NANOMAGMA Focused Report

Page 1

E EKD 'D &K h^ Z WKZd E EKƐƚƌƵĐƚƵƌĞĚ ĂĐƚŝǀĞ D 'ŶĞƚŽͲƉůĂƐŵŽŶŝĐ D ƚĞƌŝĂůƐ

NANOMAGMA FOCUSED REPORT NANOstructured active MAGnetoplasmonic MAterials

Funded by

Coordinated by

Edited by

Introduction

page 1

Plasmonics

page 5

What is plasmonics? Localized surface plasmons (LSPs) Surface plasmon polaritons (SPPs) Plasmonic applications Active plasmonics

7 9 10 11 13

Magneto-optics

page 15

Fundamentals

17

1.

17 17 18 18 18 19

2. 3.

Magneto-optical effects 1.1 Transverse configuration 1.2 Polar configuration 1.3 Longitudinal configuration Materials that exhibit MO activity Types of MO characterization

Telecom applications

19

1. 2.

19 20 20 21 24 22 23

Magneto-optical storage: basic description Integrated photonics: optical isolators 2.1 Commercial optical isolators: bulk components 2.2 Integrated optical isolators 2.3 Integrated optical isolators based on ferromagnetic materials 2.4 Nanostructuration of the optical domains 2.5 Nanostructuration of the magnetic domains

Magneto-plasmonics

page 25

Modeling

27

1.

27 27 29 31 32

2. 3.

Basic concepts. Building blocks of MO devices 1.1 MO multilayers 1.2 MO nanoparticles Magneto-Plasmonic coupling Magneto-Plasmonic active control of single-molecule fluorescence 3.1 Generalized framework for modeling FRET in the presence of a nanoparticle 3.2 Magneto-optical control of Fรถrster energy transfer 3.3 Periodic MO structures

32 33 34

Fabrication of magneto-plasmonic nanostructures

41

1.

41 42 43

Top-Down 1.1 Colloidal lithography 1.2 Atomic layer deposition

Index

1. 2. 3. 4. 5.

2.

46 46 46 47 49 49

Magnetoplasmonic properties

51

1. 2.

51 53

Continuous systems Nanostructured systems

Telecom applications

57

1.

58 59 59 59 59 63 63 64

2.

Index

Bottom-up 2.1 Ferrite based magnetic nanoparticles 2.2 Gold based plasmonic nanoparticles 2.3 Hybrid magnetic-plasmonic materials 2.4 Optically addressable thin films 2.5 Monolayers of magnetic and magnetoplasmonic nanoparticles on silicon surfaces

Investigation of non-reciprocal magneto-plasmonic waveguides 1.1 Simulation tool 1.2 Figures of merit as guidelines for design 1.3 Material parameters 1.4 Simulation results Fabrication 2.1 Proposed process flow 2.2 Challenges

Sensing

64

Sensing and Biosensing Application

66

1.

Experimental 1.1 MOSPR sensors: main performance characteristics 1.2 SPR and MOSPR coupled instrumentation based on prism coupler 1.3 Optoelectronic system for SPR activation 1.4 Detection and data processing system 1.5 Fluid control system 1.6 Experimental setup for magneto-plasmonic characterization

66 66 66 67 67 68 68

2.

Application 2.1 Bio-sensing applications 2.2 Gas- sensing applications 2.3 SPR coupling to other sensing techniques 2.4 Surface Functionalization 2.5 Flow Injection 2.6 Modeling 2.7 Data analysis 2.8 Combined EIS, SPR and MOSPR analysis

69 69 73 75 76 76 77 77 78

References

page 79

Annex

page 87

Introduction

Therefore the main idea behind NANOMAGMA is to get insight into the interplay between plasmonics and magneto-optics. The project has two main goals; the first is to prepare active magneto-plasmonic materials with tailored properties in the nanoscale and understanding the interactions of the magnetic properties with the plasmonic and optical ones, linked to electric charge oscillations. The second goal is to propose devices for applications that can benefit of this coupling. Since it is expected that the optical properties of these materials can be driven by using a magnetic field, this will allow designing and developing novel magnetoplasmonic devices. These devices will be of use in both areas mentioned above: sensing, i.e. a surface magnetoplasmon resonance (SMPR) sensor tailored on the nanoscale, and information technologies, i.e. nonreciprocal components for photonic integrated circuits based on magnetoplasmonic elements.

Introduction During the last decades a large effort has been invested in the development of a new discipline devoted to benefit from optical excitations in materials where metals are key element (Plasmonics). We will make an introduction on this topic below, but let’s anticipate that two application areas are sensing and information technologies. In the first case, it is the strong dependence of the plasmon resonance on the environment the factor that is used for the development of applications. In the second, it is the capability to confine the electromagnetic field beyond the diffraction limit when coupling to the plasmon what is put to work. In both cases there is way for improvement, and we have identified an element that can be used in both areas, and in turn make an interesting influence in another field of research: magnetooptics. Magneto-optics is a discipline that has been tied to the information technologies framework from long ago, mainly to endorse active (tunable) capabilities. There will be an introductory section to the topic below.

The following reflects, after some introductory summary of relevant concepts in plasmonics and magnetooptics, the latest developments on magneto-plasmonics within the framework of the NANOMAGMA consortium.

3

Germany

France

Page

Italy

Romania

Spain

NANOMAGMA Consortium

Plasmonics

What is plasmonics?

In the last decade, and driven by the strong development of nanotechnology, there has been an increased interest in the study of the optical properties of metallic nanostructures and nanoparticles and their ability to control and manipulate light. This constitutes a field denoted as plasmonics, which takes the name due to the fact that the main responsible of the singular properties of the metallic nanostructures is the excitation of surface plasmon resonances. These plasmon resonances were analysed as a fascinating surfacewave effect in the first and middle parts of the last century. Nowadays they have recuperated their attraction as they offer a way to beat the light diffraction limit and enable a path towards subwavelength optics [1]. Plasmons are, generically, resonant oscillations of the free charges in plasmas. From an electromagnetic perspective, metals are described as plasmas comprised of fixed positive ions and free mobile conduction electrons. This behaviour results in the negative relative permittivities (or dielectric constant, Îľ) that characterize metals in their interaction with light (up to the plasma frequency, located in the UV for most metals) [2,3]. Surface plasmons (SPs) are electromagnetic waves coupled to the collective oscillations of the surface free charges in an interface between two media with permittivities with opposite sign, typically a dielectric and a metal [1,3]. Surface plasmon modes can exist on a wide variety of metallic structures, such as single surfaces, thin films, nanoparticles, cylinders, etc. They can be classified in two main modes: localized surface plasmons (LSPs), also called particle

Surface plasmons are able to confine the electromagnetic field (i.e. the light, when we are dealing with optical frequencies) in small, nanoscopic volumes, beyond the diffraction limit, which makes them suited for the development of nanophotonic devices [1,4,5] This strong confinement can result in huge local field enhancements which increase enormously the interaction of light with molecules or emitters located there, another interesting property of SPs to manipulate the light at the nanoscale [6]. SPs are also highly sensitive to the optical properties (refractive index) of the dielectric media surrounding the metal, being therefore sensing one of the main and better established applications [7,8]. Moreover, when going beyond the classical description to enter into quantum mechanics, we can also describe plasmons as the quanta of the plasma free charges oscillations [2], in such a way that surface plasmons are the quanta of electron/photon polariton modes [9]. Therefore, SPs can experience quantum phenomena such as entanglement [10] or couple to quantum emitters [11,12], allowing also their use in quantum applications. It should be mentioned here that within this description SPs are bosons, even if charge is involved. These properties have given rise to the extraordinary appeal of surface plasmons and the development of plasmonics. Plasmonics is therefore, a subfield of nanooptics, and more generally of nanoscience and nanotechnology, that aims to understand and control light using the surface plasmon resonances sustained by metal nanostructures. It is related to the localization, guiding, and manipulation of electromagnetic waves

7

1.

Page

Plasmonics

plasmons, which are the ones sustained by nanoparticles, and surface plasmon polaritons (SPPs) or propagating surface plasmons, those supported by planar interfaces. Both kinds of SPs will be described below in more detail.

Page

8

History of plasmonics. Although the understanding of surface plasmons has not happened until the 20th century, the associated optical properties had already become familiar to the glass and ceramic makers since much earlier. Some example of vessels made of dichroic glass dating from the late Roman period testify the knowledge of red-colouring glass by means of gold at those ancient times, being the Lycurgus cup one of the most outstanding examples. This cup shows a green colour under ambient, diffuse illumination and turns to a bright red when light is shone through it (see Figure 1a). This optical effect is due to the excitation of SP resonances in goldsilver nanocolloids present in the glass [13]. However, the technique of making the red glass was later lost and, although it could have been used in some stained glasses during the Middle Ages, it was not until the 17th century where we find written references to it, under the name of “Purple of Cassius” or ruby gold glass [14] (see in Figure 1b a cup dating from that period). Although by then the empirical know-how of the red colour being associated with gold was well established, Faraday was the first to postulate that the colour was due to the presence of gold particles of small size compared to that of light [15]. Last century brought the observations and discoveries that finally allowed

identifying surface plasmons [16]: from the observation of Wood anomalies in metallic gratings diffraction by R. W. Wood in 1902 and G. Mie development of the (Mie) theory on light scattering by spherical particles in 1908 up to R. Ritchie prediction of surface plasmon polaritons in metal foils based on EELS experiments in 1957, the demonstration of several methods for optical excitation of SPPs by A. Otto, E. Kretschmann and H. Raether in 1968, and U. Kreibig and P. Zacharias description of the optical properties of metal nanoparticles in terms of SPs for the first time by in 1970, a long path has been travelled were all these phenomena had to be connected and recognized as a single physical fact, the excitation of surface plasmon resonances. An important push for the progress of plasmonics was the detection by the first time of surfaceenhanced Raman scattering (SERS) by M. Fleischmann et al. in 1974. Finally, the development of nanotechnology in the ‘90s made it possible an accurate control on the obtainment of metal (a)

(b)

Late Roman, 4th century AD

(c)

Germany, ~1690

4000 3500 Number of Articles

beyond the diffraction limit and down to the nanometer-length scale by using these metal nanostructures in novel ways.

3000

N. of Articles

2500 2000 1500 1000 500 0

Publication Year

Figure 1: (a) The Lycurgus cup, late Roman, 4th century AD (© Trustees of the British Museum). (b) Gold Ruby Goblet, Potsdam, Brandenburg (Germany), probably engraved by Gottfried Spiller, about 1690-1700 (79.3.258, Corning Museum of Glass, gift of The Ruth Bryan Strauss Memorial Foundation). (c) Evolution of the number of scientific articles related to plasmonics in the last 45 years (data obtained from Thomson Reuters Web of Knowledge).

Localized surface plasmons (LSPs)

The excitation of localized surface plasmons in metal nanoparticles makes them acquire different colours depending on their size, shape, constituting material and surroundings. When light illuminates a metal particle, the associated oscillating electric field (E0) exerts a force on the mobile electrons of the conduction band and displaces them, generating charges at opposite surfaces of the particle and inducing a dipole moment (see Figure 2a). The attraction of the charges also produces a restoring force on the displaced electrons. As a result, we have an electron oscillator characterized by a resonance frequency which depends on the restoring force and the mass of the electron [3,6,17]. This resonating electron oscillator is the localized surface plasmon. Like for any oscillator, firstly we are interested in the position of the LSP resonance. For spherical nanoparticles smaller than the metal skin depth (which is around 25 nm for the typical plasmonic metals such as gold and silver at visible frequencies), the light electric field E0 is constant across all the particle and the electrostatic approximation can be used to obtain the induced dipole (a) E0 Metal core

E0

ecloud

Resonance @ Îľm = -2Îľd (b)

+∆t Restauring force

For very small particles, a << Îť : Îľ (ω ) − Îľ d p = Îą E; with Îą (ω ) = 4Ď€ a3 m Îľ m (ω ) + 2Îľ d Figure 2: (a) Schematic representation of the action of the light electric field on a metal nanoparticle. (b) Solution of nanocolloidal gold, with its characteristic red colour resulting from the excitation of the localized surface plasmon.

( = 4Ď€a

( − , (1 ( + 2

where ξm(ω) is the frequency dependent metal permittivity; ξd the dielectric constant of the surrounding dielectric; and a the particle radius. From that expression, it can be seen that a maximum in the polarizability takes place for ξm =-2ξd (the localized surface plasmon). In this situation, the particle gives rise to a strong dipole that generates an enhanced electromagnetic field at each side, and the interaction with the light is maximized with high scattering and absorption. As ξm is strongly dependent on the frequency (or equivalently the wavelength), this is the origin of the colour in the nanoparticles [3,19]. For instance, for Au nanoparticles in solution the condition is satisfied for Ν~520 nm, so that the particles absorb strongly in the green and we see them red in transmission (Figure 2b). We can also deduce one of the most important LSP property regarding applications: its sensitivity to the refractive index of the dielectric environment. When ξd changes, the position of the LSP is modified. The expression shown above can also be extended for small particles of nonspherical shape [18]. For instance, in the case of ellipsoidal particles, we have: ( = V

( − , (2 + L( ( −

with V the nanoparticle volume and 0≤L≤1 a factor accounting for the particle aspect ratio. Therefore, the particle shape also determines the LSP wavelength position. Actually, by varying the shape of the nanoparticles from nanorods to nanodisks, the position of the LSP can be tuned over a great portion of the spectra [18]. Both Eqs. (1) and

9

2.

and thus the particle polarizability, p [18]: = ( (

Page

nanoparticles and the nanostructuration of metal films and boosted the field (see Figure 1c).

(2) are valid for particles smaller than the metal skin depth. When the nanoparticles are bigger, the electrostatic approximation does not hold any more (the electric field is no longer uniform inside the particle) and we need exact calculations in order to obtain the response of the nanoparticle to the light. Those calculations indicate that the dipolar mode position does indeed depend on the particle size and that also higher order modes such as quadrupolar can be excited [17,20,21], giving rise to multiple resonances or LSP peaks. Another important property of an oscillator is the damping, which determines the quality factor Q of the resonance. For noble metals, the ones with stronger plasmonic behaviour, this quality factor is of the order of 10-100. Q provides a measurement of the enhancement of the local electric field associated with the LSP: it is proportional to a factor of Q2, and therefore between 102 and 104 [6]. When photon emitters, such as molecules, are placed in the vicinity of particles supporting LSPs, their absorption and emission rates are enhanced due to the increased value of the local electric field [22], and therefore governed by Q2. Processes involving two photons, such as SERS, depend on the square of the electric field intensity, and therefore they scale as Q4, so that the increase in the presence of metal nanoparticles can be huge, between 104 and 108.

Page

10

3.

Surface plasmon polaritons (SPPs)

Surface plasmons can also exist in planar interfaces, and in this case they are known as surface plasmon polaritons, as in this case the collective electron oscillations are coupled to a photon. For SPPs, the surface plasmon is not localized but it consists of an electromagnetic wave propagating along the interface and decaying exponentially across it [1,3,6,23,24] (see Figure 3a).

The existence of this kind of waves, which are transverse magnetic (i.e. the magnetic field is perpendicular to the propagation direction and parallel to the interface) can be predicted by solving Maxwell equations for a metal-dielectric interface. In the simplest case of a semiinfinite planar interface, the SPP dispersion relation (relation between the energy and the wavevector) is given by [23]: = , +

being kSPP the SPP wavevector and k0 the free light wavevector, = 2 /! = /" The electric (magnetic) field of the SPP wave can be expressed as: #($ (% = # ($ (% e'(()**+,-. e,(/|1| ,

with l=x,z for E and l=y for H (see Figure 3a). As it can be seen from this expression, the SPP is an evanescent wave whose penetration at each side of the interface, 7 234,5 = 1/2 34,5 with 34,5 = 6 − 4,5 7 , is the order of nanometers (a few nanometers towards the metal and a few hundreds of nanometers towards the dielectric) 1 . This strong confinement of the waves in the interface has motivated a strong interest of using SPPs as a way to develop integrated waveguides and photonic devices [1,4,5]. Unfortunately, the SPP dispersion relation also shows us that kSPP is a complex number since Îľm is complex too. This means that the SPP is damped while propagating due to absorption in the metal (see Figure 3b). However, the propagation distance of SPP, LSPP = 1/2klm SPP, is of the order of several microns for Au or Ag, and therefore still appropriate for the

1

For instance: for Au at visible frequencies (Îť0 = 633nm)

we have (800nm),

δ zair ≃

150nm, for Au at the near infrared

δ zair ≃ 300nm, and for Ag in the red (633nm)

δ zair ≃ 400nm. In all these cases, δ zmetal ≃ 10nm.

miniaturized

optical

More importantly, from the dispersion relation it can be seen that the SPP is at the right of the light line, i.e. kSPP > k0, which implies that the SPPs cannot be simply coupled with light directly incident on the planar metal surface, but we need to implement some methods to match the momentum [23,24]. Together with some more atypical arrangements, such as impinging on a local defect or slit [25,26,27], or coupling in the near-field [28], mainly two configurations are employed: the Kretschmann-Raether or prism coupling configuration, and the use of a grating. In the last case, diffraction from the grating of period d provides the lacking momentum: . In the prism coupling configuration, light is coming through a prism of refractive

Figure 3: (a) Sketch showing the surface plasmon polariton, SPP, electromagnetic field in a planar semiinfinite metal-dielectric interface. (b) Experimental image, obtained by scanning near-field optical microscopy, of a SPP propagating on a Au surface. The decay in the field intensity as the SPP moves forward is clearly appreciated.

2

For the case of 50nm of Ag deposited on glass, 100m at 0 = 800nm and LSPP 25m at

LSPP

633nm. For 50nm of Au on glass, LSPP 0 = 800 (633)nm, respectively.

30 (6) m at

index higher than that of the dielectric of the interface where the SPP is to be excited. For a very specific angle, SPP, the projection at the interface of the wavevector of the incident light equals kSPP, being therefore able to couple to it: . Having a look again at the SPP dispersion relation, it can be seen that SPPs, similarly to LSP, are also sensitive to the optical properties of the adjacent dielectric. This is the key for one of the most extended and commercialized application of SPPs, the surface plasmon based resonance (SPR) sensors, which will be explained in more detail below. Furthermore, the dispersion relation of SPPs can be shaped by structuring the metal surface in order to design ways of controlling the surface plasmon propagation, which is crucial for the development of miniaturized plasmonic devices. A periodic metallic structure opens a gap in the dispersion relation [29], and therefore waveguides can be fabricated by opening channels in this plasmonic crystals [30,31], or SPP Bragg mirrors can be built [32]. Stripes [33], grooves [34] or wedges [35] can be patterned to provide different plasmonic waveguides, or dielectric overlayers can be used to play with the SPP effective refractive index and generate optical elements such us prisms and lenses [36], waveguides [37,38,39] or mirrors [40]. Very thin films on metal embedded in a symmetric dielectric media [41,42], or the reverse structure metal-insulatormetal [43,44], also sustain SPPs with modified dispersion relations that could offer advantages in the design of devices. 4.

Plasmonic applications

The different surface plasmon properties (mainly the high sensitivity to the optical properties of the environment, the local field enhancement and the

11

of

Page

development devices2.

Figure 4:.Schematic representation of the SPR sensing in the Kretschmann configuration.

strong confinement) have given rise to a still extending applications portfolio, most of which can be grouped in two main areas: sensing and biosensing methods, and information processing (telecom) technologies.

Page

12

The category of sensing and biosensing techniques includes some of the earliest and most developed plasmonic applications. On the one hand, and based on SPPs and the wavevector dependence on the refractive index of the adjacent dielectric media as has been exposed above, the so called surface plasmon resonance (SPR) sensors have been largely implemented and commercialized [8]. This platform has been applied in many contexts and research is still in progress towards the development of new SPR sensors of higher sensitivity. The use of LSPs and the dependence of the resonance position on the environment properties within the sensing/biosensing field is more recent but is gaining importance [7,45,46,47,48,49], and one of the most successful examples is the commercialization of pregnancy tests based on gold nanoparticles. As mentioned earlier, coupling of photons into SPs can be achieved using a coupling medium (such as a prism – Kretschmann configuration or grating) to match the photon and surface plasmon wave vectors. To this end, a light is beamed upon a metal film through a prism and, at a certain incident angle (resonance angle), the plasmons are set to resonate with light, resulting in absorption of light at that angle. Due to

SP Resonance (SPR), the intensity profile of the reflected beam at the resonance angle exhibits a dip, or minimal intensity, seen as a dark line in the reflected beam. These SPs induce the field enhancement in the near-field region. The generated field propagates along the boundary and its amplitude decays exponentially with increasing distance from the surface The penetration depth of SPs is dependent on the wavelength (λ) and angle of incidence (α) of incident light, and refractive indices of the dielectric (n2) and coupling media (prism n1 and metal layer) and is linked to the wavevector k perpendicular to the interface given by: 2 877 D = 8E ∙ G H ∙ 7 − IJ87 ! 8E

Surface Plasmon Resonance (SPR) sensors gained a wide applicability as a direct, label-free and real-time approach to analyze biomolecular reactions occurring in the vicinity of a functionalized sensor surface. Accordingly, SPR provides quantitative data on Kinetics, Affinity/Equilibrium, Thermodynamics, Concentration determination, Adsorption rates. An SPR experiment measures the position shift of the dip (the angle shift) upon molecular adsorption or change in refractive index. The resultant SPR-time curve (sensorgram) can then be used to extract the kinetic parameters of the binding between the probe and the target or simply derive from a calibration curve the refractive index of the probe (see figure 5).

50 00

40 00

Response [relative units]

RU

5000

30 00

20 00

10 00

4000

Glycerol 3%

0 1.327

1. 32 8

1.329

1.330

1. 33 1

1.332

1.333

1 .3 34

RI

3000

2000

Glycerol 1%

1000

Water 0 0

100

200

300

400

Time [seconds]

Figure 5: Injections of glycerol solutions with different refractive indices. Inset the SPR response (RU) as a function of the refractive index (RI).

Within the information technologies category, the high field confinement provided by surface plasmons makes them attractive candidates for the development of miniaturized photonic devices [1,4], which hold the promise of building logic circuits as the electronic ones but working at light speed and also with reduced consumption. In the last decade long effort has been devoted towards the implementation of plasmonic waveguides and the analysis of the most appropriate configuration in terms of low losses and high confinement [5,33,34,35,37,38,42,44,53], being engraved channel waveguides [34] and dielectric loaded ones [54] the most favourable architectures. Based on them, not only linear waveguides but more complex configurations such as couplers, splitters or ring resonators can be obtained. Full progress of plasmonic circuitry also involves other components already on the route, such as detectors [55,56,57], sources [58,59,60] including lasers [61,62] or modulators [63,64,65],

Surface plasmons also play relevant roles in other fields, such as the development of metamaterials for negative refraction and cloaking [72], the efficiency improvement of solar cells [73], or optical tweezers [74], revealing the strong versatility of metallic nanostructures and predicting a bright future for them. 5.

Active plasmonics

The design and implementation of configurations where surface plasmons can be controlled by an external agent is denoted as active plasmonics, and constitutes an important aspect for the development of a complete plasmonic circuitry including active components (modulators, switches, etc.). Preliminary demonstrations of the feasibility of active plasmonics have already been published, using different control signals. Some of the first proposals were based on the thermo-electrical effect, which has succeeded in the demonstration of compact components such as MachZehnder interferometers or directional couplers [75,76]. This configuration has

13

Glycerol 5%

Fit

60 00

6000

which require plasmons to be controlled by an external agent, in what is called active plasmonics. The subject of active plasmonics will be discussed in more detail in the next section. The coupling of plasmonic nanoantennas to conventional lasers [66] or photodetectors [67] offers a new way to improve their performance, reduce their size or shape the outcoming light beam. The quantum nature of the surface plasmons makes them also interesting for the development of quantum computing, and initial demonstrations of the feasibility of the concept are starting [10,11,12,68]. Besides information processing, surface plasmons have also potential in the field of information storage, where they can help increasing the storage density through a process denoted thermally assisted magnetic recording [69,70], or in holography [71].

Page

Surface enhanced Raman Scattering (SERS) is another aspect of molecular detection, in this case due to the local field enhancement associated with surface plasmons [50], and the increased sensitivity achieves levels of single molecule detection [51]. To finish, we mention another relevant bioapplication of surface plasmons, thermal cancer treatment [52].

Page

14

the advantage that both the optical signal and the electrical one (that provides the temperature variation) travel through the same line, but the operation speed is too low. A faster option is provided through voltage control by using the electro-optical effect [77,64] or the field effect [63] although the modulation levels are in the limit for compact devices. The fastest option (femtosecond level) for SPP control up to now has been shown with all-optical methods, where a control beam affects the absorption properties of the metal or the dielectric in the interface and the SPP transmission rate is modified [78,79,80,81,82]. The magnetic field, taking advantage of the magneto-optical effect, provides an interesting and competitive alternative for control signal, both in terms of switching speed and modulation level [65,83].

Magneto-optics

1.

