Review Cite This: Chem. Rev. XXXX, XXX, XXX-XXX
pubs.acs.org/CR
Structural Engineering in Plasmon Nanolasers Danqing Wang,† Weijia Wang,† Michael P. Knudson,‡ George C. Schatz,†,§ and Teri W. Odom*,†,‡,§ †
Graduate Program in Applied Physics, ‡Department of Materials Science and Engineering, and §Department of Chemistry, Northwestern University, Evanston, Illinois 60208, United States ABSTRACT: This review focuses on structural engineering of lasers from the macroscale to the nanoscale, with an emphasis on plasmon nanolasers. Conventional lasers based on Fabry−Pérot cavities are limited in device size. In contrast, plasmon nanolasers can overcome the diffraction limit of light and incorporate unique structural designs to engineer cavity geometries and optical band structure. Since the spaser concept was introduced in 2003, tremendous progress in nanolasing has been made on architectures that exploit metal films and nanoparticles. Theoretical approaches in both frequency and time domains have inspired the development of plasmon nanolasers based on mode analysis and time-dependent lasing buildup. Plasmon nanolasers designed by band-structure engineering open prospects for manipulation of lasing characteristics such as directional emission, real-time tunable wavelengths, and controlled multimode lasing.
CONTENTS 1. Introduction 1.1. Overview 1.2. Cavity Designs in Macroscale Lasers 1.3. Gain Considerations and Lasing Characteristics of Macroscale Lasers 2. Small Lasers 2.1. Reducing Device Sizes in Conventional Lasers 2.2. Structural Engineering in Photonic Lasers 3. Plasmon Nanolasers 3.1. Introduction to Plasmon Nanolasers 3.2. Different Plasmon Nanolaser Systems 3.2.1. Single-Particle Spasers 3.2.2. Nanowire-on-Film Lasers 3.2.3. Plasmonic Nanoarrays 4. Lasing from Spaser Nanoparticle Arrays 4.1. Characteristics of Metal Nanoparticle Arrays 4.2. Lasing from Spaser Nanoparticle Arrays 4.2.1. Directional Lasing at Room Temperature 4.2.2. Real-Time Tunable Lasing 4.2.3. Amplified Spontaneous Emission in Spaser Nanoparticle Arrays 4.3. Theoretical Modeling of Plasmon Lasing 4.4. Recent Progress 4.4.1. Lasing from Dark Modes 4.4.2. Multimodal Lasing in Plasmonic Superlattices 5. Challenges in Plasmon Nanolasers 6. Summary and Outlook Author Information Corresponding Author ORCID Notes © XXXX American Chemical Society
Biographies Acknowledgments References
A A B
M N N
1. INTRODUCTION
C C
1.1. Overview
Lasers are coherent light sources that are driving discoveries in ultrafast spectroscopy and sub-femtosecond chemical reactions1−3 and are widely used in optical communications,4 biomedical imaging,5−7 and laser printing.8 Small lasersthose with sizes close to half the lasing wavelength, or the diffraction limitcan be widely applied in nanoscale applications. For example, they can serve as coherent light sources in on-chip electro-photonic circuits9,10 and be implanted in tissues for in situ cellular imaging.11−13 Shrinking coherent light sources to subwavelength scales can also provide new insight into light− matter interactions such as fluorescence, photocatalysis, quantum optics, and nonlinear optical processes.6,14,15 This review focuses on small-laser designs that rely on structural engineering from macroscale to nanoscale dimensions, with an emphasis on plasmon nanolasers. Advances in nanofabrication have enabled unconventional architectures for plasmon nanolasers different from traditional Fabry−Pérot mode cavities. Intense, localized electric fields at plasmonic hotspots can overcome the diffraction limit of light and enable ultrafast operation at terahertz frequencies.16−18 Since the concept of the spaser (surface plasmon amplification by stimulated emission of radiation)19 was proposed, plasmon nanolasers based on either surface plasmon polariton (SPP)
C D E E E E E G H H I I J J J L L L M M M M M M
Special Issue: Plasmonics in Chemistry Received: July 14, 2017
A
DOI: 10.1021/acs.chemrev.7b00424 Chem. Rev. XXXX, XXX, XXX−XXX
Chemical Reviews
Review
