RU2502102C2 - Optical filter - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: optical filter has channel optical waveguides for input/output of optical radiation and propagation of a light beam, light beam splitters arranged in series on the radiation path and a forming means for transmitting optical radiation branched by the beam splitters. The forming means is in form of a set of connecting channel optical waveguides and/or a planar optical waveguide. The beam splitters are in form of a set of connected optical waveguides, the mutual arrangement of which is based on maintaining, at operating radiation wavelength, a phase difference which is essentially a multiple of 2π for most beams branched by different beam splitters and transmitted from the input to the output of the optical filter.

EFFECT: providing a wide adjustment range and a narrow filtration line width.

23 cl, 10 dwg

Description

The invention relates to integrated optics and more specifically relates to an optical tunable filter. It can be used as a tunable filter for frequency compression of signals in fiber-optic communication systems, a small-sized tunable optical spectrometer, or a filter element in the equipment for reading data from fiber Bragg sensors.

A device is known - an integrated multi-reflective tunable filter (AV Tsarev, "Tunable optical filters", United States Patent No. 6,999,639, February 14, 2006, Published on February 14, 2006, Foreign Application Priority Data Sep 06, 2001; A.V. Tsarev, “Multiplexers for WDM with nanophotonic reflectors - a new way to control many hundreds of optical spectral channels”, Nano- and Microsystem Technology, No. 4, pp. 51-55 (2007)), containing channel optical waveguides for input-output optical radiation and propagation of a light beam, light beam dividers arranged sequentially along radiation in the form of inclined elementary reflectors; a set of connecting channel optical waveguides for transmitting optical radiation reflected from opposite elementary reflectors. In this device, the filtering of a given wavelength of the optical spectrum is carried out due to the constructive interference of many optical beams reflected from the inclined elementary reflectors, periodically located along the channel optical waveguides. Tuning of the filtered wavelength of light is carried out due to a controlled change in the refractive index in phase-shifting optical elements located along waveguides containing inclined reflectors (for fine tuning) and a set of connecting waveguides (for wide tuning). Their task is to form a constant phase shift between any two optical beams reflected from neighboring elementary reflectors. The level of the side lobes in the passband of the filter can be reduced to a level greater than -20 dB due to apodization, by changing the beam division coefficient when reflected from elementary reflectors.

A similar type of filter is known (using many inclined reflectors), which operates on the basis of the interference of many beams propagating in a planar optical waveguide, which provide the possibility of wide wavelength tuning due to the acousto-optical (AO) effect (A.V. Tsarev. "Acousto-optic tunable filter ", patent of the Russian Federation No. 2182347, May 10, 2002, published in Bul. No. 13, 05/10/2002, AVTsarev" Acousto-optical variable filter ", United States Patent No. 7092139, Published on August 15, 2006, Foreign Application Priority Data August 04, 2000). In this device, light beam dividers in the form of periodically arranged oblique elementary reflectors that intersect the core of the channel waveguide are made in such a way that the reflected beams propagate further along the planar waveguide, and then again arrive at a similar channel waveguide with many inclined reflectors, which is called filtering an element. In this device, the filtering of a given wavelength of the optical spectrum is also carried out due to the constructive interference of many optical beams coming from a planar waveguide to inclined elementary reflectors, and which direct optical radiation along the axis of the channel optical waveguide of the filter element. Moreover, for each wavelength there is an optimal angle of the incident beam, for which constructive interference works and effective filtering is performed. This allows the wavelength to be tuned not only by changing the refractive index of the corresponding channel waveguides (as in the optical filter described above), but also by rotating the phase front during acousto-optical interaction. In this case, the optical beams that propagate along a planar waveguide interact with a surface acoustic wave (SAW) excited by an on-board transducer (IDT). The diffracted optical beam deviates from the incident beam by a double Bragg angle and enters the filter element. In this device, this angle is determined by the length of the acoustic wave, which is controlled by the frequency of the high-frequency signal (hundreds of megahertz) attached to the interdigital transducer.

The advantage of both of these types of optical filters and multiplexers lies in the fundamental possibility of narrow-band filtering and wide-range wavelength tuning. The disadvantage of these designs is the difficulty of manufacturing inclined reflectors that combine a low reflection coefficient (0.01-0.0001) and a low level of spurious scattering. Therefore, these optical elements are still not implemented as experimental devices. These devices are closest to the claimed and therefore taken as a prototype.

Also known is a tunable optical filter based on ring resonators (Magdalena S. Nawrocka, Tao Liu, Xuan Wang, and Roberto R. Panepucci, “Tunable silicon microring resonator with wide free spectral range”, Appl. Phys. Lett. 89, 071110 (2006 ) that can be grouped together to improve filtering properties (K. Yamada, T. Shoji, T. Tsuchizawa, T. Watanabe, J. Takahashi, and S. Itabashi, “Silicon-wire-based ultrasmall lattice filters with wide freespectral ranges , ”Opt. Lett. 28, 1663-1664 (2003)). Optical coupling of a ring resonator with channel optical waveguides carrying out input / output of optical radiation is carried out due to the tunnel coupling of these waves Dow with a ring resonator waveguide These filters can be manufactured on various waveguide structures, for example, based on silicon nitride, polymers, lithium niobate or silicon-on-insulator (SOI) structures.The most compact filters are implemented on SOI waveguides, because .high contrast of the refractive index (silicon oxide) allows you to get a small radius of curvature of the channel waveguides (up to several microns). In addition, it provides a wider (up to 40 nm) free spectral range (FSR-free spectral range), which determines the operating range of the filter. The wavelength tuning of such ring resonators is carried out due to phase-shifting optical elements operating on the basis of the thermo-optical effect, the electro-optical effect, or a change in the concentration of free charge carriers in the waveguide region. The disadvantage of these filters is the low wavelength tuning range, which is proportional to the magnitude of the change in the refractive index, limited by the physical properties of the waveguide material. To expand the tuning range, sometimes use the Vernier effect, i.e. filtering is carried out by a consistent change in the refractive index simultaneously for two ring resonators having a narrower and different free spectral band (J. Floriot, F. Lemarchand, and M. Lequime. Tunable double-cavity solid-spaced bandpass filter, Opt. Express, 2004 , v. 12, p. 6289-6298). The disadvantage of this type of filter is the difficulty of controlling the wavelength, as well as the presence of spurious signals at wavelengths that are multiples of the free spectral zone of each filter.

An optical filter is also known (K. Yamada, T. Shoji, T. Tsuchizawa, T. Watanabe, J. Takahashi, and S. Itabashi, “Silicon-wire-based ultrasmall lattice filters with wide freespectral ranges,” Opt. Lett. 28 , 1663-1664 (2003)), based on the use of the Mach-Zehnder (MZ) line of interferometers and tunnel coupling of channel waveguides. The filter is manufactured in SOI structures using the complementary metal-oxide-semiconductor technology (CMOS) technology and has a constant path-length difference of the different arms of the interferometer (for filtering ) and a different amount of tunnel communication (to ensure apodization necessary to reduce the level of side lobes). Unfortunately, such filters are not intended for broadband wavelength tuning, therefore, they are preferred for use as fixed filtering devices with a small number of frequency channels (i.e., with a small set of operating wavelengths).

The technical result of the invention is the creation of an optical tunable filter, which would simultaneously have a wide tuning range and a narrow width of the filtration line, and which could be made on the basis of existing and promising technologies.

The technical result is achieved in that in an optical tunable filter containing channel optical waveguides for input-output of optical radiation and propagation of the light beam, light beam dividers arranged in series along the radiation, forming means for transmitting optical radiation branched by beam dividers, wherein the forming means is made in the form of a set of connecting channel optical waveguides and / or a planar optical waveguide, according to the invention, d The beam dividers are made in the form of a set of coupled optical waveguides, the mutual arrangement of which is selected taking into account the maintenance at the working wavelength of the radiation of a phase difference essentially multiple of 2π for most beams branched out using various beam dividers and passed from the input to the output of the optical filter. Here π = 3.14159 ... is a universal constant. The requirement that most of the beams has a phase difference shift that is essentially a multiple of 2π leads to the fact that at the working wavelength all these beams will fold in phase and form an intense signal at the output of the device. That relatively small number of beams, where this condition is violated, will not make a significant contribution to the signal intensity at the working wavelength. However, their contribution can be useful for the formation of the desired shape of the spectral characteristic (for example, to suppress spurious signals outside the width of the filter line).

The device is based on channel and / or planar waveguides. The channel optical waveguide has an increased value of the refractive index both in depth and across the structure, i.e. it is a local region on / or below the surface of a solid in the form of a thin strip with a width of fractions to units of microns, with a refractive index higher than the refractive index of its surroundings. Moreover, the region with an increased value of the refractive index can be both homogeneous and inhomogeneous (the case of a gradient optical waveguide). A channel optical waveguide can support low-loss propagation of a narrow and non-diverging optical beam along its axis in the vicinity of a region with an increased refractive index. The number of guided (waveguide) waves (modes) that this structure supports, and the spatial distribution of their optical fields are determined by the profile of the change in the refractive index in depth and width. For the correct operation of this device, it is desirable to use only single-mode waveguides, i.e. those in which only one fundamental mode is propagated for the polarization used in the operation of the device. Most often it is TE polarization, i.e. when the electric field of the optical wave lies in the plane of the substrate. Sometimes TM polarization is also used, in which the magnetic field vector lies in the plane of the substrate. A planar waveguide is a thin layer with a thickness of fractions to several microns with a refractive index higher than the refractive index of the surrounding medium (substrate and surrounding upper layer, in this case, air). In a planar optical waveguide, a light beam can propagate inside this layer with very low losses (less than 1 dB / cm). Planar waveguides can be either homogeneous or gradient.