Magneto-optical effects

The response of a material to the external action of an electromagnetic field is described by the dielectric tensor. The commonly known as optical properties of any material are described by the diagonal elements of the dielectric tensor, and basically describe the reflection, transmission, dispersion and absorption response of a physical system to an electromagnetic field without any external excitation. On the other hand, the so called magnetooptical properties are related to the off diagonal elements of the dielectric tensor and describe the optical response of materials in the presence of an external magnetic field. For example, if the material is isotropic the dielectric tensor adopts the form = K −L$3 −L$D

L$3 L$M

L$D −L$M N

where “a� represents the MO activity of the material and Hi the components of the applied magnetic field (only linear terms in H are considered). For ferromagnetic materials, the non diagonal components of the dielectric tensor depend, not on the applied magnetic field, but on the components of the magnetization Mi, and therefore the dielectric tensor in this case becomes: = K −LO3 −LOD

LO3 LOM

LOD −LOM N

This implies that when light is reflected from a material, its intensity or

Here rpp, and rss are the reflectivity coefficients of the p and s-components of light, whereas rsp (and rps) represents the polarization conversion of ppolarized light into s-polarized light (and vice versa). It can be demonstrated that these coefficients are different depending on the direction of the magnetic field with respect to both the incident light and sample planes. In particular, there are three main configurations in which the MOKE effect may be classified: 1.1

Transverse configuration ( TU≠0, Ty= yU=0)

In this configuration, the magnetic field is applied parallel to the sample plane but perpendicular to the plane of incident light. In that case, the dielectric tensor becomes: =V 0 − M3

0 M3 0X 0

The transverse Kerr (TMOKE) effect manifests itself as a variation of the pcomponent of the reflected light. The scomponent does not experiment any variation (since we are restricting ourselves to linear contributions) and there is no conversion polarization from p- to s-polarization or vice versa. As a result the reflectivity coefficients RYI and RIY are zero, and therefore the TMOKE involves no change in polarization but a change in the light intensity of the ppolarized light. The quantification of the Transverse Kerr effect is performed analyzing the relative

17

Fundamentals [84]

Page

Magneto-optics

polarization sate may depend on the applied magnetic field. All the information related to this effect is summarized in the reflectivity matrix of the system. This matrix contains the Fresnel reflectivity coefficients and is defined as: RAA RAC S = QR RCC S CA

variation of the p-component reflectivity, defined as: ZO[\# =

of

]AA (+$ − ]AA (−$ ]AA (+$ + ]AA (−$

where ]YY(Âą$) represents the pcomponent of the reflectivity for the magnetized at saturation states at the two directions of the magnetic field. It can also be defined as: ZO[\# =

]AA (+$ − ]AA (−$ 2]AA (0

where ]YY(0) represents the pcomponent of the reflectivity for the demagnetized state. Typical values of this magnitude range from 10−4 to 10−2. The value of the TMOKE signal depends on the incident angle and on the optical and MO properties of the materials involved. It is worth being noted that the TMOKE signal is zero at normal incidence. 1.2

Polar configuration ( Ty≠0, Tz= yU=0)

The magnetic field in this case is applied perpendicular to the sample plane. The dielectric tensor presents the following form:

Page

18

Îľ=V− MD 0

MD 0

0 0X

When linear p-polarized (s-polarized) light reflects from a sample under such situation, the reflected beam presents a small component of s-polarized (ppolarized) light. The Polar MOKE (PMOKE) manifests itself in that linearly polarized light becomes elliptically polarized with its major axis rotated from its initial incident polarization plane. The RYY and RII coefficients are not modified. However, the coefficients related to the polarization conversion present non-zero and equal values (RYI = RIY ≠0) [85]. The

rotation of the polarization plane (? ) and the ellipticity state (_ ) are related to these coefficients as follows: RCA θ` + J_` = RAA 1.3 Longitudinal configuration ( yz≠0, Ty= xU=0) This configuration is characterized by an applied magnetic field parallel to both the sample and incident light planes. The dielectric tensor is described as: 0 Îľ = K0 0 D3

0 − D3 N

This configuration presents a similar behavior than the polar one: when ppolarized (s-polarized) light reflects from the magnetized material the light experiences a change in the rotation plane and ellipticity. Despite this similarity, the Longitudinal MOKE differs from the polar case in that the Fresnel coefficients corresponding to the polarization conversion present different signs (RYI =−RIY ≠0). Another important consideration is that in this specific configuration, there is no MO activity at normal incidence. 2.

Materials that exhibit MO activity

All materials exhibit MO effects, but the intensity of the effect varies from one another, most of them having negligible values. The main exception corresponds to the ferromagnetic transition metals, Fe, Co and Ni, which exhibit sizeable MO constants and therefore MO properties. As an extension, MO activity is also present in their alloys (FeCo, CoNi, FeNi) and the alloys with other elements. Special mention deserve the alloys with high spin orbit coupling metals, such as Pt and Pd. In these cases, specific crystalline structures of Fe and Co alloys with Pt and Pd give rise to MO activities which in fact are larger than those of the bare

3.

Types of MO characterization

There are basically two ways of applying MO measurements from a scientific (and in a longer term to an applied) point of view. The first one is to take into account the spectral dependence of the diagonal and off diagonal elements of the dielectric tensor, which in turn will necessarily imply a spectral dependence of the different MO effects. Therefore by continuously varying the photon energy of the light source used in the experiment it is possible to record the MO spectrum of a specimen. Actually, due to the direct dependence of the different MO effects with the off diagonal elements of the dielectric tensor, the combined acquisition of MO spectra together with a complementary spectral ellipsometry characterization (which gives information about the diagonal elements) allows in principle to obtain the whole dielectric tensor of a material system. This, supported with suitable theoretical calculations, makes it possible to extract information about electronic structure related modifications of the systems under investigation. On the other hand, it is also possible to use either of the MO effects to extract information of the purely magnetic properties of the system. Since currently MO investigations are carried out using light in the visible range, the limited penetration depth and its high sensitivity makes it extremely convenient to study magnetic properties in the ultrathin limit. Actually, MOKE (Magneto Optical Kerr Effect) is a term extensively used in

An extension of this can be performed if x-rays are used instead of visible light. In this case, the so called XMCD (x-ray magnetic circular dichroism) experiments allow for element specificity, since one can tune the x-ray wavelength to a specific atomic transition corresponding to an element constituent of the sample. However, this usually requires the use of large facilities such a synchrotron, where the capabilities for x-ray wavelength fine tuning are available.

Telecom applications Let us now briefly mention some applications in the telecom framework. Within NANOMAGMA we will focus on the investigation of non-reciprocal elements based on the combination of magneto-optical and plasmonic elements. Therefore, in this section we will pay special attention to those kinds of telecom elements. Nevertheless, magneto-optical storage is likely the most well known application of magnetooptics, so a brief description is also given. 1.

Magneto-optical storage: basic description

Magneto optical storage uses a static magnetic field and a laser to change polarization of a magnetic media. At temperatures below the Curietemperature the magnetic polarization is

19

Regarding non metallic systems, oxide garnets are especially important since they are dielectric systems, less optically absorbent than metals, with large Faraday rotation values.

works where magnetic properties of atomically thin ferromagnetic layers such as magnetic anisotropy, onset of ferromagnetism and magnetic couplings in ultrathin films are investigated. This is in all cases done at a fixed wavelength by recording the MO signal as a function of the magnetic field, either in the form of a hysteresis loop, or as a function of the relative orientation of the magnetic field with some characteristic axes of the specimen.

Page

ferromagnetic components, due to the intrinsic electronic structure modifications induced by the specific alloying and crystalline structure.

"frozen" into the media. However, when the system is locally heated above this temperature a static external magnetic field is used to modify the polarization state. When the media cools down below its Curie-temperature, the information is frozen again. A high power write laser is used to heat the disk surface to the appropriate temperature at which time the external field can set the polarization on the disk magnetic surface. The read-out process is based on the PMOKE effect described above. The interest in this technology was decreasing during the last years, but with the use of plasmonic elements to increase the locality of the heating (thus here the plasmon is used for local heating purposes) it is experiencing a rebirth. 2.

Integrated photonics: optical isolators

20

Creating compact on-chip nonreciprocal components is of fundamental interest in integrated optics at telecom wavelength of 1.55 µm. Today a lot of integrated functions are available commercially such as lasers, couplers, filters, modulators, etc] but there is a lack of non-reciprocal integrated functions like optical isolators and circulators (illustrated in Figure 6). In particular, optical isolators are essential to protect telecom lasers from optical feedback that can damage them or change their wavelength, thus introducing cross-talk between channels.

Page

Optical isolator / 3-port optical circulator A B C A

Today commercial optical isolators are expensive bulk components and the development of integrated optical isolators is a great challenge in the field of integrated photonics. It is currently a hot topic in this community. The challenges are compactness, low-cost and wafer-level fabrication. Industrial specifications set an isolation performance of 20 dB with insertion losses in the non-isolated direction below 0.5 dB.

4-port optical circulator : A B C D A Figure 6: Schematic of non-reciprocal components like an optical isolator and a 4-port optical circulator.

2.1

Commercial optical isolators: bulk components

Current optical isolators are bulk functions operating in the Faraday configuration illustrated in Figure 7. A longitudinal magnetic field is applied, i.e. along the propagation direction. This magnetic field causes a rotation of the incident polarization. The length of the magneto-optical (MO) material is adjusted in order to obtain a rotation of the polarization by 45°. A polarizer oriented at 45° with respect to the incident polarization is placed at the output of the MO material, so that the incident light is entirely transmitted. Reversely, a reflected light incident at 45°, i.e. along the direction of the output polarizer, undergoes a rotation of the polarization by 45° through the MO material. When it impinges the input polarizer, the polarization is at 90° and no light is transmitted.

The material used in commercial bulk isolators is iron garnet doped with rare earth like yttrium or ytterbium. This material has been chosen because it has nearly no losses at telecom wavelength. As a counterpart, its magneto-optical activity is relatively low, leading to a length of a few millimeters to achieve optical isolation. The Faraday configuration is obviously not suited for integration, due to: • •

•

The use of bulk magneto-optical material The conversion of polarization difficult to manage in an integrated photonic circuit The need of appropriate polarizers, also difficult to integrate

Commercial isolators can be found in the form of individually packaged components of cm size.

One nonreciprocal arm

x z y

A

B

M

nonreciprocal

nonreciprocal

M

A

B M Two nonreciprocal arms

Figure 8: Optical isolator based on the Mach-Zehnder configuration. Adapted from Ref. [86].

applied with respect to the propagation direction. In this case, the external magnetic field induces a non-reciprocal modification of the propagation constant in the forward and backward directions. This configuration is best suited for integration. Several configurations can then be used to form an optical isolator. The most popular one is the Mach-Zehnder configuration, as illustrated in Figure 8.

Integrated optical isolators

Integrated optical isolators based on garnet materials

2.3

Integrated optical isolators based on ferromagnetic materials

A more appropriate configuration for integration is the Kerr configuration, where a transverse magnetic field is

An alternative approach is the use a ferromagnetic material [88]. These materials have some absorption losses

21

2.2

Page

Figure 7: Principle of today’s commercial optical isolators: Faraday configuration.

There have been a large number of demonstrations of stand-alone optical isolators and circulators based on iron garnet waveguides [86], but they have not become commercial devices due to their lack of compatibility with InP and Si-based photonic platforms. Up to now, the expensive technique of wafer bonding is necessary to combine garnets with InP- or Si- based semiconductor materials [87]. However, garnet-based integrated optical isolators remain today the more mature technology with the best performances.

at telecom wavelength but from a technological point of view, they are compatible with existing photonic integrated circuits. Moreover, they have a much larger magneto-optical activity (two orders of magnitude larger than garnet materials), leading to much more compact devices and in contrast to ferromagnetic garnets, they have a permanent magnetization. They can work without applying an external magnetic field. For optical isolators using this material, there is a compromise between compactness and insertion losses. The magneto-optical Kerr effect induces a non-reciprocal shift of the complex effective index of the guided mode, as illustrated in Figure 9. Two kinds of optical isolators can be designed, either based on the non-reciprocity of the propagation constant (in a MachZehnder configuration for example) or based on the non-reciprocity of propagation losses. In this latter case, isolation effect can be obtained in a simple waveguide structure due to much higher absorption losses in the backward than in the forward direction.

Page

22

CoFe has been used experimentally in an integrated laser-isolator system operating at 1.3 µm wavelength, as illustrated in Figure 9, with an isolation ratio close to 100 dB/cm. In that particular case, absorption losses in the forward direction were compensated by current injection. At this stage, it can be concluded that ferrimagnetic iron garnets and ferromagnetic metals have their own advantages and drawbacks. In any case, there will necessarily be a compromise between magneto-optical strength for compactness and low losses. 2.4

Nanostructuration of the optical domains

A number of works are being devoted to analyze the benefit of nanostructuration

current injection laterally magnetized ferromagnetic metal contact

magnetization

current isolation

M

thin InP buffer layer tensile strained MQW active layer

toward light propatation

σeff M≠0 M=0 M≠0

backward TM mode

forward TM mode current injection

~ ISOLATION RATIO

neff Figure 9: Schematic and operation principle of a monolithic integrated optical waveguide isolator. The magnetization of the ferromagnetic layer causes nonreciprocity of the complex effective index of the structure. Adapted from Ref. [88].

to enhance the magneto-optical effects in iron-garnet materials and design much more compact devices. Periodic nanostructuration allows light localization or slow light modes and therefore a larger interaction with the magnetooptical materials. Magneto-optical enhancement was demonstrated theoretically with a 1D photonic crystal in 2003 [89], but the effect was too small for use in photonic integrated circuits. A significant theoretical work has been carried out with 2D photonic crystals consisting of air holes etched through an iron garnet membrane. [90] In this case, the large interaction with the magnetooptical material is obtained in a microresonator defined in the magneto-optical photonic crystal and coupled to three identical waveguides to form 3-port optical. Without an external magnetic field, the resonator sustains at least two degenerate counter-propagating modes

and the structure acts as a (reciprocal) add-drop filter. Conversely, when a magnetic field is applied, a mode splitting occurs and the structure becomes non reciprocal. 2.5

Nanostructuration of the magnetic domains

Page

An alternative approach to magnetic domains has been proposed very recently in Ref. [91]. It consists of an engineered cavity design, with the whole size and the position chosen in order to achieve maximal field localization at the transitions between magneto-optical material and air holes. The magnitude of the frequency splitting between two counter-propagating cavity modes is maximized. Obviously, the larger the frequency splitting, the larger the circulator’s operation bandwidth and isolation level will be. This component can be used in a uniform static magnetic field with very good performances of 20 dB isolation in a bandwidth of 80 GHz at 1.3 ¾m wavelength.

23

In order to enhance the magneto-optical effect, subdomains magnetized in opposite directions should be created, following the distribution of the magneticoptical field. This is reported in Ref. [86] for TE and TM iron garnet magnetooptical waveguides and in Ref. [90] for optical circulators based on magnetophotonic crystal resonators with 30 dB isolation over a bandwidth of 12 GHz at 1.55 Âľm wavelength. To the best of our knowledge, this remains a promising theoretical proposal for a magnetophotonic based optical circulator, one of the possible reasons might be that it is technologically very challenging to ensure a precise control of magnetic fields on the scale of ~100 nm.

Magneto-Plasmonics

Optical properties of composite materials are usually described in terms of different mean- field-like or effectivemedium approaches [92]. However, those approaches fail to give a proper description of materials when the interaction between neighboring nanoparticles is important, in general when the electromagnetic wavelength (位) becomes comparable to any characteristic length of the system. For instance, as 位 becomes of the order of the interparticle distance, multiple scattering effects (similar to mesoscopic quantum electron interference) are expected to induce strong changes in the magneto-optical properties of the system. This will be especially relevant for ordered lattices of nanoparticles where the MO properties will likely be dominated by Photonic Crystal (PC) effects. Though the first 1D MO-photonic crystals have been fabricated recently and their properties studied in some detail [93], little is known about MO-PC effects in 2 and 3 dimensions [94]. Even in disordered samples, depending on the fabrication method, the spatial distribution of nanoparticles is not fully random. The existence of spatial shortrange order in the particle distribution may induce interesting PC- like effects in the MO response of the sample. 1.

Basic concepts. Building blocks of MO devices

1.1

MO multilayers

Theoretical work on the MO properties of thin films and multilayers has long been a topic of interest mainly in connection with MO readout mechanisms for

Figure 10: Transfer Matrix Method for multilayered samples.

A generic Surface Plasmon Resonance (SPR) biosensor operates by measuring differences of the reflected light (Rpp) when the refractive index of the dielectric close to the metal surface changes. In these sensors, the angular region to perform the measurements is chosen to obtain the largest slope of the reflectivity curve and therefore the largest sensitivity. However, the slope of this

27

Modeling

Page

Magneto- plasmonics

information storage systems [95]. Though the first 1D MO-photonic crystals were fabricated more than ten years ago [94],and their properties studied in some detail [93], the optimization of plasmonic and ferromagnetic materials, to provide a high sensitivity for biosensing measurements, has stimulated our theoretical work in this direction.

curve is determined by the thickness of the metal layer, so at last the biosensor sensitivity is controlled by the thickness of the metal. The relationship between the wavevector modulation and the intensity of the SPPs EM field within the MO material is also of great fundamental interest. From a practical point of view, the knowledge of the role played by the EM field distribution would allow finding routes to enhance the modulation and thereby have a higher control on the SPP propagation. Specifically, we analyze the SPP wave-vector modulation in flat dielectric/metal and dielectric/metal/dielectric systems when a very thin film of MO material is buried inside the metal, in terms of its thickness, distance to the interface, and EM field distribution within it. In dielectric/metal systems we will establish a perfect correlation between the wavevector modulation and the EM field intensity in the MO layer. Let us consider a multilayered substrate as sketched in Figure 10 (see page 27). We assume an incoming plane wave given by

∙ ∙ E = (E , s + E , p )

with ki=k0ni= niω/c. Following Sipe’s notation, [96] k ≥ k = K + u K = u + u s (K) ≥

K Ă— u |K|

Âą p Âą

= (K ¹ γ u ) × s = (K) ≥ k × s

−|K|&u' Âą K |K|

Page

28

In terms of the reflection matrix, the “specular� reflected beam will be:

+ + + ∙ + ∙ E = (E , s + E , p )

∙ + 2 = ,-. / , + . / , 0s + -. / , + . / , 0p + 1 ()*

and the transmitted beam

∙ 6 ∙ E34(5 = (E ,3 s + E ,3 p 3 )

= ,-7 / , + 7 / , 0p

+ -7 / , + 7 / , 0s 1 ∙ 26 3

In compact “S�-matrix form: . / + + :/ = :. 9 / 3 < = 9 7 /+ 8 3 ; 8 7

. . 7 7

7> 7> .>

.>

7> /

7> = :/ = < 9 / 3+ < .>

.> ; 8/ 3 ;

To compute the Fresnel coefficients, we use the Transfer Matrix Method [97,98] •

Modeling Au/Co/Au trilayers

The combination of plasmonic and materials sensitive to magnetic fields, like ferromagnets, allows to obtain a new magnitude (∆Rpp) defined as the variation of reflectivity with the magnetic field (see TMOKE definition above) ∆Rpp = Rpp(+M) - Rpp(-M) This magnitude is also very sensitive to changes of the refractive index of the dielectric allowing developing a new kind of biosensor, the magneto-optical surface plasmon resonance sensor (MOSPRs). The sensitivity of the MOSPR sensor can be controlled by the thickness of the metal layer in the same way the SPR sensor does. Let us consider the system composed of Au/Co/Au trilayers over a glass substrate. Simulations of the MO Kerr signal, with and without SPP excitation for such structures, were performed via transfer-matrix method using the actual dielectric tensor elements of the materials obtained from ellipsometric and polar Kerr spectra [98]. We have shown that the magneto-optical activity presents a better sensitivity to changes of the refractive index than that of the optical activity in standard surface Plasmon resonance (SPR) sensors: Figure 11a and Figure 11b show respectively the reflectivity Rpp map, and the variation of reflectivity ∆Rpp with the magnetic field map for air background (n=1.33). The reflectivity map shows the total reflection condition close to 64Âş and the minimum close to 73Âş and 18nm gold thickness. The variation of the reflectivity

existing in each side of the metallic layer give rise to the (a) R (c) R well known symmetric/antisymmetric (or Low-indexmode/High-index-mode) modes. The EM field distribution of this coupled modes is still (b) ΔR (d) ΔR undisturbed by a thin enough MO film placed inside the metal. We find that the wavevector modulation depends on the difference of the EM field distribution of the non interacting SPP modes, Figure 11: (a) Reflectivity map ∆Rpp, and (b) pure magnetic part of the variation of reflectivity map ∆Rpp, as a function of angle and thickness located at the MO layer. These of the bottom gold layer for a n1=1.33 background medium. (c and d) assertions are confirmed by same as (a, b) but for a n2=1.34 background medium. Thicknesses: numerical simulations for Upper gold layer: 12nm; Middle cobalt layer: 4nm. air/gold interface and for a shows a maximum close to 70º and thick gold layer surrounded by different 25nm gold thickness. On the other hand, dielectric media, using cobalt as MO Figure 11c and Figure 11d show these material. magnitudes for a background with a change in the refractive index in the % 1.2 MO nanoparticles range (n=1.34). As it can be seen, the modification of the ∆Rpp map is much Electromagnetic scattering from stronger than that of Rpp map, confirming nanometer-scale objects has long been the higher sensitivity of the MO vs the a topic of large interest and relevance to optical measurement to changes in the fields from astrophysics or meteorology refractive index. to biophysics and material science. During the last decade nano-optics has • Magneto-plasmonic coupling developed itself as a very active field within the nanotechnology community. When an external magnetic field is Much of it has to do with plasmon applied parallel to the trilayer plane and (propagating) based subwavelength perpendicular to the SPP propagation optics and applications. Also, isolated direction, there is a non-reciprocal metallic particles supporting localized variation of the SPP wave-vector k upon plasmons have attracted a great deal of magnetization reversal [99]. We have interest due to their ability to concentrate performed a study of the relationship the electromagnetic field in between the wave-vector modulation subwavelength (some tens of and the EM field distribution in SPPs for nanometers) volumes. As a result, the dielectric/metal and dielectric/metal/dielectric studies in the field often involve the systems, when a very thin MO film is contributions of small elements or placed at different positions inside the particles where the dipole approximation metal [100]. For dielectric/metal systems may be sufficient to describe the optical the MO layer introduces a minor response. perturbation in the SPP electromagnetic field distribution meanwhile causes a The capabilities and applicability of all wave vector modulation proportional to these promising examples can be largely the amount of EM field inside the MO enhanced if some degree of tunability is layer. For dielectric/metal/dielectric added. These capacities can be systems, the non interacting SPP modes endorsed by exploiting different x-optic n1=1.33

pp

29

pp

pp

Page

pp

n2 =1.34

effect (thermo-, electro-, magneto-, piezo-) where an external agent modifies some elements of the dielectric tensor in some extent, which, in general, will be non-diagonal. Most of the previous works on small anisotropic spherical particles, consider the dipolar approximation (DA) in the electrostatic limit [101]. By taking into account the fact that the polarization within the sphere is uniform, the polarizability is usually written as ?@ ≥ 3v(C − CD I)(C + 2CD I)+G

being v the particle volume (I is the unit dyadic) and ξh the relative permittivity of the host medium at the point where the particle is placed. The host medium is assumed to be isotropic. For isotropic particles, radiative corrections to the electrostatic polarizability are well known [102]. However, these corrections had not been considered in the context of scattering from small anisotropic particles. We have recently analyzed the polarizability of small dielectrically anisotropic spheres including radiative corrections [103]. In the dynamic regime, one has to account for the field radiated by the particle itself, that also influences the particle response (radiation reaction). The dipole moment at a frequency ω is then written as the product of the exciting field E0 and an effective polarizability ιeff:

Page

30

p = C@ ÎąI33 E@

In the most general situation, the background medium can be inhomogeneous. We described the electrodynamic response of the background (i.e. the environment of the polarizable particle) by the Green dyadic G(r, r′, ω) = G0(r, r′, ω) + S(r, r′, ω) where G0(r, r′, ω) is the free-space Green dyadic and S(r, r′, ω) describes the heterogeneities. Note that the external field E0 has to be understood as the field in the background medium in absence of the particle.

In these conditions, the particle polarizability Îąeff is given by [104] ÎąI33 ≥ (I − & ?@ G )+G ?@

For the sake of clarity, we have introduced the interaction dyadic, Gp(r,r′,ω)=i Im(G0(r,r’,ω))+S(r,r’,ω). The first term in Gp describes the free-space radiation reaction and the second term describes the interaction with the structured environment. For a magneto-optical particle, the dielectric function becomes a tensor, C = CI + ∆C with ∆C = iMCA ; 0 A = O Q −Q

−Q 0 Q

Q −Q R 0

where, in most of the cases, the MO coefficient Q is much smaller than 1 and MO effects can be treated in a perturbative approach. The actual induced dipole can then be approximated as the sum of two terms [104]: p = pS + pTU = C@ (ÎąS + ÎąTU )E@

pS = C@ Îą@V W E@ , pTU = C@ W XYW

W = (Z − Îą@V & G )+G , X = 9\]MC/(C + 2)&

The first term corresponds to the polarizability of the particle in absence of the external magnetic field (i.e. the standard dielectric isotropic response), and will be denoted as “isotropic dipoleâ€?. The second term can be seen as a perturbation induced by the magnetic field, and will be denoted by “magnetooptical (MO) dipoleâ€?: This perturbative approach allows us to give a simple interpretation of the MO dipole: the electric field DpE0, due to the real exciting field E0 and the radiation of the isotropic dipole, excites an anisotropic particle with polarizability ÎąMO which by self-radiation induces the MO dipole pMO=Îľ0δDpADp. An important consequence is that the environment,

Magneto-Plasmonic coupling

In this section, we describe recent results [105] concerning the excitation of surface plasmon polaritons on a flat metallic surface by a magneto-optic nanoparticle placed at subwavelength distance from the surface, and illuminated by a linearly polarized plane wave. We show that the directivity of plasmon excitation can be controlled with the external magnetic field, through

Figure 12: Geometry of the system. In the LMOKE configuration that we consider, the external magnetic field is oriented along the x direction.

