RepProgPh70 - Pitarke Theory SPSPP

INSTITUTE OF PHYSICS PUBLISHING
REPORTS ON PROGRESS IN PHYSICS
Rep. Prog. Phys. 70 (2007) 187
doi:10.1088/0034-4885/70/1/R01
Theory of surface plasmons and surface-plasmon

polaritons
J M Pitarke1,2 , V M Silkin2 , E V Chulkov2,3 and P M Echenique2,3
1
Materia Kondentsatuaren Fisika Saila, Zientzi Fakultatea, Euskal Herriko Unibertsitatea,

644 Posta kutxatila, E-48080 Bilbo, Basque Country, Spain
2
Donostia International Physics Center (DIPC) and Unidad de Fsica de Materiales
CSIC-UPV/EHU, Manuel de Lardizabal Pasealekua, E-20018 Donostia, Basque Country,
Spain
3
Materialen Fisika Saila, Kimika Fakultatea, Euskal Herriko Unibertsitatea, 1072 Posta
kutxatila, E-20080 Donostia, Basque Country, Spain
Received 2 August 2006, in final form 9 October 2006
Published 7 December 2006
Online at stacks.iop.org/RoPP/70/1
Abstract
Collective electronic excitations at metal surfaces are well known to play a key role in a wide
spectrum of science, ranging from physics and materials science to biology. Here we focus
on a theoretical description of the many-body dynamical electronic response of solids, which
underlines the existence of various collective electronic excitations at metal surfaces, such as the
conventional surface plasmon, multipole plasmons and the recently predicted acoustic surface
plasmon. We also review existing calculations, experimental measurements and applications.
This article was invited by Professor M Finnis.

0034-4885/07/010001+87$90.00
2007 IOP Publishing Ltd
Printed in the UK
J M Pitarke et al
Contents
1. Introduction
2. Surface-plasmon polariton: classical approach
2.1. Semi-infinite system
2.1.1. The surface-plasmon condition.
2.1.2. Energy dispersion.
2.1.3. Skin depth.
2.2. Thin films
3. Nonretarded surface plasmon: simplified models
3.1. Planar surface plasmon
3.1.1. Classical model.
3.1.2. Nonlocal corrections.
3.1.3. Hydrodynamic approximation.
3.2. Localized surface plasmons: classical approach
3.2.1. Simple geometries.
3.2.2. Boundary-charge method.
3.2.3. Composite systems: effective-medium approach.
3.2.4. Periodic structures.
3.2.5. Sum rules.
4. Dynamical structure factor
5. Density-response function
5.1. Random-phase approximation (RPA)
5.2. Time-dependent density-functional theory
5.2.1. The XC kernel.
6. Inverse dielectric function
7. Screened interaction
7.1. Classical model
7.1.1. Planar surface.
7.1.2. Spheres.
7.1.3. Cylinders.
7.2. Nonlocal models: planar surface
7.2.1. Hydrodynamic model.
7.2.2. Specular-reflection model (SRM).
7.2.3. Self-consistent scheme.
8. Surface-response function
8.1. Generation-rate of electronic excitations
8.2. Inelastic electron scattering
8.3. Surface plasmons: jellium surface
8.3.1. Simple models.
8.3.2. Self-consistent calculations: long wavelengths.
8.3.3. Self-consistent calculations: arbitrary wavelengths.
Page
4
8
8
8
10
11
12
13
14
14
14
15
17
17
18
19
21
21
23
23
24
25
25
27
28
28
29
29
31
33
33
35
38
42
42
43
44
44
45
47
Theory of surface plasmons and surface-plasmon polaritons
8.3.4. Surface-plasmon linewidth.

8.3.5. Multipole surface plasmons.
8.4. Surface plasmons: real surfaces
8.4.1. Stabilized jellium model.
8.4.2. Occupied d-bands: simple models.
8.4.3. First-principles calculations.
8.5. Acoustic surface plasmons
8.5.1. A simple model.
8.5.2. 1D model calculations.
8.5.3. First-principles calculations.
8.5.4. Excitation of acoustic surface plasmons.
9. Applications
9.1. Particlesurface interactions: energy loss
9.2. STEM: valence EELS
9.2.2. Spheres.
9.2.3. Cylinders.
9.3. Plasmonics
Acknowledgments
References
49
51
54
54
55
60
63
63
65
69
69
70
70
72
75
76
77
78
78
80
80
J M Pitarke et al
1. Introduction
In his pioneering treatment of characteristic energy losses of fast electrons passing through
thin metal films, Ritchie predicted the existence of self-sustained collective excitations at metal
surfaces [1]. It had already been pointed out by Pines and Bohm [2, 3] that the long-range
nature of the Coulomb interaction between valence electrons in metals yields collective plasma
oscillations similar to the electron-density oscillations observed by Tonks and Langmuir in
electrical discharges in gases [4], thereby explaining early experiments by Ruthemann [5] and
Lang [6] on the bombardment of thin metallic films by fast electrons. Ritchie investigated
the impact of the film boundaries on the production of collective excitations and found that
the boundary effect is to cause the appearance of a new lowered loss due to the excitation
of surface collective oscillations [1]. Two years later, in a series of electron energy-loss
experiments Powell and Swan [7] demonstrated the existence of these collective excitations,
the quanta of which Stern and Ferrell called the surface plasmon [8].
Since then, there has been a significant advance in both theoretical and experimental
investigations of surface plasmons, which for researches in the field of condensed matter and
surface physics have played a key role in the interpretation of a great variety of experiments
and the understanding of various fundamental properties of solids. These include the nature of
Van der Waals forces [911], the classical image potential acting between a point classical
charge and a metal surface [1215], the energy transfer in gassurface interactions [16],
surface energies [1719], the damping of surface vibrational modes [20, 21], the energy loss
of charged particles moving outside a metal surface [22, 23] and the de-excitation of adsorbed
molecules [24]. Surface plasmons have also been employed in a wide spectrum of studies
ranging from electrochemistry [25], wetting [26] and biosensing [2729] to scanning tunnelling
microscopy [30], the ejection of ions from surfaces [31], nanoparticle growth [32,33], surfaceplasmon microscopy [34, 35] and surface-plasmon resonance technology [3642]. Renewed
interest in surface plasmons has come from recent advances in the investigation of the
electromagnetic properties of nanostructured materials [43, 44], one of the most attractive
aspects of these collective excitations now being their use to concentrate light in subwavelength
structures and to enhance transmission through periodic arrays of subwavelength holes in
optically thick metallic films [45, 46].
The so-called field of plasmonics represents an exciting new area for the application of
surface and interface plasmons, an area in which surface-plasmon based circuits merge the
fields of photonics and electronics at the nanoscale [47]. Indeed, surface-plasmon polaritons
can serve as a basis for constructing nanoscale photonic circuits that will be able to carry
optical signals and electric currents [48,49]. Surface plasmons can also serve as a basis for the
design, fabrication and characterization of subwavelength waveguide components [5064]. In
the framework of plasmonics, modulators and switches have also been investigated [65, 66],
as well as the use of surface plasmons as mediators in the transfer of energy from donor to
acceptors molecules on opposite sides of metal films [67].
According to the work of Pines and Bohm, the quantum energy collective plasma
oscillations in a free electron gas with equilibrium density n is h
p = h
(4 ne2 /me )1/2 ,
4
p being the so-called plasmon frequency . In the presence of a planar boundary, there is
a new mode (the surface plasmon), the frequency of which equals in the nonretarded region
(where the speed of light can be taken to be infinitely large) Ritchies frequency s = p / 2
at wave vectors q in the range s /c q qF (qF being the magnitude of the Fermi wave
The electron density n is usually characterized by the density parameter rs = (3/4 n)1/3 /a0 , a0 being the Bohr
radius, a0 = 0.529 . In metals the electron-density parameter of valence electrons is typically in the range 2 < rs < 6,
which corresponds to plasmon energies on the order of 10 eV and frequencies that lie in the optical regime.
vector) and exhibits some dispersion as the wave vector is increased. In the retarded region,
where the phase velocity s /q of the surface plasmon is comparable to the velocity of light,
surface plasmons couple with the free electromagnetic field. These surface-plasmon polaritons
propagate along the metal surface
with frequencies ranging from zero (at q = 0) towards the
asymptotic value s = p / 2, the dispersion relation (q) lying to the right of the light line
and the propagating vector being, therefore, larger than that of bare light waves of the same
energy. Hence, surface-plasmon polaritons in an ideal semi-infinite medium are nonradiative
in nature, i.e. cannot decay by emitting a photon and, conversely, light incident on an ideal
surface cannot excite surface plasmons.
In the case of thin films, the electric fields of both surfaces interact. As a result, there
are (i) tangential oscillations characterized by a symmetric disposition of charge deficiency or
excess at opposing points on the two surfaces and (ii) normal oscillations in which an excess
of charge density at a point on one surface is accompanied by a deficiency at the point directly
across the thin film. The phase velocity of the tangential surface plasmon is always less than
the speed of light, as it occurs in the case of a semi-infinite electron system. However, the
phase velocity of normal oscillations may surpass that of light, thereby becoming a radiative
surface plasmon that should be responsible for the emission of light [68]. This radiation was
detected using electron beam bombardment of thin films of Ag, Mg and Al with thicknesses
ranging between 500 and 1000 [69, 70]. More recently, light emission was observed in the
ultraviolet from a metaloxidemetal tunnel diode and was attributed to the excitation of the
radiative surface plasmon [71].
Nonradiative surface plasmons in both thin and thick films can couple to electromagnetic
radiation in the presence of surface roughness or a grating, as suggested by Teng and Stern [72].
Alternatively, prism coupling can be used to enhance the momentum of incident light, as
demonstrated by Otto [73] and by Kretchmann and Raether [74]. Since then, this so-called
attenuated reflection (ATR) method and variations upon it have been used by several workers
in a large variety of applications [7581].
During the last decades, there has also been a significant advance in our understanding of
surface plasmons in the nonretarded regime. Ritchie [82] and Kanazawa [83] were the first
to attack the problem of determining the dispersion (q) of the nonretarded surface plasmon.
Bennett [84] used a hydrodynamical model with a continuous decrease of the electron density
at the metal surface and found that a continuous electron-density variation yields two collective
electronic excitations: Ritchies surface plasmon at s , with a negative energy dispersion
at low wave vectors, and an upper surface plasmon at higher energies. In the direction normal
to the surface, the distribution of Ritchies surface plasmon consists of a single peak, i.e. it has
a monopole character; however, the charge distribution of the upper mode has a node, i.e. it
has a dipole character and is usually called multipole surface plasmon.
Bennetts qualitative conclusions were generally confirmed by microscopic descriptions
of the electron gas. On the one hand, Feibelman showed that in the long-wavelength
limit the classical result s = p / 2 is correct for a semi-infinite plane-bounded electron
gas, irrespective of the exact variation of the electron density in the neighbourhood of the
surface [85]. On the other hand, explicit expressions for the linear momentum dispersion of
the conventional monopole surface plasmon that are sensitive to the actual form of the electrondensity fluctuation at the surface were derived by Harris and Griffin [86] using the equation
of motion for the Wigner distribution function in the random-phase approximation (RPA)
and by Flores and Garca-Moliner [87] solving Maxwells equations in combination with an
integration of the field components over the surface region. Quantitative RPA calculations
of the linear dispersion of the monopole surface plasmon were carried out by several authors
using the infinite-barrier model (IBM) of the surface [88], a step potential [89, 90], and the
J M Pitarke et al
more realistic LangKohn [91] self-consistent surface potential [92]. Feibelmans calculations
showed that for the typical electron densities in metals (2 < rs < 6) the initial slope of the
momentum dispersion of monopole surface plasmons of jellium surfaces is negative [92], as
anticipated by Bennett [84].
Negative values of the momentum dispersion had been observed by high-energy electron
transmission on uncharacterized Mg surfaces [93] and later by inelastic low-energy electron
diffraction on the (100) and (111) surfaces of Al [94,95]. Nevertheless, Klos and Raether [96]
and Krane and Raether [97] did not observe a negative dispersion for Mg and Al films.
Conclusive experimental confirmation of the negative surface plasmon dispersion of a variety
of simple metals (Li, Na, K, Cs, Al and Mg) did not come about until several years
later [98101], in a series of experiments based on angle-resolved low-energy inelastic electron
scattering5 . These experiments showed good agreement with self-consistent dynamicalresponse calculations carried out for a jellium surface6 in a time-dependent adiabatic extension
of the density-functional theory (DFT) of Hohenberg, Kohn and Sham [102]. Furthermore,
these experiments also showed that the multipole surface plasmon was observable, its energy
and dispersion being in quantitative agreement with the self-consistent jellium calculations
that had been reported by Liebsch [103].
Significant deviations from the dispersion of surface plasmons at jellium surfaces occur
on Ag [104107] and Hg [108], due to the presence of filled 4d and 5d bands, respectively,
which in the case of Ag yields an anomalous positive dispersion. In order to describe the
observed features of Ag surface plasmons, various simplified models for the screening of d
electrons have been developed [109113]. Most recently, calculations have been found to yield
a qualitative understanding of the existing electron energy-loss measurements by combining
a self-consistent jellium model for valence 5s electrons with a so-called dipolium model in
which the occupied 4d bands are represented in terms of polarizable spheres located at the
sites of a semi-infinite face-cubic-centred (fcc) lattice [114].
Ab initio bulk calculations of the dynamical response and plasmon dispersions of noble
metals with occupied d bands have been carried out recently [115117]. However, firstprinciples calculations of the surface-plasmon energy and linewidth dispersion of real solids
have been carried out only in the case of the simple-metal prototype surfaces Mg(0001) and
Al(111) [118,119]. These calculations lead to an accurate description of the measured surfaceplasmon energy dispersion that is superior to that obtained in the jellium model, and they show
that the band structure is of paramount importance for a correct description of the surfaceplasmon linewidth.
The multipole surface plasmon, which is originated in the selvage electronic structure at the
surface, has been observed in a variety of simple metals at 0.8p [98101], in agreement
with theoretical predictions. Nevertheless, electron energy-loss spectroscopy (EELS)
measurements of Ag, Hg and Li revealed no clear evidence of the multipole surface plasmon. In
the case of Ag, high-resolution energy-loss spectroscopy low-energy electron diffraction (ELSLEED) measurements indicated that a peak was obtained at 3.72 eV by subtracting the data for
two different impact energies [120], which was interpreted to be the Ag multipole plasmon.
However, Liebsch argued that the frequency of the Ag multipole surface plasmon should be
in the 68 eV range above rather than below the bulk plasma frequency and suggested that the
observed peak at 3.72 eV might not be associated with a multipole surface plasmon [121].
5
Since the finite angular acceptance of typical energy-loss spectrometers guarantees that the momentum transfer is
associated with a plasmon wavelength larger than the wavelength of light at = s (2 c/s 103 ), only the
nonretarded region is observed in these experiments.
6 A jellium surface consists of a fixed uniform positive background occupying a halfspace plus a neutralizing cloud
of interacting electrons.
An alternative spectroscopy technique to investigate multipole surface plasmons is

provided by angle- and energy-resolved photoyield experiments (AERPY) [122]. In fact,
AERPY is more suitable than electron energy-loss spectroscopy to identify the multipole
surface plasmon, since the monopole surface plasmon of clean flat surfaces (which is the
dominant feature in electron energy-loss spectra) is not excited by photons and thus the
weaker multipole surface mode (which intersects the radiation line in the retardation regime)
can be observed. A large increase in the surface photoyield was observed at = 0.8p
from Al(100) [122] and Al(111) [123]. Recently, the surface electronic structure and optical
response of Ag has been studied using this technique [124]. In these experiments, the
Ag multipole surface plasmon is observed at 3.7 eV, while no signature of the multipole
surface plasmon is observed above the plasma frequency (p = 3.8 eV) in disagreement
with the existing theoretical prediction [121]. Hence, further theoretical work is needed on
the surface electronic response of Ag that go beyond the sd polarization model described
in [121].
Another collective electronic excitation at metal surfaces is the so-called acoustic surface
plasmon that has been predicted to exist at solid surfaces where a partially occupied
quasi-two-dimensional surface-state band coexists with the underlying three-dimensional
continuum [125, 126]. This new low-energy collective excitation exhibits linear dispersion
at low wave vectors and might therefore affect electronhole (eh) and phonon dynamics
near the Fermi level7 . It has been demonstrated that it is a combination of the nonlocality
of the 3D dynamical screening and the spill out of the 3D electron density into the vacuum
which allows the formation of 2D electron-density acoustic oscillations at metal surfaces,
since these oscillations would otherwise be completely screened by the surrounding 3D
substrate [127]. This novel surface-plasmon mode has been observed recently at the (0001)
surface of Be, showing a linear energy dispersion that is in very good agreement with firstprinciples calculations [128].
Finally, we note that metaldielectric interfaces of arbitrary geometries also support charge
density oscillations similar to the surface plasmons characteristic of planar interfaces. These
are localized Mie plasmons occurring at frequencies which are characteristic of the interface
geometry [129]. The excitation of localized plasmons on small particles has attracted great
interest over the years in scanning transmission electron microscopy [130135] and near-field
optical spectroscopy [136]. Recently, new advances in structuring and manipulating on the
nanometre scale have rekindled interest in this field [137]. In nanostructured metals and
carbon-based structures, such as fullerenes and carbon nanotubes, localized plasmons can be
excited by light and can therefore be easily detected as pronounced optical resonances [138
140]. Furthermore, very localized dipole and multipole modes in the vicinity of highly coupled
structures are responsible for surface-enhanced Raman scattering [141,142] and other striking
properties like, for example, the blackness of colloidal silver [143].
Collective electronic excitations in thin adsorbed overlayers, semiconductor heterostructures, and parabolic quantum wells have also attracted attention over the last years. The
adsorption of thin films is important, because of the drastic changes that they produce in the
electronic properties of the substrate and also because of related phenomena such as catalytic
promotion [144]; however, the understanding of adsorbate-induced collective excitations is
still incomplete [145152]. The excitation spectrum of collective modes in semiconductor
quantum wells has been described by several authors [153157]. These systems, which have
been grown in semiconductor heterostructures with the aid of molecular beam epitaxy [158],
7 The sound velocity of this acoustic mode is, however, close to the Fermi velocity of the 2D surface-state band,
which is typically a few orders of magnitude larger than the sound velocity of acoustic phonons in metals but still
about three orders of magnitude smaller than the velocity of light.
J M Pitarke et al
z=0
Figure 1. Two semi-infinite media with dielectric functions 1 and 2 separated by a planar interface
at z = 0.
form a nearly ideal free-electron gas and have been, therefore, a playground on which to test
existing many-body theories [159, 160].
Major reviews on the theory of collective electronic excitations at metal surfaces have
been given by Ritchie [161], Feibelman [162] and Liebsch [163]. Experimental reviews
are also available, which focus on high-energy EELS experiments [164], surface plasmons
on smooth and rough surfaces and on gratings [165] and angle-resolved low-energy EELS
investigations [166, 167]. An extensive review on plasmons and magnetoplasmons in
semiconductor heterostructures has been given recently given by Kushwaha [168].
This review will focus on a unified theoretical description of the many-body dynamical
electronic response of solids, which underlines the existence of various collective electronic
excitations at metal surfaces, such as the conventional surface plasmon, multipole plasmons
and the acoustic surface plasmon. We also review existing calculations, experimental
measurements and some of the most recent applications including particlesolid interactions,
scanning transmission electron microscopy and surface-plasmon based photonics, i.e.
plasmonics.
2. Surface-plasmon polariton: classical approach
2.1. Semi-infinite system
2.1.1. The surface-plasmon condition. We consider a classical model consisting of two
semi-infinite nonmagnetic media with local (frequency-dependent) dielectric functions 1 and
2 separated by a planar interface at z = 0 (see figure 1). The full set of Maxwells equations
in the absence of external sources can be expressed as follows [169]:
H i = i
1
Ei ,
c t
(2.1)
Ei =
1
Hi ,
c t
(2.2)
(i Ei ) = 0
(2.3)
and
Hi = 0,
(2.4)
where the index i describes the media: i = 1 at z < 0, and i = 2 at z > 0.
Solutions of equations (2.1)(2.4) can generally be classified into s-polarized and ppolarized electromagnetic modes, the electric field E and the magnetic field H being parallel to
the interface, respectively. For an ideal surface, if waves are to be formed that propagate along
the interface there must necessarily be a component of the electric field normal to the surface.
Hence, s-polarized surface oscillations (whose electric field E is parallel to the interface) do
not exist; instead, we seek conditions under which a travelling wave with the magnetic field H
parallel to the interface (p-polarized wave) may propagate along the surface (z = 0), with the
fields tailing off into the positive (z > 0) and negative (z < 0) directions. Choosing the x-axis
along the propagating direction, we write
(2.5)
Ei = (Eix , 0, Eiz ) ei |z| ei(qi xt)
and
(2.6)
Hi = (0, Eiy , 0) ei |z| ei(qi xt) ,
where qi represents the magnitude of a wave vector that is parallel to the surface. Introducing
equations (2.5) and (2.6) into equations (2.1)(2.4), one finds
(2.7)
i 1 H1y = + 1 E1x ,
c
i 2 H2y = 2 E2x
(2.8)
c
and

2
(2.9)
i = qi2 i 2 .
c
The boundary conditions imply that the component of the electric and magnetic fields
parallel to the surface must be continuous. Using equations (2.7) and (2.8), one writes the
following system of equations:
2
1
H1 y +
H2 = 0
(2.10)
1
2 y
and
(2.11)
H1y H2y = 0,
which has a solution only if the determinant is zero, i.e.
1 2
+
= 0.
(2.12)
1 2
This is the surface-plasmon condition.
From the boundary conditions also follows the continuity of the 2D wave vector q entering
equation (2.9), i.e. q1 = q2 = q. Hence, the surface-plasmon condition (equation (2.12)) can
also be expressed as follows [170]:

1 2
,
(2.13)
q() =
c 1 + 2
where /c represents the magnitude of the light wave vector. For a metaldielectric interface
with the dielectric characterized by 2 , the solution (q) of equation (2.13) has slope equal to
c/ 2 at the point q = 0 and is a monotonic increasing function of q, which is always smaller
than c q/ 2 and for large q is asymptotic to the value given by the solution of
(2.14)
1 + 2 = 0.
This is the nonretarded surface-plasmon condition (equation (2.12) with 1 = 2 = q), which
is valid as long as the phase velocity /q is much smaller than the speed of light.
10
J M Pitarke et al
30
25
(eV)
20
15
10
0.01
0.02
-1
q ( )
Figure 2. The solid lines represent the solutions of equation (2.16) with p = 15 eV: the dispersion
of light in the solid (upper line) and the surface-plasmon polariton (lower line). In the retarded
region (q < s /c), the surface-plasmon polariton dispersion curve approaches the light line
= cq ( ). At short wave lengths (q s /c), the surface-plasmon
polariton approaches
asymptotically the nonretarded surface-plasmon frequency s = p / 2 (- - - -).
2.1.2. Energy dispersion.

2 = 1 and [171]
1 = 1
In the case of a Drude semi-infinite metal in vacuum, one has

p2
( + i)
being a positive infinitesimal. Hence, in this case equation (2.13) yields

2 p2
.
q() =
c 22 p2
(2.15)
(2.16)
We have represented in figure 2 by solid lines the dispersion relation of equation (2.16),
together with the light line = c q (dotted line). The upper solid line represents the dispersion
of light in the solid. The lower solid line is the surface-plasmon polariton

2 (q) = p2 /2 + c2 q 2 p4 /4 + c4 q 4 ,
(2.17)
which in the retarded region (where q < s /c) couples with the free electromagnetic field
and in the nonretarded
limit (q s /c) yields the classical nondispersive surface-plasmon
frequency s = p / 2.
We note that the wave vector q entering the dispersion relation of equation (2.17) (lower
solid line of figure 2) is a 2D wave vector in the plane of the surface. Hence, if light hits the
surface in an arbitrary direction the external radiation dispersion line will always lie somewhere
between the light line c q and the vertical line, in such a way that it will not intersect the surfaceplasmon polariton line, i.e. light incident on an ideal surface cannot excite surface plasmons.
Nevertheless, there are two mechanisms that allow external radiation to be coupled to surfaceplasmon polaritons: surface roughness or gratings, which can provide the requisite momentum
via umklapp processes [72], and attenuated total reflection (ATR) which provides the external
radiation with an imaginary wave vector in the direction perpendicular to the surface [73, 74].
11
Figure 3. Schematic representation of the electromagnetic field associated with a surface-plasmon

polariton propagating along a metaldielectric interface. The field strength Ei (see equation (2.5))
decreases exponentially with the distance |z| from the surface, the decay constant i being given by
equation (2.18). + and represent the regions with lower and higher electron density, respectively.
400
vacuum
300
li (A)
1/q
200
metal
100
0.005
0.01
-1
q ( )
0.015
0.02
Figure 4. Attenuation length li = 1/i , versus q, as obtained from equation (2.18) at the surfaceplasmon polariton condition (equation (2.17)) for a Drude metal in vacuum. 1 has been taken to
be of the form of equation (2.15) with p = 15 eV and 2 has been set up to unity. The dotted line
represents the large-q limit of both l1 and l2 , i.e. 1/q.
2.1.3. Skin depth. Finally, we look at the spatial extension of the electromagnetic field
associated with the surface-plasmon polariton (see figure 3). Introducing the surface-plasmon
condition of equation (2.13) into equation (2.9) (with q1 = q2 = q), one finds the following
expression for the surface-plasmon decay constant i perpendicular to the interface:
i =
c
i2
,
1 + 2
(2.18)
which allows to define the attenuation length li = 1/i at which the electromagnetic field
falls to 1/e. Figure 4 shows li as a function of the magnitude q of the surface-plasmon
polariton wave vector for a Drude metal [1 of equation (2.15)] in vacuum (2 = 1). In the
vacuum side of the interface, the attenuation length is over the wavelength involved (l2 > 1/q),
whereas the attenuation length into the metal is determined at long-wavelengths (q 0) by
the so-called skin depth. At large q (where the nonretarded surface-plasmon condition of
equation (2.14) is fulfilled), the skin depth is li 1/q thereby leading to a strong concentration
of the electromagnetic surface-plasmon field near the interface.
12
J M Pitarke et al
2.2. Thin films

Thin films are also known to support surface collective oscillations. For this geometry, the
electromagnetic fields of both surfaces interact in such a way that the retarded surface-plasmon
condition of equation (2.12) splits into two new conditions (we only consider nonradiative
surface plasmons), depending on whether electrons in the two surfaces oscillate in phase or
not. In the case of a thin film of thickness a and dielectric function 1 in a medium of dielectric
function 2 , one finds [165]:
1
2
=0
(2.19)
+
1 tanh(1 a/2) 2
and
1
2
+
= 0.
1 coth(1 a/2) 2
(2.20)
Instead, if the film is surrounded by dielectric layers of dielectric constant 0 and equal thickness
t on either side, one finds
1
0
=0
(2.21)
+
1 tanh(1 a/2) 0
and
1
0
= 0,
+
1 coth(1 a/2) 0
(2.22)
where
=
1 e20 t
,
1 + e20 t
(2.23)
=
2 0 0 2
2 0 + 0 2
(2.24)
with
and
2
.
(2.25)
c2
Electron spectrometry measurements of the dispersion of the surface-plasmon polariton in
oxidized Al films were reported byPettit et al [172], spanning the energy range from the shortwavelength limit where p / 2 all the way to the long-wavelength limit where c q.
The agreement between the experimental measurements and the prediction of equations (2.21)
(2.25) (with a Drude dielectric function for the Al film and a dielectric constant 0 = 4 for the
surrounding oxide) is found to be very good, as shown in figure 5.
In the nonretarded regime (q s /c), where 1 = 2 = q, equations (2.19) and (2.20)
take the form
1 + 2
= eqa ,
(2.26)
1 2
0 =
q 2 0
which for a Drude thin slab (1 of equation (2.15)) in vacuum (2 = 1) yields [1]
p
= (1 eqa )1/2 .
(2.27)
2
This equation has two limiting cases, as discussed by Ferrell [68]. At short wavelengths
(qa 1), the surface waves become decoupled
and each surface sustains independent
oscillations at the reduced frequency s = p / 2 characteristic of a semi-infinite electron gas
13
(eV)
10
0
0
0.01
0.02
0.03
-1
q ( )
Figure 5. Dispersion (q) of the surface-plasmon polariton of an Al film of thickness a = 120

surrounded by dielectric layers of equal thickness t = 40 . The solid lines represent the result
obtained from equations (2.21)(2.25) with 2 = 1, 0 = 4 and a frequency-dependent Drude
dielectric function 1 (see equation (2.15)) with p = 15 eV and = 0.75 eV8 . The solid
circles represent the electron spectrometry measurements reported byPetit et al [172]. The
dashed line represents the nonretarded surface-plasmon frequency p / 5, which is the solution
of equation (2.14) with 2 = 4 and a Drude dielectric function 1 . The dotted line represents the
light line = cq.
with a single plane boundary. At long wavelengths (qa 1), there are normal oscillations at
p and tangential 2D oscillations at
2D = (2naq)1/2 ,
(2.28)
which were later discussed by Stern [173] and observed in artificially structured
semiconductors [174] and more recently in a metallic surface-state band on a silicon
surface [175].
3. Nonretarded surface plasmon: simplified models
The classical picture leading to the retarded equation (2.12) and nonretarded equation (2.14)
ignores both the nonlocality of the electronic response of the system and the microscopic spatial
distribution of the electron density near the surface. This microscopic effects can generally be
ignored at long wavelengths where q qF ; however, as the excitation wavelength approaches
atomic dimensions nonlocal effects can be important.
As nonlocal effects can generally be ignored in the retarded region where q < s /c (since
s /c qF ), here we focus our attention on the nonretarded regime where s /c < q. In
this regime and in the absence of external sources, the -components of the time-dependent
electric and displacement fields associated with collective oscillations at a metal surface satisfy
8 Although in the case of an ideal damping-free electron gas the quantity entering the Drude dielectric function
of equation (2.15) should be a positive infinitesimal, a phenomenological finite parameter is usually introduced in
order to account for the actual electron damping occurring in real solids.
14
J M Pitarke et al
the quasi-static Maxwells equations

