A compact and versatile source of coherent surface-plasmon polaritions (SPPs) is demonstrated by end-coupling a laser diode operating at 1.46 μm to a plasmonic waveguide integrated on the same microchip. With an optimized overlap between the spatial-modes of the laser and a planar-stripe waveguide, a high coupling efficiency of ~36% is achieved, that computations show could approach ~60% with smaller, readily achievable gaps between laser and waveguide. This integrated and electrically-activated source, with an available SPP power limited only by the laser diode, appears ideally suited for directly driving plasmonic circuitry or surface-enhanced sensors.
©2010 Optical Society of America
The emerging field of plasmonics promises applications from sensitive surface-enhanced sensors  to nanoscale “optical circuitry” having dimensions far below the diffraction limit . Key to many applications is the surface plasmon polariton (SPP), an optical mode that propagates along a conductor’s surface through a hybridization of free-space fields and conduction currents. While many plasmonic circuit elements [3,4] and waveguide configurations [5,6] have been constructed or proposed, applications will be hampered until the external optics now used to pump these elements can be replaced with an on-chip electrical source of SPPs, the subject of this report.
To efficiently excite SPPs, one must overcome the well-known momentum mismatch between a photon and the corresponding shorter-wavelength SPP oscillating at the same frequency . Most often this is accomplished by phase matching an external light source to the SPPs by beaming radiation at a specific angle toward either a grating impressed on the metal surface [8,9] or a prism contacting an optically thin metal film . An alternative technique places photo-  or electrically-excited luminescent molecules [12,13] or nanocrystals  tens of nanometers from the metal surface, allowing some of the luminescence to quench into SPPs through momentum components in the near field . While this approach recently achieved electrical SPP-generation in both an organic light-emitting diode electrode  and a Si-based metal-insulator-metal waveguide , it provides only a relatively weak source of incoherent SPPs. However, because this approach also underlies the “spaser,” wherein gain in the coupled fluorophore-plasmon mode is achieved [16,17], it may lead to coherent plasmon sources in the future. A third method is end-coupling, which is analogous to splicing two waveguides of different modal index where the resulting momentum mismatch produces a reflection loss. To date, this approach has been exemplified by abutting an optical fiber against a collinearly aligned plasmonic waveguide [3,18].
Here we exploit end-coupling between a laser diode and an integrated plasmonic waveguide to introduce an electrically driven and coherent source of SPPs with an output power and wavelength limited only by the diode-laser characteristics. While a variety of waveguides is possible, for the present demonstration we choose a simple Au-stripe waveguide  for its ease of fabrication and high coupling efficiency. As illustrated in Fig. 1(a) , an on-chip laser diode with an etched output facet faces an optically thick (100 nm) Au film that serves as a multi-lateral-mode plasmonic waveguide driven by the laser radiation. The modal index  of SPPs propagating on such a film closely matches that of free space, so reflectance losses are only ≈1%. The coupling efficiency ηc is instead limited by mode-shape mismatch and beam divergence as radiation propagates over the gap separating the laser facet and the metallic waveguide. The efficiencies measured in this work (see Section 3) are consistent with ηc = 36% as computed from the simulation in Fig. 1(b). Whereas this pertains to the smallest gap examined, 4.4 μm, constrained by our use of optical lithography, the computed coupling efficiency would approach 60% if electron beam lithography were used to reduce the gap to <1 μm. Combining such high efficiency with a high-power laser diode could allow a single SPP source to power multiple plasmonic circuits integrated on a single chip , through the use of power splitters  or couplers .
2. Plasmonic-source design and characterization
Our plasmonic source is pumped by a broad-area ridge laser emitting at 1.46 μm. Two InGaAs/InGaAsP active quantum-wells were grown on InP under tensile strain  to produce transverse-magnetic (TM) polarized radiation as required to match the inherently TM-polarized SPP modes at the planar metal surface. To further optimize the coupling, we positioned the 100-nm thick plasmonic guide h = 1.3 μm below the laser quantum wells [Fig. 1(a)], a distance controlled by the dry-etch depth that defines the laser ridge. This depth was designed to best match the approximately-Gaussian spatial profile of the vertically diverging laser-radiation (computed with Fraunhofer diffraction theory and verified experimentally below) to the exponential profile of the SPP mode above the Au surface. A wide 150-μm ridge was chosen for these first devices to simplify fabrication. Figure 1(c) presents a SEM micrograph of the integrated structure. Figure 1(d) shows an infrared image of the operating plasmon source, where laser radiation is seen to scatter from the gap between the etched facet and the guide while, more importantly, radiation is also observed emanating from the opposite end of the guide where SPPs convert back to free-space radiation with a computed efficiency of ~99%. Although in present devices we have cleaved the plasmonic guide short to aid characterization, future circuits could use the intense electrically-generated SPPs to perform a variety of optoelectronic functions.