Our study is based on exact numerical simulations, using Green function methods [106] and the discrete dipole approach extended to deal with magneto-optical nanoparticles, MO-DDA (see below). These methods allow computing the near field in the geometry in Figure 12, and in particular the surface plasmon field excited by scattering at the MO nanoparticle. In order to compute a quantity relevant to experiments, we introduce a (near field) LMOKE signal under a modulated external magnetization: _`ab/ =

c/defgg (+`). î c − c/defgg (−`). î c &

&

c/defgg (+`). î c + c/defgg (−`). î c &

&

where /defgg (±`) is the surface plasmon field (near field) for a magnetization ±M along the x direction. An example of calculated near-field LMOKE signal map is shown in Figure 13 (page 32). In this case the modulated near field intensity is dominated by the surface plasmon field. A striking result is that the modulated surface plasmon excitation is highly directional. Beyond the numerical computations, we have performed a perturbative analysis that provides physical insight and simple rules for the design of magnetoplasmonic components. Although for brevity we do not describe the analysis here, we stress its major output. The high directivity of the modulated surface-plasmon

31

2.

the MO response of the nanoparticle. Considering the LMOKE configuration, with a modulated external magnetic field, we predict that the modulated part of the surface- plasmon field is highly directional. It is therefore in principle possible to launch and detect a directional surface plasmon with a nanoobject placed on an otherwise flat surface, and under linearly-polarized plane-wave excitation. A sketch of the geometry is shown in Figure 12.

Page

whose influence is described by the tensor Gp (and consequently Dp), modifies differently the isotropic and the anisotropic responses of the particle. This opens the possibility, for example, to relatively enhance the MO response by maximizing the ratio between the non- diagonal and the diagonal components of the effective polarizability (αji/αii) thanks to favorable external configurations. In the following, we study the interaction between a single nanoparticle and a flat surface, in the framework of the perturbative analysis described above. We study the possibility of enhancing the MO response of the nanoparticle following two different strategies: enhancing the MO contribution in the dressed polarizability, or selecting preferentially the radiation of the induced MO dipole. Both strategies can of course be combined in practice.

field results from a subtle interference effect between the excitation driven by the isotropic polarizability of the nanoparticle (standard dielectric response) and that driven by the anisotropic magneto-optic polarizability (magneto-optical response). Therefore the angular aperture is essentially given by the phase difference between these two polarizabilities, which depends on the size and the material of the nanoparticle. This study is a first indication that magneto-optical nanoparticles could be key components of magneto-plasmonic devices. A first experimental demonstration of the directivity effect shown in Figure 13 might be a first proof of principle.

Figure 13: LMOKE signal calculated in an observation plane located 10 nm above a silver interface. The configuration is that shown in Figure 12, with a 10 nm iron nanoparticle located 10 nm above the interface. The excitation wavelength is l = 800 nm.

Page

32

3.

Magneto-Plasmonic active control of single-molecule fluorescence

In this section, we introduce a general framework to study dipole-dipole energy transfer between an emitter and an absorber in a nanostructured environment. The theory allows us to address Förster Resonant Energy Transfer (FRET) between a donor and an acceptor (D-A) in the presence of a nanoparticle with an anisotropic electromagnetic response. In the particular case of a magneto-optical anisotropy, we compute the generalized FRET rate and discuss the orders of magnitude. The distance dependence,

the FRET efficiency and the sensitivity to the orientation of the transition dipoles orientation differ from standard FRET and can be controlled using the static magnetic field as an external parameter. 3.1

Generalized framework for modeling FRET in the presence of a nanoparticle

We introduce a generalized formalism to compute the FRET rate that allows us to deal with a D-A system in interaction with a nanostructured environment. This formalism includes as a particular case the standard Förster theory. Let us consider a donor (emitter) and an acceptor (absorber) with arbitrary locations and orientations of transition dipoles, in the vicinity of a nanostructure. We denote by (rA, µA) and by (rD, µD) the position and the direction of the transition dipole of the acceptor and donor, respectively. For electric-dipole transitions, the normalized FRET rate ΓDA / Γ0 can be calculated from the electric Green function which describes the electromagnetic response of the environment. In this work, we assume the weak coupling regime, which means that the transfer rate from donor to acceptor remains smaller than the spectral broadening due to the environment [107]. This excludes strong coupling and the associated coherent FRET mechanism [108], that are unlikely when the donor and acceptor molecules are coupled to a metallic particle at room temperature, which is the situation considered in this study. In these conditions, the normalized FRET rate takes the form [109]: w Γlm tm (u)vl (u)`(u) = 18pC@& q r s xu Γ@ ur @

In this expression, ΓDA is the energy transfer rate from donor to acceptor, Γ0 the decay rate of the donor in free space and ω the emission frequency. fD(ω) is the normalized emission spectrum of the donor, and the function M is defined by

3.2

Magneto-optical control of FĂśrster energy transfer

In principle, the presence of a nanostructure close to a D-A couple will modify the emission and absorption by the transition dipoles. The modifications are accounted for by the dyadic Green function that describes the electrodynamic response of the environment, through the function M(ω). This formalism leads to a very general treatment of the FRET transfer mediated by an external nanostructure, such as a nanoparticle with an anisotropic response. It permits a study of the influence of many parameters of practical relevance, such as the orientation of the transition dipoles, or the shape and material properties of the nanoparticle. Using the expression of the Green function in the presence of the nanoparticle, the FRET rate mediated by |g the nanoparticle, that we denote by Γlm , can be cast in a compact form [109] : |g Γlm .@ ( = 1) . =} ~ } ~ Γ@ .m .l

In this expression, R0 is the usual freespace FĂśrster radius, RA (RD) is the distance between the acceptor (donor) and the nanoparticle, and Rp is a new length scale, denoted by polarization coupling radius, that accounts for the influence of the nanoparticle. An important result is that the distance dependence differs from that of standard

(free space) FRET. Moreover, the polarization coupling radius Rp allows us to compare the indirect FRET rate (i.e., mediated by the nanoparticle) and the standard free-space FRET rate. For the sake of illustration, let us consider the situation in which the three bodies are aligned (see Figure 14a), with RD = RA = 2RNP and R = 4RNP (κ = 2). In this case, we obtain |g Γlm .@ ( = 1) . 1 . ~ } ~ = } ~ . @ = } .@ ( = 1) .|g 4 .|g Γlm

This simple expression shows that the ratio Rp/RNP is the crucial parameter that describes the influence of the nanoparticle on the FRET rate. For Rp > RNP, the nanoparticle enhances the FRET transfer, while for Rp << RNP, the FRET becomes exclusively driven by the direct transfer. We show in Figure 14a (page 34) the ratio Rp/RNP (with RNP =10nm) versus the emission wavelength of the donor for different materials that are known to exhibit a magneto-optical response (Nickel, Iron and Cobalt). For these materials, the enhancement factor of the DAFRET rate is of the order of 5, showing that in this case the FRET rate is substantially influenced by the nanoparticle. Figure 14b shows a computation of the ratio Rp/RNP with the same materials, but in the case of an orthogonal configuration, i.e., with the donor and acceptor dipoles orthogonal to each other. The magnetization is orthogonal to the plane containing the D-A couple and the nanoparticle. Let us stress that, in this configuration, the FRET rate vanishes in absence of an external static magnetic field due to the orthogonality of the donor and acceptor transition dipoles (Îş = 0). Although one observes that Rp/RNP remains smaller than one, the possibility of inducing a FRET rate driven only by the polarization anisotropy of the nanoparticle is an interesting result, showing the potential of magneto-

33

where G is the electric dyadic Green function that describes the environment of the donor and acceptor. The expression of the FRET rate is very general, and establishes the basis for a generalized FRET theory. In particular, as previously discussed in ref. [110], the Green dyadic formalism allows to handle an arbitrary geometry, including nearfield interactions.

Page

`(u) = |ym . z({m , {l , u). yl |&

(a)

degree of anisotropy can be controlled by an external static magnetic field, in the context of molecular fluorescence and FRET. In particular, this might suggest schemes for FRET tuning or modulation. The study has been limited to a proof of concept. Further work should focus on enhancing the (weak) magneto-optical FRET signal. 3.3

Periodic MO structures a)

Computational methods

(b)

a.1) Discrete Dipole Approximation (DDA)

Page

34

Figure 14: Ratio Rp/RNP for Iron (blue solid line), Nickel (gold dash-dotted line), and Cobalt (red dashed line) as a function of the emission wavelength λ of the donor. RNP=10nm. The configuration is illustrated in the inset, (a) showing that the dipoles are collinear and the couple Donor- Acceptor and nanoparticle are aligned. (b) Same as in (a) but in the presence of an external magnetic field inducing a magnetization in the direction orthogonal of the plane containing the tree body D-A-NP. RNP=10nm. The configuration is illustrated in the inset ((uD = µD), (uA = µA), and µA.µD = 0).

optical nanoparticles for FRET. On the one hand, the anisotropic response allows to couple molecules for which standard FRET gives a vanishing signal due to orientational mismatch. On the other hand, the possibility of controlling the magneto-optical response with a static magnetic field as an external parameter could allow to tune or modulate the FRET rate, which can be an advantage, e.g., to increase the sensitivity of the detection process. This study shows the potential of magnetooptical nanoparticles, for which the

The discrete-dipole approximation (DDA) is a powerful method for calculating absorption and scattering by targets with arbitrary geometry that have sizes smaller than or comparable to the wavelength of the incident radiation [102,111,112]. The DDA can treat inhomogeneous targets and anisotropic materials. Developed originally to study scattering from isolated structures such as dust grains, the DDA was extended to deal with small isotropic particles on a substrate [113], inside a multilayered system [114] or to targets that are spatially periodic and (formally) infinite in extent [115]. Here we present a further extension of the DDA method for a periodic array of arbitrary shaped anisotropic targets with heterogeneous composition. Our main goal is to compute the electric field diffracted by a plasmonic structure made of a plane substrate (slab or multilayered system) in interaction with a magneto-optical nanostructured grating. The multilayered substrate can have an arbitrary number of layers of magneto-optical, metallic or dielectric materials. The field can be computed in the near and the far zone, under plane- wave illumination or using a point electric-dipole source (in order to describe single-molecule fluorescence). Conceptually, the DDA consists of approximating a finite target of interest by an array of polarizable points with specified polarizabilities. The original

The volume of each object is considered as the union of nonoverlapping, simply connected subregions of volume Vi (i=1,...,N) Each subregion Vi is homogeneous and so small that the electric and magnetic fields can be considered as approximately constant.

−

−

The electric field can be written as a sum over reciprocal lattice vectors (or diffracted modes), as E 544 =

1 & e ƒ ƒ,g …(@)-K + †e, i, i 0p 1 - †0∙( + ‡) C@ CD  ‚

& Š‰ de â‰ĄÂƒ C@ CD ˆ( )‹Œ Â… g

†e

ˆ

(@)

-b ´ , i, i@ 0ÂŽ ≥ Â?(p ∙ s )s + (p ∙ p Âą )p Âą Â?

\ ¹ 2( + ) ‘ 2

where scattered field from each object is given by the radiation field of a set of N dipoles. The actual dipoles are obtained by solving a set of 3N X 3N equations. Once we solve for the electric field inside the object, the electric field everywhere (outside the objects themselves) is given by: Žˆ = C@ CD ]ˆ

CÂ…-rˆ 0 − CD +G ƒ Aˆ E@ (r ) ≥ C@ CD ƒ Θ ,ˆ E@ (r ) CD

E(r) = E@ (r) + E 544 (r) = E@ (r) +

deˆ (r)pˆ ƒ Š‰ C@ CD &

ˆ

The aim of the present treatment is to take an arbitrary patterned multilayer structure and calculate reflectivity or transmission spectra for light incident from the outside, or emission spectra from an internal dipole source. The calculation of reflectivity properties of MO multilayers can be done in a very efficient way by using the transfer-matrix method, as we have already shown in our analysis of Au/Co/Au trilayers. However, this method is mathematically unstable when dealing with arbitrary patterned multilayer structures. The Generalized Scattering Matrix (GSM) method provides a better computational approach. While the T-matrix gives the amplitudes of both incoming and outgoing waves at the surface in terms of those in the substrate, the S-matrix relates the amplitudes of the outgoing waves at the surface and in the substrate, to those of the incoming waves on either side of the structure. The S-matrix method for electromagnetic fields in periodically patterned multilayers was originally developed for non magneto-optical systems [117] and it was extended to MO systems in both polar (PMOKE) [118]. We have recently extended the GSM approach to MO periodically patterned multilayers with arbitrary magnetic field configurations [119]. The basic structure of the GSM method is illustrated in Figure 15(a) (page 36). The system is divided in multilayers. Within each layer the system is a periodic array in x and y directions (the dielectric function (tensor) will be consider periodic in the xy plane). Within an l layer, of thickness d, the magnetic field can be expanded as a sum over the reciprocal lattice vectors, e H(r) = ƒ ℎ -†e; i0 - †0∙ †e

e E(r) = ƒ -†e; i0 - †0∙ †e

35

Let us consider a periodic array of anisotropic objects embedded in an otherwise homogeneous media. Our formulation of the coupled dipole method relies on the same direct-space discretization scheme that is widely used to study the scattering of light by finite objects.

a.2) Scattering matrix method

Page

DDA method, also known as couple dipole approximation, was shown to be equivalent to a discretized version of the integral formulation of the Maxwell equations [116]. Once the polarizabilities are specified, Maxwells equations can be solved accurately for the dipole array. In this Section we present the formalism to extend previous approaches to deal with MO particles on layered substrates.

of the amplitudes of forwards and backwards going waves in different layers of the structure ( ) ÂĄÂ , } ‚( ´) ~ = } ¥ ´, Â

ÂĄÂ , ´  ( ) ~} ~ ¥ ´, ´ ‚( ´ )

The total S-matrix in the system is obtained from the appropriate matching conditions at each interface n. Once the S-matrix has been calculated, reflectivity, transmission spectra and Kerr angles are easily obtained.

Figure 15: Sketch of the layered system (a) for the Scattering Matrix Method and (b) for the combination of Discrete Dipole approximation and Scattering Matrix approach.

Each dielectric tensor describes one layer and it does not depend on the z coordinate. Within each layer, we first obtain the field states which are, in their z-dependence, simple plane waves. The H field is expanded in basis states with zero divergence ℎ -†e; i0 = — -†e0 ˜ ℎ -†e; i0 = — -†e0 ˜

1 ℎ -†e; i0 = − š(b + † )— -†e0 + -b + † 0— -†e0› ˜ ™

Page

36

This vector field must be a solution of Maxwell equation with periodic dielectric tensor in the xy plane which leads to a nontrivial eigenvalue problem. For a general MO configuration, the mathematical problem is to solve a cubic eigenvalue problem whose solution gives a set of eigenfunctions with their associated eigenvalues for forward propagating waves, and the corresponding solutions for backwards waves. For a layer of thickness d, the inplane components of the magnetic field can then be written as

â„Ž -†e; i0 = ƒ,— -Âœ, †e0 ˜Â? + ž -Âœ, †e0‚ Â?(Â&#x;+ ) 1

â„Ž -†e; i0 = ƒ,— -Âœ, †e0 ˜Â? + ž -Âœ, †e0‚ Â?(Â&#x;+ ) 1

and the z-component follows directly from the condition of zero divergence. The scattering matrix relates the vectors

a.3) Scattering Matrix and Discrete Dipole Approximation For some periodic structures, the eigenvalue problem (specially for longitudinal and transverse MOKE configurations) is poorly convergent and involves a very large number of reciprocal vectors. As a way to overcome this problem, we propose to combine the two methods described above: Scattering Matrix and Discrete Dipole Approximation. For a periodic arrangement of objects, the total field obtained from the DDA, can be cast in a form compatible with the field expansion used in the Generalized Scattering Matrix approach. The basic idea behind this combined approach is sketched in Figure 15: Let us consider a periodic target characterized by a dielectric permittivity tensor embedded in an otherwise homogeneous layer “lâ€?. We assume that the system is homogeneous (with a constant scalar permittivity) at the two boundary layers l − 1 and l + 1. The S-matrix connecting these two layers ( +G) ÂĄÂ +G, +G }‚ ( G) ~ = } ÂĄÂ G, +G Â

ÂĄÂ +G, G  ( +G) ~} ~ ÂĄÂ G, G ‚ ( G)

can be obtained as follows: The incident field on the l-layer (zl ≤ z ≤ zl + dl) containing our periodic objects can be written as

( ) ( )

˜Â? ( + ÂŁ ) @ -†e ; i0 = š (¢ − 1)s +  (¢ − 1)p

›

( ) ( ) + š‚ (¢ + 1)s + ‚ (¢ + 1)p + › ˜Â? ( ÂŁ Â&#x;ÂŁ + )

C -r 0 − CD +G A E@ (r ) ≡ C@ CD Θ , E@ (r ) CD

¤¥4 - e ; i > i + x 0 = \ + - e 0. ( + ) 1 & ,-p ∙ s 0s + -p ∙ p

0p 1 C@ CD 2

¤¥4 - e ; i < i 0 =

1 & \ + - e 0. + ( + ) ,-p ∙ s 0s + -p ∙ p + 0p + 1 C@ CD 2

Combining the DDA and GSM equations we finally have: © de +G s + Θ s

¡ +G, +G = ¨ +G +© de

p Θs

¡ +G, G = ¨ ¡ G, +G = ¨

+G +© de +

s Θs v © de +G p + Θ s + v

Θ de s

+G vs +G © de

vp Θs

©

© Θ de +G vs s + v ¡ G, G = ¨ +G © de + vp Θs v where

© de +G s + Θ p

ª +G +© de

p Θp

+G +© de +

s Θp v ª © de +G p + Θ p + v

© Θ de +G vs p

ª +G © de

vp Θp

© Θ de +G vs p + v ª +G © de + vp Θp v

± s , ≡ - 0. ± - + £0« , p

, e

+G ­ = X ­

≡ - 0. ± - + £0 ¬ e

& \ , v = X ­ £ 2 ­

These equations provide the scattering matrix of the layer in terms of the DDA solution of the periodic layer. Once we have it, the S-matrix can be combined with those of successive layers as in the standard GSM method. b)

Periodic arrays of MO nanorods

The study of light scattering from periodic structures has been a topic of interest during the last century. Already in 1902, Wood [120] reported remarkable effects (known as Wood’s anomalies) in the reflectance of onedimensional (1D) metallic gratings. Two

±

Since the observation of enhanced transmission through a metallic film perforated by a 2D array of subwavelength holes [123], there has been a renewed interest in analyzing and understanding the underlying physics of both reflection and transmission “anomalies”. Although the enhanced transmission in noble metals is commonly associated to the excitation of surface plasmons, dynamical diffraction resonances, due to constructive interference effects, can give rise to similar resonant phenomena [124] even in periodic arrays of dielectric particles [125]. Interestingly, arrays of transparent dielectric nanorods were shown [126] to produce very large local field enhancements at specific resonant conditions. Magneto-optics could provide a new way to actively control the resonant behavior of these structures or modulate field enhancements with the MO signal which would open new perspectives in device applications. Our aim here is to summarize our current work on this topic. Let us consider the geometry sketched in Figure 16 (page 38). Our system consists of an infinite set of parallel anisotropic nanorods with their axis

37

p = C@ CD ]

different types of anomalies were definitely identified by Fano [121]. One is associated to the discontinuous change of intensity along the spectrum at sharply defined frequencies and was already discussed by Rayleigh [122]. The other is related to a resonance effect. It occurs when the incoming wave couples with quasi-stationary waves confined in the grating. The nature of the confined waves depends on the details of the periodic structure and is usually associated to surface plasmon polaritons in shallow metallic gratings, standing waves in deep grating groove or guided modes in dielectric coated metallic gratings.

Page

Once we know the incident fields the outgoing fields can be directly obtained from the DDA Eqs.

along the z-axis and radius a much smaller than the wavelength in an otherwise uniform medium. For simplicity we assume an incoming monochromatic plane wave

+ + + ‘ ∙ + ˜‘ E@ (r) = E@ e ¯‘ ∙ = -/ÂŤ, s + /‹, p 0e e

With

s = (0,0,1); p + =

1 1 (™ , b , 0); p

= (−™@ , b@ , 0); @ @

s = (0,0,1); p + = (q°¹ ² , Âą\Âœ ² , 0); p

= (−q°¹ ² , Âą\Âœ ² , 0)

The solution of the scattering problem is given by the DDA equations discussed above. E(r) = E@ (r) + E 544 (r) = E@ (r) +

& deˆ (r)pˆ ƒ Š‰ C@ CD ˆ

beam, all the diffracted modes are evanescent and 1 1 ( & − b@& ) = ™& & Im�‰ (0)� = 2 ™@ 2 ™@ @ & Im,‰ (0)1 = & Im�‰ (0)� =

Since we are dealing with nanorods much smaller than the wavelength, in the present case, the object is described by a single dipole. It is then possible to solve the DDA problem analytically [125]. The self-consistent solution can be given in terms of a renormalized polarizability such that

1 & 2 ™@

In this regime, the depolarization factor can be computed exactly as a sum over the lattice. Near the onset of the diffracted beams, both Gb,yy and Gb,zz diverge and we can be approximated the sums as: & Re,zÂľ, 1 ≈ 0 & Re,zÂľ, 1 ≈

& Re,zÂľ, 1 ≈

Figure 16: Sketch of the geometry of the MO nanorod array.

1 b& 2 ™@ @

(b@ Âą 2p/Â )&

2 š(b@ Âą 2p/ )& − & &

2 š(b@ Âą 2p/ )& − &

Close to these divergences, due to the constructive interference of the scattered fields, the real part of the polarizability is zero (in absence of losses) and the grating presents dynamic geometric resonances. The resonant coupling of the induced dipoles in the grating is illustrated in Figure 17 for non MOnanorods. As it can be seen there is no resonant coupling for the dipoles oriented along the x direction.

Page

38

³ E (0), E(r) = E@ (r) + 2… ‰(r)? … @

where E0(0) is the incident field on the rod located at the orig in and & Âł +G = ? ´ +G − &. ,z Šddde Šde ? Â… Šddde @ Âľ 1 − \ ZQ,‰(0)1 +G ´ Šddde1, =? Šddde − &,z @

Below the onset of the first diffracted

Figure 17: Coupling of dipoles in the grating.

The reflected and transmitted fields can be written in terms of the specular reflection and transmission matrices: \ & Âł ∙ s , s ∙ ? Â… 2 ™@ & \ Âł ∙ p + , = p

∙ ? … 2 ™@

{ =

{

\ & \ & + Âł Âł ∙ s , s ∙ ? Â… Âş = p ∙ ? Â… ∙ s

2 ™@ 2 ™@ \ & + Âł + \ & Âł ∙ p + =1+ p ∙ ? Â… ∙ p , Âş = s ∙ ? Â… 2 ™@ 2 ™@

Âş = 1 +

Âş

\ & Âł p ∙ ? Â… ∙ s

2 ™@ \ & Âł ∙ p + { = s ∙ ? Â… 2 ™@

{ =

Let us discuss now the complex Kerr rotation for different configurations of the external magnetic field. The complex Kerr angle, θK, for “sâ€?-polarization, is given by { , ( ) ( ) = ²(¤4 + \ž)  {

Instead of considering an arbitrary external magnetic field and, in order to simplify the discussion, we will consider each different MOKE configuration separately: b.1) PMOKE (s-polarized light, E0 // z) Let us assume that the external magnetic field is oriented perpendicular to the grating plane (parallel to the incoming wave vector at normal incidence), i.e. mx =mz =0,my =1. For spolarized fields, the Kerr angle (keeping terms up to order Q2) is given by { , ? =− =− { ? C = \M & cos ²? + ž(MÂż ) p{ (C − 1)&

( ) ¤ 5(

-² 0

In this configuration there are no MOKEresonances due to the poor coupling between dipoles. However, notice that the reflection coefficient rss does present strong resonances near the Rayleigh condition for dielectric nanorods. (p-polarized light, E0 // x; Figure 18a)

Figure 18: MO coupling of dipoles in the grating.

At normal incidence, the system is almost transparent for p-polarized waves in absence of MO effects. However, there is a resonant PMOKE effect which take place at the resonant condition of spolarized waves (see Fig. 18a). For ppolarized light and normal incidence, -² 0

( ) ¤ 5(

with ? = ¨

(² = 0) = 1

?@,

{ , C = \M & ? { p{ (C − 1)&

− & Re,zÂľ, 1 − \ImÂ?z (0)Â?ÂŞ

2 = 2p{ & ¨Re Ă€ − 2p{ & & zÂľ, Ă (C − 1)

− \ Ă‚2p{ &

+G

& 2 − Im ĂƒÂŞ (C − 1) 2 ™@

+G

At the s-resonant condition, the real part of ιz is zero. In absence of absorption, i.e. for lossless dielectric nanorods, ²(¤4

( ) ¤ 5(

≈ −M

ž) Â

C  Ä (C − 1)& p & { &

( ) ¤ 5(

≈0

39

( )

Page

² = −

changes sign and becomes positive at resonance. Notice that the LMOKE at normal incidence gives zero Kerr rotation (even in absence of absorption, the resonance width goes to zero at normal incidence [126]).

At resonance, the reflected light is linearly polarized with a huge Kerr rotation (notice that the radius r of the rod can be much smaller than both the wavelength and the period of the grating). In general, for a symmetric nanorod grating, the ellipticity does not depend on the nanorod radius and is a function of only the material permittivity.

b.3) TMOKE (p-polarized light) Let us assume that the external magnetic field is oriented along the cylinder axis, i.e. mx =my =0,mz =1. As discussed previously the quantification of the Transverse Kerr effect is performed analyzing the relative variation of the p-component of reflectivity, defined as:

For metallic rods, however, the real part of Îľ is typically negative and there are no s-resonances. This configuration should then be useful for dielectric MO materials but not very appropriate for metallic nanowires. b.2) LMOKE (s-polarized light, E0 // z)

TMOKE =

Let us assume that the external magnetic field is in the plane of incidence but oriented parallel to the grating, i.e. my =mz =0,mx =1. We then have (keeping terms up to order Q2) that, far from normal incidence the Kerr angle is given by ( )  ¤ Ă….

