REVIEW ARTICLE

PUBLISHED ONLINE: 3 JUNE 2013 | DOI: 10.1038/NPHYS2615

Quantum plasmonics
M. S. Tame1 *, K. R. McEnery1,2 , Ş. K. Özdemir3 , J. Lee4 , S. A. Maier1 * and M. S. Kim2

Quantum plasmonics is a rapidly growing field of research that involves the study of the quantum properties of light and its
interaction with matter at the nanoscale. Here, surface plasmons—electromagnetic excitations coupled to electron charge
density waves on metal–dielectric interfaces or localized on metallic nanostructures—enable the confinement of light to scales
far below that of conventional optics. We review recent progress in the experimental and theoretical investigation of the
quantum properties of surface plasmons, their role in controlling light–matter interactions at the quantum level and potential
applications. Quantum plasmonics opens up a new frontier in the study of the fundamental physics of surface plasmons and
the realization of quantum-controlled devices, including single-photon sources, transistors and ultra-compact circuitry at the
nanoscale.

P
lasmonics provides a unique setting for the manipulation outlook on some of the important challenges that remain to be
of light via the confinement of the electromagnetic field addressed and new directions for the field.
to regions well below the diffraction limit1,2 . This has
opened up a wide range of applications based on extreme light Quantization
concentration3 , including nanophotonic lasers and amplifiers4,5 , One of the most fundamental aspects in quantum plasmonics is the
optical metamaterials6 , biochemical sensing7 and antennas trans- description of surface plasmons using quantum mechanics. This is
mitting and receiving light signals at the nanoscale8 . These appli- what sets it apart from all other areas of modern plasmonics. Much
cations and their rapid development have been made possible by of the work laying the foundations for quantization was carried out
the large array of experimental tools that have become available in in the 1950s by Bohm and Pines, with work by Pines providing
recent years for nanoscale fabrication and theory tools in the form the very first model for quantizing plasma waves in metals21 .
of powerful electromagnetic simulation methods. At the same time, Here, electrons in the conduction band were considered to be free
and completely parallel to this remarkable progress, there has been electrons in an electron gas and the long-range correlations in their
a growing excitement about the prospects for exploring quantum positions treated in terms of collective oscillations of the system as a
properties of surface plasmons and building plasmonic devices whole. The quantized form of these collective matter oscillations—
that operate faithfully at the quantum level9 . The hybrid nature plasmons—were found to be bosons, with both wave-like and
of surface plasmon polaritons (SPPs) as ‘quasi-particles’ makes particle-like behaviour, as expected for quantum excitations. The
them intriguing from a fundamental point of view, with many of ‘polariton’—a joint state of light and matter—was introduced by
their quantum properties still largely unknown. In addition, their Hopfield22 , who provided a quantum model for the polarization
potential for providing strong coupling of light to emitter systems, field describing the response of matter to light. Depending on
such as quantum dots10,11 and nitrogen–vacancy (NV) centres12 , via the type of matter, Hopfield called the field a ‘phonon-polariton’,
highly confined fields offers new opportunities for the quantum ‘plasmon-polariton’ and so on, with the quanta as bosons. The
control of light, enabling devices such as efficient single-photon concept of a surface plasma wave (SPW) was proposed soon
sources13–16 and transistors17–19 to be realized. Although surface after by Ritchie23 . Several years later, Elson and Ritchie24 , and
plasmons are well known to suffer from large losses, there are others used Hopfield’s approach to provide the first quantized
also attractive prospects for building devices that can exploit this description of SPWs as ‘SPPs’, whose coupled light–matter features
lossy nature for controlling dissipative quantum dynamics20 . This are described in Fig. 1. Hydrodynamic effects were also included in
new field of research combining modern plasmonics with quantum the quantization25 . Despite its great success, Hopfield’s approach
optics has become known as ‘quantum plasmonics’. did not consider loss, which is caused by the scattering of
In this Review, we describe the wide range of research activities electrons with background ions, phonons and themselves in the
being pursued in the field of quantum plasmonics. We begin with a conduction band8,26 (ohmic loss) and at high frequencies by
short description of SPPs and their quantization. Then, we discuss interband transitions26 . A new ‘microscopic’ quantization method
one of the major strengths of plasmonic systems: the ability to was introduced by Huttner and Barnett27 , extending Hopfield’s
provide highly confined electromagnetic fields. We describe how approach to polaritons in dispersive and lossy media, including
this enables the enhancement of light–matter interactions and the waveguides. Most recently, a ‘macroscopic’ approach has been
progress that has been made so far in demonstrating a variety of developed using Green’s functions28 . Localized surface plasma
schemes that take advantage of it in the quantum regime. We also oscillations at nanoparticles have also been quantized29–31 , the
review key experiments that have probed fundamental quantum quanta of which are called localized surface plasmons (LSPs).
properties of surface plasmons and their potential for building In Box 1, we outline a basic approach to quantization for the
compact nanophotonic circuitry. We conclude by providing an waveguide32,33 (SPP) and localized30,31 (LSP) setting.

1 ExperimentalSolid State Group, Department of Physics, Imperial College London, Prince Consort Road, SW7 2BW, UK, 2 Quantum Optics and Laser
Science Group, Department of Physics, Imperial College London, Prince Consort Road, SW7 2BW, UK, 3 Department of Electrical and Systems
Engineering, Washington University, St Louis, Missouri 63130, USA, 4 Department of Physics, Hanyang University, Seoul 133-791, Korea.
*e-mail: m.tame@imperial.ac.uk; s.maier@imperial.ac.uk

NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics 329

© 2013 Macmillan Publishers Limited. All rights reserved.
REVIEW ARTICLE NATURE PHYSICS DOI: 10.1038/NPHYS2615

Photon Plasma shifted photon

ωp
+ =
ωsp

ωp
SPP

k
Plasmon

Figure 1 | SPP. The coupling of a photon and a plasmon at the interface of a material with a negative dielectric function (for example, a metal) and one with
a positive dielectric function (for example, air) leads to a splitting of the (ω − k) dispersion curves (solid lines) for the excitations, which form a plasma
shifted photon and a SPP as the joint state of light (photon) and matter (surface plasmon).

Optical confinement structure, maintained and released back out. An experiment by

The ability of SPPs to confine and guide their coupled light field Fasel et al. demonstrated that energy–time entanglement in a
within regions far below the diffraction limit is one of the major photon pair can be preserved during the photon–SPP–photon
strengths first highlighted by Takahara and colleagues1 . Here, a conversion using a gold grating and a long-range SPP (LRSPP)
nanowire with negative dielectric function, ε, was considered, waveguide44 . The preparation of energy–time entangled SPPs
and it was found that fundamental limits imposed on the field in two separate LRSPP waveguides was later reported45 , where
confinement for standard optical materials with positive ε were coherence lengths shorter than the plasmon propagation distance
no longer valid. Subsequent work34,35 showed the underlying were used, ensuring complete conversion of photons into SPPs.
difference between subwavelength confinement, which standard A single SPP in a coherent superposition of existing at two
optical materials can also achieve (using large positive ε), and different times, with a delay much larger than the SPP lifetime,
subdiffraction confinement, which is a unique feature of light was also realized45 . The work on entanglement was extended by
guided by SPPs and localized by LSPs using materials with negative Ren et al. to spatial modes46 , showing that entanglement of orbital
ε, such as metals26 , superconductors26 and graphene26,36 . The angular momentum could also survive conversion into and out
principles of these two key concepts are described in Box 2. of SPPs. Guo et al. demonstrated the preservation of two-photon
By confining light using SPPs or LSPs, one is able to significantly quantum coherence by sending two photons onto a gold grating,
alter the photonic density of states. Thus, the dynamics of light– finding that interference fringes at a Mach–Zehnder interferometer
matter interactions can be significantly modified and enhanced37 . showed the distinct two-photon de Broglie wavelength before
Several groups initiated investigations into the emission of and after the plasmonic structure47 . Huck et al. demonstrated
light from isolated matter systems placed close to metal. Most the robustness of continuous-variable states during the photon–
notably, Hecker et al. observed a 3-fold enhancement in the SPP–photon conversion process48 , hinting at the possibility of
luminescence from a single quantum well and found that it controlling continuous-variable quantum states using plasmonics.
was due to the generation of SPWs (ref. 38). Neogi et al. used These initial experiments confirmed that photonic entangle-
time-resolved photoluminescence measurements to demonstrate ment and quantum information could be encoded into the
the enhancement of spontaneous emission due to the coupling collective motion of a many-body electronic system, and that
of a quantum well to SPWs (ref. 39). The enhancement was the macroscopic nature of an SPP (involving ∼106 electrons)
quantified using the Purcell factor40 —the ratio of the spontaneous does not destroy quantum behaviour. This was surprising, as it
emission rate to that in free space—with values of up to was anticipated that the collisions in this massive collection of
92 observed. These experimental works and related quantum charges would inevitably lead to decoherence and the loss of
optical models41 provided a stimulating backdrop for researchers quantum information.
as they started to explore plasmonic systems using quantum
optics techniques. Decoherence and loss. Although quantum properties of the light
field were found to survive the photon–SPP–photon conversion
Quantum properties of SPPs process to a high degree, in several experiments42,44,45 the visibilities
Survival of entanglement. The first experimental observation of of the interference fringes (first-order coherence) were found to
quantum optical effects in a plasmonic structure was reported by decrease as a result of ohmic loss and surface scattering. In their
Altewischer and colleagues42 . In this pioneering work (Fig. 2a) it study of squeezed states, Huck et al. successfully modelled the
was shown that when one or both photons from a polarization- decoherence leading to the degradation of squeezing in the photon–
entangled pair are converted into SPPs on gold films perforated SPP–photon conversion process as a beam splitter interaction33,48 .
by subwavelength holes (gold grating), then back into photons, Di Martino et al. went further by exciting single SPPs in metallic
their entanglement survives. Although many incident photons stripe waveguides of different lengths49 (Fig. 2b). They found that
were lost owing to losses in the metal, the photons that survived the losses incurred during propagation are consistent with an
and reached the detectors were found to be highly entangled. uncorrelated Markovian linear loss model. Fujii et al., on the other
Since this experiment and its quantum description43 , many further hand, have revealed that in addition to linear loss, in quantum
experiments have been reported, suggesting that entanglement in plasmonic systems there are chromatic and dispersion effects which
other degrees of freedom could also be transferred into a plasmonic may induce temporal and spectral mode distortion50 .