That the emission from the terminus of the waveguide originates from SPPs is convincingly established by its angular distribution, described below, and by its strong TM polarization even when the guide is excited by poorly polarized electroluminescence (EL) from the laser operating below threshold. The unpolarized nature of the EL is demonstrated by the similar intensities in the micrographs of Figs. 2(a) and 2(b) for the cleaved rear laser facet when operating at 70 mA, far below the 1.4 A threshold. The views are end-on through a polarization analyzer set for polarizations normal [TM, Fig. 2(a)] and parallel [TE, Fig. 2(b)] to the waveguide surface. In contrast, the corresponding images for radiation emanating from the end of the waveguide show strong emission only for TM polarization, Fig. 2(c), while no emission is observable for TE polarization, Fig. 2(d), as expected for SPPs. A more quantitative analysis is presented in Fig. 2(e), which plots the dependence of image intensity on analyzer angle ϕ (referenced to the TM axis). Here squares represent the polarization dependence of EL from the rear laser facet (normalized to 1.1 for clarity), and circles correspond to the ratio Pg(ϕ)/Pr(ϕ), normalized to unity, where Pg and Pr are, respectively, image intensities of the guide and rear facet. The emission from the waveguide follows a cos2(ϕ)dependence (solid-line fit), as expected when the TM-polarized SPPs convert to photons at the guide termination.
To further characterize the SPP radiation, the far-field angular distributions of radiation from the driving facet and the Au waveguide were simulated, Fig. 3(a) , and compared to measurements, Fig. 3(b). The electromagnetic simulations were conducted with a vectorial two-dimensional finite-element electromagnetic solver  of Maxwell’s equations treating a geometry dimensioned via scanning electron micrographs. The diffraction in the lateral (across the ridge) dimension can be ignored because of the large ridge width. The measured distributions were obtained from images of powered devices formed with an objective restricted to an angular acceptance of 3.5°. The resulting large depth-of-field permitted simultaneous imaging of both the end of the guide and the etched facet of the driving laser. By measuring the integrated intensities of the CCD images as the sample was rotated, angular distributions such as in Fig. 3(b) were mapped. Both the shape and width of the measured quasi-Lorentzian distribution emitted from the waveguide agree well with the predictions of Fraunhofer diffraction theory  when applied to a source with an exponential SPP-mode profile. The agreement [5.5° half-width at half-maximum (HWHM) measured; 5.7° predicted] independently confirms the origin as SPP radiation. The complex angular distribution from the driving laser facet agrees quite reasonably with the simulated distribution in width, angular displacement from the plane of the device (due to the reflection from the substrate beyond the facet), and the presence of a secondary lobe at higher angles. As a control, the distribution from the cleaved rear facet was also measured [Fig. 3(c)]. This much broader, single-lobe quasi-Gaussian distribution (20° HWHM) agrees well with that computed from Fraunhofer diffraction applied to the Gaussian transverse laser-mode perpendicular to the quantum well plane (21° predicted HWHM).
3. Laser-plasmon coupling efficiency
The generated SPP power depends on the coupling efficiency ηc between the laser output facet and the waveguide. While our integrated architecture precludes a direct determination, we can compute ηc through simulation and compare to values deduced from corrected measurements as explained in the Appendix. As stated above, simulations  based on a facet-to-waveguide gap of 4.4 μm give ηc = 0.36. The exact value depends on the tilt angle of the facet from perpendicular, which for an etched facet may vary with position. Fortunately, we find the dependence to be weak, with the theoretical ηc varying only from 0.36 for the nominal experimental tilt of 7° to 0.37 for no tilt. Our measured values for ηc range from 0.29 to 0.37, which bracket the preferred simulated efficiency of 0.36.