Channel and planar waveguides can be made by diffusion of metals, proton exchange from molten salts, sputtering of materials with a higher refractive index than that of the substrate, epitaxy from the gas or liquid phase, modification of the properties of the surface layer due to irradiation, for example, by electrons and / or photons, etc. A channel waveguide can be made by etching grooves on the surface of a planar waveguide. As a result, the vertical limitation of the wave field is carried out by the properties of the initial planar waveguide, and the transverse limitation is due to a jump in the refractive index at the boundary of the etched region. The depth of the grooves can partially overlap the core of the planar waveguide and form a comb type of channel waveguide, or completely intersect its core, thereby forming a strip optical waveguide. Such waveguides are most promising for creating this type of device in silicon-on-insulator and lithium niobate structures. As materials for the manufacture of optical waveguides, polysilicon, a mixture (in the required ratio) of silicon oxide (SiO ₂ ) and titanium oxide (TiO ₂ , Titanium dioxide), chalcogenide glass (As ₂ S ₃ ), aluminum nitride (AlN, Aluminum can be used) nitride), silicon nitride (Si ₃ N ₄ , Silicon nitride), silicon oxynitride (SiON, Silicon oxynitride), gallium nitride (GaN, Gallium nitride), polymers and other materials widely used in photonics and integrated optics.

For convenience in the discussion below, a channel waveguide with corresponding beam dividers through which optical radiation is introduced will be called a forming element, and a channel waveguide with corresponding beam dividers through which optical radiation is output, will be called a filtering element.

Channel waveguides along which optical radiation propagates between the respective beam dividers of different filtering elements will be called connecting waveguides. The term coupled optical waveguides is understood to mean a situation generally accepted in the scientific literature when the energy of an optical wave can flow (partially or completely) between two (or more) optical waveguides due to the tunnel coupling of their optical fields through the space separating them. The term phase-shifting optical elements is understood to mean a situation generally accepted in the scientific literature when the phase of an optical wave propagating through an optical waveguide is controlled by an external signal, for example, due to electro-optical or thermo-optical effects, or the effect of electrostriction, or a change in the concentration of free charge carriers, or under other physical effects (deformation, radiation, etc.).

In addition, based on the present invention, additional technical results can be obtained, which are discussed below.

It is advisable to provide better filtering of optical radiation and reduce the size of the device, in the composition of the beam dividers and / or forming means, to perform curved channel waveguides and / or channel waveguides that change direction due to the effect of reflection from a region with a high reflection coefficient, for example, due to the effect of full internal reflection. This will facilitate their connection with the connecting waveguides and / or the formation of a set of phased beams in a planar waveguide, taking into account the maintenance of the phase difference at the working radiation wavelength substantially multiple of 2π.

It is advisable to provide better filtering of optical radiation in the composition of the formation means to perform beam expanders, which form a set of phased beams in a planar waveguide, taking into account the maintenance of the phase difference at the working radiation wavelength essentially multiple of 2π. The beam expanders can be made in the form of adiabatic horn elements and / or tapering channel waveguides.

It is advisable to reduce the size of the device, the waveguides of the forming and filtering elements run parallel to each other. In this case, the connecting waveguides, as a rule, are made normally (at right angles) to the specified waveguides. Such a filter design will be called rectangular (orthogonal).

It is advisable to increase the steepness of the tuning of the wavelength and / or expansion of the free spectral zone, the waveguides of the forming and filtering elements to perform at an angle to each other. Moreover, the connecting waveguides, as a rule, are made obliquely to the specified waveguides at an angle different from the straight line. This design of the filter will be called inclined.

It is advisable to ensure the simultaneous filtering of several frequency channels (multi-channel filter function - multi channel drop filter) to perform several filter elements in series (one for each frequency channel), with an effective refractive index, and / or the angle of inclination of the filter elements and / or the location of the dividers each beam is made taking into account the individual characteristic only of the filtering element of the working wavelength different from the working wavelength of other filtering elements.

For the through-passage of a broadband optical signal (through-pass function), it is advisable to perform the last filter element along the radiation taking into account the maintenance for most beams of the phase difference essentially multiple of 2π in a wide spectral range, not less than the free spectral zone of the filter.

To adjust the wavelength of the filtered radiation in the immediate vicinity of at least one channel waveguide of the filtering or forming element, it is advisable to perform at least one set of control electrodes in the form of strips of conductive material in order to create, when an electric field is applied, local changes in the refractive index in the vicinity of these waveguides due to electro-optical or thermo-optical effects, or the effect of electrostriction, or changes in the concentration of free carriers over poison.

It is advisable to perform at least one set of control electrodes in the form of strips of conductive material in order to create an electric field when applying an electric field to adjust the wavelength of the filtered radiation at the same time of all optical channels and to expand the range of the tuning of wavelengths of light of an optical filter, in the immediate vicinity of a set of connecting channel optical waveguides local changes in the refractive index in the vicinity of these waveguides due to electro-optical or thermo-optical effects or the effect of electrostriction, or a change in the concentration of free charge carriers, the length of the control electrodes and the magnitude of the voltage applied to them being chosen so as to have the same or different by an even number of π phase shift for adjacent beams branched by different beam dividers and passed from input to output optical filter.

To reduce the number of control electric signals of an optical filter with an extended range of tuning wavelengths of light, it is advisable to carry out multi-section electrode structures in the immediate vicinity of a set of connecting channel optical waveguides. The application of an electric field to electrode structures should provide a linear increase in the phase shift for neighboring beams (taking into account the possible phase shift by an even number π).

To reduce the number of control electric signals of an optical filter with an extended range of tuning wavelengths of light, it is advisable to carry out multi-section electrode structures in the immediate vicinity of a set of connecting channel optical waveguides, with at least two rows of these structures arranged in series along the optical beam. The application of an electric field to the first row of electrode structures with a variable length of electrodes provides a linear increase in phase shift for adjacent beams, and the simultaneous application of an electric field to the second row of electrode structures with a constant length of electrodes provides a linear (taking into account a possible shift by an even number π) phase shift across the entire set of strip waveguides.

In the last two cases, in order to ensure the through passage of the broadband optical signal (the through function), it is necessary to place a similar set of electrode structures in the area between the last and the last but one filtering element, the length of the electrodes and the magnitude of the applied voltage to each of the electrodes being chosen to have zero or a phase shift differing by an even number π for adjacent beams branched by different beam dividers and passed from the input to the output of the optical filter.

To ensure a wide adjustment of the wavelength, it is advisable to perform at least one source of acoustic waves to excite an acoustic wave capable of interacting with the light waves of optical beams propagating through the formation means.

It should be noted that in the case when the optical filter contains a planar waveguide and a source of acoustic waves, the through-passage of the broadband optical signal (through function) is carried out for all wavelengths, except those that were rejected by the acousto-optical wave. Part of the deflected light beams, at given wavelengths, will be filtered using one or more filter elements. However, it is desirable that all remaining (unfiltered) wavelengths be able to pass through the device through. For this purpose, it is advisable to produce an optical filter with at least one additional source of acoustic waves, configured to generate an acoustic wave directed against the acoustic wave of the main source and capable of interacting with the light waves of the beam in the area between the last and penultimate filtering elements.

By optimizing the magnitude of the division coefficient, phase and position of each of the beam dividers of the forming and / or filtering elements, both the shape of the transmission line and its envelope in the spectral range of the filter can be adjusted. For example, to ensure significant suppression (more than 20 dB) of the side lobes, it is advisable to perform beam dividers with different fission factors. This is achieved by changing the magnitude of the tunnel coupling, which is chosen so as to provide the optimal amplitude of the beams and / or phase of the waves branched by various beam dividers. At the same time, at the output of the device, the contribution to the resulting signal from different beams, as a rule, decreases from the middle part of the forming and filtering elements to their ends. It should be noted that the term “apodization” refers to the case generally accepted in the scientific and technical literature when the amplitudes and / or phases of various signal components vary, the summation of which forms the spectral response of the device. Therefore, apodization can be amplitude, phase, or mixed (amplitude-phase). Most interference optical devices use apodization to improve their specifications.

In order to provide different beam division coefficients for the forming and / or filter elements, it is advisable to perform beam dividers with different distances between coupled waveguides and / or different effective coupling lengths.

In order to simplify the manufacturing technology, beam dividers and / or channel waveguides of the forming and / or filtering elements and / or connecting waveguides can be made in one layer.

In order to improve the characteristics of the optical filter, beam dividers and / or channel waveguides of the forming and / or filter elements, and / or connecting waveguides and / or planar waveguides can be made in different layers.

To reduce spurious signals, it is advisable to connect the free ends of the channel waveguides with damping regions that have large optical propagation losses. Such damping regions can be parts of a structure with a high concentration of free charge carriers, and / or containing scattering centers, and / or submicron diffraction gratings, and / or narrowing (wedge-shaped) channel waveguides with a gradually decreasing core cross section (so that the radiation passing through them increases its spatial size (expanded) and left the area of the optical filter).

In order to reduce the size of the filter, beam dividers and / or channel waveguides of the forming and / or filtering elements and / or connecting waveguides are made in silicon-on-insulator structures.

In order to reduce filter size and improve parameters, the device is made in waveguide structures based on lithium niobate.