-² 0

=−

? =¨

1

?@,

− & Re,zÂľ, 1 − \Im,z (0)1ÂŞ − \ Ă‚2p{ &

( ) ¤ Ă….

40 Page

{ =

+G

(C + 1) b@& − Im ĂƒÂŞ (C − 1) 2 ™@

At the p-resonant condition, the real part of Îąy is zero. In absence of absorption, i.e. for lossless dielectric nanorods, (far from normal incidence) ž) Â

. = c{ c

{ , C = \M & sin ²? + ž(M Âż ) { p{ (C − 1)&

(C + 1) = 2p{ & ¨Re Ă‚ − 2p{ & & zÂľ, Ăƒ (C − 1)

²(¤4

with

≈ −M

( ) ¤ Ă….

C  Ă„ (C − 1)& p & { & º Âœ²

≈0

At resonance, the reflected light is linearly polarized with a huge Kerr rotation. For metallic rods, the absorption broadens the resonant peak. Since, for metals, the real part of Îľ is typically negative, far from the resonance the ellipticity is negative and

. (+M) − . (−M) . (+M) + . (−M)

&

\ Âł ∙ p + = p

∙ ? Â… 2 cos² &

? ? \ C }? sin& ² − ? cos& ² + \M sin 2²~ + ž(M& ) (C − 1)& p{ & 2 Â™@

We then have: TMOKE ≈ 2Re Ă‚ +G

? ? C \M Ăƒ sin 2² (C − 1)& p{ & ? sin& ² − ? cos & ²

The TMOKE signal presents a strong peak at the minima of the transmission coefficients and, in particular, divergences at the zero reflection condition ? sin& ² = ? cos& ²

Notice, however, that this is just an “optical� effect and it is not a real resonant MO effect. Near the actual resonances, there are no relevant MO enhancement effects in TMOKE configurations. c)

TMOKE modulation of resonances and field enhancement

Grating structures based on ordered

arrays of nanoparticles has been shown to lead to huge field enhancements (FE) near geometric resonances. Resonant dielectric structures are then being explored as FE platforms for both oneand two-photon fluorescence excitation [126,127]. The MO modulation of the field enhancement factor could open new ways in the active control of fluorescent emission. The electric field near the surface of a nanorod is given by the sum of incoming and outgoing fields: E(r) = c-1 + & ddde zµ (r, 0)? 0E Ê (0)c| |Ë( ³ E (0) = ? +G + & ddde zµ (r, 0) ? @

where ? is the polarizability of an ³ is the dressed isolated nanorod and ? polarizability. The field intensity enhancement factor is defined as the induced dipole FE ≡

³ E (0)c 0? +G ? zµ (r, 0)? c-1 + & ddde @ |E@ (0)|&

&

is modulated by the MO signal in a way which resembles the definition of TMOKE parameter. It would then be possible to measure a TMOKE fluorescence parameter by modulating the fluorescence signal of molecules adsorbed on a MO nanorod in TMOKE configuration.

Fabrication of magneto-plasmonic nanostructures In this case we have adopted two different strategies for the development of magneto-plasmonic structures. One is based on the so-called top-down approach, and makes use of different lithographic techniques combined with physical vapor deposition (PVD) and/or electrochemical growth. The second is a bottom-up approach, where the structures are developed via colloidal chemistry synthesis 1.

Top-Down

Here we focus on the standard MOKE configurations described above.

³ E (0)c +G ? c? @ |E@ (0)|& It is easy to see that, for s-polarized plane wave illumination, the field intensity is a function of Q2 and does not depend on the sign of the external magnetic field. However, for p-polarized waves, the FE factor, near the resonant conditions, is given by &

? C+1 & Í (sin& ² + Re \M? sin 2²) 2p{ & C − 1

At normal incidence, the field enhancement increases as we approach normal incidence. However, the resonance width goes to zero. In absence of absorption, the resonant field enhancement, which has this dependence ∝Í

C+1 & λ r Í } ~ C−1 {

In general, it is desirable that magneto-plasmonic nanostructures extend over large areas. This has been realized through a top-down approach based on colloidal lithography that can be additionally combined with other techniques, such as atomic layer deposition (Figure 19). Colloidal lithography is used to structure nanodisks over large areas while atomic

41

FE ≈ Í

Figure 19: Approach used for the fabrication of magneto-plasmonic nanostructures based on gold nanodisks by colloidal lithography and thin films of magnetic materials by atomic layer deposition.

Page

FE ≈

layer deposition allows for the deposition of thin films of magnetic materials into close contact with the plasmonic nanostructures. Both techniques have been applied in different sequences to produce either isolated magnetoplasmonic nanodisks, or plasmonic gold nanodisks that can be either deposited over a flat magnetic (dielectric in this case) material or completely coated with the magnetic material. 1.1

from Präzisions Glas & Optik GmbH. Samples size is 1x1 cm2. The fabrication proceeds as follows: i.

ii.

Colloidal lithography

Colloidal lithography is a suitable technique for the fabrication of the plasmonic nanostructures as it can be easily applied over large areas. In this method, polystyrene nanospheres are used to pattern a mask to develop the nanodisks via PVD. The diameter of the disks can be easily modified by changing the size of the polystyrene nanospheres while the thickness depends on the duration of the material deposition. Because the polydispersity of the nanoparticles is very small, the diameter of the final nanostructures is almost constant and homogeneous over the whole sample. The nanospheres can be deposited on large areas thanks to the flexibility of spin coating.

iii.

iv.

v.

Page

42

vi.

Figure 20: Process of fabrication of nanodisks on flat substrates through colloidal lithography.

The fabrication of the Au nanodisks through colloidal lithography is schematized in Figure 20. The substrates used are normally BK7 glass

vii.

viii.

A PMMA A4 film is first spinned over the glass substrate (6000 RPM – 60 s) and pre-baked at 180 ºC for 5 min. The thickness of the resulting PMMA layer is ~ 360 nm. Oxygen plasma treatment (50 W, 450 mTorr) for 15 s of time is used to make hydrophilic the surface of the PMMA A4 film. This activates the surface for its later functionalization. A single layer precursor film is absorbed on the PMMA film using 0.02 % (by weight) poly (diallyldimethylammonium chloride) (PDDA, MW 400 000-/500 000, Sigma Aldrich), by spin coating (100 RPM, 60 s), rinse in water during 60 s and finally dry with nitrogen gas. This treatment makes the surface positively charged. Negatively-charged polystyrene nanoparticles (8 % w/v, 100 nm, sulfate latex, INVITROGEN) adsorb onto the charged substrates from solution by electrostatic interactions. Particle concentrations of 0.02 % (by weight) are used. Adsorption time is 1 min using Spin coating (100 RPM) to allow the adsorption to reach saturation and a better uniform distribution in all experiments. Excess particles are rinsed off under running water and the samples are blown dry with nitrogen gas. A Au thin film is then deposited by sputtering (15 nm) on the composition Glass substrate | PMMA film | Sulfate latex Spheres. Oxygen plasma treatment (50W, 450 mTorr, 15 s) is used again to clean the Au surface. Latex spheres are retired of the surface substrate using tape striping. Then the surface presents an antidot Au thin film stopping to the overdraft circles of PMMA. This sample is exposed in an oxygen plasma treatment using a

If the deposition consists of plasmonic and magneto-optically active materials, e.g. Au/Co/Au, the magnetoplasmonic element is finished. On the other hand, one can choose to prepare purely plasmonic disks, and then the deposition of magnetic materials with adequate properties is the next step to the further preparation of the magnetoplasmonic nanostructures. 1.2

Atomic layer deposition

Atomic layer deposition (ALD) is a powerful technique for the conformal deposition of homogeneous thin films of magnetic materials over large areas and with high thickness control. In ALD the substrate is sequentially exposed to two

In our approach, Ni and Co are selected first as soft and hard magnetic materials for the preparation of the hybrid magneto-plasmonic nanostructures. The magnetic film is produced by deposition of the respective oxide by ALD and subsequent thermal reduction to the metallic phase. ALD reactants are nickelocene (NiCp2), Figure 21: Representative AFM images and extinction spectrum of typical metallic nanodisks produced by colloidal lithography. cobaltocene (CoCp2)

43

Figure 21 shows an AFM image of the resulting gold nanodisk deposited on BK7 glass substrate. The diameter of these nanodisks is around 150 nm and they are found homogeneously distributed over the whole area.

or three reactants in vapor-phase. At each exposure the incoming reactant adsorbs on the substrate surface, where it reacts with the previous adsorbed reactant. Due to this sequential procedure, no reaction happens directly in the gas phase between the reactants, which would result in undesired chemical vapor deposition (CVD). The process is thus based in gas – solid reactions which take place at the surface substrate, assuring highly conformal and smooth films. Usually, the activation energy of the reaction is overcome with temperature by heating the substrate in the reaction chamber. Then the exposure sequence of reactants is repeated several times in a cycled process in order to thicken the deposited layer. Typical growth rates are in the range 0.1 – 3 Å / sequence, depending on the deposition process and the experimental conditions. Each individual exposure consists of three steps: 1) injection of the precursor in the chamber, 2) adsorption and reaction of the precursor on the surface substrate, and 3) purge from the gas phase of the unreacted / non-adsorbed precursor as well as other products resulting from the surface reaction. The duration of each step must be optimized for each precursor in order to obtain the best conditions for a homogeneous and conformal deposition over the whole surface.

Page

RIE to transfer the pattern from the gold layer into the PMMA. The substrate is then exposed at the bottom of the drilled holes. ix. The desired elements forming the disks are then deposited by thermal evaporation. x. After lift-off of the PMMA layer using acetone and an ultrasonic bath, the top gold layer (and the excess of materials for the disks) detaches and only the material deposited at the bottom of the holes remains, forming the nanodisks.

and ozone, and the process is carried out at 200 °C and 250 °C, respectively. A homogeneous deposition of the oxide is achieved, with good conformaility and growth rate 0.9 Å/cycle for nickel oxide and 0.5 Å/cycle for cobalt oxide. After deposition, thermal treatment is carried out using an Ar/5% H2 atmosphere at 400 °C for 3 h. In the case of nickel oxide, metallic nickel is obtained, as pointed out by the magnetic properties (soft ferromagnetic nickel with coercive field around 150 Oe), but critical morphological changes result from the thermal treatment (Figure 22). The oxide film collapses and discrete metallic clusters form on the surface, destroying the continuity of the film. Further analysis on NiO reveals a polycrystalline oxide with the cubic structure, but it is oxygen rich, resulting in a nonstoichiometric oxide (NixO with x ~ 0.88). We think that the use of ozone-rich oxygen gas as oxidising agent promotes the oxygen enrichment of the oxide by introducing nickel vacancies. The large difference of the unit cell volume between both structures (43,76 Å3 for Ni and 71,89 Å3 for Ni0.88O, a decrease of 40%) results in the collapse of the structure while H2O is released by reaction with the gas H2.

CoxFe3-xO4 films with different Fe to Co atomic ratios are produced by tuning the ratio of number of pulses of cobaltocene to ferrocene in the ALD process. The composition range extend from x = 0.5 (Co2FeO4) up to x = 4 (Co0,6Fe2,4O4). The magnetic properties depend strongly on the Fe to Co atomic ratio (see Figure 23a). The process is carried out at 250 °C resulting in a growth rate of 0,5 Å/cycle. As-deposited films are polycrystalline, with a low roughness, and consist only of the spinel phase. The films show the characteristic hard ferromagnetic behaviour of this type of oxides. The smallest saturation magnetization and coercive field belong to the lower iron content films. Increasing the amount of Fe increases both the saturation magnetization and the coercive field. Typical values are in the range 100 to 320 emu/cm3 for saturation magnetization and 20 to 150 mT for coercive field, where the largest values correspond to the

Page

44

As an alternative to nickel and cobalt, thin films of NixFe3-xO4 (nickel ferrite) and CoxFe3-xO4 (cobalt ferrite) are proposed for the fabrication of magneto-plasmonic nanostructures because of their good

magnetic properties (soft and hard, respectively), chemical resistance and versatility of the deposition process. Thus, further processing can be carried out without risk of oxidation or breakdown of film continuity which would result in loss of magnetic properties. These materials are grown by combination of the ALD processes of NiO, Co3O4 and Fe2O3. Cobaltocene, nickelocene, ferrocene and ozone-rich oxygen are used as precursors.

Figure 22: SEM image of a) as-deposited nickel oxide on BK7 glass by ALD; b) metallic nickel clusters after thermal reduction of nickel oxide layer, the inset graph corresponds to the longitudinal Kerr signal (in a.u) vs applied external field parallel to the surface measured with a NanoMOKE system.

composition CoFe2O4. Post-annealing treatments can be carried out maximum at 500 °C, the highest temperature sustained by BK7 glass substrates. This results in significant improvement of the hard magnetic properties, with coercive fields up to 6000 Oe, due to the recrystallization of the sample, which is already observed at this temperature. No critical changes like in the nickel films are observed after the annealing, except the recrystallization process. The magneto-optical characterization of Co1.6Fe1.4O4 films shows that the MO peaks are linked to the CT and CF transitions of Co2+, being the last the strongest (Figure 23 c). Moreover, as in the NPs, the peak is red shifted and the coercive field of the MO loops depends on the wavelength that was used to measure.

NixFe3-xO4 with different Fe to Ni atomic ratios is also produced by tuning the amount of nickelocene to ferrocene pulses. Fe:Ni ratios extend from 0.5 (Ni2FeO4) up to 2 (NiFe2O4) and the growth rate at 200 °C is 0,3 Å/cycle. The nickel ferrites show the pure cubic spinel phase and as-deposited are only partially crystalline. The magnetic properties depend only slightly on the Fe to Ni ratio. The films are soft magnetic, with coercive fields much lower, around 10 to 20 mT. Typical saturation magnetization values of as-deposited films are around 50 to 100 emu cm-3, lower respect to the as-deposited cobalt ferrites. The smallest coercive field is obtained at the composition x = 1 (NiFe2O4) (Figure 23 b). Annealing at 500 °C resulted only in a slight improvement of the properties without significant changes in the coercive fields. This is in agreement with the slight recrystallization process observed. Only after annealing at 700 °C using a suitable substrate, complete recrystralization is achieved, reaching saturation magnetizations close to that of cobalt ferrites (around 300 emu cm-3).

Page

Figure 23: Magnetic properties of (a) cobalt ferrite with different Fe concentrations and (b) nickel ferrite before and after annealing. Down: spectral magnetic circular dichroism for a selected cobalt ferrite film.

From the previous results, we consider that the properties of the structures obtained by colloidal lithography (gold nanodisks) and atomic layer deposition (nickel and cobalt ferrite thin films) are thus appropriate for the final realization of the hybrid magneto-plasmonic nanostructures. Figure 24 shows a SEM image of a BK7 glass substrate where

45

Figure 24: SEM image of BK7 substrate with gold nanodisks surrounded by a thin film (40 nm ) of magnetic CoFe2O4.

gold nanodisks are first grown and next a 40 nm thin film of CoFe2O4 is deposited. It is clearly seen the polycrystalline nature of the annealed ferrite as well as its conformal deposition surrounding the gold nanodots. 2.

Bottom-up

The scientific activity regarding the bottom-up approach has been devoted to explore the possibilities of wet chemistry colloidal synthesis of magnetic, plasmonic and magnetoplasmonic materials and focussing on the most promising materials for the studies of their magnetoplasmonic properties. This has led to different targets: 2.1- Optimization

2.2-

2.3-

2.4-

2.5-

Page

46

2.1

of transition metal based magnetic colloidal materials, with particular attention to magnetite and cobalt ferrite nanospheres. Preparation of plasmonic gold based metallic nanoparticles of different sizes, with varying plasmon resonance frequencies. Synthesis of hybrid magneticplasmonic materials with heterodimer and core@shell geometries based on transition metal oxides and gold. Preparation of polymer matrixembedded nanosystems based on the materials described above, both in the shape of optically addressable free-standing thick films and supported thin films. Monolayers of magnetic and magnetoplasmonic nanoparticles on silicon surfaces Ferrite based magnetic nanoparticles

The variety of systems prepared can be rationalized as a continuous series of materials ranging from magnetite (Fe3O4) to cobalt ferrite (CoFe2O4), obtained through an increasing

substitution of Fe2+ ions in magnetite with Co2+ ions. As a function of cobalt content, ferrites exhibit a variation in their magnetic properties (blocking temperature and coercivity increase with Co2+ molar fraction, as a result of the high single ion anisotropy of Co2+). In a similar way, spectroscopic MCD features vary significantly. A particular effort was made to control both size distribution and shape of the nanocrystals, in order to obtain a starting material that is at the same time highly performing and easy to model. The synthesis of such systems was carried out using procedures based on the high temperature decomposition of organometallic precursors, the route that at the moment is known to yield nanocrystals of the best quality. In addition to TEM, SQUID and MCD studies, the magnetic systems were characterised by ICP, in order to assess the Fe/Co ratio and by x-ray diffraction to check crystal size and quality. Finally, an EXAFS study has been performed to gain information on the distribution of Co2+ ions between octahedral and tetrahedral sites, which has a strong influence on the magnetic properties of the materials. 2.2

Gold based plasmonic nanoparticles

The magneto optical response of systems exhibiting localized plasmon resonances was confirmed to be extremely interesting by several investigations carried out in WP4. Both for this reason and to obtain good substrates for conjugation with magnetic moieties, we delved deeper into the control of the synthesis of colloidal gold nanoparticles. The particles we prepared are varied both in size, ranging from 3 to 25 nm and capping agent, form weakly interacting amine-based ligands to thiols, which exhibit a strong chemical interaction with the gold surface. These parameters allowed us to significantly vary the optical and magneto-optical

b

Figure 25: Bright field TEM micrographs of colloidal gold nanoparticles with average diameter of 16 (a) and 13 nm (b). Scale bars are 100 and 50 nm respectively.

2.3

Hybrid magnetic-plasmonic materials

Hybrid colloidal nanomaterials have been synthesised in three geometries: TM oxide-gold heterodimers, gold core @ TM oxide shell/hollow shell concentric systems and TM core @ gold shell systems. The systems we have synthesised feature a gradual increase in the degree of interaction between the magnetic and the plasmonic moieties, ranging from the very strong interaction found in core@shell systems to a weaker regime typical of heterodimer structures, to finish with simple mixtures of the two components, in which no direct interaction is present.

Stronger interaction

Strong interaction

Using this general method, several combinations have been obtained, ranging from systems with very small gold cores (3 nm), in which the plasmon resonance is strongly damped by the oxide layer, to big gold cores (8 nm), surrounded by a thin layer of magnetite (1 nm). As mentioned above, not only Au-Magnetite core@shell systems were obtained, but also more complex systems, namely Au@CoFe@CoFeO onion systems and Au-core@CoFeOhollow shell systems. In all cases, a strong shift in the plasmon resonance peak position was observed with respect to the corresponding unconjugated Au particle due to the difference in the dielectric constant between the organic capping layer and that of the TM oxide.

Weak interaction

In addition to TEM, optical and magnetic characterisations, a thorough MCD study is underway, in order to gain a

47

a

The synthesis is based on the heterogeneous nucleation of the TM oxide on the surface of Au particle seeds. Au seeds can be either preformed or prepared in situ before the TM oxide nucleation step. Gold surface acts as a catalyst to oxide nucleation, significantly lowering the activation energy of this process; thus, homogeneous nucleation is very efficiently prevented in favour of heterogeneous nucleation. Lattice mismatch between the two lattices (Au and ferrites) is within a few percent of integer values, so the oxide can grow epitaxially on Au faces. Whether the growth takes place on a single face or on multiple faces, leading to heterodimers or core@shell structures, respectively, is related to the polarity of the solvent used as reaction medium, and the effect is believed to depend on the ability of the solvent to shield more or less efficiently the surface charges that build up on the gold surface as nucleation takes place on it.

Page

response of the systems. Some representative bright field TEM images are shown in Figure 25.

deeper understanding of the possible electron interaction/delocalisation between the conjugated.

Figure 26: Schematic configurarions of heterodimers and of Au core at iron oxide shells

The inverse type of core@shell structures has been attempted using cobalt ferrite nanoparticles as the core material; the covering with a gold layer has been tried by two different procedures, thermal decomposition and ultrasound. The thermal decomposition procedure was performed following Ref. [128], while the sonochemical procedure was performed based on Ref. [129].

Page

48

Cobalt ferrite nanoparticles were synthesized by thermal decomposition, using two different precursors. In one case, cobalt (II) acetylacetonate and iron (III) acetylacetonate were used in a 1:2 ratio. In a different scenario, cobalt (II) oleate and iron (III) oleate were used in a 1:2 ratio. This was done with the purpose to compare the acetylacetonate and oleate metal precursors. In order to follow exactly the protocol established in [128], cobalt ferrite nanoparticles were prepared with acetylacetonate metallic precursors. These are used as seeds, over which the layer of gold will grow. Then 0.18 g Co(acac)2 (ca. 0.7 mmol) and 0.49 g of

Fe(acac)3 (ca. 1.3 mmol) were mixed in 20 mL phenyl ether with 2 mL oleic acid (6 mmol) and 2 mL oleylamine (6 mmol) under nitrogen atmosphere with vigorous stirring. 1,2-Hexadecanediol (2.58 g, 10 mmol) was added into the solution. The solution was heated to 210 °C and refluxed for 2 h. The suspension was cooled to room temperature and left like that for a future use. The preparation of gold coated cobalt ferrite nanoparticles involved the subsequent reduction of gold acetate in the presence of seeds. The suspension previously prepared was used without separation. 5 mL of phenyl ether reaction suspension of cobalt ferrite nanoparticles, ca. 0.42 g (1.1 mmol) of gold acetate, 1.55 g (6 mmol) 1,2hexadecanodiol, 0.25 mL (ca. 0.8 mmol) oleic acid), 1.5 mL (4.5 mmol) oleylamine were added into 15 mL phenyl ether. The reaction solution was heated under nitrogen atmosphere to 190°C with vigorous stirring. The reaction was kept at these conditions for 1.5 h. After cooling to room temperature, the product was washed with ethanol and redispersed in hexane. In order to use the sonochemical method, cobalt ferrite nanoparticles prepared by thermal decomposition need to be stable in water. Since the “as-prepared” particles were only stable in non-polar organic solvents, a surface functionalization by ligand exchange was carried out in order to confer water stability to cobalt ferrite nanoparticles. Surface functionalization was performed with DMSA, in order to displace the oleic acid molecules present on the surface of the magnetic nanoparticles. The functionalized particles had a final concentration of 1 mg/mL cobalt-ferrite nanoparticles in a water suspension. The procedure followed is now described. An aqueous solution of 0.1 mM HAuCl4 was prepared, for which every 50 mL has 100 µL of methanol, isopropanol, oleic acid, or 0.3 g of PVP

to obtain this type of patterns has already been tested by our group as an effective method to pattern cobalt ferrite particles on silicon. Finally, nanosystems exhibiting freestanding gold surfaces (gold particles and heterodimers) could be organised as monolayers on surfaces using sulphur chemistry; preliminary tests are being carried out in this direction. 2.5

2.4

Optically addressable thin films

Polystyrene embedded nanoparticles were prepared as thick, self-supported films by drop casting and as silicon, quartz, glass and gold supported thin films by spin coating. Through careful control of spinning time and speed, as well as initial viscosity of the polymernanoparticle solution, reproducible control over film thickness in the hundreds of nanometers range can be achieved. On the other hand, an additional study is underway to obtain patterned films of particles embedded in silicon oxide thin layers on silicon. This idea stems from the preliminary results obtained on silicaprotected NP arrays on glass that have been studied in the near field range with Scanning Near-field Optical Microscopy (SNOM). Since lateral resolution of SNOM is not suited to detect single nanoparticles, a very effective source of lateral contrast would be obtained by patterning stripes of silicon oxide that embed magnetoplasmonic particles. In this way, an efficient anchoring of particles to the surface would be obtained, as required by the SNOM experiment. The technology necessary

The concept underlying the formation of thin layers of nanoparticles on a surface stems from that of the preparation of self-assembled monolayers of molecules. The key driving force in the formation of the monolayer is the strong chemical interaction between a specific headgroup on the molecule and the chosen surface; such interaction also affords surface passivation after the first layer of adsorbed molecules, since incoming molecules will not find a reactive target surface, but the unreactive tails of adsorbed molecules, thus making the process self-limiting. When shifting from simple molecules to nanoparticles, an additional step has to be carried out, i.e. the functionalisation of the particle surface with the appropriate surface-binding molecules. In this case, it is more appropriate to consider the particle as a second surface, and the bifunctional molecule as a link between two active surfaces (i.e. particle and flat surface, see Figure 28).

49

Figure 27: Graphic scheme for sonochemical gold coating of cobalt ferrite nanoparticles.

Monolayers of magnetic and magnetoplasmonic nanoparticles on silicon surfaces

Figure 28: Schematic representation of the grafting strategy for nanoparticles on surfaces.

Page

(polyvinylpirrolidone). 100 ¾L of the cobalt ferrite 1 mg/mL aqueous solution is added, and for 30 minutes the reaction mixture is deoxygenated at room temperature with N2. Next, the reaction mixture is sonicated for 1 h in an ultrasonic bath Elmasonic S30 H with a main voltage of 220-240 V (50/60 Hz), not allowing the temperature to rise to more than 30°C. The reaction product was removed magnetically (Figure 27).

Since the chosen flat surface is Si(111) and the particle is made of cobalt ferrite, we used a molecular linker that contains an alkene and a silane functional group situated at the two ends of the molecule (see Figure 29), that are capable of selectively binding silicon and the oxide respectively.

dissolving the functionalised cobalt ferrite nanoparticles in toluene under nitrogen atmosphere together with the H-terminated Si(111) wafers. The system is then heated to toluene reflux temperature (120°C) and kept under reflux for 4 hours; the wafers are then rinsed and sonicated in fresh solvent in order to remove physisorbed particles.

a Figure 29: Molecular structure of the linker molecule used in this work.