Figure 1. Development timeline of plasmon nanolasers. (a) The first plasmon laser using SPP modes as optical feedback for lasing at cryogenic temperatures, including CdS semiconducting nanowires on a silver film separated by a MgF2 layer. (b) Plasmon laser operated at room temperature with CdS nanosquares on a Ag film separated by a MgF2 layer. (c) Plasmon nanolaser using an epitaxially grown silver film for a lower lasing threshold. (d) The first theoretical study of spaser arrays for directional emission using split-ring resonators at mid-infrared wavelengths. (e) The first experimental realization of directional emission from spaser NP arrays at near-infrared wavelengths. (f) Plasmonic superlattice arrays with controlled multimodal nanolasing. (g) Timeline of plasmon nanolasers. Panel a was adapted from ref 17 with permission. Copyright 2009 Nature Publishing. Panel b was adapted from ref 20 with permission. Copyright 2011 Nature Publishing. Panel c was adapted from ref 24 with permission. Copyright 2012 American Association for the Advancement of Science. Panel d was adapted from ref 23 with permission. Copyright 2008 Nature Publishing. Panel e was adapted from ref 16 with permission. Copyright 2013 Nature Publishing. Panel f was adapted from ref 27 with permission. Copyright 2017 Nature Publishing.
modes on metal films17,20−22 or localized surface plasmon (LSP) modes from metal nanoparticles (NPs)16,23 have been reported (Figure 1). Although the first theoretical paper on spasers19 focused on plasmon amplification of LSPs in single metal NPs, semiconducting nanowires on metal films (Figure 1a) were one of the first designs realized using SPP modes as optical feedback for lasing at cryogenic temperatures.17 Roomtemperature operation was later achieved using nanosquares (Figure 1b) as resonators.20 Additionally, a larger contact area at the nanowire−insulator interface21 and integration of epitaxially grown metal films with atomic smoothness24,25 decreased the scattering losses and further improved SPP lasers by lowering lasing thresholds (Figure 1c). Nearly 10 years ago, the concept of using an array structure to couple subunits for directional lasing emission was proposed for split-ring resonators23 at mid-infrared wavelengths (Figure 1d). However, experimental realization16 was only recently achieved and with a simpler system (Figure 1e). Periodic arrays of metal NPs coupled with organic gain molecules showed room-temperature operation, spatial coherence, directional emission normal to the surface, and tunability over nearinfrared wavelengths.16,26 Advances in theoretical modeling by considering cavity modes and time-dependent lasing action have contributed to the understanding and design of emerging plasmon lasers, including switchable, multimodal nanolasers27 (Figure 1f) and dark-mode lasers.28,29 This review will focus on plasmon nanolasers from a structural engineering approach.
Figure 2. A conventional laser supporting Fabry−Pérot modes between two end mirrors (length L) and gain medium inside the cavity. Adapted from ref 6 with permission. Copyright 2014 Nature Publishing.
gain media in the cavity and are amplified to build up stimulated emission with spatial and temporal coherence.30 For a Fabry−Pérot cavity of length L and mirror reflectivities R1 and R2, the electric field amplitude of the light returning to its original value after a round trip forms a standing wave that can be expressed as6 E0 R1R 2 e(G − α)Le i4πnL / λ0 = E0
(1)
where G is the optical gain for a specific cavity mode, α is the absorption loss, λ0 is the resonance wavelength in free space, and n is the effective index of the medium. Therefore, 4πnL/λ0 must be a multiple of 2π in eq 1, leading to
1.2. Cavity Designs in Macroscale Lasers
Traditional lasers are composed of gain media sandwiched between parallel mirrors that support Fabry−Pérot modes within the optical cavity (Figure 2). Photons generated by optical or electrical pumping travel back and forth through the
L= B
λ0 m 2n
(2) DOI: 10.1021/acs.chemrev.7b00424 Chem. Rev. XXXX, XXX, XXX−XXX
Chemical Reviews
Review
where m is an integer indexing the lasing mode. The shortest possible cavity length for traditional lasers is λ0 (m = 1), which 2n represents the diffraction limit. Additionally, sufficient gain is needed to overcome the absorption losses (α) and mirror losses (R1, R2 < 1) in one round trip, resulting in L=
−ln(R1R 2) 2(G − α)
must be optically pumped; however, they can achieve a similar range of tunability based on broad absorption and emission spectra.47,48 In the miniaturization of lasers, both the physical size of the cavity and the gain needed to overcome losses are critical, and these factors will be considered in the subsequent sections.