In order to improve the filter parameters, the channel waveguides forming the beam dividers have the same or close effective refractive indices.

The structural implementation of optical filters may contain many intersections of channel waveguides of the forming and / or filtering elements with connecting waveguides. To reduce spurious signals, it is advisable to intersect the specified channel waveguides with minimal scattering losses, for example, due to the multilayer crossing method using the vertical coupling described in the scientific literature, for example, based on narrowing channel waveguides (see, for example, K.Watanabe, Y. Hashizume, Y. Nasu, Y. Sakamaki, M. Kohtoku, M. Itoh, and Y. Inoue, “Low-loss three-dimensional waveguide crossings using adiabatic interlayer coupling,” Electron. Letters 44, 1356-1357 (2008); R. Sun, M. Beals, A. Pomerene, J. Cheng, Ching-yin Hong, L. Kimerling, and J. Michel, "Impedance matching vertical optical waveguide couplers for dense high index contrast c ircuits, "Opt. Express 16, 11682-11690 (2008); or Andrei V. Tsarev," Efficient silicon wire waveguide crossing with negligible loss and crosstalk, "Opt. Express 19, 13732-13737 (2011)).

The operation of the optical filter is illustrated in the following figures, which have a continuous numbering, further explained in the course of the presentation.

Figure 1 depicts a schematic diagram of a single-channel optical filter on coupled waveguides (orthogonal orientation), according to the invention isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 7-12 are coupled channel optical waveguides of beam dividers of the forming element, 13-18 are coupled channel optical waveguides of beam dividers of the filtering element, 30-32 - curved channel optical waveguides, 36-41 - connecting channel optical waveguides, 60 - tapering channel optical waveguide, 61 - input of the optical filter, 62 - output of the optical filter for the first spectral th channel, 66 - optical filter output for input radiation control, 70 - optical beams branched beam splitters.

Figure 2 depicts a schematic diagram of an optical single-channel filter on coupled waveguides (orthogonal orientation, the case of intersection with connecting waveguides), according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 7-12 are coupled channel optical waveguides of beam dividers of the forming element, 13-18 are coupled channel optical waveguides of beam dividers of the filtering element, 30-32 - curved channel optical waveguides, 36-41 - connecting channel optical waveguides, 60 - tapering channel optical waveguide, 61 - input of the optical filter, 62 - output of the optical filter for the first spectral th channel, 66 - optical filter output for input radiation control, 70 - optical beams branched beam splitters.

Figure 3 depicts a schematic diagram of a single-channel optical filter on coupled waveguides (oblique orientation), according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 7-12 are coupled channel optical waveguides of beam dividers of the forming element, 13-18 are coupled channel optical waveguides of beam dividers of the filtering element, 30-32 - curved channel optical waveguides, 36-41 - connecting channel optical waveguides, 60 - tapering channel optical waveguide, 61 - input of the optical filter, 62 - output of the optical filter for the first spectral th channel, 66 - optical filter output for input radiation control, 70 - optical beams, beam splitters branched, θ - the angle of orientation of the waveguides and forming the filter elements, as measured relative to the normal to the connecting waveguides.

Figure 4 depicts a schematic diagram of a two-channel optical filter on coupled waveguides, according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 and 4 are channel optical waveguides of the filtering elements, 7-12 are coupled channel optical waveguides of the beam dividers of the forming element, 13-18 and 19-24 are coupled channel optical waveguides of the filtering beam elements, 30-32 - curved channel optical waveguides, 36-41 - connecting channel optical waveguides, 60 - tapering channel optical waveguide, 61 - input of the optical filter, 62 and 63 - output of the optical filter for the first and the second spectral channels, 66 — output of the optical filter for monitoring the input radiation, 70 — optical beams branched by beam dividers.

Figure 5 depicts a schematic diagram of a single-channel optical filter-multiplexer on coupled waveguides, according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 and 6 are channel optical waveguides of the filtering elements, 7-12 are coupled channel optical waveguides of the beam dividers of the forming element, 13-18 are coupled channel optical waveguides of the beam dividers of the filtering element, 71- 76 connected channel optical waveguides of beam dividers of the filtering element for the transmission function, 30-32 - curved channel optical waveguides, 36-41 - connecting channel optical waveguides, 60 - tapering channel optical waveguide, 61 — input of the optical filter, 62 — output of the optical filter for the first spectral channel, 65 — output of the optical filter for the transmission function, 66 — output of the optical filter to control input radiation, 70 — optical beams coupled by beam dividers.

6 depicts a schematic diagram of a three-channel optical filter multiplexer on coupled waveguides, according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3-6 are channel optical waveguides of the filtering elements, 7-12 are coupled channel optical waveguides of beam dividers of the forming element, 13-18, 19-23, 54-58 are coupled channel optical waveguides of beam dividers of filter elements of different spectral channels, 71-76 coupled channel optical waveguides of beam dividers of filter element for transmission function, 30-32 - curved channel optical waveguides, 36-41, 44-48, 49-53, 54-58 connecting to anal optical waveguides, 60 — tapering channel optical waveguide, 61 — input of the optical filter, 62–64 — outputs of the optical filter for different spectral channels, 65 — output of the optical filter for transmission function, 66 — output of the optical filter to control input radiation, 67- 69 — optical filter inputs for adding different optical wavelengths to the transmission channel; 70 — optical beams branched by beam dividers.

7 depicts a schematic diagram of a single-channel optical filter on coupled waveguides (orthogonal orientation with thermo-optical control), according to the invention, where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 7-12 are connected channel optical waveguides of beam dividers of the forming element, 13-18 connected channel optical waveguides of beam dividers of the filtering element, 36-41 - connecting channel optical wave s, 60 — tapering channel optical waveguide, 61 — input of the optical filter, 62 — output of the optical filter for the first spectral channel, 66 — output of the optical filter to control the input radiation, 101-106 — electrodes for heating thermo-optical phase-shifting elements of wide tuning, 107 and 108 - a set of electrode structures with a constant length, 109 and 110 - a set of electrode structures with a linearly variable length, 111-112 - electrodes for heating thermo-optical phase-shifting elements of fine tuning, 113 - curved channel optical olnovody due to the effect of total internal reflection.

Fig. 8 is a schematic diagram of a single-channel optical filter with acousto-optical control on coupled waveguides (with adiabatic expansion of horn-type channel waveguides), according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 13-18 are coupled channel optical waveguides of beam dividers of the filtering element, 30-32 are curved channel optical waveguides, 59 is a planar optical waveguide of the forming element , 60 — narrowing channel optical waveguide, 61 — input of the optical filter, 62 — output of the optical filter for the first spectral channel, 76 — optical beams branched by beam dividers, 77 — optical beams ki rejected by SAW, 79 - SAW absorber, 80 - adiabatic horn channel wave expander, 82 - SAW, 83 interdigital transducer for SAW excitation, 84 - high-frequency signal source for SAW excitation.

Fig.9 depicts a schematic diagram of a single-channel optical filter with acousto-optical control on coupled waveguides (two-layer version with adiabatic expansion of the optical beam based on a narrowing channel waveguide), according to the invention, isometry; where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 13-18 are coupled channel optical waveguides of beam dividers of the filtering element, 30-32 are curved channel optical waveguides, 59 is a planar optical waveguide of the forming element , 60 — narrowing channel optical waveguide, 61 — input of the optical filter, 62 — output of the optical filter for the first spectral channel, 76 — optical beams branched by beam dividers, 77 — optical beams ki deflected with a surfactant, 77 — optical beams passing through the filter element, 79 — absorber of a surfactant, 81 — adiabatic optical beam expander based on a narrowing channel waveguide, 82 — surfactant, 83 — interdigital transducer for excitation of a surfactant, 84 — source high-frequency signal for excitation of surfactants, 85 - adiabatic optical beam expander based on a narrowing channel waveguide at the output of the filter element.

Figure 10 depicts a schematic diagram of a simulation of a single-channel optical filter on coupled waveguides (orthogonal orientation with thermo-optical control), where 1 is a solid-state substrate, 2 is a channel optical waveguide of the forming element, 3 is a channel optical waveguide of the filtering element, 7-12 are coupled channel optical waveguides beam dividers of the forming element, 13-18 connected channel optical waveguides of the beam dividers of the filtering element, 30-32 - curved channel optical waveguides, 36-41 - connections relative channel optical waveguides, 61 — input of the optical filter, 62 — output of the optical filter for the first spectral channel, 66 — output of the optical filter to control the input radiation, 70 — optical beams branched by beam dividers, 91–95 — FDTD monitors, 97 — damping region, d — gap gap width, R — radius of curvature of curved waveguides, LL — length of phase-shifting optical elements on the connecting waveguides, A — period of arrangement of the connecting waveguides, ΔT0, ΔT1, ΔT2, ΔT3 - the value of the increase in temperature of the phase-shifting optical elements on connecting waveguides.

11 is a change in the beam division coefficient for deflected (R _c ) and transmitted (T _c ) waves, as a function of the distance d between the strip waveguides based on silicon nanowires (2D FDTD calculation). For comparison, the exponential approximation for R _c (d) is given.

12 is an example of the apodization function A _p and the beam division coefficients R _c and gap gap d between the strip waveguides based on silicon nanowires, as a function of the beam divider number M (calculation 2D FDTD), necessary for its implementation.

Fig. 13 is a numerical demonstration of the operation of an optical filter having 32 beam dividers based on silicon nanowires (2D FDTD calculation), where 2 is a channel optical waveguide of a forming element, 3 is a channel optical waveguide of a filtering element, 61 is an input of an optical filter, 62 is an output optical filter, 91, 91, 95 - FDTD monitors.