Using this efficiency, we infer that a peak SPP power of 36 mW is launched in the Au guide when 2.0 A of current is applied to the drive laser (see Appendix). The present devices were pulsed at very low duty cycle to avoid thermal damage (1.7 μs pulse duration, 1-kHz repetition rate), and no appreciable roll-over was observed at the maximum current. Comparable continuous-wave (cw) SPP powers should easily be attainable with improved thermal management, e.g., by epitaxial-side-down mounting or reducing the ridge width to ≤ 5 μm and electro-plating the top with a Au heat sink. A narrower ridge would also enable additional advantages, including (i) cw wall-plug plasmon-generation efficiencies up to ~10%; (ii) efficient coupling to narrower stripe waveguides supporting only one or a few modes [18,19]; (iii) improved coupling to subwavelength modes associated with metal-dielectric  or channel  waveguide architectures (the on-chip drive laser would then employ the more common TE polarization); and (iv) narrow spectral line widths if a distributed feedback Bragg-grating were incorporated in the ridge. These realistic methods for increasing the power and decreasing the mode volumes would also improve the prospects for driving plasmonic concentrators  and/or nonlinear interactions.
An alternative coupling scheme for an integrated surface-plasmon source could be realized by fabricating a grating-coupled waveguide  directly onto a cleaved laser-diode output facet or a vertical-cavity surface-emitting laser. While such plasmonic sources have yet to be demonstrated, the feasibility of this approach is suggested by the success of placing plasmonic antennas  and collimators  on laser output facets.
We have demonstrated a powerful approach to electrically generating surface plasmons. Such an on-chip source of coherent SPPs is a prerequisite for practical integrated plasmonic circuitry that will be used in nanophotonic and sensing applications. By end-coupling a laser diode with output polarization and optical mode matched to an integrated plasmonic waveguide, ~36% coupling efficiency and 36 mW of SPP peak power are produced at room temperature. Devices incorporating improvements to the laser, coupling, and thermal management are expected to generate tens of mW of cw plasmons, suitable for driving many plasmonic devices from a single source.
Appendix: Experimental details
The TM-polarized drive laser  was designed by NRL and grown by AdTech Optics using metal-organic chemical vapor deposition (MOCVD) on an n-InP substrate. The two 10-nm thick tensile-strained In0.39Ga0.61As active quantum wells were separated by a 15-nm In0.84Ga0.16As0.35P0.65 barrier and sandwiched between 200-nm thick separate-confining heterostructures of In0.84Ga0.16As0.35P0.65. Optical cladding was provided by an n-InP buffer layer (1-μm thick) grown on the substrate and a p-InP layer (1.7-μm thick) above, capped by a degenerately doped p-In0.53Ga0.47As contact layer (200-nm thick). After deposition of a Ti/Pt/Au wafer-wide bottom contact, standard optical lithography and reactive ion etching defined 150-μm-wide ridge lasers with an etched output facet and corrugated sidewalls to spoil totally-internally-reflected optical modes. The dry etch continued to a depth 1.3 μm below the quantum wells to place the wells a height h, Fig. 1(a), above the subsequently deposited Au waveguide. This height was computed to optimize the laser and SPP spatial-mode overlap for best photon-to-SPP coupling. A 100-μm-wide Ti/Pt/Au top-contact was then deposited to within 14 μm of the etched output facet. In the final deposition, electron beam evaporation was used to deposit the 100-nm-thick Au plasmonic waveguide 150 μm wide atop a 6 nm Ti adhesion layer. Lasers with 2 mm cavity lengths were then defined by cleaving the chip 2 mm from the etched output facet. A final cleave through the Au waveguide yielded guides about 300 μm long with the terminal end exposed for diagnostic microscopy. The finished drive lasers were operated at room temperature, typically with 1.7 μs pulses at a 1 kHz repetition-rate to minimize thermal effects. The lasers emitted at a wavelength of 1.46 μm with a lasing threshold of ~1.4 A.
The laser and waveguide emissions were characterized by quantitative image-analyses of infrared micrographs obtained with a custom microscope equipped with a thermoelectrically cooled InGaAs charge-coupled device camera (Xenics, Inc.), a linear polarizer, and an objective with a measured numerical aperture (NA) of 0.32, giving a theoretical resolution of ~2.7 μm at 1.46 μm. Angular emission distributions were obtained with a separate objective that had its angular acceptance restricted by a 1-mm pinhole near its back focal plane to produce a measured 1.8° acceptance (HWHM).
