EP1875179A2

EP1875179A2 - Ultra-sensitive optical detector with high time resolution, using a surface plasmon

Info

Publication number: EP1875179A2
Application number: EP06743825A
Authority: EP
Inventors: Laurent Frey; Jean-Claude Villegier
Original assignee: Commissariat a lEnergie Atomique CEA
Current assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2005-04-25
Filing date: 2006-04-21
Publication date: 2008-01-09
Also published as: WO2006114540A2; US20080272302A1; FR2884916B1; US7928386B2; FR2884916A1; WO2006114540A3

Abstract

The invention relates to an ultra-sensitive optical detector with high time resolution, which uses a surface plasmon. The inventive optical detector, which is intended to detect at least one photon, comprises a dielectric substrate (SB) having at least one biometric detection element (39a) disposed thereon in order to generate an electric signal from the energy of the photon(s) received. In addition, at least one coupling element is formed on the substrate, which is different from the detection element and which comprises a metallic element (G) that is provided to generate a surface plasmon by interaction with the photon(s) and to guide said plasmon to the detection element which then absorbs the energy of said surface plasmon.

Description

ULTRA-SENSITIVE ULTRASENSITIVE DETECTOR WITH HIGH TIME RESOLUTION USING SURFACE PLASMON

DESCRIPTION

TECHNICAL AREA

The present invention relates to an ultrasensitive optical detector, with high temporal resolution, and in particular a superconducting single photon detector.

The invention applies in particular to the detection and localization of malfunctions of a semiconductor integrated circuit, by detecting the emission of specific photons by failed transistors.

It also applies to reception and routing circuits for very high speed telecommunications, in particular those intended for telecommunication satellites, because of the very low heat dissipation of the detector object of the invention. The invention is also applicable to the detection of the key of the quantum code in a cryptographic system.

It also applies to the detection of biological or medical objects by tomography as well as the astronomical observation of distant objects. STATE OF THE PRIOR ART

Emerging techniques such as quantum computation and cryptography, the detection of very weak signals for telecommunications or astronomy, the non-contact diagnosis of integrated circuits in microelectronics or the detection of biological objects, require radiation detectors. especially in the visible domain or the near infra-red domain, which are very fast, have a very weak jitter, make very little noise and are extremely sensitive: they must be able to detect a single photon .

In this regard, reference is made to the document [Verevkin 03] which, like the other documents cited later, is mentioned at the end of the present description.

Single photon superconducting detectors, also known as SSPDs, potentially have all the above mentioned qualities and are good candidates for replacing current avalanche photodiodes and photomultiplier tubes, whose performance is limited, particularly in the infrared.

The different types of SSPDs can be classified into two categories, namely the superconducting tunnel junctions (STJs) and the bolometers, the HEBs, the hot electron bolometers (HEBs). fastest class. HEBs use ultra-thin superconducting films whose thickness is less than lOnm, to obtain very short characteristic times, of the order of 10 ps. For these films, the preferred material is niobium nitride (NbN) in cubic phase B1. In a known HEB, a very thin film of

NbN is epitaxially grown on a sapphire substrate whose orientation is 1102 (plane R); and after structuring, this film forms meanders whose width ranges from 100 nm to 200 nm. The manufacture, operation and characterization of such SSPDs are described for example in [Gol'tsman 03] and [Korneev 03]. Detection efficiency, or the conversion efficiency of the photon into an electrical signal, is a key parameter that must be optimized and that would be worth 100% for an ideal detector.

It depends on the filling rate (coverage rate of the incident beam and the active area of the detector), the optical absorption in the NbN layer and the hot spot capacity, which is formed as a result of the absorption of the photon, to create a resistive barrier over the entire width of the strip (in English, strip) superconducting. We distinguish the detection efficiency of quantum efficiency, which is defined for a recovery rate of 100%.

The meander structure has been designed to have a correct fill rate. In the latter embodiments, the interval between two strips of this structure is as small as the width of the superconducting tape and the fill rate is about 50%. It can hardly exceed this value of makes current technological limitations associated with the electronic lithography that is used to form the meander structure.

Better recovery rates can be envisaged with other techniques, but a structure whose meanders are too close together favors crosstalk and leads to an increased length of superconducting material. Absorption by NbN is of the order of

20% for a thickness of 4 nm. The only way to increase it would be to increase the thickness of the NbN layer, but the ratio between the hot spot section and that of the superconducting band would decrease, since the width of this band is limited to 100 nm by the beam of electrons, and this would cause a collapse in overall efficiency (see [Korneev 03]).

The detection efficiency can not exceed about 15% with this architecture. It reaches 5% to 7% with the wavelengths used for telecommunications in the latest generation of detectors (see [Korneev 04]).

It would probably be difficult to do better because the very long length of the NbN band

(150 μm to 200 μm when this strip is "unwound") has two disadvantages, namely the non-uniformity of the width of the line and the classical or quantum fluctuations of the superconducting fluid, which leave little hope of approaching the line. maximum efficiency. Several approaches have been proposed to increase the coupling of light in the absorbent region of the detector.

For example, it has been proposed to use a mirror and a possible anti-reflection layer. The light which has passed through the NbN without having been absorbed is then returned to the NbN by a concave mirror (see [FR 2812455]) or by a plane mirror (see [LeCoupanec 03]). The NbN film is optionally coated with an anti-reflection layer to eliminate the reflection losses of the incident wave.

The expected detection efficiency is 40% but it is in any case limited by the fill rate (50%) that is inherent to any film design that is attacked at normal incidence.

It has also been proposed to use waveguide coupling (see [JPL 03]). The incident light then propagates in a silicon plane waveguide. The serpentine deposited over absorbs the evanescent wave associated with the guided mode and the photon is likely to be absorbed at each passage of the evanescent wave by a line of the meander. The thickness of NbN traversed at each passage (100 nm to 200 nm) is much greater than in the case of a normal incidence where this thickness is 4 nm. But as most of the intensity passes through the guide, many meanders are required for absorption to be complete.

The efficiency can in principle reach very high values but it is, here too, limited by the very great length of the superconducting band. Moreover, from a point of view technology, the feasibility of monocrystalline NbN growth on Si has not yet been clearly demonstrated.