Fig - spectral properties of an optical filter having 32 beam divider, based on silicon nanowires, for different combinations of control temperatures of phase-shifting optical elements (calculation 2D FDTD); a) for fine tuning the wavelength; b) for a wide adjustment of the wavelength; to indicate different values of ΔT1, ΔT2 and ΔT3, integer numbering in the form * i *** was used. Here, the first value indicates the value of ΔT, and the subsequent digits indicate the values of ΔT1, ΔT2 and ΔT3 in fractions of ΔT00. In particular, the designation 50i123 corresponds to the case: ΔT = 50 ° C, ΔT1 = 1 × ΔT00, ΔT2 = 2 × ΔT00, and ΔT3 = 3 × ΔT00.

Fig. 15 shows the beam dividing coefficients R as a function of their M number for different filter elements, which are necessary for creating a narrow-band optical filter multiplexer with a 25 GHz frequency grid; here R _c2 , R _c3 , R _{c6 are} the beam dividing coefficients for different waveguides 2, 3-5, and 6, respectively.

Fig. 16 shows the spectral properties of a narrow-band optical filter-multiplexer based on silicon nanowires, for different combinations of control temperatures ΔT of phase-shifting fine-tuning optical elements (calculation by the radiation model); a) for orthogonal orientation; b) for oblique orientation.

Fig. 17 shows beam dividing coefficients R as a function of their number M for different filtering elements (case of an acousto-optical filter with a frequency grid of 25 GHz); here, R _c2 and R _c3 are the beam fission factors for different waveguides 2 and 3, respectively.

Fig - spectral dependence of the filtering efficiency for different values of the frequency of the surfactant (in MHz). AO filter case with a grid of frequencies of 25 GHz and 250 GHz. d = 10 μm, the optical aperture is 0.58 cm and 0.058 cm, respectively. The 25 GHz filter spectra are located inside the 250 GHz filter spectrum.

Fig. 19 is a spectral dependence of filtering efficiency for an AO filter with a 25 GHz frequency grid. d = 10 μm, SAW frequency 1225 MHz, optical aperture 0.58 cm.

Structurally, the device is as follows. On the surface of the solid-state substrate 1 (Figs. 1–9) or in the immediate vicinity below it (the case of the so-called buried waveguide), channel optical waveguides of the forming (2) and filtering (3-6) elements are made. A set of beam dividers is made on the surface of the solid-state substrate 1 (Figs. 1–9) or below it in the immediate vicinity of the channel optical waveguides (2, 3) of the forming and filtering elements, including adjacent channel optical waveguides (7-29, 71-76 ), which due to tunnel coupling are able to branch part of the energy of the incident beam into these waveguides. The rest of the energy propagates further along the original waveguide to the next similar beam splitter until it reaches the end of the structure.

The channel waveguides of the beam dividers have curved parts (30-32, 34, 35), which, on the one hand, serve to smoothly change the coupling coefficient (30-32) to reduce spurious reflection, and, on the other hand, make it possible to optimally arrange all the conclusions of the beam dividers for subsequent connection with the means of formation. Alternatively (see Fig. 7), the direction of the channel waveguide changes due to the effect of specular reflection from a region with a high reflection coefficient, for example, due to the effect of total internal reflection (113) on a vertically etched border. Such elements are described in the literature for creating ring resonators (Doo Gun Kim, Jae Hyuk Shin, Cem Ozturk, Jong Chang Yi, Youngchul Chung, Nadir Dagli, "Rectangular Ring Lasers Based on Total Reflection Mirrors and Three Waveguide Couplers," Photonics Technology Letters, IEEE, vol.19, no.5, pp.306-308, (2007)). In the general case, the forming means comprises a set of curved (34, 35) and direct (36-41, 44-58) channel waveguides, and / or a planar waveguide (59).

In the latter case, a planar waveguide 59 is made on the surface of the solid-state substrate 1 (see Figs. 8 and 9) or beneath it in close proximity to the waveguides of the forming 2 and filtering 3 elements, which is optically matched with the indicated strip waveguides (to reduce stray scattering). For optical matching of the fields of the strip and planar waveguides, a horn expansion of the beam (see Fig. 8) and / or an adiabatically tapering waveguide 81 (see Fig. 9) were made. The latter is usually performed on the upper layer of the multilayer structure and ensures the expansion of the beam and the transition of the optical wave between two waveguide layers (from a channel waveguide to a planar waveguide). This optical element operates on the well-known principle that the optical field of the guided wave (fundamental waveguide mode) adiabatically increases in size as the wave propagates to the narrow end (VR Almeida, RR Panepucci, and M. Lipson, "Nanotaper for compact mode conversion," Opt. Lett. 28, 1302-1304 (2003)).

The channel optical waveguide 2 is an integral part of the forming element, and through its input 61 an optical beam is introduced into the filter, which may contain different wavelengths of the optical spectrum. The input of light into the channel optical waveguide 2 can be carried out in various ways, for example, by docking with a fiber optical fiber, focusing the optical radiation on the end of the structure, using an I / O grating element, etc. Similarly, channel optical waveguides 3-6 are an integral part of the filter element, and through their outputs 62-65 output optical beams that have passed through the forming means.

In the immediate vicinity of each of the channel optical waveguides 2-6, channel waveguides of beam dividers 7-29, 71-76 are made, which are tunnel connected with these waveguides 2-6. The coefficient for dividing the beam energy by coupled waveguides depends on the length of the communication region (usually increases with its increase), on the overlap of the mode fields of each waveguide (usually decreases with increasing distance between the waveguides) and the phase mismatch of their fields (depends on the effective refractive indices of the modes waveguides). Note that in some cases, for example, for technological reasons of manufacturing convenience, the beam splitter may contain more than two waveguides, for example, three channel waveguides (see (Doo Gun Kim, Jae Hyuk Shin, Cem Ozturk, Jong Chang Yi, Youngchul Chung, Nadir Dagli, "Rectangular Ring Lasers Based on Total Reflection Mirrors and Three Waveguide Couplers," Photonics Technology Letters, IEEE, vol.19, no.5, pp.306-308, (2007)). It is important that the coefficient fission varies widely (from 0 to 1) by changing the parameters of the coupling element of the optical waveguides, which performs the function of a beam splitter. 1 to 9, the channel waveguides of the beam dividers 7-29, 71-76 are made at different distances from the waveguides 2-6, which provides the necessary apodization.

To control the working wavelength of the optical filter, phase-shifting optical elements are used that operate on the basis of electro-optical or thermo-optical effects, the effect of electrostriction, or changes in the concentration of free charge carriers. The technology for the manufacture of electrodes and the design of phase-shifting optical elements are described in detail in the scientific and technical literature. In Fig.7 presents embodiments of the device based on the thermo-optical effect. For this purpose, electrode structures 101-112 are made on the surface of the structure, the heating of which, when an electric current flows, causes local changes in the refractive index in the region of the optical waveguide. The electrodes 101-106 are controlled independently of multi-channel direct or alternating current sources (not shown) or are grouped, for example, with a period of 4 (i.e., each 4 is energized in parallel) to reduce the number of channels at the current source. In a variant of the device, groups of electrodes 107-108 and 109-110 are made sequentially along the path of optical radiation, moreover, electrodes 107-108 are made with a constant length of electrodes, and group 109-110 has a linearly varying length from the number of the connecting channel waveguide on which they are made . Heating the electrodes 109-110 gives a linear phase change within each group. Heating of the electrodes 107-108 provides a linear phase change (accurate to a constant bias of 2π) throughout the filter structure. To ensure fine tuning of the wavelength, the electrodes 111 and 112 are made along the optical waveguides forming 2 and filtering 3 elements.

To ensure acousto-optic control on the structures of Figs. 8 and 9, strip 2 and 3, as well as planar 59 waveguides are made. To excite the surface acoustic wave 82, at least one electrode structure of the interdigital transducers 83 is made, which is fed by a high-frequency radiation source 84. To form optical beams at a Bragg angle to the surfactant, adiabatic beam expanders of horn type 80 are made (see Fig. 8) and based on tapering channel waveguides 81 (see Fig.9). To attenuate the reflected wave, an acoustic absorber 79 is made, for example, based on a rubberized compound. To quench spurious and unused radiation, tapered tapered waveguides 60 are made. The device in Fig. 9 is expediently made in two waveguide layers separated by a buffer layer with a low refractive index. Moreover, most of the device is made on the main substrate, and channel waveguides of beam dividers 7-18 are made on the upper layer.

The operation of the optical filter can be characterized as follows. Having arrived at the input of the device 61 (see Fig. 1-9), the light beam sequentially passes through the beam dividers and enters the forming means in the form of a set of beams (see arrows) with a strictly specified amplitude and phase, which are determined by the parameters of the communication elements of the corresponding channel optical waveguides 2-6 with channel waveguides 7-29, 71-76. Part of the radiation passes through the waveguide 2 to its end 66 and can be used to control the level of the input signal (out function). However, for the optimal design of the device, most of the energy of the input beam is transferred to the means of formation in the form of coherent light beams 70, which provide filtering properties of the device.