The photon-to-plasmon coupling efficiency, ηc, was inferred from two approaches that circumvented the inability to directly access the full emission from the etched driving facet. In the first approach, an efficiency of 0.37 was determined from ηc=(Pgtg−1/Pr)(P′r/P′f)(1- f′sub). Here Pg and Pr are, respectively, the emissions from the Au waveguide and rear cleaved facet, which serves as a reference. These emissions were measured either from micrographs as in Fig. 2(c), or from angular distributions as in Figs. 3(b) and (c), which agree to better than 3%. The factor tg−1 corrects for the SPP propagation loss in the guide as described below. The remaining terms account for the smaller power emitted from the cleaved rear reference-facet relative to the front etched face. This ratio was determined with the aid of an ancillary device from which the Au guide had been cleaved to make the etched facet more accessible. P′r and P′f are the powers measured from the rear and front ends of the ancillary device as measured by a 1-cm-square Ge photodiode that intercepted the full angular emission widths. The ratio P′r/P′f does not fully account for the difference in emitted powers because a fraction f′sub of the emission from the etched facet was absorbed by 4.4 μm of substrate that remained in front of the facet after the guide was cleaved off. For a 7° tilted facet, simulation yields f′sub=0.4, the veracity of which is supported by good agreement between the simulated and measured angular emission distributions from the front of the ancillary device (not shown), which was influenced by substrate reflection and minor diffraction effects.
The second approach, requiring only an intact device, determined ηc=0.29 from ηc=[Pgtg−1/(Pgtg−1 + Pe)] (1-fsub). Here Pg and Pe are the integrated areas under the angular distributions emitted by the guide and etched facet, shown normalized to unity in Fig. 2b. The term (1-fsub) corrects for the computed fraction fsub of laser emission lost to the substrate [see Fig. 1(b)], which is computed from simulations to be 0.49 for a 7° tilted facet.
The peak SPP power of 36 mW was inferred by referencing to the average power emitted by the cleaved rear facet, as measured with a calibrated 1-cm square Ge photodiode positioned 3.2 mm from the facet to intercept the full angular width of the emission. This average reference power was then scaled to the higher power emitted from the etched facet (discussed above), adjusted for the duty factor (1.7 μs pulses every 1 ms), and corrected for the computed photon-to-plasmon coupling efficiency of 0.36.
The SPP transmittance along the guide, tg=0.46, is determined from tg=exp(-Lg/LSPP), where Lg is the length of the guide (276 μm) and LSPP is the SPP propagation length. A value LSPP ≈ 353 μm is computed using the standard expression  for a thick plane with optical constants (ε=-97.04+i6.12) measured on Au witness films with variable-angle spectroscopic ellipsometry. This value of LSPP is consistent with the exponential decay of SPP radiation scattered from holes drilled with a focused ion beam along the Au guide of a test device, although the limited attenuation of less than one decay length prevented a precise determination of LSPP with this technique .
This work was funded by the Office of Naval Research. The research was performed while R. A. Flynn held a National Research Council Research Associateship.
References and links
1. J. J. Baumberg, T. A. Kelf, Y. Sugawara, S. Cintra, M. E. Abdelsalam, P. N. Bartlett, and A. E. Russell, “Angle-resolved surface-enhanced Raman scattering on metallic nanostructured plasmonic crystals,” Nano Lett. 5(11), 2262–2267 (2005). [CrossRef] [PubMed]
2. H. A. Atwater, S. Maier, A. Polman, J. A. Dionne, and L. Sweatlock, ““The new ‘p-n junction’: plasmonics enables photonic access to the nanoworld,” MRS Bull. 30, 385 (2005). [CrossRef]
3. 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(7083), 508–511 (2006). [CrossRef] [PubMed]
4. S. Passinger, A. Seidel, C. Ohrt, C. Reinhardt, A. Stepanov, R. Kiyan, and B. N. Chichkov, “Novel efficient design of Y-splitter for surface plasmon polariton applications,” Opt. Express 16(19), 14369–14379 (2008). [CrossRef] [PubMed]
5. R. F. Oulton, G. Bartal, D. F. P. Pile, and X. Zhang, “Confinement and propagation characteristics of subwavelength plasmonic modes,” N. J. Phys. 10(10), 105018 (2008). [CrossRef]
6. R. F. Oulton, V. J. Sorger, D. A. Genov, D. F. P. Pile, and X. Zhang, “A hybrid plasmonic waveguide for subwavelength confinement and long-range propagation,” Nat. Photonics 2(8), 496–500 (2008). [CrossRef]
8. H. Ditlbacher, J. R. Krenn, A. Hohenau, A. Leitner, and F. R. Aussenegg, “Efficiency of local light-plasmon coupling,” Appl. Phys. Lett. 83(18), 3665 (2003). [CrossRef]
9. E. Verhagen, A. L. Tchebotareva, and A. Polman, “Erbium luminescence imaging of infrared surface plasmon polaritons,” Appl. Phys. Lett. 88(12), 121121 (2006). [CrossRef]