I / INVENTION STATEMENT

The meandering structure, optical coupling device currently used in HEB in NbN, is not optimized because of the low filling rate, the weak absorption by the NbN layer, and the long length meanders which increases the noise and degrades the electrical characteristics due to meander width non-uniformities.

As a result, the detection efficiency is limited to 20% in the visible range. With the wavelengths used for telecommunications, it is even lower (less than 10%) and the noise is greater because the energy of the photon is lower. A larger pixel, better suited to longer wavelengths, would require even longer meanders.

The design of the detector (for example SSPD, more precisely HEB), which is proposed in the present invention, abandons the meandering structure with its limitations. Its originality lies in the optical addressing of the photon towards a detector, in particular a superconducting detector, with optimized geometry, via surface plasmons which are guided in a metal guide. The zones of optical coupling and detection by the HEB are now separated. The present invention has various advantages.

In particular, the optical coupling is significantly more efficient than in the meander structure. Indeed, the filling rate is close to 100%.

In addition, the energy is fully absorbed in the active element of the detector.

In addition, the length of the detection band, which constitutes the thermometer of the HEB or the STJ junction, can be at least ten times smaller than that of the meander structure. Problems related to the non-uniformity of superconducting bandwidth and classical or quantum electronic fluctuations largely disappear.

The detection efficiency can therefore reach very high values, the noise is lower and the jitter (in English, jitter) is reduced. Indeed, there is less phase noise and better temperature uniformity in a shorter superconducting band.

In addition, the invention can be implemented by existing technological means.

Specifically, the subject of the present invention is an optical detector, intended to detect at least one photon (from one to ten photons), this detector comprising a dielectric substrate and, on this substrate, at least one bolometric detection element. which is intended to generate an electrical signal from the energy of the photon or photons received, this detector being characterized in that further comprises at least one coupling element which is formed on the substrate, is distinct from the bolometric detection element and comprises a metallic element, designed to generate a surface plasmon by interaction with the photon (s) and to guide this plasmon surface to the bolometric detection element which then absorbs the energy of this surface plasmon.

It is specified that the bolometric detection element is a bolometric element that is at thermal equilibrium or out of thermal equilibrium (as for hot electron bolometers).

According to a particular embodiment of the invention, the bolometric detection element forms a thin and short strip.

According to a first particular embodiment of the detector which is the subject of the invention, the bolometric detection element comprises a monolayer of a material which is chosen from bolometric materials and superconducting materials.

According to a second particular embodiment, the bolometric detection element comprises a superconducting multilayer forming a tunnel junction and / or a Josephson junction.

The tunnel junction may be a magnetic tunnel junction.

According to a first particular embodiment of the invention, the bolometric detection element and the metal element are substantially rectilinear and the metal element extends. longitudinally by the bolometric detection element.

Preferably, in this case, the metallic element has a first width and the bolometric detection element has a second width which is smaller than the first width, and the metal element is connected to the bolometric detection element by a end.

According to a second particular embodiment of the invention, the bolometric detection element and the metal element are substantially rectilinear and the bolometric detection element is perpendicular to the metal element.

Preferably, the generation of the surface plasmon is obtained by coupling in an Otto or Kretschmann type configuration or by coupling by means of a diffraction grating, which is formed on the metallic element, or by means of a dielectric waveguide which is formed on or formed between the metallic element and the substrate and has a refractive index greater than that of the substrate.

The detector object of the invention may further comprise an optical fiber for transporting the photon or photons for coupling. The end of the optical fiber can be cleaved at a cleavage angle that allows generation of the coupling surface plasmon in the Otto configuration.

Alternatively, the end of the optical fiber is cleaved at a cleavage angle which allows the generation of surface plasmon by coupling into the Kretschmann type configuration and the refractive index of the substrate is less than that of the core of the optical fiber.

According to another variant, the end of the optical fiber is cleaved at a cleavage angle which allows generation of the surface plasmon by coupling in the Kretschmann type configuration and the refractive index of the substrate is greater than that of the core of the optical fiber, and a dielectric layer, whose refractive index is greater than that of the substrate, is provided on the metal element, between the latter and the optical fiber.

According to a preferred embodiment of the invention, the metal element of the coupling element constitutes a guide for coupling a focused incident light beam, formed by the photon or photons, towards a long-distance surface plasmon. (English, "long range surface plasmon"), this coupling being done by the edge of the guide, at a first end thereof.

The guide can then be a tapered guide (in English, "type") whose second end is narrower than the first end.

In this case, the bolometric detection element may be substantially rectilinear, placed facing the second end of the guide and perpendicular to this second end; or at least a part of the bolometric detection element may be substantially rectilinear, this part being placed under the second end of the guide, parallel to this second end. The guide can also be a straight guide to long-distance surface plasmon, placed between two plasmon mirrors, namely two reflecting mirrors

(For example in gold or silver), arranged at the entrance and exit of the guide and for example deposited on the facets of the chip (in English, chip) of the detector. The guide provided with the two mirrors thus constitutes a plasmon resonator. The bolometric detection element is then substantially rectilinear and placed at least partially under the guide, for example parallel or perpendicular to this guide.

According to a particular embodiment of the invention, the detector comprises a set of coupling elements and a set of bolometric detection elements, the set of coupling elements forming a matrix of elements on the substrate and being intended coupling a collimated incident light beam, formed by the photon or photons, which coupling takes place by attenuated total reflection, in the Otto or Kretschmann configuration, and wherein each of the detection elements constitutes a bridge which connects two adjacent coupling elements .

Preferably, each coupling element constitutes a resonator in a direction parallel to the plane of incidence of the light beam on the substrate.

Then, preferably, each coupling element also constitutes a resonator in a direction perpendicular to the plane of incidence of the light beam on the substrate. According to a preferred embodiment of the detector which is the subject of the invention, each bolometric detection element comprises a superconducting nitride phase. This phase may be niobium nitride.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be better understood on reading the description of exemplary embodiments given below, for purely indicative and non-limiting purposes, with reference to the appended drawings in which:

FIG. 1A is a schematic longitudinal sectional view of an example of the detector which is the subject of the invention,

FIG. 1B is a schematic top view of the detector object of FIG.