Cases when the forming means is made only of channel optical waveguides are presented in figures 1-7. First, consider the options for simple single-channel filters (see Figs. 1-3, 7). Here, the optical beams come from the forming element to the connecting channel waveguides 36-41 in the form of coherent light beams 70 and then sequentially pass through the beam dividers of the filter element formed by coupled waveguides 3 and 13-18. On each such element, the optical beam is divided into two parts, one passes further along the curved waveguides 32 to the tapering ends 60, which are used as damping elements (for removing unnecessary radiation from the structure). The other part tunnels into waveguide 3 and sequentially passes further through similar beam dividers. On each of the beam dividers, the optical fields (taking into account their amplitudes and phases) are added to the beam dividers from the side of the corresponding connecting waveguides 36-41, and the fields already entered into waveguide 3 from the previous (along the optical radiation) beam dividers.

According to the invention, the mutual arrangement of the beam dividers is selected taking into account the maintenance of the phase difference at the working radiation wavelength for most beams branched out using various beam dividers, essentially a multiple of 2π. Therefore, at the working wavelength, all such beams will be summed in phase and their amplitudes will increase as they propagate along waveguide 3 (see the increase in the width of the arrows, which illustrate the increase in amplitude). For all other wavelengths, the phase incursion for any pair of beams will be arbitrary (not a multiple of 2π) and the condition of constructive interference will be violated, therefore, the intensity that has passed before the output of the 62 optical wave will be small. Thus, this device will perform the function of optical filtering for those wavelengths for which the above condition of multiplicity 2π is fulfilled.

Note that, structurally, the optical filter can be implemented both in the orthogonal orientation (see FIGS. 1 and 2) and inclined (see FIG. 3). Its distinctive feature is that it allows you to increase the steepness of the wavelength tuning and / or to expand the size of the free spectral zone. Indeed, the path of the waveguides of the forming and filtering elements are located at an angle θ relative to the normal to the connecting waveguides. Then the normal orientation corresponds to the case θ = 0 °, and the inclined orientation corresponds to the case θ ≠ 0 °. Typically, an angle range of 45 ° <θ <65 ° is used for oblique orientation. According to the invention, the condition of constructive interference at the operating wavelength λ _{0 is} described by the expression:

2π / λ ₀ [A × (N ₁ + N ₂ ) / cosθ + L × (N _{i + 1} -N _i ) - 2A × N _i tg (θ)] = 2π m, (one)

where A is the location step of the connecting waveguides; N ₁ and N ₂ are the effective refractive indices of the fundamental mode of the channel waveguides of the forming and filtering elements, respectively; N _{i + 1} and N _i are the effective refractive indices of the fundamental mode of the channel waveguides of two adjacent connecting waveguides (with numbers i + 1 and i); L are the same along the length of the region of the connecting waveguides; m is the interference order, which determines the magnitude of the free spectral zone Δλ of the optical interference filter:

Δλ = λ ₀ / m (2)

Based on expressions (1) - (2), the basic rules are formulated for controlling the wavelength of the proposed type of optical filter, which can be controlled by changing the refractive index of the channel waveguides of the forming and filtering elements and / or connecting waveguides. For clarity, we consider the case when the refractive index of the channel waveguides of the forming and filtering elements changes simultaneously by δN. Then from expression (1) we can find the corresponding change in the working wavelength:

δλ ₀ = λ ₀ δN / (N ₁ -N _i sin (θ)). (3)

It can be seen from this expression that the steepness of wavelength tuning (δλ ₀ / δN) depends on the orientation angle of the waveguides θ, as well as the refractive indices of the connecting waveguides and waveguides of the forming and filtering elements. If the waveguides are the same (N ₁ = N _i ), then the steepness of the tuning increases as 1 / (1-sin (θ)), which for characteristic angles θ of 45 ° and 60 ° gives an increase in the steepness of 3.4 and 7.5 times, respectively, compared with the case of orthogonal orientation (θ = 0 °). If the waveguides are different, with N ₁ <N _i , then a similar effect can be achieved at smaller angles θ, or an increase in the slope of the wavelength tuning will be even more significant. Note that the effect of increasing the steepness is associated with an increase in the location period A of the connecting waveguides and a corresponding increase in the length of the waveguides of the forming and filtering elements, which are used to maintain the same value Δλ, namely:

A = λ ₀ ² / (2Δλ) cos (θ) / (N ₁ -N _i sin (θ)). (four)

The optical filter control described above can be carried out monotonously along the wavelength using phase-shifting elements located along the waveguides of the forming and filtering elements, which we will call fine tuning elements.

From the expressions (1) - (2) it follows that the working filter wavelength by δλ ₀ can be changed if a constant step of changing the indicator ΔN _L = (N _{i + 1} -N _i ) in the connecting channel waveguides is formed.

δλ ₀ = (Δλ / λ ₀ ) ΔN _L L (5)

It is very convenient to change the wavelength in a discrete manner with a step of δλ ₀ = Δλ / p, where p is an integer (usually 4 or 8). In this case, it is necessary to have a change ΔN _L LL = λ ₀ / p, which corresponds to a phase shift of 2π / p in phase-shifting elements of length LL located along the connecting waveguides. Since the phase is determined with an accuracy of 2π, a discrete change in the wavelength of the optical filter with a step Δλ / p can be achieved by a set of p-1 phase-shifting devices arranged in series (with period p) on different channel waveguides in such a way as to create an additional change in the phase 2π / p between any adjacent waveguides. Such phase-shifting elements will be called elements of a wide wavelength tuning. The combination of elements of fine and wide tuning allows you to rebuild the working length of the filter with the limits of the free spectral zone with minimal changes in the refractive index δN and ΔN _L in the corresponding optical waveguides, which is an important advantage of the proposed device. This property will be described in more detail below.

The difference between the single-channel filters in FIGS. 1 and 2 lies in the location of the tapering regions 60, which are used to remove unfiltered radiation from the structure (to reduce spurious signals). Figure 1 is more compact, but in the structure of figure 2, the radiation leaves the filter region, which leads to an additional reduction of spurious interference, however, for this it is necessary to suppress waveguides with low losses and a low level of crosstalk (see, for example, K.Watanabe, Y . Hashizume, Y. Nasu, Y. Sakamaki, M. Kohtoku, M. Itoh, and Y. Inoue, “Low-loss three-dimensional waveguide crossings using adiabatic interlayer coupling,” Electron. Letters 44, 1356-1357 (2008) ; R. Sun, M. Beals, A. Pomerene, J. Cheng, Ching-yin Hong, L. Kimerling, and J. Michel, "Impedance matching vertical optical waveguide couplers for dense high index contrast circuits," Opt. Express 16 , 11682-11690 (2008); or Andrei V. Tsarev, "Efficient silicon wire waveguide crossing with negligible loss and crosstalk, "Opt. Express 19, 13732-13737 (2011)).

The use of several structures similar to that in figure 2 allows you to design filter elements with advanced functionality. Figure 4 presents a General view of a two-channel optical filter, in which two filters from figure 2 are combined in a single device. Their work is similar to that of a single-channel filter, however, here the optical beams leaving the filter are not damped using a tapering waveguide 60 (see Fig. 2), but pass further along the corresponding channel waveguides (see arrows) and are connected tunnelly with the channel waveguide 4 of the second filter element. At its working wavelength (different from the wavelength of the first filtering element), the phase incursion for any pair of beams branched by two beam dividers will be a multiple of 2π, and therefore, all such beams will be summed in phase and increase as they propagate along the waveguide 4 (see the increase in the width of the arrows, which illustrate the increase in amplitude). It is important that the condition of multiplicity 2π described above, which determines the working wavelengths of the first and second filtering elements, depends on the phase delays of the optical microbeams from input 61 to outputs 62 and 63, and therefore, the lengths of the filtered waves can be controlled by the method described above by changing the parameters and the location of the corresponding channel waveguides forming the structure of the optical filter.

It is easy to see that if you change the position of the waveguide 4 (see figure 4) to the position of the waveguide 6 (see figure 5), then this device is transformed into an optical multiplexer. Indeed, it is easy to see from the geometry of FIG. 5 that all the microbeams that pass from the input 61 of the waveguide 2 to the output 65 of the waveguide 6 undergo the same phase shift, which does not depend on the wavelength of the optical radiation. Therefore, they will all be summed in phase and at the output of the device 65 will generate an intense signal, thereby realizing the through function for all wavelengths that were not filtered at the output 62 using a filtering optical element.

The combination of technical solutions presented in figures 4 and 5, allows you to design various types of multi-channel optical multiplexers, which simultaneously implements the function of multi-channel filtering and through passage of all unfiltered wavelengths. An example of such a device is shown in Fig.6. This device implements the function of a three-channel multiplexer. It contains one forming element with an input 61, three filtering elements with outputs 62-64, from which optical radiation can be output at three different wavelengths, and also a filtering element of the transmission function with terminal 65.

The operation of the device is as follows. An optical beam containing a plurality of optical wavelengths enters the input 61 as they propagate along the channel waveguide 2 and branch into the channel waveguides 7-11 due to the tunnel coupling. Each of the branched beams propagates along the corresponding channel waveguides and, in the course of propagation, pass through the coupling elements of various filter elements. According to the invention, the device is made taking into account the maintenance of the phase difference at the working wavelengths of radiation for any pair of beams of each of the filter elements branched out using different beam dividers, essentially a multiple of 2π. In this orthogonal design, any optical beams undergo the same phase shift along the path from the beam dividers of the forming and filtering elements. For example, this can be seen from a comparison of the optical path (shown by arrows) along the following arrangement of channel waveguides: 7-36-44-49-54, 8-37-45-50-55, 9-38-46-51-56 and t .d. In this case, if the requirement of the same delay for any pair of adjacent beam dividers of waveguides 2 and 6, for example, 7-8 and 71-72, is fulfilled, then the phase delay will be the same for all beams from input 61 to output 65, and, therefore, the device implements the passage function. On the other hand, the required phase shift (a multiple of 2π) for filtering is mainly due to the optical delay between the respective beam dividers, for example, 7 and 8, as well as 26 and 25 (from input 61 to output 64). For different filter elements, the required phase shift 2π is achieved at different wavelengths, which is determined by the period of location of the respective beam dividers, as well as the refractive index of the channel waveguides that form the structure of the optical filter. This allows you to organize tunable filtering of optical radiation at the output 62-64 of the corresponding filter elements using controlled phase-shifting optical elements, as well as the passage to the exit of 65 of all remaining (unfiltered wavelengths).