The grafting process consists of three steps: I. Functionalisation of the NPs with the appropriate molecular linker. II. Cleaning of the silicon surface and treatment to prepare a H-terminated reactive surface. III. Grafting of the particles to the surface.

Page

50

Particle functionalization was achieved by a wet chemistry ligand exchange reaction: oleylamine/oleic acid capped cobalt ferrite NPs (Figure 30a) were suspended in toluene in the presence of an excess of the ligand “trimethoxy-7octen-1-yl-silane” and kept at 80°C for an hour under nitrogen atmosphere. After purification, cobalt ferrite nanoparticles suitable for grafting to silicon were obtained. Before proceeding with the deposition of the thin layer, the silicon surface is prepared in such a way that Si-bound hydrogen atoms are protruding; such hydrogen atoms are bound to react with the alkene functional groups situated around the cobalt ferrite particles. The treatment used to obtain such Hterminated surface consists of a brief etching with an aqueous fluoride buffer solution (HF-NH4F 40%-60%) after a thorough cleaning with solvents and oxidizing agents (H2O2-H2SO4 30%70%). The grafting step is carried out by

b

Figure 30: Bright field TEM micrographs of the cobalt ferrite (a) and gold@cobalt ferrite (b) nanoparticles used. Scale bar is 50 nm in both images.

This process is compatible with all ferritebased magnetic particles; in particular complex hybrids gold@ferrite core-shells and gold-ferrite heterodimers since both feature the same surface chemistry of plain ferrite nanoparticles. For this reason the process described above can be applied to complex hydrids. As an example we present in Figure 30b the grafting of gold-core@cobalt ferrite hollow shell nanoparticles. These particles were functionalized using the same linker molecule as described above (see Figure 29) and subsequently grafted to a Hterminated Si(111) surface.

1. Continuous systems

Due to the lower optical absorption of Ag compared to Au, one should expect better performance of Ag-based systems than Au-based ones. However, as Ag deteriorates easily in ambient conditions, a capping layer is mandatory. Pt and Au capping layers have been tested [132,133], and the best results were obtained in the latter case: Au capped Ag/Co/Ag trilayers structures have exhibited a TMOKE signal upon SPP excitation that is 150 times higher than that obtained without SPP excitation, and such enhancement is five times larger than the one obtained in Au/Co/Au trilayers in similar conditions (Figure 31, right).

∆R/R*103

Continuous systems were firstly studied in the pioneering work by Safarov and co-workers [130]. In particular, they fabricated Au/Co/Au trilayers and analyzed the effect of SPP excitation in the MO properties, finding an enhancement in the MO activity in resonance conditions. These results inspired the first experimental works of Another issue that has been explored is the Nanomagma consortium, in which the effect of both crystallinity (grain size similar trilayers but with varying Co layer or grain boundary scattering) and thickness were studied [131]. It was surface and interface roughness in the shown that the MO signal (TMOKE in that case) does not AuCoAu 80 PtAgCoAg*10 exhibit a monotonic AuAgCoAg increase as the Co 40 layer thickness 0 increases, but it -40 reaches a maximum for a -80 42 43 44 45 46 47 specific thickness incidence angle θ (º) (Figure 31, left). Such a thickness is Figure 31: Left panel: Kerr effect signal in transverse configuration, ∆R/R, as a function of the Co thickness under SPP excitation (squares for the experiment, blue the one that line for simulations) and without SPP excitation (circles and green line). Right panel: produces optimum Comparison of the SPP induced MO enhancement, for optimum Co thickness in each SPP excitation, case, between Au/Co/Au trilayers (green), Pt capped Ag/Co/Ag trilayers (blue) and Au capped Ag/Co/Ag trilayers. which has two

51

In this section we summarize our current understanding on the correlations between the magnetooptical response, and the optical one. Along the development of the project we have been dealing with near and far field properties. One of the conclusions we have extracted is that the extinction suffers a noticeable increase when the plasmon resonance (either extended or localized) is excited. Our experimental work has been focused on both continuous systems supporting Surface Plasmon Polaritons (SPPs) and nanostructured ones exhibiting Localized Surface Plasmons (LSPs).

Page

Magnetoplasmonic properties

consequences: 1) a strong decrease of the reflectivity of the system, and 2) an enhancement of the electromagnetic field at the Co layer (i.e. the effective magnetic field acting on the Co increases). The combined action of both effects gives rise to maximum MO response. It is important to notice that this MO enhancement depends strongly on the excitation conditions of the SPP, in particular on the refractive index of the dielectric environment: this phenomenon can be used to develop sensing devices, as it will be explained later.

52 Page

MP properties of continuous systems. By comparing the magnetoplasmoni c (MP) properties of epitaxial and polycrystalline Au/Fe/Au trilayers Figure 32: Left panel: Sketch of the magnetoplasmonic microinterferometers. Right panel: Wavevector modulation in Au/Co/Au trilayers as a function of the Co depth for [134], it has been air (black) and PMMA (red) on top of the structures. found that the magnetic field modulation in MP interferometers has induced SPP wavevector modulation in been already provided by depositing a the epitaxial structures doubles that dielectric layer on top of the MP obtained in their equivalent multilayer [136], leading to a seven fold polycrystalline ones, since the effective enhancement despite a reduction in the MO constants of the latter structures are SPP propagation length (Figure 32). reduced due to their rough interfaces. Also from the viewpoint of applications, However, both kinds on structures the magnetic modulation as well as the exhibited a similar MO activity upon SPP enhancement of the MO signal upon excitation, the key factor being in this SPP excitation have been exploited to case the achievement of the adequate develop the so-called magneto-optic thickness that leads to optimum SPP surface plasmon resonance (MO-SPR) excitation with independence of their sensors [137]. Unlike intensitycrystalline nature. interrogated SPR sensors that are based Other important aspect reported in the on measuring the reflectivity, the MOaforementioned seminal work is how the SPR sensors are based on the SPP properties can be altered by measurement of the TMOKE signal applying a magnetic field. In fact, it is when the SPP is excited. Compared to a shown that the SPP wavevector traditional SPR sensor, the MO-SPR depends on the magnetization of the device requires two changes that can be ferromagnetic layer, which can be done easily: 1) adding an electromagnet switched by an adequate magnetic field to apply a magnetic field in transverse of easily reachable intensity. So, the direction; 2) using a MP multilayer as magnetic field induced SPP wavevector transducer instead of a simple gold layer. modulation can be deduced by By doing so, two advantages are comparing the TMOKE signal and the obtained: on the one hand, for the same angular derivative of the reflectivity. A variation in the external refraction index, more straightforward quantification of the TMOKE signal changes more such modulation is obtained by MP abruptly than the reflectivity; on the other interferometry [135], in which SPPs are hand, the use of a modulation technique launched in a groove and propagate improves the signal-to-noise ratio, towards a slit where interfere with therefore enhancing the limit of detection. directly transmitted light: a weak Up to now, several MO-SPR sensing magnetic field is used to switch the experiments have been performed. In magnetization, therefore modifying the liquid environment, Au/Cr/Co/Cr [137] interference pattern. This modulation and Au/Fe/Au/Cr [138] transducers have capability of MP systems opens an been used: in both cases the MO-SPR avenue to the development of active sensor exhibited improved sensitivity plasmonic devices. compared with traditional SPR sensors with Au transducers (twofold for Actually, a first approach to increase the

Au/Cr/Co/Cr and threefold for Au/Fe/Au/Cr). Very recently, both Au/Co/Au transducers and Au transducers have been functionalized with porous TiO2 layers to perform gas sensing experiments [139], and the MOSPR sensor with Au/Co/Au transducers showed an improved sensitivity (twofold) compared to the traditional SPR sensor with Au transducers.

the electromagnetic field inside the nanostructure. To that end we designed an experiment to analyze the MO activity (a far field measured magnitude) of Au/Co/Au nanodiscs as a function of the Co layer position in the vertical direction. Placing the Co layer in different positions within the nanodisk allows probing the inhomogeneous distribution of the EM field inside it upon plasmon excitation.

MO enhancement appears to be highly correlated with a redistribution of the electromagnetic field to be spatially localized in the region of the metal nanostructure. On the other hand, we have also observed that an increase of the magneto-optical (MO) response takes place in plasmon excitation conditions, and thus highly correlated with an increase of the extinction and, as just mentioned, with an increase of the localization of the electromagnetic (EM) field in the region of the nanostructure. In most cases this will suffice to give a basic rule of thumb for most of the applications.

2. Nanostructured systems Nanostructured MP systems exhibit localized surface plasmon resonances (LSPRs), which take place at specific wavelengths depending on the materials, shape and morphology. Within the Nanomagma framework, two shapes have been studied: nanodisks and nanohole arrays (also known as antidots). Nanodisk arrays of multilayered composition, such as Au/Co/Au and Au/Co/Au/Cr, have been prepared in large area by colloidal lithography. The excitation of LSP resonances is revealed by peaks in the optical extinction spectra, although when compared to those of pure Au nanodisks with similar dimensions, the peaks exhibit broadening due to the large optical absorption of cobalt.

However, it is pertinent to go a bit further, since in some cases the nanostructures we have been dealing with present a complex structure (such as core shell, or multilayered nanodisks). In that case, it would be important to develop an intuition about the far field response, based on a more than near field property: the spatial distribution of

The interesting phenomenology is observed when the magneto-optical

44nmAu/6nmCo/2nmAu/2nmCr

12nmAu/6nmCo/34nmAu/2nmCr

0.2

0.0 ellipticity rotation

-0.1 -0.2

0.1 0.0

-0.2 1.5

2.0 2.5 3.0 Photon energy (eV)

3.5

ellipticity rotation

-0.1

1.5

2.0 2.5 3.0 Photon energy (eV)

3.5

Figure 33: From left to right: AFM image of an array of magnetoplasmonic nanodisks prepared by e-beam lithography, optical extinction spectra and polar Kerr spectroscopy spectra of an array of Au/Co/Au nanodiscs with Co near the surface (composition: 12 nm Au/ 6 nm Co/ 34 nm Au /2 nm Cr), idem for Co near the substrate (44 nm Au/ 6 nm Co/ 2 nm Au /2 nm Cr).

53

0.1

Page

400nm

Elipticity&rotation (ยบ)

Elipticity&rotation (ยบ)

0.2

54 Page

Kerr-effect spectra in polar configuration (PMOKE spectra) are measured (Figure 33, page 53). In the spectral region where the LSPR takes place, both the rotation and ellipticity spectra present anomalous features: the Kerr rotation (θ) spectrum exhibits an S-like shape and the ellipticity (ϕ) spectrum exhibits a peak. In fact, if the modulus of the complex Kerr rotation (|Φ| = ¹θ& + φ& ) is calculated, a clear resonance peak is observed at the wavelength of the LSPR. Moreover, the position of such maximum is redshifted as the nanodisk diameter increases, in agreement with the redshift in the extinction peak, which confirms the link between the enhancement of the MO response of the nanostructures and the presence of LSPRs. It is also worth noticing the evolution of the maximum MO activity with the Co position [140] (i.e. the depth which is located at), which is completely different to that observed in continuous multilayers with the same composition. In the nanodisks, the MO activity increases when the Co is closer to one of the planar surfaces (the bases of the nanodisk), whereas in the multilayers increases when the Co is closer to the upper surface. Such difference is due to the different mechanisms involved. In the multilayers, the optical absorption of the gold layer above the cobalt one plays the main role, and as a result the MO activity increases when the Co layer receives a higher amount of incident light, i.e. when it is closer to the upper surface; in other words: the deeper the Co layer, the smaller the MO response. However, in the nanodisks, the key factor is the electromagnetic field distribution associated to the LSPR excitation, and such field is enhanced near the two metal/dielectric interfaces: the upper surface between the metallic disk and the air, and the lower surface between the disk and the substrate. As a consequence, the MO activity of the nanodisk increases when the Co layer is placed near one of those surfaces, and

takes a minimum when it is placed at intermediate depth.

Figure 34: Left panel: Evolution with the angle of incidence of the TMOKE spectra for Au 16nm/Co 10nm/Au 6nm/SiO2 20nm/ thin films with Au nanodisk arrays (disk diameter: 110nm, height: 20nm) on top, with 400 nm of periodicity. Right panel: (a) Experimental dispersion relations of the LSP and SPP modes extracted from the TMOKE spectra (solid symbols), together with the dispersion relation extracted from calculations (continuous lines); (b) Evolution of the magnetic field-induced modulation for the different plasmon modes: LSP (dot-dashed black line, theory), SPP (red continuous line for theory, triangles for experimental data) and SPP for a SiO2/Au/Co/Au multilayer system without disks array on top (dotted red line, theory).

The structures made of Au/Co/Au thin films with a SiO2 spacer and a Au nanodisk array on top exhibit a more complex behavior, since they exhibit LSP resonances on the Au nanodisks and SPPs on the Au/Co/Au trilayer. Upon certain excitation conditions these two plasmon modes can interact and consequently modify the dispersion curves. Through the analysis of the magneto-optical Kerr effect spectra in transverse configuration (TMOKE) as a function of the incident angle, such interaction can be followed and the dispersion curves can be obtained. As is observed in Figure 34, left panel, the spectral location of a S-like structure (labeled with a triangle) depends on the incident angle whereas that of a peak (circle) is basically independent of it, which allows to associate the peaks with the LSP of gold disks and the S-like features with a SPP of the trilayer. From the comparison between the TMOKE spectra and the energy derivative of the reflectivity, the magnetic field-induced

modulation of the SPP wavevector (∆E) can be extracted (Figure 34, right panel). For the smallest wavevectors ∆E is nearly constant, whereas starts to decrease at higher wavevectors, those at which the LSP and SPP modes start to interact. This indicates that the interaction of the localized and propagating modes produces a strong reduction of the dependence of the SPP dispersion curve with the Co magnetization, this dependence being transferred to the LSP mode. [141] Finally, a surprising phenomenon has been reported in the case of Au nanodisks, i.e. without ferromagnetic material. By comparing optical extinction spectra with PMOKE spectra (Figure 35), it is unambiguously shown that such nanostructures made out of pure noble metals exhibit measurable magnetooptic activity when the LSPRs are

excited in the presence of a static magnetic field parallel to the propagation of the incident light [142]. It has been explained that such large magnetooptical response comes from an increase of the magnetic Lorentz force that is induced by the large collective movement of the conduction electrons in the nanostructures when the LSPR is excited. In the case of nanohole (or antidot) arrays, pure ferromagnetic structures made of Fe [143] and Ni [144] (i.e. Fe membranes and Ni ones) have shown to exhibit magnetoplasmonic properties. For instance, by measuring the PMOKE spectra (Figure 36) it is found a noticeable enhancement of the Kerr rotation with respect to that of a continuous film in the spectral region where LSPRs take place: such resonances can be ascribed to surface

200nm

Page

Figure 36: Left panel: AFM image of a Fe membrane. Central panel: relative reflectivity (top) and MO polar Kerr spectra (bottom) for two different Fe membranes (both 100 nm thick and with 470 nm periodicity) with holes diameter of 248 nm (open circles) and 297 nm (open squares), respectively. Right panel: experimental (top) and simulated (bottom) polar Kerr rotation spectra of a Ni membrane (200 nm thick, holes diameter: 45 nm, periodicity: 90 nm) with PMMA (n=1.5, open circles) and with air (n=1, black circles) inside the pores.

55

Figure 35: From left to right: AFM image of a representative array, optical extinction spectra and polar Kerr ellipticity spectra of two arrays of Au nanodiscs prepared by colloidal lithography : in black with aspect ratio AR= 2.2 (diameter D=70nm, height h=32nm) and in red with AR=6 (D=120nm, h=20nm).

plasmons propagating along the pores. Actually, it has been observed that the spectral location of the enhanced Kerr rotation region varies as the refractive index of the material inside the pore (n) is modified, being redshifted when n increases. A similar behavior has been obtained if the pore radius changes while keeping the pore concentration unchanged. Both effects are clear signatures indicating that localized surface plasmon resonances propagating along the pores govern the magneto-optical response of the membranes.

TMOKE spectra. One further degree of correlation between plasmonic and magneto-optical signal was investigated in single NPs that can exhibit plasmonic and magnetic properties (Figure 37). These NPs consist of a solid-solution of Au and Fe that is ferromagnetic. Both structural and magnetic properties confirm the Au-Fe alloying at the nanoscale. The OA spectra do not present any structure that can be related to the excitation of plasmon resonances in the nUV-vis-nIR range. Neither at room temperature nor at 1.7 K have the MO spectra exhibited a line-shape that can be associated to the plasmon resonance. Through experiments using different light powers, temperatures and time relaxation of the MO signal we determined that the

Finally, for the case of Fe membranes, the TMOKE spectra have been also analyzed [145]. The excitation of SPPBloch waves has been observed, producing clear signatures in the 1.0

Optical absorption

(b) Au50Fe50 0.5

Fe Au3

0.0

300

400

500

600

700

800

900

Wavelength (nm)

Page

MCD (mrad) at 2T

56

2.0

Fe

T=1.7 K

1.0

0.0

Au50Fe50 -1.0

(c) -2.0

400

500

600

700

800

900

Wavelength (nm) Figure 37: (a) TEM micrograph of Au50Fe50 NPs dispersed in silica, (b) OA spectra of Au50Fe50 NPs, Au NPs and Fe NPs, (c) MCD spectra measured at 1.7 K of Au50Fe50 NPs and Fe NPs and (d) MO loops measured at 632.8 nm and at 1.7 K simultaneously irradiated with light at 514 nm of different powers.

contributions of the thermal effects cannot explain the observed results. In this way we identify one new mechanism of magnetic relaxation in NPs based in the excitation of localized surface plasmon resonances that had not been described previously [146].

be very difficult to integrate on siliconbased photonic chips. Moreover their magneto-optical activity is rather low, leading to mm or even cm long devices. These are most likely the reasons why there are no integrated optical isolators commercially available yet.

In conclusion we observe that MM and MO loops in NPs are not equivalent. Such difference has not been previously reported. The origin of the effect can be related to the fact that coupling between light propagation and the magnetization depends on the direction of the magnetocrystalline easy axis that can be different for each type of MO transitions. A new result is that the magnetic and magneto-optical properties both in single NPs as in hybrid nanostructures are not correlated: the shapes of the MO hysteresis loops are different to that of the magnetometric measurement, it changes with the wavelength and can be composed by different contributions at specific wavelengths.

2D magneto-photonic crystals based on iron garnet materials are a very efficient approach to enhance magneto-optical effects and therefore drastically reduce the footprint of the devices. In particular micro-resonator based devices with smart engineering of the micro-resonator have been proposed theoretically, but the fabrication of so-called magnetophotonic crystals raises new technological challenges and the technological integration issue remains unsolved, so that there has been so far no experimental realization.

Magneto-optical iron garnets, transparent at telecom wavelengths and used in current commercial bulk isolators based on the Faraday effect, turn out to

•

By using non-reciprocity of propagation. An interferometric structure like a Mach-Zehnder interferometer is required to achieve optical isolation

57

Telecom applications

For those reasons, we performed a study of non-reciprocal waveguides based on magneto-plasmonic materials. Then optical isolation can be achieved in two ways:

Page

Considering these analysis we found that in the case of Noble metal/Metal transition/ Noble metal trilayers nanostructures the coupling between the magnetic and plasmonic effects can be reasonably understood in terms on the influence of the strong electromagnetic field on the MO properties of the magnetic parts. On the opposite, in the nanoparticulated based hybrid systems such correlation is not possible as the MO and Magnetic properties of the magnetic moiety are strongly changed in the heteronucleation processes. Bimetallic MT-NM nanoparticles could exhibit simultaneous magneto-plasmonic effects but only at low temperatures.

Therefore, we believe that there is room for the development of isolators based on magneto-plasmonic metals. Ferromagnetic metals like FeCo were little considered due to their absorption losses at telecom wavelength but they are compatible with silicon chips. In addition the magneto-optical activity of ferromagnetic metals is larger than for garnets and it should potentially be further enhanced by the combination with plasmonic materials sustaining lowloss plasmons. If we can find a suitable design to observe the enhancement effect, this means that at the same time the isolator length and the propagation losses should be reduced by using magneto-plasmonic materials instead of pure MO materials.

•

Or by using non-reciprocity of propagation losses, if it is large enough to achieve a good isolation ratio in a simple waveguide section.

1. Investigation of non-reciprocal magneto-plasmonic waveguides

preferential magnetization in the direction perpendicular to the layer. TbFeCo is a well-known material which has been widely used in magneto-optical data storage a few years ago. The objectives of these simulations are: •

We propose a generic non-reciprocal waveguide configuration shown in Figure 38. It is based on a standard silicon waveguide with a magneto-plasmonic cladding on top. The magneto-plasmonic stack used has to be compatible with standard microelectronics CMOS fabrication techniques.

Figure 38: Schematic waveguide structure.

of

magneto-plasmonic

Page

58

The more appropriate magnetoplasmonic metal can be typically an alloy of FeCo which can be easily deposited by PVD (Physical Vapour Deposition). The equiatomic composition leads to the best compromise between magnetooptical activity and absorption losses [147]. FeCo is a magneto-plasmonic material as it combines both magneto-optical (as a ferromagnetic material) and plasmonic properties (as a metal with relatively high conductivity). The waveguide is operated in the Kerr configuration with transverse magnetization with respect to the propagation direction and to the electric field in order to maintain the same polarization along the waveguide. As FeCo is spontaneously magnetized in the layer plane, TM guided modes will be used. However, if needed, TE guided modes could also be used with other MO materials such as TbFeCo with

•

to maximize the non-reciprocity of the waveguide and to assess the performances in terms of nonreciprocity (or equivalently isolation ratio), transmission in the forward direction (or equivalently insertion losses). The key parameter will be the FeCo cladding thickness. to understand whether the plasmonic properties of this material, provided that they are smartly used, can enhance its magneto-optical properties.

Once optimized, the non-reciprocal waveguide sections can be inserted in interferometric structures, like a MachZehnder interferometer as illustrated in Figure 39, or like a resonant ring sidecoupled to a waveguide. The Mach-Zehnder length necessary to achieve isolation is given by: •

•

_TÓ = Ä/(8pΔ )** ) if the two arms of the Mach-Zehnder comprise non-reciprocal waveguide sections with opposite magnetization _TÓ = Ä/(4pΔ )** ) if only one arm of the Mach-Zehnder comprises a non-reciprocal waveguide section

Alternatively, optical isolation can be achieved in a simple non-reciprocal waveguide section by using the nonreciprocal absorption losses like in Ref. 87.

Figure 39: Example of integrated optical isolator structure consisting of a Mach-Zehnder interferometer. Each arm of the interferometer comprises a nonreciprocal waveguide section. The sections are magnetized in opposite directions.

Simulation tool

1.2

Figures of merit as guidelines for design

First we investigate the guided modes supported by the magneto-plasmonic waveguide. For each guided mode of interest, we study the effective index neff and the propagation losses ι. Then we define several figures of merit to guide the optimization: •

the non-reciprocity of the effective index of the guided mode in the forward and backward directions or equivalently the non-reciprocity for opposite magnetization directions Ă•. D5 ) D *4 = =

•

−

Âľ5 Ă—Ă–5(Â&#x; Âœ)**

œ)** (` = 0)

œ)** (+`) − œ)** (−`) œ)** (` = 0)

the non-reciprocity of the propagation losses of the guided mode in the forward and backward directions Ă•. ¤ ) =

•

*¤(Ă–5(Â&#x; Âœ)**

?)**

*¤(Ă–5(Â&#x;

Âľ5 Ă—Ă–5(Â&#x; − ?)**

?)** (` = 0)

?)** (+`) − ?)** (−`) = ?)** (` = 0)

the product ? Ă— _TĂ“ where _TĂ“ = Ă„@ /4(Âœ)** (+`) − Âœ)** (−`)) is the Mach-Zehnder length required to achieve optical isolation. The

1.3

Material parameters

The ferromagnetic material chosen for this work is FeCo. Today, the full dielectric permittivity tensor of FeCo is not well-known. There is only one publication reporting its value at 1.33 Âľm wavelength [147]. Taking this uncertainty into account, we have studied the influence of the complex refractive index of FeCo on the properties of the nonreciprocal waveguides. The diagonal elements of the permittivity are written ´ ´´ Ă˜ = (ÂœĂ™)Ăš¤ + \ Ă™)Ăš¤ )& = Ă˜Ă™)Ăš¤ + \Ă˜Ă™)Ăš¤ Ă› ´´ with Ă˜Ă™)Ăš¤ = ĂœĂ?ĂžĂ&#x; For a conductivity à å‘

´´ tĂ™)Ăš¤ > 5 ∙ 10ĂŁ (equivalent to Ă˜Ă™)Ăš¤ > 46), the Si/FeCo waveguide supports a plasmonic guided mode with the maximum of the field intensity located at the interface between Si and FeCo. We ´´ have chosen the value of Ă˜Ă™)Ăš¤ = 46 in the following of the study and we study ´ the influence of Ă˜Ă™)Ăš¤ .

1.4

Simulation results

a)

Guided modes

First we investigate the guided modes supported by the Si waveguide with a FeCo cladding as a function of the FeCo thickness in the range 0-150 nm. The effective indices of the guided modes together with the mode profiles are shown in Figure 40 (page 60) for Ă˜ = −10 − 46\ . The genuine Si waveguide without FeCo cladding supports TE and TM fundamental modes with an effective index of 2.43 and 1.85 respectively. With a thin FeCo cladding the effective indices of these TE and TM guided modes decrease slowly. At 21 nm FeCo thickness a third guided mode appears. The mode profile is typical of a plasmonic guided mode with

59

A detailed modal analysis is a powerful tool to determine the non-reciprocal properties (the difference between forward and backward propagation constants and the difference between forward and backward propagation losses) of the waveguides and to determine the device length required to form an optical isolator. COMSOL 4.1 software, based on the finite element method, has been used for this study.

minimization of this figure of merit allows minimizing simultaneously the propagation losses and the length of the isolator device.