FIG. 1C is a schematic view of control and processing means associated with the detector of FIGS. 1A and 1B,

FIG ID is a schematic top view of another example of the detector object of the invention,

FIG. 1E is a schematic top view of a variant of FIG.

FIG. 2 is a schematic longitudinal sectional view of another example of the detector which is the subject of the invention,

FIG. 3 is a schematic longitudinal sectional view of another example of the detector which is the subject of the invention, FIG. 4A is a schematic longitudinal sectional view of another example of the detector which is the subject of the invention,

FIG. 4B is a schematic top view of the detector of FIG. 4A,

FIG. 5A is a schematic longitudinal sectional view of another example of the detector which is the subject of the invention,

FIG. 5B is a schematic top view of the detector of FIG. 5A,

FIG. 6A is a schematic perspective view of a preferred embodiment of the detector which is the subject of the invention,

FIG. 6B is a schematic and partial top view of the detector of FIG. 6A,

FIG. 7 is a schematic and partial plan view of a variant of this detector of FIG. 6A,

FIG. 8A is a schematic perspective view of another embodiment of FIG.

The invention,

FIG. 8B is a schematic and partial sectional view of the detector of FIG. 8A,

FIG. 9A is a schematic longitudinal sectional view of another example of the detector which is the subject of the invention,

FIG. 9B is a schematic top view of the detector of FIG. 6A,

FIG. 10A is a schematic top view of another detector according to FIG.

The invention, FIG. 10B is a schematic and partial sectional view of the detector of FIG. 10A,

FIG. 10C is a schematic view of two adjacent coupling elements of this detector of FIG. 10A, which are connected by a superconducting material bridge, and

- Figures HA to HE schematically illustrate steps of a method of manufacturing an element made of superconductive material, used to manufacture a detector according to the invention.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

Various detectors according to the invention are described below, as well as the operation of these detectors, in the case of a focused incident light beam and then of a collimated incident light beam.

It is first of all in the context of a focused incident beam, whose width is typically a few micrometers and which is derived from a monomode optical fiber or a focusing optics such as a microscope objective.

A first example of a detector according to the invention is seen in longitudinal section in FIG. 1A and from above in FIG. 1B.

This detector 2 comprises a dielectric substrate 4 and a bolometric detection element 6 formed on this substrate and made of a superconducting material, preferably NbN. This element forms a thin strip, short and substantially straight. Typically, the thickness E of this band is less than 10 nm, its width 1 is less than 1 μm and its length L is less than 100 μm but advantageously greater than the propagation length δ _sp of plasmon at the surface of the superconducting material , the NbN in the example, relative to the air.

In Example E is 4 nm, 1 is 0.2 μm and L is 10 μm. The detector 2 also comprises a coupling element 8 which is formed on the substrate. This element 8 is distinct from the bolometric detection element 6 but in contact with it. The coupling element 8 comprises a substantially straight metal element 10 which is extended longitudinally by the bolometric detection element 6.

The coupling element 8 is designed to receive one or a few photons that it is desired to detect, for example a single photon, and to communicate the energy of this photon to the bolometric detection element 6. In the example, the photon is fed to the metal element 10 by an optical fiber 11, more precisely by the heart (in English, core) (not shown) of this fiber. The bolometric sensing element is provided to generate an electrical signal from the photon energy and to permit fast and undistorted electrical propagation of the detected signal.

Note that the optical fiber 11 is disposed obliquely in the example of Figures IA and IB. In this example, the energy of the photon is communicated to the bolometric detection element in the form of a surface plasmon 12: the metal element 10 generates this plasmon by interaction with the photon and guides this plasmon to the detection element 6.

Two electrodes (not shown) are formed respectively at both ends of the formed structure of the bolometric detection element 6 and the coupling element 8. Each electrode consists of one or more layers of a suitable metal, for example example gold. These electrodes allow the connection of the detection element 6 to control and processing means 15 which can be seen in FIG. 1C, on which these electrodes respectively have the references 14 and 16.

These electrodes have the references 3 and 4 in Figure ID which will be discussed later and which is a schematic top view of a detector according to the invention.

In FIG. 1C, a cryostat 18 is shown to hold the detector 2 at a temperature much lower than the critical temperature of NbN. The control and processing means 15 are also seen. These means 15 comprise a bias tee 20, which is connected to the electrodes 14 and 16 and which makes it possible to separate the fast detection signals, the frequency of which is of the order of 1 GHz, DC power supply of the detector, by means of propagation lines whose structure is coplanar and whose impedance is adapted (see figure

ID).

These means 15 also include a source of bias current which is connected to the tee and provided for biasing the detector 2 by a direct current whose intensity is less than the critical intensity of the detector. Specific means (not shown) for signal amplification or multiplexing may be provided in the cryostat. Figure ID is a schematic top view of a particular embodiment of the detector object of the invention.

This detector further comprises the substrate 4 and, on the latter, two electrically conductive structures 26 and 27. The structure 27 forms a substantially rectilinear strip. It is extended by the coupling element 8 which still includes the element

10 followed by zone 28.

This zone 28 is followed by the linear bolometric detection element 6, the ends of which have the reference 25.

The structure 26 extends on either side of the assembly comprising the band 27, the element 8 and the element 6. In addition, as can be seen, this structure 26 is connected to the end of the element 6, opposite to that which is in contact with the zone 28.

The structure 26 and the band 4 thus delimit two zones 29 of the substrate which are substantially rectilinear and constitute coplanar lines of propagation of the detected electrical signal. The structure 26 is grounded (in English "grounded").

The band 4 allows the polarization of the current detector and the output of the detected signal to the polarization tee 20 (Figure IC).

Element 6 constitutes a bridge in NbN. As a purely indicative and in no way limiting, this NbN bridge has a length of 5 μm, a width of

0.2 μm and a thickness of 4 nm; and the width of the lines 29 is 30 μm.

Figure IE is a variant of Figure ID, which will be discussed later.

Consider again the example of Figures IA and IB. The energy conveyed by the incident wave coming from the optical fiber 11, whose core has a diameter that is typically a few microns, is confined in two directions of space before being guided to the surface of the metallic element. 10.

Indeed, there is a vertical confinement: the incident light excites a plasmon mode of the interface between the dielectric 32 and the element 10, the latter constituting a metal plasmon guide, and the excited plasmon 12 propagates to the interface between the metal and the dielectric which overcomes this metal, namely the air in the example.