To implement the function of a tunable optical input-output multiplexer (Add / Drop multiplexor (ROADM)), signals at optical wavelengths of the corresponding filter elements can be input to pins 67, 68 and 69. To better suppress the signal that was filtered in the previous step, for example, through terminal 62, the subsequent filter element 4 and / or 5 may have the same operating wavelength. This provides better suppression of a given wavelength in the transmitted signal, which reduces the level of spurious interference for optical signals at the same wavelength that can come from input 68 or 69.

As discussed above, this device can be controlled using phase-shifting optical elements for thin and wide tuning of the wavelength, as illustrated in Fig.7. The first of them are located along the channel optical waveguides forming 2 and filtering 3-6 elements (see Fig.6). They change the refractive index and, therefore, the phase delay on the path between adjacent beam reflectors. As a result, the required phase shift (a multiple of 2π) is achieved at wavelengths different for different channels, which depend on a change in the refractive index in the respective waveguides 2-6. Phase-shifting elements on waveguides 3-5 allow you to independently change the operating wavelength of each of the filter elements, however, the range of wavelength tuning is limited by the smallest possible change in the exponent and phase shift at a small distance (usually of the order of 10 μm) between adjacent beam dividers. For a larger change in the working wavelength, phase-shifting elements of wide tuning on connecting waveguides are used (36-41, 44-48, 49-53, 54-58). Their length can be large enough, which removes the restriction on the magnitude of the phase shift and, therefore, on the range of wavelength tuning. Their task is to create a constant change in the phase shift as the number of the corresponding connecting waveguide increases (36-41, 44-48, 49-53, 54-58). At operating wavelengths, this phase shift is compensated (due to the detuning of the wavelength) by changing the phase delay in the respective waveguides 2-5. Any predetermined working wavelength can be filtered by sharing phase-shifting optical elements for fine and wide tuning of the wavelength.

The case when the channel optical waveguides 2 and 3 are made in the immediate vicinity of the planar optical waveguide 59 are shown in Figs. 8 and 9. Here, the optical beams branched on beam dividers pass into the planar waveguide 59 using adiabatic beam expanders (80 or 81), which the corresponding coupling elements of the beam dividers end. In this case, a set of coherent optical beams 76 are formed in the planar waveguide, which pass through the region along which the acoustic wave 82 propagates, which is generated by the source 83 of acoustic waves. The period of arrangement of the beam dividers is determined by the free zone Δλ according to expression (4). As a source of acoustic waves 83, one or more phased interdigital transducers (IDTs) are usually used, which are a comb of electrodes connected to a high-frequency source 84 of a high-frequency alternating electric field. Source 83, due to the piezoelectric effect, effectively excites a surface acoustic wave, which propagates in the surface region occupied by the optical waveguide 59, and can enter into effective interaction with guided optical waves. For this purpose, the phase fronts of beams formed and guided by adiabatic beam expanders should be located at a Bragg angle θ _B to the phase front of the acoustic wave:

θ _B = arcsin (λ ₀ / (2ΛN)), (6)

where Λ is the SAW wavelength, Λ = v / F, v and F are the SAW speed and frequency, N is the effective refractive index of the guided mode of the planar waveguide.

An acoustic absorber 79 is used to exclude SAW reflection, which can lead to the appearance of spurious signals. At a working wavelength of light, coherent optical beams 76 satisfy the conditions of Bragg phase matching and diffract to the SAW, as a result of which they change the propagation direction to a double Bragg angle and form coherent beams 77. Further, the diffracted optical beams 77 with the help of similar adiabatic beam expanders (80 or 81) get from a planar optical waveguide 59 a filtering optical element, made in the form of a channel optical waveguide 3 and a set of beam dividers on coupled waveguides 13-18, and is output from the device through output 62. The spectral characteristics of the optical filter are optimized by the apodization method by using different beam fission factors determined by their coupling coefficients, for example, by changing the relative position (gap) of the channel waveguides 7-18, 2, and 3 (see Fig. 8 and 9).

The main difference between the devices in Fig. 8 and 9 is the features of their technical implementation. The device in Fig. 8 is proposed to be implemented in a single layer, for example, on waveguides in lithium niobate. First, single-mode planar waveguides are created, for example, due to high-temperature diffusion of titanium, which creates a layer in the surface region (of the order of several microns) with a higher refractive index (by a value of the order of 0.01) with respect to the bulk material. Further, the structure of channel waveguides is implemented by deep etching technology through the entire waveguide layer. In this design, each communication element on the channel waveguides ends on one side with an adiabatically expanding region 80 (horn type) to form a relatively wide (about 10 μm) optical beam 76, and on the other hand with a narrowing waveguide 60 for outputting unused radiation from the device. In this sense, this device is very similar to the device shown in Fig.1. The difference is that in Fig. 1, wide-range wavelength tuning is performed by linearly varying the refractive index in channel waveguides 36-41. In this device of FIG. 8, it is proposed to tune the wavelength due to the rotation of the phase front during acousto-optic interaction in the planar optical waveguide 59.

The structure of FIG. 9 is proposed to be manufactured in two waveguide layers with a high refractive index (for example, Ti: LiNbO3 and SiO2: TiO2) separated by a buffer layer with a lower refractive index (for example, from silicon dioxide). Its advantage is that due to the manufacture of beam dividers on vertically coupled waveguides, it is possible to remove unfiltered radiation from the region of the optical filter, where it can be damped, for example, by using narrowing waveguides 85 with a lower level of spurious scattering (as in FIG. 2) or used for further processing (similar to Figs. 4-6). In the multilayer version, the adiabatic expansion of the beam is carried out on the basis of a tapering channel waveguide 81, which simultaneously implements the optical matching function of the channel and planar waveguides. The principle of operation of such elements has been well studied in the scientific literature (V. R. Almeida, R. R. Panepucci, and M. Lipson, "Nanotaper for compact mode conversion," Opt. Lett. 28, 1302-1304 (2003)). With an adiabatic decrease in the size of the channel waveguide, the field region occupied by its optical mode uniformly increases. As a result, not only the expansion of the optical beam takes place, but also due to the tunnel coupling, the bulk of its energy gradually flows from the channel waveguide to the planar waveguide. A combined version of the device is possible, combining the based elements of Figs. 8 and 9. Here, the aforementioned tapering channel waveguide 81 converts optical radiation into a downstream channel waveguide, in which it expands to the desired size using the adiabatic region 80 (horn type). Its task is to form a slightly diverging optical beam, which propagates at a Bragg angle to the front of the acoustic wave in order to ensure effective acousto-optical interaction.

As mentioned above, the mutual arrangement of the beam dividers of the devices of Figs. 8 and 9 are selected taking into account the maintenance of the phase difference at the working radiation wavelength for any pair of beams 76 or 77, essentially a multiple of 2π. In this case, the phase shift for any beams passing from the input 61 to the output 62 is essentially a multiple of 2π. As a result, at the working wavelength of the light, most of the optical waves will form in phase and the filter will skip the specified wavelength of light. At all other wavelengths of light, the condition of constructive interference is violated, and the signal transmission to output 62 decreases by several orders of magnitude (signal obstruction). These beams (at all other wavelengths) are damped (see Fig. 8) or pass further (see Fig. 9) in the form of beams 78. The latter can leave the filter or be used for further processing, for example, in the case of several similar filters that perform multichannel filtering (similar to figure 4-6). The working wavelength of the filter is tuned by changing the wavelength of the surfactant, which is controlled by the frequency set by the source 84. For different frequencies of the surfactant, the double Bragg angle will change, by which the beams 76 deflect during diffraction by the surfactant, and therefore the phase shift between by diffracted beams 77. As a result, the coherent summation condition (phase shift 2π) will change, and it will be observed for a different wavelength, depending on the frequency of the surfactant.

In the design of the tunable filter shown in FIG. 9, only one filter element based on waveguide 3 and one source 83 of acoustic waves are shown. However, the functionality of the device is significantly expanded if, by analogy with Fig.6, it includes several filter elements in series, and also use additional sources of acoustic waves. In this case, the device acquires the property of a multi-channel narrow-band tunable optical filter.

 It should be emphasized that all of these devices (see Fig.1-Fig.9) use the principle of interference and their operation is completely similar to the operation of optical filters based on multi-reflective elements, which are described in detail in the scientific literature and patents (United States Patent No. 7092139 and No. 6999639) of the author of this invention. All the basic properties of new devices can be obtained from the description of the operation of multi-reflective filters by replacing weakly reflecting mirrors with beam dividers based on coupled waveguides. In the main expressions, the reflection coefficient must be replaced by the beam division coefficient. It is important that all the advantages of multi-deposition filtering elements are fully transferred to the proposed new types of optical filters, but the latter have another important advantage - greater manufacturability. It is connected with the fact that the technology of beam division on coupled waveguides is well developed and used in photonics, and experimental devices based on multi-deposition elements have not yet been created. Although a qualitative description of the operation of new filters on coupled waveguides is clear and understandable, their quantitative description is very complicated and possible only using numerical simulation methods.