E(r, ) = 4 n(r, ),
(3.1)
or, equivalently,
2 (r, ) = 4 n(r, )
(3.2)
D(r, ) = 0,
(3.3)
and
n(r, ) being the fluctuating electron density associated with the surface plasmon and (r, )
being the component of the time-dependent scalar potential.
3.1. Planar surface plasmon
3.1.1. Classical model. In the classical limit, we consider two semi-infinite media with local
(frequency-dependent) dielectric functions 1 and 2 separated by a planar interface at z = 0, as
in section 2.1 (see figure 1). In this case, the fluctuating electron density n(r, ) corresponds
to a delta-function sheet at z = 0:
n(r, ) = n(r , ) (z),
(3.4)
where r defines the position vector in the surface plane, and the displacement field D(r, )
takes the following form:
z < 0,
1 E(r, ),
D(r, ) =
(3.5)
2 E(r, ),
z > 0.
Introducing equation (3.4) into equation (3.2), one finds that self-sustained solutions of
Poissons equation take the form
(r, ) = 0 eqr eq|z| ,
(3.6)
where q is a 2D wave vector in the plane of the surface, and q = |q|. A combination of
equations (3.3), (3.5) and (3.6) with E(r, ) = (r, ) yields the nonretarded surfaceplasmon condition of equation (2.14), i.e.
1 + 2 = 0.
(3.7)
3.1.2. Nonlocal corrections. Now we consider a more realistic jellium model of the solid
surface consisting of a fixed semi-infinite uniform positive background at z 0 plus a
neutralizing nonuniform cloud of interacting electrons. Within this model, there is translational
invariance in the plane of the surface; hence, we can define 2D Fourier transforms E(z; q, )
and D(z; q, ), the most general linear relation between them being

D(z; q, ) = dz (z, z ; q, ) E(z ; q, ),
(3.8)
where the tensor (z, z ; q, ) represents the dielectric function of the medium.
In order to avoid an explicit calculation of (z, z ; q, ), one can assume that far from
the surface and at low wave vectors (but still in the nonretarded regime, i.e. s /c < q < qF )
equation (3.8) reduces to an expression of the form of equation (3.5):
z < z1 ,
1 E(z; q, ),
D(z; q, ) =
(3.9)
2 E(z; q, ),
z > z2 ,
15
where z1 0 and z2 0. Equations (3.2) and (3.3) with E(r, ) = (r, ) then yield
the following integration of the field components Ez and Dx in terms of the potential (z) at
z1 and z2 [where it reduces to the classical potential of equation (3.6)]:
z2
dz Ez (z; q, ) = (z2 ; q, ) (z1 ; q, )
(3.10)
z1
and

i
z2
dz Dz (z; q, ) = 2 (z2 ; q, ) 1 (z1 ; q, ).
(3.11)
z1
Neglecting quadratic and higher-order terms in the wave vector, equations (3.10) and
(3.11) are found to be compatible under the surface-plasmon condition [87]
1 + 2
= q[d () d ()],
(3.12)
1 2
d () and d () being the so-called d-parameters introduced by Feibelman [162]:

d
d
d () = dz z Ez (z, ) / dz Ez (z, ) = dz z n(z, ) / dz n(z, )
dz
dz
and

d () =
d
dz z Dz (z, ) /
dz

dz
d
Dz (z, ),
dz
(3.13)
(3.14)
where Ez (z, ), Dx (z, ) and n(z, ) represent the fields and the induced density evaluated
in the q 0 limit.
For a Drude semi-infinite metal in vacuum [2 = 1 and equation (2.15) for 1 ], the
nonretarded surface-plasmon condition of equation (3.12) yields the nonretarded dispersion
relation
= s {1 qRe[d (s ) d (s )]/2 + . . .},
(3.15)
where s is Ritchies frequency: s = p / 2. For neutral jellium surfaces, d () coincides

with the jellium edge and the linear coefficient of the surface-plasmon dispersion (q),
therefore, only depends on the position d (s ) of the centroid of the induced electron density
at s (see equation (3.13)) with respect to the jellium edge.
3.1.3. Hydrodynamic approximation. In a hydrodynamic model, the collective motion of
electrons in an arbitrary inhomogeneous system is expressed in terms of the electron density
n(r, t) and the hydrodynamical velocity v(r, t), which assuming irrotational flow we express
as the gradient of a velocity potential (r, t) such that v(r, t) = (r, t). First of all, one
writes the basic hydrodynamic Blochs equations (the continuity equation and the Bernoullis
equation) in the absence of external sources [176]:
d
n(r, t) = [n(r, t) (r, t)]
dt
(3.16)
and
1
G[n]
d
(r, t) = |(r, t)|2 +
+ (r, t),
dt
2
n
and Poissons equation:
2 (r, t) = 4 n(r, t),
(3.17)
(3.18)
16
J M Pitarke et al
where G[n] is the internal kinetic energy, which is typically approximated by the Thomas
Fermi functional
3
G[n] =
(3 2 )2/3 [n(r, t)]5/3 .
(3.19)
10
The hydrodynamic equations (equations (3.16)(3.18)) are nonlinear equations, difficult
to solve. Therefore, one typically uses perturbation theory to expand the electron density and
the velocity potential as follows:
n(r, t) = n0 (r) + n1 (r, t) + . . .
(3.20)
(r, t) = 0 + 1 (r, t) + . . . ,
(3.21)
and
so that equations (3.16)(3.18) yield the linearized hydrodynamic equations

d
n1 (r, t) = [n0 (r) 1 (r, t)],
dt
(3.22)
n1 (r, t)
d
1 (r, t) = [(r)]2
+ 1 (r, t),
dt
n0 (r)
(3.23)
and
2 1 (r, t) = 4 n1 (r, t),
(3.24)
where n0 (r) is the unperturbed electron density and (r) = 1/3 [3 2 n0 (r)]1/3 represents the
speed of propagation of hydrodynamic disturbances in the electron system9 .
We now consider a semi-infinite metal in vacuum consisting of an abrupt step of the
unperturbed electron density at the interface, which we choose to be located at z = 0:
z 0,
n,
n0 (z) =
(3.25)
0,
z > 0.
Hence, within this model n0 (r) and (r) are constant at z 0 and vanish at z > 0.
Introducing Fourier transforms, equations (3.22)(3.25) yield the basic differential
equation for the plasma normal modes at z 0:
2 (2 p2 + 2 2 )1 (r, ) = 0
(z 0)
(3.26)
and Laplaces equation at z > 0:

2 1 (r, ) = 0
(z > 0),
(3.27)
where both n1 (r, ) and 1 (r, ) vanish. Furthermore, translational invariance in the plane
of the surface allows to introduce the 2D Fourier transform 1 (z; q, ), which according to
equation (3.26) must satisfy the following equation at z 0:
(q 2 + d2 /dz2 )[2 p2 (q 2 + d2 /dz2 )]1 (z; q, ) = 0,
(3.28)
where q represents a 2D wave vector in the plane of the surface.

Now we need to specify the boundary conditions. Ruling out exponential increase at
z and noting that the normal component of the hydrodynamical velocity should
9 While the use of the ThomasFermi functional G[n] of equation (3.19), which assumes static screening, predicts
= 1/3 (3 2 n0 )1/3 , the value = 3/5 (3 2 n0 )1/3 should be more appropriate when high-frequencies of the
order of the plasma frequency are involved; see, e.g. [176].
17
vanish at the interface, for each value of q one finds solutions to equation (3.28) with
frequencies [177, 178]
2 p2 + 2 q 2
(3.29)
and
1 2
(3.30)
p + 2 q 2 + q 2p2 + 2 q 2 .
2
Equations (3.29) and (3.30) represent a continuum of bulk normal modes and a surface normal
mode, respectively. At long wavelengths, where q/p 1 (but still in the nonretarded
regime where s /c < q), equation (3.30) yields the surface-plasmon dispersion relation
= p / 2 + q/2,
(3.31)
2 =
which was first derived by Ritchie [82] using Blochs equations, and later by Wagner [179] and
by Ritchie and Marusak [180] by assuming, within a Boltzmann transport-equation approach,
specular reflection at the surface.
3.2. Localized surface plasmons: classical approach
Metaldielectric interfaces of arbitrary geometries also support charge density oscillations
similar to the surface plasmons characteristic of planar interfaces. In the long-wavelength (or
classical) limit, in which the interface separates two media with local (frequency-dependent)
dielectric functions 1 and 2 , one writes
Di (r, ) = i Ei (r, ),
(3.32)
where the index i refers to the media 1 and 2 separated by the interface. In the case of
simple geometries, such as spherical and cylindrical interfaces, equations (3.1)(3.3) can be
solved explicitly with the aid of equation (3.32) to find explicit expressions for the nonretarded
surface-plasmon condition.
3.2.1. Simple geometries.
Spherical interface. In the case of a sphere of dielectric function 1 in a host medium of
dielectric function 2 , the classical (long-wavelength) planar surface-plasmon condition of
equation (3.7) is easily found to be replaced by [129]
l 1 + (l + 1) 2 = 0,
l = 1, 2, . . . ,
(3.33)
which in the case of a Drude metal sphere (1 of equation (2.15)) in vacuum (2 = 1) yields
the Mie plasmons at frequencies

l
l = p
.
(3.34)
2l + 1
Cylindrical interface. In the case of an infinitely long cylinder of dielectric function
1 in a host medium of dielectric function 2 , the classical (long-wavelength) surfaceplasmon condition depends on the direction of the electric field. For electromagnetic waves
with the electric field normal to the interface (p-polarization), the corresponding longwavelength (and nonretarded) surface-plasmon condition coincides with that of a planar
surface, i.e. [181, 182, 183]
1 + 2 = 0,
(3.35)
18
J M Pitarke et al
which for Drude cylinders (1

of equation (2.15)) in vacuum (2 = 1) yields the planar surfaceplasmon frequency s = p / 2.
For electromagnetic waves with the electric field parallel to the axis of the cylinder (spolarization), the presence of the interface does not modify the electric field and one easily
finds that only the bulk mode of the host medium is present, i.e. one finds the plasmon condition
2 = 0.
(3.36)
In some situations, instead of having one single cylinder in a host medium, an array of parallel
cylinders may be present with a filling fraction f . In this case and for electromagnetic waves
polarized along the cylinders (s-polarization), the plasmon condition of equation (3.36) must
be replaced by [184]10
f 1 + (1 f ) 2 = 0,
(3.37)
which for Drude cylinders
(1 of equation (2.15)) in vacuum (2 = 1) yields the reduced
plasmon frequency = f p .
3.2.2. Boundary-charge method. In the case of more complex interfaces, a so-called
boundary-charge method (BCM) has been used by several authors to determine numerically
the classical (long-wavelength) frequencies of localized surface plasmons. In this approach,
one first considers the -component of the time-dependent surface charge density arising from
the difference between the normal components of the electric fields inside and outside the
surface:

1
s (r, ) =
(3.38)
E(r, ) n|r=r + E(r, ) n|r=r+ ,
4
which noting that the normal component of the displacement vector (see equation (3.32)) must
be continuous yields the following expression:
1 1 2
s (r, ) =
E(r, ) n|r=r+ ,
(3.39)
4
1
where n represents a unit vector in the direction perpendicular to the interface.
An explicit expression for the normal component of the electric field at a point of medium
2 that is infinitely close to the interface (r = r+ ) can be obtained with the use of Gauss theorem.
One finds
E(r, ) n|r=r+ = n (r, ) + 2 s (r, ),
(3.40)
where (r, ) represents the scalar potential. In the absence of external sources, this potential
is entirely due to the surface charge density itself:

s (r , )
(r, ) = d2 r
.
(3.41)
|r r |
Combining equations (3.39)(3.41), one finds the following integral equation:

1 + 2
r r
2
s (r, ) d2 r
n s (r , ) = 0,
(3.42)
1 2
|r r |3
which describes the self-sustained oscillations of the system.
The boundary-charge method has been used by several authors to determine the normalmode frequencies of a cube [185, 186] and of bodies of arbitrary shape [187, 188]. More
recent applications of this method include investigations of the surface modes of channels cut
on planar surfaces [189], the surface modes of coupled parallel wires [190] and the electron
energy loss near inhomogeneous dielectrics [191,192]. A generalization of this procedure that
includes relativistic corrections has been reported as well [193].
10 It is not necessary that the cylinders are circular and the same result is found in the case of plane parallel layers
aligned along the electric field.
19
3.2.3. Composite systems: effective-medium approach. Composite systems with a large

number of interfaces can often be replaced by an effective homogeneous medium that in
the long-wavelength limit is characterized by a local effective dielectric function eff ().
Bergman [194] and Milton [195] showed that in the case of a two-component system with
local (frequency-dependent) dielectric functions 1 and 2 and volume fractions f and 1 f ,
respectively, the long-wavelength effective dielectric function of the system can be expressed
as a sum of simple poles that only depend on the microgeometry of the composite material
and not on the dielectric functions of the components:

B
,
(3.43)
eff () = 2 1 f
u m
where u is the spectral variable

u = [1 1 /2 ]1 ,
(3.44)
m are depolarization factors and B are the strengths of the corresponding normal modes,
which all add up to unity:

B = 1.
(3.45)
Similarly,

1
() = 21 1 + f
eff
with

C
,
u n
C = 1.
(3.46)
(3.47)
The optical absorption and the long-wavelength energy loss of moving charged particles are
known to be dictated by the poles of the local effective dielectric function eff () and inverse
1
dielectric function eff
(), respectively. If there is one single interface, these poles are known
to coincide.
In particular, in the case of a two-component isotropic system composed of identical
inclusions of dielectric function 1 in a host medium of dielectric function 2 , the effective
dielectric function eff () can be obtained from the following relation:
(eff 2 ) E = f (1 2 ) Ein ,
(3.48)
where E is the macroscopic electric field averaged over the composite:

E = f Ein + (1 f )Eout ,
(3.49)
Ein and Eout representing the average electric field inside and outside the inclusions,
respectively11 .
Simple geometries. If there is only one mode with strength different from zero, as occurs
(in the long-wavelength limit) in the case of one single sphere or cylinder in a host medium,
equations (3.43) and (3.46) yield

1
eff () = 2 1 f
(3.50)
um
11 In the case of identical inclusions composed of an anisotropic material, as occurs in the case of an array of fullerenes
or carbon nanotubes, equation (3.48) still holds, as long as the scalar dielectric function 1 is replaced by its tensorial
counterpart. In this case, the electric fields E and Ein would have, in general, different directions.
20
J M Pitarke et al
and
1
eff
()
21

1
,
1+f
un
(3.51)
normal modes occurring, therefore, at the frequencies dictated by the following conditions:
m 1 + (1 m) 2 = 0
(3.52)
n 1 + (1 n) 2 = 0.
(3.53)
and
For Drude particles
(1 of equation
(2.15)) in vacuum (2 = 1), these frequencies are easily
found to be = m p and = n p , respectively.
Indeed, for a single 3D spherical or 2D circular12 inclusion in a host medium, an elementary
analysis shows that the electric field Ein in the interior of the inclusion is
u
Ein =
E,
(3.54)
um
where m = 1/D, D representing the dimensionality of the inclusions, i.e. D = 3 for spheres
and D = 2 for cylinders. Introduction of equation (3.54) into equation (3.48) leads to an
effective dielectric function of the form of equation (3.50) with m = 1/D, which yields (see
equation (3.52)) the surface-plasmon condition dictated by equation (3.33) with l = 1 in the
case of spheres (D = 3) and the surface-plasmon condition of equation (3.35) in the case of
cylinders (D = 2). This result indicates that in the nonretarded long-wavelength limit (which
holds for wave vectors q such that s a/c < q a 1, a being the radius of the inclusions) both
the absorption of light and the energy-loss spectrum of a single 3D spherical or 2D circular
inclusion exhibit one single strong maximum at the dipole resonance where 1 + 22 = 0 and
1 + 2 = 0, respectively, whichfor a Drude sphere
and cylinder (1 of equation (2.15)) in
vacuum (2 = 1) yield = p / 3 and = p / 2.
In the case of electromagnetic waves polarized along one single cylinder or array of
parallel cylinders (s-polarization), the effective dielectric function of the composite is simply
the average of the dielectric functions of its constituents, i.e.
eff = f 1 + (1 f ) 2 ,
(3.55)
which can also be written in the form of equations (3.50) and (3.51), but now with m = 0
and n = f , respectively. Hence, for this polarization the absorption of light exhibits no
maxima (in the case of a dielectric host medium with constant dielectric function 2 ) and
the long-wavelength energy-loss spectrum exhibits a strong maximum at frequencies dictated
by the plasmon condition of equation (3.37), which in the case of Drude cylinders
(1 of
equation (2.15)) in vacuum (2 = 1) yields the reduced plasmon frequency = f p .

MaxwellGarnett approximation. The interaction among spherical (or circular) inclusions in
a host medium can be introduced approximately in the framework of the well-known Maxwell
Garnett (MG) approximation [129].
The basic assumption of this approach is that the average electric field Ein within a particle
located in a system of identical particles is related to the average field Eout in the medium outside
as in the case of a single isolated (noninteracting) particle, thereby only dipole interactions
being taken into account. Hence, in this approach the electric field Ein is taken to be of the form
of equation (3.54) but with the macroscopic electric field E replaced by the electric field Eout
12 In the case of p-polarized electromagnetic waves, infinitely long cylinders can be represented by 2D circular
inclusions.
21
Figure 6. Complementary systems in which the regions of plasma and vacuum are interchanged.
The top panel represents the general situation. The bottom panel represents a half-space filled with
metal and interfaced with vacuum. The surface-mode frequencies s1 and s2 of these systems
fulfil the sum rule of equation (3.58).
outside, which together with equations (3.48) and (3.49) yields the effective dielectric function
and effective inverse dielectric function of equations (3.50) and (3.51) with the depolarization
factors m = n = 1/D (corresponding to the dilute limit, where f 0) replaced by
m=
1
(1 f )
D
(3.56)
n=
1
[1 + (D 1)f ] .
D
(3.57)
and
3.2.4. Periodic structures. Over the years, theoretical studies of the normal modes of complex
composite systems had been generally restricted to mean-field theories of the MaxwellGarnett
type, which approximately account for the behaviour of localized dipole plasmons [129].
Nevertheless, a number of methods have been developed recently for a full solution of
Maxwells equations in periodic structures [196201]. The transfer matrix method has been
used to determine the normal-mode frequencies of a lattice of metallic cylinders [202] and
rods [203], a so-called on-shell method has been employed by Yannopapas et al to investigate
the plasmon modes of a lattice of metallic spheres in the low filling fraction regime [204]
and a finite difference time domain (FDTD) scheme has been adapted to extract the effective
response of metallic structures [201].
Most recently, an embedding method [199] has been employed to solve Maxwells
equations, which has allowed to calculate the photonic band structure of three- and twodimensional lattices of nanoscale metal spheres and cylinders in the frequency range of
the Mie plasmons [143]. For small filling fractions, there is a surface-plasmon polariton
which
in the nonretarded region yields the nondispersive Mie plasmon with frequency
p / D. As the filling fraction increases, a continuum of plasmon modes is found to exist
between zero frequency and the bulk metal plasmon frequency [143], which yield strong
absorption of incident light and whose energies can be tuned according to the particleparticle
separation [205].
3.2.5. Sum rules. Sum rules have played a key role in providing insight in the investigation
of a variety of physical situations. A useful sum rule for the surface modes in complementary
22
J M Pitarke et al
media with arbitrary geometry was introduced by Apell et al [206], which in the special case
of a metal/vacuum interface implies that [207]
s21 + s22 = p2 ,
(3.58)
where s1 is the surface-mode frequency of a given system and s2 represents the surface
mode of a second complementary system in which the regions of plasma and vacuum are
interchanged (see figure 6).
For example, a half-space filled with a metal of bulk plasma frequency p and interfaced
with vacuum maps into itself (see bottom panel of figure 6), and therefore equation (3.58)
yields
(3.59)
s1 = s2 = p / 2,
which is Ritchies frequency of plasma oscillations at a metal/vacuum planar interface.
Other examples are a Drude metal sphere in vacuum, which sustains localized Mie
plasmons at frequencies given by equation (3.34), and a spherical void in a Drude metal,
which shows Mie plasmons at frequencies

l+1
.
(3.60)
l = p
2l + 1
The squared surface-mode frequencies of the sphere (equation (3.34)) and the void
(equation (3.60)) add up to p2 for all l, as required by equation (3.58).
The splitting of surface modes that occurs in thin films due to the coupling of the
electromagnetic fields in the two surfaces (see equation (2.26)) also occurs in the case of
localized modes. Apell et al [206] proved a second sum rule, which relates the surface modes
corresponding to the in-phase and out-of-phase linear combinations of the screening charge
densities at the interfaces. In the case of metal/vacuum interfaces this sum rule takes the form
of equation (3.58) but now s1 and s2 being in-phase and out-of-phase modes of the same
system.
For a Drude metal film with equal and abrupt planar surfaces, the actual values of the
nonretarded s1 and s2 are those given by equation (2.27), which fulfil the sum rule dictated
by equation (3.58). For a spherical fullerene molecule described by assigning a Drude dielectric
function to every point between the inner and outer surfaces of radii r1 and r2 , one finds the
following frequencies for the in-phase and out-of-phase surface modes [208]:

p2
1
2
2l+1
s =
,
(3.61)
1
1 + 4l(l + 1)(r1 /r2 )
2
2l + 1
also fulfilling the sum rule of equation (3.58).
Another sum rule has been reported recently [202, 203], which relates the frequencies
of the modes that can be excited by light (as dictated by the poles of the effective dielectric
function of equation (3.43)) and those modes that can be excited by moving charged particles (as
dictated by the poles of the effective inverse dielectric function of equation (3.46)). Numerical
calculations for various geometries have shown that the depolarization factors m and n
entering equations (3.43) and (3.46) satisfy the relation [202, 203]
n = 1 (D 1) m ,
(3.62)
where D represents the dimensionality of the inclusions.

Furthermore, combining equations (3.50) and (3.51) (and assuming, therefore, that only
dipole interactions are present) with the sum rule of equation (3.62) yields equations (3.56)
and (3.57), i.e. the MG approximation. Conversely, as long as multipolar modes contribute
to the spectral representation of the effective response (see equations (3.43) and (3.46)), the
23
strength of the dipolar modes decreases (see equations (3.45) and (3.47)) and a combination
of equations (3.43) and (3.46) with equation (3.62) leads to the conclusion that the dipolar
resonances must necessarily deviate from their MG counterparts dictated by equations (3.56)
and (3.57). That a nonvanishing contribution from multipolar modes appears together with a
deviation of the frequencies of the dipolar modes with respect to their MG counterparts was
shown explicitly in [202].
4. Dynamical structure factor
The dynamical structure factor S(r, r ; ) represents a key quantity in the description of both
single-particle and collective electronic excitations in a many-electron system [209]. The rate
for the generation of electronic excitations by an external potential, the inelastic differential
cross section for external particles to scatter in a given direction, the inelastic lifetime of excited
hot electrons, the so-called stopping power of a many-electron system for moving charged
particles and the ground-state energy of an arbitrary many-electron system (which is involved
in, for example the surface energy and the understanding of Van der Waals interactions) are
all related to the dynamical structure factor of the system.
The dynamical structure factor, which accounts for the particle-density fluctuations of the
system, is defined as follows:

0n (r1 ) n0 (r2 ) ( En + E0 ).
(4.1)
S(r, r ; ) =
n
Here, n0 (r) represent matrix elements, taken between the many-particle ground state |0 of
energy E0 and the many-particle excited state |n of energy En , of the operator (r)
n0 (r),
where (r)
is the electron-density operator [210]:
(r)
=
N

ri ),
(r
(4.2)
i=1
with and ri describing the Dirac-delta operator and electron coordinates, respectively, and
n0 (r) represents the ground-state electron density, i.e.
n0 (r) =< 0 |(r)|

0 >.
(4.3)
The many-body ground and excited states of a many-electron system are unknown and
the dynamical structure factor is, therefore, difficult to calculate. Nevertheless, one can use
the zero-temperature limit of the fluctuationdissipation theorem [211], which relates the
dynamical structure factor S(r, r ; ) to the dynamical density-response function (r, r ; )
of linear-response theory. One writes,

S(r, r ; ) = Im (r, r ; ) (),
(4.4)
where represents the normalization volume and (x) is the Heaviside step function.
5. Density-response function
Take a system of N interacting electrons exposed to a frequency-dependent external potential
ext (r, ). Keeping terms of first order in the external perturbation and neglecting retardation
effects, time-dependent perturbation theory yields the following expression for the induced
electron density [210]:

n(r, ) = dr (r, r ; ) ext (r , ),
(5.1)
24
J M Pitarke et al
where (r, r ; ) represents the so-called density-response function of the many-electron

system:

1
1
,
(5.2)
n0
(r)n0 (r )
(r, r ; ) =
E0 E n + h
( + i) E0 + En + h
( + i)
n
being a positive infinitesimal.
The imaginary part of the true density-response function of equation (5.2), which accounts
for the creation of both collective and single-particle excitations in the many-electron system,
is known to satisfy the so-called f -sum rule:

d Im (r, r ; ) = n0 (r)(r, r ) ,
(5.3)
with n0 (r) being the unperturbed ground-state electron density of equation (4.3).
5.1. Random-phase approximation (RPA)
In the so-called random-phase or, equivalently, time-dependent Hartree approximation, the
electron density n(r, ) induced in an interacting electron system by a small external potential
ext (r, ) is obtained as the electron density induced in a noninteracting Hartree system (of
electrons moving in a self-consistent Hartree potential) by both the external potential ext (r, )
and the induced potential

H
(r, ) = dr v(r, r ) n(r , ),
(5.4)
with v(r, r ) representing the bare Coulomb interaction. Hence, in this approximation one
writes

n(r, ) = dr 0 (r, r ; )

ext

(r , ) + dr v(r , r ) n(r , ) ,
(5.5)
which together with equation (5.1) yields the following Dyson-type equation for the interacting
density-response function:

(5.6)
(r, r ; ) = (r, r ; ) + dr1 dr2 0 (r, r1 ; )v(r1 , r2 ) (r2 , r ; ),
where 0 (r, r ; ) denotes the density-response function of noninteracting Hartree electrons:
0 (r, r ; ) =
i (r)j (r)j (r )i (r )
2
(fi fj )
.
i,j
j + i + i
(5.7)
Here, fi are FermiDirac occupation factors, which at zero temperature take the form
fi = (F i ), F being the Fermi energy and the single-particle states and energies i (r)
and i are the eigenfunctions and eigenvalues of a Hartree Hamiltonian, i.e.

1
(5.8)
2 + vH [n0 ](r) i (r) = i i (r),
2
where

vH [n0 ](r) = v0 (r) +
25
dr v(r, r )n0 (r ),
(5.9)
with v0 (r) denoting a static external potential and n0 (r) being the unperturbed Hartree electron
density:
n0 (r) =
N

|i (r)|2 .
(5.10)
i=1
5.2. Time-dependent density-functional theory

In the framework of time-dependent density-functional theory (TDDFT) [212], the exact
density-response function of an interacting many-electron system is found to obey the following
Dyson-type equation:

0

(r, r ; ) = (r, r ; ) + dr1 dr2 0 (r, r1 ; )
{v(r1 , r2 ) + fxc [n0 ](r1 , r2 ; )} (r2 , r ; ).
(5.11)
Here, the noninteracting density-response function 0 (r, r ; ) is of the form of equation (5.7)
but with the single-particle states and energies i (r) and i being now the eigenfunctions and
eigenvalues of the KohnSham DFT, i.e:

1
(5.12)
2 + vKS [n0 ](r) i (r) = i i (r),
2
where
with
vKS [n0 ](r) = vH [n0 ](r) + vxc [n0 ](r)
(5.13)

Exc [n]
.
vxc [n0 ](r) =
n(r) n=n0
(5.14)
Exc [n] represents the unknown XC energy functional and n0 (r) denotes the exact unperturbed
electron density of equation (4.3), which the DFT shows to coincide with that of equation (5.10)
but with the Hartree eigenfunctions i (r) of equation (5.8) being replaced by their KohnSham
counterparts of equation (5.12). The XC kernel fxc [n0 ](r, r ; ) denotes the Fourier transform
of

vxc [n](r, t)
,
(5.15)
fxc [n0 ](r, t; r , t ) =
n(r , t )
n=n0
with vxc [n](r, t) being the exact time-dependent XC potential of TDDFT.