Page

1.1

thickness, the value of neff is higher for the plasmonic mode than for the photonic modes. This indicates the higher confinement of the plasmonic mode. Consequently, a better lightmatter interaction and therefore a better non-reciprocity are expected for the plasmonic mode.

Figure 40: Effective indices of the three guided modes supported by the Si/FeCo waveguide (TM plasmonic mode, TE fundamental photonic mode, TM fundamental photonic mode) as a function of the FeCo thickness.

the maximal intensity at the interface between the dielectric Si and the metal FeCo. Such modes are called DL-SPP modes (for Dielectric Loaded Surface Plasmon Polariton) because they are located at the interface of a metal layer “loaded” by a dielectric material. The examination of the components of the electric field shows that this mode is TM polarized. The effective index of the plasmonic mode increases with increasing FeCo thickness.

Page

60

Figure 41 presents the evolution of the effective index of the three different ´ guided modes for three values of ØTg (and therefore three different values of nMP) in order to investigate the influence of the relative value of the refractive index of the MP material with respect to the refractive index of silicon. For the usual photonic modes (TE & TM modes), the evolution of neff versus the MP material thickness is independent on the refractive index difference between the MP material and Si. For the plasmonic mode, the situation is different: the evolution of neff (particularly close to the cut-off thickness of 20 nm) depends on the refractive index difference between the MP material and Si. For nMP>nSi, neff increases with MP material thickness. For nMP<nSi, this is the opposite. Whatever the MP material

Figure 41: Evolution of the effective index of the guided TE, TM and DLSPP modes of the Si/MP waveguide as a function of the MP material thickness, for different values of the complex refractive index of the MP material.

b)

Propagation losses

Figure 42 shows the propagation losses (expressed in µm-1) of the different guided modes as a function of the MP material thickness. Whatever the value ´ of ØTg , the propagation losses of the

The propagation losses of the photonic modes reach a local maximum around 25 nm for the TE mode and 14 nm for the TM mode. The evolution of the propagation losses of the plasmonic mode is completely different. They decrease when the MP material thickness increases (decreasing red curves). Close to the cut-off thickness, the losses increase when the imaginary part of the refractive index of the MP material increases. For large MP material thickness above 60 nm, this is the opposite: the losses decrease when the refractive index of the MP material decreases. Regarding propagation losses, large MP material thickness is preferable. c)

The non-reciprocal effective index variations are plotted in Figure 43 for the three guided modes (photonic TE mode, photonic TM mode and plasmonic TM mode). For each mode, the magnetization is applied in the Kerr configuration, i.e. perpendicular to the electric field (in the y direction for the TE mode and in the x direction for the TM photonic and plasmonic modes). The first observation is that nonreciprocity is much larger for the plasmonic mode than for the photonic modes.

photonic modes are smaller than the losses of the plasmonic mode. This is not surprising because this is most likely due to the fact that the overlap of the electric field with the MP material is lower for the Figure 43: Evolution of the non-reciprocity of propagation for the guided TE, TM and photonic modes. DLSPP modes of the Si/MP waveguide as a function of the MP material thickness, for different values of the complex refractive index of the MP material.

61

For TE and TM photonic modes, there is a local maximum of non-reciprocity. For , the values of non-

Page

Figure 42: Evolution of the propagation losses of the guided TE, TM and DLSPP modes of the Si/MP waveguide as a function of the MP material thickness, for different values of the complex refractive index of the MP material.

Non-reciprocity of propagation

reciprocity are :  

-4

2 10 for the TE mode for a MP material thickness of 27 nm. 4 10-3 for the TM mode for a MP material thickness of 12 nm.

For the photonic modes, the values of non-reciprocity decrease when the refractive index of the MP material decreases. For the plasmonic mode, the evolution of non-reciprocity versus the MP material thickness highly depends on the relative value of the refractive index of the MP material with respect to the value of the refractive index of Si: 

For high refractive index of the MP material, non-reciprocity increases when the thickness of the MP material increases. For refractive index close to Si, the value of non-reciprocity is nearly independent on the MP material thickness. For low refractive index of the MP material, non-reciprocity decreases when the thickness of the MP material increases.

Page

62

The optimal value of non-reciprocity is obtained for a MP material refractive index close to Si and large value of MP material thickness (above 60-70 nm). Typically, the non-reciprocal phase shift is in the range 7 10-3 – 1.1 10-2 whatever the value of the refractive index of the MP material. It is higher by a factor of 2 to 3 with respect to the plasmonic mode. It can be noted that non-reciprocity of propagation and propagation losses do not follow the same evolution. This means that non-reciprocity is not directly linked to the overlap of the mode with the MP material. This point will require more

investigation in the remaining time of the project in order to understand more clearly what governs non-reciprocity. d)

Non-reciprocity of losses

The non-reciprocal variations of losses are plotted in Figure 44. This figure of merit is also interesting to compute, because if non-reciprocity is large enough, a simple non-reciprocal waveguide section can be enough to achieve optical isolation with a significant isolation ratio. Again, the non-reciprocity of propagation losses NR() is larger for the plasmonic modes than for the photonic modes whatever the MP material thickness and whatever the refractive index of this material with respect to Si. For the plasmonic mode, the value of nonreciprocity is in the range 1 - 5 10-2. The evolution of NR() greatly depends on nMP. There is an inversion of nonreciprocity, depending on the value of nMP with respect to nSi. When nFeCo>nSi, NR()>0 and when nMP<nSi,. NR()<0. e)

Global Mach-Zehnder performances

As stated above, the product has to be minimized to achieve the best performances of Mach-Zehnder based isolator devices. Indeed, minimizing this product will ensure the more compact devices with the maximum transmitted power in the forward direction.

Figure 44: Evolution of the non-reciprocity of propagation losses for the guided TE, TM and DLSPP modes of the Si/MP waveguide as a function of the MP material thickness, for different values of the complex refractive index of the MP material.

photonic modes whatever nMP. The optimum is achieved for nMP close to nSi In this case, the Mach-Zehnder length is the shortest with only 11µm. For all three kinds of modes, the best performances are achieved for nMP larger than nSi. In conclusion, better isolation performances are achieved when using the plasmonic TM mode of the Si/MP waveguide instead of the standard photonic modes. Propagation losses are larger for the plasmonic modes but nonreciprocity is much larger, leading to a smaller global product . This means that the larger non-reciprocity leads to a significant reduction of the required device length which more than compensates the larger propagation losses. 2. Fabrication 2.1

For the plasmonic mode, the product decreases when the MP material thickness increases. Thus it is preferable to have a thickness above 60 nm. This product is smaller than for

The principle is to fabricate the structures with a single lithographic step. The process flow proposed is the following: 

Deposition of the magnetoplasmonic stack directly on the SOI wafers. The magneto-plasmonic stack can simply consist of iron cobalt FeCo (combining magnetooptical and plasmonic properties at 1.55 µm wavelength) or a more complex combination of FeCo and copper which shows better plasmonic properties. The stack can be deposited either by IBD (Ion Beam Deposition) or by PVD

63

For the photonic modes there is a local optimum of MP material thickness to minimize the product : 27 nm for TE and 12 nm for TM for . The corresponding Mach-Zehnder lengths defined by are 247µm for TE mode and 55µm for TM mode.

We propose a process flow to fabricate such non-reciprocal waveguides on 200 mm Si wafers using the standard technologies of microelectronics and microsystems. The most convenient base wafers to fabricate the waveguides are SOI wafers, which are Si wafers with a thick layer of silicon oxide and a thin layer of 220 nm Si on top.

Page

Figure 45: Evolution of the product  x LMZ of the guided TE, TM and DLSPP modes of the Si/MP waveguide as a function of the MP material thickness, for different values of the complex refractive index of the MP material.

Proposed process flow

(Physical Vapor Deposition), preferably under a magnetic field of around 100 Gauss, in order to orientate the magnetic field of the ferromagnetic layer in a given direction. •

Deposition of a thin Ta cap of ~6 nm is deposited to prevent material oxidation

•

Deposition of a hard mask material (usually 100 nm of SiH4-based oxide grown at low temperature (< 300°C)

•

Photolithography to define the waveguides using deep UV projection lithography at 248 nm wavelength.

•

Reactive ion etching of hard mask using ICP tool (Inductively Coupled Plasma).

•

Stripping of photoresist (by O2 plasma etching and solvent chemical stripping)

•

Ion beam etching (IBE) of Ta cap and magneto-plasmonic stack

•

Reactive ion etching (RIE) of Si waveguide.

These process steps are illustrated in Figure 46.

Figure 46: Process flow of silicon waveguide with a magneto-plasmonic layer on top.

Page

64

2.2

Challenges

The critical point in this process flow is to ensure that the magnetization direction is orthogonal to the direction of propagation inside the wave guide in order to avoid polarization conversion during propagation in the waveguide. For that purpose one can force the

direction of magnetization by using the exchange anisotropy between ferromagnetic (FM) and antiferromagnetic (AFM) layers. Indeed a ferromagnetic film which is strongly exchange-coupled to an antiferromagnet will have its interfacial spins pinned. If the ferromagnetic layer is thin enough (below 30 nm), all the spins will be aligned in the same direction, and hence the anisotropy of the global magnetic field of the material. This phenomenon of exchange anisotropy is widely used in magnetic recording.

Figure 47: Optimization of the magneto-plasmonic stack to impose the direction of magnetization

Simulation results have shown that optimal performances of non-reciprocal Si/FeCo waveguides are obtained for an FeCo thickness above 50-60 nm. In this case, the 50-60 nm FeCo layer can be replaced by an IrMn/CoFe/IrMn trilayer or by an IrMn/CoFe/IrMn/CoFe/IrMn multilayer, as illustrated in Figure 47. The exchange anisotropy is set by annealing the ferromagnetic/antiferromagnetic stack at a temperature higher than the ordering temperature of the antiferromagnet and at high magnetic field (ex 260°C and 1T for IrMn). Other AFM materials could also be used like FeMn, PtMn, NiMnh.

Sensing There is a great interest in the world in gas sensing and biosensing for many

Various approaches have been proposed to enhance the sensitivity of such technique. These include, for example, phase-sensitive detection schemes, use of metallic nanostructures such as metal nanoparticles, line gratings, or hole arrays in metal films, modulation techniques can be used to improve the signal to noise ratio (SNR)

Very recently, a magneto-optical SPR sensor, based on the combination of magneto-optic effects and SPR has been proposed [153]. This sensor is based on a magneto-plasmonic (MP) modulation technique produced in multilayers of noble and ferromagnetic metals. This combination has shown a great enhancement of magneto optic (MO) Kerr effects of p-polarized light when the surface plasmon resonance condition is satisfied. Such enhancement strongly depends on the excitation conditions of the SPP and therefore on the refractive index of the dielectric in contact with the metal layer, thus providing the sensing principle of the MO-SPR device. If in traditional SPR sensor, the metal layer (usually Au) acts as the passive transducing layer, in this case it is represented by the active MP multilayer structure combining noble metal and ferromagnets with a suitable thickness. In Figure 48 a sketch of a typical experimental setup with test chamber for gas sensing and biosensing is shown.

Figure 48: Scheme of the experimental setup for MOSPR gas sensing

As discussed before, for the specific gas sensing application of SPR technique, an “active sensing layer� is needed onto the transducing substrate. Its active role consists in the change of its optical

65

Application of SPR technology can be found also in biosensors field. Generally, biosensors devices are systems that allow the analysis of biomolecular interactions; the investigated reactions are detected by a transducer that transforms a biological signal in a measurable signal. The fundamental basis of the biomolecular analysis is the specificity of the biomolecular recognition between the two present species, allowing the formation of a stable complex when one of the species is immobilized onto a solid interface. SPR biosensor provides real-time information on the course of the binding, it can be applied to interactions within a broad range of affinities, and it uses small sample volumes.

and enhance the limit of detection [150,151,152].

Page

purposes. A variety of sensing transduction methodologies are employed, depending on the target, cost and other factors. An appropriate transducing technique has to provide the most efficient transformation of the changes in the sensitive layer into an output signal. Surface Plasmon Resonance (SPR) technique is one of them. SPR is sensitive to the refractive index of the dielectric (gas or liquid) in contact with the supporting metal (Au or Ag), as well as to its thickness and refractive index of the thin films (the transducing layer) located at the interface between the conductor and the dielectric [148]. Various materials and technologies have been used to realize suitable transducing layers in chemical SPR sensors. Improvements are made in the development of new chemically sensitive materials and perfectly adapted transducer [149].

constants upon the interaction with the investigated gas which can be sensitively detected as a change in the reflectivity SPR curve. Organic, inorganic or mixed active sensing materials can be chosen, with various structural and morphological features to these purposes. Because of this, gas sensing capabilities depend critically on the synthesis method and parameters that allow for tailoring of selectivity and sensitivity toward the target species. In this project a novel application of such MO-SPR sensor is proposed in chemical gas and bio-sensing purpose. Thin films of nanoporous columnar TiO2 deposited by Glancing angle deposition (GLAD) [154] or thin films of synthesised TiO2 brukite phase nanorods deposited by Matrix Assisted Pulsed Laser Deposition (MAPLE) [155] have been chosen as sensing material to detect the presence of different alcohol vapours of interest in food applications. As concern ordered organic thin film multilayers of ethanebridged Zn porphyrins dimers macrocycles has been deposited by Langmuir-Shafer technique has also proposed as active layers in the detection of volatile organic compounds.

comprises an optical system, a transducing medium which interrelates the optical and (bio) chemical domains, and an electronic system supporting the optolectronic components of the sensor and allowing data processing. Major properties of an SPR sensor are determined by properties of the sensor’s subsystems. The main performance characteristics of SPR sensors include sensitivity, resolution and the lowest detection limit. Sensor sensitivity is the ratio of the change in sensor output to the change in the value of the measurand. Resolution is the smallest increment in the measurand that can be resolved by the sensor. Both the resolution and lowest detection limit of an SPR sensor are ultimately limited by the accuracy with which the SPR sensor’s output can be determined, which is limited by the stability of sensor’s baseline and noise of the sensor output. In this context the SPR optical characterization was performed using Kretschmann’s prism configuration in a home-made experimental setup (partially reported in Figure 49).

As concern gas and biosensing application, optimisation of the biological functionalisation procedures of these novel trasductors and increasing biosensing performance has been demonstrate finding superior sensitivity values than the conventional SPR sensors by using commercial biological probes and target proteins. Figure 49: Experimental setup for SPR and MOSPR characterization

Sensing and Biosensing Application

Page

66

1.2 1.

Experimental

1.1

MOSPR sensors: main performance characteristics

Generally,

an

SPR

optical

sensor

SPR and MOSPR coupled instrumentation based on prism coupler

Attenuated Total Reflection (ATR) in prism couplers has been widely used for the excitation of surface plasma waves

in SPR sensors. Particularly the Kretschmann geometry has been found to be very suitable for sensing and has become the most widely used geometry in SPR sensors and biosensors. The excitation of an SP wave results in a drop in the intensity of the reflected light. 1.3

Optoelectronic system for SPR activation

The optoelectronic system (Figure 50) includes surface plasmon resonance activation instruments and the related system for the optical signal detection. The p-polarized radiation (electric field parallel to the plane of incidence) is generated by a monochromatic He-Ne laser (λ=632,8 nm) with a beam section of 1 mm2 and emission power of 1 mW. A beam splitter provides two identical beams, whose intensities are measured before and after the reflection from the sensor surface, in order to calculate the reflectivity.

reflectivity (ratio between the incident and reflected light intensity) as function of the incidence angle (SPR signal). The SPR sensor is coupled to a commercial glass prism (BK7), transparent to a wavelength of 632.8 nm, with refractive index equal to n=1.515 also experimentally determined by measuring the Brewster’s angle. The prism-sample combination is placed on a rotating structure θ-2θ driven by a microprocessor-controlled stepping motor (Figure 51). This motor is directly linked to the computer via a dedicated control board, and allows the user to control the angular position of the sample (resolution of 0.01°).

Figure 51: Prism-Sample combination and test chamber for sensing activities

1.4

Detection and data processing system

A dedicated software developed with LabView handle the experimental data acquisition and processing. Through a simple graphical interface, this program allows to control the angular position of the sample and to obtain a continuous reading of the generated photocurrents. Choosing the right measurement mode, is possible to record both the SPR curve as function of the incidence angle or the variations of the reflectivity as function of time in a precise angular position.

67

The photons of the reflected beam reach the surface of a first silicon photodiode, which converts the light intensity in a photocurrent signal measured by a digital Pico-ammeter (KEITHLEY 485, µA scale). This system is directly linked to the user PC using a National Instruments PCI GPIB parallel interface, and allows knowing instantaneously the intensity of the radiation that reaches the prism-sample system. The beam component that passes through the splitter, reaches the sensor surface undergoing a further reflection. A second photodiode linked with another picoammeter, collects the photons of this radiation and allows to calculate the

Page

Figure 50: Optical unit of excitation of surface Plasmon resonance

As is possible to see in Figure 51, the sensor surface is placed into a test chamber connected to the fluid control system. This structure allows analyzing the optical and magneto-optical response of the sample in a controlled environment.

All these operations are displayed in a control panel (front panel) that acts as an interface between the user and the real program (block diagram), and several graphs allow to perform real time readings of the current values detected by pico-ammeters (see Figure 52).

centimeters per minute). Channel 1 controls a dry air flow that reaches directly the test chamber on the sample, while the second channel flow is deviated in a vial containing the analyte in liquid phase. Driven by this dry air flow, the saturated vapors of the analyte, present in the headspace of the vial, move first through a second empty vial and then mix with the flow coming from the first channel.

Figure 53: Fluid control unit scheme

1.6

Figure 52: Software interface for experimental data acquisition and processing

Page

68

1.5

Fluid control system

The assembled experimental setup allows obtaining different vapor concentrations using a home-made gas mixing station composed by a mass flow controller connected to two mass flow meters through a particular system of stainless steel pipelines and switching valves. This unit (shown in Figure 53) provides an excellent seal to prevent accidental contamination or any unwanted losses. Both flowmeters used (Channel 1 and 2) are connected to an outside line that supplies dry air and allow a maximum total flow of 50 sccm (standard cubic

Experimental setup for magnetoplasmonic characterization

The realized setup was equipped with an electromagnet in TMOKE configuration, in order to investigate the combination of SPR and MO effects (Figure 54). Bringing the magnet as near as possible to the sample, the magnetic field H present in the multilayer sensor can be approximated to a homogenous one. With an appropriate field’s intensity, the magnetic domains of the ferromagnetic material present in the MO-SPR sensors reach a saturation state, with a magnetization vector M directed parallel to the sample surface and perpendicular to the radiation’s plane of incidence. When the sample is illuminated with a ppolarized light, the physical quantity observed in this configuration is the reflectivity variation that occurs when the magnetization vector changes its value from +M to -M, changing the external magnetic field orientation: Δ. . (+`) − . (−`) = . 2. (0)

The magnetic field intensity must be sufficient to saturate the magnetic domains of the used ferromagnetic layer (see Figure 55), and this value can be obtained by measuring the Kerr effect signal as function of the transversal magnetic field intensity.

Figure 55: Magnetic domains saturation by an oscillating external magnetic field

Once the minimum field value is determined, the magnetization vector variation from +M to -M, can be obtained by simply changing the direction of H, as shown in Figure 55. Since the ∆R/R reflectivity variation is a very small signal, it’s necessary to use a particular measurement technique that allows achieving a great improvement of

In order to perform magneto-plasmonic characterization, the photodiode receiving the radiation reflected from the sensor is no longer linked to the picoammeter used in SPR mode, but to a low noise pre-amplifier. This instrument receives the generated photo-current and converts it into an amplified voltage signal. After this process the preamplifier output signal reaches the lockin where, thanks to a reference signal provided by a pulse function generator (HP 8116A), can be filtered and further amplified. In order to obtain a reference signal with the same frequency of the signal to detect, the pulse function generator, set with 800 Hz and 5 V values, has been used to both modulate the magnetic field variations and give a reference signal to the Lock-In. This purpose was achieved using a homemade electronic card (H bridge realized in CNR IMM laboratories) whose task is to change the electromagnet’s polarity with the same frequency of the pulsed function received from the generator. The amplified output signal of the LockIn is sent to a data acquisition board (PCI-6034 DAC CH1 National Instruments) installed on a PCI slot of the used computer. This board allow to convert the analog signal coming from the lock-in into a 16 bits digital signal (65536 possible values), sampling the input signal with a speed of 250 KSample/s divided between two channels (CH1 and CH2 with individual speeds of 125 KSample/s). 2.

Application

2.1

Bio-sensing applications

In order to have a first characterization of the proposed biosensing platform, a series of different concentration of ethanolic solution have been sent into the test chamber and the signal

69

Figure 54: MO-SPR setup with TMOKE configuration

the signal to noise ratio. This can be done using lock-in amplification techniques.

Page

In order to measure the ∆R/R signal, was used in an external magnetic field with a 50 Oe intensity (measured with a magnetic field probe GM07, Hirst Magnetic Instruments), oscillating with an 800Hz frequency.

Lock-in signal

-7

2.4x10

2.0x10-7 -7

0.4

MOSPR AuC oAu SPR AuCoAu SPR Au

0.3 0.2 0.1 0.0

1.6x10

-20

Normaliz ed signal

0.5

MOSPR

0

20

40

60

time (min)

80

100

1.3329 1.3338 1.3347 1.3356 1.3365 refractive index (RIU)

Figure 56: Calibration curves of SPR and MOSPR sensors and typical refractometric measurement.

correspondent signal have been recorded. Thus a first calibration curve has been build by comparing the sensor responses with respect to each change in refractive index in RIU (refractive index unit, the value for the considered ethanolic solution can be extracted from literature data) for the three investigated configuration: SPR and MOSPR using the multilayer Au/Co/Au on glass as transducing layer, and traditional SPR using the classical Au/glass substrate as transducing layer. The obtained results are reported in Figure 56 with a typical refractometric measurement recorded in MOSPR configuration.

Page

70

The increase in the sensitivity have been demonstrated either in comparison of MOSPR and SPR using the same transducing layer Au/Co/Au/glass and with respect to traditional SPR setup using as transducer layer a “classical� Au/glass substrate. It is important to observe how the calibration curves are obtained by a variation in the fourth decimal place of the refractive index. Once calibrated the MOSPR system for biosensing tests, a comparison between traditional SPR technique and MOSPR have been performed by using a typical protein-antibody interaction like BSA and anti-BSA. 2.1.1 Optimization of gold functionalization with BSA probes and Bio-sensing tests results

In order to attain this result a first step aimed at the optimization of the biological assay is needed. It consists in the development of a proper experimental procedure of functionalization of the transducing surfaces and of the following immobilization of the biological probes and successive hybridization with the selected antibody. One optimized this procedure, the following step are monitored by the two SPR and MOSPR signal in order to extract information about the sensor response, sensitivity, specificity and stability. As concern the biosensors analysis, synthetic BSA-antiBSA tests were performed to demonstrate the increasing in the sensitivity of MOSPR measurements respect to SPR. The protocol is described as follows: Goldcoated SPR or MOSPR chips were immersed for at least 48 hours in 150 mM solution of 11-MUA in glycerol/ethanol 1:1(v/v). Afterwards chips were rinsed with ethanol 95%, again with ultra pure water before placing in SPR or MOSPR chamber. Freshly 11-MUA-modified chip was placed in the test chamber and conditioned in ethanol 95%, followed ultra pure water. After conditioning, when stable readout was achieved, the carboxyl groups of 11-MUA SAM were activated by treating with 50 mM NHS in water/ethanol (10:1 v/v) for 5 minutes followed by 30 mM EDAC in water/ethanol (10:1 v/v) followed by a

measurements to demonstrate increasing in the sensitivity.

Figure 58: Main steps of surface functionalization.

−

SPR or MOSPR curve is recorded in water and PBS buffer condition prior any functionalization

−

The investigated transducing layer is properly functionalized following the experimental procedure described in the previous paragraph.

−

Then the SPR or MOSPR signal is again recorded in order to check the occurred functionalization step. From now all the experimental passages are carried out in the test chamber in order to monitor each step and have a precise control of them. After the activation of carbossilic groups by the proper solution, the following passage is the immobilization of BSA protein (50 ppm in a water solution). The chosen concentration has been experimentally proven to be suitable for a whole coverage of the functionalized Au layer.

−

After a blocking step with ethanolamine (1M) aiming to avoid a-specific interactions, a sequence of increasing concentration of antiBSA antibodies are sent into the test chamber and the signal recorded during the time in order to follow the dynamics of the interactions.

This protocol was used comparison of SPR and

in the MOSPR

−

Finally, when a saturation of all available sites is reached (namely when no further increase in the

71

All the stages concerning the measurement can be thus summarized:

Figure 57: Real time functionalization of the gold surface. In green, the variation in reflectivity of the immobilized BSA (delta R=3,22%) by covalent binding at time 130 min, and the hybridization of antibody antiBSA (delta R=9,55%) at time 205 min.

The 11-MUA-modified SPR and MOSPR gold chip was utilized for the covalent binding of 100 ppm BSA in watery solution. The immobilization procedure scheme is shown in Figure 57. The immobilization of BSA probe, performed via EDC/NHS coupling chemistry, was monitored constantly and the SPR angle shift was recorded.

the

Page

mixture of NHS/EDAC at the same concentration for 30 minutes. Subsequently chip was washed with 10% ethanol followed by ultra pure water until stable readout was obtained. BSA (100 ppm) was diluted with ultra pure water and was injected in the test chamber for at least 45 minutes. After this step a water solution was injected in the SPR cell to move away the unbound BSA until to obtain a stable readout. The chip was incubated in 1M ethanolamine in watery solution. Now the sensor is ready to hybridization with antibody. In this protocol has been used 20 ppm of anti-BSA in Posphate Buffer Saline (PBS). The following graph shows a real time functionalization of the chip. In the previous protocol we have used 10 mM of 11-MUA in ethanol solution for at least 48 hours, and 2 mM NHS in water solution, 6 mM EDAC in water solution for 30 minutes, holding the same steps and the same times. An example of graphs:

surface during all tests performed in liquid phase. The obtained results are really encouraging.

signal results from the correspondent increase in antibodies concentrations), a solution of HCl (1mM) is passed for few minutes into the test chamber in order to regenerate the sensor and eventually repeat the experiment.