There is also a lateral confinement: the surface plasmon is then guided towards a focusing zone 28 where it undergoes a lateral confinement.

In the example, the metal element 10 has a length of the order of 10 microns and a width of the order of 5 microns. Several configurations are possible for the coupling of the plasmon.

A first possible configuration (FIG. 1A) is an Otto type coupling, the first variant of ATR (Attenuated Total Reflection) technique.

The end 30 of the fiber 11 is cleaved at an angle which allows the guided mode in the fiber to undergo a total reflection on the end 30 and to generate, in a layer 32 of low refractive index, an evanescent wave which comes exciting the plasmon mode at the upper interface of the element 10 metal.

This layer 32 may be an air layer or a low index dielectric layer between the cleaved end 30 and the element 10.

This low index layer must have a refractive index lower than the effective index of the guided mode of the fiber, to ensure total reflection; its thickness is typically 1 μm. The cleavage angle of the fiber must be greater than the critical angle i _c of total reflection between the fiber and the dielectric 32. The value of this cleavage angle is determined by the plasmon wave vector, which depends on the dielectric constant of the metal of element 10 (for example silver), and the refractive index of the low index layer.

With regard to the choice of the metal of the element 10, the propagation length of the plasmon, which depends on the dielectric constant of the metal of the element 10 and the index of the dielectric layer

32, should be sufficient, typically a few tens of microns, to limit losses. This condition is fulfilled by gold and silver, common metals in plasmon applications. These weakly resistive metals are also suitable for producing the electrical structures connected to the detector.

The thickness of the element, or film, metal 10 is typically a few tens of nanometers. It is specified that the thicknesses and indices are adjusted to optimize the coupling, in particular so that the angular width of the resonance covers the angular dispersion of the mode of the fiber.

It is also possible to interpose one or more dielectric layers, having appropriate optical indices, between the fiber and the element 10, to allow good optical coupling.

The range of wavelengths that can be envisaged for the detector 2 corresponds to the domain of existence of surface plasmons in the metals considered. For silver and gold, the wavelength should be approximately equal to or greater than 350 nm, and for NbN it should be approximately equal to or greater than 600 nm. It should be noted that the propagation length of plasmons is an increasing function of the wavelength.

It is specified that the zone delimited by an ellipse 34 in FIG. 1B diagrammatically represents the intersection of the incident light beam and the plane of the element 10. In addition, the length of the zone 28 must be less than the propagation length of the plasma. Alternatively, a Kretschmann type coupling (another variant of the ATR technique) can be used as shown in FIG.

In this case, the cleaved end 30 of the fiber 11 is directly in contact with the metal of the element 10. There is therefore a total reflection of the light that propagates in the fiber. The evanescent wave excites the plasmon at the lower metal / substrate interface, which is possible provided that the substrate index is lower than that of the core of the fiber.

This condition is not fulfilled with a sapphire substrate, whose index is 1.75, and a standard fiber whose core has an index equal to 1.45. But we can then consider interposing a high index material (index greater than that of the substrate) between the metal element 10 and the cleaved face of the fiber.

In another variant, which is schematically illustrated in FIG. 3, a coupling is used by means of a diffraction grating (in English, diffraction grating).

A plasmon is excited from the fiber, which is cleaved perpendicularly to its axis in this case, with a normal or oblique incidence on the metal film 10. The incidence is oblique in the example of FIG. diffraction 36 is provided on the surface of the element, or film, metal 10 and the end of the fiber is placed facing this network, at an angle less than or equal to 90 ° with the latter (see [Ditlbacher 03 ] about a network coupling). The characteristics of the network, namely its pitch and its amplitude are chosen so as to optimize the coupling.

Other couplings can be envisaged, according to which the optical fiber is coupled to the detector by the wafer of the latter.

In a variant, waveguide coupling is used (see FIGS. 4A-4B and 5A-5B).

The light from the fiber 11, which is cleaved perpendicular to its axis, is coupled in an amorphous dielectric waveguide 38, which is formed on the metal element 10. The evanescent wave propagating at the lower interface of the waveguide 38 couples to the plasmon mode at the lower metal interface, provided that the effective index of the guided mode is greater than the index of the substrate.

This is schematically illustrated in FIG. 4A, in which the detector of the example considered is seen in longitudinal section, and in FIG. 4B, on which it is seen from above. It can be seen that the waveguide 38 also extends above the bolometric detection element 6 and that its width is substantially equal to that of the metal element 10. Alternatively, the metal element 10 can be deposited. and the NbN over the dielectric waveguide 38, which index must then be greater than that of the substrate 4. This is schematically shown in longitudinal section in Figure 5A and in plan view in Figure 5B. For example, it is possible to form the guide by implantation in an Al ₂ O ₃ or MgO substrate. It is also possible to use an edge waveguide (in English, ridge waveguide) in LiNbO ₃ , which is formed on the substrate. It is known that the epitaxy of LiNbO ₃ films on Al ₂ O ₃ and on MgO is possible.

With regard to the guidance and focusing of plasmon, propagation over several tens of micrometers, with excellent lateral confinement, as well as focusing by a triangular termination were obtained on waveguides of a few micrometers in width. silver (see [Weeber 01]), and gold (see [Krenn 02]).

Another example of wafer coupling is schematically illustrated in FIG. 6A, which corresponds to a preferred embodiment in the case of a focused incident light beam. Indeed, the angular divergence of this beam is important, which makes the coupling techniques previously mentioned (with oblique incidence) limited. More precisely, the diffraction-limited beam (the most favorable case) is typically a few micrometers wide and this beam is derived from a single-mode optical fiber or focusing optics, for example a microscope objective. The angular divergence is important (11 ° for the total angle at 1.55 μm), which makes the traditional techniques of coupling by ATR, in Otto or Kretschmann configuration, or network coupling, poorly adapted to the coupling of light to a surface plasmon, whose angular resonance is fine and typically ranges from 0.1 ° to 1 °. To solve this problem, one solution is to couple the light L (Figure 6A) of the optical fiber (not shown) by the wafer (end-fire coupling type) of a metal guide G to the large mode. propagation distance or LRSP mode (for "long range plasmon area"), which exists in particular for a very thin guide (whose thickness is from 5 to 10 nm) in a symmetrical structure, including the substrate SB and the superstrate SP have the same index (see [Nikolajsen 04]). The polarization of the light beam to be detected must be of TM type, that is to say perpendicular to the guide G.