If short-range XC effects are ignored altogether by setting the unknown XC potential
vxc [n0 ](r) and XC kernel fxc [n0 ](r, r ; ) equal to zero, the TDDFT density-response function
of equation (5.11) reduces to the RPA equation (5.6).
5.2.1. The XC kernel. Along the years, several approximations have been used to evaluate
the unknown XC kernel of equation (5.15).
26
J M Pitarke et al
Random-phase approximation (RPA). Nowadays, one usually refers to the RPA as the result
of simply setting the XC kernel fxc [n0 ](r, r ; ) equal to zero:
fxcRPA [n0 ](r, r ; ) = 0,
(5.16)
but still using in equations (5.7) and (5.10) the full single-particle states and energies i (r) and
i of DFT (i.e. the solutions of equation (5.12)) with vxc [n0 ](r) set different from zero. This
is sometimes called the DFT-based RPA.
Adiabatic local-density approximation (ALDA). In this approximation, also called timedependent local-density approximation (TDLDA) [213], one assumes that both the unperturbed
n0 (r) and the induced n(r, ) electron densities vary slowly in space and time and, therefore,
one replaces the dynamical XC kernel by the long-wavelength (Q 0) limit of the static XC
kernel of a homogeneous electron gas at the local density:

d2 [nxc (n)]
ALDA

(r r ),
(5.17)
fxc [n0 ](r, r ; ) =

dn2
n=n0 (r)
where xc (n) is the XC energy per particle of a homogeneous electron gas of density n.
PGG and BPG. In the spirit of the optimized effective-potential method [214], Petersilka,
Gossmann and Gross (PGG) [215] derived the following frequency-independent exchangeonly approximation for inhomogeneous systems:

2

2
i fi i (r) i (r )
PGG

,
(5.18)
fx [n0 ](r, r ; ) =
n0 (r)n0 (r )
|r r |
where i (r) denote the solutions of the KohnSham equation (5.12).
More recently, Burke, Petersilka, and Gross (BPG) [216] devised a hybrid formula for the
XC kernel, which combines expressions for symmetric and antisymmetric spin orientations
from the exchange-only PGG scheme and the ALDA. For an unpolarized many-electron system,
one writes [216]

1 ,PGG
(5.19)
fxcBPG [n0 ](r, r ; ) =
+ fxc,ALDA ,
fxc
2
where fxc and fxc represent the XC kernel for electrons with parallel and antiparallel spin,
respectively.
Average approximation. The investigation of short-range XC effects in solids has been
focused to a great extent onto the simplest possible many-electron system, which is the
homogeneous electron gas. Hence, recent attempts to account for XC effects in inhomogeneous
systems have adopted the following approximation [217, 218]:
|r r |; ),
fxcav [n0 ](r, r ; ) = fxchom (n;
(5.20)

where n represents a function of the electron densities at points r and r , typically the
arithmetical average

1
(5.21)
n0 (r) + n0 (r ) ,
n =
2
|r r |; ) denotes the XC kernel of a homogeneous electron gas of density n,
and fxchom (n;

whose 3D Fourier transform fxchom (n;
Q, ) is directly connected to the so-called local-field
factor G(n;
Q, ):
4
Q, ) = 2 G(n;
Q, ).
(5.22)
fxchom (n;
Q
27
In the ALDA, one writes

G
ALDA
(n;
Q, ) = G(n;
Q 0, = 0)

2
Q d2 [nxc (n)]
=
,

4
dn2
(5.23)
n=n
which in combination with equations (5.20)(5.22) yields the ALDA XC kernel of

equation (5.17). However, more accurate nonlocal dynamical expressions for the local-field
factor G(n;
Q, ) are available nowdays, which together with equations (5.20)(5.22) should
yield an accurate (beyond the ALDA) representation of the XC kernel of inhomogeneous
systems.
During the last decades, much effort has gone into the determination of the static local-field
factor Gstatic (n;
Q) = G(n;
Q, = 0) [219226], the most recent works including diffusion
Monte Carlo (DMC) calculations [227,228] and the parametrization of the DMC data of [228]
given by Corradini et al [229]:
2 + BQ
2 /(g + Q
2) + Q
4 e Q 2 ,
Gstatic (n;
Q) = C Q
(5.24)
= Q/qF , and the parameters B, C, g, and are the dimensionless functions of n

where Q
listed in [229].
Calculations of the frequency dependence of the local-field factor G(n;
Q, ) have been
carried out mainly in the limit of long wavelengths (Q 0) [230235], but work has also
been done for finite wave vectors [236239].
6. Inverse dielectric function
In the presence of a many-electron system, the total potential (r, ) of a unit test charge at
point r that is exposed to the external potential ext (r, ) can be expressed in the following
form:
(r, ) = ext (r, ) + H (r, ),
(6.1)
where (r, ) represents the induced potential of equation (5.4). Using equations (5.1) and
(5.4), the total potential (r, ) of equation (6.1) is easily found to take the following form:

(r, ) = dr 1 (r, r ; ) ext (r , ),
(6.2)
where
1 (r, r ; ) = (r r ) +
dr v(r r ) (r , r ; ).
(6.3)
This is the so-called inverse longitudinal dielectric function of the many-electron system, whose
poles dictate the occurrence of collective electronic excitations and which can be evaluated in
the RPA or in the framework of TDDFT from the knowledge of the density-response function
of equations (5.6) and (5.11), respectively.
Some quantities, such as the optical absorption and the electron energy loss of charged
particles moving in arbitrary inhomogeneous media, can be described by the so-called effective
1
inverse dielectric function eff
(Q, ), which is defined as a 3D Fourier transform of the inverse
1

dielectric function (r, r ; ):

1

1
eff (Q, ) =
dr
dr eiQ(rr ) 1 (r, r ; )
(6.4)

28
J M Pitarke et al
and which at long wavelengths (Q 0) should take the form of equation (3.46)13 .
In particular, in the case of a homogeneous system and in the classical long-wavelength
limit, where the total potential (r, ) of a unit test charge at point r only depends on the
external potential ext (r, ) at that point, the inverse dielectric function takes the following
form:
1 (r, r ; ) = 1 () (r r ),
(6.5)
which in combination with equation (6.2) yields the classical formula

(r, ) = ext (r, )/(),
(6.6)
() representing the so-called local dielectric function of the medium.

7. Screened interaction
Another key quantity in the description of electronic excitations in a many-electron system,
which also dictates the occurrence of collective electronic excitations, is the frequencydependent complex screened interaction W (r, r ; ). This quantity yields the total potential
(r, ) of a unit test charge at point r in the presence of an external test charge of density
next (r , ) at point r :

(r, ) = dr W (r, r ; ) next (r , ).
(7.1)
The potential ext (r, ) due to the external test charge density next (r, ) is simply

ext (r, ) = dr v(r, r ) next (r , ).
(7.2)
Hence, a comparison of equations (6.2) and (7.1) yields

W (r, r ; ) = dr 1 (r, r ; ) v(r , r ),
(7.3)
and using equation (6.3):

W (r, r ; ) = v(r, r ) + dr1 dr2 v(r, r1 ) (r1 , r2 ; ) v(r2 , r ).
(7.4)
From equations (6.4) and (7.3) one easily finds the following representation of the effective
inverse dielectric function:

1

1
eff (Q, ) =
dr dr eiQ(rr ) W (r, r ; ),
(7.5)
vQ
where vQ = 4/Q2 denotes the 3D Fourier transform of the bare Coulomb interaction v(r, r ).
7.1. Classical model
In a classical model consisting of two homogeneous media characterized by local (frequencydependent) dielectric functions 1 and 2 and separated by an interface of arbitrary geometry,
the total potential at each medium is simply given by equation (6.6) and is, therefore, a solution
of Poissons equation
4 ext
2 (r, ) =
n (r, ),
(7.6)
i ()
13 In the long-wavelength (Q 0) limit, longitudinal and transverse dielectric functions with the same polarization
coincide.
29
i () being 1 or 2 depending on whether the point r is located in medium 1 or in medium 2,

respectively. Hence, the screened interaction W (r, r ; ) entering equation (7.1) is a solution
of the following equation:
2 W (r, r ; ) =
4
(r r ).
i ()
(7.7)
For simple geometries, such as the planar, spherical and cylindrical interfaces,
equation (7.7) can be solved explicitly by imposing the ordinary boundary conditions of
continuity of the potential and the normal component of the displacement vector at the interface.
7.1.1. Planar surface. In the case of two semi-infinite media with local (frequencydependent) dielectric functions 1 (at z < 0) and 2 (at z > 0) separated by a planar interface
at z = 0 (see figure 1), there is translational invariance in two directions, which we take to be
normal to the z axis. Hence, one can define the Fourier transform W (z, z ; q, ), q being the
magnitude of a 2D wave vector in the plane of the interface, and imposing the ordinary boundary
conditions of continuity of the potential and the normal component of the displacement vector
at the interface, one finds:

q|zz |
q(|z|+|z |)
e
/1 ,
+
g
e
z < 0, z < 0,

2

2 g eq|zz | /(1 2 ),
z< < 0, z> > 0,
(7.8)
W (z, z ; q, ) =
q

q|zz |
e
g eq(|z|+|z |) /2 ,
z > 0, z > 0,
where z< (z> ) is the smallest (largest) of z and z , and g is the classical surface-response
function:
g() =
1 () 2 ()
,
1 () + 2 ()
(7.9)
or, equivalently,
g() =
n
,
un
(7.10)
where u is the spectral variable of equation (3.44) and n = 1/2.

An inspection of equations (7.8) and (7.9) shows that the screened interaction
W (z, z ; q, ) has poles at the classical bulk- and surface-plasmon conditions dictated by
i = 0 and by equation (3.7), respectively.
7.1.2. Spheres. In the case of a sphere of radius a and local (frequency-dependent) dielectric
function 1 embedded in a host medium of local (frequency-dependent) dielectric function 2 ,
we first expand the screened interaction W (r, r ; ) in spherical harmonics:
W (r, r ; ) =

l,m
Wl (r, r ; ) Yl,m
() Yl,m ( ),
2l + 1
(7.11)
30
J M Pitarke et al
and we then derive the coefficients of this expansion by imposing the boundary conditions.
One finds [240]14 :
< l

(r r )l
(r )
+ (l + 1) gl 2l+1 /1 ,
r, r < a,
> )l+1
(r
a
(r < )l

Wl (r, r ; ) = (l + 1) gl > l+1 /(1 2 ),
(7.12)
r < < a, r > > a,
(r
)
(r < )l
a 2l+1
/2 ,
l gl
r, r > a,
(r > )l+1
(r r )l+1
where r < (r > ) is the smallest (largest) of r and r , and
gl () =
1 () 2 ()
,
l 1 () + (l + 1) 2 ()
(7.13)
or, equivalently,
gl () =
nl
,
u nl
(7.14)
with u being the spectral variable of equation (3.44) and

l
.
(7.15)
2l + 1
As in the case of the planar surface, the screened interaction of equations (7.11)(7.13)
has poles at the classical bulk- and surface-plasmon conditions, which in the case of a single
sphere in a host medium are dictated by i = 0 and by equation (3.33), respectively.
Introducing equations (7.11)(7.13) into equation (7.5), one finds the following expression
for the effective inverse dielectric function [182]:

3
1
1
1
1
eff (Q, ) = 2 + f (1 2 ) 1 +
(2l + 1) gl jl (x) l (x) ,
(7.16)
x l=0
nl =
where
l (x) =
l jl1 (x) 1 (l + 1) jl+1 (x) 2

1 2
(7.17)
and x = Qa. Here, f represents the volume fraction filled by the sphere and jl (x) are spherical
Bessel functions of the first kind [241]. This equation represents the dilute (f 0) limit of
the effective inverse dielectric function derived by Barrera and Fuchs for a system composed
of identical interacting spheres in a host medium [242].
In the limit as Qa 1, an expansion of equation (7.16) yields

1 2
1
eff
,
(7.18)
(Q, ) = 21 1 3f
1 + 22
which is precisely the long-wavelength effective inverse dielectric function obtained in
section 3.2.3 from equations (3.48) and (3.54) with D = 3 and which admits the spectral
representation of equation (3.51) with n = 1/3. This result demonstrates the expected result
14 The direct contribution to the screened interaction is missing in the first line of equation (5) of this reference, as
pointed out in [182]. This contribution, together with part of the term represented in the second line of equation (5),
would give the bulk contribution 1 1 to equation (7) of the same reference. Hence, equation (5) of this reference
must be replaced by equations (7.11)(7.13) of the present manuscript with 2 = 1.
31
that in the limit as Qa 1 a broad beam of charged particles interacting with a single sphere
of dielectric function 1 in a host medium of dielectric function 2 can only create collective
excitations at the dipole resonance where
1 + 22 = 0 (equation (3.33) with l = 1), which for
a Drude sphere in vacuum yields = p / 3.
7.1.3. Cylinders. In the case of an infinitely long cylinder of radius a and local (frequencydependent) dielectric function 1 embedded in a host medium of local (frequency-dependent)
dielectric function 2 , we expand the screened interaction W (r, r ; ) in terms of the modified
Bessel functions Im (x) and Km (x) [241], as follows:

2
W (r, r ; ) =
dqz cos qz (z z )
0
m Wm (, ; ) cos m( ) ,
(7.19)
m=0
where z and represent the projections of the position vector along the axis of the cylinder
and in a plane perpendicular to the cylinder, respectively, qz denotes the magnitude of a wave
vector along the axis of the cylinder and mm are Neumann numbers
m = 0,
1,
m =
(7.20)
m 1.
2,

The coefficients Wm (, ; ) are then derived by imposing the boundary conditions, i.e. by
requiring that the total scalar potential and the normal component of the displacement vectors
be continuous at the interface. One finds:
Wm (, ; )

Im (qz < ) Km (qz > ) Km (x) gm Im (qz > )/Im (x) /1 ,

1
=
x Im (x) Km (x)
gm Im (qz < ) Km (qz > )/(1 2 ),

Im (qz < ) Im (x) gm Km (qz < )/Km (x) Km (qz > )/2 ,
, < a,
< < a, > > a,
, > a,
(7.21)
where x = qz a and
gm (x, ) =
Im (x) Km (x) [1 () 2 ()]

,

Im (x) Km (x) 1 () Im (x) Km (x) 2 ()
(7.22)
or, equivalently,
nm
,
u nm
with u being the spectral variable of equation (3.44) and
gm (x, ) =
nm = x Im (x) Km (x).
(7.23)
(7.24)
In the limit as x 0 (p-polarization) , the depolarization factors of equation (7.24) are

easily found to be n0 = 0 (corresponding to the plasmon condition 2 = 0)16 and nm = 1/2
15
15
As we are considering longitudinal fields, where the electric field E and the wave vector q have the same direction,
in the limit as x = qz a 0 the electric field lies in the plane perpendicular to the axis of the cylinder.
16 This mode, however, does not contribute to the effective inverse dielectric function and, therefore, to the energy
loss of moving charged particles, since when x = qz a = 0 (p-polarization) the strength of this mode is equal to zero
for all values of the total wave vector Qa (see [183]).
32
J M Pitarke et al
0.5
m=2
0.4
m=1
nm
0.3
m=0
0.2
0.1
0
0
q za
10
Figure 7. Depolarization factors nm = x Im (x) Km (x), as a function of x = qz a, for

m = 0, 1, 2, 3, 4, 5 (thin solid lines), and m = 10 (thick solid line). As m , the depolarization
factor nm equals the planar surface-plasmon value nm = 1/2 for all values of x = qz a.
(corresponding to the planar surface-plasmon condition 1 + 2 = 0) for all m = 0; in the limit

as x , equation (7.24) yields nm = 1/2 for all m. For the behaviour of the depolarization
factors nm of equation (7.24) as a function of x see figure 7. This figure shows that the
energies of all modes are rather close to the planar surface-plasmon energy (corresponding
to nm = 1/2), except for m = 0. The m = 0 mode, which corresponds to a homogeneous
charge distribution around the cylindrical surface, shifts downwards from the planar surfaceplasmon energy (n0 = 1/2) as the adimensional quantity x = qz a decreases, as occurs with
the symmetric low-energy mode in thin films (see equations (2.26) and (2.27)).
Introducing equations (7.19)(7.22) into equation (7.5), one finds the following expression
for the effective inverse dielectric function [182]:

2
1
1
1
1
m Jm (y)gm m (x, y) ,
(7.25)
eff (Q, ) = 2 + f (1 2 ) 1 + 2
x + y 2 m=0
where
m (x, y) =
Im (x) fm(1) (x, y) 1 + Km (x) fm(2) (x, y) 2

,
Im (x) Km (x) [1 2 ]
(7.26)
fm(1) (x, y) = x Jm (y) Km1 (x) + y Jm1 (y) Km (x),
(7.27)
fm(2) (x, y) = x Jm (y) Im1 (x) y Jm1 (y) Im (x),
(7.28)
x = qz a, and y = qa, qz and q representing the components of the total wave vector Q along
the axis of the cylinder and in a plane perpendicular to the cylinder, respectively. The volume
fraction filled by the cylinder is denoted by f , and Jm (x) are cylindrical Bessel functions of
the first kind [241]. A spectral representation of the effective inverse dielectric function of
equations (7.25)(7.28) was reported in [183].
In the limit as Qa 1, an expansion of equations (7.25)(7.28) yields

1 2 x 2
y2
1
1
eff (Q, ) = 2 1 f 2
,
+2
x + y 2 2
1 + 2
33
(7.29)
which admits the spectral representation of equation (3.46) with two nonvanishing spectra
strengths: C0 = x 2 /(x 2 + y 2 ) and C1 = y 2 /(x 2 + y 2 ), the corresponding depolarization factors
being n0 = 0 and n1 = 1/2, respectively.
Equation (7.29) demonstrates that in the limit as Qa 1 and for a wave vector normal
to the cylinder (x = 0), moving charged particles can only create collective excitations at the
dipole resonance where n1 = 1/2, i.e.
1 + 2 = 0, which for a Drude cylinder in vacuum
yields Ritchies frequency s = p / 2. Conversely, still in the limit as Qa 1 but for a
wave vector along the axis of the cylinder (y = 0), moving charged particles can only excite
the bulk mode of the host medium dictated by the condition u = 0 (corresponding to n0 = 0),
i.e. 2 = 0, in agreement with the discussion of section 3.2.3 (Simple geometries).
7.2. Nonlocal models: planar surface
Nonlocal effects that are absent in the classical model described above can be incorporated in
a variety of semiclassical and quantal approaches, which we here only describe for a planar
surface.
7.2.1. Hydrodynamic model.
Semiclassical hydrodynamic approach. Within a semiclassical hydrodynamic approach,
the screened interaction W (r, r ; ) (as defined in equation (7.1)) can be obtained from
the linearized hydrodynamic equations (3.22)(3.24). For a semi-infinite metal in vacuum
consisting of an abrupt step of the unperturbed electron density n0 (z) (see equation (3.25)),
we can assume translational invariance in the plane of the surface, and noting that the normal
component of the hydrodynamical velocity should vanish at the interface equations (3.22)
(3.24) yield the following expression for the 2D Fourier transform W (z, z ; q, ):
s (z) s (z )
s (z z ) + s (z + z ) 2 g
,
z < 0, z < 0,
2
s (z< ) qz>
W (z, z ; q, ) =
2
g
e
,
z< < 0, z> > 0,
q
1 s0
q|zz |

e
g eq(z+z ) ,
z > 0, z > 0,
(7.30)
where z< (z> ) is the smallest (largest) of z and z ,
s (z; q, ) =
g(q, ) =
=
( + i) eq|z| q p2 e|z|

,
( + i) p2
p2
2 2 ( + q) p2
1 2
p + 2 q 2 ( + i),
(7.31)
(7.32)
(7.33)
34
J M Pitarke et al
1/3(3 2 n0 )1/3 (as in equations (3.29) and (3.30)) and

s0 (q, ) = s (z = 0; q, ).
(7.34)
An inspection of equations (7.30)(7.34) shows that the hydrodynamic surface-response

function g(q, ) and, therefore, the hydrodynamic screened interaction W (z, z ; q, ) become
singular at the hydrodynamic surface-plasmon condition dictated by equation (3.30). We also
note that the second moment of the imaginary part of the hydrodynamic surface-response
function g(q, ) is found to be

d Im g(q, ) = 2 2 n,
(7.35)
where n represents the electron density: n = p2 /4 .

Finally, we note that in the long-wavelength (q 0) limit the hydrodynamic screened
interaction of equations (7.30)(7.34) reduces to the classical screened interaction of
equations (7.8) and (7.9) with the dielectric functions 1 and 2 being replaced by the Drude
dielectric function (equation (2.15)) and unity, respectively. The same result is also obtained
by simply assuming that the electron gas is nondispersive, i.e. by taking the hydrodynamic
speed equal to zero.
Quantum hydrodynamic approach. Within a quantized hydrodynamic model of a manyelectron system, one first linearizes the hydrodynamic Hamiltonian with respect to the induced
electron density and then quantizes this Hamiltonian on the basis of the normal modes of
oscillation (bulk and surface plasmons) corresponding to equations (3.29) and (3.30). One
finds
H = HG + H0B + H0S ,
(7.36)
where HG represents the ThomasFermi ground state of the static unperturbed electron
system [176], and H0B and H0S are free bulk and surface plasmon Hamiltonians, respectively:

1
B
H0B =
(7.37)
1/2 + Q
aQ (t)aQ (t)
q,q
z
and
H0S =

1
1/2 + qS bq (t)bq (t).
A q
(7.38)
Here, and A represent the normalization volume and the normalization area of the surface,
respectively, aQ (t) and bq (t) are BoseEinstein operators that annihilate bulk and surface
B
and qS represent the
plasmons with wave vectors Q = (q, qz ) and q, respectively, and Q
dispersion of bulk and surface plasmons:
B 2
Q = p2 + 2 Q2
(7.39)
and
1 2
(7.40)
p + 2 q 2 + q 2p2 + 2 q 2 .
2
Hence, within this approach one can distinguish the separate contributions to the imaginary
part of the hydrodynamic surface-response function g(q, ) of equation (7.32) coming from
the excitation of either bulk or surface plasmons. One finds [243]

qS
2
Im g(q, ) = Im g B (q, ) + Im g S (q, ),
(7.41)
35
where
Im g B (q, ) =
1
q
2
B
dqz ( Q
)
B
) qz2
(p2 /Q
qz4 + qz2 (q 2 + p2 / 2 ) + p4 /(4 4 )
(7.42)
and
Im g S (q, ) =
with Q =
p2
q
( qS ),
2 q + 2q qS

q 2 + qz2 and
q =

1
q + 2p2 + 2 q 2 .
2
For the second moments of Im g B (q, ) and Img S (q, ), one finds

q
d Im g B (q, ) =
2
4
q
+
2q p
0
and
(7.43)
d Im g S (q, ) =
2q
2 ,
4 q + 2q p
(7.44)
(7.45)
(7.46)
which add up to the second moment of equation (7.35).

In the limit as q 0 the bulk contribution to the so-called energy-loss function Img(q, )
(see equations (7.41)(7.44)) vanishes, and the imaginary part of both equations (7.32) and
(7.43) yields the classical result:
Im g(q, )
s ( s ),
2
(7.47)
which can also be obtained from equation (7.9) with 1 replaced by the Drude dielectric function
of equation (2.15) and 2 set equal to unity. Equation (7.47) shows that in the classical (longwavelength)
limit the energy loss is dominated by the excitation of surface plasmons of energy
s = p / 2, as predicted by Ritchie.
7.2.2. Specular-reflection model (SRM). An alternative scheme to incorporate nonlocal
effects, which has the virtue of expressing the screened interaction W (z, z ; q, ) in terms of
the dielectric function (Q, ) of a homogeneous electron gas representing the bulk material,
is the so-called specular-reflection model reported independently by Wagner [179] and by
Ritchie and Marusak [180]. In this model, the medium is described by an electron gas in
which all electrons are considered to be specularly reflected at the surface, thereby the electron
density vanishing outside.
For a semi-infinite metal in vacuum, the unperturbed electron density n0 (z) is taken to
be of the form of equation (3.25), and the SRM yields a screened interaction of the form of
equation (7.30) but with the quantities s (z; q, ) and g(q, ) being replaced by the more
general expressions:

q + dqz iqz z 1
s (z; q, ) =
e (Q, )
(7.48)
Q2
36
J M Pitarke et al
and
g(q, ) =
1 s0 (q, )
,
1 + s0 (q, )
(7.49)

with s0 (q, ) defined as in equation (7.34), and Q = q 2 + qz2 .
The inverse dielectric function 1 (Q, ) entering equation (7.48) represents the 3D
Fourier transform of the inverse dielectric function 1 (r, r ; ) of a homogeneous electron
gas. From equation (6.3), one finds
1 (Q, ) = 1 + vQ (Q, ),
(7.50)
where (Q, ) represents the 3D Fourier transform of the density-response function

(r, r ; ).
In the framework of TDDFT, one uses equation (5.11) to find
(Q, ) = 0 (Q, ) + 0 (Q, )

vQ + fxc (n;
Q, ) (Q, ),
(7.51)
with (Q, ) and fxc (n;

Q, ) being the 3D Fourier transforms of the noninteracting densityresponse function and the XC kernel of equations (5.7) and (5.15), respectively. For a
homogeneous electron gas, the eigenfunctions i (r) entering equation (5.7) are all plane
waves; thus, the integrations can be carried out analytically to yield the well-known Lindhard
function 0 (Q, ) [244]. If one sets the XC kernel fxc (n;
Q, ) equal to zero, introduction
of equation (7.51) into equation (7.50) yields the RPA dielectric function
0
RPA (Q, ) = 1 vQ 0 (Q, ),
(7.52)
which is easy to evaluate.