The improved sensing performance reached by MOSPR setup with respect to SPR configurations is clearly evidenced in Figure 60 where the calibration curves relative to the interaction BSA protein and anti-BSA antibody is reported for different concentration of the selected antibodies. In both cases a typical Langmuir adsorption isotherm can be recognized which typically describes the dynamics of such kind of interactions.

The different experimental step can be clearly distinguished in Figure 59a. In Figure 59b the sensorgram relative to the detection of different concentration of BSA antibodies is reported. (a)

5.0x10

-7

4.0x10

-7

3.0x10

-7

2.0x10

-7

1.0x10

-7

Substrate Tiol BSA Anti-BSA

M O S P R s 66 S P R sA u

Lock-in

1 .8

1 .2

-1.0x10

-7

-2.0x10

-7

-3.0x10

-7

∆R/R

0.0

0 .6

0

2

4

6

8

0 .0

10

0

Incidence angle (deg)

1

2

3

4

5

C on c en tratio n (p p m )

(b) Figure 60: Calibration curves relative to MOSPR and SPR configuration for the BSA-antiBSA biological assay

-7

4.5x10

Lock-in signal (mA)

Sample 64 -7

4.0x10

-7

Ab10 PBS

Ab5 Et.amin.

3.5x10

Acq

PBS

Ab2 PBS

-7

3.0x10

HCl

Ab0.33

-7

2.5x10

PBS PBS

-7

2.0x10

PBS

-7

1.5x10

Acq Acq

A great increase of the sensor response can be recognized in the MOSPR configuration besides the improved sensitivity which can be extracted from the slope of linear part of the calibration curve.

-7

1.0x10

0

50

100

150

200

250

300

350

Time (mins)

Page

72

Figure 59: MOSPR signal recorded at each step of the biosensing experiment for the study of BSA protein- antiBSA interaction (a); typical sensorgram following the binding of different concentration of antibodies solution with the sensing probe. Regeneration step is also clearly evidenced (b).

The transducing layers used in this study are Au/Co/Au/Ti/glass, traditional Au/Ti/glass Au/Ag/Co/Au/Ti/glass. Different series of the transducing layers (prepared by IMM group in Madrid) have been tested with the aim to verify the reproducibility of the results and stability of the measurements and of the gold

The sensitivity values than can be extracted from the linear fit are of 0,47 ppm-1 and 0,81 ppm-1 for SPR and MOSPR sensor respectively, confirming the achievement of the goal at least of the deliverable. In order to take into consideration the stability of the sensor signal during time, the MOSPR signal has been monitored after two days in water conditions. The Au/Co/Au deposited on glass transducing layers have demonstrated a very good stability, reporting the same signal even after two or more days without any signal of ageing.

2.2

Once deposited onto Au/Co/Au multilayers (deposited in turns onto Corning glass substrates), the TiO2 layers have been characterized by a magneto-plasmonic point of view. SPR and MOSPR curves of the Au/Co/Au multilayers before and after the deposition of the TiO2 sensing layers are recorded, here reported in Figure 61 for the TiO2 thin films deposited by GLAD. The following gas sensing tests with the same analytes have confirmed the increased sensing performance of the MOSPR sensor with respect to SPR sensors as clearly represented in Figure 61 where the calibration curve for methanol vapours is reported as an example.

Gas- sensing applications

In order to complete the gas sensing characterization by using the novel magneto-plasmonic materials and the novel MOSPR technology, proper gas sensing test have been carried out by using different “sensing layersâ€? and different deposition techniques. As a first attempts, the main idea was to compare gas-sensing results obtained with TiO2 sensing layers deposited by Glancing angle deposition (GLAD) with nanocrystalline TiO2 prepared by colloidal routes. To this purpose sizetunable brookite (orthorhombic) TiO2 nanorods, capped with an organic shell of oleate/oleyl amine surfactants, are synthesized by a colloidal non-hydrolytic sol-gel route based on aminolysis of titanium oleate complexes at 280°C under air-free conditions [156].

Porous TiO2 layer deposited by GLAD have reported the best sensing performance with respect to colloidal TiO2 nanocrystal layers [158]. In fact the porous nature of the sensing layer makes him more sensitive to local refractive index changes at the Au/air interface. The effect of the thickness of the deposited layer onto the sensing performances has been investigated thoroughly (not reported here).

Moreover, thin nanocrystal layers suitable for sensing devices are realized by using the Matrix-Assisted PulsedLaser Deposition (MAPLE) technique [157], which has recently been developed for the deposition of a variety of organic and biological materials, starting from frozen precursor solutions as the laser targets. An extensive structural, morphological, and optical characterization of both- TiO2 thin films has been performed by combining highresolution scanning and transmission electron microscopy investigations, and UV-Vis spectroscopy.

In the figures are reported typical sensing measurements performed onto the investigated transductors (TiO2/Au/Co/Au/Ti/glass). Similar measurements have been also obtained for different Volatile Organic Compounds (VOCs) vapours. In any case an increasing in the sensitivity has been

140

SPR signal MO-SPR signal

Signal/noise

120 100 80 60

0 0

20

40

60

80

100

120

time (min)

Figure 61: MOSPR curve comparison of TiO2 thin films deposited by GLAD onto Au/Co/Au multilayers. Sensing test comparison relative to isopropanol and methanol alcohol vapours

Page

20

73

40

Langmuir–Schafer technique has been used to deposit organic chemicals like Ethane-Bridged Zinc Porphyrin Dimers (Figure 62), a prospective candidate as sensing element for volatile organic compounds detection. The surface layer containing the organic chemicals was transferred from the water surface by the LS deposition technique by lowering the substrate (Au/Co/Au/Ti/glass) horizontally until contact with the floating film. During the transfer we typically used surface pressure of 15 mN/m. Three monolayers (about 10 nm thick) has been transferred onto the transductor to perform sensing tests.

100 80

R(%)

In order to complete gas sensing characterization and to test the MOSPR performance also in the presence of a more selective sensing layer, organic materials (porphyrins macrocycles) have been also chosen. To this purpose a proper deposition technique has to be found in order to have a whole and perfect coverage of the transducing layer and ensuring a suitable sensing activity.

towards smaller angles demonstrates the deposition of the sensing layer.

60 40 20

AuCoAu AuCoAu+ZnPc

0 0

2

4

6

8

10

12

angle (°) -3

2.0x10

AuCoAu AuCoAu+ZnPc

-3

1.5x10

-3

1.0x10

∆R

obtained by using MOSPR methodology respect to the SPR transduction method.

-4

5.0x10

0.0 -4

-5.0x10

-3

-1.0x10

0

2

4

6

8

10

angle (°) Figure 63: SPR and MOSPR curves relative to of ethane-bridged Zinc porphyrins dimmers onto Au/Co/Au substrates

Page

74

Figure 62: Molecular structure and schematic representation of ethane-bridged Zinc porphyrins dimers

In our investigation we considered some VOCs which are of interest in food analysis. In particular those compounds with a well-known toxicity such as amines. The responses obtained from our sensing layers are reported together with calibration curves. In the following figures it is possible to see the SPR and MOSPR signals coming from the LS layers deposited onto Au/Co/Au transductor. The shift

Typical dynamic response curves toward try-butylammine vapours at increasing concentration are reported in the figures above. The improvement of the signal to noise ratio is apparent in Figure 64 in the MOSPR signal with respect to classical one, which cause the increase in the sensing performances of the gas sensor, namely the sensitivity and the resolution. This improvement is clearly represented in the calibration curves (Figure 65) where the sensor response is reported with respect to the concentration of the tested analytes, particularly by the slope of the curve, in its linear part, which gives important information about the sensitivity.

0.75

G

tri-Butiyammine vapours G

0.70 Reflectivity

G

0.65

G G Gas

0.60

dry air

dry air

0.55 0.50

0

50

100

150

200

250

300

Time (min)

-3

1.4x10

tri-Butylammine vapours

G

G dry air

G -3

∆R

1.2x10

dry air

G G

-3

1.0x10

Gas

-4

8.0x10

0

50

100

150

200

250

300

Combined SPR and electrochemical measurements can provide additional insights into various electrochemical phenomena and processes taking place at the solid-liquid interfaces. Examples include charge-induced molecular adsorption/desorption and structural changes of adsorbates, electrodeposition, and anodic stripping. Impedance evaluation has matured into a powerful technique for monitoring cellular systems (an area where SPR assays are notoriously difficult to perform). It is based on measuring the response (current and its phase) as a function of frequency of an electrochemical system to an applied oscillating potential (up to 50 mV) in the frequency range mostly spanning 1 Hz 10 MHz.

Time (min)

Figure 64: SPR and MOSPR sensorgrammes relative to of ethane-bridged Zinc porphyrins dimmers deposited onto Au/Co/Au substrates and exposed at different concentration of try-butylammine vapours

SPR coupling to other sensing techniques

Fostering combined Surface MagnetoPlasmon Resonance and Electrochemical Impedance Sensing assays within a unitary platform, involving measurements based on applying oscillating magnetic fields, at ICB were developed two highly sensitive frequency impedance analyzers. Each of

75

As surface plasmons can be excited by both electrons and photons, coupling of SPR with: ● Electrochemistry (including impedance); ● Fluorescence (to increase the fluorescence intensity or signal-to noise ratio) ● Photochemistry ● Magnetic actuation, holds potential for improved sensitivity and versatility.

Figure 65: Comparison of SPR (red) and MOSPR (Black) Sensor response relative to of ethane-bridged Zinc porphyrins dimmers deposited onto Au/Co/Au substrates towards different ammine vapours.

Page

2.3

The combination of SPR with impedance investigations (that expand the applicative potential of SPR assays) is apparently straightforward as the thin gold films used for the generation of surface plasmon waves can be simultaneously used as sensing electrodes.

them is able to perform fast EIS assays at a selected frequency, with capability to be coupled via a dedicated frontend (also accomplished) for simultaneous highly sensitive measurements at two frequencies (e.g., 50kHz and 500kHz, respectively). The system enables optical and electric addressing of interfacial processes related to a wide choice of functional surfaces and flow conditions. Polymer thin films or matching liquids (this is the convenient approach used in the following) enable optical interfacing with chips with variable configurations. 2.4

Surface Functionalization

Thus far, most bioaffinity studies involve monitoring the analyte or target in solution with the affinity (probe) species immobilized onto the SPR sensor chip. As the quality of the assay is highly dependent on the quality and stability of the sensing area, optimizations of the functionalization, patterning protocols as well as the regeneration protocols in order to extend the usage of the sensors are prerequisites. A typical SPR sensorgram for surface functionalization based on activated carboxy moieties is presented below.

mass transport with little carry-over from run to run, (3) automatic operation (4) low sample consumptions. The injected sample plug is diverted to the SPR flow cell by a carrier (buffer) solution, while the change in the SPR signal is monitored as a function of time. We developed an automated flow injection system compatible to biosensing assays based on SPR and Electrochemical methods. The dimensions and type of tubing, the syringe pumps and the valves as well as the flow rates were chosen to assure a smooth flow for an effective sensing. The entire FIA system is controlled by user-friendly software able to control the fluidics related to all specific preparation and sensing steps including injection of the sample, washing and regeneration.

Figure 67: The holder integrating the electromagnet and the flow cell.

36000 35800

Activation EDC/NHS

Response [relative units]

35600

Blocking Etanolamine

Imobilization Antibody

35400 35200 35000 10000

5000

5240 RU 0

0

500

1000

1500

2000

2500

3000

3500

Time [seconds]

Page

76

Figure 66: SPR monitoring of HIgG immobilization

2.5

Flow Injection

Flow injection (FI) SPR allows one to carry out SPR analysis in a continuous liquid stream. The advantages of FI include (1) higher versatility (2) improved

The flow injection addresses a measurement chamber with two channels made of PDMS - compatible with electrical measurements (tightness, biocompatibility, it is an insulating material, biochemically inert, easily replicable) with pairs for ports for fluid access. Electrical contacts are provided via spring loaded pins, with tips coming into contact with the surface, ensuring good electric contact via elastic pressure. A PMMA plate is used for strengthening mechanical structure and provide interface with the fluidics through Teflon ferules, while a PEEK pressure plate assembles all parts of the device and houses the electrical contacts.

The effective thickness di and dielectric constant εi of each layer in the system shape the reflectivity spectrum and influence the resonance angle and are important parameters in the construction of a transfer matrix (see Figure 10, page 27). Transfer Matrix [159] involves repeated application of the Fresnel equation [160], and relates the SPR angle shift to the complex permittivities of compounds in the multilayer system associated with the experimental platform. The transfer matrix combines the entire set of field components and involved layers and can therefore be used to calculate the reflectivity of the complete system provided that thickness and refractive index (permittivity) of all layers, the wavelength and the angle of incidence α are given. Different metal structures reshape the reflectivity spectra as revealed in the figures below, therefore the transfer matrix approach is a powerful tool in chip design (to select the best configuration in terms of thickness and arrangement of sensing layers – a critical issue for magneto-optic chips).

Figure 68: Letf: The SPR curve on Cr (2 nm) Au (50 nm) layers. Right: The SPR curve on Cr2nm/Au26nm/Co6nm/Au12nm layers

2.7

Data analysis

SPR analyses related the SPR sensorgrams (i.e. time variation of the reflectance dip position, or SPR angle) to the quantity of interest, assuming one effective layer (characterized by an effective thickness deff, and dielectric

constant εeff) on top of the SPR chip. Traditionally, simplified kinetic models, two-state or parallel reaction models, have been proposed based on equilibrium values of the overall SPR signal, without any reference to the dynamics of interacting partners. In contrast, a detailed kinetic model and a realistic fitting procedure [161,162], relates the evolution of interacting compounds to the “evolving” layers on the chip and is able to provide the dynamic assessment of the whole process of interaction. The transfer matrix can be used to calculate the reflectivity of the complete system. While this approach is used in chip design it can also be used, in conjunction with kinetic models of the dielectric permittivity changes, to derive the surface concentration of compounds in the multilayer system comprising a mixture of different dielectric media. The equivalent dielectric permittivities for each layer can be considered within the transfer matrix to compute the variation of the position of the reflectance minimum (the SPR angle) and relate it with the SPR data. This algorithm and the proposed model have been used to fit the experimental data, to derive the concentration thresholds and kinetic parameters for each constitutive phase (association, insertion and lipid membrane destabilization) and provide time evolutions of actual peptide to lipid ratios within each layer for the specific case of melittin lipid interaction [163]. Alternatively, In contrast to the approach where the refractive index of the media above the surface of the sensor is related to the angle corresponding to the minimum value of the SPR curve, the angle pertaining for the largest reflectivity variation within the SPR curve is selected via the corresponding photodetector in the detection array. The data from this photodetector is collected at a high rate and analyzed using the Fast Fourier Transform. Based

77

Modeling

Page

2.6

Figure 69: Left: Calibration curves for Escherichia coli detection: simultaneous EIS, SPR and SMPR detection. EIS measurements were made using a continuous active electrode and a platinum counter electrode. Right: Calibration curves for Escherichia coli detection: the SPR and SMPR same as within the left figure, but with EIS measurements on coplanar electrodes

on the known frequency of the oscillating magnetic field, from the amplitude of the oscillation of the light intensity for the chosen angle of incidence one can straightforwardly derive the reflectivity for each time point. 2.8

Combined EIS, SPR and MOSPR analysis

Page

78

Based on both novel instrumentation (enabling both MOSPR and EIS assays) and on measuring set-ups & sensing protocols advanced by ICB [164,165] successful proof of concept of combined EIS and SMPR assays to detect bacterial cells were performed by ICB. The figure below emphasizes calibration curves related to SPR, MOSPR and EIS response related to various concentrations of Escherichia coli (at 4 5 6 10 , 10 and 10 cell/ml) as well as a reference (0 cell/ml) (figure 69).

References

[1]

[2] [3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

W. L. Barnes, A. Dereux, and T. W. Ebbesen, “Surface Plasmon Subwavelength Optics”, Nature 424, 824 (2003). C. Kittel, Introduction to Solid State Physics (Wiley, New York, 2005). W. A. Murray and W. L. Barnes, “Plasmonic Materials”, Adv. Mater. 19, 3771 (2007). E. Ozbay, “Plasmonics: Merging Photonics and Electronics at Nanoscale Dimensions”, Science 311, 189 (2006). D. K. Gramotnev and S. I. Bozhevolnyi, “Plasmonics beyond the diffraction limit”, Nat. Photon. 4, 83 (2010). M. I. Stockman, “Nanoplasmonics: The physics behind the applications”, Phys. Today 64, 39 (2011). J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao, and R. P. V. Duyne, “Biosensing with plasmonic nanosensors”, Nat. Mater. 7, 442 (2008). J. Homola, “Surface Plasmon Resonance Sensors for Detection of Chemical and Biological Species”, Chem. Rev. 108, 462 (2008). R. Kolesov, B. Grotz, G. Balasubramanian, R. J. Stöhr, A. A. L. Nicolet, P. R. Hemmer, F. Jelezko, and J. Wrachtrup, “Wave–particle duality of single surface plasmon polaritons”, Nat. Phys. 5, 470 (2009). E. Altewischer, M. P. van Exter, and J. P. Woerdman, “Plasmon-assisted transmission of entangled photons,” Nature 418, 304 (2002). Y. Fedutik, V. V. Temnov, O. Schöps, U. Woggon, and M. V. Artemyev, “Exciton-Plasmon-Photon Conversion in Plasmonic Nanostructures”, Phys. Rev. Lett. 99, 136802 (2007). D. Martín-Cano, L. Martín-Moreno, F. J. García-Vidal, and E. Moreno, “Resonance Energy Transfer and Superradiance Mediated by Plasmonic Nanowaveguides”, Nano Lett. 10, 3129 (2010). I. Freestone, N. Meeks, M. Sax, and C. Higgitt, “The Lycurgus Cup — A Roman nanotechnology”, Gold Bulletin 40, 270 (2007).

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

81

[15]

References

L. Hunt, “The true story of Purple of Cassius”, Gold Bulletin 9, 134 (1976). M. Faraday, “Experimental Relations of Gold (and Other Metals) to Light”, Philos. Trans. R. Soc. London 147, 145 (1857) P. G. Kik and M. L. Brongersma, “Surface Plasmon Nanophotonics”, in Surface Plasmon Nanophotonics, M. L. Brongersma and P. G. Kik (ed.), Springer Series in Optical Sciences, vol. 131, Ch. 1 (2007). K. L. Kelly, E. Coronado, L. L. Zhao, and G. C. Schatz, “The Optical Properties of Metal Nanoparticles: The Influence of Size, Shape, and Dielectric Environment”, J. Phys. Chem. 107, 668 (2003). C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (Wiley, New York, 1998). P. Mulvaney, “Surface Plasmon Spectroscopy of Nanosized Metal Particles”, Langmuir 12, 788 (1996). J. Rodríguez-Fernández, J. PérezJuste,,F. J. García de Abajo, and L. M. Liz-Marzán, “Seeded Growth of Submicron Au Colloids with Quadrupole Plasmon Resonance Modes”, Langmuir 22, 7007 (2006). G. Schider, J. R. Krenn, A. Hohenau, H. Ditlbacher, A. Leitner, F. R. Aussenegg, W. L. Schaich, I. Puscasu, B. Monacelli, and G. Boreman, “Plasmon dispersion relation of Au and Ag nanowires”, Phys. Rev. B 68, 155427 (2003). J. R. Lakowicz, K. Ray, M. Chowdhury, H. Szmacinski, Y. Fu, J. Zhang, and K. Nowaczyk, “Plasmon-controlled fluorescence: a new paradigm in fluorescence spectroscopy”, Analyst 133, 1308 (2008). H. Raether, Surface Plasmons on Smooth and Rough Surfaces and on Gratings, Vol. 111 of Springer Tracts in Modern Physics (Springer-Verlag, Berlin, 1988). A. V. Zayats and I. I. Smolyaninov, “Near-field photonics: surface plasmon polaritons and localized surface plasmons”, J. Opt. A: Pure Appl. Opt. 5, S16 (2003). H. Ditlbacher, J. R. Krenn, N. Felidj, B. Lamprecht, G. Schider, M. Salerno, A. Leitner, and F. R. Aussenegg, “Fluorescence Imaging of Surface Plasmon Fields”, Appl. Phys. Lett. 80, 404 (2002).

Page

[14]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

Page

82

[34]

[35]

F. López-Tejeira, S. G. Rodrigo, L. Martín-Moreno, F. J. García-Vidal, E. Devaux, T. W. Ebbesen, J. R. Krenn, I. P. Radko, S. I. Bozhevolnyi, M. U. González, J. C. Weeber, and A. Dereux, “Efficient unidirectional nanoslit couplers for surface plasmons”, Nat. Phys. 3, 324 (2007). V. V. Temnov, U. Woggon, J. Dintinger, E. Devaux, and T. W. Ebbesen, “Surface plasmon interferometry: measuring group velocity of surface plasmons”, Opt. Lett. 32, 1235 (2007). B. Hecht, H. Bielefeldt, L. Novotny, Y. Inouye, and D. W. Pohl, “Local Excitation, Scattering, and Interference of Surface Plasmons”, Phys. Rev. Lett. 77, 1889 (1996). S. C. Kitson, W. L. Barnes, and J. R. Sambles, “Full Photonic Band Gap for Surface Modes in the Visible”, Phys. Rev. Lett. 77, 2670 (1996). S. I. Bozhevolnyi, J. Erland, K. Leosson, P. M. W. Skovgaard, and J. M. Hvam, “Waveguiding in Surface Plasmon Polariton Band Gap Structures”, Phys. Rev. Lett. 86, 3008 (2001). A. L. Baudrion, J. C. Weeber, A. Dereux, G. Lecamp, P. Lalanne, and S. I. Bozhevolnyi, “Influence of the filling factor on the spectral properties of plasmonic crystals”, Phys. Rev. B 74, 125406 (2006). M. U. González, J.-C. Weeber, A.-L. Baudrion, A. Dereux, A. L. Stepanov, J. R. Krenn, E. Devaux, and T. W. Ebbesen, “Design, near-field characterization, and modeling of 45° surface-plasmon Bragg mirrors” Phys. Rev. B 73, 155416 (2006). J.-C. Weeber, J. R. Krenn, A. Dereux, B. Lamprecht, Y. Lacroute, and J. P. Goudonnet, “Near-Field Observation of Surface Plasmon Polariton Propagation on Thin Metal Stripes”, Phys. Rev. B 64, 45411 (2001). S. I. Bozhevolnyi, V. S. Volkov, E. Devaux, J. Y. Laluet, and T. W. Ebbesen, “Channel plasmon subwavelength waveguide components including interferometers and ring resonators”, Nature 440, 508 (2006). E. Moreno, S. G. Rodrigo, S. I. Bozhevolnyi, L. Martín-Moreno, and F. J. García-Vidal, “Guiding and Focusing of Electromagnetic Fields with Wedge Plasmon Polaritons”, Phys. Rev. Lett. 100, 023901 (2008).

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45] [46]

[47]

A. Hohenau, J. R. Krenn, A. L. Stepanov, A. Drezet, H. Ditlbacher, B. Steinberger, A. Leitner, and F. R. Aussenegg, “Dielectric Optical Elements for Surface Plasmons”, Optics Letters 30, 893 (2005) B. Steinberger, A. Hohenau, H. Ditlbacher, A. L. Stepanov, A. Drezet, F. R. Aussenegg, A. Leitner, and J. R. Krenn, “Dielectric stripes on gold as surface plasmon waveguides”, Appl. Phys. Lett. 88, 094104 (2006). T. Holmgaard and S. I. Bozhevolnyi, “Theoretical analysis of dielectricloaded surface plasmon-polariton waveguides”, Phys. Rev. B 75, 245405 (2007) J. Grandidier, S. Massenot, G. C. des Francs, A. Bouhelier, J.-C. Weeber, L. Markey, A. Dereux, J. Renger, M. U. González, and R. Quidant, "Dielectricloaded surface plasmon polariton waveguides: Figures of merit and mode characterization by image and Fourier plane leakage microscopy", Phys. Rev. B 78, 245419 (2008). S. Randhawa, M. U. González, J. Renger, S. Enoch, and R. Quidant, "Design and properties of dielectric surface plasmon Bragg mirrors", Opt. Express 18, 14496-14510 (2010). J. J. Burke, G. I. Stegeman, and T. Tamir, "Surface-polariton-like waves guided by thin, lossy metal films", Phys. Rev. B 33, 5186-5201 (1986). R. Charbonneau, P. Berini, E. Berolo, and E. Lisicka-Shrzek, "Experimental Observation of Plasmon Polariton Waves Supported by a Thin Metal Film of Finite Width", Opt. Lett. 25, 844-846 (2000). R. Zia, M. D. Selker, P. B. Catrysse, and M. L. Brongersma, "Geometries and Materials for Subwavelength Surface Plasmon Modes", J. Opt. Soc. Am. A 21, 2442-2446 (2004). J. A. Dionne, L. A. Sweatlock, H. A. Atwater, and A. Polman, "Plasmon slot waveguides: Towards chip-scale propagation with subwavelength-scale localization", Phys. Rev. B 73, 035407 (2006). A. D. McFarland and R. P. V. Duyne, Nano Lett. 3, 1057 (2003). C. Sonnichsen, B. M. Reinhard, J. Liphardt, and A. P. Alivisatos, Nature Biotechnol. 23, 741 (2005). M. P. Kreuzer, R. Quidant, G. Badenes, and M.-P. Marco, Biosens. Bioelectron.