10. H. Raether, Surface Plasmons (Springer-Verlag, 1988).
12. P. A. Hobson, S. Wedge, J. A. E. Wasey, I. Sage, and W. L. Barnes, “Surface plasmon mediated emission from organic light-emitting diodes,” Adv. Mater. 14(19), 1393–1396 (2002). [CrossRef]
13. D. M. Koller, A. Hohenau, H. Ditlbacher, N. Galler, F. Reil, F. R. Aussenegg, A. Leitner, E. J. W. List, and J. R. Krenn, “Organic plasmon-emitting diode,” Nat. Photonics 2(11), 684–687 (2008). [CrossRef]
14. R. J. Walters, R. V. A. van Loon, I. Brunets, J. Schmitz, and A. Polman, “A silicon-based electrical source of surface plasmon polaritons,” Nat. Mater. 9(1), 21–25 (2010). [CrossRef]
15. G. W. Ford and W. H. Weber, “Electromagnetic interactions of molecules with metal surfaces,” Phys. Rep. 113(4), 195–287 (1984). [CrossRef]
16. M. T. Hill, M. Marell, E. S. P. Leong, B. Smalbrugge, Y. Zhu, M. Sun, P. J. van Veldhoven, E. J. Geluk, F. Karouta, Y.-S. Oei, R. Nötzel, C.-Z. Ning, and M. K. Smit, “Lasing in metal-insulator-metal sub-wavelength plasmonic waveguides,” Opt. Express 17(13), 11107–11112 (2009). [CrossRef] [PubMed]
18. A. Berini, R. Charbonneau, N. Lahoud, and G. Mattiussi, “Characterization of long-range surface-plasmon-polariton waveguides,” J. Appl. Phys. 98(4), 043109 (2005). [CrossRef]
19. R. Zia, A. Schuller, and M. L. Brongersma, “Near-field characterization of guided polariton propagation and cutoff in surface plasmon waveguides,” Phys. Rev. B 74(16), 165415 (2006). [CrossRef]
20. G. Jones, A. D. Smith, E. P. O’Reilly, M. Silver, A. T. R. Briggs, M. J. Fice, A. R. Adams, P. D. Greene, K. Scarrott, and A. Vranic, “The influence of tensile strain on differential gain and auger recombination in 1.5- um multiple-quantum-well lasers,” J. Quant. Electron. 34(5), 822–833 (1998). [CrossRef]
21. COMSOL, Inc., Burlington MA.
22. M. Born, and E. Wolf, Principles of Optics (Cambridge Univ. Press, 2002).
23. V. S. Volkov, S. I. Bozhevolnyi, S. G. Rodrigo, L. Martín-Moreno, F. J. García-Vidal, E. Devaux, and T. W. Ebbsen, “Nanofocusing with channel plasmon polaritons,” Nano Lett. 9(3), 1278–1282 (2009). [CrossRef] [PubMed]
24. E. Cubukcu, E. A. Kort, K. B. Crozier, and F. Capasso, “Plasmonic laser antenna,” Appl. Phys. Lett. 89(9), 093120 (2006). [CrossRef]
25. N. Yu, R. Blanchard, J. Fan, Q. J. Wang, C. Pflügl, L. Diehl, T. Edamura, M. Yamanishi, H. Kan, and F. Capasso, “Quantum cascade lasers with integrated plasmonic antenna-array collimators,” Opt. Express 16(24), 19447–19461 (2008). [CrossRef] [PubMed]
26. R. A. Flynn, K. Bussmann, B. S. Simpkins, I. Vurgaftman, C.-S. Kim, and J. P. Long, “Propagation length of surface plasmon polaritons determined by emission from introduced surface discontinuities,” J. Appl. Phys. 107(1), 013109 (2010). [CrossRef]