To demonstrate the principle of operation and the main properties of the proposed optical filter on coupled waveguides, we conducted a numerical experiment using the finite difference time domain (FDTD) method, which is implemented in the popular FullWave commercial package from RSoft Design Group, Inc. (www.rsoftdesign.com). As a model for detailed analysis, we used the basic structure of the filter shown in figure 1.

A general view of a new filter on coupled strip silicon waveguides (Si wire) is shown in FIG. 10. The input signal comes from input 61 to the right waveguide 2 and splits into micro-beams on a set of adiabatic guided couplers (beam dividers) on waveguides 7-12 with different coupling coefficients, which are responsible for filter apodization. Each of the micro-beams propagates further in structure and are jointly combined into an output waveguide 3 using an appropriate set of similar guided couplers on waveguides 13-18. The structure has a constant difference in optical length. Thus, at the working wavelength, all micro-beams are summed in phase along the output waveguide and effectively filter at output 62. All other wavelengths pass further along the curved parts of the guided couplers and leave the filter structure. To reduce spurious signals, the curved part of each coupler ends with local damping regions with a large imaginary part (equal to 0.4) of the refractive index. It provides a slight reflection back and leads to a complete attenuation of the incoming light wave. For real three-dimensional (3D) structures, at the end of the curved part of the waveguide, a tapering waveguide 60 (see FIGS. 1, 3-5) must be manufactured, which emits (from the structure) the optical beam entering it, or it must be extended to the right ( not shown) channel waveguides, cross it and exit further from the filter (as in figure 2). It can be damped or used for additional processing, say, to build a multiplexer or multichannel filter (as in FIGS. 4-6). For this case, it is preferable to use the intersection of multilayer waveguides with a low level of crosstalk (see, for example, Andrei V. Tsarev, "Efficient silicon wire waveguide crossing with negligible loss and crosstalk," Opt. Express 19, 13732-13737 (2011)).

The parameters of all waveguides correspond to typical silicon waveguides with a width of 450 nm and a height of 250 nm, fabricated on silicon-on-insulator structures. The input and output waveguides have curved parts with a small radius of curvature R, as well as different gap widths d of the gap for different communication elements (see Fig. 10), for the implementation of applause. When designing, we use the Gaussian apodization function A _p (to ensure high suppression of the side lobes) and calculated in advance by the FDTD method the dependence of the beam fission coefficient (R _c ) and the transmission coefficient (T _c ) for optical waves (see Fig. 11) by measuring fundamental mode power (determined by FDTD monitors 91-96, see FIG. 10) for various values of the radius of curvature of the waveguides R and the length of the straight part Lg of the directional coupler. Then we determined the necessary dependences of the beam dividing coefficient (R _c ) and the slit width d on the number of the coupler M (see Fig. 12). A typical distribution of the electromagnetic field at the operating wavelength (Drop) of such a filter is shown in Fig. 13. It can be seen that at this wavelength, all optical micro-beams coming from various couplers add up constructively and create an intense signal in the output waveguide 3, which is measured at the output of the device 62 using the FDTD monitor 95.

As mentioned above, a complete restructuring of the working optical wavelength of the filter can be arranged using phase shifters for fine and wide tuning. The purpose of these phase shifters is to create a controlled linearly variable (from the number of the connecting waveguide) phase delay between the optical micro-beams passing along different paths from the input to the output of the device. In this filter, we carry out filtering due to the thermo-optical effect, namely by changing the temperature in the corresponding sets of waveguides shown in Fig. 10.

Thin adjustment of the device is carried out by increasing the temperature (ΔT) at the input 2 and output 3 waveguides. The results presented in figa show that this filter has a moderate slope of the wavelength tuning Δλ / ΔT = 0.093 nm / ° C, which is typical for other types of silicon optical filters, say, based on ring resonators. For reasonable values of possible temperature changes (<100 ° C), phase-shifting fine-tuning elements (see curves * i000 for ΔT = 0 ° C, 50 ° C, and 100 ° C) can provide only ¼ of the free spectral zone of the filter, which in this case is about 37 nm.

For a wider range of adjustment of the filter wavelength, phase-shifting elements of wide tuning are used, controlled by the scheme of the 2π module (manifold module (2π) scheme). For this scheme, a linearly growing (from the waveguide number) phase shift is formed for all micro-beams, which varies from zero to 2π. We considered this scheme for the case of discrete wavelength switching with an FSR / 4 step. For this, the thermo-optical phase-shifting elements, located along the waveguides 36-41 along the X axis, were organized into 4 periodic groups (see Fig. 10). Each of them has its own value of temperature increase (ΔT0, ΔT1, ΔT2, ΔT3) in channel optical waveguides, which is conveniently determined in units of ΔT00, i.e. the characteristic temperature necessary to create a phase shift π / 2 at the operating wavelength of the device. For compact phase shifters of different lengths LL = 11 μm and LL == 40 μm, the FDTD simulation yields ΔT00 ~ 177 ° C and ΔT00 ~ 48 ° C, respectively. In practice, it is advisable to use longer structures that are controlled by lower temperature values. For example, we obtained ΔT00 ~ 4.8 C ° for the case LL = 400 μm.

The results of calculating the spectral characteristics of the optical filter for different combinations of control temperature values in each of the groups of waveguides 36-41 are shown in Fig.14b. In all calculations, it was assumed that ΔT0 = 0, and to indicate different values of ΔT1, ΔT2, and ΔT3, integer numbering in the form * i *** was used. Here, the first value indicates the value of ΔT, and the subsequent digits indicate the values of ΔT1, ΔT2 and ΔT3 in shares of ΔT00. In particular, the designation 50i123 corresponds to the case: ΔT = 50 ° C, ΔT1 = 1 × ΔT00, ΔT2 = 2 × ΔT00, and ΔT3 = 3 × ΔT00. The data of Fig. 14 show that the proposed device can be used as a widely tunable optical filter with a small level of internal insertion loss (-1 dB) and high side-lobe suppression (below -26 dB). An arbitrary working optical wavelength (within the entire FSR) can be easily tuned using only 4 control signals, namely, 3 for discrete wavelength tuning in increments of FSR / 4 by changing the temperature ΔT1, ΔT2 and ΔT3 of phase-shifting elements on the connecting waveguides 36-41, as well as by fine continuous tuning within the FSR / 4 range by changing the temperature ΔT in the phase-shifting fine tuning elements located along the waveguides 2 and 3 of the forming and filtering elements, respectively. FSR depends on the path-length difference, and for our case (the bond length Lg = 2 μm and the radius of curvature R = 3 μm) is about 37 nm. The total maximum line half-width (FWHM) of an 0.32 mm optical filter using 32 directional couplers with a variable gap d (from 100 nm to 417 nm to ensure filter apodization) was 1.7 nm. In the general case, FWHM depends on the design of the optical filter and the number of beam dividers M _c , and in our case it can be estimated as 0.68 × FSR / Mc. Scaling of these data for large structures shows that a filter with FWHM less than 0.05 nm can be implemented in a device about 1 cm in size. The filter can be tuned within FSR around 37 nm with a moderate temperature change (<100 ° C) in four groups of phase-shifting thermo-optical elements .

To describe devices of such a large size, it is not possible to apply methods of direct numerical simulation of the FDTD type. Therefore, for such cases it is convenient to use the beam model, which correctly takes into account the amplitudes and phases of all micro beams that have passed in different ways from the input to the output of the device, taking into account the division coefficients of amplitudes and phase delay on each of the communication elements. Such algorithms have been successfully used by the patent author to describe multi-reflective filter elements. Examples of spectral characteristics of the proposed optical filters with thermo-optical and acousto-optical control are shown in Fig.15-19. They demonstrate that these filters have properties that are of practical importance, and the filters themselves can be used in data processing and transmission devices, fiber communication and sensor devices.

In particular, FIGS. 15 and 16 describe the expected parameters of a multi-channel tunable filter multiplexer, similar to that shown in FIG. 6, for cases if it were implemented in orthogonal and inclined configurations. To construct a filter with optimal parameters, the beam division coefficient was optimized for different waveguides (see Fig. 15), and it turned out that the dependences of the beam division coefficients R _c2 , R _c3 , R _c6 should be different for different waveguides 2, 3-5 and 6, respectively. The spectral dependences corresponding to such a distribution of the beam fission coefficients are shown in FIG. 16. For orthogonal orientation, wavelength tuning due to phase-shifting elements of a thin substring is carried out in a relatively small range (Fig. 16a), therefore, for tuning in the entire FSR range, it is necessary to additionally apply phase-shifting elements of wide tuning. For oblique orientation with a 60 ° angle, fine tuning elements can tune the filter wavelength over the entire FSR range (see Fig.16b), with a relatively small change in the temperature of the phase-shifting elements on waveguides 2-6.