This mode has the particularity of having a large spatial extension and a field profile which is compatible with those of a monomode optical fiber, for a guide section e x 1 typically equal to

(5 nm to 10 nm) x (5 μm to 10 μm). This metal guide G plays the role of a "rail" which guides the light without confining it strongly. The coupling losses mentioned in [Nikolajsen 04] are limited to 10%, and they could be reduced by filling the space between the fiber and the guide by a medium having an optical index close to that of the substrate, to limit retro-reflection.

The metal guide G is made of a noble metal, such as silver or gold, metals which have a low extinction coefficient, thus low losses by dissipation, which allows large propagation distances. It should be noted that the distance between the fiber and the guide must be less than the Rayleigh distance, equal to 50 μm for a single-mode fiber at 1.55 μm. The LRSP mode is then confined in a sub-micron section s of the tapered guide (in English, "taper") that constitutes the metal guide G, using a low angle taper, which performs an adiabatic transition (thus loss-free or almost) from the LRSP to a localized surface plasmon ([Stockman 04]).

The tapered guide G is relatively long (several hundred micrometers). At its output end, the energy density is very important because of the very strong confinement of the field at least in the plane of the metal layer constituting the guide, and can be up to 1000 times greater than the density existing at the guide entry ([Stockman 04]). This intensification of the field is of interest for efficiently transferring the energy to the absorbing material (for example NbN) of the bolometric detection element 39a in the form of a straight strip, whose el x 11 section is also very small and is typically 4 nm) x (100 to 200 nm). In FIGS. 6A and 6B, we also see the contacts 39b which are provided at the ends of the strip 39a.

If the exit end of the tapered guide is wider than the thickness of this guide, the bias remains TM type throughout the guide. One can thus excite a localized plasmon mode of the NbN band 39a which is placed near the outlet end of the guide G, perpendicular to this guide ([Weeber 99]). The length of this band 39a, forming a nanowire, is sufficiently large to absorb the plasmon (in NbN, one can choose a length of ten micrometers at a wavelength of 1.55 microns), or is fixed at a multiple of half the wavelength of the plasmon ([Schider 03]), so as to constitute a resonant and at the same time absorbing antenna.

Such a structure can be produced by successively depositing on the substrate, the guide, a very thin layer of a dielectric material and then the NbN band. Other techniques can be envisaged, for example based on hybridization.

The band NbN 39a may also be positioned under the outlet end 40 of the guide G, parallel to this guide (Figure 7). The absorbent material (NbN in the example) absorbs the evanescent wave which circulates near the guide. The fact of using a tapered guide makes it possible to increase the field and thus to reduce the necessary length of NbN.

Note that in the example of Figure 7, the band 39a is in two parts 39c and 39d, which are perpendicular to each other, and only the portion 39d is parallel to the guide G.

Instead of a tapered guide, another solution is to use a straight LRSP guide placed between two mirrors as a resonator (Figures 8A and 8B). The two mirrors (not shown) are two reflecting mirrors, for example gold or in silver, arranged at the entrance or exit guide and for example deposited on the facets of the detector chip. The length of this straight guide 41 is a multiple of half the wavelength of the LRSP, itself very close to the wavelength in the vacuum of the detected light. The straight strip 39a of NbN is placed at least partially under the guide 41, for example perpendicularly to this guide. The LRSP undergoes multiple round trips into the cavity and is partially absorbed into the NbN at each pass.

The use of an LRSP mode makes it possible to avoid dissipation losses in the metal. This solution requires precise dimensional control for the structure to resonate at a given wavelength.

By way of example, various numerical values relating to a detector according to the invention are given below:

thickness of the coupling element 8: typically 10 nm, possibly a few nanometers as the superconducting strip 6;

- width of the coupling element: a few micrometers, width well adapted to the dimensions of an optical fiber; - length of the coupling element: much smaller than the plasmon propagation distance;

- Output section of the focusing point 28: substantially identical to that of the superconducting band 6 (a few tenths of micrometers by a few nanometers); length of the focussing tip 28: also less than the propagation distance of the plasmon; the shape of this focusing point (triangular or other shape) is optimized to minimize losses.

The following gives details of the guidance, absorption and detection in the superconducting band 6.

This superconducting band is a plasmon guide with absorption losses. Several configurations are possible:

superconducting nano-guide aligned with the guide, or metal element 8 (see [Dickson 00] for gold or silver nano-guides): this configuration corresponds to the examples of FIGS. 1A-1B, 2, 3, 4A -4B and 5A-5B; the light propagates in the NbN band in a guided plasmon mode, at the NbN / air or NbN / substrate interface, depending on the coupling used; the length of this nano-guide must be greater than the characteristic propagation length of the plasmon at the interface between NbN and air or a low-index dielectric, which length is of the order of 5 μm to 10 μm;

- nano-guide in NbN perpendicular to the guide, or element, metallic: the band of NbN then behaves like a nano-wire which guides the light and which is attacked by a polarization parallel to its axis [Weeber 99], according to a process localized plasmon type; the length of the nano-guide is then greater than the propagation length; an example of detector corresponding to such a configuration is diagrammatically illustrated by FIG. 9A (sectional view longitudinal) and in Figure 9B (seen from above); the NbN 6 band is shown, the metal element 42 provided with a focusing zone 44, by which the element 42 is connected to the band 6, and the substrate 4 on which the detector is formed.