The RPA dielectric function RPA (Q, ) of a homogeneous electron gas can be further
approximated in the framework of the hydrodynamic scheme described in section 3.1.3. One
finds
p2
hydro (Q, ) = 1 + 2 2
,
(7.53)
Q ( + i)
which in the classical (long-wavelength) limit yields the local Drude dielectric function of
equation (2.15). Introduction of equation (7.53) into equation (7.48) yields the hydrodynamic
screened interaction of equations (7.30)(7.34).
We know from equation (7.1) that collective excitations are dictated by singularities in the
screened interaction or, equivalently, maxima in the imaginary part of this quantity. For z and
z coordinates well inside the solid (z, z ), one finds

dqz iqz (zz )
in

W (z, z ; q, ) =
vQ 1 (Q, ),
(7.54)
e
2
which in the case of the Drude dielectric function (Q, ) of equation (2.15) and for positive
frequencies ( > 0) yields
2

p ( p ) eq|zz | .
q
For z and z coordinates both outside the solid (z, z > 0), one finds
2 q|zz |

W out (z, z ; q, ) =
g(q, ) eq(z+z ) ,
e
q
which in the classical (q 0) limit and for positive frequencies ( > 0) yields
Im W in (z, z ; q, )
ImW out (z, z ; q, )
2

s ( s ) eq(z+z ) .
q
(7.55)
(7.56)
(7.57)
37
Im[-W(z,z;q,)] (a.u.)
50
25
0.5
(a.u.)
Figure 8. The solid line represents the energy-loss function, Im[W (z, z ; q, )], versus , as
obtained at z = z , q = 0.4qF , and rs = 2.07 from equation (7.54) by using the full RPA dielectric
function RPA (Q, ). The thick dashed and dotted lines represent separate contributions from the
B
B
excitation of bulk collective modes and eh pairs occurring at energies Q=q
< < Q=Q
and
c
qqF + q 2 /2, respectively. The vertical dotted line represents the energy p = 15.8 eV at
which collective oscillations would occur in a Drude metal.
Figures 8 and 9 show the energy-loss function Im[W (z, z ; q, )] that we have obtained
at z = z , q = 0.4qF and rs = 2.07 from equations (7.54) and (7.56), respectively, by using
the full RPA dielectric function RPA (Q, ). For z coordinates well inside the solid (figure 8),
instead of the single classical collective excitation at p (dotted vertical line) predicted by
equation (7.55) the RPA energy-loss spectrum (solid line) is composed of (i) a continuum of
B
B
17
bulk collective excitations (dashed line) occurring at energies Q=q
< < Q=Q
and (ii)
c
the excitation of electronhole (eh) pairs represented by a thick dotted line.
For z coordinates that are outside the surface, it had been generally believed that only
surface plasmons and eh pairs can be excited. However, it was shown explicitly in [243, 245]
that the continuum of bulk-plasmon excitations dominating the energy-loss spectrum inside
the solid (see figure 8) is still present for z coordinates outside, as shown in figure 9 for
B
z = z = F 18 . This continuum, which covers the excitation spectrum at energies Q=q
and is well separated from the lower-energy spectrum arising from the excitation of surface
plasmons and eh pairs, is accurately described by using the quantal hydrodynamic surface
energy-loss function of equation (7.42), which has been represented in figure 9 by a thick
dotted line.
Nonetheless, the main contribution to the energy-loss spectrum outside the solid comes
from the excitation of surface plasmons, which are damped by the presence of eh pairs. These
eh pair excitations are not present in the classical and hydrodynamic schemes described above,
which predict the existence of long-lived
surface plasmons at the energies represented in figure 9
by thin dotted vertical lines: s = p / 2 (see equation (7.57)) and qS of equation (7.40) (see
equations (7.43)(7.44)), respectively.
B denotes the energy of bulk plasmons with momentum Q, and Q denotes the critical momentum for which the
Q
c
B enters the electronhole pair excitation spectrum. For r = 2.07 and Q = 0.4q , the
bulk-plasmon dispersion Q
s
F
B and B are 17.6 eV and 23.6 eV, respectively.
RPA values of Q
Qc
18 The Fermi wavelength is defined as follows: = 2/q , q being the magnitude of the Fermi wave vector.
F
F
F F
17
38
J M Pitarke et al
Im[-W(z,z;q,)] (a.u.)
0.5
0.5
(a.u.)
Figure 9. The solid line represents the energy-loss function, Im[W (z, z ; q, )], versus , as
obtained at z = z = F , q = 0.4qF , and rs = 2.07 from equation (7.56) by using the full
RPA dielectric function RPA (Q, ). The thick dotted line, which is nearly indistinguishable from
the solid line covering the same part of the spectrum, represents the hydrodynamic prediction
from equation (7.42). The thin dotted vertical lines represent the energies s = 11.2 eV and
S
Q
= 15.4 eV of equation (7.40) at which long-lived surface plasmons would occur in a semiinfinite metal described by a Drude and a hydrodynamic model, respectively. Introduction of the
full RPA dielectric function RPA (Q, ) into equation (7.49) yields a maximum of the surfaceloss function Img(q, ) (and, therefore, Im[W (z, z ; q, )]) at the surface-plasmon energy
S
Q
= 16.0 eV, which is slightly larger than its hydrodynamic counterpart.
7.2.3. Self-consistent scheme. For an accurate quantal description of the electronic excitations
that can occur in a semi-infinite metal, we need to consider the true density-response function
(r, r ; ) entering equation (7.4), which is known to fulfil the f -sum rule of equation (5.3).
Jellium surface. In the case of a free-electron gas bounded by a semi-infinite positive
background of density
z 0,
n,
n+ (z) =
(7.58)
0,
z > 0,
translationally invariance in the plane of the surface allows one to define the 2D Fourier
transform W (z, z ; q, ), which according to equation (7.4) can be obtained as follows:

W (z, z ; q, ) = v(z, z ; q) + dz1 dz2 v(z, z1 ; q) (z1 , z2 ; q, ) v(z2 , z ; q),
(7.59)
where v(z, z ; q) is the 2D Fourier transform of the bare Coulomb interaction v(r, r ):
2 q|zz |
v(z, z ; q) =
e
,
(7.60)
q
and (z, z ; q, ) denotes the 2D Fourier transform of the interacting density-response function
(r, r ; ). In the framework of TDDFT, one uses equation (5.11) to find

0

(z, z ; q, ) = (z, z ; q, ) + dz1 dz2 0 (z, z1 ; q, )
{v(z1 , z2 ; q) + fxc [n0 ](z1 , z2 ; q, )} (z2 , z ; q, ),
(7.61)
39
where 0 (z, z ; q, ) and fxc [n0 ](z, z ; q, ) denote the 2D Fourier transforms of the
noninteracting density-response function 0 (r, r ; ) and the XC kernel fxc [n0 ](r, r ; ),
respectively. Using equation (5.7), and noting that the single-particle orbitals i (r) now take
the form
k,i (r) = eikr i (z),
one finds
0 (z, z ; q, ) =
(7.62)

fk,i fk+q,j
2
i (z)j (z)j (z )i (z )
,
A i,j
Ek,i Ek+q,j + + i
(7.63)
where
k2
,
(7.64)
2
the single-particle orbitals i (z) and energies i now being the solutions of the one-dimensional
KohnSham equation

1 d2
+ vKS [n0 ](z) i (z) = i i (z),

(7.65)
2 dz2
with
Ek,i = i +
vKS [n0 ](z) = vH [n0 ](z) + vxc [n0 ](z),

vH (z) = 2
dz |z z | n0 (z ) n+ (z ) ,
(7.66)
(7.67)

Exc [n]
vxc [n0 ](z) =
,
n(z) n=n0
and
n0 (z) =
1
(F i ) i2 (z) (F i ).
i
(7.68)
(7.69)
From equation (5.3), the imaginary part of the density-response function (z, z ; q, ) is easily
found to fulfil the following sum rule:

d2
n0 (z) (z z ).
d Im (z, z ; q, ) = q 2 +
(7.70)
dzdz
Within this scheme, the simplest possible approximation is to neglect XC effects altogether
and set the XC potential vxc [n0 ](z) and kernel fxc [n0 ](z, z ; q, ) equal to zero. In this case,
the one-dimensional single-particle wave functions i (z) and energies i are the self-consistent
eigenfunctions and eigenvalues of a one-dimensional Hartree Hamiltonian. The calculation
of the density-response function is further simplified if the Hartree potential vH [n0 ](z) of
equation (7.67) is replaced by
z z0 ,
v0 ,
vIBM (z) =
(7.71)
,
z > z0 ,
where the value z0 = (3/16)F is chosen so as to ensure charge neutrality. This is the so-called
inifinite-barrier model (IBM) [246], in which the single-particle orbitals i (z) are simply sines.
If one further neglects interference between incident and scattered electrons at the surface, this
model yields the classical IBM (CIBM) [247] which can be shown to be equivalent to the SRM
described in section 7.2.2.
40
J M Pitarke et al
Alternatively, and with the aim of incorporating band-structure effects (such as the
presence of energy gaps and surface states) approximately, the self-consistent jellium-like
KohnSham potential of equation (7.66) can be replaced by a physically motivated model
potential vMP (z). Examples are the parametrized model potential reported by Chulkov et al
[248], which was successful in the description of the lifetimes of image and Shockley states
in a variety of metal surfaces [249255], and the q-dependent model potential that has been
reported recently to investigate the momentum-resolved lifetimes of Shockley states at the
Cu(111) surface [256].
At this point, we note that for z and z coordinates that are far from the surface into the
vacuum, where the electron density vanishes, equation (7.59) takes the form of equation (7.56)
(which within the SRM is true for all z, z > 0), i.e.
W (z, z ; q, ) = v(z, z ; q)
2 q(z+z )
g(q, ),
e
q
(7.72)
but with the surface-response function g(q, ) now being given by the general expression [257]:

2
(7.73)
dz1 dz2 eq(z1 +z2 ) (z1 , z2 ; q, ),
g(q, ) =
q
which according to equation (5.1) can be expressed as follows

g(q, ) = dz eqz n(z; q, ),
(7.74)
with n(z; q, ) being the electron density induced by an external potential of the form
ext (z; q, ) =
2 qz
e .
q
(7.75)
In the framework of TDDFT, the induced electron density is obtained as in the RPA (see
equation (5.5)), but with the XC kernel fxc [n0 ](r, r ; q, ) added to the bare Coulomb
interaction v(r, r ). Hence, after Fourier transforming one writes

0

ext
n(z; q, ) = dz (z, z ; q, ){ (z ; q, ) + dz

v(z , z ; q) + fxc [n0 ](z , z ; q, ) n(z ; q, )}.
(7.76)
Using equations (7.70) and (7.73), the surface loss function Img(q, ) is easily found to
fulfil the following sum rule:

2
d Img(q, ) = 2 q dz e2qz n0 (z),
(7.77)
0
which for a step-like electron density n0 (z) of the form of equation (3.25) reduces to
equation (7.35), as expected.
Periodic surface.
For a periodic surface, single-particle wave functions are of the form
k,n;i (r) = k,n (r ) i (z),
(7.78)
where k,n (r ) are Bloch states:

1
k,n (r ) = ekr uk,n (r ),
A
(7.79)
41
with r and k being 2D vectors in the plane of the surface. Hence, one may introduce the
following Fourier expansion of the screened interaction:
1 i(q+g)r i(q+g )r
e
e
Wg,g (z, z ; q, ),
A q g,g
SBZ
W (r, r ; ) =
(7.80)
where q is a 2D wave vector in the surface Brillouin zone (SBZ), and g and g denote
2D reciprocal-lattice vectors. According to equation (7.4), the 2D Fourier coefficients
Wg,g (z, z ; q, ) are given by the following expression:

Wg,g (z, z ; q, ) = vg (z, z ; q) g,g + dz1 dz2
vg (z, z1 ; q) g,g (z1 , z2 ; q, ) vg (z2 , z ; q),
(7.81)
where vg (z, z ; q) denote the 2D Fourier coefficients of the bare Coulomb interaction v(r, r ):
vg (z, z ; q) =
2

e|q+g| |zz | ,
|q + g|
(7.82)
and g,g (z, z ; q, ) are the Fourier coefficients of the interacting density-response function
(r, r ; ). In the framework of TDDFT, one uses equation (5.11) to find:

0

g,g (z, z ; q, ) = g,g (z, z ; q, ) + dz1 dz2

0
g,g
+fgxc1 ,g2 [n0 ](z1 , z2 ; q, ) g2 ,g (z2 , z ; q, ),
(z, z1 ; q, ) vg (z1 , z2 ; q) g ,g
1
1 2
(7.83)
0
xc

where g,g
(z, z ; q, ) and fg,g [n0 ](z, z ; q, ) denote the Fourier coefficients of the
noninteracting density-response function 0 (r, r ; ) and the XC kernel fxc [n0 ](r, r ; ),
respectively. Using equation (5.7), one finds:

fk,n;i fk+q,n ;j
2
i (z)j (z)j (z )i (z )
A i,j
k+q,n ;j + h
( + i)
n,n k,n;i
SBZ
0

g,g
(z, z ; q, ) =
k,n |e
i(q+g)r
|k+q,n k+q,n |e
i(q+g )r
|k,n ,
(7.84)
the single-particle orbitals k,n;i (r) = k,n (r ) i (z) and energies k,n;i being the
eigenfunctions and eigenvalues of a 3D KohnSham Hamiltonian with an effective potential
that is periodic in the plane of the surface.
As in the case of the jellium surface, we can focus on the special situation where both z
and z coordinates are located far from the surface into the vacuum. Equation (7.4) shows that
under such conditions the Fourier coefficients Wg,g (z, z ; q, ) take the following form:
Wg,g (z, z ; q, ) = vg (z, z ; q) g,g
where
gg,g (q, ) =
2
q
2 q

gg,g (q, ) e|q+g|z e|q+g |z ,

|q + g| |q + g |
(7.85)

dz1
dz2 eq(z1 +z2 ) g,g (z1 , z2 ; q, ).
(7.86)
42
J M Pitarke et al
In particular,

gg=0,g=0 (q, ) =
dz eqz ng=0 (z; q, ),
(7.87)
with ng (z; q, ) being the Fourier coefficients of the electron density induced by an external
potential of the form
gext (z; q, ) =
2 qz
e g,0 .
q
(7.88)
Finally, one finds from equation (5.3) that the Fourier coefficients g,g (z, z ; q, ) fulfil
the following sum rule:

d2
n0 (z; q) (z z ),
d Img,g (z, z ; q , ) = q 2 +
(7.89)
dzdz gg
where the coefficients n0g (z; q) denote the 2D Fourier components of the ground-state electron
density n0 (r). Furthermore, combining equations (7.86) and (7.89), one writes

d Imgg,g (q, ) = 2 2 q dz e2qz ngg (z; q),
(7.90)
0
which is a generalization of the sum rule of equation (7.35) to the more general case of a real
solid in which the crystal structure parallel to the surface is taken into account.
8. Surface-response function
The central quantity that is involved in a description of surface collective excitations is the
surface-response function introduced in the preceding section: g(q, ) for a jellium surface
(see equation (7.73)) and gg,g (q, ) for a periodic surface (see equation (7.86)).
8.1. Generation-rate of electronic excitations
In the framework of TDDFT, an interacting many-electron system exposed to a frequencydependent external potential ext (r, ) is replaced by a fictitious system of noninteracting
electrons exposed to an effective self-consistent potential sc (r, ) of the form
sc (r, ) = ext (r, ) + (r, ),
where

(r, ) =

dr v(r, r ) + fxc [n0 ](r, r ; ) n(r , ).
(8.1)
(8.2)
Hence, the rate at which a frequency-dependent external potential ext (r, ) generates
electronic excitations in the many-electron system can be obtained, within lowest-order
perturbation theory, as follows

2
w() = 2
fi (1 fi ) j (r)| sc (r, )i (r) (i f ),
(8.3)
i,j
with i (r) and i being the eigenfunctions and eigenvalues of a 3D KohnSham Hamiltonian.
In terms of the induced electron density n(r, ) of equation (5.1), one finds [163]

w() = 2 Im dr ext
(r, ) n(r, ),
(8.4)
43
Figure 10. A schematic drawing of the scattering geometry in angle-resolved inelastic electron
scattering experiments. On exciting a surface mode of frequency (q), the energy of detected
electrons becomes f = i (q), with the momentum q being determined by equation (8.9).
which in the case of a periodic surface takes the following form:

w() =
SBZ

g
wg (q, ),
(8.5)
where wg (q, ) denotes the rate at which the external potential generates electronic excitations
of frequency and parallel wave vector q + g:

2
(8.6)
wg (q, ) = Im
dz gext (z, q) ng (z, q),
A
gext (z, q) and ng (z, q) being the Fourier coefficients of the external potential and the induced
electron density, respectively.
In particular, if the external potential is of the form of equation (7.88), the rate wg (q, )
can be expressed in terms of the surface-response function gg,g (q, ) of equation (7.86), as
follows
4
(8.7)
wg (q, ) =
Imgg,g (q, ) g,0 ,
qA
which for a jellium surface reduces to
4
Img(q, ),
(8.8)
w(q, ) =
qA
with g(q, ) being the surface-response function of equation (7.73). We note that although
only the coefficient g=0,g =0 (z, z ; q, ) enters into the evaluation of the more realistic
0

equation (8.7), the full g,g
(z, z ; q, ) matrix is implicitly included through equation (7.83).
8.2. Inelastic electron scattering
The most commonly used experimental arrangement for the detection of surface collective
excitations by the fields of moving charged particles is based on angle-resolved inelastic
electron scattering [258]. Figure 10 shows a schematic drawing of the scattering geometry.
A monochromatic beam of electrons of energy i , incident on a flat surface at an angle i , is
44
J M Pitarke et al
back scattered and detected by an angle-resolved energy analyzer positioned at an angle f and
energy f . Inelastic events can occur, either before or after the elastic event, on exciting a surface
mode of frequency (q) = i f . The energy and lifetime of this mode are determined by
the corresponding energy-loss peak in the spectra, and the momentum q parallel to the surface
is obtained from the measured angles i and f , as follows:

(8.9)
q = 2 i sin i f sin f .
The inelastic scattering cross section corresponding to a process in which an electronic
excitation of energy and parallel wave vector q is created at a semi-infinite solid surface
can be found to be proportional to the rate given by equation (8.7) [or equation (8.8) if the
medium is represented by a jellium surface] and is, therefore, proportional to the imaginary
part of the surface-response function, i.e. the so-called surface-loss function. Hence, apart
from kinematic factors (which can indeed vary with energy and momentum [259, 260]), the
inelastic scattering cross section is dictated by the rate wg (q, ) [or w(q, ) if the medium is
represented by a jellium surface] at which an external potential of the form of equation (7.88)
generates electronic excitations of frequency and parallel wave vector q.
In the following sections, we focus on the behaviour of the energy-loss function
Imgg=0,g=0 (q, ) [or Img(q, )] and its maxima, which account for the presence and
momentum-dispersion of surface collective excitations.
8.3. Surface plasmons: jellium surface
8.3.1. Simple models. In the simplest possible model of a jellium surface in vacuum, in
which a semi-infinite medium with local dielectric function () at z 0 is terminated at
z = 0, the surface-response function g(q, ) is obtained from equation (7.9) with 2 = 1, or,
equivalently, from equation (7.49) with q = 0, i.e.
g(q, ) =
() 1
,
() + 1
(8.10)
which for a Drude dielectric function (see equation (2.15)) leads to the surface-loss function
(8.11)
Img(q, ) = s ( s )
2
peaked at the surface-plasmon energy s = p / 2.

The classical energy-loss function of equation (8.10) represents indeed the true longwavelength (q 0) limit of the actual self-consistent surface-loss function of a jellium surface.
Nevertheless, the classical picture leading to equation (8.10) ignores both the nonlocality of
the electronic response of the system and the microscopic spatial distribution of the electron
density near the surface. Nonlocal effects can be incorporated within the hydrodynamic and
specular-reflection models described in sections 7.2.1 and 7.2.2.
Within a one-step hydrodynamic approach, the surface-loss function is also dominated by
a delta function (see equation (7.43)) but peaked at the momentum-dependent surface-plasmon
energy of equation (7.40) (see also equation (3.30)):
1 2
2 =
(8.12)
p + 2 q 2 + q 2p2 + 2 q 2 ,
2
which at long wavelengths yields
(8.13)
= p / 2 + q/2,
representing the speed of propagation of hydrodynamic disturbances in the electron system9 .
45
In the SRM (with the bulk dielectric function being described within the RPA),
surface plasmons, which occur at a momentum-dependent energy slightly different from its
hydrodynamic counterpart, are damped by the presence of eh pair excitations, as shown in
figure 919 .
The one-step hydrodynamic equations (8.12) and (8.13) and a numerical evaluation of the
imaginary part of the SRM surface-response function of equation (7.49) (see figure 9) both
yield a positive surface-plasmon energy dispersion at all wave vectors. Nonetheless, Bennett
used a hydrodynamic model with a continuum decrease of the electron density at the metal
surface, and found that a continuous electron-density variation yields a monopole surface
plasmon with a negative dispersion at low wave vectors [84].
8.3.2. Self-consistent calculations: long wavelengths. Within a self-consistent longwavelength description of the jellium-surface electronic response, Feibelman showed that
up to first order in an expansion in powers of the magnitude q of the wave vector, the surfaceresponse function of equation (7.73) can be written as [162]
[() 1] [1 + qd ()]
g(q, ) =
,
(8.14)
() + 1 [() 1] qd ()
where () represents the long-wavelength limit of the dielectric function of the bulk material 20
and d () denotes the centroid of the induced electron density (see equation (3.13)) with
respect to the jellium edge21 . Equation (8.14) shows that at long wavelengths the poles of
the surface-response function g(q, ) are determined by the non-retarded surface-plasmon
condition of equation (3.12) with 2 = 1, which for a semi-infinite free-electron metal in
vacuum yields the surface-plasmon dispersion relation of equation (3.15), i.e.
= s (1 + q) ,
(8.15)
= Re [d (s )] /2.
(8.16)
with
Equations (8.15)(8.16) show that the long-wavelength surface-plasmon energy dispersion
is dictated by the position of the centroid of the induced electron density with respect to the
jellium edge. This can be understood by noting that the potential associated with the surfaceplasmon charge attenuates on either side of Re[d (s )] with the attenuation constant q. If
the fluctuating charge lies inside the jellium edge (Re[d (s )] < 0), as q increases (thus
the surface-plasmon potential attenuating faster) more of the field overlaps the metal giving
rise to more interchange of energy between the electric field and the metal and resulting in
a positive energy dispersion. However, if the fluctuating charge lies outside the jellium edge
(Re[d (s )] > 0), as q increases less of the metal is subject to the plasmons electric field
(i.e. there is a decreasing overlap of the fluctuating potential and the unperturbed electron
density) which results in less interchange of energy between the electric field and the metal
and a negative dispersion coefficient [261, 262].
Quantitative RPA calculations of the surface-plasmon linear-dispersion coefficient
entering equation (8.15) were carried out by several authors by using the specular-reflection
and infinite-barrier models of the surface [82, 83, 88], a step potential [89, 90], and the more
19 The quantity Im[W (z, z; q, )] represented in figure 9 (see equation (7.56)) coincides, apart from the factor
2 exp(2qz)/q, with the energy-loss function Img(q, ).
20 For an ideal free-electron gas, () is given by the Drude dielectric function of equation (2.15).
21 In general, the d-parameter d () entering equation (8.14) should be replaced by d () d (). However, as for

neutral jellium surfaces d () coincides with the jellium edge, d () d () reduces to the centroid d () of the
induced electron density with respect to the jellium edge.
46
J M Pitarke et al
0.5
()
-0.5
-1
-1.5
-2
3
rs
Figure 11. Long-wavelength surface-plasmon dispersion coefficient entering equation (8.15),

versus the electron-density parameter rs , as obtained from equation (8.17) (dotted line) and from
SRM (thick dashed line), IBM (thin dashed line), and self-consistent RPA and ALDA (thin and
thick solid lines, respectively) calculations of the centroid of the induced electron density at = s .
The solid circles represent the angle-resolved low-energy inelastic electron scattering measurements
reported in [98] for Na and K, in [100] for Cs, in [101] for Li and Mg, and in [264] for Al. The
IBM and self-consistent (RPA and ALDA) calculations have been taken from [88] and [103],
respectively.
realistic LangKohn self-consistent surface potential [92, 103, 263]. Both Feibelmans RPA
self-consistent calculations [92] and the ALDA calculations carried out later by Liebsch [103]
and by Kempa and Schaich [263] demonstrated that in the range of typical bulk densities
(rs = 26) the centroid d of the induced electron density at s lies outside the jellium
edge, which leads to a negative long-wavelength dispersion of the surface plasmon. These
calculations, which corroborated Bennetts prediction [84], also demonstrated that the longwavelength surface-plasmon dispersion is markedly sensitive to the shape of the barrier and to
the presence of short-range XC effects.
Existing calculations of the long-wavelength dispersion coefficient entering
equation (3.15) are shown in figure 11 and summarized in table 1. At one extreme, the
dotted line of figure 11 gives the single-step hydrodynamic coefficient HD obtained from
equation (3.31), i.e.
HD =
,
(8.17)
2s
3/5(3 2 n0 )1/3 ; within this model, the long-wavelength surface-plasmon

with =
dispersion is always positive. As in the single-step hydrodynamic approach, the SRM
equilibrium density profile is of the form of equation (3.25), and in the IBM the electron
density still varies too rapidly; as a result, the corresponding SRM and IBM coefficients
(represented by thick and thin dashed lines, respectively) are both positive. At the other
extreme are the more realistic self-consistent RPA and ALDA calculations, represented by
thin and thick solid lines, respectively, which yield a linear-dispersion coefficient that is
negative in the whole metallic density range.
Conclusive experimental confirmation that the original Bennetts prediction [84] was
correct came with a series of measurements based on angle-resolved low-energy inelastic
electron scattering [98101, 264]. It has been demonstrated that the surface-plasmon energy
of simple metals disperses downward in energy at small momentum q parallel to the surface, the
47
Table 1. Long-wavelength surface-plasmon dispersion coefficient in for various simple metal

surfaces, as obtained from equation (8.17) (HD) and from SRM, IBM, and self-consistent RPA and
ALDA calculations of the centroid of the induced electron density at = s (see equation (8.16)),
and from angle-resolved low-energy inelastic electron scattering measurements reported in [98] for
Na and K, in [100] for Cs, in [101] for Li and Mg, and in [264] for Al. As in figure 11, the IBM and
self-consistent (RPA and ALDA) calculations have been taken from [88] and [103], respectively.
Also shown in this table are the measured valuesof the
surface-plasmon energy s at q = 0, which
are all slightly below the jellium prediction: p / 2 = 3/2rs3 e2 /a0 , due to band-structure effects.
s
HD
SRM
IBM
RPA
ALDA
Exp.
2.07
2.66
3.25
3.93
4.86
5.62
10.86
7.38
4.28
3.99
2.74
1.99
0.46
0.52
0.58
0.64
0.71
0.76
0.50
0.57
0.63
0.70
0.78
0.84
0.36
0.21
0.30
0.40
0.45
0.35
0.26
0.32
0.50
0.70
0.85
1.10
1.14
0.32
0.41
0.24
0.39
0.39
0.44
Im g(q,) (arbitrary units)
Al
Mg
Li
Na
K
Cs
rs
0.33
0.33
-1
q=0.10 A
-1
q=0.15 A
-1
q=0.20 A
-1
q=0.30 A
-1
q=0.40 A
-1
q=0.50 A
0.6
0.7
0.8
/p
0.9
Figure 12. Self-consistent RPA calculations (thin solid lines) of the surface loss function Img(q, ),
as obtained from equation (7.73), versus the energy for a semi-infinite free-electron gas with the
electron density equal to that of valence electrons in Al (rs = 2.07) and for various magnitudes of
the 2D wave vector q. The thick solid line represents the SRM surface loss function for rs = 2.07
and q = 0.5 1 .
dispersion coefficients (represented in figure 11 by solid circles) being in reasonable agreement

with self-consistent jellium calculations, as shown in figure 11 and table 1.
8.3.3. Self-consistent calculations: arbitrary wavelengths. At arbitrary wavelengths, surfaceplasmon energies can be derived from the maxima of the surface loss function Img(q, )
and compared with the peak positions observed in experimental electron energy-loss spectra.
Figure 12 shows the self-consistent calculations of Img(q, ) that we have obtained in the
RPA for a semi-infinite free-electron gas (jellium surface) with the electron density equal to
that of valence electrons in Al (rs = 2.07). For 2D wave vectors of magnitude in the range
q = 0 0.5 1 , all spectra are clearly dominated by a surface-plasmon excitation, which is
seen to first shift to lower frequencies as q increases [as dictated by equations (8.15)(8.16)]
48
J M Pitarke et al
and then, from about q = 0.15 1 on, towards higher frequencies.

For the numerical evaluation of the spectra shown in figure 12 we have first computed
the interacting density-response function (z, z ; q, ) of a sufficiently thick jellium slab
by following the method described in [265], and we have then derived the surface-loss
function from equation (7.73). Alternatively, the self-consistent energy-loss spectra reported
in [100] (see also [163]) were obtained by first computing the noninteracting density-response
function 0 (z, z ; q, ) of a semi-infinite electron system in terms of Greens functions, then
performing a matrix inversion of equation (7.76) with the external potential ext (z; q, )
of equation (7.75), and finally deriving the surface-loss function from equation (7.74). As
expected, both approaches yield the same results.
Figure 13 shows the calculated and measured dispersion of surface plasmons in Al, Mg,
Li, Na, K, and Cs. The jellium calculations presented here do not include effects due to
band-structure effects; thus, all frequencies have been normalized to the measured value s
of the q = 0 surface-plasmon energy (see table 1). This figure shows that the single-step HD
and SRM surface-plasmon dispersions (thin and thick dashed lines, respectively) are always
upward and nearly linear. However, self-consistent RPA and ALDA calculations (thin and
thick solid lines, respectively), which are based on a self-consistent treatment of the surface
density profile, show that the surface-plasmon dispersion is initially downward (as also shown
in figure 11 and table 1), then flattens out, and rises thereafter, in agreement with experiment
(solid circles). A comparison between self-consistent RPA and ALDA calculations shows that
although there is no qualitative difference between them XC effects tend to reduce the surfaceplasmon energy, thereby improving the agreement with the measured plasmon frequencies.
This lowering of the surface-plasmon energies shows that dynamic XC effects combine to
lower the energy of the electron system, which is due to the weakening of the Coulomb ee
interaction by these effects.
Recently, XC effects on the surface-plasmon dispersion of Mg and Al were introduced
still in the framework of TDDFT (but beyond the ALDA) by using the nonlocal (momentumdependent) static XC local-field factor of equation (5.24) [118,119]. At low wave vectors, this
calculation (thick solid line with open circles) nearly coincides with the ALDA calculation
(thick solid line), as expected. At larger wave vectors, however, the nonlocal calculation
begins to deviate from the ALDA, bringing the surface-plasmon dispersion to nearly perfect
agreement with the data for all values of the 2D wave vector. Ab initio calculations of the
surface-plasmon dispersion of real Al and Mg were also reported in [118, 119]. For plasmon
frequencies that are normalized to the measured value s at q = 0 (as in figure 13), ab initio
and jellium calculations are found to be nearly indistinguishable; however, only the ab initio
calculations account for an overall lowering of the surface-plasmon dispersion that is due to
core polarization.
In the case of the alkali metals Li, Na, K and Cs, there is a mismatch in the linear (low q)
region of the surface-plasmon dispersion curve between jellium ALDA calculations and the
experiment. The theoretical challenges for the future are therefore to understand the impact
of band-structure and many-body effects on the energy of surface plasmons in these materials,
which will require to pursue complete first-principles calculations of the surface electronic
response at the level of those reported in [118, 119] for Mg and Al, respectively.
Thin films of jellium have been considered recently, in order to investigate the influence of
the slab thickness on the excitation spectra and the surface-plasmon energy dispersion [266].
Oscillatory structures were found, corresponding to electronic interband transitions, and it was
concluded that in the case of a slab thickness larger than 100 a.u. surface plasmons behave
like the surface plasmon of a semi-infinite system.
49
Al
Mg
0.8
/p
/p
0.8
0.7
0.7
0.2
0.4
0.2
-1
q (A )
Na
0.8
/p
/p
Li
0.8
0.7
0.7
-1
0.2
-1
q (A )
0.2
q (A )
/p
Cs
0.8
/p
K
0.8
0.7
0.7
0.4
-1
q (A )
-1
q (A )
0.2
-1
0.2
q (A )
Figure 13. Surface-plasmon energy dispersion for the simple metals Al, Mg, Li, Na, K, and
Cs, as obtained from equation (8.12) (thin dashed lines) and from SRM (thick dashed lines) and
self-consistent RPA (thin solid lines) and ALDA (thick solid lines) calculations of the surface loss
function Img(q, ), and from the peak positions observed in experimental electron energy-loss
spectra (solid circles). The dotted lines represent the initial slope of the HD surface-plasmon
energy dispersion, as obtained from equation (3.31). In the case of Mg, the thick solid line with
open circles represents the jellium TDDFT calculations reported in [118] and obtained with the
use of the nonlocal (momentum-dependent) static XC local-field factor of equation (5.24). All
frequencies have been normalized to the measured value s of the q = 0 surface-plasmon energy.
8.3.4. Surface-plasmon linewidth. At jellium surfaces, the actual self-consistent surface

