adjusted (in an interferometric manner) by the wedge and the autocorrelation signal read out right in the exit area of the interferometer.
So far, no statement has been made regarding the order of the detected autocorrelation signal. As the signal is probed in nonlinear (two-photon) photoemission it seems at first glance likely that the autocorrelation signal is of second order, therefore providing information about the temporal profile and the group velocity dispersion (GVD) of the SPP. However, as the excitation laser is illuminating the entire detection area, one has to discriminate between two different possible excitation processes. Two-photon photoemission can either be induced by the subsequent absorption of two plasmon quanta of the SPP-SPP superposition field (yielding a second order autocorrelation) or, alternatively, by absorption of a plasmon quantum in the first excitation step and a photon of the laser field in the second excitation step. In the latter case the actually probed SPP-SPP interference signal would correspond to a first order autocorrelation, which provides information about the spectrum and the temporal profile of a transform-limited SPP pulse.
In the following a pump-probe laser experiment on the wedge structure A using a Mach-Zehnder interferometer is presented to determine the order of the autocorrelation. The pump pulse excites SPP pulses in the same manner as described before. These SPP pulses propagate along the gold-vacuum interface and are probed by the second laser pulse, which is incident with a time delay δ. Moreover, it probes the SPP pulse pair generated by the pump pulse. The setup corresponds to the one described in detail by Lemke et al. in [10, 25] without the active stabilization of the Mach-Zehnder interferometer.
Figures 5(a) and 5(b) compare interference patterns over an extended area next to the wedge for laser time delays of δ = 0 fs and δ = 200 fs and recorded with 48 fs laser pulses at a center wavelength of λL = 800 nm. The signal for δ = 0 fs, shown in Fig. 5(a), essentially corresponds to what is observed for the single pulse excitation experiment shown in Fig. 4(a) exhibiting the SPP autocorrelation pattern next to the wedge. At δ = 200 fs, however, an additional interference pattern is observed, located at a distance of ≈ 30 μm from the wedge structure (Fig. 5(b)). This signal can be ascribed to the interference pattern formed by the two SPP pulses, which are propagating along the gold-vacuum interface at the SPP group velocity. It is obviously probed by the second laser pulse arriving at a delayed time at the surface.
These data particularly provide evidence that in the present study the laser pulse is an indispensable part in the two-photon photoemission process and that, in conclusion, the measured SPP-SPP autocorrelation signal is of first order. For comparison, Fig. 5(c) (δ = 0 fs) and 5(d) (δ = 200 fs) show the same sample area now excited and monitored with 14 fs laser pulses at λL = 793 nm. The overall qualitative behavior is very similar although with some differences. Note, that the lateral extension of the propagating interference pattern appears narrower than for 48 fs excitation. This can be understood if one considers that here essentially a cross-correlation between laser field and SPP-SPP interference pattern is recorded. Note that these results additionally demonstrate that the tracking of SPP pulses in space and time does not necessarily require sub-diffraction spatial resolution and an active stabilized Mach-Zehnder interferometer.
We would like to add that further experiments performed in the single pulsed excitation scheme and using chirped laser pulses confirm our interpretation in terms of a first order autocorrelation.
The comparison between experimental and simulated first order autocorrelation traces as shown in Fig. 4 can provide approximate information on the bandwidth and the temporal profile of the probed SPP pulses. In principle such information can also be directly deduced from a Fourier analysis of the experimental data. In the present study the quantitative information from such an analysis is, however, of rather limited value due to the restricted time delay range that can be probed with the wedge interferometers (see particularly Fig. 4). An extension of the time delay range can, however, be achieved by a lateral extension of the wedge or an increase of the used wedge angle α. The qualitatively convincing agreement between experiment and simulation provide strong indication that the full bandwidth of the laser pulse is transferred to the SPP at excitation. For a gold-vacuum interface and λSPP = 780 nm dispersion effects should barely affect the temporal profile of the SPP pulse on the short propagation distances relevant in the presented study. We therefore expect the SPP pulse-lengths to be close to the values of the excitation laser pulse.