The properties of an acousto-optical filter, similar to that shown in Fig. 8, illustrate the data of Figs. 17-19. Fig. 17 shows the optimal distribution of the beam dividing coefficient on waveguides 2 and 3 for a narrow-band filter, for operation with a 25 GHz frequency grid. Its spectral characteristics for different SAW frequencies are shown in Fig. 18. It can be seen that different SAW frequencies correspond to different lengths of the filtered waves (see the upper scale on the graph) having different optical frequencies (see the lower scale on the graph). For clarity, the figure shows the parameters of two filters of different sizes, with 10 times different number of beam dividers, 58 and 580, respectively. A detailed form of the transmission line of a narrow-band acousto-optical filter is shown in Fig. 18. It can be seen that the device has a narrow line width, a high level of suppression of the side lobes and small dimensions (less than 1 cm) at the same time. All these results were obtained in the framework of the ray model and spectral approximation (to describe the wave field in the region of a planar waveguide) according to the algorithm that we developed earlier for the description of acousto-optic filters on multi-reflective elements (A.V. Tsarev, E.A. Kolosovsky “Compact narrow-band tunable acousto-optic filter ”, Avtometriya, Volume 42, No. 6, pp. 93-104 (2006)).

It should be noted that all these filters, like most other optical devices, are polarization-dependent, i.e. their characteristics depend on the polarization of the radiation passing through them. For these types of optical elements, polarization diversification methods are used, in which the incoming radiation is separated by polarization, and each polarization is filtered in parallel, with the same polarization-dependent devices. If necessary, the processing elements, in addition to the polarization dividers, also contain polarization conversion elements so that the filters use optical radiation of the same polarization. Modern technologies make it possible to realize the function of polarization diversification in a monolithic design with an optical filter (see, e.g., W. Bogaerts, D. Taillaert, P. Dumon, D. Van Thourhout, and R. Baets, “A polarization-diversity wavelength duplexer circuit in silicon-on-insulator photonic wires, Opt. Express 15, 1567-1578 (2007)).

An important note regarding examples of devices shown in figures 1-9. These figures demonstrate the schematic diagrams of new optical filters and, for clarity, contain a small number of beam dividers on coupled optical waveguides. Devices with practically significant characteristics look similar, but contain a larger number of beam dividers, from several units to several hundred.

Industrial applicability

The proposed tunable optical filter can be used in the design of frequency division multiplexing (DWDM) systems used in fiber-optic communication, as well as to create small tunable tunable optical radiation spectrometers, for example, when creating remote sensing devices - sensors for the composition of gases, liquids and solids, and also as a part of data reading elements from Bragg fiber sensors.

The tunable optical filter can be manufactured using well-known technology developed to create integrated optics and microelectronics devices. As a material for the manufacture of the acousto-optical version of the device, any transparent solid body can be used for which there is a technology for manufacturing channel and planar optical waveguides with low losses (less than about 1 dB / cm) and effective excitation of acoustic waves. Such materials include lithium niobate and tantalate, semiconductor heteroepitaxial structures A _III B _V , layered dielectric structures containing a piezoelectric layer for excitation of surfactants, for example, ZnO / SiO2 / Si, etc. The most promising are devices based on optical waveguides based on lithium niobate, which has good optical, acousto-optical, and electro-optical properties.

As materials for the manufacture of a variant of a device with thermo-optical control, any transparent solid body for which there is a technology for manufacturing channel optical waveguides with low losses can be used. More promising are waveguides with a high refractive index, because channel optical waveguides with a small radius of curvature can be easily implemented in them. Such materials include waveguides based on silicon-on-insulator structures, as well as semiconductor heteroepitaxial structures A _III B _V. Moreover, silicon-based waveguides, the so-called silicon wires, are considered the most promising, because they are the most technologically advanced (manufactured using CMOS-compatible technology), cheap and have high thermo-optical properties. Waveguides based on semiconductor materials are also interesting in that they can realize a high wavelength switching rate due to a change in the concentration of free charge carriers. Devices based on polymer optical waveguides can also be implemented. Their disadvantage is a low refractive index, which makes it difficult to realize small radii of curvature. However, such waveguides are cheap, technologically advanced and there is a wide selection of different materials, including those with very good control properties (high values of thermo-optical or electro-optical coefficients). The specific choice of design and material of the optical filter depends on the technical task. This invention allows to realize different tasks in the most flexible way based on known technologies that have been successfully used in photonics for other types of optical elements, for example, ring resonators and trellis filters at SOI, and / or ring resonators and acousto-optical filters on lithium niobate.

The author is grateful to the RSOFT Design Group. Inc, USA (www.rsoftdesign.com), which provided a user license and technical support for a photon analysis software suite (Rsoft Photonic CAD Suite 8.0, including FullWAVE 6.0 for FDTD computing).

Claims (23)

1. An optical filter containing channel optical waveguides for input-output of optical radiation and propagation of the light beam, light beam dividers arranged in series along the radiation, and forming means for transmitting optical radiation branched by beam dividers, wherein the forming means is made in the form a set of connecting channel optical waveguides and / or a planar optical waveguide, characterized in that the beam dividers are made in the form of a set of coupled optical olnovodov, the mutual arrangement of which is selected taking into account the maintenance at the working wavelength of the radiation, the phase difference is essentially a multiple of 2π for most beams branched out using various beam dividers and passed from the input to the output of the optical filter, where π = 3,14159 ... - universal constant. 2. The optical filter according to claim 1, characterized in that the beam dividers or forming means further comprise curved channel waveguides, or channel waveguides that change direction due to the effect of reflection from a region with a high reflection coefficient. 3. The optical filter according to claim 1 or 2, characterized in that the forming means further comprises beam expanders made in the form of adiabatic horn elements or tapering channel waveguides designed to form a set of phased beams in a planar waveguide while maintaining the phase difference at the working length radiation waves, essentially a multiple of 2π. 4. The optical filter according to claim 1, characterized in that the channel waveguide with corresponding beam dividers through which optical radiation is input is made in the form of a forming element, and the channel waveguide with corresponding beam dividers through which optical radiation is output, is made in at least one filter element. 5. The optical filter according to claim 4, characterized in that the waveguides of the forming and filtering elements are made parallel to each other. 6. The optical filter according to claim 4, characterized in that the waveguides of the forming and filtering elements are made at an angle to each other. 7. The optical filter according to claim 4, characterized in that several filter elements are made sequentially along the radiation path. 8. The optical filter according to claim 4, characterized in that the last filter element along the radiation is made taking into account the maintenance of a difference essentially multiple of 2π in a wide spectral range, not less than the free spectral zone of the filter. 9. The optical filter according to claim 4, characterized in that in the immediate vicinity of at least one channel waveguide of the filtering or forming element, at least one set of control electrodes is made in the form of strips of conductive material to create local changes in the refractive index when an electric field is applied in the vicinity of these waveguides due to electro-optical, or thermo-optical effects, or the effect of electrostriction, or changes in the concentration of free charge carriers. 10. The optical filter according to claim 4, characterized in that in the immediate vicinity of the set of connecting channel optical waveguides, at least one set of control electrodes in the form of strips of conductive material is made to create, when an electric field is applied, local changes in the refractive index in the vicinity of these waveguides due to electro-optical or thermo-optical effects, or the effect of electrostriction, or a change in the concentration of free charge carriers, the length of the control electrodes and the value of the voltage applied to them are selected such as to have the same or differing by an even number of π phase shift for adjacent beams, the branched different beam splitters and transmitted from the input to the output of the optical filter. 11. The optical filter according to claim 4, characterized in that in the immediate vicinity of the set of connecting channel optical waveguides, multi-sectional electrode structures are made, at least two rows of these structures being arranged in series along the optical beam. 12. The optical filter according to claim 4 or 11, characterized in that in the area between the last and the penultimate filter element there is a similar set of electrode structures, the length of the electrodes and the magnitude of the applied voltage to each of the electrodes being chosen to be zero or even the number π is the phase shift for adjacent beams branched by various beam dividers and passed from the input to the output of the optical filter. 13. The optical filter according to claim 4, characterized in that it contains at least one source of acoustic waves to excite an acoustic wave capable of interacting with light waves of optical beams propagating through the forming means. 14. The optical filter according to claims 4, 9 or 13, characterized in that it contains at least one additional source of acoustic waves for exciting an acoustic wave directed against the acoustic wave of the main source and capable of interacting with light waves of optical beams propagating through means of formation between the last and penultimate filter elements. 15. The optical filter according to claim 4, characterized in that the beam dividers are made with different division factors, and the amplitudes of the beams branched with different beam dividers, as a rule, decrease from the middle part of the forming and filtering elements to their ends. 16. The optical filter according to claim 4, characterized in that the beam dividers of the forming or filtering elements are made with different distances between the connected waveguides and / or different effective coupling lengths. 17. The optical filter according to claim 4, characterized in that the forming or filtering elements are made in one layer. 18. The optical filter according to claim 4, characterized in that the forming or filtering elements are made in different layers. 19. The optical filter according to claim 4, characterized in that the free ends of the channel waveguides are connected to damping regions having large optical propagation losses, in the form of structures with a high concentration of free charge carriers, or structures containing scattering centers, or structures containing submicron diffraction gratings, or structures containing tapering (wedge-shaped) channel waveguides with a gradually decreasing core cross section. 20. The optical filter according to claim 4, characterized in that the forming or filtering elements, or connecting waveguides, or a planar waveguide are made in silicon-on-insulator structures. 21. The optical filter according to claim 4, characterized in that the forming, or filtering elements, or connecting waveguides, or a planar waveguide are made in structures based on lithium niobate. 22. The optical filter according to claim 1, characterized in that the waveguides forming the beam dividers have the same or close effective refractive indices. 23. The optical filter according to claim 1, characterized in that the beam dividers contain at least two connected channel optical waveguides.

Images