The detection of the energy carried by the plasmon in the superconductor can be done:

- Or according to the principle of a hot electron bolometer threshold, the bolometer is then polarized by a subcritical current (see [Korneev 04]); this case corresponds to the example of figure IC;

either by a superconducting tunnel junction according to the principle of SIS (superconductor-insulator-superconductor) receivers, by suppressing the Josephson current by means of a magnetic field; this case is schematically illustrated by the example of Figure IE which is a schematic and partial top view of a variant of Figure ID; the superconductor may for example be NbN and the MgO insulator or a tantalum nitride whose thickness is typically 1 nm; in the example of Figure IE, the bridge 6 is thus replaced by a Josephson tunnel junction 6a which is for example of NbN-MgO-NbN type; the layer I of MgO is then between the two layers II and III of NbN; one of these two NbN layers is electrically connected to the zone 28 and the other layer III of NbN is connected to the structure 26;

- either by detecting the Josephson current of a superconducting junction or a SQUID (for

Superconducting Quantum Interference Device). Other materials than superconductors can be envisaged for the detection: one can use a bolometric material, for example a manganite or a cuprate, or a semiconductor material whose gap (band gap) is adapted to the length of photon wave to be detected.

In addition, the bolometric detection element may consist of a single layer of the material in question (for example in the case of bolometers) or consist of a multilayer (in the case of a superconducting tunnel junction and / or a Josephson junction or a magnetic tunnel junction).

In what follows, the case of a collimated incident light beam is considered.

With a collimated beam whose width is at least 100 μm, the angular divergence is reduced to less than 1 ° and makes it possible to envisage efficient coupling of the light to a surface plasmon, by a traditional technique of attenuated total reflection coupling. (ATR), in Otto or Kretschmann configuration, for flat metal layers. The beam is then in oblique incidence above the structure. However, with a beam whose width is typically 100 microns, it is not possible to efficiently couple to a LRSP mode whose angular resonance is extremely fine. Light can only be effectively coupled to modes that are confined to a single metal interface, ie a superior interface or lower interface (Otto / Kretschmann); however these modes attenuate on characteristic distances that are less than 100 microns. Under these conditions, the energy can not be transferred to the absorbent element. It can then be envisaged, as shown in FIGS. 10A and 10B, to fragment the zone of noble metal (for example Au or Ag), covered by the beam to be detected 45, into multiple coupling elements 46 forming a matrix on the substrate. 47. In addition, two contiguous elements 45 are connected by a filamentary bridge 48 in NbN which rests on two NbN / metal junctions, each having an area of the order of 1 μm ² , and which absorbs the energy of the plasmon. .

FIG. 10C is an enlarged view of one of the bridges 48 in NbN, which connects two elements or cells 46. The two corresponding NbN / metal junctions have the reference 50 for this figure.

Since the energy transfer from a metal cell 46 to an associated NbN filamentary bridge is not optimized because of the absence of a tapered guide, it is necessary to choose the length of the cell, in the direction of the incident beam ( or, more precisely, parallel to the plane of incidence of the beam on the substrate carrying the matrix, this plane being indicated by the line I), for this cell to be a resonator, and thus to give this length a value equal to a multiple of λ / 2, where λ represents the wavelength of the beam, while giving this length a value well below the propagation distance of the plasmon in the metal of the cell. With each passage of the plasmon sp at the end of the cell, the energy is transferred to the NbN in the form of localized plasmon which attenuates very rapidly, after a few micrometers, or is simply absorbed by the NbN which constitutes a material absorbent.

The width of the cell is also a multiple of λ / 2 so that the cell also resonates in the direction perpendicular to the previous one (more precisely perpendicular to the plane of incidence of the beam), in order to collect the energy efficiently to take account of the angular width of the beam also in this direction, even if this width is reduced. From an electrical point of view, all the cells 46 of the same column of the matrix can be connected in series and be traversed by the same bias current. Such a configuration is used in the example of FIG. 10A, where two contacts 51 are also seen which are respectively formed at the two ends of the circuit thus obtained. However, other electrical configurations are possible. The cumulative length of NbN is very short (a few micrometers) because the space between two cells is very small, less than 1 μm.

The following is an indication of a method of manufacturing a single photon monolithic detector including the layers and structures that allow the photogeneration of a surface plasmon, the propagation of the latter and the detection of this plasmon by a HEB element in NbN.

The steps of the method for manufacturing such a superconducting bolometric detector HEB in NbN are given below.

We can also refer to the document [Romestain 2004].

1. Preparation of the support, or substrate, for depositing the very thin layer of the superconductor (NbN) therein and for producing the structure of the detector

A monocrystalline corundum or sapphire substrate is preferably used, the crystalline orientation of which is preferably plane-R

(1102). The face on which the deposition of the bolometer layers will be performed has an epitaxial quality polish, adapted to the epitaxy of silicon. It is also possible to use an SOS substrate whose diameter is typically around 75 mm. This diameter can be up to 300 mm. Different surfaces and shapes of the substrate are also possible.

The surface of the substrate is cleaned, typically by immersion of the latter in an acetone bath, then in an alcohol bath, and then by exposure to ultrasound. In a variant, a high temperature surface treatment of the substrate is carried out, this temperature typically being greater than 1000 ^° C. in order to reconstruct the crystalline quality of its surface.

Other monocrystalline substrates may be used, for example monocrystalline oxide films whose crystalline mesh does not exhibit a too much parameter disagreement with that of the NbN layer whose mesh parameter a is about 0.44 nm. It is possible, for example, to use monocrystalline oxides of cubic phase oxides of the NaCl or B1 type, such as MgO (at about 0.42 nm), or LaAlO ₃ or ZrO ₂ -Y ₂ O ₃ substrates.

It is also possible to use monocrystalline fluoride substrates such as CaF ₂ for example. It is also possible to use blades made of semiconductors such as silicon or germanium, or SOI structures. However, the use of semiconductor substrates such as silicon requires a combination of specific surface treatments ex situ and in situ prior to performing the next step.

2. Deposition of the very thin active layer of the detector, made of niobium nitride, by sputtering

The superconducting layer of NbN, having a few nanometers in thickness and reproducible superconducting properties (critical temperature of the order of 1OK, or greater than 10 K, for a thickness of about 3.5 nm), and constituting the element detector, possibly obtained by continuous magnetron sputtering of a high purity niobium target (> 99.99%), in a mixture of argon and nitrogen, when the substrate (of the type sapphire-plane-R or other according to the criteria defined in paragraph 1) is uniformly heated to a temperature of about 600 ⁰ C during deposition. For the conditions of the deposit as well as the procedure of nitriding of the target of niobium, one will refer to [Villégier 01].

A step of degassing the spray chamber provides the vacuum quality required for deposition. It should be noted that the target material is preferably niobium but that other variants described in the literature allow to stabilize other thin superconducting layers for example niobium, TiN, MoN or YBaCuO.