In summary we demonstrated experimentally a plasmonic autocorrelator for SPP pulse characterization based on a wedge shaped structure acting as an interferometer. The 2P-PEEM studies revealed that a first order autocorrelation was detected, which could successfully be modeled within an analytical framework. Autocorrelation measurements for different SPP pulse lengths clearly have proven the sensitivity of the technique to SPP pulse characteristics. As a noteworthy side aspect, our data showed that PEEM is capable of mapping SPP field amplitudes in a direct manner, which is particularly helpful for comparison with other SPP imaging techniques. Past PEEM-studies provided in this context only indirect information as exclusively superposition fields between excitation laser and SPP were recorded [9, 10, 29, 30].
Finally, both, the presented approach of the plasmon autocorrelator as well as the pump-probe measurements may be of general interest in the field of ultrafast plasmonics because it is in principle applicable to other SPP sensitive detection schemes which provide moderate spatial resolution.
This work was funded by the German Research Foundation (DFG) through Priority Program 1391 “Ultrafast Nanooptics” as well as by the Danish Council for Independent Research (FTP project “ANAP”, Contract No. 09-072949).
References and links
1. D. K. Gramotnev and S. I. Bozhevolnyi, “Plasmonics beyond the diffraction limit,” Nat. Photonics 4, 83–91 (2010). [CrossRef]
2. K. F. Macdonald, Z. L. Samson, M. I. Stockman, and N. I. Zheludev, “Ultrafast active plasmonics,” Nat. Photonics 3, 55–58 (2009). [CrossRef]
3. M. Kauranen and A. V. Zayats, “Nonlinear plasmonics,” Nat. Photonics 6, 737–748 (2012). [CrossRef]
4. V. V. Temnov, “Ultrafast acousto-magneto-plasmonics,” Nat. Photonics 6, 728–736 (2012). [CrossRef]
5. N. Rotenberg, M. Spasenović, T. L. Krijger, B. le Feber, F. J. García de Abajo, and L. Kuipers, “Plasmon scattering from single subwavelength holes,” Phys. Rev. Lett. 108, 127402 (2012). [CrossRef]
6. M. Sandtke, R. J. P. Engelen, H. Schoenmaker, I. Attema, H. Dekker, I. Cerjak, J. P. Korterik, F. B. Segerink, and L. Kuipers, “Novel instrument for surface plasmon polariton tracking in space and time.” Rev. Sci. Instrum. 79, 013704 (2008). [CrossRef]
7. T. Hattori, A. Kubo, K. Oguri, H. Nakano, and H. T. Miyazaki, “Femtosecond laser-excited two-photon fluorescence microscopy of surface plasmon polariton,” Jpn. J. Appl. Phys. 51, 04DG03 (2012). [CrossRef]
8. P. Andrew and W. L. Barnes, “Energy transfer across a metal film mediated by surface plasmon polaritons.” Science (New York, N.Y.) 306, 1002–1005 (2004). [CrossRef]
10. C. Lemke, C. Schneider, T. Leißner, D. Bayer, J. W. Radke, A. Fischer, P. Melchior, A. B. Evlyukhin, B. N. Chichkov, C. Reinhardt, M. Bauer, and M. Aeschlimann, “Spatiotemporal characterization of spp pulse propagation in two-dimensional plasmonic focusing devices.” Nano Lett. 13, 1053–1058 (2013). [CrossRef]
11. F.-J. Meyer zu Heringdorf, L. Chelaru, S. Möllenbeck, D. Thien, and M. Horn-von Hoegen, “Femtosecond photoemission microscopy,” Surf. Sci. 601, 4700–4705 (2007). [CrossRef]
12. C. Rewitz, T. Keitzl, P. Tuchscherer, J.-S. Huang, P. Geisler, G. Razinskas, B. Hecht, and T. Brixner, “Ultrafast plasmon propagation in nanowires characterized by far-field spectral interferometry.” Nano Lett. 12, 45–9 (2012). [CrossRef]