[50]

[51] [52] [53]

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64] [65]

[66]

[67]

[68]

[69]

[70]

[71] [72]

[73] [74]

[75]

[76]

[77]

[78]

[79] [80]

[81]

Martin, J. M. Garcia-Martin, T. Thomay, A. Leitenstorfer, and R. Bratschitsch, Nat. Photon. 4, 107 (2010). N. Yu, R. Blanchard, J. Fan, Q. J. Wang, C. Pflügl, L. Diehl, T. Edamura, S. Furuta, M. Yamanishi, H. Kan, and F. Capasso, IEEE Trans. Nanotechnol. 9, 11 (2010). L. Tang, S. E. Kocabas, S. Latif, A. K. Okyay, D.-S. Ly-Gagnon, K. C. Saraswat, and D. A. B. Miller, Nat. Photon. 2, 226 (2008). D. E. Chang, A. S. Sorensen, E. A. Demler, and M. D. Lukin, Nat. Phys. 3, 807 (2007). W. A. Challener, C. Peng, A. V. Itagi, D. Karns, W. Peng, Y. Peng, X. Yang, X. Zhu, N. J. Gokemeijer, Y.-T. Hsia, G. Ju, R. E. Rottmayer, M. A. Seigler, and E. C. Gage, Nat. Photon. 3, 220 (2009). B. C. Stipe, T. C. Strand, C. C. Poon, H. Balamane, T. D. Boone, J. A. Katine, J.-L. Li, V. Rawat, H. Nemoto, A. Hirotsune, O. Hellwig, R. Ruiz, E. Dobisz, D. S. Kercher, N. Robertson, T. R. Albrecht, and B. D. Terris, Nat. Photon. 4, 484 (2010). M. Ozaki, J. Kato, and S. Kawata, Science 332, 218 (2011). J. Valentine, S. Zhang, T. Zentgraf, E. Ulin-Avila, D. A. Genov, G. Bartal, and X. Zhang, Nature 455, 376 (2008). H. A. Atwater and A. Polman, Nat. Mater. 9, 205 (2010). M. Righini, A. S. Zelenina, C. Girard, and R. Quidant, Nat. Phys. 3, 477 (2007). T. Nikolajsen, K. Leosson, and S. I. Bozhevolnyi, Appl. Phys. Lett. 85, 5833 (2004). J. Gosciniak, S. I. Bozhevolnyi, T. B. Andersen, V. S. Volkov, J. KjelstrupHansen, L. Markey, and A. Dereux, Opt. Express 18, 1207 (2010). M. J. Dicken, L. A. Sweatlock, D. Pacifici, H. J. Lezec, K. Bhattacharya, and H. A. Atwater, Nano Lett. 8, 4048 (2008). A. V. Krasavin, K. F. MacDonald, N. I. Zheludev, and A. V. Zayats, Appl. Phys. Lett. 85, 3369 (2004). D. Pacifici, H. J. Lezec, and H. A. Atwater, Nat. Photon. 1, 402 (2007). R. A. Pala, K. T. Shimizu, N. A. Melosh, and M. L. Brongersma, Nano Lett. 8, 1506 (2008). K. F. MacDonald, Z. L. Sámson, M. I. Stockman, and N. I. Zheludev, Nat. Photon. 3, 55 (2009).

83

[49]

21, 1345 (2006). S. S. Aćimović, M. P. Kreuzer, M. U. González, and R. Quidant, ACS Nano 3, 1231 (2009). M. A. Otte, B. Sepúlveda, W. Ni, J. Pérez Juste, L. M. Liz-Marzán, and L. M. Lechuga, ACS Nano 4, 349 (2010). K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari, and M. S. Feld, J. Phys. C 14, R597 (2002). S. Nie and S. R. Emory, Science 275, 1102 (1997). S. Lal, S. E. Clare, and N. J. Halas, Acc. Chem. Res. 41, 1842 (2008). A. Boltasseva, T. Nikolajsen, K. Leosson, K. Kjaer, M. Larsen, and S. Bozhevolnyi, J. Lightwave Technol. 23, 413 (2005). T. Holmgaard, Z. Chen, S. I. Bozhevolnyi, L. Markey, and A. Dereux, Opt. Express 17, 2968 (2009). H. Ditlbacher, F. R. Aussenegg, J. R. Krenn, B. Lamprecht, G. Jakopic, and G. Leising, Appl. Phys. Lett. 89, 161101 (2006). J.-C. Weeber, K. Hassan, A. Bouhelier, G. Colas-des-Francs, J. Arocas, L. Markey, and A. Dereux, Appl. Phys. Lett. 99, 031113 (2011). A. L. Falk, F. H. L. Koppens, C. L. Yu, K. Kang, N. de Leon Snapp, A. V. Akimov, M.-H. Jo, M. D. Lukin, and H. Park, Nat Phys 5, 475 (2009). D. Koller, A. Hohenau, H. Ditlbacher, N. Galler, F. Reil, F. R. Aussenegg, A. Leitner, E. List, and J. R. Krenn, Nat. Photon. 2, 684 (2008). A. V. Akimov, A. Mukherjee, C. L. Yu, D. E. Chang, A. S. Zibrov, P. R. Hemmer, H. Park, and M. D. Lukin, Nature 450, 402 (2007). R. J. Walters, R. V. A. van Loon, I. Brunets, J. Schmitz, and A. Polman, Nat Mater 9, 21 (2010). M. A. Noginov, G. Zhu, A. M. Belgrave, R. Bakker, V. M. Shalaev, E. E. Narimanov, S. Stout, E. Herz, T. Suteewong, and U. Wiesner, Nature 460, 1110 (2009). R. F. Oulton, V. J. Sorger, T. Zentgraf, R.-M. Ma, C. Gladden, L. Dai, G. Bartal, and X. Zhang, Nature 461, 629 (2009). J. A. Dionne, K. Diest, L. A. Sweatlock, and H. A. Atwater, Nano Lett. 9, 897 (2009). W. Cai, J. S. White, and M. L. Brongersma, Nano Lett. 9, 4403 (2009). V. V. Temnov, G. Armelles, U. Woggon, D. Guzatov, A. Cebollada, A. Garcia-

Page

[48]

[82]

[83]

[84]

[85] [86] [87] [88] [89] [90] [91] [92]

[93] [94]

[95] [96]

Page

84

[97]

[98]

V. V. Temnov, K. Nelson, G. Armelles, A. Cebollada, T. Thomay, A. Leitenstorfer, and R. Bratschitsch, Opt. Express 17, 8423 (2009). D. Martin-Becerra, J. B. Gonzalez-Diaz, V. V. Temnov, A. Cebollada, G. Armelles, T. Thomay, A. Leitenstorfer, R. Bratschitsch, A. Garcia-Martin, and M. U. Gonzalez, Appl. Phys. Lett. 97, 183114 (2010). J.B.González Díaz, PhD Thesis. Universidad Autónoma de Madrid, 2010. R.P.Hunt, J.Appl.Phys 38, 1652(1967). H. Dötsch et al., J. Opt. Soc. Am. B 22, 240 (2005). R. Espinola et al. Opt. Lett. 29, 941 (2004). W. Van Parys et al., Appl. Phys. Lett. 88, 071115 (2006). H. Kato et al., J. of Appl. Phys. 93, 3906 (2003). Z. Wang and S. Fan, Opt. Lett. 30, 1989 (2005). W. Smigaj et al., Opt. Lett. 35, 568 (2010). J. C. Maxwell-Garnett, Philos. Trans. R. Soc. London, Ser. A 203, 385 (1904); 205, 237 (1906); D. A. G. Bruggemann, Ann. Phys. 24, 636 (1935); D. Stroud, Phys. Rev. B 12, 3368 (1975); J.I. Gittleman and B. Abeles, Phys. Rev. B 15, 3273 (1977); J. S. Ahn, et al., Phys. Rev. B 52, 15244 (1995). M .Inoue, et al., J. Appl. Phys. 83, 6768 (1998). P. Halevi, A.A. Krokhin, J.Arriaga, Phys. Rev. Lett. 82 (1999); M. Diwekar, et al., Appl. Phys. Lett. 84, 3112 (2004); A. García-Martín, G. Armelles and S. Pereira, Phys. Rev. B 71, 205116 (2005); Z. Wang and S. Fan, Appl. Phys. B 81, 369 (2005). R.P. Hunt, J. Appl. Phys. 38, 1652 (1967). J. E. Sipe, J. Opt. Soc. Am. B 4, 481 (1987). M. Schubert, Phys. Rev. B 53, 4265 (1996) ; M. Schubert, T. E. Tiwald, and J. A. Woollam, Appl. Opt. 38, 177 (1999); H. Wöhler, M. Fritsch, G. Haas and D. A. Mlynski, J. Opt. Soc. Am. A 5, 1554 (1988). E. Ferreiro- Vila, J. B. Gonzalez- Díaz, R. Fermento, M. U. González, A. García- Martín, J. M. García- Martín, A. Cebollada, G. Armelles, D. MenesesRodríguez, and E. Muñoz Sandoval, Phys. Rev. B 80, 125132 (2009).

[99]

[100] [101]

[102] [103]

[104] [105]

[106]

[107]

[108] [109] [110]

[111] [112]

[113]

P. R. McIsaac, Microwave Theory and Techniques, IEEE Transactions on Microwave Theory and Techniques 39, 1808 (1991). J.F. Torrado et al., to be published (2011). D. Bedeaux and P. Mazur, Physica, 67, 23 (1973); V.V. Varadan, A. Lakhtakia and V.K. Varadan, IEEE Trans. Antennas Prop. 37, 800 (1989); A. Lakhtakia, Astrophys. J. 394, 494 (1992); A.H. Sihvola, Opt. Lett. 19, 430 (1994); K. Hinsen, A. Bratz and B.U. Felderhof, Phys. Rev. B. 23, 14995 (1992); D.A. Smith and K.L. Stokes Opt. Express 14, 5746 (2006); A. Lakhtakia, Int. J. Infrared Millim. Waves 14, 2269-2275 (1993). B.T. Draine, Astrophys. J. 333, 848 (1988). S. Albaladejo, R. Gómez-Medina, L. S. Froufe-Pérez, H. Marinchio, R. Carminati, J. F. Torrado, G. Armelles, A. García-Martín and J.J. Sáenz . Optics Express, 18, 3556-3567 (2010). H. Marinchio, R. Carminati and J.J. Sáenz, to be published (2011). R. Vincent, J.J. Sáenz and R. Carminati, “Local and directional excitation of surface surface waves in magneto-plasmonics”, in preparation (2011). C. Vandenbem, L. S. Froufe-Pérez, R. Carminati, J. Opt. A 11, 114007 (2009); C. Vandenbem, D. Brayer, L. S. FroufePérez, R. Carminati, Phys. Rev. B 85, 085444 (2010). L. Novotny, B. Hecht, Principles of Nano-optics, Cambridge University Press, 2006. V. Klimov, S.K. Sekatskii and G. Dietler, J. Mod. Opt. 51, 1919 (2004). R. Vincent and R. Carminati, Phys. Rev. B 83, 165426 (2011). G. Colas des Francs, C. Girard and O.J.F. Martin, Phys. Rev. A, 67, 053805 (2003). E.M. Purcell and C.R. Pennypacker, Astrophys. J. 186, 705-714 (1973). B. T. Draine, “The discrete dipole approximation for light scattering by irregular targets,” in Light Scattering by Nonspherical Particles: Theory, Measurements, and Applications, M. I. Mishchenko, J. W. Hovenier, and L. D. Travis, eds. (Academic, 2000), pp. 131145. R. Schmehl, B. M. Nebeker, and E. D. Hirleman, J. Opt. Soc. Am. A 14,

[116] [117] [118]

[119] [120]

[121] [122] [123]

[124]

[125]

[126]

[127]

[128]

[129]

85

[115]

[130] V. I. Safarov, V. A. Kosobukin, C. Hermann, G. Lampel, J. Peretti, and C. Marlière, Phys. Rev. Lett. 73, 3584 (1994). [131] J.B. Gonzalez-Diaz, A. García- Martín, G. Armelles, J. M. García-Martín, C. Clavero, A. Cebollada, R. A. Lucaszew, J. R. Skuza, D. P. Kumah, R. Clarke, Phys. Rev. B. 2007, 76, 153402. [132] E. Ferreiro-Vila, X. M. Bendana Sueiro, J. B. Gonzalez-Diaz,A. García-Martin, J. M. García-Martin, A. Cebollada Navarro, G. Armelles Reig, D. Meneses Rodriguez, and E.Munoz Sandoval, IEEE Trans. Magn. 44, 3303 (2008). [133] E. Ferreiro- Vila, J. B. Gonzalez- Díaz, R. Fermento, M. U. González, A. García- Martín, J. M. García- Martín, A. Cebollada, G. Armelles, D. MenesesRodríguez, and E. Muñoz Sandoval, Phys. Rev. B 80, 125132 (2009). [134] E. Ferreiro-Vila, M. Iglesias, E. Paz, F. J. Palomares, F. Cebollada, J. M. González, G. Armelles, J. M. GarcíaMartín and A. Cebollada, Phys. Rev. B 83, 205120 (2011). [135] V.V.Temnov, G.Armelles, U.Woggon, D.Guzatov, A.Cebollada, A.Garcia-Martin, J.M. Garcia-Martin, T.Thomay, A.Leitenstorfer and R.Bratschitsch, Nature Photonics 4 (2010) 107. [136] D. Martín-Becerra, J. B. González-Díaz, V. V. Temnov, A. Cebollada, G. Armelles, T. Thomay, A. Leitenstorfer, R. Bratschitsch, A. García-Martín, and M. U. González, Appl. Phys. Lett. 97, 183114 (2010). [137] B. Sepulveda, A. Calle, L.M.Lechuga, G. Armelles, Optics Letters 2006, 8, 1085 [138] D. Regatos, D. Fariña, A. Calle, A. Cebollada, B. Sepúlveda, G. Armelles, and L. M. Lechuga, J. Appl. Phys. 108, 054502 (2010). [139] M.G.Manera, G. Montagna, E.FerreiroVila, L. González-García, J. R. Sánchez-Valencia, A. R. GonzálezElipe, A. Cebollada, J.M. Garcia-Martin, A. Garcia-Martin, G. Armelles, and R. Rella, J. Mater. Chem. 2011 (to appear). [140] D. Meneses-Rodríguez, E. FerreiroVila, P. Prieto, J. V. Anguita, M. U. González, J. M. García-Martín, A. Cebollada, A. García-Martín, and G. Armelles, Small (2011 to appear) [141] J. F. Torrado, J. B. González-Díaz, M. U. González, A. García-Martín, and G.

Page

[114]

30263036 (1997); M. Paulus and O. J. F. Martin, Phys. Rev. E 63, 066615 (2001); P.C. Chaumet and M. NietoVesperinas, Phys. Rev. B 61, 14 119 (2000); 62, 11 185 (2000); 64, 035422 (2001). A. Rahmani, P.C. Chaumet, and F. de Fornel, Phys. Rev. A 63, 023819 (2001). B. T. Draine and P. Flatau. J. Opt. Soc. Am. A 25, 2693-2703 (2008); P. C. Chaumet, A. Rahmani, and G. W. Bryant, , Phys. Rev. B 67, 165404 (2003). A. Lakhtakia, Opt. Commun. 79, 1-5 (1990). D.M. Whittaker and I.S. Culshaw, Phys. Rev. B 60, 2610 (1999). A. Garcia-Martin, G. Armelles and S. Pereira, Phys. Rev. B 71, 205116 (2005). B. Caballero, A. Garcia-Martin, J.C. Cuevas, to be published. R.W. Wood, Proc. R. Soc. London A 18, 269 (1902); Phys. Rev. 15, 928 (1935). U. Fano, J. Opt. Soc. Am. 31, 213 (1941). Lord Rayleigh, Proc. Roy. Soc. (London) A79, 399 (1907). T.W. Ebbesen, H.J. Lezec, H.F. Ghaemi, T. Thio and P.A. Wolf, Nature 391, 667 (1998). F. J. García de Abajo, R. GómezMedina and J. J. Sáenz, Phys. Rev. E, 72, 016608, (2005); F. J. García de Abajo and J. J. Sáenz, Phys. Rev. Lett. 95, 233901 (2005); F. J. García de Abajo, J. J. Sáenz, I Campillo and J.S. Dolado, Opt. Express 14, 7 (2006). R. Gómez-Medina, M. Laroche and J.J. S áenz, Opt. Express 14, 3730 (2006); M. Laroche, S. Albaladejo, R. GómezMedina and J.J. Sáenz, Phys. Rev. B 74, 245422 (2006). M. Laroche, S. Albaladejo, R. Carminati and J.J. Sáenz, Opt. Lett. 32, 2762 (2007). S. Soria, T. Katchalski, E. Teitelbaum, A. A. Friesem, and G. Marowsky, Opt. Lett. 29, 1989 (2004). L. Y. Wang, J. Luo, M. M. Maye, Q. Fan, Q. Rendeng, M. H. Engelhard,. C. M. Wang, Y. H. Lin and C. J. Zhong.,J. Mater. Chem. 15, 1821 (2005). A. Pradhan, R.C. Jones, D. Caruntu, C.J. O'Connor, and M.A. Tarr, Ultrason. Sonochem. 15, 891 (2008).

[142]

[143]

[144]

[145]

[146]

[147]

[148] [149]

[150]

[151]

[152]

[153]

[154]

Page

86

[155]

[156]

[157]

Armelles, Opt. Express 18, 15635 (2010). B. Sepulveda, J. B. Gonzalez-Diaz, A. Garcia-Martin, L. M. Lechuga, and G. Armelles, Phys. Rev. Lett. 104, 147401 (2010). J. B. Gonzalez-Diaz, J. M. GarciaMartin, A. Garcia-Martin, D. Navas, A. Asenjo, M. Vazquez, M. HernandezVelez, and G. Armelles, Appl. Phys. Lett. 94, 263101 (2009). E. Th. Papaioannou, V. Kapaklis, P. Patoka, A. Garcia-Martin, E. FerreiroVila, G. Ctistis, Phys. Rev. B 81, 054424 (2010). J. F. Torrado, E. Th. Papaioannou, G. Ctistis, P. Patoka, M. Giersig, G. Armelles, and A. Garcia-Martin, Physica status solidi (RRL) 4, 271 (2010). C. de Julián Fernández, G. Mattei, E. Paz, R. L. Novak, L. Cavigli, L. Bogani, F. J. Palomares, P. Mazzoldi, A. Caneschi, Nanotechnology 21 (2010) 165701-165709. A. Lessufleur et al., J. of Magnetism and Magnetic Materials 305, 284 (2006). J. Homola, Chem. Rev., 2008, 108, 462. M. G. Manera, J. Spadavecchia, A. Taurino, R. Rella, Journal of Optics A, 2010, 12, 035003. S. Y. Wu, H. P. Ho, W. C. Law, C. Lin, and S. K. Kong, Opt. Lett. 2004, 29, 2378. F. Romanato, K. H. Lee, H. K. Kang, G. Ruffato, and C. C.Wong, Opt. Express 2009, 17, 12145 . W. L. Barnes, W. A. Murray, J. Dintinger, E. Devaux, T. W. Ebbesen, Phys. Rev. Lett. 2004, 92, 107401. B. Sepulveda, A. Calle, L.M.Lechuga, G. Armelles, Optics Letters 2006, 8, 1085. L. González-García, G. Lozano, A. Barranco, H. Míguez and A. R. González-Elipe, J. Mater. Chem., 2010, 20, 6408. D.B. Chrisey, A. Piqué, R.A. McGill, J.S. Horwitz, B.R. Ringelsen, D.M. Bubb, P.K. Wu, Chem. Rev. 2003, 103, 553. R. Buonsanti, V. Grillo, E. Carlino, C. Giannini, T. Kipp, R. Cingolani, P. D. Cozzoli, J. Am. Chem. Soc. 2008, 130, 11223–11233. A.P. Caricato, R. Buonsanti, M. Catalano, M . Cesaria, P.D. Cozzoli,

[158]

[159]

[160]

[161]

[162]

[163]

[164]

[165]

A. Luches, M.G. Manera, M. Martino, A. Taurino, R. Rella, Applied Physics A, 2011, 104, 963. M. G. Manera, G. Montagna, E. Ferreiro-Vila, L. González-García, J. R. Sánchez-Valencia, A. R. GonzálezElipe, A. Cebollada, J. M. GarciaMartin, A. Garcia-Martin, G. Armelles and R. Rella, Journal of Materials Chemistry 21 , 16049 (2011) Born, M and Wolf, E (1980) Principles of optics: electromagnetic theory of propagation, interference and diffraction of light. 6th ed. Pergamon Press. 808 p. Reitz, JR, Milford, FJ and Christy, RW (1993) Foundations of electromagnetic theory. 4th ed., Addison-Wesley 630p. Gheorghiu, M, et al (2009) Sensing based on assessment of nonmonotonous effect determined by target analyte: case study on pore-forming compounds. Biosens Bioelectron 24(12): 3517-3523. Olaru, A, et al (2009) Assessment of the multiphase interaction between a membrane disrupting peptide and a lipid membrane. J Phys Chem B 113(43): 14369-14380. Gheorghiu, M, et al (2009) Sensing based on assessment of nonmonotonous effect determined by target analyte: case study on pore-forming compounds. Biosens Bioelectron 24(12): 3517-3523. Gheorghiu E. RO Patent Application A00136/2011: “Method to assess the amount of target analytes by controlled periodic actuation” Gheorghiu E., David S., Polonschii C., Bratu D., RO Patent Application A00135/2011: “Device to assess the amount of target micro-organisms by controlled periodic actuation”.

Annex

Annex

In short the S&T objectives can be regarded as four, in which we will consider both bottomup and top-down approaches to obtain the desired magneto-plasmonic materials: a)

b)

c)

Development of nanomaterials that combine plasmons and magnetic properties (films, nanoparticles, core-shell structures). Investigate the correlation between the optical, magnetic, magneto-optical and magneto-plasmonic properties. Carry out theoretical calculations of the optical response considering the magneto-optical contribution. Provide proof of concept for applications based in the magneto-plasmonic activity, and testing for specific applications in the field of chemical sensors and biosensors. Identification of applications for microelectronics and information technology.

The purpose of this European funded project is the study, development and application of a novel concept of nanostructured materials formed by the combination of components with plasmonic and magneto-optic (MO) activity. This smart combination will produce “magneto-plasmonic� nanomaterials tailored on the nanoscale.

d)

The project has two main goals; the first is to prepare active magneto-plasmonic materials with tailored properties in the nanoscale and understanding the interactions of the magnetic properties with the plasmonic and optical ones, linked to electric charge oscillations. The second goal is to provide proof of concept for applications that can benefit of this coupling. Since it is expected that the optical properties of these materials can be driven by using a magnetic field, this will allow designing and developing novel magneto-plasmonic devices. In particular, as a proof of the applicability of this concept, we will design, fabricate and test a new kind of surface plasmon resonance (SPR) sensor with MO elements, i.e. a surface magneto-plasmon resonance (SMPR) sensor, comparing its performance against standard sensors.

About Us nanoICT

NANOstructured active MAGnetoplasmonic MAterials

EC contribution

2,96 M Euros

Contract number

FP7-214107-2

NÂş. of partners

10

Coordinator

IMM / CSIC (Spain) / Antonio Garcia-Martin

Start date

November 01, 2008

Duration

36 months

Website

www.phantomsnet.net/Nanomagma

Consortium

Page

Depending on the materials properties several application routes will be proposed in either opto-electronic, spin-tronic, spin-photonic domains and for each electromagnetic simulation and integration analysis will be performed. Preliminary manufacturing flows will be proposed based on 200-to-300mm silicon standards for microelectronic (CMOS) and microsystem (System On Chip SOC) uses.

89

The project also includes prospective tasks for silicon-oriented uses. This part includes the identification of relevant applications of magneto-plasmonic materials for microelectronics and information technology.

Page

94

NANOMAGMA Partners

Institution

Country

Contact Person

Consorzio Interuniversitario Nazionale per la Scienza e Technologia dei Materiali

Italy

Caneschi, Andrea andrea.caneschi@unifi.it

Centre National de la Recherche Scientifique - Ecole Supérieure de Physique et de Chimie Industrielles – ParisTech

France

Carminati, Rémi remi.carminati@espci.fr

Biotecgen SRL

Italy

Chiesa, Maurizio m.chiesa@biotecgen.it

Phantoms Foundation

Spain

Correia, Antonio antonio@phantomsnet.net

Instituto de Microelectronica de Madrid CSIC

Spain

Garcia-Martin, Antonio antonio@imm.cnm.csic.es

International Centre of Biodynamics

Romania

Gheorghiu, Eugen egheorghiu@biodyn.ro

Hamburg University

Germany

Nielsch, Kornelius knielsch@physnet.uni-hamburg.de

CEA

France

Olivier, Segolene segolene.olivier@cea.fr

CNR - IMM Institute for Microelectronic and Microsystems

Italy

Rella, Roberto roberto.rella@le.imm.cnr.it

Universidad Autónoma de Madrid

Spain

Sáenz Gutiérrez, Juan José juanjo.saenz@uam.es

95

Page

&ƵŶĚĞĚ ďLJ

ĚŝƚĞĚ ďLJ

WŚĂŶƚŽŵƐ &ŽƵŶĚĂƚŝŽŶ ůĨŽŶƐŽ 'ŽŵĞnj ϭϳ ϮϴϬϯϳ DĂĚƌŝĚ Ͳ ^ƉĂŝŶ ŝŶĨŽΛƉŚĂŶƚŽŵƐŶĞƚ͘ŶĞƚ ǁǁǁ͘ƉŚĂŶƚŽŵƐŶĞƚ͘ŶĞƚ

Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.