response function Img(q, ) reduces in the long-wavelength (q 0) limit to equation (8.11),
so that long-wavelength surface plasmons are expected to be infinitely long-lived excitations.
At finite wave vectors, however, surface plasmons are damped (even at a jellium surface) by
the presence of eh pair excitations [267].
Figure 14 shows the results that we have obtained from SRM (dashed lines) and selfconsistent (solid lines) RPA calculations of the full width at half maximum (FWHM), , of
the surface loss function Img(q, ). Also shown in this figure are the corresponding linewidths
that have been reported in [98, 100, 101, 264] from experimental electron energy-loss spectra
50
J M Pitarke et al
4
Al
Mg
2.5
3
(eV)
(eV)
2
2
1.5
1
1
0.5
0
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.2
0.3
0.4
q (A-1)
2
0.8
Na
Li
1.5
(eV)
0.6
(eV)
0.1
(A-1)
0.4
0.5
0.2
0
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.1
(A-1)
0.2
q (A-1)
0.7
0.6
0.8
Cs
(eV)
(eV)
0.5
0.6
0.4
0.4
0.3
0.2
0.2
0.1
0
0
0.1
0.2
q (A-1)
0
0
0.05
0.1
0.15
0.2
0.25
0.3
q (A-1)
Figure 14. Surface-plasmon full width at half maximum (FWHM), , for the simple metals
Al, Mg, Li, Na, K, and Cs, as obtained from SRM (thick dashed lines) and self-consistent RPA
(thin solid lines) calculations of the surface loss function Img(q, ), and from the experimental
electron energy-loss spectra at different scattering angles (solid circles) reported in [98] for Na and
K, in [100] for Cs, in [101] for Li and Mg, and in [264] for Al. In the case of Mg, the thick solid
line with open circles represents the jellium TDDFT calculations reported in [118] and obtained
with the use of the nonlocal (momentum-dependent) static XC local-field factor of equation (5.24).
at different scattering angles. This figure clearly shows that at real surfaces the surfaceplasmon peak is considerably wider than predicted by self-consistent RPA jellium calculations,
especially at low wave vectors. This additional broadening should be expected to be mainly
caused by the presence of short-range many-body XC effects and interband transitions, but
also by scattering from defects and phonons.
Many-body XC effects on the surface-plasmon linewidth of Mg and Al were incorporated
in [118,119] in the framework of TDDFT with the use of the nonlocal (momentum-dependent)
static XC local-field factor of equation (5.24). These TDDFT calculations have been plotted in
figure 14 by solid lines with circles; a comparison of these results with the corresponding RPA
calculations (solid lines without circles) shows that short-range XC effects tend to increase
the finite-q surface-plasmon linewidth, bringing the jellium calculations into nice agreement
51
Table 2. Relative widths /s of surface plasmons, as derived from the imaginary part of the
surface-response function of equation (8.10) with measured values of the bulk dielectric function
() (theory) [163] and from the surface-loss measurements at q = 0 reported in [98,100,101,264]
(experiment).
Theory
Experiment
Al
Mg
Li
0.035
0.22
0.16
0.19
0.33
0.35
Na
K
0.035
0.07
Cs
Ag
Hg
0.027
0.027
0.18
0.16
with experiment at the largest values of q. Nevertheless, jellium calculations cannot possibly
account for the measured surface-plasmon linewidth at small q, which deviates from zero even
at q = 0.
In the long-wavelength (q 0) limit, surface plasmons are known to be dictated by bulk
properties through a long-wavelength bulk dielectric function (), as in equation (8.10).
Hence, the experimental surface-plasmon widths at q = 0 should be approximately
described by using in equation (8.10) the measured bulk dielectric function (). Table 2
exhibits the relative widths /s derived in this way [163], together with available surfaceloss measurements at q = 0. Since silver (Ag) and mercury (Hg) have partially occupied
d-bands, a jellium model, like the one leading to equation (8.10), is not, in principle, appropriate
to describe these surfaces. However, table 2 shows that the surface-plasmon width of these
solid surfaces is very well described by introducing the measured bulk dielectric function
(which includes band-structure effects due to the presence of d-electrons) into equation (8.10).
Nevertheless, the surface-plasmon widths of simple metals like K and Al, with no d-electrons,
are considerably larger than predicted in this simple way22 . This shows that an understanding
of surface-plasmon broadening mechanisms requires a careful analysis of the actual band
structure of the solid.
Approximate treatments of the impact of the band structure on the surface-plasmon energy
dispersion have been developed by several authors, but a first-principles description of the
surface-plasmon energy dispersion and linewidth has been reported only in the case of the
simple-metal prototype surfaces Mg(0001) [118] and Al(111) [119]; these calculations will be
discussed in section 8.4.3.
8.3.5. Multipole surface plasmons. In his attempt to incorporate the smoothly decreasing
electron density profile at the surface, Bennett [84] solved the equations of a simple
hydrodynamic model with a density profile which decreases linearly through the surface
region and found that in addition to Ritchies surface plasmon at s , with a negative
energy dispersion at low q wave vectors, there is an upper surface plasmon at higher energies.
This is the so-called multipole surface plasmon, which shows a positive wave vector dispersion
even at small q.
The possible existence and properties of multipole surface plasmons was later investigated
in the framework of hydrodynamical models for various choices of the electron-density profile
at the surface [269271]. According to these calculations, higher multipole excitations could
indeed exist, for a sufficiently diffuse surface, in addition to the usual surface plasmon at
s . However, approximate quantum-mechanical RPA calculations gave no evidence
for the existence of multipole surface plasmons [9, 89], thereby leading to the speculation that
multipole surface plasmons might be an artifact of the hydrodynamic approximation [271,272].
22 Beck and Dasgupta [268] reported a two-band model calculation of the surface-plasmon linewidth of Al, but they
found /s 0.07 which is also too small.
52
J M Pitarke et al
Ep = 50 eV
s
i = 45
Intensity (arb.units)
45
47
49
51
53
55
57
0
10
12
14
16
18
20
Electron Energy Loss (eV)
Figure 15. Angle-resolved high-resolution electron energy-loss spectra of the Al(111) surface at
various scattering angles s . The primary beam energy is 50 eV and the incident angle is i = 45
(from [264], used with permission).
The first experimental sign for the existence of multipole surface modes was established
by Schwartz and Schaich [273] in their theoretical analysis of the photoemission yield spectra
that had been reported by Levinson et al [122]. Later on, Dobson and Harris [274] used
a DFT scheme to describe realistically the electron-density response at a jellium surface, to
conclude that multipole surface plasmons should be expected to exist even for a high-density
metal such as Al (which presents a considerably abrupt electron density profile at the surface).
Two years later, direct experimental evidence of the existence of multipole surface plasmons
was presented in inelastic reflection electron scattering experiments on smooth films of the
low-density metals Na, K and Cs [99], the intensity of these multipole surface plasmons being
in agreement with the DFT calculations reported later by Nazarov and Nishigaki [275]. The
Al multipole surface plasmon has been detected only recently by means of angle-resolved
high-resolution electron-energy-loss spectroscopy (HREELS) [264].
Figure 15 shows the loss spectra of the Al(111) surface, as obtained by Chiarello et al [264]
with an incident electron energy of 50 eV and an incident angle of 45 with respect to the surface
normal. The loss spectrum obtained in the specular geometry (s = 45 ) is characterized
mainly by a single peak at the conventional surface-plasmon energy s = 10.55 eV. For offspecular scattering angles (s = 45 ), the conventional surface plasmon exhibits a clear energy
dispersion and two other features arise in the loss spectra: the multipole surface plasmon and
the bulk plasmon. The loss spectrum obtained at s = 53 , which corresponds to q = 0
(see equation (8.9)) is represented again in figure 16, but now together with the background
substraction and Gaussian-fitting procedure reported in [264]. The peak at p = 15.34
corresponds to the excitation of the Al bulk plasmon, and the multipole surface plasmon
is located at 13.20 eV.
The calculated and measured energies and linewidths of long-wavelength (q 0)
multipole surface plasmons in simple metals are given in table 3. On the whole, the ratio
m /p agrees with ALDA calculations. Good agreement between ALDA calculations and
experiment is also obtained for the entire dispersion of multipole surface plasmons, which is
53
Figure 16. Angle-resolved high-resolution electron energy-loss spectra of the Al(111) surface at
s = 53 , for i = 45 and i = 50 eV (as in figure 15) but now together with the deconvolution into
contributions corresponding to the excitation of the conventional surface plasmon at s = 10.55 eV,
multipole surface plasmon at 13.20 eV, and bulk plasmon at p = 15.34 eV (from [264], used with
permission).
Table 3. Angle-resolved low-energy inelastic electron scattering measurements of the energy m
and width of multipole surface plasmons at q = 0, as reported in [99] for Na and K, in [100]
for Cs, in [101] for Li and Mg, and in [264] for Al. Also shown is the ratio m /p calculated in
the RPA and ALDA and reported in [100].
Al
Mg
Li
Na
K
Cs
rs
m (eV)
m /p
RPA
ALDA
(eV)
/p
2.07
2.66
3.25
3.93
4.86
5.62
13.20
0.86
0.14
0.81
0.84
0.83
0.782
0.784
0.789
0.798
0.814
0.837
2.1
4.67
3.20
2.40
0.821
0.825
0.833
0.849
0.883
0.914
1.23
0.68
0.64
0.21
0.18
0.22
found to be approximately linear and positive. This positive dispersion is originated in the
fact that the centroid of the induced electron density, which at s is located outside the
jellium edge, is shifted into the metal at the multipole resonance frequency m . We also note
that multipole surface plasmons have only been observed at wave vectors well below the cutoff
value for Landau damping, which has been argued to be due to an interplay between Coulomb
and kinetic energies [276].
In a semi-infinite metal consisting of an abrupt step of the unperturbed electron density at
the surface (as in the hydrodynamic and specular-reflection models described above), surface
plasmons would be localized at the surface and would propagate like plane waves, as shown in
figure 3, with positive and negative surface charge regions alternating periodically. In the real
situation in which the electron density decays smoothly at the surface, surface plasmons also
propagate along the surface as illustrated in figure 3, but the finite width of the electron-density
profile yields fluctuating densities with finite widths, as illustrated in figure 17 for a model
surface electron density profile with linear decay over a finite region.
Figure 17 qualitatively represents the real situation in which apart from small Friedel
oscillations the distribution of the conventional surface plasmon at s consists in the direction
normal to the surface of a single peak (i.e. it has a monopole character), while the charge
J M Pitarke et al
n (arb. units)
54
m
0
0
z (arb. units)
Figure 17. Real part of the electron density induced at = s (solid line) and = m (dashed
line), as reported in [163] for a model unperturbed electron density profile with linear decay over
a finite region.
distribution of the upper mode at m has decreasing oscillating amplitude towards the interior
of the metal, i.e. it has a multipole character. Along the direction normal to the surface, the
electronic density associated with this multipole surface plasmon integrates to zero.
In the retarded region, where q < s /c, the surface-plasmon dispersion curve deviates
from the non-retarded limit (where q s /c) and approaches the light line = cq, as
shown by the lower solid line of figure 2, thus going to zero at q = 0; hence, in a light
experiment the external radiation dispersion line will never intersect the surface-plasmon line,
i.e. in an ideal flat surface the conventional monopole surface plasmon cannot be excited in
a photoyield experiment. However, the multipole surface plasmon dispersion curve crosses
the light line at q m /c and goes to m at q = 0. Consequently, angle- and energyresolved photoyield experiments are suitable to identify the multipole surface plasmon. A large
increase in the surface photoyield was observed from K and Rb at m = 3.15 eV and
2.84 eV, respectively [277], from Al(100) at m = 12.5 eV [122], and from Al(111) at
m = 13 eV [123], in nice agreement with the multipole-plasmon energy observed with
the HREELS technique by Chiarello et al [264]. HREELS measurements yield, however,
a FWHM of 2.1 eV for the q = 0 multipole surface plasmon in Al(111) [264], which is
considerably smaller than the FWHM of 3 eV measured by photoyield experiments [123].
8.4. Surface plasmons: real surfaces
8.4.1. Stabilized jellium model. A simple way of including approximately the lattice
potential that is absent in the jellium model is the so-called stabilized jellium or structureless
pseudopotential model [278, 279], which yields energy stability against changes in the
background density.
In this model, a solid surface is assumed to be translationally invariant in the plane of the
surface, as in the jellium model. Hence, single-particle wave functions can be separated as in
equation (7.62) into a plane wave along the surface and a component i (z) describing motion
normal to the surface. In the framework of DFT and TDDFT, this component is obtained by
solving self-consistently a KohnSham equation of the form of equation (7.65) but with the
55
Table 4. Measured values of the surface-plasmon energy s at q = 0 and the long-wavelength

surface-plasmon dispersion coefficient of equation (8.16) for the noble metal Ag and the transition
metals Hg and Pd. Also shown in this table is the coefficient obtained from RPA and ALDA
calculations of the centroid of the electron density induced at = s (see equation (8.16)) in a
homogeneous electron gas with the electron density equal to that of valence sp electrons in Ag and
Hg. Note that the measured values
surface-plasmon energy s at q = 0 are considerably
of the
below the jellium prediction: p / 2 = 3/2rs3 e2 /a0 , mainly due to the presence of d electrons.
s and are given in eV and , respectively.
rs
Ag(100)
Ag(111)
Ag(110)
3.02
3.02
3.02
3.7
3.7
3.7
Hg
Pd
2.65
6.9
7.4
Direction
exp
RPA
ALDA
100
110
0.377
0.162
0.305
0.114
0.167
1.02
0.370
0.370
0.370
0.370
0.32
0.609
0.609
0.609
0.609
0.50
effective KohnSham potential of equation (7.66) being replaced by

vKS [n0 ](z) = vH [n0 ](z) + vxc [n0 ](z) + vWS ,
(8.18)
vWS representing the difference between a local pseudopotential and the jellium potential:
vWS =
3rc2
3Z 2/3
,
2rs3
10rs
(8.19)
where Z is the chemical valence of the solid and rc is a core radius that is chosen to stabilize
the metal for given values of the parameters rs and Z.
The stabilized jellium model was used by Ishida and Liebsch [280] to carry out RPA
and ALDA calculations of the dispersion of the energy and linewidth of surface plasmons
in Mg and Li. It is well known that the stabilized jellium model gives considerably better
work functions and surface energies than the standard jellium model; however, the impact of a
structureless pseudopotential on the properties of surface plasmons is found to be very small.
In particular, this simple pseudopotential cannot explain the presence of core polarization
lowering the surface-plasmon frequency for all wave vectors and does not account for the fact
that the measured q dependence of the Li surface-plasmon energy is very much flatter than
calculated within the jellium model (see figure 13). This discrepancy (not present in the case
of Al and Mg) is attributed to the presence of an interband transition in Li at 3.2 eV, which
is only slightly below the measured surface-plasmon energy at q = 0 (s = 4.28 eV) and
allows, therefore, for interference between bulk single-particle and surface collective modes,
as discussed in [280].
8.4.2. Occupied d-bands: simple models. Significant deviations from the dispersion of
surface plasmons at jellium surfaces occur on the noble metal Ag [104107], and the transition
metals Hg [108] and Pd [281] (see table 4). These deviations are mainly due to the presence
in these metals of filled 4d and 5d bands, which in the case of Ag yields an anomalous positive
and strongly crystal-face dependent dispersion.
Ag. In order to describe the observed features of Ag surface plasmons, Liebsch [111]
considered a self-consistent jellium model for valence 5s electrons and accounted for the
presence of occupied 4d bands via a polarizable background of d electrons characterized by a
local dielectric function d () that can be taken from bulk optical data (see figure 18) [282].
56
J M Pitarke et al
-2
3
3.5
(eV)
4.5
Figure 18. The thin solid line represents the real part of the d band contribution d () to the
measured optical dielectric function of Ag [282]. The thick solid line represents (q, )+d ()1
at q = 0, i.e. d () p2 /2 . The vertical dotted lines represent the energies at which the
plasmon conditions of equations (8.22) and (8.29) are fulfilled: p = 3.78 eV and s = 3.62 eV,
respectively. At these frequencies: d (p ) = 5.65 and d (p ) = 5.15.
For a homogeneous electron gas, one simply replaces in this model the bare Coulomb
interaction v(r, r ) by
v (r, r ; ) = v(r, r ) d1 (),
(8.20)
which due to translational invariance yields the following expression for the Fourier transform
W (q, ) of the screened interaction W (r, r ; ) of the form of equation (7.3):
W (q, ) =
v(q)
,
(q, ) + d () 1
(8.21)
with (q, ) being the dielectric function of a homogeneous system of sp valence electrons
(5s electrons in the case of Ag). Hence, the screened interaction W (q, ) has poles at the
bulk-plasmon condition
(q, ) + d () 1 = 0,
(8.22)
which in the case of Ag (rs = 2.02) and in the absence of d electrons (d = 1) yields the
long-wavelength (q 0) bulk plasmon energy p = 8.98 eV. Instead, if d electrons are
characterized by the frequency-dependent dielectric function d () represented by a thin solid
line in figure 18, equation (8.22) yields the observed long-wavelength bulk plasmon at
p
= 3.78 eV.
(8.23)
p =
d (p )
In the case of a solid surface, one still uses a modified (d-screened) Coulomb interaction
v (r, r ; ) of the form of equation (8.20), but with d () being replaced by
z zd
d (),
d (z, ) =
(8.24)
1,
z > zd ,
57
which represents a polarizable background of d electrons that extends up to a certain plane at

z = zd . Using equation (8.24), the 2D Fourier transform of v (r, r ; ) yields [283]:
v (z, z ; q , ) =
2

[eq |zz | + sgn(zd z )d () eq |zzd | eq |zd z | ],
q d (z , )
(8.25)
where
d = [d () 1]/[d () + 1].
(8.26)
Introduction of equations (8.25) and (8.26) into an equation of the form of equation (7.59)
yields the following expression for the modified RPA surface-response function:

n(z; q, )
+ a(q, ),
(8.27)
g (q, ) = dz eqz
d (z, )
where n(z; q, ) represents the RPA induced density of sp valence electrons which is given
by equation (7.76) with fxc [n0 ](z, z ; q, ) = 0, and

n(z; q, )
qzd
q|zd z|
sgn(zd z)
.
(8.28)
a(q, ) = d () e + dze
d (z, )
In the long-wavelength (q 0) limit, equation (8.27) takes the form of equation (8.10)
but with () replaced by () + d () 1, which leads to the surface-plasmon condition:
() + d () = 0.
(8.29)
For a Drude dielectric function () with p = 8.98 eV (rs = 3.02) and in the absence of d
electrons (d = 1), the surface-plasmon condition of equation (8.29) yields the surface-plasmon
energy s = 6.35 eV. However, in the presence of d electrons characterized by the frequencydependent dielectric function d (), equation (8.29) leads to a modified (d-screened) surface
plasmon at
p
= 3.62 eV,
(8.30)
s =
1 + d (s )
which for p = 8.98 eV and the dielectric function d () represented by a thin solid line in
figure 18 yields s = 3.62 eV, only slightly below the energy 3.7 eV of the measured Ag
surface plasmon [106].
At finite wavelengths, the surface-plasmon dispersion was derived by Liebsch from the
peak positions of the imaginary part of the surface-response function of equation (8.27), as
obtained with a self-consistent RPA calculation of the induced density n(z; q, ) of 5s1
valence electrons. The surface-plasmon dispersion was found to be positive for zd 0 and
to best reproduce the observed linear dispersion of Ag surface plasmons for zd = 0.8 .
Hence, one finds that (i) the sd screened interaction is responsible for the lowering of the
surface-plasmon energy at q = 0 from the free-electron value of 6.35 eV to 3.62 eV, and (ii)
the observed blueshift of the surface-plasmon frequency at increasing q can be interpreted as
a reduction of the sd screened interaction in the selvedge region that is due to a decreasing
penetration depth of the induced electric field.
With the aim of describing the strongly crystal-face dependence of the surface-plasmon
energy dispersion in Ag, Feibelman [110] thought of the Ag surface plasmon as a collective
mode that is split off the bottom of the 4d-to-5s electronhole excitation band, and considered
the relation between the surface-plasmon dispersion and the 4d-to-5s excitations induced by
the surface-plasmons field. He calculated the sd matrix elements using a one-dimensional
surface perturbation of the jellium LangKohn potential, and he argued that for Ag(100) the
58
J M Pitarke et al
4
(100)
(111)
(110)[001]
(110)[011]
h (eV)
3.9
(100)
(111)
(110)[001]
(110)[011]
3.8
3.7
(a)
0
0.05
0.1
Q (1/ )
0.15
(b)
0.2 0
0.05
0.1
0.15
0.2
Q (1/ )
Figure 19. Surface-plasmon energy dispersion for the low-index faces (100), (111), and (110) of
Ag. In the case of Ag (110), the exhibited surface-plasmon energy dispersions correspond to wave
vectors along the [001] and [110] directions. (a) Experimental data from [106]. (b) The results
obtained within the jelliumdipolium model (from [114], used with permission).
dispersion coefficient is increased relative to Ag(111) because in the case of Ag(100) the 4d
electrons lie closer to the centroid of the oscillating free-electron charge, and the probability
of 4d-to-5s excitation is, therefore, enhanced.
Most recently, a self-consistent jellium model for valence 5s electrons in Ag was combined
with a so-called dipolium model in which the occupied 4d bands are represented in terms
of polarizable spheres located at the sites of a semi-infinite fcc lattice [114] to calculate
electron energy-loss spectra for all three low-index faces of Ag. The surface-plasmon energy
dispersions obtained from these spectra are exhibited in figure 19 together with the experimental
measurements [106]. This figure shows that the trend obtained for the different crystal
orientations is in qualitative agreement with the data. On the one hand, the surface-plasmon
dispersion of Ag(100) lies above that of the (111) surface; on the other hand, the slope for
Ag(110) is larger for wave vectors along the [001] direction than when the wave vector is
taken to have the [110] direction, illustrating the effect of the interplanar geometry on the
effective local fields. The observed overall variation of the measured positive slope with
crystal orientation is, however, considerably larger than predicted by the jelliumdipolium
model. This must be a signature of genuine band structure effects not captured in the present
model, which calls for a first-principles ab initio description of the electronic response of real
Ag surfaces.
The impact of the band structure on the surface-plasmon energy dispersion in Ag was
addressed by Moresco et al [284] and by Savio et al [285]. These authors performed test
experiments with the K/Ag(110) and O/Ag(001) systems and concluded that (i) the surfaceplasmon dispersion can be modified at will by manipulating the surface electronic structure
near s and (ii) surface interband transitions between Shockley states should be responsible for
the anomalous linear behaviour of the surface-plasmon dispersion in Ag(100). Furthermore,
HREELS experiments on sputtered and nanostructured Ag(100) have shown that the anomaly
exhibited by surface plasmons on Ag(100) can indeed be eliminated by modifying the surface
structure [286].
Hg. The sd polarization model devised and used by Liebsch to describe the positive energy
dispersion of surface plasmons in Ag [111] was also employed for a description of surface
plasmons in Hg [108].
59
Figure 20. Hg surface-plasmon energy dispersion, as obtained from standard jellium calculations
of the RPA and ALDA surface loss function g(q, ) of equation (7.74) (upper dashed and solid
curves, respectively), from stabilized jellium calculations of the RPA and ALDA surface loss
function g (q, ) of equation (8.27) (upper dashed and solid curves, respectively), and from the peak
positions observed in experimental electron energy-loss spectra (from [108], used with permission).
At q = 0, the measured surface-plasmon energy of this transition metal ( 6.9 eV) lies
about 1 eV below the value expected for a bounded electron gas with the density equal to the
average density of 6s2 valence electrons in Hg (rs = 2.65). This can be explained along the
lines described above for Ag (see equation (8.30)), with the use of a polarizable background
with d (s ) = 1.6.
At finite wave vectors, the sd screening in Hg is found to considerably distort the surfaceplasmon dispersion, as in the case of Ag, but now the linear coefficient of the low-q dispersion
being still negative though much smaller than in the absence of d electrons. This is illustrated
in figure 20, where the calculated (RPA and ALDA) and measured energy dispersions of
the Hg surface plasmon are compared with the corresponding dispersions obtained within a
standard jellium model (with rs = 2.65) in the absence of a polarizable medium. This figure
clearly shows that in addition to the overall lowering of the surface-plasma frequencies by
about 12% relative to the jellium calculations (with no d electrons) the sd screening leads to
a significant flattening of the energy dispersion, thereby bringing the jellium calculations into
nice agreement with experiment.
Simple calculations have also allowed to describe correctly the measured broadening of the
Hg surface plasmon. As shown in table 2, the Hg surface-plasmon width at q = 0 is very well
described by simply introducing the measured bulk dielectric function into equation (8.10).
For a description of the surface-plasmon broadening at finite q, Kim at [108] introduced
into equation (8.10) a Drude dielectric function of the form of equation (2.15), but with a
q-dependent finite defined as
(q) = (0) eqa ,
(8.31)
with a = 3 and (0) = s /s being the measured relative width at q = 0. This procedure
gives a surface-plasmon linewidth of about 1 eV for all values of q, in reasonable agreement
with experiment.
Pd. Collective excitations on transition metals with both sp and d-bands crossing the Fermi
level are typically strongly damped by the presence of interband transitions. Pd, however, is
known to support collective excitations that are relatively well defined.
60
J M Pitarke et al
Surface plasmons in Pd(110) were investigated with angle-resolved electron-energy-loss

spectroscopy by Rocca et al [281]. These authors observed a prominent loss feature with a
strongly negative linear initial dispersion, which they attributed to a surface-plasmon excitation.
At q = 0, they found s = 7.37 eV and s 2 eV, in agreement with optical data [287].
At low wave vectors from q = 0 up to q = 0.2 1 , they found a linear surface-plasmon
dispersion of the form of equation (8.15) with = 1 . This linear dispersion is considerably
stronger than in the case of simple metals (see table 1), and calls for a first-principles description
of this material where both occupied and unoccupied sp and d states be treated on an equal
footing.
Multipole surface plasmons. Among the noble and transition metals with occupied d bands,
multipole surface plasmons have only been observed recently in the case of Ag. In an
improvement over previous HREELS experiments, Moresco et al [120] performed ELS-LEED
experiments with both high-momentum and high-energy resolution, and by substracting the
data for two different impact energies they found a peak at 3.72 eV, which was interpreted
to be the Ag multipole plasmon. However, Liebsch argued that the frequency of the Ag
multipole surface plasmon should be in the 68 eV range above rather than below the bulk
plasma frequency, and suggested that the observed peak at 3.72 eV might not be associated
with a multipole surface plasmon [121].
Recently, the surface electronic structure and optical response of Ag have been studied on
the basis of angle- and energy-resolved photoyield experiments [124]. In these experiments,
the Ag multipole surface plasmon is observed at 3.7 eV, but no signature of the multipole
surface plasmon is observed above the plasma frequency (p = 3.8 eV) in disagreement with
the existing theoretical prediction [121].
8.4.3. First-principles calculations. First-principles calculations of the surface-plasmon
energy and linewidth dispersion of real solids have been carried out only in the case of the
simple-metal prototype surfaces Mg(0001) [118] and Al(111) [119]. These calculations were
performed by employing a supercell geometry with slabs containing 16(27) atomic layers
0

of Mg(Al) separated by vacuum intervals. The matrix g,g
(z, z ; q, ) was calculated from

equation (7.84), with the sum over n and n running over bands up to energies of 30 eV
above the Fermi level, the sum over k including 7812 points, and the single-particle orbitals
k,n;i (r) = k,n (r )i (z) being expanded in a plane-wave basis set with a kinetic-energy
cutoff of 12 Ry. The surface-response function gg=0,g =0 (q, ) was then calculated from
equation (7.86) including 300 3D reciprocal-lattice vectors in the evaluation of the RPA
or TDDFT Fourier coefficients g,g (z, z ; q, ) of equation (7.83). Finally, the dispersion of
the energy and linewidth of surface plasmons was calculated from the maxima of the imaginary
part of gg=0,g =0 (q, ) for various values of the magnitude and the direction of the 2D wave
vector q.
The ab initio calculations reported in [118,119] for the surface-plasmon energy dispersion
of Mg(0001) and Al(111) with the 2D wave vector along various symmetry directions show
that (i) there is almost perfect isotropy of the surface-plasmon energy dispersion, (ii) there is
excellent agreement with experiment (thereby accurately accounting for core polarization not
presented in jellium models), as long as the nonlocal (momentum-dependent) static XC localfield factor of equation (5.24) is employed in the evaluation of the interacting density-response
function, and (iii) if the corresponding jellium calculations are normalized to the measured
value s at q = 0 (as shown in figure 13 by the thick solid line with open circles) ab initio and
jellium calculations are found to be nearly indistinguishable.
61
Figure 21. Linewidth of surface plasmons in Mg(0001), as a function of the magnitude of the 2D
wave vector q. The filled (open) diamonds represent ab initio calculations [118], as obtained with
direction by employing the static XC local-field factor of equation (5.24)
q along the M ( K)
in the evaluation of the Fourier coefficients of the interacting density-response function. The solid
and long dashed lines represent the best fit of the ab initio calculations. The short dashed and dotted
lines correspond to jellium and 1D model-potential calculations, also obtained by employing the
static XC local-field factor of equation (5.24). Stars represent the experimental data reported
in [101].
Ab initio calculations also show that the band structure is of paramount importance for a
correct description of the surface-plasmon linewidth. First-principles TDDFT calculations of
the Mg(0001) and Al(111) surface-plasmon linewidth dispersions along various symmetry
directions are shown in figures 21 and 22, respectively. Also shown in this figures are
the experimental measurements reported in [95, 101] (stars) and the corresponding jellium
calculations (dashed lines). For small 2D wave vectors, the agreement between theory and
experiment is not as good as in the case of the surface-plasmon energy dispersion (which can be
attributed to finite-size effects of the supercell geometry). Nevertheless, ab initio calculations
for Mg(0001) (see figure 21) yield a negative slope for the linewidth dispersion at small q, in
agreement with experiment, and properly account for the experimental linewidth dispersion at
intermediate and large wave vectors. Figure 21 also shows that the Mg(0001) surface-plasmon
linewidth dispersion depends considerably on the direction of the 2D wave vector, and that
the use of the parametrized one-dimensional model potential reported in [248] (dotted line)
does not improve the jellium calculations. As for Al(111), we note that band-structure effects
bring the jellium calculations closer to experiment; however, ab initio calculations are still
in considerable disagreement with the measurements reported in [95, 264], especially at the
lowest wave vectors. At the moment it is not clear the origin of such a large discrepancy
between theory and experiment. On the other hand, considerable disagreement is also found
between the measured surface-plasmon linewidth at q 0 reported in [95] (s 1.9 eV)
and [264] (s 2.3 eV) and the value derived (see table 2) from the imaginary part of
the surface-response function of equation (8.10) with measured values of the bulk dielectric
function (s = 0.38 eV).
62
J M Pitarke et al
Figure 22. Linewidth of surface plasmons in Al(111), as a function of the magnitude of the 2D wave
vectors q. The filled (open) diamonds represent ab initio ALDA calculations, as obtained with q
[119]. The thin solid line represents the best fit of the ab initio calculations. The
along the M ( K)
dashed (solid) line represents the jellium model (1D model potential) dispersion. Long dashed and
dotted lines are the experimental data of [95] for Al(111) and (100) respectively. The open circle at
q = 0 is the linewidth derived from optical measurements of the Al bulk dielectric function [163].
The open square at q = 0 shows the linewidth obtained from surface-loss measurements [163].
The measured values from [264] are represented by dots with error bars.
Figure 23. Schematic representation of the surface band structure on Cu(111) near the point.
The shaded region represents the projection of the bulk bands.
63
Table 5. 2D Fermi energy (F2D ) of surface states at the point of Be(0001) and the (111) surfaces
of the noble metals Cu, Ag, and Au. vF2D and m2D represent the corresponding 2D Fermi velocity
and effective mass, respectively. vF2D is expressed in units of the Bohr velocity v0 = e2 /h.
Be(0001)
Cu(111)
Ag(111)
Au(111)
F2D (eV)
vF2D /v0
m2D
2.75
0.44
0.065
0.48
0.41
0.28
0.11
0.35
1.18
0.42
0.44
0.28
8.5. Acoustic surface plasmons