2. SMALL LASERS
(3)
2.1. Reducing Device Sizes in Conventional Lasers
Therefore, the cavity sizes of conventional lasers need to be large enough to provide sufficient gain to overcome the losses. On the basis of propagation and reflection losses of the cavity, the quality factor (Q) can indicate the residence time of photons. The Purcell factor F is typically used to quantify the enhanced spontaneous emission rate in dielectric resonators and it influences the buildup dynamics of population inversion for lasing.31 A simplified expression for F is30 F=
3 3 ⎛⎜ Q ⎞⎟⎜⎛ λ ⎟⎞ 4π 2 ⎝ V ⎠⎝ 2n ⎠
Advances in nanofabrication have facilitated the miniaturization of conventional macroscale lasers by decreasing internal cavity losses and increasing interactions between gain and cavity. Although overall device sizes can be reduced, the diffraction limit still sets the minimum sizes of dielectric cavities. Epitaxial growth methods can create distributed Bragg reflectors (DBRs) consisting of stacks of alternating layers of high and low refractive index materials.51 Traditional edge emitting lasers show large beam divergences and high thresholds.52 Verticalcavity surface-emitting lasers (VCSELs) exploit DBRs with very high reflectivities (>99%) to reduce internal losses and hence the amount of gain required in the system53 to produce a spatially confined beam normal to the surface with high optical power (Figure 3a). These devices are several microns tall because of the multilayer DBR structure.
(4)
where Q depends on line width and mode volume V defines electromagnetic field confinement. Unlike the process of spontaneous emission, where photons are emitted with random direction and phase, stimulated emission is the result of coherent amplification of photons by interaction with the excited states of gain media.30 The critical characteristics of lasing include spatial and temporal coherence, where emitted photons from the lasing cavity share the same wavelength, phase, amplitude, polarization, and directionality.30 As generally agreed upon in the community,32 criteria for lasing include (1) a deterministic lasing threshold in the input power−output intensity curves, (2) line width narrowing [full width at half-maximum (fwhm) < 1−2 nm)] above the threshold, (3) a nonlinear increase in the output intensity above the threshold, and (4) minor divergence of the emission beam. Other competing phenomena, such as amplified spontaneous emission (ASE), need to be eliminated to verify lasing action, especially in unconventional nanolasers.32,33 1.3. Gain Considerations and Lasing Characteristics of Macroscale Lasers
Figure 3. Scheme of small lasers including (a) VSCEL, (b) microdisk, and (c) nonplasmonic metal cavity lasers. Panel a was adapted from ref 53 with permission. Copyright 1989 AIP Publishing. Panel b was adapted from ref 14 with permission. Copyright 2014 American Chemical Society. Panel c was adapted from ref 55 with permission. Copyright 2007 Nature Publishing.
Fabry−Pérot cavities are standard for a range of lasers with different gain material, including solid-state, gas, semiconductor, and dye media.34−49 Gain material affects wavelength, tunable range, pumping mechanisms, and time scales of lasing dynamics.6,50 The earliest lasers were based on insulators doped with ions: Cr:Al2O3 (ruby),34 Ti:Al2O3 (Ti:sapphire),35 and Nd:Y3Al5O12 (Nd:YAG).36,37 Advantages of solid-state lasers include tunable emission over hundreds of nanometers35 and generation of ultrashort pulses for femtosecond measurements.38,39 The development of other types of gain materials produced different lasing signatures. Gas and semiconductor lasers can be electrically pumped, and so they are frequently used to optically pump other laser systems that cannot be driven electrically.40−42,44,45 Gas lasers rely on electrical discharge40−42 and have pulse durations