A surface passivating layer (in this case a very thin layer of amorphous aluminum nitride) is deposited on the nitride layer so that the electrical, physico-chemical, mechanical and optical properties of the very thin layer of nitride niobium are not subsequently altered by the formation of a thin layer of oxide (hydroxide or / and carboxide) native or / and thermal. In this regard, reference is made to the document [Romestain 2004].

3. Lithography of the nano-bridge HEB-NbN (size about 10μmxl00nmx4nm) by electronic insolation A positive or negative resin, which is sensitive to electrons, is used. The thickness of this resin can range from 50nm to 500nm but depends on the nature of the resin. It is possible to use a resin which disappears after revelation in the areas exposed by the electrons. In this case, the shape to be cut in the resin and this shape is then cut in the very thin layer of NbN by means of a RIE etching, that is to say a reactive ion etching (in English reactive ion etching), using a gas type SF6. The remaining resin layer is then removed by the same technique, using an oxygen plasma (delamination).

Alternatively, an electronic lithography in reverse polarity may be used. The NbN bridge and the contact pads are then described by an electronic insolation which hardens the resin.

Taking the example of Figures ID, the NbN bridge corresponds to the reference 6 in this figure. The contact pads correspond to reference 2 in this figure and serve to connect the ground structure 26, the signal propagation structure 27 and the optical coupling structure 8.

According to another variant, optical lithography is used with deep ultraviolet radiation, that is to say a radiation whose wavelength is less than 0.3 microns. Such lithography also makes it possible to obtain the desired bridge dimension

According to another variant, an NbN nanobridge is formed by anodizing the NbN under the tip of an atomic force microscope.

An exemplary method is schematically illustrated by Figures HA to HE.

Firstly, a layer 60 of NbN is formed on a substrate 62 in sapphire plane-R. To do this, we performs an epitaxial deposition of NbN by continuous cathode sputtering at 600 ^° C. (FIG.

A positive electronic resin layer 64 is then deposited on the NbN layer 62 (FIG. HB). This layer 64 is then etched by an electron beam to define the desired shape for the NbN element used in the detector.

Then, a fluorinated reactive ion etching of NbN (FIG. HC) is carried out. Then the remaining resin 66 is removed by means of an oxygen plasma (Figure HD). This gives the desired element in NbN 68 on the substrate 62.

Gold contacts 70 and 72 (FIG. HE) are then deposited at the two ends of this element by a peeling technique (in English, lift off), by immersion in a solvent, for example acetone.

4. Formation of the structure of surface plasmon generation, DC polarization and coplanar propagation of the detected signal by means of deposition (s) of layer (s) of gold or suitable alloys, through a mask peel (in English, lift-off).

This step consists of bringing one or more layers of structured gold onto the NbN layer, in a suitable way. These layers will be used for the absorption of the photon converted to plasmon, its propagation and the electronic implementation of the detector. Successive deposits of layers of suitable thickness make it possible to improve the coupling and propagation of plasmon. In addition, the critical elements of the HEB detector are checked in NbN. The main parameters of the layers and structures resulting from the process used are estimated by means of physicochemical characterization and measurement tools, of the kind used in microelectronics, in particular to make measurements of nanometric thicknesses. using grazing X-rays, or X-ray diffraction.

A detector according to the invention is not limited to the detection of a single photon: such a detector can detect up to ten photons. In addition, in the present invention, higher NbN thicknesses than those given in the examples can be used. In addition, other superconductors than NbN are usable, for example other superconducting nitride phases or high temperature superconducting cuprate phases. Examples which may be mentioned are MoN, TaN, TiN, VN, MgB ₂ , Yba ₂ Cu ₃ O ₇ -x, ReBa ₂ Cu ₃ O ₇ -X, TiNNb, MoGe, MoRe, NbSi and HgBaCaCuO. More generally, as already mentioned, it is possible to use other materials than superconductors. It is indeed possible to use any material capable of detecting a low light energy and of restoring an electrical signal corresponding to this energy, for example a bolometric material or a semiconductor material. This material may be in the form of a monolayer or be used in a stack constituting a tunnel junction and / or Josephson, this stack being for example of the NbN / MgO / NbN or NbN / MgO / NbN or manganite / barrier layer / manganite, or ferromagnetic metal / barrier layer / ferromagnetic metal.

The documents cited in this description are as follows:

[FR 2812455] FR 2812455 A. Invention of R. Sobolewski et al. (Schlumberger and Rochester University, 2/2002)

[Dickson 00] R. M. Dickson et al., "Unidirectional plasmon propagation in metallic nanowires", J. Phys. Chem. B 104, 6095 (2000)

[Ditlbacher 03] H. Ditlbacher et al., "Efficiency of local light-plasmon coupling", Appl. Phys. Lett. 83, 3665 (2003)

[Gol'tsman 03] Gol'tsman et al., "Manufacturing of Nanostructured Superconducting Single Photon Detectors", IEEE Transactions on Applied Superconductivity, 13 (2), 192, June 2003

[JPL 03] D. Jackson, J. Stem, "High bandwidth, improved quantum efficiency detector for multi-GHz class OKD throughput", Jet Propulsion Laboratory, California Institute of Technology, Single Photon Detector Workshop, NIST Gaithersburg, 4/2003

[Korneev 03] A. Korneev et al., "GHz counting rate NbN single-photon detector for IR diagnoses of VLSI CMOS circuits", Microelectronics Engineering 69, 274 (2003) [Korneev 04] A. Korneev et al., "Sensitivity and gigahertz counting performance of NbN superconducting single-photon detectors", Appi. Phys. Lett. 84, 5338 (2004)

[Krenn 02] J. R. Krenn et al., "Non-diffraction Iimited light transport by gold nanowires", Europhysics Letters, 60, 663 (2002)

[LeCoupanec 03] LeCoupanec P., W. K. Lo, K.