330 NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics

© 2013 Macmillan Publishers Limited. All rights reserved.
NATURE PHYSICS DOI: 10.1038/NPHYS2615 REVIEW ARTICLE
Box 1 | Quantization of SPWs.

A basic approach to the quantization of SPWs involves quantizing a b

the electromagnetic field by accounting for the dispersive proper-
ties of the metal via the collective response of the electrons32,33 . A
mathematically equivalent, but more rigorous approach including a^ a^ c^
loss uses the microscopic Hopfield27 or macroscopic Green’s b^
function28 formalism. There are three steps to quantization: (1) b^
classical mode description, (2) discretization of classical modes,
and (3) quantization via the correspondence principle. We briefly
present these steps for SPPs and LSPs. SPP LSP

Classical mode description. In Fig. B1a the SPW at a plane

interface between a metal and vacuum (or air) is shown. The Figure B1 | Quantization of SPPs on waveguides and LSPs at
metal has a dielectric constant ε(ω) and initially loss is neglected. nanoparticles. a, The magnitude of the transverse z component of the
One can describe the total electromagnetic field in terms of a mode function uK (r) for a single SPP excitation, denoted by â, is shown
vector potential, A(r,t ), where the electric and magnetic fields are along with a selection of alternative waveguide geometries. b, The
recovered in the usual way using Coulomb’s gauge (∇ ·A = 0), that magnitude of the radial r component of the mode function ui (r) for a
is E = −∂A/∂t and B = ∇ × A. By solving Maxwell’s equations a single LSP excitation. Reservoir modes, denoted by b̂ and ĉ, are also
general form of the vector potential for the SPW is found to be shown for both SPPs and LSPs, enabling loss to be included in
Z the quantization.
1
A(r,t ) = d2 KαK uK (r)exp(−iωt ) + c.c
(2π)2 responsible for the different behaviour of physical observables
Here, c.c denotes the complex conjugate, K is a real wave compared with the classical regime. For example, the widely
vector parallel to the interface and the frequency ω is linked used second-order coherence function at a fixed position145 ,
2
to the√wavenumber, K = |K|, by the dispersion relation K = g (2) (τ ) = (hÊ − (0)Ê − (τ )Ê + (τ )Ê + (0)i/hÊ − (0)Ê + (0)i ), quantifies
(ω/c) ε(ω)/(ε(ω) + 1). In addition, the term αK is an amplitude the probability of measuring an excitation at time t = 0 and
and the mode function uK (r) is given by another at t = τ . Here, hX̂ i represents the expectation value of the
operator X̂ and Ê + (t ) (Ê − (t )) is the positive (negative) electric
1

K
 field—a function of annihilation (creation) operators. For single
uK (r) = √ exp(−κj z) b K−i b z exp(iK ·r) SPPs, g (2) (0) = 0, whereas for SPWs, as there is no commutation,
L(ω) κj
the numerator factorizes to give g (2) (0) ≥ 1. Examples of g (2) (0)
where L(ω) is a length normalization and κj2 = K 2 − εj ω2 /c 2 can be seen in Figs 2b and 3b.
characterizes the decay of the field in the z direction (±ve solution The above quantization procedure can be carried out for more
of κ1,2 chosen for ±z, respectively), with ε1 = 1 and ε2 = ε(ω). In complex waveguides, such as channels and nanowires, with the
the above,bv denotes the unit vector for the vector v. only change being the mode function uK (r) which represents the
classical wavelike properties of the excitation. In most cases a
Discretization of classical modes. To discretize the SPW mode continuum limit is used for the wave vector K. To include loss
functions, a virtual square of area S = Lx × Ly is introduced on in the quantization, one couples the SPP to a reservoir of bath
the surface. This gives discretized values for the wavenumbers modes33 , b̂i , as depicted in Fig. B1a, whose coupling strength is
Kx = nx 2π/Lx and Ky = nRy 2π/Ly , where nx and ny are integers. determined in a phenomenological approach from the imaginary
By substituting (1/(2π)2 ) d 2 K → (1/S)6K and αK → SAK one part of ε(ω) for the metal, which is a result of the damping
obtains a discretized form for A(r,t ). Using the formula for the experienced by the electrons. This is mathematically equivalent to
total Renergy
R of the electromagnetic field in the virtual square the more rigorous reservoir method27 .
U = dt dr(E(∂D/∂t ) + H(∂B/∂t )), where D = ε0 E + P = εj E A similar procedure is used to quantize the near field of
and H = µ0 B are used, one finds localized plasma oscillations at nanoparticles31 , as shown in
Fig. B1b.P The vector potential for the field can be written as
X A(r,t ) = i αi ui (r)sin(ω0 t ), where the mode function is given by
U= ε0 ω2 S[AK A∗K + A∗K AK ]
ui (r) =b i for r < R and ui (r) = −(R3 /r 3 )[3(b i ·b
r)br −bi] for r > R.
K
Here, the subscript i represents the three-dimensional coordinates
which has exactly the structure of the energy of a harmonic (i = x,y,z), R is the radius of the nanoparticle and r is the radial
oscillator for each mode K. coordinate of the position vector r, taken with respect to the
centre of the nanoparticle. Following similar steps as for SPWs,
Quantization via the correspondence principle. Using one obtains bosonic annihilation and creation operators â and
the
P quantized Hamiltonian of a harmonic oscillator Ĥ = ↠, where |1i = ↠|vaci represents a single quantized localized
√K (h̄ω K /2)[â K â †
K + â †
K â K ]
√ with the correspondence AK → surface plasma oscillation, or LSP, corresponding to the creation
(h̄/2ε0 ωS)âK and A∗K → (h̄/2ε0 ωS)â†K , the field of the SPW is of a quantum of energy h̄ω in the near field of the nanoparticle.
quantized by the association of a quantum mechanical oscillator Internal damping is then modelled as a reservoir of bath modes,
with each mode K. The operators âK and â†K are annihilation and b̂i , as in the SPP case. The far-field radiation can also be treated
creation operators which destroy and create a quantum of energy, as a reservoir of bath modes, ĉi , the evolution of which can be
h̄ωK , and obey bosonic commutation relations [âK ,â†K0 ] = δK,K0 . A tracked and measured if desired, such that they do not constitute a
single quantized SPW excitation, or SPP (now both a wave and a fundamental loss channel. A more rigorous approach30,104 enables
particle), is then written as |1K i = â†K |vaci, where |vaci represents the treatment of LSPs at arbitrary shaped nanostructures and with
the vacuum state of the system. The commutation relations are hydrodynamic effects included.

NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics 331

© 2013 Macmillan Publishers Limited. All rights reserved.
REVIEW ARTICLE NATURE PHYSICS DOI: 10.1038/NPHYS2615

Box 2 | SPP field confinement: subwavelength versus subdiffraction.

The field associated with SPP quanta can be highly confined to finds the 3D diffraction limit still applies in the xy plane in the
both subwavelength and subdiffraction dimensions. To see this, upper material, with an evanescent decay given by an exponential
one can consider the k-space surfaces for three different scenarios, function with 1/e length δD = 1/κz > ∼ (λ0 /2n), which can be
as shown in Fig. B2. subwavelength for large n.
For light in a bulk three-dimensional (3D) material with By replacing the lower material with one that has a neg-
positive ε, as shown in Fig. B2a (bottom), the spatial spread ative dielectric function, as shown in Fig. B2c (bottom), one
of a beam in a plane (yz) transverse to the direction of prop- is able to ‘break’ the 3D diffraction limit. Here, noble metals
agation (x) must satisfy 4ki 4i = 2π, where i = y,z, 4i is the such as gold can be used, where the effective response of the
spatial spread in direction i and 4ki the corresponding spread electrons at the surface to the coupled field can be described by
in wavenumber. This inequality is due to the Fourier reciprocity a Drude–Lorentz dielectric function26 , ε(ω), which is negative
that occurs when arbitrary fields are expanded as a synthesis of for frequencies below the plasma frequency. In this negative
plane waves. Using Maxwell’s equations one finds the relation regime, using Maxwell’s equations, one finds kx2 + ky2 − κz2 = k02 ε2
between the wavenumber components, kx2 + ky2 + kz2 = k02 ε = k 2 , and kx2 + ky2 − κ 0 2z = k02 ε(ω). Here, the z components of both
where k0 is the freespace wavenumber, k0 = 2π/λ0 , and λ0 is the wave vectors have become imaginary—the field has become 2D.
freespace wavelength. The k-space surface for 3D waves is shown The k-space surfaces of these two equations that represent the
in Fig. B2a (top), where the maximum variation of a wavenumber combined light field supported by the electrons (the SPP) are
is 4ki = 2k. This leads to the well-known 3D diffraction limit, shown in Fig. B2c (top). Although kx2 + ky2 must match across
4i = λ0 /2n. Subwavelength confinement (compared to λ0 ) can be the boundary, its value is no longer limited, which in principle
achieved by simply using a material with a larger refractive index, enables confinement to arbitrary spatial extent in the xy plane.
n = ε 1/2 . However, an additional constraint comes from the maximum
At an interface between two materials with positive dielectric value that kx2 + ky2 can take, given by the dispersion relation
functions, ε > ε2 , as shown in Fig. B2b (bottom), where total for the SPP, as shown √ in Fig. 1. √
For the geometry consid-
internal reflection takes place, for example at the interface between ered we have kSPP = kx2 + ky2 = (ω/c) ε2 ε(ω)/(ε2√ + ε(ω)). From
core and cladding in an optical fibre or a cavity wall, one finds Fourier reciprocity this gives 4x, 4y ≥ (λ0 /2n) 1 − ε2 /|ε(ω)|
kx2 +ky2 +kz2 = k02 ε and kx2 +ky2 −κz2 = k02 ε2 . Here, the z component and δD = 1/κz ≥ (λ0 /2n)f (ε2 ,ε(ω)). Both can be made sig-
of the wave vector in the upper material has become imaginary nificantly smaller than their positive dielectric counterparts.
(kz = iκz , with κz real) representing the evanescent decay of the The amount depends on the materials and geometry, with
field. The k-space surfaces for these two equations are shown in nanowires and channel waveguides providing even larger field
Fig. B2b (top). As kx2 + ky2 must match across the interface one confinement34,35 .