14. C. Rulliere, ed., Femtosecond Laser Pulses: Principles and Experiments, 2nd ed. (Springer, 2004).
15. A. Drezet, A. Hohenau, A. L. Stepanov, H. Ditlbacher, B. Steinberger, F. R. Aussenegg, A. Leitner, and J. R. Krenn, “Surface plasmon polariton machzehnder interferometer and oscillation fringes,” Plasmonics 1, 141–145 (2006). [CrossRef]
16. S. I. Bozhevolnyi, V. S. Volkov, E. Devaux, J.-Y. Laluet, and T. W. Ebbesen, “Channel plasmon subwavelength waveguide components including interferometers and ring resonators,” Nature 440, 508–511 (2006). [CrossRef]
17. A. Brun, P. Georges, G. L. Saux, and F. Salin, “Single-shot characterization of ultrashort light pulses,” J. Phys. D: Appl. Phys. 24, 1225–1233 (1991). [CrossRef]
18. V. V. Temnov, K. Nelson, G. Armelles, A. Cebollada, T. Thomay, A. Leitenstorfer, and R. Bratschitsch, “Femtosecond surface plasmon interferometry.” Opt. Express 17, 8423–8432 (2009). [CrossRef]
19. W. Swiech, G. Fecher, C. Ziethen, O. Schmidt, G. Schönhense, K. Grzelakowski, C. M. Schneider, R. Fromter, H. Oepen, and J. Kirschner, “Recent progress in photoemission microscopy with emphasis on chemical and magnetic sensitivity,” J. Electron Spectrosc. Relat. Phenom. 84, 171–188 (1997). [CrossRef]
20. O. Schmidt, M. Bauer, C. Wiemann, R. Porath, M. Scharte, O. Andreyev, G. Schönhense, and M. Aeschlimann, “Time-resolved two photon photoemission electron microscopy,” Appl. Phys. B. 74, 223–227 (2002). [CrossRef]
21. M. Cinchetti, A. Gloskovskii, S. Nepjiko, G. Schönhense, H. Rochholz, and M. Kreiter, “Photoemission electron microscopy as a tool for the investigation of optical near fields,” Phys. Rev. Lett. 95, 47601 (2005). [CrossRef]
22. A. Kubo, K. Onda, H. Petek, Z. Sun, Y. S. Jung, and H. K. Kim, “Femtosecond imaging of surface plasmon dynamics in a nanostructured silver film.” Nano Lett. 5, 1123–1127 (2005). [CrossRef]
23. L. Douillard, F. Charra, Z. Korczak, R. Bachelot, S. Kostcheev, G. Lerondel, P.-M. Adam, and P. Royer, “Short range plasmon resonators probed by photoemission electron microscopy,” Nano Lett. 8, 935–940 (2008). [CrossRef]
24. R. C. Word, J. P. S. Fitzgerald, and R. Konenkamp, “Direct imaging of optical diffraction in photoemission electron microscopy,” Appl. Phys. Lett. 103, 021118 (2013). [CrossRef]
25. C. Lemke, T. Leißner, S. Jauernik, A. Klick, J. Fiutowski, J. Kjelstrup-Hansen, H.-G. Rubahn, and M. Bauer, “Mapping surface plasmon polariton propagation via counter-propagating light pulses,” Opt. Express 20, 12877 (2012). [CrossRef]
26. T. Leißner, K. Thilsing-Hansen, C. Lemke, S. Jauernik, J. Kjelstrup-Hansen, M. Bauer, and H.-G. Rubahn, “Surface plasmon polariton emission prompted by organic nanofibers on thin gold films,” Plasmonics 7, 253–260 (2012). [CrossRef]
27. R. Olmon, B. Slovick, T. Johnson, D. Shelton, S.-H. Oh, G. Boreman, and M. Raschke, “Optical dielectric function of gold,” Phys. Rev. B 86, 235147 (2012). [CrossRef]
28. P. B. Johnson and R. W. Christy, “Optical constants of the noble metals,” Phys. Rev. B 6, 4370–4379 (1972). [CrossRef]
29. N. Buckanie, P. Kirschbaum, S. Sindermann, and F. Meyer zu Heringdorf, “Interaction of light and surface plasmon polaritons in ag islands studied by nonlinear photoemission microscopy,” Ultramicroscopy 130, 1–6 (2013). [CrossRef]
30. T. Leißner, C. Lemke, J. Fiutowski, J. W. Radke, A. Klick, L. Tavares, J. Kjelstrup-Hansen, H.-G. Rubahn, and M. Bauer, “Morphological tuning of the plasmon dispersion relation in dielectric-loaded nanofiber waveguides,” Phys. Rev. Lett. 111, 46802 (2013). [CrossRef]