A variety of metal surfaces, such as Be(0001) and the (111) surfaces of the noble metals Cu,
Ag, and Au, are known to support a partially occupied band of Shockley surface states within
a wide energy gap around the Fermi level (see figure 23) [288, 289]. Since these states are
strongly localized near the surface and disperse with momentum parallel to the surface, they
can be considered to form a quasi 2D surface-state band with a 2D Fermi energy F2D equal to
the surface-state binding energy at the point (see table 5).
In the absence of the 3D substrate, a Shockley surface state would support a 2D collective
oscillation, the energy of the corresponding plasmon being given by [173]
2D = (2n2D q)1/2 ,
2D
with n
(8.32)
being the 2D density of occupied surface states, i.e,

n2D = F2D /.
(8.33)
Equation (8.32) shows that at very long wavelengths plasmons in a 2D electron gas have low
energies; however, they do not affect eph interaction and phonon dynamics near the Fermi
level, due to their square-root dependence on the wave vector. Much more effective than
ordinary 2D plasmons in mediating, e.g. superconductivity would be the so-called acoustic
plasmons with sound-like long-wavelength dispersion.
Recently, it has been shown that in the presence of the 3D substrate the dynamical screening
at the surface provides a mechanism for the existence of a new acoustic collective mode, the
so-called acoustic surface plasmon, whose energy exhibits a linear dependence on the 2D wave
number [125127]. This novel surface-plasmon mode has been observed at the (0001) surface
of Be, showing a linear energy dispersion that is in very good agreement with first-principles
calculations [128].
8.5.1. A simple model. First of all, we consider a simplified model in which surface-state
electrons comprise a 2D electron gas at z = zd (see figure 24), while all other states of the semiinfinite metal comprise a 3D substrate at z 0 represented by the Drude dielectric function of
equation (2.15). Within this model, one finds that both eh and collective excitations occurring
within the 2D gas can be described with the use of an effective 2D dielectric function, which
in the RPA takes the form [127]
2D
0
eff
(q, ) = 1 W (zd , zd ; q, ) 2D
(q, ),
(8.34)
0
W (z, z ; q, ) being the screened interaction of equation (7.59), and 2D
(q, ) being the
noninteracting density-response function of a homogeneous 2D electron gas [173].
In the absence of the 3D substrate, W (z, z ; q, ) reduces to the bare Coulomb interaction
2D
v(z, z ; q), and eff
(q, ) coincides, therefore, with the RPA dielectric function of a 2D electron
64
J M Pitarke et al
Figure 24. Surface-state electrons comprise a 2D sheet of interacting free electrons at z = zd . All
other states of the semi-infinite metal comprise a plane-bounded 3D electron gas at z 0. The
metal surface is located at z = 0.
gas, which in the long-wavelength (q 0) limit has one single zero corresponding to collective
excitations at = 2D .
2D
In the presence of a 3D substrate, the long-wavelength limit of eff
(q, ) is found to have
two zeros. One zero corresponds to a high-frequency ( vF q) oscillation of energy
2
,
2 = s2 + 2D
(8.35)
in which 2D and 3D electrons oscillate in phase with one another. The other zero corresponds
to a low-frequency acoustic oscillation in which both 2D and 3D electrons oscillate out of
phase; for this zero to be present, the long-wavelength (q 0) limit of the low-frequency
( 0) screened interaction W (zd , zd ; q, ),
I (zd ) = lim W (zd , zd ; q, 0),
q0
(8.36)
must be different from zero. The energy of this low-frequency mode is then found to be of the
form [127]
= vF2D q,
where

=
1+
[I (zd )]2
.
[ + 2 I (zd )]
(8.37)
(8.38)
Local 3D response. If one characterizes the 3D substrate by a Drude dielectric function of

the form of equation (2.15), the 3D screened interaction W (zd , zd ; q, vF q) is easily found to
be
zd 0
0
W (zd , zd ; q, vF q) =
.
(8.39)
4 zd
zd > 0
Hence, in the presence of a 3D substrate that is spatially separated from the 2D sheet
(zd > 0), introduction of I (zd ) = 4 zd into equation (8.38) yields at zd 1:

(8.40)
= 2zd ,
65
which is the result first obtained by Chaplik in his study of charge-carrier crystallization in
low-density inversion layers [290].
If the 2D sheet is located inside the 3D substrate (zd 0), I (zd ) = 0, which means
that the effective dielectric function of equation (8.34) has no zero at low energies ( < s ).
Hence, in a local picture of the 3D response the characteristic collective oscillations of the 2D
electron gas would be completely screened by the surrounding 3D substrate, and no low-energy
acoustic mode would exist. This result had suggested over the years that acoustic plasmons
should only exist in the case of spatially separated plasmas, as pointed out by Das Sarma and
Madhukar [291].
Nevertheless, Silkin et al [125] have demonstrated that metal surfaces where a partially
occupied quasi-2D surface-state band coexists in the same region of space with the underlying
3D continuum can indeed support a well-defined acoustic surface plasmon. This acoustic
collective oscillation has been found to appear as the result of a combination of the nonlocality
of the 3D dynamical screening and the spill out of the 3D electron density into the
vacuum [127].
Self-consistent 3D response. Figures 25(a) and 25(b) exhibit self-consistent RPA calculations
of the effective dielectric function of equation (8.34) corresponding to the (0001) surface of
Be, with q = 0.01 a01 (figure 25a) and q = 0.1 a01 (figure 25b). The real and imaginary parts
of the effective dielectric function (thick and thin solid lines, respectively) have been displayed
for zd = 0, as approximately occurs with the quasi-2D surface-state band in Be(0001).
Also

2D
shown in these figures (dotted lines) is the energy-loss function Im 1/eff
(q) for the 2D
sheet located inside the metal at zd = F , at the jellium edge (zd = 0), and outside the metal
at zd = F /2 and zd = F .
2D
2D
Collective excitations are related to a zero of Re eff
(q, ) in a region where Im eff
(q, )
2D
is small, which yields a maximum in the energy-loss function Im 1/eff (q, ) . For the 2D
electron density under study (rs2D = 3.12), in the absence of the 3D substrate a 2D plasmon
would occur at 2D = 1.22 eV for q = 0.01a01 and 2D = 3.99 eV for q = 0.1a01 . However,
in the presence of the 3D substrate (see figure 25) a well-defined low-energy acoustic plasmon
occurs at energies (above the upper edge u = vF q + q 2 /2 of the 2D eh pair continuum)
dictated by equation (8.37) with the sound velocity vF being just over the 2D Fermi velocity
vF ( > 1). When the 2D sheet is located far inside the metal surface, the sound velocity is
found to approach the Fermi velocity ( 1). When the 2D sheet
is located far outside the
metal surface, the coefficient approaches the classical limit ( 2zd ) of equation (8.40).
Finally, we note that apart from the limiting case zd = F and q = 0.01 a01 (which
yields a plasmon linewidth negligibly small) the small width of the plasmon peak exhibited in
figure 25 is entirely due to plasmon decay into eh pairs of the 3D substrate.
8.5.2. 1D model calculations. For a more realistic (but still simplified) description of
electronic excitations at the (0001) surface of Be and the (111) surface of the noble metals Cu,
Ag, and Au, the 1D model potential vMP (z) of [248] was employed in [125, 126]. The use
of this model potential allows to assume translational invariance in the plane of the surface
and to trace, therefore, the presence of collective excitations to the peaks of the imaginary
part of the jellium-like surface response function of equation (7.73). The important difference
between the screened interaction W (z, z ; q, ) used here to evaluate g(q, ) and the screened
interaction used in equation (8.34) in the framework of the simple model described above
lies in the fact that the single-particle orbitals i (z) and energies i are now obtained by
solving equation (7.65) with the jellium KohnSham potential vKS [n0 ](z) replaced by the
66
J M Pitarke et al
50
(a)
zd=F
zd=-F
25
zd=0
(q,)
zd=F/2
x 0.2
x 0.2
-25
0.1
0.2
(eV)
0.3
0.4
0.5
20
(b)
zd=F
10
zd=-F
zd=F/2
(q,)
zd=0
-10
(eV)
Figure 25. Effective dielectric function of a 2D sheet that is located at the jellium edge (zd = 0), as
obtained from equation (8.34) and self-consistent RPA calculations of the 3D screened interaction
W (zd , zd ; q, ) and the 2D density-response function 2D (q, ) with (a) q = 0.01 and (b)
q = 0.1 [127]. The real and imaginary parts of eff (q, ) are represented by thick and thin solid
1
lines, respectively. The dotted lines represent the effective 2D energy-loss function Im[eff
(q, )]
for zd = F , zd = F /2 and zd = F . The vertical dashed line represents the upper edge
u = vF q + q 2 /2m of the 2D eh pair continuum, where 2D eh pairs can be excited. The
calculations presented here for zd = F and q = 0.01 a01 have been carried out by replacing the
energy by a complex quantity + i with = 0.05 eV. All remaining calculations have been
carried out for real frequencies, i.e. with = 0. The 2D and 3D electron-density parameters have
been taken to be rs2D = 3.14 and rs = 1.87, corresponding to Be(0001).
model potential vMP (z). Since this model potential has been shown to reproduce the key
features of the surface band structure and, in particular, the presence of a Shockley surface
state within an energy gap around the Fermi level of the materials under study, it provides a
realistic description of surface-state electrons moving in the presence of the 3D substrate.
Figure 26 shows the imaginary part of the surface-response function of equation (7.73), as
obtained for Be(0001) and for increasing values of q by using the 1D model potential vMP (z)
67
0.6
-1
Img(q,)/ (eV )
0.4
0.2
0
0.4
0.2
0
0.4
0.2
0
0
10
12
14
16
Energy (eV)
Figure 26. Energy loss function Im[g(q, )]/ of Be(0001) versus the excitation energy ,
obtained from equation (7.73) with the use of the 1D model potential of [248], as reported in [125].
The magnitude of the wave vector q has been taken to be 0.05 (top panel), 0.1 (middle panel), and
0.15 (bottom panel), in units of the inverse Bohr radius a01 . In the long-wavelength limit (q 0),
g(q, ) is simply the total electron density induced by the potential of equation (7.75).
Figure 27. Energy loss function Im[g(q, )]/ of Be(0001) versus the excitation energy obtained
from equation (7.73) with the use of the 1D model potential of [248] and for various values of q [293].
of [248]. As follows from the figure, the excitation spectra is clearly dominated by two distinct
features: (i) the conventional surface plasmon at h
s 13 eV, which can be traced to the
characteristic pole that the surface-response function g(q, ) of a bounded 3D free electron
gas with rs3D = 1.87 exhibits at this energy23 , and (ii) a well-defined low-energy peak with
linear dispersion.
That the low-frequency mode that is visible in figure 26 has linear dispersion is clearly
shown in figure 27, where the imaginary part of the surface-response function g(q, ) of
Be(0001) is displayed at low energies for increasing values of the magnitude of the wave
vector in the range q = 0.01 0.12 a01 . The excitation spectra is indeed dominated at low
energies by a well-defined acoustic peak at energies of the form of equation (8.37) with an
coefficient that is close to unity, i.e. the sound velocity being at long wavelengths very close
to the 2D Fermi velocity vF2D (see table 5).
The energy-loss function Im g(q, ) of the (111) surfaces of the noble metals Cu, Ag,
23
This surface plasmon would also be present in the absence of surface states, i.e. if the 1D model potential vMP
of [248] were replaced by the jellium KohnSham potential of equation (5.13).
68
J M Pitarke et al
Figure 28. Energy-loss function Im[g(q, )]/ of the (111) surfaces of the noble metals Cu, Ag
and Au [126], shown by solid, dashed, and dashed-dotted lines, respectively, versus the excitation
energy , as obtained from equation (7.73) with the use of the 1D model potential of [248] and
for q = 0.01 a01 and = 1 meV. The vertical solid lines are located at the energies = vF2D q,
which would correspond to equation (8.37) with = 1.
Figure 29. The solid line shows the energy of the acoustic surface plasmon of Be(0001) [125], as
obtained from the maxima of the calculated surface-loss function Im g(q, ) shown in figure 27.
The thick dotted line and the open circles represent the maxima of the energy-loss function
Im[1/eff (q, )] obtained from equation (8.34) with zd far inside the solid (thick dotted line)
and with zd = 0 (open circles). The dashed line is the plasmon dispersion of a 2D electron gas in
the absence of the 3D system. The grey area indicates the region of the (q,) plane (with the upper
up
limit at 2D = vF2D q + q 2 /2m2D ) where eh pairs can be created within the 2D Shockley band of
Be(0001). The area below the thick solid line corresponds to the region of momentum space where
min and q min are determined
transitions between 3D and 2D states cannot occur. The quantities inter
from the surface band structure of Be(0001). 2D and 3D electron densities have been taken to be
those corresponding to the Wigner radii rs2D = 3.14 and rs3D = 1.87, respectively.
and Au is displayed in figure 28 for q = 0.01 a01 . This figure shows again the presence of a
well-defined low-energy collective excitation whose energy is of the form of equation (8.37)
with 1.
Figure 29 shows the energy of the acoustic surface plasmon of Be(0001) versus q (solid
line), as derived from the maxima of the calculated Im g(q, ) of figure 27 (solid line) and
69
Figure 30. A periodic grating of constant L. The grating periodic structure can provide an
impinging free electromagnetic radiation with additional momentum 2/L.
from the maxima of the effective energy-loss function Im[1/eff (q, )] (see figure 25)
obtained from equation (8.34) for zd = 0 (dashed line). Little discrepancies between these
two calculations should be originated in (i) the absence in the simplified model leading to
equation (8.34) of transitions between 2D and 3D states, and (ii) the nature of the decay
and penetration of the surface-state orbitals, which in the framework of the model leading to
equation (8.34) are assumed to be fully localized in a 2D sheet at z = zd .
8.5.3. First-principles calculations. First-principles calculations of the imaginary part of the
surface-response function gg=0,g=0 (q, ) of Be(0001) have been carried out recently [128],
and it has been found that this metal surface is indeed expected to support an acoustic surface
plasmon whose energy dispersion agrees with the solid line represented in figure 29 (if the
dispersion of figure 29 calculated for surface state effective mass m = 1 is scaled according to
the ab initio value m = 1.2 [292]). Furthermore, these calculations have been found to agree
closely with recent high-resolution EELS measurements on the (0001) surface of Be [128]
(under grazing incidence), which represent the first evidence of the existence of acoustic
surface plasmons.
8.5.4. Excitation of acoustic surface plasmons. As in the case of the conventional surface
plasmon at the Ritchies frequency s , acoustic surface plasmons should be expected to
be excited not only by moving electrons (as occurs in the EELS experiments reported
recently [128]) but also by light. Now we focus on a possible mechanism that would lead
to the excitation of acoustic surface plasmons by light in, e.g. vicinal surfaces with high
indices [293].
At long wavelengths (q 0), the acoustic surface-plasmon dispersion curve is of the
form of equation (8.37) with 1. As the 2D Fermi velocity vF2D is typically about three
orders of magnitude smaller than the velocity of light, there is, in principle, no way that incident
light can provide an ideal surface with the correct amount of momentum and energy for the
excitation of an acoustic surface plasmon to occur. As in the case of conventional surface
plasmons, however, a periodic corrugation or grating in the metal surface should be able to
provide the missing momentum.
Let us consider a periodic grating of constant L (see figure 30). If light hits such a
surface, the grating periodic structure can provide the impinging free electromagnetic waves
with additional momentum arising from the grating periodic structure. If free electromagnetic
radiation hits the grating at an angle , its wave vector along the grating surface has magnitude
2
q=
sin
n,
(8.41)
c
L
where L represents the grating constant, and n = 1, 2, . . .. Hence, the linear (nearly vertical)
dispersion relation of free light changes into a set of parallel straight lines, which can match
the acousticplasmon dispersion relation as shown in figure 31.
For a well-defined acoustic surface plasmon in Be(0001) to be observed, the wave number q
needs to be smaller than q 0.06 a01 (see figure 29)24 . For q = 0.05 a01 , equation (8.41) with
24
Acoustic surface plasmons in Be(0001) have been observed for energies up to 2 eV [128].
70
J M Pitarke et al
Figure 31. A schematic representation of the dispersion relation of acoustic surface plasmons
(solid line) and free light impinging on a periodic grating of constant L (essentially vertical dotted
lines).
n = 1 yields a grating constant L = 66 . Acoustic surface plasmons of energy 0.6 eV

could be excited in this way. Although a grating period of a few nanometres sounds unrealistic
with present technology, the possible control of vicinal surfaces with high indices could provide
appropriate grating periods in the near future.
9. Applications
Surface plasmons have been employed over the years in a wide spectrum of studies ranging
from condensed matter and surface physics [924] to electrochemistry [25], wetting [26],
biosensing [27, 28, 29], scanning tunnelling microscopy [30], the ejection of ions from
surfaces [31], nanoparticle growth [32, 33], surface-plasmon microscopy [34, 35] and
surface-plasmon resonance technology [3642]. Renewed interest in surface plasmons has
come from recent advances in the investigation of the optical properties of nanostructured
materials [43, 44], one of the most attractive aspects of these collective excitations now being
their use to concentrate light in subwavelength structures and to enhance transmission through
periodic arrays of subwavelength holes in optically thick metallic films [45, 46], as well as the
possible fabrication of nanoscale photonic circuits operating at optical frequencies [48] and
their use as mediators in the transfer of energy from donor to acceptor molecules on opposite
sides of metal films [67].
Here we focus on two distinct applications of collective electronic excitations at metal
surfaces: the role that surface plasmons play in particlesurface interactions and the new
emerging field called plasmonics.
9.1. Particlesurface interactions: energy loss

Let us consider a recoilless fast point particle of charge Z1 moving in an arbitrary
inhomogeneous many-electron system at a given impact vector b with nonrelativistic velocity
v, for which retardation effects and radiation losses can be neglected25 . Using Fermis golden
rule of time-dependent perturbation theory, the lowest-order probability for the probe particle
25 This approximation is valid for heavy charged particles, e.g. ions and also for swift electrons moving with
nonrelativistic velocities that are large compared to the velocity of target electrons. In the case of electrons, Z1 = 1.
to transfer momentum q to the medium is given by the following expression [294]:

4 2
dq ib(q+q )
Z1
d
e
Pq =
LA
(2)3
0
ImW (q, q ; ) ( q v) ( + q v),
71
(9.1)
where L and A represent the normalization length and area, respectively, and W (q, q ; ) is
the double Fourier transform of the screened interaction W (r, r ; ) of equation (7.4):

(9.2)
W (q, q ; ) = dr dr ei(qr+q r ) W (r, r ; ).
Alternatively, the total decay rate 1 of the probe particle can be obtained from the
knowledge of the imaginary part of the self-energy. In the GW approximation of many-body
theory [295], and replacing the probe-particle Green function by that of a non-interacting
recoilless particle, one finds [250]:

dr dr i (r) f (r )ImW (r, r ; i f ) i (r ) f (r),
(9.3)
1 = 2 Z12
f
where i (r) represents the probe-particle initial state of energy i , and the sum is extended
over a complete set of final states f (r) of energy f . Describing the probe-particle initial and
final states by plane waves in the direction of motion and a Dirac function in the transverse
direction, i.e.

1
(9.4)
(r) = eivr (r b),
A
where r represents the position vector perpendicular to the projectile velocity, one finds that
the decay rate of equation (9.3) reduces indeed to a sum over the probability Pq of equation (9.1),
i.e.
1
Pq ,
(9.5)
1 =
T q
T being a normalization time.
For a description of the total energy E that the moving probe particle loses due to
electronic excitations in the medium, one can first define the time-dependent probe-particle
charge density
ext (r, t) = Z1 (r b v t),
(9.6)
and one then obtains the energy that this classical particle loses per unit time as follows [87]

V ind (r, t)
dE
= dr ext (r, t)
,
(9.7)
dt
t
where V ind (r, t) is the potential induced by the probe particle at position r and time t, which
to first order in the external perturbation yields (see equation (7.1)):
+
+

d i(tt ) ind
W (r, r ; ) ext (r , t )
dt
(9.8)
e
V ind (r, t) = dr
2
with
W (r, r ; ) = W (r, r ; ) v(r, r ),
(9.9)
72
J M Pitarke et al
Figure 32. Particle of charge Z1 moving with constant velocity at a fixed distance z from the
surface of a plane-bounded electron gas. Inside the solid, the presence of the surface causes (i) a
decrease of loss at the bulk-plasmon frequency p varying with z as K0 (2p z/v) and (ii) an
additional loss at the surface-plasmon frequency s varying with z as K0 (2s z/v). Outside the
solid, energy losses are dominated by a surface-plasmon excitation at s .
Finally, one writes:

E =
dE
dt
dt

.
(9.10)
Introducing equations (9.6) and (9.8) into equation (9.7), and equation (9.7) into
equation (9.10), one finds that the total energy loss E can indeed be written as a sum over
the probability Pq of equation (9.1), i.e.

E =
(9.11)
(q v) Pq ,
q
where q v is simply the energy transferred by our recoilless probe particle to the medium.
9.1.1. Planar surface. In the case of a plane-bounded electron gas that is translationally
invariant in two directions, which we take to be normal to the z axis, equations (9.6)(9.8)
yield the following expression for the energy that the probe particle loses per unit time:
+

Z2
dE
d2 q

dt
d ei(qv )(tt ) W [z(t), z(t ); q, ],
(9.12)
=i 1
dt
(2)2
0
where q is a 2D wave vector in the plane of the surface, v represents the component of the
velocity that is parallel to the surface, z(t) represents the position of the projectile relative to
the surface, and W (z, z ; q, ) is the 2D Fourier transform of W (r, r ; ).
In the simplest possible model of a bounded semi-infinite electron gas in vacuum, in which
the screened interaction W (z, z ; q, ) is given by the classical expression equation (7.8) with
1 being the Drude dielectric function of equation (2.15) and 2 = 1, explicit expressions can
be found for the energy lost per unit path length by probe particles that move along a trajectory
that is either parallel or normal to the surface.
Parallel trajectory. In the case of a probe particle moving with constant velocity at a fixed
distance z from the surface (see figure 32), introduction of equation (7.8) into equation (9.12)
yields [293]
dE
=
dx
Z12
v2

2
2
p ln(kc v/p ) K0 (2p z/v) + s K0 (2s z/v),
73
z < 0,
(9.13)
s2 K0 (2s |z|/v),
z > 0,
where K0 () is the zero-order modified Bessel function [241], and kc denotes the magnitude
of a wave vector above which long-lived bulk plasmons are not sustainable.
For particle trajectories outside the solid (z > 0), equation (9.13) reproduces the classical
expression of Echenique and Pendry [22], which was found to describe correctly EELS
experiments [131] and which was extended to include relativistic corrections [296]. For
particle trajectories inside the solid (z > 0), equation (9.13) reproduces the result first obtained
by Nunez et al [297]. Outside the solid, the energy loss is dominated by the excitation of
surface plasmons at s . When the particle moves inside the solid, the effect of the boundary
is to cause (i) a decrease in loss at the bulk plasma frequency p , which in an infinite electron
gas would be dE/dx = Z12 p2 ln(kc v/p )/v 2 and (ii) an additional loss at the surface-plasma
frequency s .
Nonlocal effects that are absent in the classical equation (9.13) were incorporated
approximately by several authors in the framework of the hydrodynamic approach and the
specular-reflection model described in sections 7.2.1 and 7.2.2 [298303]. More recently,
extensive RPA and ALDA calculations of the energy-loss spectra of charged particles moving
near a jellium surface were carried out [245] within the self-consistent scheme described in
section 7.2.3. At high velocities (of a few Bohr units) and for charged particles moving far from
the surface into the vacuum, the actual energy loss was found to converge with the classical
limit dictated by the first line of equation (9.13). However, at low and intermediate velocities
substantial changes in the energy loss were observed as a realistic description of the surface
response was considered.
Corrections to the energy loss of charged particles (moving far from the surface into the
vacuum) due to the finite width of the surface-plasmon resonance that is not present, in principle,
in jellium self-consistent calculations, have been discussed recently [304]. These corrections
have been included to investigate the energy loss of highly charged ions undergoing distant
collisions at grazing incidence angles with the internal surface of microcapillary materials, and
it has been suggested that the correlation between the angular distribution and the energy loss
of transmitted ions can be used to probe the dielectric properties of the capillary material.
For a more realistic description of the energy loss of charged particles moving near
a Cu(111) surface, the KohnSham potential vKS (z) used in the self-consistent jellium
calculations of [245] was replaced in [305] by the 1D model potential vMP of [248]. It was
shown, however, that although the Cu(111) surface exhibits a wide band gap around the Fermi
level and a well-defined Shockley surface state the energy loss expected from this model does
not differ significantly from its jellium counterpart. This is due to the fact that the presence of
the surface state compensates the reduction of the energy loss due to the band gap.
Existing first-principles calculations of the interaction of charged particles with solids
invoke periodicity of the solid in all directions and neglect, therefore, surface effects and,
in particular, the excitation of surfaces plasmons [306, 307]. An exception is a recent firstprinciples calculation of the energy loss of ions moving parallel with a Mg(0001) surface [308],
which accounts naturally for the finite width of the surface-plasmon resonance that is present
neither in the self-consistent jellium calculations of [245] nor in the 1D model calculations
of [305].
A typical situation in which charged particles can be approximately assumed to move along
a trajectory that is parallel to a solid surface occurs in the glancing-incidence geometry, where
74
J M Pitarke et al
ions penetrate into the solid, they skim the outermost layer of the solid, and are then specularly
repelled by a repulsive, screened Coulomb potential, as discussed by Gemell [309]. By first
calculating the ion trajectory under the combined influence of the repulsive planar potential and
the attractive image potential, the total energy loss can be obtained approximately as follows

1

dE
dz,
(9.14)
(z) vz (z)
E = 2v
ztp dx
ztp and vz (z) denoting the turning point and the value of the component of the velocity normal
to the surface, respectively, which both depend on the angle of incidence.
Accurate measurements of the energy loss of ions being reflected from a variety of solid
surfaces at grazing incidence have been reported by several authors [310314]. In particular,
Winter et al [312] carried out measurements of the energy loss of protons being reflected from
Al(111). From the analysis of their data at 120 keV, these authors deduced the energy loss
dE/dx(z) and found that at large distances from the surface the energy loss follows closely
the energy loss expected from the excitation of surface plasmons. Later on, RPA jellium
calculations of the energy loss from the excitation of valence electrons were combined with a
first-Born calculation of the energy loss due to the excitation of the inner shells and reasonably
good agreement with the experimental data was obtained for all angles of incidence [315].
Normal trajectory. Let us now consider a situation in which the probe particle moves along a
normal trajectory from the vacuum side of the surface (z > 0) and enters the solid at z = t = 0.
The position of the projectile relative to the surface is then z(t) = vt. Assuming that the
electron gas at z 0 can be described by the Drude dielectric function of equation (2.15) and
introducing equation (7.8) into equation (3.8) yields [293]:

2
ln(kc v/p ) h(p z/v) + s2 h(s z/v),
z<0
Z12 p
dE
= 2
(9.15)
dz
v 2
s f (2s |z|/v),
z > 0,
where
h() = 2 cos() f () f (2),
(9.16)
with f () being given by the following expression:

x ex
dx.
(9.17)
f () =
1 + x2
0
Equation (9.15) shows that (i) when the probe particle is moving outside the solid the effect
of the boundary is to cause energy loss at the surface-plasmon energy s , and (ii) when the
probe particle is moving inside the solid the effect of the boundary is to cause both a decrease
in loss at the bulk-plasmon energy p and an additional loss at the surface-plasmon energy s ,
as predicted by Ritchie [1].
Now we consider the real situation in which a fast charged particle passes through a finite
foil of thickness a (see figure 33). Assuming that the foil is thick enough for the effect of each
boundary to be the same as in the case of a semi-infinite medium, and integrating along the
whole trajectory from minus to plus infinity, one finds the total energy that the probe particle
loses to collective excitations:

Z2
kc v
(9.18)
E = 21 a p2 ln
p + s .
v
p
2
This is the result first derived by Ritchie in a different way [1], which brought him to
the realization that surface collective excitations exist at the lowered frequency s . The first
75
Figure 33. Particle of charge Z1 passing perpendicularly through a finite foil of thickness a. The
presence of the boundaries leads to a decrease in the energy loss at the bulk-plasmon frequency
p and an additional loss at the surface-plasmon frequency s , the net boundary effect being an
increase in the total energy loss in comparison to the case of a particle moving in an infinite medium
with no boundaries.
term of equation (9.18), which is proportional to the thickness of the film represents the bulk
contribution, which would also be present in the absence of the boundaries. The second and
third terms, which are both due to the presence of the boundaries and become more important
as the foil thickness decreases, represent the decrease in the energy loss at the plasma frequency
p and the energy loss at the lowered frequency s , respectively. Equation (9.18) also shows
that the net boundary effect is an increase in the total energy loss above the value which would
exist in its absence, as noted by Ritchie [1]. A more accurate jellium self-consistent description
of the energy loss of charged particles passing through thin foils has been performed recently
in the RPA and ALDA [316].
9.2. STEM: valence EELS
The excitation of both surface plasmons on solid surfaces and localized Mie plasmons on
small particles has attracted great interest over the years in the fields of scanning transmission
electron microscopy [130135] and near-field optical spectroscopy [136].
EELS of fast electrons in STEM shows two types of losses, depending on the nature of
the excitations that are produced in the sample: atomically defined core-electron excitations
at energies > 100 eV and valence-electron (mainly collective) excitations at energies up
to 50 eV. Core-electron excitations occur when the probe moves across the target, and
provide chemical information about atomic-size regions of the target [317]. Conversely,
valence-electron excitations provide information about the surface structure with a resolution
of the order of several nanometres. One advantage of valence EELS is that it provides a
strong signal, even for non-penetrating trajectories (the so-called aloof beam energy loss
spectroscopy [318, 319]), and generates less specimen damage [131].
The central quantity in the interpretation of valence EELS experiments is the total
probability P () for the STEM beam to exchange energy with the sample. In terms of
the screened interaction W (r, r ; ) and for a probe electron in the state 0 (r) with energy 0 ,
first-order perturbation theory yields:

P () = 2
dr dr f (r ) 0 (r ) f (r) 0 (r)ImW (r, r ; ) [ 0 F ],
(9.19)
f
where the sum is extended over a complete set of final states f (r) of energy f . For probe
electrons moving on a definite trajectory (and having, therefore, a charge density of the form
76
J M Pitarke et al
of equation (9.6) with no beam recoil), the total energy loss E of equations (9.10) and (9.11)
can be expressed in the expected form

d P ().
(9.20)
E =
0
Ritchie and Howie [320] showed that, in EELS experiments where all the inelastic scattering
is collected, treating the fast electrons as a classical charge of the form of equation (9.6) is
indeed adequate. Nonetheless, quantal effects due to the spatial extension of the beam have
been addressed by several authors [133, 321, 322].
9.2.1. Planar surface. In the case of a classical beam of electrons moving with constant
velocity at a fixed distance z from a planar surface (see figure 32), the initial and final states
can be described by taking a function in the transverse direction and plane waves in the
direction of motion. Neglecting recoil, the probability P () of equation (9.19) then takes the
following form:

1
dqx ImW (z, z; q, ),
(9.21)
P () = 2
v 0

with q = qx2 + (/v)2 .
In a classical model in which the screened interaction W (z, z ; q, ) is given by the
classical expression equation (7.8), the probability P () is easy to calculate. In particular, if
the beam of electrons is moving outside the sample, one finds:

2
(9.22)
1 g e2qz /2 ,
W (z, z; q, ) =
q
with the surface-response function g given by equation (7.9). Introduction of equation (9.22)
into equation (9.21) yields the classical probability
2L
K0 (2z/v) Img(),
(9.23)
P () =
v 2
which in the case of a Drude metal in vacuum yields
s
(9.24)
P () = L 2 ( s ),
v
and, therefore (see equation (9.20)), the energy loss per unit path length given by equation (9.13)
with Z1 = 1 and z > 0.
The classical equation (9.22) can be easily extended to the case in which the sample is
formed by a semi-infinite medium characterized by 1 and covered by a layer of dielectric
function 3 and thickness a. One finds:

32 + 13 e2qa
2
,
(9.25)
1 e2qz
W (z, z; q, ) =
q
32 13 + e2qa
where
3 2
(9.26)
32 =
3 + 2
and
1 3
.
(9.27)
13 =
1 + 3
Equations (9.25)(9.27) show that while for a clean surface (a = 0) the energy-loss function
P () is dominated by the excitation of surface plasmons at = s (see equation (9.24)),
EELS should be sensitive to the presence of sub-surface structures. This effect was observed
by Batson [130] in the energy-loss spectra corresponding to an Al surface coated with an Al2 O3
layer of increasing thickness.
77
9.2.2. Spheres.
Definite trajectories. In the case of a classical beam of electrons moving with constant
velocity at a fixed distance b from the centre of a single sphere of radius a and dielectric
function 1 () that is immersed in a host medium of dielectric function 2 (), the initial and
final states can be described (as in the case of the planar surface) by taking a function in the
transverse direction and plane waves in the direction of motion. If recoil is neglected and the
classical screened interaction of equations (7.11)(7.13) is used, then equation (9.19) yields
the following expression for the energy-loss probability [132]:
P () =
a
2l
4a
m
v 2 l=1 m=0 (l m)(l + m) v
Km2 (b /v) Imgl (),
(9.28)
for outside trajectories (a b), and [323]

P () =

2

4a (l m) 0
m
Alm + Ailm Imgl () + Ailm A0lm + Ailm Im 1 () ,
2
v l=1 m=0
(l + m)
(9.29)
for inside trajectories (a b), with

l+1

b2 + x 2
x
1
0
m
clm (x/v)
dx
Pl
Alm () =
a c
a
b2 + x 2
(9.30)
and
l

x
1 c
b2 + x 2
clm (x/v).
=
dx
Plm
(9.31)
a 0
a
b2 + x 2
Here, c = a 2 b2 , Plm (x) denote Legendre functions, clm (x) = cos(x), if (l + m) is even,
os sin(x), if (l + m) is odd, and gl () denotes the classical function of equation (7.13), which
has poles at the classical surface-plasmon condition of equation (3.33).
Equations (9.28) and (9.29) show that an infinite number of multipolar modes can be
excited, in general, which contribute to the energy loss of moving electrons. It is known,
however, that for small spheres with a v/s the dipolar mode
(l = 1) dominates, which in
the case of a Drude sphere in vacuum occurs at 1 = p / 3. For spheres with a v/s ,
many multipoles contribute with similar weight. For very large spheres (a v/s ) and a < b,
the main contribution arises from high multipolar modes occurring approximately at the planar
surface plasmon energy s , since in the limit of large l gl () of equation (7.13) reduces to the
energy-loss function of equation (7.9). Indeed, this is an expected result, due to the fact that
for very large spheres the probe electron effectively interacts with an almost planar surface.
Ailm ()
Broad beam. We now consider a broad beam geometry, and we therefore describe the probe
electron states by plane waves of the form:
1
0 (r) = eik0 r
(9.32)

and
1
(9.33)
f (r) = eikf r ,

78
J M Pitarke et al
where represents the normalization volume. Then, introducing equations (9.32), (9.33) and
(7.11)(7.13) into equation (9.19), and neglecting recoil, one finds

1

dQ
1
Im eff
(Q, ) ( Q v).
(9.34)
P () = 2
2
Q
This is precisely the energy-loss probability corresponding to a probe electron moving in a 3D
homogeneous electron gas but with the inverse dielectric function 1 (Q, ) replaced by the
1
(Q, ) of equation (7.16).
effective inverse dielectric function eff
9.2.3. Cylinders.
Definite trajectory. In the case of a classical beam of electrons moving with constant velocity
at a fixed distance b from the axis of a single cylinder of radius a and dielectric function
1 () that is immersed in a host medium of dielectric function 2 (), the initial and final
states can be described (as in the case of the planar surface) by taking a function in the
transverse direction and plane waves in the direction of motion. If recoil is neglected and
the classical screened interaction of equations (7.19)(7.22) is used, then equation (9.19)
yields the following expression for the energy-loss probability [135]:
!
2
P () = Im (11 21 )
v

"

)Km ( a
)Im2 b
2 Km ( a
b
b
v
v
v
Km
+
,
(9.35)
m Im
v
v
1 Im ( a
)Km ( a
) 2 Im ( a
)Km ( a
)
v
v
v
v
m=0
for inside trajectories (a b), and
!
"

a
a
2 b

I
(
)I
(
)K
2
1
m
m
m
v
v
v
m
,
P () = Im (11 21 )
( a )K ( a ) I ( a )K ( a )
v

I
1
m
2
m
m
m v
v
v
v
m=0
(9.36)
for outside trajectories (a b). In particular, for axial trajectories (b = 0), the energy-loss
probability P () is due exclusively to the m = 0 mode.
Broad beam. For a broad beam geometry the probe electron states can be described by plane
waves of the form of equations (9.32) and (9.33), which after introduction into equation (9.19),
neglecting recoil, and using the classical screened interaction of equations (7.19)(7.22)
yield an energy-loss probability of the form of equation (9.34) but now with the inverse
1
dielectric function 1 (Q, ) replaced by the effective inverse dielectric function eff
(Q, )
of equation (7.25).
9.3. Plasmonics
Renewed interest in surface plasmons has come from recent advances in the investigation of
the electromagnetic properties of nanostructured materials [43, 44], one of the most attractive
aspects of these collective excitations now being their use to concentrate light in subwavelength
structures and to enhance transmission through periodic arrays of subwavelength holes in
optically thick metallic films [45, 46]. Surface-plasmon polaritons are tightly bound to metal
dielectric interfaces penetrating around 10 nm into the metal (the so-called skin-depth) and
typically more than 100 nm into the dielectric (depending on the wavelength), as shown in
figure 4. Indeed, surface plasmons of an optical wavelength concentrate light in a region
that is considerably smaller than their wavelength, a feature that suggests the possibility of
79
using surface-plasmon polaritons for the fabrication of nanoscale photonic circuits operating
at optical frequencies [47, 48].
Here we discuss only a few aspects of the most recent research that has been carried
out in the so-called field of plasmonics. This constitutes an important area of research, since
surface-plasmon based circuits are known to merge the fields of photonics and electronics at
the nanoscale, thereby enabling to overcome the existing difficulties related to the large size
mismatch between the micrometre-scale bulky components of photonics and the nanometrescale electronic chips. Indeed, the surface-plasmon polariton described in section 2 can serve
as a base for constructing nano-circuits that will be able to carry optical signals and electric
currents. These optoelectronic circuits would consist of various components such as couples,
waveguides, switches, and modulators.
In order to transmit optical signals to nanophotonic devices and to efficiently increase
the optical far-field to near-field conversion, a nanodot coupler (fabricated from a linear
array of closely spaced metallic nanoparticles) has been combined recently with a surfaceplasmon polariton condenser (working as a phase array) fabricated from hemispherical metallic
nanoparticles [49]. By focusing surface-plasmon polaritons with a spot size as small as 400 nm
at = 785 nm, their transmission length through the nanodot coupler was confirmed to be
three times longer than that of a metallic-core waveguide, owing to the efficient near-field
coupling between the localized surface plasmon of neighboring nanoparticles.
Achieving control of the lightsolid interactions involved in nanophotonic devices requires
structures that guide electromagnetic energy with a lateral mode confinement below the
diffraction limit of light. It was suggested that this so-called subwavelength-sized wave guiding
can occur along chains of closely spaced metal nanoparticles that convert the optical mode
into surface (Mie) plasmons [50]. The existence of guided long-range surface-plasmon waves
was observed experimentally by using thin metal films [5153], nanowires [54, 55], closely
spaced silver rods [56] and metal nanoparticles [57].
The propagation of guided surface plasmons is subject to significant ohmic losses that
limit the maximum propagation length. In order to avoid these losses, various geometries
have been devised using arrays of features of nanosize dimensions [58, 59]. The longest
propagation length (13.6 mm) has been achieved with a structure consisting of a thin lossy
metal film lying on a dielectric substrate and covered by a different dielectric superstrate [60].
The main issue in this context is to strongly confine the surface-plasmon field in the cross
section perpendicular to the surface-plasmon propagation direction, while keeping relatively
low propagation losses. Recently, it has been pointed out that strongly localized channel
plasmon polaritons (radiation waves guided by a channel cut into a planar surface of a
solid characterized by a negative dielectric function [61]) exhibit relatively low propagation
losses [62]. The first realization and characterization of the propagation of channel plasmon
polaritons along straight subwavelength metal grooves was reported by Bozhevilnyi et al
[63]. More recently, the design, fabrication, and characterization of channel plasmon
polariton based subwavelength waveguide components have been reported; these are Ysplitters, Mach-Zehnder interferometers and waveguide-ring resonators operating at telecom
wavelengths [64].
In the framework of plasmonics, modulators and switches have also been investigated.
Switches should serve as an active element to control surface-plasmon polariton waves [65,66].
This approach takes advantage of the strong dependence of the propagation of surfaceplasmon polaritons on the dielectric properties of the metal in a thin surface layer that may
be manipulated using light. This idea was realized introducing a few-micron long gallium
switching section to a gold-on-silica waveguide [66]. An example of an active plasmonic
device has been demonstrated by using a thin silver film covered from both sides by thin
80
J M Pitarke et al
polymer films with molecular chromophores [67]. In this case, coupled surface-plasmon
polaritons provide an effective transfer of excitation energy from donor molecules to acceptor
molecules on the opposite sides of a metal film up to 120 nm thick.
Another emerging area of active research in the field of plasmonics is based on the
generation and manipulation of electromagnetic radiation of various wavelengths from
microwave to optical frequencies. For instance, the coating of semiconductor quantum wells by
nanometre metal films results in an increased spontaneous emission rate in the semiconductor
that leads to the enhancement of light emission. This enhancement is due to an efficient
energy transfer from electronhole pair recombination in the quantum well to surface-plasmon
polaritons at the surface of semiconductor heterostructures coated by metal. Recently, a 32-fold
increase in the spontaneous emission rate at 440 nm in an InGaN/GaN quantum well has been
probed by time-resolved photoluminescence spectroscopy [324]. Also probed has been the
enhancement of photoluminescence up to an order of magnitude through a thin metal film from
organic light emitting diodes, by removing the surface-plasmon polariton quenching with the
use of a periodic nanostructure [325].
Recent theoretical and experimental work also suggest that surface-plasmon polaritons
play a key role both in the transmission of electromagnetic waves through a single aperture
and the enhanced transmission of light through subwavelength hole arrays [45,326328]. This
enhanced transmission has also been observed at millimetre-waves and micro-waves [329,330]
and at THz-waves [331].
Finally, we note that surface-plasmon polaritons have been used in the field of
nanolithography. This surface-plasmon based nanolithography can produce subwavelength
structures at surfaces, such as sub100 nm lines with visible light [332335]. On the other
hand, the enhancement of evanescent waves through the excitation of surface plasmons led
Pendry [336] to the concept of the so-called superlens [337].
Acknowledgments
Partial support by the University of the Basque Country, the Basque Unibertsitate eta Ikerketa
Saila, the Spanish Ministerio de Educacion y Ciencia and the EC 6th framework Network of
Excellence NANOQUANTA (Grant No NMP4-CT-2004-500198) is acknowledged. One of
us (JMP) would like to acknowledge fruitful discussions with A Howie and the hospitality of
Churchill College and the Cavendish Laboratory of the University of Cambridge, where part
of this review was written.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
Ritchie R H 1957 Phys. Rev. 106 874

Pines D and Bohm D 1952 Phys. Rev. 85 338
Pines D 1956 Rev. Mod. Phys. 28 184
Tonks L and Langmuir I 1929 Phys. Rev. 33 195
Ruthemann G 1942 Naturwiss 30 145
Ruthemann G 1948 Ann. Phys. 2 113
Lang W 1948 Optik 3 233
Powell C J and Swan J B 1959 Phys. Rev. 115 869
Powell C J and Swan J B 1959 Phys. Rev. 116 81
Stern E A and Ferrell R A 1960 Phys. Rev. 120 130
Inglesfield J E and Wikborg E 1975 J. Phys. F: Metal Phys. 5 1475
Zaremba E and Kohn W 1976 Phys. Rev. B 13 2270
Sernelius B E 2005 Phys. Rev. B 71 235114
Feibelman P J 1971 Surf. Sci. 27 438

[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
81
Ritchie R H 1972 Phys. Lett. A 38 189

Ray R and Mahan G D 1972 Phys. Lett. A 42 301
Sunji
c M, Toulouse G and Lucas A A 1972 Solid State Commun. 11 1629
Gadzuk J W and Metiu H 1980 Phys. Rev. B 22 2603
Schmit J and Lucas A 1972 Solid State Commun. 11 415
Schmit J and Lucas A 1972 Solid State Commun. 11 419
Wikborg E and Inglesfield J E 1977 Phys. Scr. 15 37
Lang N D and Sham L J 1975 Solid State Commun. 17 581
Langreth D C and Perdew J P 1977 Phys. Rev. B 15 2884
Chabal Y J and Sievers A J 1980 Phys. Rev. Lett. 44 944
Persson B and Ryberg R 1985 Phys. Rev. Lett. 54 2119
Echenique P M and Pendry J B 1975 J. Phys. C: Solid State Phys. 8 2936
Echenique P M, Ritchie R H, Barberan N and Inkson J 1981 Phys. Rev. B 23 6486
Ueba H 1992 Phys. Rev. B 45 3755
Knoll W 1998 Annu. Rev. Phys. Chem. 49 569
Herminghaus S, Vorberg J, Gau H, Conradt R, Reinelt D, Ulmer H, Leiderer P and Przyrembel M 1997 Ann.
Phys. Lpz. 6 425
Malmqvist M 1993 Nature 261 186
Braguglia C M 1998 Chem. Biochem. Eng. Q. 12 183
Chien F C and Chen S J 2004 Biosensors Bioelectron. 20 633
Berndt R, Gimzewski J K and Johansson P 1991 Phys. Rev. Lett. 67 3796
Shea M J and Compton R N 1993 Phys. Rev. B 47 9967
Jin R C, Cao Y W, Mirkin C A, Kelly K L, Schatz G C and Zheng J G 2001 Science 294 1901
Jin R, Cao C, Hao E, Metraux G S, Schatz G C and Mirkin C 2003 Nature 425 487
Rothenhausler B and Knoll W 1988 Nature 332 615
Flatgen G, Krischer K, Pettinger B, Doblhofer K, Junkes H and Ertl G 1995 Science 269 668
Gordon J G and Ernst S 1980 Surf. Sci. 101 499
Liedberg B, Nylander C and Lundstrom I 1983 Sensors Actuators 4 299
Schuster S C, Swanson R V, Alex L A Bourret R B and Simon M I 1993 Nature 365 343
Schuck P 1997 Curr. Opin. Biotechnol. 8 498
Schuck P 1997 Annu. Rev. Biophys. Biomol. Struct. 26 541
Homola J, Yee S S and Gauglitz G 1999 Sensors Actuators B 54 3
Mendelsohn A R and Brend R 1999 Science 284 1948
Green R J, Frazier R A, Shakesheff K M, Davies M C, Roberts C J and Tendler S J B 2000 Biomaterials
21 1823
Pendry J B 1999 Science 285 1687
Prodan E, Radloff C, Halas N J and Nordlander P 2003 Science 302 419
Ebbesen T W, Lezec H J, Ghaemi H F, Thio T and Wolff P A 1998 Nature 391 667
Lezec H J, Degiron A, Devaux E, Linke R A, Martin-Moreno L, Garcia-Vidal F J and Ebbesen T W 2002
Science 297 820
Ozbay E 2006 Science 311 189
Barnes W L, Dereux A and Ebbesen T W 2003 Nature 424 824
Nomura W, Ohtsu M and Yatsui T 2005 Appl. Phys. Lett. 86 181108
Quinten M, Leitner A, Krenn J R and Aussenegg F R 1998 Opt. Lett. 23 1331
Charbonneau R, Berini P, Berolo E and Lisicka-Shrzek E 2000 Opt. Lett. 25 844
Lamprecht B, Krenn J R, Schider G, Ditlbacher H, Salerno M, Felid j N, Leitner A, Aussenegg F R and
Weeber J C 2001 Appl. Phys. Lett. 79 51
Nikolajsen T, Leosson K, Salakhutdinov I and Bozhevolnyi S I 2003 Appl. Phys. Lett. 82 668
Krenn J R, Lamprecht B, Ditlbacher H, Schider G, Saleno M, Leitner A and Aussenegg F R 2002 Europhys.
Lett. 60 663
Krenn J R and Weeber J C 2004 Philos. Trans. R. Soc. Lond. Ser. A 362 739
Maier S A, Kik P G, Atwater H A, Meltzer S, Harel E, Koel B E and Requicha A A G 2003 Nature Mater.
2 229
Murray W A, Astilean S and Barens W L 2004 Phys. Rev. B 69 165407
Maier S A, Barclay P E, Johnson T J, Friedman M D and Painter O 2004 Appl. Phys. Lett. 84 3990
Maier S A, Friedman M D, Barclay P E and Painter O 2005 Appl. Phys. Lett. 86 071103
Berini P, Charbonneau R, Lahoud N and Mattiussi G 2005 J. Appl. Phys. 98 043109
Novikov I V and Maradudin A A 2002 Phys. Rev. B 66 035403
82
J M Pitarke et al
[62] Pile D F P and Gramotnev D K 2004 Opt. Lett. 29 1069

Pile D F P and Gramotnev D K 2004 Appl. Phys. Lett. 85 6323
[63] Bozhevilnyi S I, Volkov V S, Devaux E and Ebbesen T W 2005 Phys. Rev. Lett. 95 046802
[64] Bozhevilnyi S I, Volkov V S, Devaux E, J.Y. Laluet and Ebbesen T W 2006 Nature 440 508
[65] Krasavin A V and Zheludev N I 2004 Appl. Phys. Lett. 84 1416
[66] Krasavin A V, Zayats A V and Zheludev N I 2005 J. Opt. A: Pure Appl. Opt. 7 S85
[67] Andrew P and Barnes W L 2004 Science 306 1002
[68] Ferrell R A 1958 Phys. Rev. 111 1214
[69] Brown R W, Wessel P and Trounson E P 1960 Phys. Rev. Lett. 5 472
[70] Arakawa E T, Herickhoff R J and Birkhoff R D 1964 Phys. Rev. Lett. 12 319
[71] Donohue J F and Wang E Y 1986 J. Appl. Phys. 59 3137
Donohue J F and Wang E Y 1987 62 1313
[72] Teng Y-Y and Stern E A 1967 Phys. Rev. Lett. 19 511
[73] Otto A 1968 Z. Phys. 216 398
[74] Kretschmann E and Raether H 1968 Z. Naturf. A 23 2135
[75] Bruns R and Raether H 1970 Z. Phys. 237 98
[76] Kretschmann E 1971 Z. Phys. 241 313
[77] Mills D L and Burstein E 1974 Rep. Prog. Phys. 37 817
[78] Agranovich V M and Mills D L (ed) 1982 Surface Polaritons (Amsterdam: North-Holland)
[79] Boardman A D (ed) 1982 Electromagnetic Surface Modes (New York: Wiley)
[80] Sambles J R, Bradbery G W and Yang F Z 1991 Contemp. Phys. 32 173
[81] Zayats A V and Smolyaninov I I 2003 J. Opt. A: Pure Appl. Opt. 5 S16
[82] Ritchie R H PhD Thesis 1959 University of Tennessee (unpublished)
Ritchie R H 1963 Prog. Theor. Phys. 29 607
[83] Kanazawa H 1961 Prog. Theor. Phys. 26 851
[84] Bennett A J 1970 Phys. Rev. B 1 203
[85] Feibelman P J 1971 Phys. Rev. B 3 220
[86] Harris J and Griffin A 1971 Phys. Lett. A 34 51
[87] Flores F and Garca-Moliner F Solid State Commun. 11 1295
Garca-Moliner F and Flores F 1979 Introduction to the Theory of Solid Surfaces (Cambridge: Cambridge
University Press)
[88] Beck D E 1971 Phys. Rev. B 4 1555
[89] Beck D E and Celli V 1972 Phys. Rev. Lett. 28 1124
[90] Inglesfield J E and Wikborg E 1973 J. Phys. C: Solid State Phys. 6 L158
Inglesfield J E and Wikborg E 1974 Solid State Commun. 14 661
[91] Lang N D and Kohn W 1970 Phys. Rev. B 1 4555
Lang N D and Kohn W 1971 Phys. Rev. B 3 1215
[92] Feibelman P J 1973 Phys. Rev. Lett. 30 975
Feibelman P J 1974 Phys. Rev. B 9 5077
Feibelman P J 1975 Phys. Rev. B 12 1319
[93] Kunz C 1966 Z. Phys. 196 311
[94] Bagchi A, Duke C B, Feibelman P J and Porteus J O 1971 Phys. Rev. Lett. 27 998
[95] Duke C B, Pietronero L, Porteus J O and Wendelken J F 1975 Phys. Rev. B 12 4059
[96] Kloos T and Raether H 1973 Phys. Lett. A 44 157
[97] Krane K J and Raether H 1976 Phys. Rev. Lett. 37 1355
[98] Tsuei K-D, Plummer E W and Feibelman P J 1989 Phys. Rev. Lett. 63 2256
[99] Tsuei K-D, Plummer E W, Liebsch A, Kempa K and Bakshi P 1990 Phys. Rev. Lett. 64 44
[100] Tsuei K-D, Plummer E W, Liebsch A, Pehlke E, Kempa K and Bakshi P 1991 Surf. Sci. 247 302
[101] Sprunger P T, Watson G M and Plummer E W 1992 Surf. Sci. 269270 551
[102] Hohenberg P and Kohn W 1964 Phys. Rev. 136 B864
Kohn W and Sham L J 1965 Phys. Rev. 140 A1133
[103] Liebsch A 1987 Phys. Rev. B 36 7378
[104] Contini R and Layet J M 1987 Solid State Commun. 64 1179
[105] Suto S, Tsuei K-D, Plummer E W and Burstein E 1989 Phys. Rev. Lett. 63 2590
Lee G, Sprunger P T, Plummer E W and Suto S 1991 Phys. Rev. Lett. 67 3198
Lee G, Sprunger P T and Plummer E W 1993 Surf. Sci. 286 L547
[106] Rocca M and Valbusa U 1990 Phys. Rev. Lett. 64 2398
Rocca M, Lazzarino M and Valbusa U 1991 Phys. Rev. Lett. 67 3197
83
Rocca M, Lazzarino M and Valbusa U 1992 Phys. Rev. Lett. 69 2122