a b c

kz κ z'
kz
κz κz

kx ky

ε2
z ε2
x
ε ε (ω )
ε

¬y
3D photon 2D photon SPP

Figure B2 | k-space surfaces. a, k-space surface for a photon in a bulk 3D material. One can see the maximum spread for any wavenumber is 2k,
leading to the diffraction limit. b, k-space surfaces for a 2D photon, where total internal reflection has taken place. Here, the total transverse
wavenumber, kx2 + ky2 , must match across the interface so that the maximum spread of the individual wavenumbers kx and ky in the upper material is
again diffraction limited. c, k-space surfaces for the field associated with an SPP, where the total transverse wavenumber is no longer diffraction
limited. Its value now depends on the waveguide geometry and material used.

332 NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics

© 2013 Macmillan Publishers Limited. All rights reserved.
NATURE PHYSICS DOI: 10.1038/NPHYS2615 REVIEW ARTICLE
a 50 b
1.5

Coincidence
40

rate (s¬1)
30

g2 (τ)
1.0
20
10 0.5
0 0.0
0 90 180 270 360
¬4 ¬2 0 2 4
Angle of polarizer P1 (°)
Delay τ (ns)
TEL A1
A1 1
C L
BBO P1 IF
&

HWP P2 IF
C
L
2
A2
c d Light
Photons
(I) d > dQR + ¬¬
BS near-field +
+ ¬
interaction + ¬
BS 1 μm d R
Plasmons NV I
(II) d ≈ dQR
quantum
d 1 d2 tunnelling
II
Tunnelling
Photons channel
Pc B
PB A Wire ends (III) d ≤ 0 + ¬
contact +¬ III
BS
PA
Plasmons Contact
crevice
Wire

Figure 2 | Probing fundamental quantum properties of SPPs. a, Plasmon-assisted transmission of polarization entangled photons through a metal grating
consisting of a gold film perforated by an array of subwavelength holes42 . The inset shows the fourth-order quantum interference fringes indicating that
the entanglement survives the photon–SPP–photon conversion process. BBO: β-barium borate nonlinear crystal for photon generation via parametric down
conversion, C: compensating crystal to adjust the phase between the components of the entangled state, HWP: half-wave plate, L: lens, TEL: confocal
telescope, A1 and A2: metal grating, P1 and P2: polarizer, IF: interference filter, P1 and P2: single-photon detectors. b, Single SPPs excited in a metallic stripe
waveguide by single photons from parametric down conversion49 are found to preserve their photon-number statistics, as witnessed by the second-order
quantum coherence, g(2) (τ ) (inset). At the single-quanta level, SPPs are observed to experience loss consistent with an uncorrelated Markovian loss
model, as suggested by the classical exponential behaviour of the count rates and the unchanged value of the second-order coherence function with
increasing waveguide length. c, Wave-particle duality of SPPs excited by a NV centre in diamond placed in close proximity to a nanowire12 . A single SPP
interferes with itself (wave-like, top) and shows sub-Poissonian statistics using a beamsplitter (particle-like, bottom). NV: nitrogen–vacancy centre, BS:
beamsplitter, d1 and d2 are the distance between the NV centre and the close and far end of the nanowire, respectively, PA and PB are photodiodes, Pc is a
photon correlator, A and B are the wire ends. d, Evolution of plasmonic modes as the inter-particle distance is varied from the classical regime through to
the quantum regime68 . The onset of quantum tunnelling determines a quantum limit of plasmonic confinement. Here, dQR denotes the critical distance
below which the plasmon interactions enter the quantum regime and R is the radius of the nanoparticle. Figure reproduced with permission from: a,
ref. 42 © 2002, NPG; b, ref. 49 © 2012, ACS; c, ref. 12 © 2009, NPG; and d, ref. 68 © 2012, NPG.

Wave-particle duality. One of the fundamental features of electron microscope has been used51,55 . Scholl et al. have found
quantum mechanics is that a single quantum excitation exhibits that as the diameter of a nanoparticle approaches a critical size,
both wave-like and particle-like behaviour. Kolesov et al. demon- the plasmon resonance undergoes a blue shift with linewidth
strated wave-particle duality for SPPs (ref. 12; Fig. 2c). Here, broadening, which is drastically different to the predictions of
single SPPs on a silver nanowire were generated by driving NV classical electromagnetism56,57 . Although analytical models are still
centres with an external field. The SPPs were found to self-interfere, used52,53,56,58 , recent work has employed numerical density func-
clearly showing the wave-like behaviour. They then showed the tional theory (DFT) to model the many-body electron system,
particle-like behaviour via the measurement of the second-order obtaining quantum corrected dielectric functions for predicting
quantum coherence function. experimental observations. DFT accounts for the spill-out of elec-
trons outside a nanoparticle and the gradual change of the dielectric
Quantum size effect. Depending on the size of a metal nanos- properties at the surface. Using DFT, Prodan et al. have shown
tructure, microscopic quantum effects can be significant in the that the electron spill-out in nanoshells can introduce new modes
description of the electrodynamics. The continuous electronic and a broadening of the plasmon resonances59 , in addition to
conduction band, valid at macroscopic scales, breaks up into strong changes in the plasmon line shapes due to the interplay
discrete states when the dimensions are small enough, making the between plasmons and single-electron excitations60 . Zuloaga et al.
Drude model for the dielectric function no longer valid51–54 . Many have also used DFT to investigate quantum plasmonic behaviour in
experiments have optically probed this quantum size effect51,54,55 , nanorods61 and dimers62 . Townsend and Bryant have found that in
which manifests itself as a shift and broadening of the plasmon small nanospheres there can be two types of collective oscillations,
resonance, in addition to the appearance of a fine structure, cor- quantum core plasmons in the centre and classical surface plasmons
responding to transitions between the discrete energy levels. Here, throughout63 . Quantum size effects in thin films64,65 and graphene66
electron-energy-loss spectroscopy with a scanning transmission have also been studied. These works have shown that quantum size

NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics 333

© 2013 Macmillan Publishers Limited. All rights reserved.
REVIEW ARTICLE NATURE PHYSICS DOI: 10.1038/NPHYS2615

effects need to be taken into account when designing ultracompact a plasmonic mode. The high quality factors, Q, or long interaction
nanophotonic devices based on plasmonics. times associated with traditional cavities limit the speed at which
photons can be emitted once collected into the cavity. Plasmonics
Quantum tunnelling. When metallic nanostructures are placed does not suffer from this problem and thus promises single-photon
close to each other quantum tunnelling can occur. Zuloaga et al. sources on a chip at optical frequencies with high operation speed.
have shown that electron tunnelling effects can play an important This plasmon-induced Purcell enhancement can also be used to
role in the optical resonances between two nanoparticles with encourage quantum interference between the transitions of a multi-
separation distances d < 1 nm (ref. 62). Moreover, for distances level emitter, leading to an enhancement in phenomena such as
d < 0.5 nm the dimer enters a conductive regime, where a charge electromagnetic-induced transparency, coherent population trap-
transfer plasmon mode appears involving electrons flowing back ping and lasing without inversion79,80 .
and forth between the particles. Mao et al. have investigated Recently, a 2.5-fold enhancement in the emission of a single
quantum tunnelling between two silver plates, showing that it is quantum dot into an SPP mode of a silver nanowire was
responsible for a reduction in surface-enhanced Raman scattering67 . demonstrated by Akimov et al.11 (Fig. 3a). Moreover, they observed
Savage et al. have experimentally revealed the quantum regime that the light scattered from the end of the nanowire was anti-
of tunnelling plasmonics in subnanometre plasmonic cavities bunched (Fig. 3b), confirming that the SPP mode could collect
formed by two nanostructures68 (Fig. 2d). They found that as and radiate single photons from the quantum dot. Subsequent
the nanostructure separation decreases below a critical size, the experiments have shown Purcell enhancements of single emitters
plasmon interactions enter the quantum regime, manifested by a coupled to SPP (refs 12,81–85) and LSP (refs 78,86) modes. Further
blue shift of the resonances, attributed to the screening of localized efforts have also been made to exploit more advanced designs
surface charges by quantum tunnelling and a consequent reduction to improve collection and control. One example is hybrid SPPs
in the plasmonic coupling. The results agree well with the predic- (refs 87,88), where a waveguide gap is used to achieve Purcell factors
tions of the quantum corrected model of Esteban et al.69 and recent as high as 60. The growing use of nanoantenna to control the
experiments by Scholl and colleagues70 . Nonlinear effects in quan- emission direction of the collected light89–92 is another example.
tum tunnelling have also been investigated71 . In a recent study72 , These efforts point towards the exciting prospect of single-photon
Wu et al. considered Fowler–Nordheim tunnelling, which occurs antennas93 that can efficiently absorb light from emitters and
in the presence of an external high electric field. Here, electrons subsequently emit the photons in a well-controlled manner.
from the conduction band of one nanoparticle tunnel into the gap The second principal regime is the strong-coupling regime.
between the nanoparticles and are swept into the other nanoparti- Here, the interaction√between light and matter can be described
cle. This process occurs when the barrier has a sloped energy–space by the coupling, g ∝ 1/Veff . Although confined plasmonic modes
profile. The strength and damping rate of plasmonic oscillations can couple very strongly to matter, unfortunately, because of large
be controlled by tuning the intensity of the incident light. Thus, the ohmic losses, it is not easy to enter the strong-coupling regime in
charge transfer can be modulated by an external source, which may plasmonic systems, where light–matter interactions must be dealt
be useful for developing novel quantum devices such as switches. with non-perturbatively. There is, however, a regime where the
coupling strength is intermediate between the mode and the emitter
Single emitters coupled to SPPs dissipation. This is known as the bad-cavity limit in cavity quantum
The large size mismatch between light and single emitters ensures electrodynamics (CQED) and exhibits interesting physics, such as
that their light–matter interaction is inherently weak. This is a cavity-induced transparency94 . A similar effect has been studied
problem as strong, coherent coupling between single photons and in coupled metal nanoparticle-emitter systems, where very large
emitters is critical for developing future quantum technology73 . enhancements in response have been predicted31,95,96 .
There are several strategies to circumvent this problem. High- In general, the strong-coupling regime is characterized by the
quality cavities have been used to boost interaction times and reversible exchange of energy between the light field and the
encourage stronger coupling. However the use of cavities places a emitter—Rabi oscillations. These oscillations manifest themselves
restriction on the bandwidth and the size of devices. An alternative in an energy splitting of the light–matter energy levels. There have
strategy is to use an interface to bridge the size gap. Confining the been experimental observations of these splittings in the spectra of
light field to small effective volumes in this way enables stronger ensembles of molecules due to plasmonic interactions97–100 . Exper-
coupling with the emitter. Plasmonic modes can be squeezed into imental evidence for strong coupling between a single emitter and a
volumes far below the diffraction limit, and therefore provide an plasmonic mode, however, is still elusive. Classical predictions have
excellent interface between single photons and emitters10 . suggested strong coupling could be achieved between an emitter
and a metallic dimer antenna101 . There have also been theoretical
Weak and strong coupling. Light–matter interactions can be split examinations of the strong-coupling regime based on a fully quan-
into two principal regimes, the weak-coupling and the strong- tum mechanical framework30,102,103 . These works take into consider-
coupling regime (Box 3). The weak-coupling regime is associated ation higher order modes whose relevance cannot be ignored as the
with the Purcell enhancement of spontaneous emission. This effect metal-emitter separation decreases past the point where the dipole
has been found to be particularly strong when an emitter is placed approximation is valid. As a result, the intuitive CQED analogy31
next to a metallic surface or nanostructure74–76 , where the emitter is replaced with macroscopic QED techniques better suited to
couples to confined plasmonic modes77 . Plasmonic modes are able more complex systems104 . Trügler and Hohenester30 predicted
to strongly enhance the fluorescence of emitters, despite having low the strong-coupling regime’s characteristic anti-crossing of energy
quality factors owing to ohmic losses. This enhancement is due levels for an emitter placed next to cigar-shaped nanoparticles.
to two simultaneous processes78 . First, the intense plasmonic field To increase the Q-factor of the plasmonic modes so that the
increases the excitation rate of the emitter. Second, the subwave- strong-coupling regime can be entered more easily, two main
length confinement of the light field enhances the decay rate of the strategies have been pursued. The first concentrates on reducing the
emitter into the plasmonic mode via the Purcell effect40 (Box 3). damping of the material. The high confinement and long lifetimes
The fluorescent enhancement is tempered by the non-radiative of graphene plasmons have been proposed in this regard105 . In the
excitation of lossy surface waves at the metal surface78 . This process, second, cavities have been incorporated into plasmonic structures
known as fluorescence quenching, occurs close to the surface and (Fig. 3c,d). These plasmonic resonators combine the benefits of
therefore leads to an optimal distance for coupling the emitter into a high Q-factor and small mode volume16,106–110 . De Leon et al.

334 NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics

© 2013 Macmillan Publishers Limited. All rights reserved.
NATURE PHYSICS DOI: 10.1038/NPHYS2615 REVIEW ARTICLE
Box 3 | Weak and strong coupling in plasmonic CQED.

The spontaneous emission of an emitter is strongly dependent on SPP and LSP plasmonic modes offer subwavelength and sub-
the electromagnetic environment it resides in40 . CQED involves diffraction field confinement that enables extreme light–matter
the interaction of emitters with tailored electromagnetic fields145 . coupling. In particular, the resonant LSP modes supported by
Typically these fields are monomodes with high quality factors (Q) metal nanoparticles can be described effectively as a leaky cavity in
and small effective volumes (Veff ). These properties provide emis- the quantum optics formalism, as shown in Fig. B3a. Recent work
sion enhancement, which is formally defined by the Purcell factor on adding resonators to waveguide SPP systems brings these types
of modes into the quasimode regime of CQED as well16,106,107,146 .
γcavity λ
 3 
Fp = ∝Q In the weak-coupling regime the excitation within the emitter-
γfree space Veff cavity system is irreversibly lost to the outside environment before
any coherent exchange of energy can occur. In this case the light
where γ is the decay rate of the emitter. The strength of the field has a perturbative effect on the emitter, which manifests
interaction between the emitter
√ and the field is characterized by itself as a modification of the decay rate, as described above.
a coupling frequency, g ∝ 1/Veff . CQED can be split into two In the strong-coupling regime a reversible exchange of energy,
regimes which are dependent on the comparison of g and the known as Rabi oscillations, exists between the emitter and the
damping rates of both the emitter and the cavity (γ ,κ). These cavity field147 , as shown in Fig. B3b for an emitter coupled to a
regimes are classified as the weak-coupling regime (g  γ ,κ) and metal nanoparticle. At this point the subsystems can no longer
the strong-coupling regime (g  γ ,κ; ref. 145). be treated separately. The composite system is described as a
CQED has been a popular platform for proof-of-principle dressed emitter whose eigenenergies show a degeneracy splitting
implementations of quantum information processing73 . However, in comparison to the undressed emitter145 , as shown in Fig. B3c.
the diffraction-limited optical cavities place a lower bound on The dressed emitter’s energy spacings provide a nonlinearity that
the size of these systems. The drive to bring CQED down to the enables single-photon nonlinear optics via the phenomenon of
nanoscale has opened the door to plasmonic CQED. Here, both photon blockade111 .

κ
γ
γ κ2

κ1

c
|2, +〉

b g
√2 g

ωp |2, ¬〉
ω0

|1, +〉
g
ωp |1, ¬〉

ω0

g
|0〉

Figure B3 | Weak and strong coupling. a, Analogy between an atom in a single-mode leaky cavity (left) and an atom residing in the near-field of a
resonant LSP mode supported by a metal nanoparticle (right). The principal difference between the two is their dissipation channels. The cavity
loses photons by transmission through its side walls. The LSP mode on the other hand is dissipated through radiative losses and ohmic losses
associated with the metal nanoparticle. b, Rabi oscillation describing the transfer of an excitation between the LSP mode and the emitter at a Rabi
frequency, g. c, Schematic of the dressed emitter’s energy levels. Each energy manifold of excitation number N has two states associated with√it,
|N,+i and |N,−i. The magnitude of the difference between the dressed emitter and undressed emitter energy levels for equal N is given by g N.
This anharmonic splitting is shown in the diagram and explains how a photon of a certain frequency may only excite the N = 1 manifold and nothing
greater, creating a photon blockade.

NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics 335

© 2013 Macmillan Publishers Limited. All rights reserved.
REVIEW ARTICLE NATURE PHYSICS DOI: 10.1038/NPHYS2615

a b

Self-correlation coincidences
80
Γ rad
60

40
Γ pl In√2/(R +Γ tot)
20
Heat

¬40 ¬20 0 20 40
Time delay, τ (ns)
c d
|e〉 Dipole
γs G
|g〉 d
κ1 κ0
+
er rm
e tap
r Microtoroid
Fib x
y Km
¬ κr
z x
z
y MNP

Figure 3 | Coupling of single emitters to SPPs. a, Quantum dot emission into SPP modes of a silver nanowire, as demonstrated in Akimov et al.’s
experiment11 . The dot can radiate into free space modes or SPP modes with rates Γrad or Γpl , respectively. Alternatively, it can non-radiatively excite lossy
surface modes, which quench the fluorescence. The quantum statistics of the fluorescence were investigated by observing the scattered light from the end
of the nanowire. b, Self-correlation coincidences of the scattered light from the SPP modes. At τ = 0, the coincidence counts almost reach zero. This
indicates that the SPP mode scatters into single photons. The temporal width of the anti-bunching curve depends on the pumping rate R of a quantum dot
from its ground state to excited state, and the total decay rate Γtot back to the ground state. c, Schematic (top) and scanning electron microscope image
(bottom) of a plasmon distributed Bragg reflector resonator16 (scale bar 1 µm). d, Sketch of a hybrid system of a whispering gallery mode (WGM) in a
microtoroid resonator and a metal nanoparticle (MNP) cavity, including a zoomed-in view of the nanoparticle and the emitter108 . Here, κ0 is the intrinsic
damping of the WGM, κ1 is the coupling between the tapered fibre and the WGM, rm is the radius of the MNP, d is the distance between the MNP and the
emitter, G is the vacuum Rabi frequency of the emitter, γs is the spontaneous emission rate of the emitter, κr is the radiative damping rate of the WGM due
to scattering from the MNP and Km is the ohmic damping rate of the MNP. In this composite resonator the nanoparticle acts as an antenna efficiently
coupling the emitter into the high-Q WGM cavity. Figure reproduced with permission from: a,b, ref. 11 © 2007, NPG; c, ref. 16 © 2012, APS; and d,
ref. 108 © 2009, APS.

have proposed a plasmonic resonator composed of a silver problem of high loss must be resolved for plasmonics to fulfil its
nanowire surrounded by dielectric Bragg reflectors16 (Fig. 3c), and potential. Bergman and Stockman119 have proposed a plasmonic
demonstrated Purcell factors exceeding 75. version of a laser for providing amplification via stimulated emis-
One of the main properties that make photons attractive for sion. This ‘spaser’ could produce stimulated emission of SPPs by
carrying quantum information is that they are weakly interacting. placing gain material around resonant metallic structures. The work
However, it also means that they do not interact with each other paved the way for the creation and preservation of strong, coherent
very well. Nonlinear materials can be used to boost this interaction, plasmonic fields at the nanoscale. Many proposals have since
however the nonlinearity requires a high light intensity. This is been put forward to exploit the spaser’s novel effects, including
unattractive as single-photon interactions are needed for quantum the creation of subwavelength nanolasers, which out-couple the
photonic devices. A strongly coupled light-emitter system has a spaser’s near-field as propagating radiation120–122 . As a result of the
nonlinear energy structure that allows photon–photon interactions Purcell enhancement, these nanolasers can exhibit threshold-less
at the single-photon level (Box 3). In CQED this is known as photon lasing122 . Spasers have also been considered in the design of
blockade111 . An analogy has been found for plasmonics112 and was metamaterials to eliminate damping. As metamaterials have been
used to devise the idea of a single-photon transistor17–19 . As well as brought from the microwave to the optical regime they have
applications in photonics, the strong coupling regime in plasmonics increasingly relied on plasmonic components123 . Incorporating
has also been shown to be useful in the field of physical chemistry gain will be essential for the practical realization of their novel
for enhancing chemical reactions113 . effects6 . Metamaterials have also been considered for controlling
In addition to single emitters, recent work has studied the quantum dynamics. Recent work has shown how negative-index
interaction of multiple emitters mediated through a strong metamaterials can aid nonlinear interactions124 and entanglement
interaction with a plasmonic mode114 . There have been predictions generation125 . Experimental probing of metamaterials in the
of a plasmonic Dicke effect where emitters coupled to a common quantum regime has also been demonstrated126 .
plasmonic mode experience cooperative emission115 . In a similar One of the key successes of quantum optics over the past few
scenario, mediated interactions via a plasmonic mode generate decades has been the precise control of single quantum systems in a
entanglement between emitters116–118 . This is a powerful insight, as range of settings. Cold atom trapping in optical lattices, for instance,
the proposed entanglement generation is induced from dissipative has helped shed light on a number of physical phenomena127 .
processes. In this way a perceived weakness of plasmonics has been However, optical lattices are not easily scalable and the lattice period
converted into a virtue. is restricted to half the wavelength of the trapping laser. Plasmonics
has emerged as an alternative route towards scalable solid-state
Nanolasers, metamaterials and many-body systems. Despite the systems for trapping atoms and molecules128,129 . Owing to the strong
remarkable progress in studying light–matter interactions using coupling between the emitter and the plasmonic mode, the metallic
plasmonic systems and a host of promising applications, the trap serves the dual purpose of trapping the atom as well as an

336 NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics

© 2013 Macmillan Publishers Limited. All rights reserved.
NATURE PHYSICS DOI: 10.1038/NPHYS2615 REVIEW ARTICLE
a b
Electron
¬¬¬
+++ 15
¬¬¬
+++
¬¬¬
Hole
+++

Y (µm)
¬¬¬ Ag nanowire 10
Plasmons +++
¬¬¬ Ge nanowire
+++
¬¬¬ 5
+++

0
Vb Vgate I 0 5 10 15
X (µm)

c Port 3

na noribbon
Silver

Port 1
a Port 4

b eg uide
on wav
Silic
Port 2

tr ate
subs
Silica

Figure 4 | Quantum plasmonic circuitry. a,b, Two approaches that have recently been used for realizing on-chip detection of single SPPs.
In a, electron–hole pair production in a germanium nanowire (inset), from the light field of SPPs on the silver nanowire, generates a current that can be
used for detection83 . In b, superconducting nanowire detectors are placed on top of gold stripe waveguides to achieve single-SPP detection in the near
field134 . Here, the waveguide splits to form an integrated plasmonic Hanbury Brown–Twiss interferometer that is used to demonstrate antibunching of
single SPPs by measurement of the second-order coherence. c, Hybrid metal–dielectric beamsplitter, where the excitation of SPPs enables an integrated
polarization sensitive beamsplitter143 that can be used for quantum information processing. Here unique properties of SPPs are used, including high-field
confinement (compactness), polarization dependence (variable splitting) and broadband nature (fast operation). Figure reproduced with permission from:
a, ref. 83 © 2009, NPG; b, ref. 134 © 2010, ACS; and c, ref. 143 © 2011, OSA.

efficient probe. The prospect of creating a plasmonic lattice with a By placing quantum dots in plasmonic cavities, enhanced SPP
nanometre period has been proposed130 . These lattices would serve generation rates and frequency selectivity have been investigated
as an interesting playground to examine many-body physics in a theoretically15 and observed experimentally16 . The coupling of
parameter regime that is unavailable to traditional optical lattices. light from a single fluorophore molecule to a Yagi–Uda antenna
structure has also been studied for generating single plasmonic exci-
Quantum plasmonic circuitry tations as coupled LSPs (ref. 13). The generation of single LSPs at a
Plasmonic circuitry opens up a route toward nanophotonic silver nanostructure has recently been demonstrated133 . To achieve
quantum control with compact device footprints1,2,35 , enhanced truly integrated systems without external driving fields, however,
coupling to emitter systems17–19 and an electro-optical behaviour the development of on-chip electrically driven SPP sources136 will
enabling interfacing with quantum photonic2 and electronic need to be pursued in the quantum regime.
components131 . Quantum plasmonic circuitry can be decomposed
into three principal stages132 : generation, manipulation, and Manipulation. To manipulate quantum states of SPPs, a range of
detection. The combination of these enables a self-contained ‘dark’ waveguides have been considered for guiding, the most popular
on-chip setting, where external far-field control is not required. being nanowires11,12,16–18,78,81,83 . Although these provide a highly
confined field that can be exploited for coupling the light to emit-
Generation. The generation of SPPs on waveguides has been ters, owing to ohmic losses the propagation length—the distance
achieved using various types of external quantum sources, includ- over which the SPP field intensity drops to 1/e of its initial value—is
ing parametric down-conversion42,44–47,49,50 , an optical paramet- small and of the order 10 µm at optical wavelengths. On the other
ric oscillator48 and emitters in cryostats133,134 . A more integrated hand, LRSPP waveguides have been probed in the quantum regime,
approach has been to embed emitters11,12,16,81–83 directly on the where propagation lengths of up to 1 cm have been reported44,45,48,50 .
waveguides and excite them with an external classical source, LRSPP waveguides provide relatively large propagation distances,
thereby generating single SPPs from the spontaneous emission. The but the field confinement is small. Thus, a combination of different
high field confinement of the plasmonic mode enhances the pro- waveguides may be required to reach an all-plasmonic solution for
cess, providing an efficient method to generate single SPPs. A more guiding. Materials such as graphene might also help reduce loss,
flexible approach, allowing the ‘deterministic’ launching of single while maintaining a high field confinement36,105 . An alternative
SPPs, was recently demonstrated by Cuche et al., where NV centres approach is a hybrid platform of metallic and dielectric waveguides,
were fixed onto the tip-apex of a near-field optical microscope, where the metal provides localized ‘hotspots’ for enhancing the
enabling the generation of single SPPs at freely chosen positions135 . coupling of light to emitters10,118 . Another approach is using

NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics 337

© 2013 Macmillan Publishers Limited. All rights reserved.
REVIEW ARTICLE NATURE PHYSICS DOI: 10.1038/NPHYS2615

of problems are still to be solved on the way to realizing fully

Microscopic
functioning and reliable quantum devices that take advantage of
quantization the intense light–matter interactions that plasmonics offers. The
Quantum plasmonic Dissipative-driven most pressing issue is how to deal with loss. Although recent work
metamaterials quantum dynamics has shown that loss compensation and gain can be achieved in
basic plasmonic waveguides in the classical regime4 , it remains to
be seen how these techniques can be translated into the quantum
regime and in what way noise can be accommodated. It might
ROUTE
QUANTUM
be, however, that hybrid quantum plasmonic-photonic systems
Unknown
SPP will be the optimal solution in the trade-off between confinement
Unknown topic
Unknown topic
and loss10,118,143 , perhaps even exploiting the loss when needed for
investigating dissipative effects in quantum systems20 . Even with
Many-body Loss compensation the problem of loss resolved, several important issues remain,
quantum simulation such as fundamental limits on the quality of transistor-based
Hybrid plasmonic¬photonic
Ultra-strong light¬matter quantum networks quantum optical logic gates due to phase noise imparted to the
interactions signals during the nonlinear interaction144 . Moreover, as we start
to look at miniaturizing plasmonic components further, several
Figure 5 | Quantum plasmonics roadmap. A range of topics on the horizon questions are already beginning to appear. At what scale do
for the field of quantum plasmonics. This includes the development of new current quantization methods based on a macroscopic approach
quantum plasmonic applications, such as ‘dissipative driven quantum break down? When will nonlocal microscopic effects, requiring
dynamics’ and probing deeper into the fundamental properties of light– density functional theory54,56,59–72 combined with quantum optics,
matter systems, such as their microscopic quantization and their potential need to be considered in the design of new quantum plasmonic
for ultra-strong quantum interactions with emitters at the nanoscale. components? In Fig. 5 we highlight some exciting and unexplored
topics related to these questions: the search for the answers promises
nanoparticles supporting coupled LSPs (refs 137,138). Several to make the next stage of research in quantum plasmonics a very
studies have used cavity-QED to investigate energy transfer139 , fruitful and productive time.
quantum state transfer140 , entanglement generation141 and ultra-
fast switching19 . Finally, recent work has shown that, by embedding Received 30 January 2013; accepted 22 March 2013;
gain material, loss can be compensated in plasmonic waveguides in published online 3 June 2013
the classical regime4 . Such techniques could potentially be used for
guiding in the quantum regime. References
More complex waveguide structures involving the quantum 1. Takahara, J., Yamagisha, S., Taki, H., Morimoto, A. & Kobayashi, T. Guiding
interference of SPPs have also been investigated. The excitation of a one-dimensional optical beam with nanometer diameter. Opt. Lett. 22,
of SPPs into two directions on a gold nanowire provided the first 475–477 (1997).
2. Gramotnev, D. K. Bozhevolnyi, Plasmonics beyond the diffraction limit.
realization of a quantum plasmonic beamsplitter12 . The SPPs were Nature Photon. 4, 83–91 (2010).
also made to reflect at one end of the nanowire and propagate 3. Schuller, J. A. et al. Plasmonics for extreme light concentration. Nature Mater.
back to interfere in a Mach–Zehnder interferometer. More recent 9, 193–204 (2010).
work has investigated the possibility of interfering two SPPs in 4. Berini, I. & De Leon, Surface plasmon-polariton amplifiers and lasers.
continuous142 and discrete140 waveguides. Such a setting, if realized, Nature Photon. 6, 16–24 (2011).
5. Ma, R-M., Oulton, R. F., Sorger, V. J. & Zhang, X. Plasmon lasers: Coherent
would show evidence of the bosonic nature of SPPs, and initial light source at molecular scales. Laser Photon. Rev. 7, 1–21 (2012).
experimental work has hinted at this50 . Moreover, the interference 6. Hess, O. et al. Active nanoplasmonic metamaterials. Nature Mater. 11,
of two SPPs forms an important first step in building up to more 573–584 (2012).
complex circuitry for the control of quantum states. In a more 7. Anker, J. N. et al. Biosensing with plasmonic nanosensors. Nature Mater. 7,
hybrid scenario, an integrated polarization sensitive beamsplitter 442–453 (2008).
8. Giannini, V., Fernández-Dominguez, A. I., Heck, S. C. & Maier, S. A.
has been designed143 . Plasmonic nanoantenas: Fundamentals and their use in controlling the
radiative properties of nanoemitters. Chem. Rev. 111, 3888–3912 (2011).
Detection. The near-field detection of SPPs generated from a 9. Jacob, Z. & Shalaev, V. M. Plasmonics goes quantum. Science 334,
quantum dot source was demonstrated recently using a silver 463–464 (2011).
10. Chang, D. E., Sørensen, A. S., Hemmer, P. R. & Lukin, M. D. Quantum optics
nanowire waveguide placed on top of a germanium field-effect with surface plasmons. Phys. Rev. Lett. 97, 053002 (2006).
transistor83 (Fig. 4a). Here, the a.c. electric field of the SPP generates 11. Akimov, A. V. et al. Generation of single optical plasmons in metallic
electron–hole pairs in the germanium nanowire. A d.c. electric field nanowires coupled to quantum dots. Nature 450, 402–406 (2007).
then separates these electron–hole pairs into free charges before re- 12. Kolesov, R. et al. Wave-particle duality of single surface plasmon polaritons.
combination takes place. The separated electron–hole pairs are de- Nature Phys. 5, 470–474 (2009).
13. Koenderink, A. F. Plasmon nanoparticle array waveguides for single photon
tected as current, with a sensitivity of up to 50 electrons per SPP. The and single plasmon sources. Nano Lett. 9, 4228–4233 (2009).
detection of single SPPs propagating on gold stripe waveguides has 14. Chen, Y., Lodahl, P. & Koenderink, A. F. Dynamically reconfigurable
also been demonstrated using superconducting nanowire detectors, directionality of plasmon-based single photon sources. Phys. Rev. B 82,
providing a faster operation and reduced dead-times134 (Fig. 4b). 081402 (2010).
Future work on integrated generation, manipulation and 15. Gan, C. H., Hugonin, J. P. & Lalanne, P. Proposal for compact solid-state
III–V single-plasmon sources. Phys. Rev. X 2, 021008 (2012).
detection, in addition to schemes for loss compensation, promises 16. de Leon, N. P. et al. Tailoring light–matter interaction with a nanoscale
the realization of more complex nanoscale interference devices and plasmon resonator. Phys. Rev. Lett. 108, 226803 (2012).
coincidence-based quantum plasmonic operations. 17. Chang, D. E., Sørensen, A. S., Demler, E. A. & Lukin, M. D. A single-photon
transistor using nanoscale surface plasmons. Nature Phys. 3, 807–812 (2007).
Perspectives 18. Kolchin, P., Oulton, R. F. & Zhang, X. Nonlinear quantum optics in a
waveguide: Distinct single photons strongly interacting at the single atom
A huge amount of progress has been made in the growing field level. Phys. Rev. Lett. 106, 113601 (2011).
of quantum plasmonics. However, many quantum properties of 19. Frank, R. Coherent control of Floquet-mode dressed plasmon polaritons.
surface plasmons are still to be fully explored and a number Phys. Rev. B 85, 195463 (2012).

338 NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics

© 2013 Macmillan Publishers Limited. All rights reserved.
NATURE PHYSICS DOI: 10.1038/NPHYS2615 REVIEW ARTICLE
20. Verstraete, F., Wolf, M. M. & Cirac, J. I. Quantum computation and 55. Kreibig, U. & Genzel, L. Optical absorption of small metallic particles.
quantum-state engineering driven by dissipation. Nature Phys. 5, Surf. Sci. 156, 678–700 (1985).
633–636 (2009). 56. Scholl, J. A., Koh, A. L. & Dionne, J. A. Quantum plasmon resonances of
21. Pines, D. A collective description of electron interactions: IV. Electron individual metallic nanoparticles. Nature 483, 421–427 (2012).
interaction in metals. Phys. Rev. 92, 626–636 (1953). 57. Haberland, H. Looking from both sides. Nature 494, E1 (2013).
22. Hopfield, J. J. Theory of the contribution of excitons to the complex dielectric 58. Ouyang, F., Batson, P. E. & Isaacson, M. Quantum size effects in the
constant of crystals. Phys. Rev. 112, 1555–1567 (1958). surface-plasmon excitation of small metallic particles by electron-energy-loss
23. Ritchie, R. H. Plasma losses by fast electrons in thin films. Phys. Rev. 106, spectroscopy. Phys. Rev. B 46, 15421–15425 (1992).
874–881 (1957). 59. Prodan, E. & Nordlander, P. Structural tunability of the plasmon resonances
24. Elson, J. M. & Ritchie, R. H. Photon interactions at a rough metal surface. in metallic nanoshells. Nano Lett. 3, 543–547 (2003).
Phys. Rev. B 4, 4129–4138 (1971). 60. Prodan, E., Nordlander, P. & Halas, N. J. Electronic structure and optical
25. Nakamura, Y. O. Quantization of non-radiative surface plasma oscillations. properties of gold nanoshells. Nano Lett. 3, 1411–1415 (2003).
Prog. Theor. Phys. 70, 908–919 (1983). 61. Zuloaga, J., Prodan, E. & Nordlander, P. Quantum plasmonics: Optical
26. Tassin, P., Koschny, T., Kafesaki, M. & Soukoulis, C. M. A Comparison of properties and tenability of metallic nanorods. ACS Nano 4, 5269–5276 (2010).
graphene, superconductors and metals as conductors for metamaterials and 62. Zuloaga, J., Prodan, E. & Nordlander, P. Quantum description of the plasmon
plasmonics. Nature Photon. 6, 259–264 (2012). resonances of a nanoparticle dimer. Nano Lett. 9, 887–891 (2009).
27. Huttner, B. & Barnett, S. M. Quantization of the electromagnetic field in 63. Townsend, E. & Bryant, G. W. Plasmonic properties of metallic nanoparticles:
dielectrics. Phys. Rev. A 46, 4306–4322 (1992). The effects of size quantization. Nano Lett. 12, 429–434 (2012).
28. Philbin, T. G. Canonical quantization of macroscopic electromagnetism. 64. Yuan, Z. & Gao, S. Linear-response study of plasmon excitation in metallic
New. J. Phys. 12, 123008 (2010). thin films: Layer-dependent hybridization and dispersion. Phys. Rev. B 73,
29. Crowell, J. & Ritchie, R. H. Radiative decay of Coulomb-stimulated plasmons 155411 (2006).
in spheres. Phys. Rev. 172, 436–440 (1968). 65. Jaklevic, R. C., Lambe, J., Mikkor, M. & Vassell, W. C. Observation of electron
30. Trügler, A. & Hohenester, U. Strong coupling between a metallic nanoparticle standing waves in a crystalline box. Phys. Rev. Lett. 26, 88–92 (1971).
and a single molecule. Phys. Rev. B 77, 115403 (2008). 66. Thongrattanasiri, S., Manjavacas, A. & Garcia de Abajo, F. J. Quantum
31. Waks, E. & Sridharan, D. Cavity QED treatment of interactions between a finite-size effects in graphene plasmons. ACS Nano 6, 1766–1775 (2012).
metal nanoparticle and a dipole emitter. Phys. Rev. A 82, 043845 (2010). 67. Mao, L., Li, Z., Wu, B. & Xu, H. Effects of quantum tunneling in metal
32. Archambault, A., Marquier, F., Greffet, J-J. & Arnold, C. Quantum theory of nanogap on surface-enhanced Raman scattering. Appl. Phys. Lett. 94,
spontaneous and stimulated emission of surface plasmons. Phys. Rev. B 82, 243102 (2009).
035411 (2010). 68. Savage, K. J. et al. Revealing the quantum regime in tunelling plasmonics.
33. Tame, M. S. et al. Single-photon excitation of surface plasmon polaritons. Nature 491, 574–577 (2012).
Phys. Rev. Lett. 101, 190504 (2008). 69. Esteban, R., Borisov, A. G., Nordlander, P. & Aizpurua, J. Bridging quantum
34. Takahara, J., Kusunoki, F. & Kobayashi, T. Guiding of two-dimensional and classical plasmonics. Nature Commun. 3, 825 (2012).
optical waves in a negative dielectric gap having a dielectric core. Proc. SPIE
70. Scholl, J. A., Garcia-Etxarri, A., Koh, A. L. & Dionne, J. A. Observation of
5928, 1–10 (2005).
quantum tunneling between two plasmonic nanoparticles. Nano Lett. 13,
35. Takahara, J. in Plasmonic Nanoguides and Circuits (ed. Bozhevolnyi, S. I.)
564–569 (2013).
Ch. 2 (Pan Stanford Publishing, 2009).
71. Marinica, D. C., Kazansky, A. K., Nordlander, P., Aizpurua, J. &
36. Vakil, A. & Engheta, N. Transformation optics using graphene. Science 332,
Borisov, A. G. Quantum plasmonics: Nonlinear effects in the field
1291–1294 (2011).
enhancement of a plasmonic nanoparticle dimer. Nano Lett. 12,
37. Barnes, W. L. Fluorescence near interfaces: The role of photonic mode density.
1333–1339 (2012).
J. Mod. Opt. 45, 661–699 (1998).
72. Wu, L. et al. Fowler–Nordheim tunneling induced charge transfer plasmons
38. Hecker, N. E., Höpfel, R. A., Sawaki, N., Maier, T. & Strasser, G. Surface
between nearly touching nanoparticles. ACS Nano 7, 707–716 (2013).
plasmon-enhanced photoluminescene from a single quantum well. Appl.
Phys. Lett. 75, 1577–1579 (1999). 73. Monroe, C. Quantum information processing with atoms and photons.
39. Neogi, A. et al. Enhancement of spontaneous recombination rate in a Nature 416, 238–246 (2002).
quantum well by resonant surface plasmon coupling. Phys. Rev. B 66, 74. Drexhage, K. H., Kuhn, H. & Schäfer, F. P. Variation of the fluorescence
153305 (2002). decay time of a molecule in front of a mirror. Ber. Bunsenges. Phys. Chem. 72,
40. Purcell, E. M. Spontaneous emission probabilities at radio frequencies. 329 (1968).
Phys. Rev. 69, 674 (1946). 75. Chance, R. R., Prock, A. & Silbey, R. Lifetime of an emitting molecule near a
41. Yeung, M. S. & Gustafson, T. K. Spontaneous emission near an absorbing partially reflecting surface. J. Chem. Phys. 60, 2744–2748 (1974).
dielectric surface. Phys. Rev. A 54, 5227–5242 (1996). 76. Gersten, J. & Nitzan, A. Spectroscopic properties of molecules interacting
42. Altewischer, E., van Exter, M. P. & Woerdman, J. P. Plasmon-assisted with small dielectric particles. J. Chem. Phys. 75, 1139–1152 (1981).
transmission of entangled photons. Nature 418, 304–306 (2002). 77. Barnes, W. L. Electromagnetic crystals for surface plasmon polaritons
43. Moreno, E., García, F. J., Erni, D., Ignacio Cirac, J. & Martín-Moreno, L. and the extraction of light from emissive devices. J. Lightw. Technol. 17,
Theory of plasmon-assisted transmission of entangled photons. Phys. Rev. 2170–2182 (1999).
Lett. 92, 236801 (2004). 78. Anger, P., Bharadwaj, P. & Novotny, L. Enhancement and quenching of single
44. Fasel, S. et al. Energy-time entanglement preservation in plasmon-assisted molecule fluorescence. Phys. Rev. Lett. 96, 113002 (2006).
light transmission. Phys. Rev. Lett. 94, 110501 (2005). 79. Yannopapas, V., Paspalakis, E. & Vitanov, N. V. Plasmon-induced
45. Fasel, S., Halder, M., Gisin, N. & Zbinden, H. Quantum superposition and enhancement of quantum interference near metallic nanostructures. Phys.
entanglement of mesoscopic plasmons. New J. Phys. 8, 13–20 (2006). Rev. Lett. 103, 063602 (2009).
46. Ren, X. F., Guo, G. P., Huang, Y. F., Li, C. F. & Guo, G. C. Plasmon-assisted 80. Hatef, A. & Singh, M. R. Plasmonic effect on quantum coherence and
transmission of high-dimensional orbital angular-momentum entangled state. interference in metallic photonic crystals doped with quantum dots. Phys.
Europhys. Lett. 76, 753–759 (2006). Rev. A 81, 063816 (2010).
47. Guo, G. P. et al. Observation of two-photon coherence in plasmon-assisted 81. Huck, A., Kumar, S., Shakoor, A. & Andersen, U. L. Controlled coupling
transmission. Phys. Lett. A 361, 218–222 (2007). of single nitrogen-vacancy center to a silver nanowire. Phys. Rev. Lett. 106,
48. Huck, A. et al. Demonstration of quadrature-squeezed surface plasmons in a 096801 (2011).
gold waveguide. Phys. Rev. Lett. 102, 246802 (2009). 82. Fedutik, Y., Temnov, V. V., Schöps, O., Woggon, U. & Artemyev, M. V.
49. Di Martino, G. et al. Quantum statistics of surface plasmon polaritons in Exciton-plasmon-photon conversion in plasmonic nanostructures. Phys. Rev.
metallic stripe waveguides. Nano Lett. 12, 2504–2508 (2012). Lett. 99, 136802 (2007).
50. Fujii, G. et al. Preservation of photon indistinguishability after 83. Falk, A. L. et al. Near-field electrical detection of optical plasmons and
transmission through surface-plasmon-polariton waveguide. Opt. Lett. single-plasmon sources. Nature Phys. 5, 475–479 (2009).
37, 1535–1537 (2012). 84. Ropp, A. et al. Nanoscale imaging and spontaneous emission control with a
51. Halperin, W. P. Quantum size effects in metal particles. Rev. Mod. Phys. 58, single nano-positioned quantum dot. Nature Commun. 4, 1447 (2013).
533–606 (1986). 85. Pfaff, W., Vos, A. & Hanson, R. Top-down fabrication of plasmonic
52. Genzel, L., Martin, T. P. & Kreibig, U. Dielectric function and plasma nanostructures for deterministic coupling to single quantum emitters.
resonances of small metal particles. Z. Physik B 21, 339–346 (1975). J. Appl. Phys. 113, 024310 (2013).
53. Keller, O., Xiao, M. & Bozhevolnyi, S. Optical diamagnetic polarizability of 86. Kühn, S., Håkanson, U., Rogobete, L. & Sandoghdar, V. Enhancement
a mesoscopic metallic sphere: Transverse self-field approach. Opt. Commun. of single-molecule fluorescence using a gold nanoparticle as an optical
102, 238–244 (1993). nanoantenna. Phys. Rev. Lett. 97, 017402 (2006).
54. Halas, N. J., Lal, S., Chang, W-S., Link, S. & Nordlander, P. Plasmons in 87. Sorger, V. J. et al. Strongly enhanced molecular fluorescence inside a nanoscale
strongly coupled metallic nanostructures. Chem. Rev. 111, 3913–3961 (2011). waveguide gap. Nano Lett. 11, 4907–4911 (2011).

NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics 339

© 2013 Macmillan Publishers Limited. All rights reserved.
REVIEW ARTICLE NATURE PHYSICS DOI: 10.1038/NPHYS2615

88. Jun, Y. C., Kekatpure, R. D., White, J. S. & Brongersma, M. L. Nonresonant 119. Bergman, D. J. & Stockman, M. I. Surface plasmon amplification by
enhancement of spontaneous emission in metal–dielectric–metal plasmon stimulated emission of radiation: Quantum generation of coherent surface
waveguide structures. Phys. Rev. B 78, 153111 (2008). plasmons in nanosystems. Phys. Rev. Lett. 90, 027402 (2003).
89. Curto, A. G. et al. Unidirectional emission of a quantum dot coupled to a 120. Noginov, M. A. et al. Demonstration of a spaser-based nanolaser. Nature 460,
nanoantenna. Science 329, 930–911 (2010). 1110–1112 (2009).
90. Farahani, J. N., Pohl, D. W., Eisler, H-J. & Hecht, B. Single quantum dot 121. Oulton, R. F. et al. Plasmon lasers at deep subwavelength scale. Nature 461,
coupled to a scanning optical antenna: A tunable superemitter. Phys. Rev. Lett. 629–632 (2009).
95, 017402 (2005). 122. Ma, R., Oulton, R. F., Sorger, V., Bartal, G. & Zhang, X. Room-temperature
91. Kinkhabwala, A. et al. Large single-molecule fluorescence enhancements sub-diffraction-limited plasmon laser by total internal reflection. Nature
produced by a bowtie nanoantenna. Nature Photon. 3, 654–657 (2012). Mater. 10, 110–113 (2011).
92. Koenderink, A. F. On the use of Purcell factors for plasmon antennas. 123. Shalaev, V. M. Optical negative-index metamaterials. Nature Photon. 1,
Opt. Lett. 35, 4208–4210 (2010). 41–48 (2007).
93. Chen, Y., Wubs, M., Mørk, J. & Koenderink, A. F. Coherent single-photon 124. Moiseev, S. A., Kamli, A. & Sanders, B. C. Low-loss nonlinear polaritonics.
absorption by single emitters coupled to one-dimensional nanophotonic Phys. Rev. A. 81, 033839 (2010).
waveguides. New J. Phys. 13, 103010 (2011). 125. Siomau, M., Kamli, A. A., Moiseev, A. A. & Sanders, B. C. Entanglement
94. Rice, P. R. & Brecha, R. J. Cavity induced transparency. Opt. Commun. 126, creation with negative index metamaterials. Phys. Rev. A 85, 050303 (2012).
230–235 (1995). 126. Wang, S. M. et al. Hong–Ou–Mandel interference mediated by the magnetic
95. Ridolfo, A., Di Stefano, O., Fina, N., Saija, R. & Savasta, S. Quantum plasmon waves in a three-dimensional optical metamaterial. Opt. Lett. 20,
plasmonics with quantum dot-metal nanoparticle molecules: Influence of the 5213–5218 (2012).
Fano effect on photon statistics. Phys. Rev. Lett. 105, 263601 (2010). 127. Bloch, I., Dalibard, J. & Nascimbène, S. Quantum simulations with ultracold
96. Zhang, W., Govorov, A. O. & Bryant, G. W. Semiconductor-metal quantum gases. Nature Phys. 8, 267–276 (2012).
nanoparticle molecules: Hybrid excitons and the nonlinear Fano effect. 128. Chang, D. E. et al. Trapping and manipulation of isolated atoms using
Phys. Rev. Lett. 97, 146804 (2006). nanoscale plasmonic structures. Phys. Rev. Lett. 103, 123004 (2009).
97. Dintinger, J., Klein, S., Bustos, F., Barnes, W. L. & Ebbesen, T. W. Strong 129. Stehle, C. et al. Plasmonically tailored micropotenials for ultracold atoms.
coupling between surface plasmon-polaritons and organic molecules in Nature Photon. 5, 494–498 (2011).
subwavelength hole arrays. Phys. Rev. B 71, 035424 (2005). 130. Gullans, M. et al. Nanoplasmonic Lattices for Ultracold atoms. Phys. Rev. Lett.
98. Fofang, N. T. et al. Plexitonic nanoparticles: Plasmon-exciton coupling in 109, 235309 (2012).
131. Ozbay, E. Plasmonics: Merging photonics and electronics at nanoscale
nanoshell-J aggregate complexes. Nano Lett. 8, 3481–3487 (2008).
dimensions. Science 311, 189–193 (2006).
99. Passmore, B. S. et al. Observation of Rabi splitting from surface plasmon
132. De Leon, N. P., Lukin, M. D. & Park, H. Quantum plasmonic circuits.
coupled conduction state transitions in electrically excited InAs quantum dots.
IEEE Sel. Top. Quant. Elec. 18, 1781–1791 (2012).
Nano Lett. 11, 338–342 (2011).
133. Celebrano, M. et al. Efficient coupling of single photons to single plasmons.
100. Vasa, P. et al. Real time observations of ultrafast Rabi oscillations between
Opt. Exp. 18, 13829–13835 (2010).
excitons and plasmons in metal nanostructures with J-aggregates. Nature
134. Heeres, R. W. et al. On-chip single plasmon detection. Nano Lett. 10,
Photon. 7, 128–132 (2013).
661–664 (2010).
101. Savasta, S. et al. Interaction nanopolaritons: Vacuum Rabi splitting with a
135. Cuche, A., Mollet, O., Drezet, A. & Huant, S. Deterministic quantum
single quantum dot in the center of a dimer. ACS Nano 4, 6369–6376 (2010). plasmonics. Nano Lett. 10, 4566–4570 (2011).
102. Van Vlack, C., Kristensen, P. T. & Hughes, S. Spontaneous emission spectra 136. Koller, D. M. et al. Surface plasmon coupled electroluminescent emission.
and quantum light–matter interactions from a strongly coupled quantum dot Appl. Phys. Lett. 92, 103304 (2008).
metal nanoparticle system. Phys. Rev. B 85, 075303 (2012). 137. Quinten, M., Leitner, A., Krenn, J. R. & Aussenegg, F. R. Electromagnetic
103. Gonzalez-Tudela, A., Rodríguez, F. J., Quiroga, L. & Tejedor, C. Dissipative energy transport via linear chains of silver nanoparticles. Opt. Lett. 23,
dynamics of a solid-state qubit coupled to surface plasmons: From Markov to 1331–1333 (1998).
non-Markov regimes. Phys. Rev. B 82, 115334 (2010). 138. Maier, S. A. et al. Local detection of electromagnetic energy transport below
104. Dung, H., Knöll, L. & Welsch, D. Three-dimensional quantization of the the diffraction limit in metal nanoparticle plasmon waveguides. Nature Mater.
electromagnetic field in dispersive and absorbing inhomogeneous dielectrics. 2, 229–232 (2003).
Phys. Rev. A 57, 3931–3942 (1998). 139. Yurke, B. & Kuang, W. Passive linear nanoscale optical and molecular
105. Koppens, F. H. L., Chang, D. E. & García de Abajo, F. J. Graphene electronics device synthesis from nanoparticles. Phys. Rev. A 81,
plasmonics: A platform for strong light–matter interactions. Nano Lett. 11, 033814 (2010).
3370–3377 (2011). 140. Lee, C., Tame, M. S., Lim, J. & Lee, J. Quantum plasmonics with a metal
106. Gong, Y. & Vučković, J. Design of plasmon cavities for solid-state cavity nanoparticle array. Phys, Rev. A 85, 063823 (2012).
quantum electrodynamics applications. Appl. Phys. Lett. 90, 033133 (2007). 141. Lee, C. et al. Robust-to-loss entanglement generation
107. Min, B. et al. High-Q surface-plasmon-polariton whispering-gallery using a quantum plasmonic nanoparticle array. Preprint at
microcavity. Nature 457, 455–458 (2009). http://arxiv.org/abs/1303.5092 (2013).
108. Xiao, Y. et al. Strongly enhanced light matter interaction in a hybrid 142. Ballester, D., Tame, M. S. & Kim, M. S. Quantum theory of surface plasmon
photonic-plasmonic resonator. Phys. Rev. A 85, 031805 (2009). polariton scattering. Phys. Rev. A 82, 012325 (2010).
109. Choy, J. T. et al. Enhanced single-photon emission from a diamond–silver 143. Zou, C-L. et al. Broadband integrated polarization beamsplitter with surface
aperture. Nature Photon. 5, 738–743 (2011). plasmon. Opt. Lett. 36, 3630–3632 (2011).
110. Russell, K. J., Liu, T., Cui, S. & Hu, E. L. Large spontaneous emission 144. Shapiro, J. Single-photon Kerr nonlinearities do not help quantum
enhancement in plasmonic nanocavities. Nature Photon. 6, 459–462 (2012). computation. Phys. Rev. A 73, 062305 (2006).
111. Birnbaum, K. M. et al. Photon blockade in an optical cavity with one trapped 145. Walls, D. F. & Milburn, G. J. Quantum Optics (Springer, 2008).
atom. Nature 436, 87–90 (2005). 146. Hümmer, T., García-Vidal, F. J., Martín-Moreno, L. & Zueco, D. Weak and
112. Manjavacas, A., Nordlander, P. & García de Abajo, F. J. Plasmon blockade in strong coupling regimes in plasmonic-QED. Phys. Rev. B 87, 115419 (2013).
nanostructured graphene. ACS Nano 6, 1724–1731 (2012). 147. Brune, M. et al. Quantum Rabi oscillations: A direct test of field quantization
113. Hutchison, J. A., Schwartz, T., Genet, C., Devaux, E. & Ebbesen, T. W. in a cavity. Phys. Rev. Lett. 76, 1800–1803 (1996).
Modifying chemical landscapes by coupling to vacuum fields. Angew. Chem.
Int. Ed. 51, 1592–1596 (2012). Acknowledgements
114. Dzsotjan, D., Kaestel, J. & Fleischhauer, M. Dipole–dipole shift of quantum We thank J. Takahara and C. Lee for comments on the manuscript. This work was
emitters coupled to surface plasmons of a nanowire. Phys. Rev. B 84, supported by the UK’s Engineering and Physical Sciences Research Council, the
075419 (2011). Leverhulme Trust, the National Research Foundation of Korea grants funded by the
115. Pustovit, V. N. & Shahnazyan, T. V. Cooperative emission of light by an Korean Government (Ministry of Education, Science and Technology; grant numbers
ensemble of dipoles near a metal nanoparticle: The plasmonic Dicke effect. 2010-0018295 and 2010-0015059), the European Office of Aerospace Research and
Phys. Rev. Lett. 102, 077401 (2009). Development (EOARD), and the Qatar National Research Fund (Grant NPRP
116. Martín-Cano, D. et al. Dissipation-driven generation of two-qubit 4-554-1-D84). S.K.O. thanks L. Yang and F. Nori for their support.
entanglement mediated by plasmonic waveguides. Phys. Rev. B 84,
235306 (2011). Additional information
117. Lin, Z-R. et al. Quantum bus of metal nanoring with surface plasmon Reprints and permissions information is available online at www.nature.com/reprints.
polaritons. Phys. Rev. B 82, 241401 (2010). Correspondence and requests for materials should be addressed to M.S.T. or S.A.M.
118. Chen, G-Y., Lambert, N., Chou, C-H., Chen, Y-N. & Nori, F. Surface
plasmons in a metal nanowire coupled to colloidal quantum dots: Scattering Competing financial interests
properties and quantum entanglement. Phys. Rev. B 84, 045310 (2011). The authors declare no competing financial interests.

340 NATURE PHYSICS | VOL 9 | JUNE 2013 | www.nature.com/naturephysics

© 2013 Macmillan Publishers Limited. All rights reserved.