R. Wilsher, "An ultra-low dark count and jitter, superconducting, single-photon detector for emission timing analysis of integrated circuits," Microelectronics Reliability, 43 (2003), 1621

[S. J. McNab 03] S. J. McNab et al., "Ultra-low loss photonic integrated circuit with photonic crystal-type waveguides", Optics Express 11 (22), 2927 (2003)

[Nikolajsen 04] T. Nikolajsen et al., "Polymer-based surface-plasmon-polariton stripe waveguides at telecommunications wavelengths", Appl. Phys. Lett. 82, 668 (2004)

[Romestain 2004] R. Romestain et al., "Fabrication of superconducting niobium nitride hot electron bolometer for single photon counting", New Journal of Physics, 6, 129, 2004 [Schider 03] G. Schider et al., "Plasmon Dispersal Relationship of Au and Ag Nanowires," Phys. Rev. B 68; 155427 (2003)

[Stockman 04] M. I. Stockman, "Nanofusing of optical energy in tapered plasmonicidegguides", Phys. Rev. Lett. 93, 137404 (2004)

[Verevkin 03] A. Verevkin et al., "GHz-Rate Superconducting Photon Counting Detector," Single Photon Detector Workshop, NIST Gaithersburg, 4/2003

[Villégier 01] J. C. Villégier et al., IEEE Transactions on Applied Superconductivity 11 (2001), 68

[Weeber 99] J. C. Weeber et al., "Plasmon polaritons of metallic nanowires for controlling submicron propagation of light", Phys. Rev. B 60 (12), 9061

[Weeber 01] J. C. Weeber et al., "Near-field observation of plasmon polariton propagation on thin metal stripes", Phys. Rev. B 64, 045411 (2001)

Claims

An optical detector for detecting at least one photon, said detector comprising a dielectric substrate (4, 47, 52, 62, SP) and, on said substrate, at least one bolometric detection element (6, 6a, 39a, 48, 50, 68) which is provided for generating an electrical signal from the energy of the received photon or photons, said detector being characterized in that it further comprises at least one coupling element (8, 42) which is formed on the substrate, is distinct from the bolometric detection element and comprises a metallic element (10, 41, 42, 46, G), intended to generate a surface plasmon by interaction with the photon (s) and to guide this Plasmon surface to the bolometric detection element (6, 6a, 39a, 48, 50, 68) which then absorbs the energy of this surface plasmon.

2. Detector according to claim 1, wherein the bolometric detection element forms a thin and short strip.

The detector of any of claims 1 and 2, wherein the bolometric detection element comprises a monolayer of a material which is selected from bolometric materials and superconducting materials.

4. Detector according to any one of claims 1 and 2, wherein the element of bolometric detection comprises a superconducting multilayer forming a tunnel junction and / or a Josephson junction (6a).

5. Detector according to claim 4, wherein the tunnel junction is a magnetic tunnel junction.

6. Detector according to any one of claims 1 to 5, wherein the bolometric detection element and the metal element (10) are substantially rectilinear and the metal element extends longitudinally by the bolometric detection element (6). , 6a).

The detector according to claim 6, wherein the metallic element (10) has a first width and the bolometric detection element (6, 6a) has a second width which is smaller than the first width, and the metal element connects to the bolometric detection element at one end (28).

8. Detector according to any one of claims 1 to 5, wherein the bolometric detection element and the metal element (42) are substantially rectilinear and the bolometric detection element (6) is perpendicular to the metal element .

The detector according to any one of claims 1 to 8, wherein the generation of the surface plasmon is obtained by coupling in an Otto or Kretschmann type configuration or by coupling by means of a diffraction grating (36). which is formed on the metal member (10), or by means of a dielectric waveguide (38) which is formed on or formed between the metal member (10) and the substrate (4). ) and has a refractive index greater than that of the substrate.

10. Detector according to any one of claims 1 to 9, further comprising an optical fiber (11) for transporting the photon or photons for coupling.

The detector of claim 10, wherein the end (30) of the optical fiber is cleaved at a cleavage angle that allows generation of the coupling surface plasmon in the Otto configuration.

A detector according to claim 10, wherein the end (30) of the optical fiber is cleaved at a cleavage angle which allows generation of the coupling surface plasmon in the Kretschmann-type configuration and the refractive index of the substrate (4) is smaller than that of the core of the optical fiber (11).

A detector according to claim 10, wherein the end (30) of the optical fiber is cleaved at a cleavage angle which allows the coupling of surface plasmon generation in the Kretschmann type configuration and the refractive index of the substrate (4) is greater than that of the core of the optical fiber (11), and a dielectric layer, whose refractive index is greater than that of the substrate, is provided on the metal element (10), between the latter and the optical fiber.

14. A detector according to claim 1, wherein the metal element (41, G) of the coupling element constitutes a guide for coupling a focused incident light beam, formed by the photon or photons, to a plasmon of surface at a great distance, this coupling being done by the edge of the guide, at a first end thereof.

15. The detector of claim 14, wherein the guide is a tapered guide (G) whose second end (s) is narrower than the first end.

16. Detector according to claim 15, wherein the bolometric detection element (39a) is substantially rectilinear, placed opposite the second end of the guide (G) and perpendicular to this second end.

17. Detector according to claim 15, wherein at least a portion (39d) of the bolometric detection element (39a) is substantially rectilinear and this portion is placed under the second end of the guide (G), parallel to this second end. .

18. A detector according to claim 14, wherein said guide is a long-distance surface plasmon right guide (41), placed between two plasmon mirrors, and the bolometric detection element (39a) is substantially rectilinear and placed at less partially under this guide.

The detector of claim 1, comprising a set of coupling elements (46) and a set of bolometric detection elements (48), the set of coupling elements forming a matrix of elements on the substrate (47). ) and being intended for the coupling of a collimated incident light beam, formed by the photon or photons, this coupling taking place by attenuated total reflection, in Otto or Kretschmann configuration, and in which each of the detection elements constitutes a bridge which connects two adjacent coupling elements.

20. The detector of claim 19, wherein each coupling element (46) constitutes a resonator in a direction parallel to the plane of incidence of the light beam on the substrate.

21. The detector of claim 20, wherein each coupling element (46) further constitutes a resonator in a direction perpendicular to the plane of incidence of the light beam on the substrate.

22. A detector according to any one of claims 1 to 21, wherein each bolometric detection element (6, 6a, 39a, 48, 50, 68) comprises a superconducting nitride phase.

23. Detector according to claim 22, wherein this phase is niobium nitride.