[107] Moresco F, Rocca M, Zielasek V, Hildebrandt T and Henzler M 1997 Surf. Sci. 388 1
Moresco F, Rocca M, Zielasek V, Hildebrandt T and Henzler M 1997 Surf. Sci. 388 24
[108] Kim B-O, Lee G, Plummer E W, Dowben P A and Liebsch A 1995 Phys. Rev. B 52 6057
[109] Tarriba J and Mochan W L 1992 Phys. Rev. B 46 12902
[110] Feibelman P J 1993 Surf. Sci. 282 129
Feibelman P J 1994 Phys. Rev. Lett. 72 788
[111] Liebsch A 1993 Phys. Rev. Lett. 71 145
Liebsch A 1993 Phys. Rev. B 48 11317
Liebsch A Phys. Rev. Lett. 72 789
[112] Liebsch A and Schaich W L 1995 Phys. Rev. B 52 14219
[113] Lopez-Bastidas C and Mochan W L 1999 Phys. Rev. B 60 8348
[114] Lopez-Bastidas C, Liebsch A and Mochan W L 2001 Phys. Rev. B 63 165407
[115] Campillo I, Rubio A and Pitarke J M 1999 Phys. Rev. B 59 12188
[116] Cazalilla M A, Dolado J S, Rubio A and Echenique P M 2000 Phys. Rev. B 61 8033
[117] Gurtubay I G, Pitarke J M, Campillo I and A. Rubio 2001 Comput. Mater. Sci. 22 123
[118] Silkin V M, Chulkov E V and Echenique P M 2004 Phys. Rev. Lett. 93 176801
[119] Silkin V M and Chulkov E V 2006 Vacuum 81 186
[120] Moresco F, Rocca M, Zielasek V, Hildebrandt T and Henzler M 1996 Phys. Rev. B 54 R14333
[121] Liebsch A 1998 Phys. Rev. B 57 3803
[122] Levinson H J, Plummer E W and Feibelman P 1979 Phys. Rev. Lett. 43 952
[123] Barman S R, Haberle P and Horn K 1998 Phys. Rev. B 58 R4285
[124] Barman S R, Biswas C and Horn K 2004 Phys. Rev. B 69 045413
[125] Silkin V M, Garca-Lekue A, Pitarke J M, Chulkov E V, Zaremba E and Echenique P M 2004 Europhys. Lett.
66 260
[126] Silkin V M, Pitarke J M, Chulkov E V and Echenique P M 2005 Phys. Rev. B 72 115435
[127] Pitarke J M, Nazarov V U, Silkin V M, Chulkov E V, Zaremba E and Echenique P M 2004 Phys. Rev. B
70 205403
[128] Diaconescu B, Pohl K, Vattuone L, Savio L, Hofmann P, Silkin V, Pitarke J M, Chulkov E, Echenique P M,
Farias D and Rocca M submitted
[129] See, e.g. Bohren C F and Huffman D R 1983 Absorption and Scattering of Light by Small Particles (New York:
Wiley)
[130] Batson P E 1982 Phys. Rev. Lett. 49 936
[131] Howie A and Milne R H 1985 Ultramicroscopy 18 427
[132] Ferrell T L and Echenique P M 1985 Phys. Rev. Lett. 55 1526
[133] Rivacoba A Zabala N and Echenique P M 1992 Phys. Rev. Lett. 69 3362
[134] McComb D W and Howie A 1995 Nucl. Instrum. Methods B 96 569
[135] Rivacoba A, Zabala N and Aizpurua J 2000 Prog. Surf. Sci. 65 1
[136] Klar T, Perner M, Grosse S, von Plessen G, W. Spirkl and Feldmann J 1998 Phys. Rev. Lett. 80 4249
[137] Shipway A N, Katz E and Willner I 2000 Chem. Phys. Chem. 1 18
[138] Lucas A A, Henrard L and Ph. Lambin 1994 Phys. Rev. B 49 2888
[139] de Heer W A, Bacsa W S, Chatelain A, Gerfin T, R. Humphrey-Baker, Forro L and Ugarte D 1995 Science
268 845
[140] Garca-Vidal F J, Pitarke J M and Pendry J B 1997 Phys. Rev. Lett. 78 4289
Pitarke J M and Garca-Vidal F J 2001 Phys. Rev. B 63 73404
Garca-Vidal F J and Pitarke J M 2001 Eur. Phys. J. B 22 257
[141] Garca-Vidal F J and Pendry J B 1996 Phys. Rev. Lett. 77 1163
Garca-Vidal F J and Pendry J B 1995 Prog. Surf. Sci. 50 55
[142] Garca-Vidal F J, Pitarke J M and Pendry J B 1998 Phys. Rev. B 58 6783
[143] Inglesfield J E, Pitarke J M and Kemp R 2004 Phys. Rev. B 69 233103
Pitarke J M and Inglesfield J E submitted
[144] Aruga T and Murata Y 1989 Prog. Surf. Sci. 31 61
[145] Liebsch A 1991 Phys. Rev. Lett. 67 2858
[146] Ishida H and Liebsch A 1992 Phys. Rev. B 45 6171
[147] Kim J-S, Chen L, Kesmodel L L and Liebsch A 1997 Phys. Rev. B 56 R4402
[148] Barman S R, Horn K, Haberle P, Ishida H and Liebsch A 1998 Phys. Rev. B 57 6662
84
J M Pitarke et al
[151] Liebsch A, Kim B-O and Plummer W 2001 Phys. Rev. B 63 125416
[152] Zielasek V, Ronitz N, Henzler M and Pfnur H 2006 Phys. Rev. Lett. 96 196801
[153] Dobson J F 1992 Phys. Rev. B 46 10163
Dobson J F 1993 Aust. J. Phys. 46 391
[154] Schaich W L and Dobson J F 1994 Phys. Rev. B 49 14700
[155] Zaremba E and Tso H C 1994 Phys. Rev. B 49 8147
[156] Guo Q, Feng Y P, Poon H C and Ong C K 1996 J. Phys.: Condens. Matter 8 9037
[157] Tokatly I V and Pankratov O 2002 Phys. Rev. B 66 153103
[158] See, e.g. Parker E H C (ed) 1985 The Technology and Physics of Molecular Beam Epitaxy (New York: Plenum)
[159] Dempsey J and Halperin B I 1992 Phys. Rev. B 45 1719
[160] Monarkha Y P, Teske E and Wyder P 2002 Phys. Rep. 370 1
[161] Ritchie R H 1973 Surf. Sci. 34 1
[162] Feibelman P J 1982 Prog. Surf. Sci. 12 287
[163] See, e.g. Liebsch A 1997 Electronic Excitations at Metal Surfaces (New York: Plenum)
[164] Raether H 1980 Excitation of Plasmons and Interband Transitions by Electrons (Springer Tracks in Modern
Physics vol 88) (New York: Springer)
[165] Raether H 1988 Surface Plasmons on Smooth and Rough Surfaces and on Gratings (Springer Tracks in Modern
Physics) vol 111 (New York: Springer)
[166] Plummer W, Tsuei K-D and Kim B-O 1995 Nucl. Instrum. Methods B 96 448
[167] Rocca M 1995 Surf. Sci. Rep. 22 1
[168] Kushwaha M S 2001 Surf. Sci. Rep. 41 1
[169] Jackson J D 1962 Classical Electrodynamics (New York: Wiley)
[170] Ritchie R H and Eldridge H B 1962 Phys. Rev. 126 1935
[171] Ashcroft N W and Mermin N D 1976 Solid State Physics (Philadelphia: Saunders)
[172] Pettit R B, Silcox J and Vincent R 1975 Phys. Rev. B 11 3116
[173] Stern F 1967 Phys. Rev. Lett. 18 546
Ando T, Fowler A B and Stern F 1982 Rev. Mod. Phys. 54 437
[174] Allen S J Jr, Tsui D C and Logan R A 1977 Phys. Rev. Lett. 38 980
[175] Nagao T, Heldebrandt T, Henzler M and Hasegawa S 2001 Phys. Rev. Lett. 86 5747
[176] See, e.g. Lundqvist S 1983 in Theory of the Inhomogeneous Electron Gas ed S Lundqvist and N H March (New
York: Plenum) p 149
[177] Ritchie R H and Wilems R E 1969 Phys. Rev. 178 372
[178] Barton G 1979 Rep. Prog. Phys. 42 65
[179] Wagner D 1966 Z. Naturf. a 21 634
[180] Ritchie R H and Marusak A L 1966 Surf. Sci. 4 234
[181] Schmeits M 1989 Phys. Rev. B 39 7567
[182] Pitarke J M, Pendry J B and Echenique P M 1997 Phys. Rev. B 55 9550
[183] Pitarke J M and Rivacoba A 1997 Surf. Sci. 377379 294
[184] See, e.g. Bergman D J and Stroud D 1992 Solid State Phys. 46 147
[185] Fuchs R 1974 Phys. Lett. A 48 353
[186] Fuchs R 1975 Phys. Rev. B 11 1732
[187] Ouyang F and Isaacson M 1989 Phil. Mag. B 60 481
[188] Ouyang F and Isaacson M 1989 Ultramicroscopy 31 345
[189] Lu J Q and Maradudin A A 1990 Phys. Rev. B 42 11159
[190] Goloskie R, Thio T and Ram-Mohan L R 1996 Comput. Phys. 10 477
[191] Garca de Abajo F J and Aizpurua J 1997 Phys. Rev. B 56 15873
[192] Aizpurua J, Howie A and Garca de Abajo F J 1999 Phys. Rev. B 60 11149
[193] Garca de Abajo F J and Howie A 1998 Phys. Rev. Lett. 80 5180
[194] Bergman D J 1978 Phys. Rep. 43 377
[195] Milton G 1981 J. Appl. Phys. 52 5286
[196] Yablonovitch E, Gmitter T J, Meade R D, Rappe A M, Brommer K D and Joannopoulos J D 1991 Phys. Rev.
Lett. 67 3380
[197] Ho K M, Chan C T and Soukoulis C M 1990 Phys. Rev. Lett. 65 3152
[198] Pendry J B and MacKinnon A 1992 Phys. Rev. Lett. 69 2772
[199] Inglesfield J E 1998 J. Phys. A: Math. Gen. 31 8495
Kemp R and Inglesfield J E 2002 Phys. Rev. B 65 115103
[200] Garca de Abajo F J 1999 Phys. Rev. B 59 3095
[201] English R A, Pitarke J M and Pendry J B 2000 Surf. Sci. 454456 1090

[202]
[203]
[204]
[205]
[206]
[207]
[208]
[209]
[210]
[211]
[212]
[213]
[214]
[215]
[216]
[217]
[218]
[219]
[220]
[221]
[222]
[223]
[224]
[225]
[226]
[227]
[228]
[229]
[230]
[231]
[232]
[233]
[234]
[235]
[236]
[237]
[238]
[239]
[240]
[241]
[242]
[243]
[244]
[245]
[246]
[247]
[248]
[249]
[250]
[251]
85
Pitarke J M, Garca-Vidal F J and Pendry J B 1998 Phys. Rev. B 57 15261

Pitarke J M, Garca-Vidal F J and Pendry J B 1999 Surf. Sci. 433435 605
Yannopapas V, Modinos A and Stefanou N 1999 Phys. Rev. B 60 5359
Riikonen S, Romero I and Garca de Abajo F J 2005 Phys. Rev. B 71 235104
Apell S P, Echenique P M and Ritchie R H 1995 Ultramicroscopy 65 53
Ronveaux A, Moussiaux A and Lucas A A 1977 Can. J. Phys. 55 1407
Ostling
D, Apell P and Rosen A 1993 Europhys. Lett. 21 539
Pines D and Nozieres 1989 The Theory of Quantum Liquids (New York: Addison-Wesley)
Fetter A L and Walecka J D 1971 Quantum Theory of Many-Particle Systems (New York: McGraw-Hill)
Callen H B and Welton T A 1951 Phys. Rev. 83 34
See, e.g. Gross E K U and Kohn W 1990 Adv. Quantum Chem. 21 255
Zangwill A and Soven P 1980 Phys. Rev. A 21 1561
Ullrich C A, Gossmann U J and Gross E K U 1995 Phys. Rev. Lett. 74 872
Petersilka M, Gossmann U J and Gross E K U 1996 Phys. Rev. Lett. 76 1212
Burke K, Petersilka M and Gross E K U 2000 A hybrid functional for the exchange-correlation kernel in timedependent density functional theory Recent Advances in Density Functional Methods vol III, ed P Fantucci
and A Bencini (Singapore: World Scientific)
Pitarke J M and Perdew J P 2003 Phys. Rev. B 67 045101
Jung J, Garca-Gonzalez P, Dobson J F and R. Godby W 2004 Phys. Rev. B 70 205107
Hubbard J 1958 Proc. R. Soc. Lond. Ser. A 243 336
Singwi K S, Tosi M P, Land R H and Sjolander A 1968 Phys. Rev. 176 589
Singwi K S, Sjolander A, Tosi M P and Land R H 1970 Phys. Rev. B 1 1044
Vashishta P and Singwi K S 1972 Phys. Rev. B 6 875
Singwi K S and Tosi M P 1981 Solid State Phys. 36 177
Ichimaru S 1982 Rev. Mod. Phys. 54 1017
Iwamoto N and Pines D 1984 Phys. Rev. B 29 3924
Engel E and Vosko S H 1990 Phys. Rev. B 42 4940
Khodel V A, Shaginyan V R and Khodel V V 1994 Phys. Rep. 249 1
Bowen C, Sugiyama G and Alder B J 1994 Phys. Rev. B 50 14838
Moroni S, Ceperley D M and Senatore G 1995 Phys. Rev. Lett. 75 689
Corradini M, Del Sole R 1987 Onida G and Palummo M 1998 Phys. Rev. B 57 14569
Gross E K U and Kohn W 1985 Phys. Rev. Lett. 55 2850
Iwamoto N and U E K Gross. Phys. Rev. B 35 3003
Vignale G and Kohn W 1996 Phys. Rev. Lett. 77 2037
Vignale G, Ullrich C A and Conti S 1997 Phys. Rev. Lett. 79 4878
Nifosi R, Conti S and Tosi M P 1998 Phys. Rev. B 58 12758
Qian Z and Vignale G 2002 Phys. Rev. B 65 235121
Brosens F, Lemmens L F and Devreese J T 1976 Phys. Status Solidi B 74 45
Devreese J T, Brosens F and Lemmens L F 1980 Phys. Rev. B 21 1349
Brosens F, Devreese J T and Lemmens L F 1980 Phys. Rev. B 21 1363
Richardson C F and Ashcroft N W 1994 Phys. Rev. B 50 8170
Sturm K and Gusarov A 2000 Phys. Rev. B 62 16474
Echenique P M, Bausells J and Rivacoba A 1987 Phys. Rev. B 35 1521
Abramowitz M and Stegun I A 1964 Handbook of Mathematical Functions (New York: Dover)
Barrera R G and Fuchs R 1995 Phys. Rev. B 52 3256
Bergara A, Pitarke J M and Ritchie R H 1999 Phys. Lett. A 256 405
Lindhard J 1954 K. Dan. Vidensk. Selsk. Mat.-Fys. Medd. 28 1
Garca-Lekue A and Pitarke J M 2001 Phys. Rev. B 64 035423
Garca-Lekue A and Pitarke J M 2003 Phys. Rev. B 67 089902(E)
Garca-Lekue A and Pitarke J M Nucl. Instrum. Methods B 182 56
See, e.g. Lang N 1973 Solid State Phys. 28 225
Reuter G E H and Sonheimer E H 1948 Proc. R. Soc. Lond. Ser. A 195 336
Griffin A and Harris J 1976 Can. J. Phys. 54 1396
Chulkov E V, Silkin V M and Echenique P M 1997 Surf. Sci. 391 L1217
Chulkov E V, Silkin V M and Echenique P M 1999 Surf. Sci. 437 330
Chulkov E V, Sarria I, Silkin V M, Pitarke J M and Echenique P M 1998 Phys. Rev. Lett. 80 4947
Echenique P M, Pitarke J M, Chulkov E V and Rubio A 2000 Chem. Phys. 251 1
Kliewer J, Berndt R, Chulkov E V, Silkin V M, Echenique P M and Crampin S 2000 Science 288 1399
86
J M Pitarke et al
[252] Echenique P M, Osma J, Silkin V M, Chulkov E V and Pitarke J M 2000 Appl. Phys. A 71 503
Echenique P M, Osma J, Machado M, Silkin V M, Chulkov E V and Pitarke J M 2001 Prog. Surf. Sci. 67 271
[253] Eiguren A, Hellsing B, Reinert F, Nicolay G, Chulkov E V, Silkin V M, Hufner S and Echenique P M 2002
Phys. Rev. Lett. 88 066805
[254] Echenique P M, Berndt R, Chulkov E V, Th. Fauster, A. Goldmann and Hofer U 2004 Surf. Sci. Rep. 52 219
[255] Crampin S 2005 Phys. Rev. Lett. 95 046801
[256] Vergniory M G, Pitarke J M and Crampin S 2005 Phys. Rev. B 72 193401
[257] Persson B N J and Zaremba E 1984 Phys. Rev. B 30 5669
[258] Ibach H and Mills D L 1982 Electron Energy Loss Spectroscopy and Surface Vuvrations (New York: Academic)
[259] Mills D L 1975 Surf. Sci. 48 59
[260] Gaspar J A, Eguiluz A G, Tsuei K-D and Plummer E W 1991 Phys. Rev. Lett. 67 2854
[261] Forstmann F and Stenschke H 1978 Phys. Rev. B 17 1489
[262] Feibelman P J 1989 Phys. Rev. B 40 2752
Feibelman P J and Tsuei K D 1990 Phys. Rev. B 41 8519
[263] Kempa K and Schaich W L 1988 Phys. Rev. B 37 6711
[264] Chiarello G, Formoso V, Santaniello A, Colavita E and Papagno L 2000 Phys. Rev. B 62 12676
[265] Eguiluz A G 1983 Phys. Rev. Lett. 51 1907
Eguiluz A G 1985 Phys. Rev. B 31 3303
[266] Marusic L, Despoja V and Sunji

c M 2006 J. Phys.: Condens. Matter 18 4253
[267] Nazarov V U, Nishigaki S and Nagao T 2002 Phys. Rev. B 66 092301
[268] Beck D E and Dasgupta B B 1975 Phys. Rev. B 12 1995
[269] Eguiluz A, Ying S C and Quinn J J 1975 Phys. Rev. B 11 2118
[270] Eguiluz A G and Quinn J J 1976 Phys. Rev. B 14 1347
[271] Schwartz C and Schaich W L 1982 Phys. Rev. B 26 7008
[272] Ahlqvist P and Apell P 1982 Phys. Scr. 25 587
[273] Schwartz C and Schaich W L 1984 Phys. Rev. B 30 1059
[274] Dobson J F and Harris G H 1988 J. Phys. C: Solid State Phys. 21 L729
Dobson J J, Harris G H and OConnor A J 1990 J. Phys.: Condens. Matter 2 6461
[275] Nazarov V U and Nishigaki S 2001 Surf. Sci. 482 640
[276] Sellares J and Barberan N 1994 Phys. Rev. B 50 1879
[277] Monin J and Boutry G-A 1974 Phys. Rev. B 9 1309
[278] Perdew J P, Tran H Q and Smith E D 1990 Phys. Rev. B 42 11627
[279] Shore H B and Rose J H 1991 Phys. Rev. Lett. 66 2519
Rose J H and Shore H B 1991 Phys. Rev. B 43 11605
Shore H B and Rose J H 1999 Phys. Rev. B 59 10485
[281] Rocca M, Lizzit S, Brena B, Cautero G, Comelli G and Paolucci G 1995 J. Phys.: Condens. Matter 7 L611
[282] Johnson P B and Christy R W 1972 Phys. Rev. B 6 4370
[283] Lopez-Bastidas C, Maytorena J A and Liebsch A 2002 Phys. Rev. B 65 035417
[284] Moresco F, Rocca M, Hildebrandt T, Zielasek V and Henzler M 1998 Europhys. Lett. 43 433
[285] Savio L, Vattuone L and Rocca M 2000 Phys. Rev. B 61 7324
[286] Savio L, Vattuone L and Rocca M 2003 Phys. Rev. B 67 045406
[287] Vehse R C, Arakawa E T and Williams M W 1970 Phys. Rev. B 1 517
[288] Inglesfield J E 1982 Rep. Prog. Phys. 45 223
[289] Chulkov E V, Silkin V M and Shirykalov E N 1987 Surf. Sci. 188 287
[290] Chaplik A V 1972 Zh. Eksp. Teor. Fiz. 62 746 (Sov. Phys.JETP 35 395)
[291] Das Sarma S and Madhukar A 1981 Phys. Rev. B 23 805
[292] Silkin V M, Balasubramanian T, Chulkov E V, Rubio A and Echenique P M 2001 Phys. Rev. B 64 085334
[293] Pitarke J M, Silkin V M, Chulkov E V and Echenique P M 2005 J. Opt. A.: Pure Appl. Opt. 7 S73
[294] Pitarke J M and Campillo I 2000 Nucl. Instrum. Methods B 164 147
[295] Hedin L and Lundqvist S 1969 Solid State Phys. 23 1
[296] Milne R H and Echenique P M 1985 Solid State Commun. 55 909
[297] Nunez R, Echenique P M and Ritchie R H 1980 J. Phys. C: Solid State Phys. 13 4229
[298] Yubero F and Tougaard S 1992 Phys. Rev. B 46 2486
[299] Chen Y F and Chen Y T 1996 Phys. Rev. B 53 4980
[300] Juaristi J I, Garca de Abajo F J and Echenique P M 1996 Phys. Rev. B 53 13839
[301] Ding Z J 1997 Phys. Rev. B 55 9999
Ding Z J 1998 J. Phys.: Condens. Matter 10 1733
[302]
[303]
[304]
[305]
[306]
[307]
[308]
[309]
[310]
[311]
[312]
[313]
[314]
[315]
[316]
[317]
[318]
[319]
[320]
[321]
[322]
[323]
[324]
[325]
[326]
[327]
[328]
[329]
[330]
[331]
[332]
[333]
[334]
[335]
[336]
[337]
87
Ding Z J 1998 J. Phys.: Condens. Matter 10 1753

Nagatomi T, Shimizu R and Ritchie R H 1999 Surf. Sci. 419 158
Song Y H, Wang Y N and Miskovic Z L 2001 Phys. Rev. A 63 052902
Tokesi K, Tong X-M, Lemell C and Burgdorfer J 2005 Phys. Rev. A 72 022901
Alducin M, Silkin V M, Juaristi J I and Chulkov E V 2003 Phys. Rev. A 67 032903
Campillo I, Pitarke J M and Eguiluz A G 1998 Phys. Rev. B 58 10307
Campillo I, Pitarke J M, Eguiluz A G and Garc a A 1998 Nucl. Instrum. Methods B 135 103
Campillo I and Pitarke J M 2000 Nucl. Instrum. Methods B 164 161
Mathar R J, Sabin J R and Trickey S B 1999 Nucl. Instrum. Methods B 155 249
Garca-Vergniory M, Gurtubay I G, Silkin V M and Pitarke J M in preparation
Gemmell D S 1974 Rev. Mod. Phys. 46 129
Kimura K, Hasegawa M and Mannami M 1987 Phys. Rev. B 36 7
Fuji Y, Fujiwara S, Narumi K, Kimura K and Mannami M 1992 Surf. Sci. 277 164
Kimura K, Kuroda H, Fritz M and Mannami M 1992 Surf. Sci. 277 164
Kimura K, Kuroda H, Fritz M and Mannami M 1995 Nucl. Instrum. Methods B 100 356
Winter H, Wilke M and Bergomaz M 1997 Nucl. Instrum. Methods B 125 124
Robin A, Hatke N, Narmann A, Grether M, Plackhe D, Jensen J and Heiland W 2000 Nucl. Instrum. Methods
B 164 566
Winter H 2002 Phys. Rep. 367 387
Cazalilla M A and Juaristi J I 1999 Nucl. Instrum. Methods B 157 104
Garc a-Lekue A and Pitarke J M 2003 J. Electron. Spectrosc. 129 223
Berger S D, Bruley J, Brown L M and McKenzie D R 1989 Ultramicroscopy 28 43
Lecante J, Ballu Y and Newns D M 1977 Phys. Rev. Lett. 38 36
Warmack R J, Becker R S, Anderson V E, Ritchie R H, Chu Y T, Little J and Ferrell T 1984 Phys. Rev. B
29 4375
Ritchie R H and Howie A 1988 Phil. Mag. A 58 753
Kohl H and Rose H 1985 Adv. Electron. Electron Phys. 65 173
Cohen H, Maniv T, Tenne R, Rosenfeld Hacohen Y, Stephan O and Colliex C 1998 Phys. Rev. Lett. 80 782
Echenique P M, Howie A and Ritchie R H 1999 Phys. Rev. Lett. 83 658
Cohen H, Maniv T, Tenne R, Rosenfeld Hacohen Y, Stephan O and Colliex C 1999 Phys. Rev. Lett. 83 659
Rivacoba A and Echenique P M 1990 Scanning Microsc. 4 73
Okamoto K, Niki I, Scherer A, Narukawa Y, Mukai T and Kawakami Y 2005 Appl. Phys. Lett. 87 071102
Wedge S, Wasey J A E, Barnes W L and Sage I 2004 Appl. Phys. Lett. 85 182
Thio T, Pellerin K M, Linke R A, Lezec H J and Ebbesen T W 1972 Opt. Lett. 26 1972
Pendry J B, Martin-Moreno L and Garcia F J-Vidal 2004 Science 305 847
Barnes W L, Murray W A, Dintinger J, Devaux E and T. W. Ebbesen 2004 Phys. Rev. Lett. 92 107401
Beruete M, Sorolla M, Campillo I, Dolado J S, Martin-Moreno L, Bravo-Abad J and Garca-Vidal F J 2004
Opt. Lett. 29 2500
Akarca-Biyikli S S, Bulu I and Ozbay E 2004 Appl. Phys. Lett. 85 1098
Cao H and Nahata A 2004 Opt. Express 12 3664
Kik P G, Maier S A and Atwater H A 2004 Phys. Rev. B 69 045418
Luoa X and Ishihara T 2004 Appl. Phys. Lett. 84 4780
Shao D B and Chen S C 2005 Appl. Phys. Lett. 86 253107
Wang L, Uppuluri S M, Jin E X and Xu X F 2006 Nano Lett. 6 361
Pendry J B 2000 Phys. Rev. Lett. 85 3966
Taubner T, Korobkin D, Urzhumov Y, Shvets G and Hillenbrand R 2006 Science 313 1595

RepProgPh70 - Pitarke Theory SPSPP

Uploaded by

Copyright:

Available Formats

RepProgPh70 - Pitarke Theory SPSPP

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

RepProgPh70 - Pitarke Theory SPSPP

Uploaded by

Copyright:

Available Formats

INSTITUTE OF PHYSICS PUBLISHING

REPORTS ON PROGRESS IN PHYSICS

Rep. Prog. Phys. 70 (2007) 187

Theory of surface plasmons and surface-plasmon

Materia Kondentsatuaren Fisika Saila, Zientzi Fakultatea, Euskal Herriko Unibertsitatea,

This article was invited by Professor M Finnis.

2007 IOP Publishing Ltd

Theory of surface plasmons and surface-plasmon polaritons

8.3.4. Surface-plasmon linewidth.

Theory of surface plasmons and surface-plasmon polaritons

Theory of surface plasmons and surface-plasmon polaritons

An alternative spectroscopy technique to investigate multipole surface plasmons is

Theory of surface plasmons and surface-plasmon polaritons

c/ 2 at the point q = 0 and is a monotonic increasing function of q, which is always smaller

asymptotically the nonretarded surface-plasmon frequency s = p / 2 (- - - -).

2.1.2. Energy dispersion.

In the case of a Drude semi-infinite metal in vacuum, one has

being a positive infinitesimal. Hence, in this case equation (2.13) yields

Theory of surface plasmons and surface-plasmon polaritons

Figure 3. Schematic representation of the electromagnetic field associated with a surface-plasmon

2.2. Thin films

oscillations at the reduced frequency s = p / 2 characteristic of a semi-infinite electron gas

Theory of surface plasmons and surface-plasmon polaritons

Figure 5. Dispersion (q) of the surface-plasmon polariton of an Al film of thickness a = 120

the quasi-static Maxwells equations

Theory of surface plasmons and surface-plasmon polaritons

dz Dz (z; q, ) = 2 (z2 ; q, ) 1 (z1 ; q, ).

where s is Ritchies frequency: s = p / 2. For neutral jellium surfaces, d () coincides

so that equations (3.16)(3.18) yield the linearized hydrodynamic equations

and Laplaces equation at z > 0:

where q represents a 2D wave vector in the plane of the surface.

Theory of surface plasmons and surface-plasmon polaritons

which for Drude cylinders (1

Theory of surface plasmons and surface-plasmon polaritons

3.2.3. Composite systems: effective-medium approach. Composite systems with a large

where u is the spectral variable

where E is the macroscopic electric field averaged over the composite:

equation (2.15)) in vacuum (2 = 1) yields the reduced plasmon frequency = f p .

Theory of surface plasmons and surface-plasmon polaritons

where D represents the dimensionality of the inclusions.

Theory of surface plasmons and surface-plasmon polaritons

n0 (r) =< 0 |(r)|

where (r, r ; ) represents the so-called density-response function of the many-electron

Theory of surface plasmons and surface-plasmon polaritons

5.2. Time-dependent density-functional theory

vKS [n0 ](r) = vH [n0 ](r) + vxc [n0 ](r)

with vxc [n](r, t) being the exact time-dependent XC potential of TDDFT.

and fxchom (n;

Theory of surface plasmons and surface-plasmon polaritons

In the ALDA, one writes

which in combination with equations (5.20)(5.22) yields the ALDA XC kernel of

= Q/qF , and the parameters B, C, g, and are the dimensionless functions of n

which in combination with equation (6.2) yields the classical formula

() representing the so-called local dielectric function of the medium.

Hence, a comparison of equations (6.2) and (7.1) yields

and using equation (6.3):

Theory of surface plasmons and surface-plasmon polaritons

i () being 1 or 2 depending on whether the point r is located in medium 1 or in medium 2,

where u is the spectral variable of equation (3.44) and n = 1/2.

where s is Ritchies frequency: s = p / 2. For neutral jellium surfaces, d () coincides

n0 (r) =< 0 |(r)|

(corresponding to the planar surface-plasmon condition 1 + 2 = 0) for all m = 0; in the limit

k,n;i (r) = k,n (r ) i (z),

where k,n (r ) are Bloch states: